SLEEP

Martin, Jennifer L; Song, Yeonsu; Hughes, Jaime; Jouldjian, Stella; Dzierzewski, Joseph M.; Fung, Constance H; Rodriguez Tapia, Juan Carlos; Mitchell, Michael N.; Alessi, Cathy A

doi: 10.1093/sleep/zsx079pmid: 28482053

Abstract Study Objective To test the effectiveness of a 4-week behavioral Sleep Intervention Program (SIP: sleep compression, modified stimulus control, and sleep hygiene) compared to a 4-week information-only control (IC) among older adults attending a VA Adult Day Health Care (ADHC) program in a double-blind, randomized, clinical trial. Methods Forty-two individuals (mean age: 77 years, 93% male) enrolled in a VA ADHC program were randomized to receive SIP or IC. All completed in-person sleep and health assessments at baseline, post-treatment and 4-months follow-up that included 3 days/nights of wrist actigraphy, the Pittsburgh Sleep Quality Index (PSQI), and the Insomnia Severity Index (ISI). Mixed repeated measures analysis was used to compare sleep outcomes at post-treatment and 4-months follow-up, with baseline values as covariates. Results SIP participants (n = 21) showed significant improvement on actigraphy sleep efficiency (p = .007), number of nighttime awakenings (p = .016), and minutes awake at night (p = .001) at post-treatment, compared to IC participants (n = 21). Benefits were slightly attenuated but remained significant at 4-month follow-up (all p’s < .05). There were no differences in total sleep time between groups. There was significant improvement on PSQI factor 3 (daily disturbances) at 4-month follow-up (p = .016), but no differences were observed between SIP and IC on other PSQI components or ISI scores at post-treatment or 4-month follow-up. Conclusions A short behavioral sleep intervention may have important benefits in improving objectively measured sleep in older adults participating in ADHC. Future studies are needed to study implementation of this intervention into routine clinical care within ADHC. adult day health care, aging, veterans, sleep, behavioral interventions Statement of Significance This work demonstrates the positive impact of a brief behavioral sleep intervention for older adults participating in adult day health care. Improved sleep may be associated with improvement in other symptoms and ultimately may prolong independence and prevent functional decline. INTRODUCTION Sleep problems are common in older adults, particularly among those with functional limitations, and functional impairments in older adults are associated with poor sleep quality.1,2 Sleep difficulties have been studied extensively in institutional long-term care settings, including nursing homes and other settings;3–5 however, little is known about how best to improve sleep quality among older adults living in the community who require noninstitutional support, such as Adult Day Health Care (ADHC). Older adults with functional limitations are increasingly using ADHC services to maintain independence.6 More than 4600 ADHC centers exist across the United States with a 35% increase since 2002. More than 260,000 individuals and their family caregivers use ADHC services, which can include care planning, assistance with activities of daily living, chronic health condition oversight and management, nursing care, physical, occupational, and speech therapy, meals, transportation, social services, and personal care activities.6,7 ADHC participants are often cognitively impaired, physically disabled, and typically have multiple chronic medical conditions (eg, hypertension, diabetes) and mental health issues (eg, depression). Given that veterans are more medically complex than the population at large,8 their need for ADHC services may be greater than the general population of older adults, and VA considers ADHC one component of its overall plan for noninstitutional long-term care for aging veterans.9 Numerous studies have shown that sleep problems are associated with depression, low quality of life, functional decline, nursing home placement, and mortality among older adults4,10–13 including older adults attending ADHC programs.14 We previously found that over two-thirds of VA ADHC participants have sleep-related complaints, and over one-third meet basic criteria for insomnia disorder.15 Available studies also show that sleep problem is typically not addressed within routine clinical care of older patients,16 and treatment of sleep issues is often limited to medications (eg, hypnotics, sedating antidepressants), which are not recommended for older adults.17 Both untreated insomnia18 and pharmacological treatment of insomnia can be associated with increased risk of falls and other adverse health events among older persons.19 On the other hand, nonpharmacological interventions, such as cognitive-behavioral therapy for insomnia (CBT-I) do not show these adverse effects. CBT-I has been shown to be as effective for older as for younger adults;17 however, studies of older adults have typically been conducted in outpatient settings with participants who do not have significant functional impairments.20–23 It also is not clear whether adapting CBT-I for patients with limited physical abilities will reduce potency. ADHC participants may not be able to adhere to all of the traditional recommendations of CBT-I, such as getting out of bed at night (because of high fall risk) or completing complex sleep diaries (due to visual or cognitive difficulties). Studies have not been done to evaluate whether sleep improvements can be achieved with behavioral interventions delivered within ADHC programs. The goal of this study was to test the effectiveness of a four-session manual-based Sleep Intervention Program (SIP), which was based on CBT-I, adapted for older adults with sleep difficulties who attend an ADHC program. The basic components of the SIP were: individualized sleep education, sleep compression, modified stimulus control, and targeted sleep hygiene recommendations. The intervention was designed to facilitate translation into routine care and application in other similar ADHC programs. The main outcomes were patient-reported sleep quality and objectively measured sleep parameters (based on actigraphy). The aims of the study were to evaluate whether the nonpharmacological SIP, delivered in the context of ADHC, lead to significant improvements in self-reported and objectively measured (by wrist actigraphy) sleep quality, and whether treatment-related improvements were maintained at 4-month follow-up. The main hypothesis was that greater improvements would be shown in the SIP group compared to the information-only control (IC) group in patient-reported sleep quality, insomnia symptoms and in actigraphy-measured total time awake, number of nighttime awakenings, total sleep time, and sleep efficiency from baseline to post-treatment. We also hypothesized that these improvements would be maintained at 4-month follow-up. METHODS Study Design and Participants This 3-year randomized controlled trial was conducted among older adults in an ADHC program at the VA Greater Los Angeles Healthcare System (Clinical Trials Identifier: NCT01259401). All veterans aged 60 years or older who had been enrolled in the ADHC program for at least 1 month were invited to complete a screening questionnaire to assess basic study eligibility (ie, the ability to understand screening items and to communicate verbally during the screening process). Individuals who met these basic criteria and were interested in participating were asked to provide written informed consent. Study data were collected between November 2010 and June 2012. A total of 123 veterans were screened for the study, 72 of whom were enrolled. Among the 51 individuals who were not enrolled, the most common reasons were discharged from the ADHC, significant dementia based on reports from ADHC nursing or social work staff or research staff observation during the screening process suggesting the participant could not provide informed consent and refusal to complete the study screening. The remaining 72 individuals consented to participate, of whom, 30 did not meet eligibility criteria for randomization, 10 due to dementia or low cognitive function (based on Mini-Mental State Examination [MMSE] scores below 20; described below), six due to unstable medical or psychiatric conditions (based on medical record review by the PI and a study physician), five refused, three did not have any sleep difficulties or complaints (based on baseline sleep questionnaires, described below), two were discharged from the ADHC, two had irregular attendance, one withdrew, and one was enrolled in a conflicting clinical trial. A total of 42 individuals were randomized to intervention or control (described below). Figure 1 shows the flow of participants through the study. This study was reviewed and approved by the Institutional Review Board of the VA Greater Los Angeles Healthcare System. Figure 1 Open in new tabDownload slide Study participant recruitment, screening, and enrollment. ADHC = Adult Day Health Care. Figure 1 Open in new tabDownload slide Study participant recruitment, screening, and enrollment. ADHC = Adult Day Health Care. Assessment Procedures Enrolled participants completed a baseline assessment interview, conducted by a trained research staff member. To minimize participant burden, we used brief and abbreviated measures whenever possible. The assessment included administration of self-report sleep questionnaires (ie, Pittsburgh Sleep Quality Index [PSQI]24 and Insomnia Severity Index [ISI])25 and other clinical metrics. These metrics included assessment of depression (Patient Health Questionnaire 9; PHQ-9),26 post-traumatic stress disorder (primary care–post traumatic stress disorder; PC-PTSD),27 cognitive function (MMSE),28 physical function (activities of daily living and instrumental activities of daily living; ADL/IADL),29 fatigue (Flinders Fatigue Scale; FFS),30 and health-related quality of life (Short Form-12 Health Survey [SF-12] physical and mental health component scores).31 Participants also wore a wrist actigraph for 3 days and nights to objectively estimate sleep (Phillips Respironics Actiwatch Spectrum with default settings; 1-minute epochs and medium threshold for sleep scoring). Initially, participants were asked to complete a 12-item sleep diary while wearing the actigraph; however, this was overly burdensome, and completion rates were low. We therefore simplified the sleep diary and asked participants to maintain a four-item sleep diary documenting only bedtime, rise time, sleep quality, and actigraph removal for each recorded day. A review of the patient’s electronic health record was completed to obtain health history information and medications prescribed. Determination of Randomization Eligibility and Randomization Procedure The complete baseline assessment measures were reviewed by the study coordinator and PI (a clinical sleep psychologist), and participants who were excluded from randomization if they had dementia or low cognitive function (based on MMSE < 20 or documentation of moderate to severe dementia in the electronic health records), were medically unstable or exhibited behavioral problems in the ADHC program, did not follow study instructions, withdrew during baseline or refused the intervention, had irregular ADHC attendance or were discharged from the ADHC program prior to randomization, were enrolled in a conflicting VA study, or did not have sleep complaints. In terms of sleep complaints, multiple data sources were included. Participants had to indicate sleep disturbances on screening items (described below) or on baseline questionnaires (reporting poor sleep quality, total sleep time <6 hours per night, sleep efficiency <85%, or have a PSQI score >5 or an ISI score >7) or show evidence of poor sleep quality on actigraphy (sleep efficiency <85%, total sleep time <6 hours per night). Eligible participants were randomly assigned, using random allocation concealment, to receive the manualized SIP or an IC. Participants were randomized in three strata to help insure that SIP and IC groups were balanced in baseline severity of sleep problems. The strata were based on the number of sleep disturbance items endorsed during preconsent screening (0/1, 2, or 3 items). The three items were based on the PSQI and assessed: (1) taking more than 30 minutes to fall asleep, (2) sleeping less than 6 hours a night, and (3) fairly bad or very bad self-rated sleep quality. A total of 42 participants were randomized to the SIP (N = 21) and IC (N = 21) groups. Intervention The SIP and IC treatments were provided by Master’s-level trained Health Educators (HEs) under the supervision of the PI (JLM), a licensed clinical psychologist who is certified by the American Board of Sleep Medicine as a Behavioral Sleep Medicine Specialist. Staff members involved in outcome assessment were blinded to group assignment. The HE could not be blinded to treatment condition, so multiple steps were taken to insure that the fidelity of both the SIP and IC were maintained. First, interventionists were carefully trained to avoid overlap between the SIP and IC conditions. Second, structured patient materials were provided to guide each session and reduce risk of contamination of the control condition with content from the active treatment. Third, interventionists maintained checklists and notes during each session to document use of the study materials in that particular session. Fourth, intervention fidelity was monitored throughout the study via direct observation by the PI and ongoing feedback during weekly supervision. We were not able to obtain permission to record intervention sessions within the ADHC program. Table 1 outlines the content of each session of the SIP. The SIP involved four weekly sessions, each lasting approximately 45 minutes with the HE. Sessions focused on: (1) individualized education about sleep, (2) sleep compression therapy, (3) targeted sleep hygiene education, (4) modified stimulus control, and (5) maintenance of sleep improvements over time, and (6) coping with future bouts of insomnia. The SIP approach was based on CBT-I, which is an empirically supported treatment for insomnia disorder.32 The most significant modifications to traditional CBT-I included (1) substitution of sleep compression33 in place of sleep restriction therapy and (2) modifications of standard stimulus control instructions (eg, no instruction to get out of bed at night due to high fall risk among ADHC participants).33 To implement sleep compression (rather than sleep restriction), the initial time in bed window was set to equal the number of hours the patient was spending in bed (rather than the number of hours of sleep). That was then gradually reduced by 15–30 minutes per week until sleep quality improved without increasing daytime sleepiness. Stimulus control principles were followed and informed recommendations to move all nonsleep activities out of the bed (eg, read or watch TV in another part of the house) before bedtime and after morning rise times. Participants were not specifically instructed to get out of bed during the night if they had awakenings with difficulty returning to sleep; however, they were encouraged to do other relaxing activities (eg, listen to music, read) in or near the bed if they had trouble sleeping during the night. Because of high fall risk and mobility limitations in many ADHC participants, they were not instructed to get out of bed and go to another room if they had trouble sleeping during the night. In addition, this program was similar to CBT-I in terms of the number of sessions, but each session was shorter than many CBT-I intervention programs, which are typically 4–8 sessions lasting 60 minutes per session.20,22,23 Some studies of brief interventions are also effective with older patients.21,34 Consistent with traditional CBT-I interventions, all recommendations were tailored to the individual patient, addressing their specific circumstances, sleep patterns, and level of motivation to engage in the intervention. At the conclusion of each session, a written outline of what was discussed was provided to the participant with the specific, individualized recommendations written down in clear language. Participants maintained the simple four-item daily sleep diary throughout the intervention period. Table 1 The Four-Session Sleep Intervention Program: Session by Session Content and At-Home Activities. Session and topics covered . At-home Activities . Session 1: How sleep works Program overview Education: Healthy sleep, sleep stages, circadian clock (rationale for regular schedule) Stimulus control: limiting nonsleep activities in the bed (note: no instruction to get out of bed at night given due to fall risk) Education: Daily-light exposure and regular daytime activities; impact of daytime napping on nighttime sleep quality Sleep compression: Initial sleep schedule (bedtime and rise time) established based on preferred routine and current time in bed Instruction in use of daily sleep diary Summary and review Follow sleep schedule Move nonsleep activities out of bed Complete sleep diary Session 2: Steps to getting sleep Session overview Check in and review sleep diary Education: sleep drive (rationale for sleep compression) Sleep compression: Reduce time in bed by 30 minutes and adjust sleep schedule accordingly, avoid naps or limit to <30 minutes before 5 pm Behavioral activation: Identify engaging activities for the evening hours to maintain wakefulness until scheduled bedtime 1–2 individualized sleep-hygiene related recommendations Sleep compression: Revised sleep schedule (if indicated) Reminder to complete the daily sleep dairy Summary and review Follow sleep schedule Keep nonsleep activities out of bed Implement 1–2 sleep hygiene practices Complete sleep diary Session 3: Healthy habits for healthy sleep Session overview Check in and review sleep diary Sleep compression: Reduce time in bed by 15–30 minutes and adjust sleep schedule accordingly, Bedtime routine: Identify activities for the last hour before bedtime (engaging but not activating). Sleep hygiene principles: physical activity, light exposure, social activities during the day. Sleep hygiene principles: diet, liquids near bedtime, alcohol, smoking, sleep environment Sleep compression: Revision of planned sleep schedule (if indicated) Summary and review Follow sleep schedule Keep nonsleep activities out of bed Follow sleep-hygiene practices Complete sleep diary Session 4: Preventing the return of chronic insomnia Session overview Check in and review sleep diary Long-term strategies If sleeping well, discuss regular, but less rigid sleep schedule If not sleeping well, discuss continued sleep compression Discuss how to get sleep back on track after health or life events Education: When to contact a health care provider about sleep Sleep compression: Revision of planned sleep schedule (if indicated) to align with long-term sleep plan Summary and review Recommend following sleep schedule (with some flexibility) Recommend keeping nonsleep activities out of bed Recommend continuing to follow good sleep-hygiene practices Complete sleep diary Recommend continuing to follow sleep hygiene–related recommendations Session and topics covered . At-home Activities . Session 1: How sleep works Program overview Education: Healthy sleep, sleep stages, circadian clock (rationale for regular schedule) Stimulus control: limiting nonsleep activities in the bed (note: no instruction to get out of bed at night given due to fall risk) Education: Daily-light exposure and regular daytime activities; impact of daytime napping on nighttime sleep quality Sleep compression: Initial sleep schedule (bedtime and rise time) established based on preferred routine and current time in bed Instruction in use of daily sleep diary Summary and review Follow sleep schedule Move nonsleep activities out of bed Complete sleep diary Session 2: Steps to getting sleep Session overview Check in and review sleep diary Education: sleep drive (rationale for sleep compression) Sleep compression: Reduce time in bed by 30 minutes and adjust sleep schedule accordingly, avoid naps or limit to <30 minutes before 5 pm Behavioral activation: Identify engaging activities for the evening hours to maintain wakefulness until scheduled bedtime 1–2 individualized sleep-hygiene related recommendations Sleep compression: Revised sleep schedule (if indicated) Reminder to complete the daily sleep dairy Summary and review Follow sleep schedule Keep nonsleep activities out of bed Implement 1–2 sleep hygiene practices Complete sleep diary Session 3: Healthy habits for healthy sleep Session overview Check in and review sleep diary Sleep compression: Reduce time in bed by 15–30 minutes and adjust sleep schedule accordingly, Bedtime routine: Identify activities for the last hour before bedtime (engaging but not activating). Sleep hygiene principles: physical activity, light exposure, social activities during the day. Sleep hygiene principles: diet, liquids near bedtime, alcohol, smoking, sleep environment Sleep compression: Revision of planned sleep schedule (if indicated) Summary and review Follow sleep schedule Keep nonsleep activities out of bed Follow sleep-hygiene practices Complete sleep diary Session 4: Preventing the return of chronic insomnia Session overview Check in and review sleep diary Long-term strategies If sleeping well, discuss regular, but less rigid sleep schedule If not sleeping well, discuss continued sleep compression Discuss how to get sleep back on track after health or life events Education: When to contact a health care provider about sleep Sleep compression: Revision of planned sleep schedule (if indicated) to align with long-term sleep plan Summary and review Recommend following sleep schedule (with some flexibility) Recommend keeping nonsleep activities out of bed Recommend continuing to follow good sleep-hygiene practices Complete sleep diary Recommend continuing to follow sleep hygiene–related recommendations Open in new tab Table 1 The Four-Session Sleep Intervention Program: Session by Session Content and At-Home Activities. Session and topics covered . At-home Activities . Session 1: How sleep works Program overview Education: Healthy sleep, sleep stages, circadian clock (rationale for regular schedule) Stimulus control: limiting nonsleep activities in the bed (note: no instruction to get out of bed at night given due to fall risk) Education: Daily-light exposure and regular daytime activities; impact of daytime napping on nighttime sleep quality Sleep compression: Initial sleep schedule (bedtime and rise time) established based on preferred routine and current time in bed Instruction in use of daily sleep diary Summary and review Follow sleep schedule Move nonsleep activities out of bed Complete sleep diary Session 2: Steps to getting sleep Session overview Check in and review sleep diary Education: sleep drive (rationale for sleep compression) Sleep compression: Reduce time in bed by 30 minutes and adjust sleep schedule accordingly, avoid naps or limit to <30 minutes before 5 pm Behavioral activation: Identify engaging activities for the evening hours to maintain wakefulness until scheduled bedtime 1–2 individualized sleep-hygiene related recommendations Sleep compression: Revised sleep schedule (if indicated) Reminder to complete the daily sleep dairy Summary and review Follow sleep schedule Keep nonsleep activities out of bed Implement 1–2 sleep hygiene practices Complete sleep diary Session 3: Healthy habits for healthy sleep Session overview Check in and review sleep diary Sleep compression: Reduce time in bed by 15–30 minutes and adjust sleep schedule accordingly, Bedtime routine: Identify activities for the last hour before bedtime (engaging but not activating). Sleep hygiene principles: physical activity, light exposure, social activities during the day. Sleep hygiene principles: diet, liquids near bedtime, alcohol, smoking, sleep environment Sleep compression: Revision of planned sleep schedule (if indicated) Summary and review Follow sleep schedule Keep nonsleep activities out of bed Follow sleep-hygiene practices Complete sleep diary Session 4: Preventing the return of chronic insomnia Session overview Check in and review sleep diary Long-term strategies If sleeping well, discuss regular, but less rigid sleep schedule If not sleeping well, discuss continued sleep compression Discuss how to get sleep back on track after health or life events Education: When to contact a health care provider about sleep Sleep compression: Revision of planned sleep schedule (if indicated) to align with long-term sleep plan Summary and review Recommend following sleep schedule (with some flexibility) Recommend keeping nonsleep activities out of bed Recommend continuing to follow good sleep-hygiene practices Complete sleep diary Recommend continuing to follow sleep hygiene–related recommendations Session and topics covered . At-home Activities . Session 1: How sleep works Program overview Education: Healthy sleep, sleep stages, circadian clock (rationale for regular schedule) Stimulus control: limiting nonsleep activities in the bed (note: no instruction to get out of bed at night given due to fall risk) Education: Daily-light exposure and regular daytime activities; impact of daytime napping on nighttime sleep quality Sleep compression: Initial sleep schedule (bedtime and rise time) established based on preferred routine and current time in bed Instruction in use of daily sleep diary Summary and review Follow sleep schedule Move nonsleep activities out of bed Complete sleep diary Session 2: Steps to getting sleep Session overview Check in and review sleep diary Education: sleep drive (rationale for sleep compression) Sleep compression: Reduce time in bed by 30 minutes and adjust sleep schedule accordingly, avoid naps or limit to <30 minutes before 5 pm Behavioral activation: Identify engaging activities for the evening hours to maintain wakefulness until scheduled bedtime 1–2 individualized sleep-hygiene related recommendations Sleep compression: Revised sleep schedule (if indicated) Reminder to complete the daily sleep dairy Summary and review Follow sleep schedule Keep nonsleep activities out of bed Implement 1–2 sleep hygiene practices Complete sleep diary Session 3: Healthy habits for healthy sleep Session overview Check in and review sleep diary Sleep compression: Reduce time in bed by 15–30 minutes and adjust sleep schedule accordingly, Bedtime routine: Identify activities for the last hour before bedtime (engaging but not activating). Sleep hygiene principles: physical activity, light exposure, social activities during the day. Sleep hygiene principles: diet, liquids near bedtime, alcohol, smoking, sleep environment Sleep compression: Revision of planned sleep schedule (if indicated) Summary and review Follow sleep schedule Keep nonsleep activities out of bed Follow sleep-hygiene practices Complete sleep diary Session 4: Preventing the return of chronic insomnia Session overview Check in and review sleep diary Long-term strategies If sleeping well, discuss regular, but less rigid sleep schedule If not sleeping well, discuss continued sleep compression Discuss how to get sleep back on track after health or life events Education: When to contact a health care provider about sleep Sleep compression: Revision of planned sleep schedule (if indicated) to align with long-term sleep plan Summary and review Recommend following sleep schedule (with some flexibility) Recommend keeping nonsleep activities out of bed Recommend continuing to follow good sleep-hygiene practices Complete sleep diary Recommend continuing to follow sleep hygiene–related recommendations Open in new tab The IC group was structured to closely resemble the “observable” aspects of the SIP to enhance credibility and to assist with blinding of the research assessment staff members and ADHC providers. The IC involved four meetings with the HE, lasting up to 45 minutes each. During these meetings, two educational brochures (published by the American Academy of Sleep Medicine, Darien, IL) were reviewed and discussed. One brochure focused on changes in sleep with age, and the other focused on sleep hygiene education. Measures Sleep Outcomes Primary outcomes for the study included patient-reported and objective sleep measures (actigraphy). Sleep outcomes were assessed at each time point (ie, baseline, post-treatment, and 4-month follow-up). Patient-reported sleep quality was assessed with the PSQI (total and three-factor subscale scores)35 and the ISI total score. The PSQI is a 19-item questionnaire that assesses self-reported sleep quality and disturbances over the past month. The items include hours of sleep, ratings for frequency of sleep concerns, general sleep quality, and daytime factors related to poor sleep. A total score greater than 5 indicates poor quality sleep.18 We modified the PSQI by asking about participants’ sleep over the past week (rather than the past month) because the duration of the intervention itself was fairly short (ie, about 4 weeks). We also elected to use a three-factor scoring system, which has superior psychometric properties compared to the originally developed seven-factor PSQI scoring system.35 In addition to the single total score, the three dimensional assessments (sleep efficiency, perceived sleep quality, and daily disturbances) from the PSQI were used to obtain more nuanced information regarding the nature of sleep problems. We used a total of four main outcome variables from the PSQI: (1) PSQI total score, (2) PSQI Factor 1 (sleep efficiency), (3) PSQI Factor 2 (perceived sleep quality), and (4) PSQI Factor 3 (daily disturbances).35 The ISI was used to measure severity of insomnia symptoms over the past week.19,20 The ISI is a seven-item questionnaire assessing the nature, severity, and impact of insomnia. A five-point Likert-like scale is used to rate each item, ranging from 0 (no problem) to 4 (very severe problem). Total scores range from 0 to 28 and are interpreted as follows: no insomnia (0–7), subthreshold insomnia (8–14), moderate insomnia (15–21), and severe insomnia (22–28). Objective sleep was measured by actigraphy (Actiwatch Spectrum, Philips Respironics) on the dominant wrist for three consecutive days (ie, 72 hours). Actigraphs are small, watch-sized devices useful in longitudinal, naturalistic (ie, not in a sleep laboratory) assessment of sleep-wake patterns.36 Actigraphy devices used for estimating sleep parameters contain subminiature solid-state accelerometers, and in general, wrist activity below an established threshold is interpreted as sleep, whereas high-wrist activity is interpreted as wakefulness, using mathematical algorithms within commercially available software accompanying the devices used. In this study, sleep was scoring using medium threshold settings and default parameters for sleep scoring based on 1-minute epochs. Participants were also asked to complete a sleep diary to record bedtime and rise time for each night while they were wearing the actigraph. Information from this diary was used to identify the “in bed” period for actigraphy scoring. Four outcomes from the actigraphy were used for our main analyses: nighttime sleep efficiency (sleep time divided by total time in bed), total sleep time, number of nighttime awakenings, and total nighttime wake time. Other Measures Participant demographic information was collected at baseline and included age, gender, race/ethnicity, marital status, living arrangement, employment status, level of education, and duration of ADHC enrollment. Comorbidity was assessed by the number of medical conditions on the problem list within the participants’ electronic health records that were defined by the International Classification of Diseases, Ninth Revision (ICD-9).37 We also collected health information at each of the three time points. Information on medication use was collected using electronic health record review and a structured interview with the participant based on the patient’s medication list in the medical record. Pain was measured using one subscale (seven items of pain intensity) from the Geriatric Pain Measure.38 PTSD was screened using the four-item Primary Care PTSD (PC-PTSD).39 The PC-PTSD is considered positive if an individual answers yes to three or more items. Cognitive function was assessed using the MMSE.28 This 19-item scale assesses orientation, registration, attention/calculation, recall, language, and construction and has standard instructions. MMSE scores range from 0 to 30, and higher scores indicate better functioning. A score below 24 is consistent with at least mild cognitive impairment; scores below 20 are consistent with moderate-to-severe cognitive impairment. Physical function was assessed using components of the Older Americans Resources and Services (OARS) multidimensional functional assessment questionnaire.29 These components were comprised of seven items of ADL and seven items on IADL. ADLs assessed included eating, dressing, grooming, walking, getting in and out of bed, taking a bath or shower, and continence. IADLs included telephone use, going places beyond walking distance, shopping, preparing meals, doing housework, handling money, and taking medications. Each item is scored as 0 = completely dependent, 1 = can do with some help, or 2 = completely independent. Total scores range from 0 to 28 with higher scores indicating greater independence. Depression was assessed using the PHQ-9.40 The PHQ-9 is the nine-item depression module from the PHQ (a self-administered diagnostic instrument for common mental disorders). The PHQ-9 total score ranges from 0 to 27 with each item ranging from 0 (not at all) to 3 (nearly every day). Self-rated health data were obtained from the SF-12.31 The 12 items are divided into eight subscales: (1) physical functioning, (2) role limitations due to physical problems and (3) emotional problems, (4) general health perceptions, (5) vitality, (6) social functioning, (7) general mental health, and (8) bodily pain. It also produces two summary component scores: physical and mental health component summary score, with scores ranging from 0 to 100. The FFS was used to screen for the presence and severity of daytime fatigue associated with insomnia over the past 7 days.30 It is a seven-item questionnaire in which six items are presented in Likert-like format with responses ranging from 0 (not at all) to 4 (extremely) and one item presented as the sum of all times of day when fatigue is experienced. Total scores range from 0 to 31. Finally, we examined patient-reported total sleep time and sleep efficiency using items within the PSQI. Sleep efficiency was computed by dividing total sleep time by time in bed (ie, time from reported bedtime to reported rise time), converted to a percent. Statistical Analysis All analyses were performed using Stata version 13.1, and the mixed repeated measures analysis was performed using the Stata mixed command. In preparation for our main analysis, we examined scatterplots of each outcome at baseline versus post-treatment and baseline versus 4-month follow-up to identify potentially influential observations for sensitivity analyses. A two (group) by three (time) mixed repeated measures analysis was performed using an unstructured residual covariance to account for the covariance of the residuals across the three time points. There were no missing data at the post-treatment time point. All available data were analyzed at the 4-month follow-up, following intention-to-treat principles; that is, all participants’ data were included regardless of treatment completion. We computed the marginal means for each outcome as a function of group membership (SIP and control) and time (baseline, post-treatment, and 4-month follow-up) as obtained from the mixed repeated measures analysis. The treatment effect at post-treatment was assessed by computing an interaction contrast that compared the change (post-treatment vs. baseline) for the treatment versus control groups. Likewise, the treatment effect at 4 months was estimated via interaction contrasts that compared the change (4-month follow-up vs. baseline) for the treatment versus control groups. Based on identification of six potentially influential observations, we conducted sensitivity analyses by repeating our original analysis omitting observations that appeared influential. RESULTS Baseline Patient Characteristics Table 2 shows demographic and health characteristics of the overall sample and the two experimental conditions. Participants included 39 men and three women with a mean (standard deviation [SD]) age of 77.1 (9.9) years (71.4% white). Participants typically had multiple comorbid conditions, high levels of pain, and poor health. Sleep characteristics are shown in Table 3. Sleep characteristics are shown in Table 3. Self-reported sleep quality was poor as measured by the PSQI total score (mean = 6.8; SD = 4.2) and subthreshold insomnia on the ISI total score (mean = 8.5; SD = 6.6). However, objective (actigraphy-assessed) sleep efficiency was relative high, with a mean of 84.9% (SD = 7.9%) but with multiple nighttime awakenings (mean = 23.6; SD = 7.9). As expected due to randomization, there were no significant differences between the SIP and IC groups in terms of demographic variables, with the exception of current living arrangement, and there were no significant differences between either objective or patient-reported sleep quality between the two groups at baseline. Table 2 Demographic and Health Characteristics of Randomized Participants and Differences Between Treatment Groups at Baseline. Variable . Overall (N = 42) . SIP (N = 21) . IC (N = 21) . p-value . Age in years [M (SD)] 77.1 (9.9) 77.7 (10.2) 76.4 (9.9) .680 Gender [n (%) male] 39 (92.9%) 18 (85.7%) 21 (100%) .232 Race [n (%) non-Hispanic white] 30 (71.4%) 13 (61.9%) 17 (81%) .306 Years of education [M (SD)] 14.5 (2.5) 14.7 (2.9) 14.2 (2.0) .463 Marital status [n (%) married] 21 (50%) 10 (47.6%) 11 (52.4%) 1.00 Employment status .509  Unable to work [n (%)] 6 (14.3%) 4 (19%) 2 (9.5%)  Volunteer [n (%)] 4 (9.5%) 1 (4.8%) 3 (14.3%)  Retired [n (%)] 32 (76.2%) 16 (76.2%) 16 (76.2%) Current living arrangement .012  Own home [n (%)] 26 (63.4%) 12 (57%) 14 (70%)  Relative or friend’s home [n (%)] 4 (9.8%) 0 (0%) 4 (20%)  Board and care home/ assisted living facility [n (%)] 11 (26.8%) 9 (43%) 2 (10%) Years since ADHC enrollment [M (SD)] 2.0 (2.7) 2.3 (2.9) 1.7 (2.6) .473 Number of diagnosed conditions in the electronic health record [M (SD)] 24.3 (15.9) 23.8 (16.9) 24.8 (15.2) .841 Geriatric Pain Measure Score [M (SD)] 16.9 (12.4) 16.1 (12.0) 17.7 (13.1) .674 Primary Care PTSD score [n (%) with score ≥3] 5 (13.2%) 3 (15.0%) 2 (11.1%) .723 MMSE score [M (SD)] 25.9 (2.8) 26.0 (2.7) 25.8 (2.9) .870 OARS ADL/IAD total score [M (SD)] 19.7 (4.2) 19.3 (3.1) 20.1 (5.2) .547 PHQ-9 score [M (SD)] 6.3 (5.5) 5.4 (4.1) 7.2 (6.6) .298 SF-12 Physical health component score [M (SD)] 36.3 (8.5) 35.1 (6.1) 37.6 (10.4) .353 SF-12 Mental health component score [M (SD)] 50.1 (11.9) 53.3 (10.3) 46.9 (12.7) .081 Flinders Fatigue scale score [M (SD)] 8.7 (7.9) 8.0 (6.9) 9.5 (9.0) .555 Variable . Overall (N = 42) . SIP (N = 21) . IC (N = 21) . p-value . Age in years [M (SD)] 77.1 (9.9) 77.7 (10.2) 76.4 (9.9) .680 Gender [n (%) male] 39 (92.9%) 18 (85.7%) 21 (100%) .232 Race [n (%) non-Hispanic white] 30 (71.4%) 13 (61.9%) 17 (81%) .306 Years of education [M (SD)] 14.5 (2.5) 14.7 (2.9) 14.2 (2.0) .463 Marital status [n (%) married] 21 (50%) 10 (47.6%) 11 (52.4%) 1.00 Employment status .509  Unable to work [n (%)] 6 (14.3%) 4 (19%) 2 (9.5%)  Volunteer [n (%)] 4 (9.5%) 1 (4.8%) 3 (14.3%)  Retired [n (%)] 32 (76.2%) 16 (76.2%) 16 (76.2%) Current living arrangement .012  Own home [n (%)] 26 (63.4%) 12 (57%) 14 (70%)  Relative or friend’s home [n (%)] 4 (9.8%) 0 (0%) 4 (20%)  Board and care home/ assisted living facility [n (%)] 11 (26.8%) 9 (43%) 2 (10%) Years since ADHC enrollment [M (SD)] 2.0 (2.7) 2.3 (2.9) 1.7 (2.6) .473 Number of diagnosed conditions in the electronic health record [M (SD)] 24.3 (15.9) 23.8 (16.9) 24.8 (15.2) .841 Geriatric Pain Measure Score [M (SD)] 16.9 (12.4) 16.1 (12.0) 17.7 (13.1) .674 Primary Care PTSD score [n (%) with score ≥3] 5 (13.2%) 3 (15.0%) 2 (11.1%) .723 MMSE score [M (SD)] 25.9 (2.8) 26.0 (2.7) 25.8 (2.9) .870 OARS ADL/IAD total score [M (SD)] 19.7 (4.2) 19.3 (3.1) 20.1 (5.2) .547 PHQ-9 score [M (SD)] 6.3 (5.5) 5.4 (4.1) 7.2 (6.6) .298 SF-12 Physical health component score [M (SD)] 36.3 (8.5) 35.1 (6.1) 37.6 (10.4) .353 SF-12 Mental health component score [M (SD)] 50.1 (11.9) 53.3 (10.3) 46.9 (12.7) .081 Flinders Fatigue scale score [M (SD)] 8.7 (7.9) 8.0 (6.9) 9.5 (9.0) .555 M = mean; SD = standard deviation; ADHC = Adult Day Health Care; MMSE = Mini-Mental State Examination; PTSD = post-traumatic stress disorder; OARS = the Older Americans Resources and Services; ADL = activities of daily living; IADL = instrumental activities of daily living; SF-12 = Short-Form v12. Open in new tab Table 2 Demographic and Health Characteristics of Randomized Participants and Differences Between Treatment Groups at Baseline. Variable . Overall (N = 42) . SIP (N = 21) . IC (N = 21) . p-value . Age in years [M (SD)] 77.1 (9.9) 77.7 (10.2) 76.4 (9.9) .680 Gender [n (%) male] 39 (92.9%) 18 (85.7%) 21 (100%) .232 Race [n (%) non-Hispanic white] 30 (71.4%) 13 (61.9%) 17 (81%) .306 Years of education [M (SD)] 14.5 (2.5) 14.7 (2.9) 14.2 (2.0) .463 Marital status [n (%) married] 21 (50%) 10 (47.6%) 11 (52.4%) 1.00 Employment status .509  Unable to work [n (%)] 6 (14.3%) 4 (19%) 2 (9.5%)  Volunteer [n (%)] 4 (9.5%) 1 (4.8%) 3 (14.3%)  Retired [n (%)] 32 (76.2%) 16 (76.2%) 16 (76.2%) Current living arrangement .012  Own home [n (%)] 26 (63.4%) 12 (57%) 14 (70%)  Relative or friend’s home [n (%)] 4 (9.8%) 0 (0%) 4 (20%)  Board and care home/ assisted living facility [n (%)] 11 (26.8%) 9 (43%) 2 (10%) Years since ADHC enrollment [M (SD)] 2.0 (2.7) 2.3 (2.9) 1.7 (2.6) .473 Number of diagnosed conditions in the electronic health record [M (SD)] 24.3 (15.9) 23.8 (16.9) 24.8 (15.2) .841 Geriatric Pain Measure Score [M (SD)] 16.9 (12.4) 16.1 (12.0) 17.7 (13.1) .674 Primary Care PTSD score [n (%) with score ≥3] 5 (13.2%) 3 (15.0%) 2 (11.1%) .723 MMSE score [M (SD)] 25.9 (2.8) 26.0 (2.7) 25.8 (2.9) .870 OARS ADL/IAD total score [M (SD)] 19.7 (4.2) 19.3 (3.1) 20.1 (5.2) .547 PHQ-9 score [M (SD)] 6.3 (5.5) 5.4 (4.1) 7.2 (6.6) .298 SF-12 Physical health component score [M (SD)] 36.3 (8.5) 35.1 (6.1) 37.6 (10.4) .353 SF-12 Mental health component score [M (SD)] 50.1 (11.9) 53.3 (10.3) 46.9 (12.7) .081 Flinders Fatigue scale score [M (SD)] 8.7 (7.9) 8.0 (6.9) 9.5 (9.0) .555 Variable . Overall (N = 42) . SIP (N = 21) . IC (N = 21) . p-value . Age in years [M (SD)] 77.1 (9.9) 77.7 (10.2) 76.4 (9.9) .680 Gender [n (%) male] 39 (92.9%) 18 (85.7%) 21 (100%) .232 Race [n (%) non-Hispanic white] 30 (71.4%) 13 (61.9%) 17 (81%) .306 Years of education [M (SD)] 14.5 (2.5) 14.7 (2.9) 14.2 (2.0) .463 Marital status [n (%) married] 21 (50%) 10 (47.6%) 11 (52.4%) 1.00 Employment status .509  Unable to work [n (%)] 6 (14.3%) 4 (19%) 2 (9.5%)  Volunteer [n (%)] 4 (9.5%) 1 (4.8%) 3 (14.3%)  Retired [n (%)] 32 (76.2%) 16 (76.2%) 16 (76.2%) Current living arrangement .012  Own home [n (%)] 26 (63.4%) 12 (57%) 14 (70%)  Relative or friend’s home [n (%)] 4 (9.8%) 0 (0%) 4 (20%)  Board and care home/ assisted living facility [n (%)] 11 (26.8%) 9 (43%) 2 (10%) Years since ADHC enrollment [M (SD)] 2.0 (2.7) 2.3 (2.9) 1.7 (2.6) .473 Number of diagnosed conditions in the electronic health record [M (SD)] 24.3 (15.9) 23.8 (16.9) 24.8 (15.2) .841 Geriatric Pain Measure Score [M (SD)] 16.9 (12.4) 16.1 (12.0) 17.7 (13.1) .674 Primary Care PTSD score [n (%) with score ≥3] 5 (13.2%) 3 (15.0%) 2 (11.1%) .723 MMSE score [M (SD)] 25.9 (2.8) 26.0 (2.7) 25.8 (2.9) .870 OARS ADL/IAD total score [M (SD)] 19.7 (4.2) 19.3 (3.1) 20.1 (5.2) .547 PHQ-9 score [M (SD)] 6.3 (5.5) 5.4 (4.1) 7.2 (6.6) .298 SF-12 Physical health component score [M (SD)] 36.3 (8.5) 35.1 (6.1) 37.6 (10.4) .353 SF-12 Mental health component score [M (SD)] 50.1 (11.9) 53.3 (10.3) 46.9 (12.7) .081 Flinders Fatigue scale score [M (SD)] 8.7 (7.9) 8.0 (6.9) 9.5 (9.0) .555 M = mean; SD = standard deviation; ADHC = Adult Day Health Care; MMSE = Mini-Mental State Examination; PTSD = post-traumatic stress disorder; OARS = the Older Americans Resources and Services; ADL = activities of daily living; IADL = instrumental activities of daily living; SF-12 = Short-Form v12. Open in new tab Table 3 Sleep Characteristics of Randomized Study Participants at Baseline. Variable . Overall (N=42) . SIP (N=21) . IC (N = 21) . p-value . Objective sleep (actigraphy)  Sleep efficiency [M (SD)] 84.9% (7.9%) 83.1% (9.3%) 86.6% (5.9%) .152  Total sleep time [M (SD)] minutes 467.6 (86.2) 459.7 (86.8) 475.4 (86.9) .563  Number of nighttime awakenings [M (SD)] 23.6 (7.9) 23.3 (8.0) 23.9 (8.0) .793  Total nighttime wake time [M (SD)] minutes 82.6 (43.6) 93.5 (52.2) 71.6 (30.3) .104 Patient-reported sleep  PSQI total score [M (SD)] 6.8 (4.2) 6.6 (4.2) 7.0 (4.4) .721  PSQI factor 1 (sleep efficiency) [M (SD)] 2.5 (2.3) 2.7 (2.5) 2.3 (2.2) .647  PSQI factor 2 (perceived sleep quality) [M (SD)] 2.6 (2.1) 2.1 (1.8) 3.2 (2.2) .088  PSQI factor 3 (daily disturbances) [M (SD)] 1.7 (0.9) 1.8 (0.9) 1.5 (0.9) .310  ISI total score [M (SD)] 8.5 (6.6) 8.1 (6.5) 8.9 (7.0) .715  PSQI Total sleep time [M (SD)] 6.6 (2.0) 6.6 (1.9) 6.6 (2.0) .950  PSQI Sleep efficiency [M (SD)] 73.6 (18.0) 73.7 (17.4) 73.6 (19.1) .981 Variable . Overall (N=42) . SIP (N=21) . IC (N = 21) . p-value . Objective sleep (actigraphy)  Sleep efficiency [M (SD)] 84.9% (7.9%) 83.1% (9.3%) 86.6% (5.9%) .152  Total sleep time [M (SD)] minutes 467.6 (86.2) 459.7 (86.8) 475.4 (86.9) .563  Number of nighttime awakenings [M (SD)] 23.6 (7.9) 23.3 (8.0) 23.9 (8.0) .793  Total nighttime wake time [M (SD)] minutes 82.6 (43.6) 93.5 (52.2) 71.6 (30.3) .104 Patient-reported sleep  PSQI total score [M (SD)] 6.8 (4.2) 6.6 (4.2) 7.0 (4.4) .721  PSQI factor 1 (sleep efficiency) [M (SD)] 2.5 (2.3) 2.7 (2.5) 2.3 (2.2) .647  PSQI factor 2 (perceived sleep quality) [M (SD)] 2.6 (2.1) 2.1 (1.8) 3.2 (2.2) .088  PSQI factor 3 (daily disturbances) [M (SD)] 1.7 (0.9) 1.8 (0.9) 1.5 (0.9) .310  ISI total score [M (SD)] 8.5 (6.6) 8.1 (6.5) 8.9 (7.0) .715  PSQI Total sleep time [M (SD)] 6.6 (2.0) 6.6 (1.9) 6.6 (2.0) .950  PSQI Sleep efficiency [M (SD)] 73.6 (18.0) 73.7 (17.4) 73.6 (19.1) .981 M = mean; SD = standard deviation; PSQI = Pittsburgh Sleep Quality Index; ISI = Insomnia Severity Index. Open in new tab Table 3 Sleep Characteristics of Randomized Study Participants at Baseline. Variable . Overall (N=42) . SIP (N=21) . IC (N = 21) . p-value . Objective sleep (actigraphy)  Sleep efficiency [M (SD)] 84.9% (7.9%) 83.1% (9.3%) 86.6% (5.9%) .152  Total sleep time [M (SD)] minutes 467.6 (86.2) 459.7 (86.8) 475.4 (86.9) .563  Number of nighttime awakenings [M (SD)] 23.6 (7.9) 23.3 (8.0) 23.9 (8.0) .793  Total nighttime wake time [M (SD)] minutes 82.6 (43.6) 93.5 (52.2) 71.6 (30.3) .104 Patient-reported sleep  PSQI total score [M (SD)] 6.8 (4.2) 6.6 (4.2) 7.0 (4.4) .721  PSQI factor 1 (sleep efficiency) [M (SD)] 2.5 (2.3) 2.7 (2.5) 2.3 (2.2) .647  PSQI factor 2 (perceived sleep quality) [M (SD)] 2.6 (2.1) 2.1 (1.8) 3.2 (2.2) .088  PSQI factor 3 (daily disturbances) [M (SD)] 1.7 (0.9) 1.8 (0.9) 1.5 (0.9) .310  ISI total score [M (SD)] 8.5 (6.6) 8.1 (6.5) 8.9 (7.0) .715  PSQI Total sleep time [M (SD)] 6.6 (2.0) 6.6 (1.9) 6.6 (2.0) .950  PSQI Sleep efficiency [M (SD)] 73.6 (18.0) 73.7 (17.4) 73.6 (19.1) .981 Variable . Overall (N=42) . SIP (N=21) . IC (N = 21) . p-value . Objective sleep (actigraphy)  Sleep efficiency [M (SD)] 84.9% (7.9%) 83.1% (9.3%) 86.6% (5.9%) .152  Total sleep time [M (SD)] minutes 467.6 (86.2) 459.7 (86.8) 475.4 (86.9) .563  Number of nighttime awakenings [M (SD)] 23.6 (7.9) 23.3 (8.0) 23.9 (8.0) .793  Total nighttime wake time [M (SD)] minutes 82.6 (43.6) 93.5 (52.2) 71.6 (30.3) .104 Patient-reported sleep  PSQI total score [M (SD)] 6.8 (4.2) 6.6 (4.2) 7.0 (4.4) .721  PSQI factor 1 (sleep efficiency) [M (SD)] 2.5 (2.3) 2.7 (2.5) 2.3 (2.2) .647  PSQI factor 2 (perceived sleep quality) [M (SD)] 2.6 (2.1) 2.1 (1.8) 3.2 (2.2) .088  PSQI factor 3 (daily disturbances) [M (SD)] 1.7 (0.9) 1.8 (0.9) 1.5 (0.9) .310  ISI total score [M (SD)] 8.5 (6.6) 8.1 (6.5) 8.9 (7.0) .715  PSQI Total sleep time [M (SD)] 6.6 (2.0) 6.6 (1.9) 6.6 (2.0) .950  PSQI Sleep efficiency [M (SD)] 73.6 (18.0) 73.7 (17.4) 73.6 (19.1) .981 M = mean; SD = standard deviation; PSQI = Pittsburgh Sleep Quality Index; ISI = Insomnia Severity Index. Open in new tab Treatment Adherence Given the health status of ADHC patients, treatment adherence was thoroughly measured. All 42 randomized participants attended all four intervention sessions. One participant in the SIP group did not complete the third intervention session because he had another appointment to attend (missed 13% of the content). Interventionists’ ratings indicated that participants had “good” or “excellent” participation and comprehension during all sessions except one (ie, 167 out of 168 sessions). Nineteen out of 21 individuals (90.4%) assigned to the SIP program completed at least one weekly sleep diary during the intervention. Based on these diaries, the interventionist noted whether the participant went to bed and got out of bed within 15 minutes of their scheduled times, and based on that definition (ie, no more than 15 minutes deviation from recommended time), the percentage of nights on which each participant adhered to their assigned schedule was calculated. On average, participants went to bed more than 15 minutes earlier than their assigned bedtime on only 16% of nights (ie, the adhered to their schedule bedtime on 84% of nights). Similarly, they got out of bed more than 15 minutes later than their scheduled rise time 19% of nights (ie, they adhered to their scheduled rise time on 81% of nights). These metrics were not available for IC participants because they were not assigned a specific sleep schedule and did not monitor their sleep schedule during the intervention period. Outcomes Supplemental Table S1 includes marginal means for each outcome as a function of group membership (SIP, IC) and time (baseline, post-treatment, and 4 months) as obtained from the mixed model estimation. Objective Sleep Outcomes Three out of four actigraphy-measured sleep outcomes showed statistically significant differences between the treatment and control groups when assessing (1) the average change from baseline to post-treatment and (2) the average change from baseline to 4 months. Those outcomes were: sleep efficiency, number of nighttime awakenings, and total nighttime wake time, described in more detail below (also see Table 4). Table 4 Study Outcomes at Post-Treatment in SIP group Versus IC Group, Controlling for Baseline. Variable . Sleep Intervention Program (SIP) vs. information-only control (IC) . Difference at post-treatment vs. baseline [mean (95% CI] . p-value . Difference at 4-month follow-up vs. baseline [mean (95% CI] . p-value . Objective sleep (actigraphy)  Sleep efficiency [M (SD)]1 4.2% (1.1%, 7.2%) .007 4.1% (0.5%, 7.7%) .025  Total sleep time [M (SD)] minutes1 −2.0 (−38.1, 34.1) .913 17.6 (−36.1, 71.3) .521  Number of nighttime awakenings [M (SD)]2 −5.3 (−9.6, −1.0) .016 −4.5 (−8.6, −0.3) .035  Total nighttime wake time [M (SD)] minutes2 −30.8 (−49.2, −12.4) .001 −25.3 (−47.0, −3.5) .023 Patient-reported sleep quality  PSQI total score [M (SD)]2 0.5 (−1.2, 2.1) .571 −1.5 (−3.5, 0.4) .129  PSQI factor 1 (sleep efficiency) [M (SD)]2 0.4 (−0.8, 1.7) .508 −0.7 (−1.8, 0.5) .237  PSQI factor 2 (perceived sleep quality) [M (SD)]2 0.3 (−0.5, 1.1) .492 0.0 (−1.0, 1.0) .992  PSQI factor 3 (daily disturbances) [M (SD)]2 −0.3 (−0.9, 0.2) .242 −0.7 (−1.3, −0.1) .016  ISI total score [M (SD)]2 −1.7 (−4.3, 1.0) .217 −0.2 (−3.0, 2.5) .862 Secondary outcomes  PSQI hours of sleep [M (SD)] 0.3 (−0.5, 1.2) .458 0.3 (−0.6, 1.3) .517  PSQI sleep efficiency [M (SD)] −9.5% (−26.0%, 6.9%) .256 2.8% (−7.3%, 13.0%) .582  Flinders fatigue scale [M (SD)]3 −3.8 (−7.6, −0.0) .048 −1.1 (−4.5, 2.3) .537  PHQ-9 score [M (SD)]3 −1.1 (−3.9, 1.8) .459 −1.6 (−3.6, 0.5) .128  SF-12 PCS subscale [M (SD)]4 −0.5 (−6.7, 5.6) .864 −2.1 (−8.4, 4.2) .510  SF-12 MCS subscale [M (SD)]4 0.4 (−4.8, 5.6) .883 −0.6 (−6.2, 4.9) .822 Variable . Sleep Intervention Program (SIP) vs. information-only control (IC) . Difference at post-treatment vs. baseline [mean (95% CI] . p-value . Difference at 4-month follow-up vs. baseline [mean (95% CI] . p-value . Objective sleep (actigraphy)  Sleep efficiency [M (SD)]1 4.2% (1.1%, 7.2%) .007 4.1% (0.5%, 7.7%) .025  Total sleep time [M (SD)] minutes1 −2.0 (−38.1, 34.1) .913 17.6 (−36.1, 71.3) .521  Number of nighttime awakenings [M (SD)]2 −5.3 (−9.6, −1.0) .016 −4.5 (−8.6, −0.3) .035  Total nighttime wake time [M (SD)] minutes2 −30.8 (−49.2, −12.4) .001 −25.3 (−47.0, −3.5) .023 Patient-reported sleep quality  PSQI total score [M (SD)]2 0.5 (−1.2, 2.1) .571 −1.5 (−3.5, 0.4) .129  PSQI factor 1 (sleep efficiency) [M (SD)]2 0.4 (−0.8, 1.7) .508 −0.7 (−1.8, 0.5) .237  PSQI factor 2 (perceived sleep quality) [M (SD)]2 0.3 (−0.5, 1.1) .492 0.0 (−1.0, 1.0) .992  PSQI factor 3 (daily disturbances) [M (SD)]2 −0.3 (−0.9, 0.2) .242 −0.7 (−1.3, −0.1) .016  ISI total score [M (SD)]2 −1.7 (−4.3, 1.0) .217 −0.2 (−3.0, 2.5) .862 Secondary outcomes  PSQI hours of sleep [M (SD)] 0.3 (−0.5, 1.2) .458 0.3 (−0.6, 1.3) .517  PSQI sleep efficiency [M (SD)] −9.5% (−26.0%, 6.9%) .256 2.8% (−7.3%, 13.0%) .582  Flinders fatigue scale [M (SD)]3 −3.8 (−7.6, −0.0) .048 −1.1 (−4.5, 2.3) .537  PHQ-9 score [M (SD)]3 −1.1 (−3.9, 1.8) .459 −1.6 (−3.6, 0.5) .128  SF-12 PCS subscale [M (SD)]4 −0.5 (−6.7, 5.6) .864 −2.1 (−8.4, 4.2) .510  SF-12 MCS subscale [M (SD)]4 0.4 (−4.8, 5.6) .883 −0.6 (−6.2, 4.9) .822 Significant differences are shown in bold typeface. 1 Greater scores imply better sleep quality and positive differences represent improvements in sleep quality from baseline. 2 Lower scores imply better sleep quality, and negative differences represent improvements in sleep quality from baseline. 3 Higher scores indicate more depression/fatigue. 4 Higher scores indicate better quality of life. SIP = Sleep Intervention Program; IC = information-only control; CI = confidence interval; M = mean; SD = standard deviation; PSQI = Pittsburgh Sleep Quality Index; ISI = Insomnia Severity Index; PHQ = Patient Health Questionnaire; PCS = Physical Component Score; MCS = Mental Component Score. Open in new tab Table 4 Study Outcomes at Post-Treatment in SIP group Versus IC Group, Controlling for Baseline. Variable . Sleep Intervention Program (SIP) vs. information-only control (IC) . Difference at post-treatment vs. baseline [mean (95% CI] . p-value . Difference at 4-month follow-up vs. baseline [mean (95% CI] . p-value . Objective sleep (actigraphy)  Sleep efficiency [M (SD)]1 4.2% (1.1%, 7.2%) .007 4.1% (0.5%, 7.7%) .025  Total sleep time [M (SD)] minutes1 −2.0 (−38.1, 34.1) .913 17.6 (−36.1, 71.3) .521  Number of nighttime awakenings [M (SD)]2 −5.3 (−9.6, −1.0) .016 −4.5 (−8.6, −0.3) .035  Total nighttime wake time [M (SD)] minutes2 −30.8 (−49.2, −12.4) .001 −25.3 (−47.0, −3.5) .023 Patient-reported sleep quality  PSQI total score [M (SD)]2 0.5 (−1.2, 2.1) .571 −1.5 (−3.5, 0.4) .129  PSQI factor 1 (sleep efficiency) [M (SD)]2 0.4 (−0.8, 1.7) .508 −0.7 (−1.8, 0.5) .237  PSQI factor 2 (perceived sleep quality) [M (SD)]2 0.3 (−0.5, 1.1) .492 0.0 (−1.0, 1.0) .992  PSQI factor 3 (daily disturbances) [M (SD)]2 −0.3 (−0.9, 0.2) .242 −0.7 (−1.3, −0.1) .016  ISI total score [M (SD)]2 −1.7 (−4.3, 1.0) .217 −0.2 (−3.0, 2.5) .862 Secondary outcomes  PSQI hours of sleep [M (SD)] 0.3 (−0.5, 1.2) .458 0.3 (−0.6, 1.3) .517  PSQI sleep efficiency [M (SD)] −9.5% (−26.0%, 6.9%) .256 2.8% (−7.3%, 13.0%) .582  Flinders fatigue scale [M (SD)]3 −3.8 (−7.6, −0.0) .048 −1.1 (−4.5, 2.3) .537  PHQ-9 score [M (SD)]3 −1.1 (−3.9, 1.8) .459 −1.6 (−3.6, 0.5) .128  SF-12 PCS subscale [M (SD)]4 −0.5 (−6.7, 5.6) .864 −2.1 (−8.4, 4.2) .510  SF-12 MCS subscale [M (SD)]4 0.4 (−4.8, 5.6) .883 −0.6 (−6.2, 4.9) .822 Variable . Sleep Intervention Program (SIP) vs. information-only control (IC) . Difference at post-treatment vs. baseline [mean (95% CI] . p-value . Difference at 4-month follow-up vs. baseline [mean (95% CI] . p-value . Objective sleep (actigraphy)  Sleep efficiency [M (SD)]1 4.2% (1.1%, 7.2%) .007 4.1% (0.5%, 7.7%) .025  Total sleep time [M (SD)] minutes1 −2.0 (−38.1, 34.1) .913 17.6 (−36.1, 71.3) .521  Number of nighttime awakenings [M (SD)]2 −5.3 (−9.6, −1.0) .016 −4.5 (−8.6, −0.3) .035  Total nighttime wake time [M (SD)] minutes2 −30.8 (−49.2, −12.4) .001 −25.3 (−47.0, −3.5) .023 Patient-reported sleep quality  PSQI total score [M (SD)]2 0.5 (−1.2, 2.1) .571 −1.5 (−3.5, 0.4) .129  PSQI factor 1 (sleep efficiency) [M (SD)]2 0.4 (−0.8, 1.7) .508 −0.7 (−1.8, 0.5) .237  PSQI factor 2 (perceived sleep quality) [M (SD)]2 0.3 (−0.5, 1.1) .492 0.0 (−1.0, 1.0) .992  PSQI factor 3 (daily disturbances) [M (SD)]2 −0.3 (−0.9, 0.2) .242 −0.7 (−1.3, −0.1) .016  ISI total score [M (SD)]2 −1.7 (−4.3, 1.0) .217 −0.2 (−3.0, 2.5) .862 Secondary outcomes  PSQI hours of sleep [M (SD)] 0.3 (−0.5, 1.2) .458 0.3 (−0.6, 1.3) .517  PSQI sleep efficiency [M (SD)] −9.5% (−26.0%, 6.9%) .256 2.8% (−7.3%, 13.0%) .582  Flinders fatigue scale [M (SD)]3 −3.8 (−7.6, −0.0) .048 −1.1 (−4.5, 2.3) .537  PHQ-9 score [M (SD)]3 −1.1 (−3.9, 1.8) .459 −1.6 (−3.6, 0.5) .128  SF-12 PCS subscale [M (SD)]4 −0.5 (−6.7, 5.6) .864 −2.1 (−8.4, 4.2) .510  SF-12 MCS subscale [M (SD)]4 0.4 (−4.8, 5.6) .883 −0.6 (−6.2, 4.9) .822 Significant differences are shown in bold typeface. 1 Greater scores imply better sleep quality and positive differences represent improvements in sleep quality from baseline. 2 Lower scores imply better sleep quality, and negative differences represent improvements in sleep quality from baseline. 3 Higher scores indicate more depression/fatigue. 4 Higher scores indicate better quality of life. SIP = Sleep Intervention Program; IC = information-only control; CI = confidence interval; M = mean; SD = standard deviation; PSQI = Pittsburgh Sleep Quality Index; ISI = Insomnia Severity Index; PHQ = Patient Health Questionnaire; PCS = Physical Component Score; MCS = Mental Component Score. Open in new tab Sleep efficiency Relative to baseline values, the average change in sleep efficiency was greater for the SIP versus IC group at post-treatment (p = .007) and greater for the SIP versus IC group at 4-month follow-up (p = .025). The average improvement in sleep efficiency (relative to baseline) for the SIP group (vs. IC) at post-treatment was 4.2%. Similarly, relative to baseline, sleep efficiency improved in the SIP group by an average of 4.1% more than in the IC group at 4-month follow-up (see Table 4 and Figure 2, Panel A). Figure 2 Open in new tabDownload slide Mean values for actigraphy (ACTI) outcome variables for the Sleep Intervention Program (SIP) and control groups at baseline, post-treatment and 4-month follow-up for sleep efficiency (panel A), number of nighttime awakenings (panel B), total nighttime wake time (panel C) and Total sleep time (panel D). Data presented here includes all available observations for actigraphy at each time point (N = 42 at baseline, N = 38 post-treatment and, N = 38 at 4-month follow-up; also see Supplementary Figures S27–S30 for evaluation of potentially influential data points). Figure 2 Open in new tabDownload slide Mean values for actigraphy (ACTI) outcome variables for the Sleep Intervention Program (SIP) and control groups at baseline, post-treatment and 4-month follow-up for sleep efficiency (panel A), number of nighttime awakenings (panel B), total nighttime wake time (panel C) and Total sleep time (panel D). Data presented here includes all available observations for actigraphy at each time point (N = 42 at baseline, N = 38 post-treatment and, N = 38 at 4-month follow-up; also see Supplementary Figures S27–S30 for evaluation of potentially influential data points). Number of Awakenings Compared to baseline values, the change in number of nighttime awakenings was greater for the SIP versus IC group at post-treatment (p = .016) and greater for the SIP versus IC group at 4-months follow-up (p = .035). From baseline to post-treatment, the SIP group averaged 5.3 fewer nighttime awakenings compared to the IC group. Likewise, from baseline to 4-month follow-up, the SIP group averaged 4.5 fewer nighttime awakenings (see Table 4 and Figure 2, Panel B). Total Nighttime Wake Time As compared to the baseline values, the average number of nighttime awakenings decreased more for the SIP than the IC group at post-treatment (p = .001) and at 4-month follow-up (p = .023). The average reduction in total nighttime wake time from baseline to post-treatment was 30.8 minutes more for the SIP compared to the IC group. Similarly, from post-treatment to 4-month follow-up, the average reduction in total nighttime awake time was 25.3 minutes more for the SIP (vs. IC) group (see Table 4 and Figure 2; Panel C). Total Nighttime Sleep Time No significant differences between two groups were observed for total sleep time at either post-treatment or 4-month follow-up (see Table 4 and Figure 2; Panel D). Patient-Reported Sleep Outcomes There were no significant differences between the SIP and IC groups at either post-treatment or at 4-month follow-up for the PSQI total score, PSQI Factor 1 (sleep efficiency) or PSQI Factor 2 (perceived sleep quality see Table 4 and Supplementary Figures S1–S3). The PSQI scores on Factor 3 (daytime disturbances) showed greater reductions (improvement) from baseline to 4-month follow-up for the SIP versus IC group (p = .016 see Table 4 and Supplementary Figure S4). The average change in Factor 3 of the PSQI for the SIP (vs. IC) group was −.7 (−0.18, 0.6). The difference for this outcome at post-treatment was not significant. The treatment effects on ISI at either post-treatment or 4-month follow-up were not significant (see Table 4 and Supplementary Figure S5). There also were no significant treatment effects in terms of patient-reported hours of sleep or sleep efficiency (based on PSQI items; see Table 4 and Supplementary Figures S6–S7). Secondary Outcomes We also tested whether the SIP improved health (ie, fatigue, depression, health-related quality of life) at post-treatment and 4-month follow-ups, compared to the IC (see Supplementary Figures S8–S11). Only fatigue measured by FFS showed significant improvement at post-treatment in SIP group compared to the IC group (4.5 vs. 9.9, p = .048). No other differences were observed in these secondary outcome measures. Sensitivity Analyses Sensitivity analyses are presented in Supplementary Table S2 and Supplementary Figures S12–S26. At post-treatment, findings for sleep efficiency, number of nighttime awakenings, and total nighttime time awake were robust to omission of potentially influential data points with one exception. If two data points were excluded from the analysis of number of nighttime awakenings, the p-value increased from .016 to .053. At 4 months, the impact of exclusion of potentially influential data points was still minimal but had some impact on statistical significance, with p-values ranging from .020 to .083 depending on the outcome variable and the number of observations excluded (1–3). Total sleep time remained nonsignificant after omitting a potentially influential observation. DISCUSSION The overall pattern of results suggests that the SIP, delivered by a trained HE under the supervision of a sleep psychologist was feasible, with high levels of attendance and engagement by ADHC participants. The SIP resulted in relative improvements in objectively measured sleep based on wrist actigraphy compared to the IC group. Importantly, these differences in sleep were largely maintained at 4-month follow-up. The pattern of results suggest that the control group may have experienced gradually worsening sleep (based on actigraphy) over the study period, whereas the intervention group either improved slightly or declined at a slower rate. This is a common phenomenon in research on older adults, and a recent study found that older adults, particularly if they were using benzodiazepines, showed deterioration in sleep quality over a 1-year period.42 Improvements in patient-reported outcomes were modest, with only one PSQI component (daily disturbances) showing significantly better (lower) scores in the SIP condition at the 4-month follow-up but not at post-treatment. It is possible that, given the duration of sleep difficulties for many of these older patients, sustained improvements in sleep were needed before patients began to feel better during the daytime hours. Another possible explanation is that we did not use a specific cutoff score on either the ISI or PSQI to determine eligibility for randomization, and these two commonly used questionnaires did not appear to reflect the sleep experience of these patients. Older veterans who attend ADHC may not report sleep disturbances in the same way as healthier older adult populations and in fact, may tolerate significantly more sleep disturbance before noting poor sleep quality. Some participants, for example, reporting taking longer than 1 hour to fall asleep, but then reported that their sleep quality was “very good.” Furthermore, in our recent insomnia treatment study of older veterans in the same health care system who were not participating in the ADHC program,23 the mean baseline PSQI score was 9.1 and the mean baseline ISI score was 11.1, while their baseline sleep efficiency based on actigraphy was 83%. ADHC Patients in the current study had similar actigraphically assessed sleep efficiency (85%); however, their questionnaires reflected substantially less “complaint.” The mean baseline PSQI was only 6.8 (slightly above the clinical cutoff of 6 for sleep disturbance), and the mean baseline ISI was only 8.5 (indicating only mild insomnia, on average). It is possible that measures like the PSQI and ISI do not capture the way in which these older veterans describe their difficulties with sleep, subjectively. This may account for our discrepant findings. We also were not able to obtain completed sleep diaries, which are typically used in studies of behavioral treatments for insomnia. We asked the first 11 enrolled participants to complete a daily sleep diary based on the American Academy of Sleep Medicine consensus sleep diary42 during their baseline assessment, and none of the patients fully completed the diary during the 3-day baseline. As a result, we used a very simple four-item diary (bedtime, rise time, sleep quality, and daytime sleepiness), which did not allow for computation of traditional sleep diary measures such as sleep efficiency or total sleep time from daily diaries but could be used for actigraphy scoring and to establish adherence to the assigned sleep schedule in the SIP program. One study of older adults recruited from primary care found improvements in both patient-reported (daily sleep dairy, questionnaires) and objective (actigraphy) outcomes,21 although baseline sleep complaints were more severe (baseline PSQI = 11) in that study compared to our study as well. We did find improvements in fatigue, which was one of our secondary patient-reported outcomes, at post-treatment. This finding is consistent with the reduced impact of sleep disturbance at 4-month follow-up observed on the PSQI; however, we did not find improvements in depression or quality of life at either post-treatment or 4-month follow-up. It is important to note, however, that this study was not powered to detect the impact of the sleep intervention on these outcomes and additional research is needed. This may be important to emphasize because improvements in daytime symptoms may be seen as a significant benefit to patients (in terms of self-reported outcomes) even if sleep quality itself is not perceived to change. This study has several strengths, including the implementation of the intervention within the ADHC program; however, despite attempts to minimize costs of participation, it was difficult to identify patients interested in participating in the intervention program. Many individuals did not feel their poor sleep was worthy of clinical attention. Interestingly, improvements in patient-reported sleep quality on the PSQI or insomnia symptoms on the ISI were not seen. Only improvements in objectively measured sleep were observed and maintained over time. One consideration in the design of the intervention was the tolerability of sleep restriction therapy, which is a common, evidence-based component of cognitive-behavioral interventions. Because we anticipated that many of the study participants would find it difficult to dramatically and quickly reduce their time in bed at the first intervention session, we instead elected to use sleep compression therapy, in which time in bed is slowly reduced, rather than contracted, and then slowly expanded. We found that participants were receptive to this approach and had success adhering to the assigned sleep schedule. In fact, they stayed up until their assigned bedtimes on 84% of nights and got out of bed at or before their assigned rise time on 81% of nights. This level of adherence is similar to what has been seen in studies of healthy, younger individuals in receiving CBT-I.43,44 We also significantly modified the standard stimulus control instructions.45 The main reason for this modification was concern about nighttime fall risk in older adults with functional limitations. Despite this modification, the intervention remained effective in reducing total time awake at night. Rather than instructing participants to get of bed if awake at night, we focused on confining sleep to the bed and bedroom and eliminating nonsleep activities from the sleep environment outside of the nighttime sleep period (eg, watch TV in the family room rather than in bed in the afternoon). While this study had multiple strength, including a high participant retention rate and implementation of the intervention in the context of an ongoing clinical program, there are also several limitations. One key limitation is that we were not able to screen participants for sleep-disordered breathing, despite data to suggest this is very common in older adults with functional and/or cognitive impairments.46,47 Participants found completion of a home sleep apnea test overly burdensome and were unwilling to complete an overnight study in the sleep laboratory; therefore, while we had planned to identify and exclude participants with severe sleep apnea, this was not possible. In addition, findings from this study may not directly generalize to ADHC programs outside of VA. Veterans are predominantly male and have more complex comorbidities than nonveterans, which may impact the delivery and benefits of the SIP.8 Although the definition of ADHC is similar regardless of where the programs are located,6 there are likely to be differences in ADHC programming and resources outside of VA, and those differences might make it challenging to implement our SIP within community ADHC programs. In summary, a brief, structured sleep improvement program may improve objectively assessed nighttime sleep in older Veterans participating in an ADHC program, and these improvements were maintained at 4-month follow-up. Modest improvements in daytime functioning, including reduced fatigue and reduced impact of sleep on daytime functioning may also be achieved. Additional research is needed to better understand how to assess patient-reported outcomes and to confirm our findings of relatively improvements in objectively measured sleep. Future research should also evaluate how best to implement this intervention program into routine care at ADHCs and to consider whether it can be delivered in group formats because therapeutic interventions are often delivered to patients in groups in ADHC settings. SUPPLEMENTARY MATERIAL Supplementary material is available at SLEEP online. FUNDING This study was funded by the Veterans Affairs (VA) Rehabilitation Research and Development Service 1RX000135-01 (PI: Martin). Additional funding support for authors includes: VA Advanced Geriatrics Fellowship Program, VA Greater Los Angeles Healthcare System Geriatric Research, Education and Clinical Center (Dzierzewski, Song); UCLA Claude Pepper Older Americans Independence Center (5P30AG028748, PI: Dzierzewski); National Center for Advancing Translational Sciences UCLA CTSI (UL1TR000124, PI: Dzierzewski); National Institute on Aging, National Institutes of Health (K23AG045937, PI: Fung); National Institute on Aging, National Institutes of Health (K23AG049955, PI: Dzierzewski). ADDRESS WHERE WORK WAS CONDUCTED VA Greater Los Angeles Healthcare System, Geriatric Research, Education and Clinical Center, 16111 Plummer St. (11E), North Hills, CA 91343. CLINICAL TRIAL “Treating Sleep Problems in VA Adult Day Health Care.” NCT01259401. DISCLOSURE STATEMENT None declared. The content is solely the responsibility of the authors and does not necessarily represent the official views of Department of Veterans Affairs, National Institutes of Health, or the U.S. Government. ACKNOWLEDGMENTS The authors wish to thank ADHC program director Jo Ellen Baur, MSW, Maureen Burruel, MSW and all members of the ADHC staff for their support of this project. The authors also wish to acknowledge research staff members Sergio Martinez, Simone Vukelich, Sandra Fontal, MPA, and Diane Lee, MSW. Most of all, we wish to posthumously recognize the many contributions of Terry Z. Vandenberg, MA, who served as the lead health educator for the study. REFERENCES 1. Dam TT , Ewing S, Ancoli-Israel S, Ensrud K, Redline S, Stone K; Osteoporotic Fractures in Men Research Group. Association between sleep and physical function in older men: the osteoporotic fractures in men sleep study . J Am Geriatr Soc . 2008 ; 56 ( 9 ): 1665 – 1673 . Google Scholar Crossref Search ADS PubMed WorldCat 2. Nóbrega PV , Maciel AC, de Almeida Holanda CM, Oliveira Guerra R, Araújo JF. Sleep and frailty syndrome in elderly residents of long-stay institutions: a cross-sectional study . Geriatr Gerontol Int . 2014 ; 14 ( 3 ): 605 – 612 . Google Scholar Crossref Search ADS PubMed WorldCat 3. Neikrug AB , Ancoli-Israel S. Sleep disturbances in nursing homes . J Nutr Health Aging . 2010 ; 14 ( 3 ): 207 – 211 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Martin JL , Webber AP, Alam T, Harker JO, Josephson KR, Alessi CA. Daytime sleeping, sleep disturbance, and circadian rhythms in the nursing home . Am J Geriatr Psychiatry . 2006 ; 14 ( 2 ): 121 – 129 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Martin JL , Fiorentino L, Jouldjian S, Josephson KR, Alessi CA. Sleep quality in residents of assisted living facilities: effect on quality of life, functional status, and depression . J Am Geriatr Soc . 2010 ; 58 ( 5 ): 829 – 836 . Google Scholar Crossref Search ADS PubMed WorldCat 6. The MetLife National Study of Adult Day Services . MMI00156(1010). 2013 . New York, NY : Mature Market Institute . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 7. Office of Public Affairs and Media Relations . VA Long-term care. 2005 . Washington, DC : Department of Veterans Affairs . Google Scholar PubMed PubMed OpenURL Placeholder Text Google Preview WorldCat COPAC 8. Agha Z , Lofgren RP, VanRuiswyk JV, Layde PM. Are patients at Veterans Affairs medical centers sicker? A comparative analysis of health status and medical resource use . Arch Intern Med . 2000 ; 160 ( 21 ): 3252 – 3257 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Chapko M , Ehreth J, Hedrick SC, Rothman ML. Effects of adult day health care on utilization and cost of care for subgroups of patients . Med Care . 1993 ; 31 ( 9 Suppl ): SS62 – SS74 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 10. Martin JL , Alam T, Harker JO, Josephson KR, Alessi CA. Sleep patterns in assisted living facilities: a comparison to home-dwelling elders . J Gerontol A Biol Sci Med Sci . 2008 ; 163A : 1407 – 1409 . Google Scholar Crossref Search ADS WorldCat 11. Webber AP , Martin JL, Harker JO, Josephson KR, Rubenstein LZ, Alessi CA. Depression in older patients admitted for postacute nursing home rehabilitation . J Am Geriatr Soc . 2005 ; 53 ( 6 ): 1017 – 1022 . Google Scholar Crossref Search ADS PubMed WorldCat 12. Alessi CA , Webber AP, Josephson KR, Rubenstein L, Harker JO, Martin JL. Sleep and functional improvement in the nursing home setting [abstract] . Gerontologist . 2003 ; 43 : 490 Google Scholar OpenURL Placeholder Text WorldCat 13. Stone KL , Blackwell TL, Ancoli-Israel Set al. Sleep disturbances and risk of falls in older community-dwelling men: the outcomes of sleep disorders in older men (MrOS Sleep) study . J Am Geriatr Soc 2014 . Google Scholar OpenURL Placeholder Text WorldCat 14. Song Y , Dzierzewski JM, Fung CHet al. Association between sleep and physical function in older veterans in an adult day healthcare program . J Am Geriatr Soc . 2015 ; 63 ( 8 ): 1622 – 1627 . Google Scholar Crossref Search ADS PubMed WorldCat 15. Hughes J , Martin JL. Sleep characteristics of veterans affairs adult day health care participants . Behavioral Sleep Medicine . 2013 ; 11 : 258 – 274 . Google Scholar Crossref Search ADS PubMed WorldCat 16. Meissner HH , Riemer A, Santiago SM, Stein M, Goldman MD, Williams AJ. Failure of physician documentation of sleep complaints in hospitalized patients . West J Med . 1998 ; 169 ( 3 ): 146 – 149 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 17. Bloom HG , Ahmed I, Alessi CAet al. Evidence-based recommendations for the assessment and management of sleep disorders in older persons . J Am Geriatr Soc . 2009 ; 57 ( 5 ): 761 – 789 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Avidan AY , Fries BE, James ML, Szafara KL, Wright GT, Chervin RD. Insomnia and hypnotic use, recorded in the minimum data set, as predictors of falls and hip fractures in Michigan nursing homes . J Am Geriatr Soc . 2005 ; 53 ( 6 ): 955 – 962 . Google Scholar Crossref Search ADS PubMed WorldCat 19. Le Couteur DG , Latimer Hill E, Cumming RG, Lewis R, Carrington S. Sleep disturbances and falls in older people . J Gerontol A Biol Sci Med Sci . 2007 ; 62A : 62 – 66 . Google Scholar OpenURL Placeholder Text WorldCat 20. Sivertsen B , Omvik S, Pallesen Set al. Cognitive behavioral therapy vs zopiclone for treatment of chronic primary insomnia in older adults . JAMA . 2006 ; 295 : 2851 – 2858 . Google Scholar Crossref Search ADS PubMed WorldCat 21. Buysse DJ , Germain A, Moul DEet al. Efficacy of brief behavioral treatment for chronic insomnia in older adults . Arch Intern Med . 2011 ; 171 ( 10 ): 887 – 895 . Google Scholar Crossref Search ADS PubMed WorldCat 22. Jacobs GD , Pace-Schott EF, Stickgold R, Otto MW. Cognitive behavior therapy and pharmacotherapy for insomnia: a randomized controlled trial and direct comparison . Arch Intern Med . 2004 ; 164 ( 17 ): 1888 – 1896 . Google Scholar Crossref Search ADS PubMed WorldCat 23. Alessi CA , Martin JL, Fiorentino Let al. Cognitive behavioral therapy for insomnia in older veterans using non-clinician sleep coaches: a randomized controlled trial . J Am Geriatr Soc 2016 ; in press. Google Scholar OpenURL Placeholder Text WorldCat 24. Buysse DJ , Reynolds CF III, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research . Psychiatry Res . 1989 ; 28 ( 2 ): 193 – 213 . Google Scholar Crossref Search ADS PubMed WorldCat 25. Bastien CH , Vallières A, Morin CM. Validation of the Insomnia Severity Index as an outcome measure for insomnia research . Sleep Med . 2001 ; 2 ( 4 ): 297 – 307 . Google Scholar Crossref Search ADS PubMed WorldCat 26. Kroenke K , Spitzer RL, Williams JBW. Validity of a brief depression severity measure . J Gen Intern Med . 2001 ; 16 : 606 – 613 . Google Scholar Crossref Search ADS PubMed WorldCat 27. Ouimette P , Wade M, Prins A, Schohn M. Identifying PTSD in primary care: comparison of the Primary Care-PTSD screen (PC-PTSD) and the General Health Questionnaire-12 (GHQ) . J Anxiety Disord . 2008 ; 22 ( 2 ): 337 – 343 . Google Scholar Crossref Search ADS PubMed WorldCat 28. Folstein MF , Folstein SE, McHugh PR. “Mini-mental state.” A practical method for grading the cognitive state of patients for the clinician . J Psychiatr Res . 1975 ; 12 ( 3 ): 189 – 198 . Google Scholar Crossref Search ADS PubMed WorldCat 29. Duke University Center for the Study of Aging and Human Development . Duke Older Americans Resources and Services Program: An Information System for Functional Assessment, Program Evaluation and Resource Allocation . Center Reports on Advances in Research 1985 ; 9:whole issue. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 30. Gradisar M , Lack L, Richards Het al. The Flinders Fatigue Scale: preliminary psychometric properties and clinical sensitivity of a new scale for measuring daytime fatigue associated with insomnia . J Clin Sleep Med . 2007 ; 3 ( 7 ): 722 – 728 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 31. Ware J Jr, Kosinski M, Keller SD. A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity . Med Care . 1996 ; 34 ( 3 ): 220 – 233 . Google Scholar Crossref Search ADS PubMed WorldCat 32. National Institutes of Health . Manifestations and management of chronic insomnia in adults . Sleep . 2005 ; 28 : 1049 – 1057 . Crossref Search ADS PubMed WorldCat 33. Lichstein KL , Riedel BW, Wilson NM, Lester KW, Aguillard RN. Relaxation and sleep compression for late-life insomnia: a placebo-controlled trial . J Consult Clin Psychol . 2001 ; 69 ( 2 ): 227 – 239 . Google Scholar Crossref Search ADS PubMed WorldCat 34. McCrae CS , McGovern R, Lukefahr R, Stripling AM. Research Evaluating Brief Behavioral Sleep Treatments for Rural Elderly (RESTORE): a preliminary examination of effectiveness . Am J Geriatr Psychiatry . 2007 ; 15 ( 11 ): 979 – 982 . Google Scholar Crossref Search ADS PubMed WorldCat 35. Cole JC , Motivala SJ, Buysse DJ, Oxman MN, Levin MJ, Irwin MR. Validation of a 3-factor scoring model for the Pittsburgh sleep quality index in older adults . Sleep . 2006 ; 29 ( 1 ): 112 – 116 . Google Scholar Crossref Search ADS PubMed WorldCat 36. Standards of Practice Committee , Morgenthaler T, Alessi CAet al. Practice parameters for the use of actigraphy in the assessment of sleep and sleep disorders: an update for 2007 . Sleep . 2007 ; 30 : 519 – 529 . Google Scholar Crossref Search ADS PubMed WorldCat 37. Centers for Disease Control . International Classification of Diseases, Ninth Revision (ICD-9) . 1979 . CDC. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 38. Ferrell BA , Stein WM, Beck JC. The Geriatric Pain Measure: validity, reliability and factor analysis . J Am Geriatr Soc . 2000 ; 48 ( 12 ): 1669 – 1673 . Google Scholar Crossref Search ADS PubMed WorldCat 39. Ouimette P , Wade M, Prins A, Schohn M. Identifying PTSD in primary care: comparison of the Primary Care-PTSD screen (PC-PTSD) and the General Health Questionnaire-12 (GHQ) . J Anxiety Disord . 2008 ; 22 ( 2 ): 337 – 343 . Google Scholar Crossref Search ADS PubMed WorldCat 40. Kroenke K , Spitzer RL, Williams JBW. Validity of a brief depression severity measure . J Gen Intern Med . 2001 ; 16 : 606 – 613 . Google Scholar Crossref Search ADS PubMed WorldCat 41. Bourgeois J , Elseviers MM, Van Bortel L, Petrovic M, Vander Stichele RH. One-year evolution of sleep quality in older users of benzodiazepines: a longitudinal cohort study in belgian nursing home residents . Drugs Aging . 2014 ; 31 ( 9 ): 677 – 682 . Google Scholar Crossref Search ADS PubMed WorldCat 42. Carney CE , Buysse DJ, Ancoli-Israel Set al. The consensus sleep diary: standardizing prospective sleep self-monitoring . Sleep . 2012 ; 35 ( 2 ): 287 – 302 . Google Scholar Crossref Search ADS PubMed WorldCat 43. Matthews EE , Schmiege SJ, Cook PF, Berger AM, Aloia MS. Adherence to cognitive behavioral therapy for insomnia (CBTI) among women following primary breast cancer treatment: a pilot study . Behav Sleep Med . 2012 ; 10 ( 3 ): 217 – 229 . Google Scholar Crossref Search ADS PubMed WorldCat 44. Bouchard S , Bastien C, Morin CM. Self-efficacy and adherence to cognitive-behavioral treatment of insomnia . Behav Sleep Med . 2003 ; 1 ( 4 ): 187 – 199 . Google Scholar Crossref Search ADS PubMed WorldCat 45. Bootzin RR , Epstein D. Stimulus control . In: Lichstein KL, Morin CM, eds. Treatment of late-life insomnia . Thousand Oaks, California : Sage Publications, Inc .; 2000 ; 167 – 184 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC 46. Song Y , Blackwell T, Yaffe K, Ancoli-Israel S, Redline S, Stone KL; Osteoporotic Fractures in Men (MrOS) Study Group. Relationships between sleep stages and changes in cognitive function in older men: the MrOS Sleep Study . Sleep . 2015 ; 38 ( 3 ): 411 – 421 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 47. Spira AP , Blackwell T, Stone KLet al. Sleep-disordered breathing and cognition in older women . J Am Geriatr Soc . 2008 ; 56 ( 1 ): 45 – 50 . Google Scholar Crossref Search ADS PubMed WorldCat Author notes Address correspondence to: Jennifer L. Martin, PhD, VA Greater Los Angeles Healthcare System, Geriatric Research, Education and Clinical Center, 16111 Plummer St. (11E), North Hills, CA 91343, USA. Telephone: 818-891-7711 × 36080; Fax: 818-895-9519; Email: [email protected] Published by Oxford University Press on behalf of Sleep Research Society (SRS) 2017. This work is written by (a) US Government employee(s) and is in the public domain in the US. Published by Oxford University Press on behalf of Sleep Research Society (SRS) 2017. This work is written by (a) US Government employee(s) and is in the public domain in the US.

Ferri,, Raffaele;Rundo,, Francesco;Silvani,, Alessandro;Zucconi,, Marco;Bruni,, Oliviero;Ferini-Strambi,, Luigi;Plazzi,, Giuseppe;Manconi,, Mauro

doi: 10.1093/sleep/zsx080pmid: 28482056

Abstract Study Objectives We aimed to analyze quantitatively rapid eye movement (REM) sleep electroencephalogram (EEG) in controls, drug-naïve idiopathic REM sleep behavior disorder patients (iRBD), and iRBD patients treated with clonazepam. Methods Twenty-nine drug-naïve iRBD patients (mean age 68.2 years), 14 iRBD patients under chronic clonazepam therapy (mean age 66.3 years), and 21 controls (mean age 66.8 years) were recruited. Power spectra were obtained from sleep EEG (central derivation), using a 2-second sliding window, with 1-second steps. The power values of each REM sleep EEG spectral band (one every second) were normalized with respect to the average power value obtained during sleep stage 2 in the same individual. Results In drug-naïve patients, the normalized power values showed a less pronounced REM-related decrease of power in all bands with frequency <15 Hz than controls and an increase in the beta band, negatively correlated with muscle atonia; in patients treated with clonazepam there was a partial return of all bands <15 Hz toward the control values. The standard deviation values of the normalized power were higher for untreated patients in all EEG bands and were almost completely normalized in patients treated with clonazepam. Conclusions The REM sleep EEG structure changes found in this study disclose subtle but significant alterations in the cortical electrophysiology of RBD that might represent the early expression of the supposed neurodegenerative processes already taking place at this stage of the disease and might be the target of better and effective future therapeutic strategies for this condition. REM Sleep, REM Sleep Behavior Disorder, REM sleep without atonia, Electroencephalography, Neurodegeneration, Synucleinopathy Statement of Significance Previous studies on the sleep EEG in patients with REM sleep behavior disorder have been able to detect changes only during non-REM (NREM) sleep, and also the effects of clonazepam in these patients have been detected only during NREM sleep. This study shows EEG changes during REM sleep in patients with REM sleep behavior disorder, characterized by less pronounced difference from NREM sleep and increased instability, that are partially recovered after chronic therapy with clonazepam. The knowledge of the mechanisms underlying the effects of clonazepam in this disorder might prove to be essential to better understanding its physiopathology and to arrange new and more effective (neuroprotective) drug approaches. INTRODUCTION Rapid eye movement (REM) sleep behavior disorder (RBD) is a parasomnia1 characterized by abnormal behaviors emerging during REM sleep, often causing injury.2,3 RBD is frequently associated or precedes neurodegenerative diseases such as synucleinopathies and is regarded as an early marker and an heraldic symptom of neurodegeneration.4–6 The term idiopathic RBD (iRBD) is used for patients without other clinical conditions.7 The presence of REM sleep without atonia (RSWA) is the main polygraphic feature of RBD.1 Bedtime clonazepam, together with melatonin, is a first-line treatment for iRBD because of the response rate of close to 90%3 and its relative safety (it is not indicated in patients with dementia, gait disturbance, or obstructive sleep apnea),8 even after years of nightly therapy;9 however, there are no double-blind, placebo-controlled, randomized trials with clonazepam in iRBD.10,11 Moreover, it is difficult to devise such a study in an ethically feasible manner, given the risk of recurrent injuries (and potential lethality) usually associated with iRBD.12 Only few studies have quantitatively analyzed the electroencephalogram (EEG) changes during REM sleep in iRBD patients13,14 and none, to our knowledge, has evaluated the eventual modifications induced by clonazepam on REM sleep EEG. The aim of this study was to analyze the differences in quantitative EEG features during REM sleep between normal controls and drug-naïve iRBD patients and another group of iRBD patients under a long-lasting regular therapy with clonazepam. METHOD Subjects and Experimental Design For this observational study, we retrospectively collected recordings that followed a standardized protocol in consecutive iRBD patients who were considered for participation in previous studies published by our groups15–20 and whose recordings corresponded to the technical specifications reported below in the “Nocturnal polysomnography” section. The diagnosis was based on the International Classification of Sleep Disorders, Third Edition criteria1 for RBD. Secondary forms of RBD were excluded on the basis of historical data, neurologic examination, and encephalic magnetic resonance imaging findings. For this study, we then identified a subgroup of consecutive patients with a video-polysomnography (vPSG) carried out when they had never been treated before with clonazepam and another subgroup with a vPSG recorded after a period of at least 1 year of regular and effective treatment with clonazepam (0.5–2 mg at bedtime). None of the normal volunteers recruited had any physical, neurological, or psychiatric disorder or history of sleep problems, and none was taking medication at the time of recording or had ever used a neuroleptic agent or selective serotonin reuptake inhibitors, or venlafaxine. The original studies were approved by the local ethics committees and all subjects had provided informed consent before entering the study. Nocturnal Polysomnography Standard nocturnal vPSG was carried out which included EEG, electrooculogram, electromyogram (EMG) of the submentalis muscle and of both tibialis anterior muscles, and electrocardiogram (ECG). Sleep signals were sampled at 128 Hz and stored on hard disk for further analysis. The sleep respiratory pattern of each patient was monitored using oral and nasal airflow thermistors and⁄or nasal pressure cannula, thoracic and abdominal respiratory effort strain gauge, and by monitoring oxygen saturation. Patients with an apnea⁄hypopnea index >5 were not included. Sleep stages were scored following standard criteria21 on 30-second epochs; because muscle atonia can be absent in RBD, REM sleep was scored without submental EMG atonia, using EEG and electrooculogram only. Onset and offset of a REM sleep period were defined according to a method specifically developed for RBD22,23 and available only for the old Rechtschaffen and Kales sleep staging criteria.21 Epochs containing technical artifacts or extremely elevated muscle activity causing saturation of amplifiers were carefully detected and marked for exclusion from the subsequent quantitative EEG analysis. A quantitative analysis of the submentalis muscle EMG activity was carried out using an established automatic scoring algorithm (REM sleep atonia index),15,16 which correlates significantly with the percentage of epochs of RSWA detected with the visual method by Lapierre and Montplaisir.15,22,24 Finally, periodic leg movements were also detected and analyzed using international standard criteria.25 Computation and Analysis of EEG Power Spectra For this study, the C3/A2 or C4/A1 EEG derivation was used, for each recording, sampled at 128 Hz. Sleep epochs containing artifacts were carefully excluded from the analysis, as specified above. A fast Fourier Transform was performed with a 2-second sliding window, every second, on EEG signals from all sleep stages, after Welch windowing. Subsequently, the total absolute power (0.5–32 Hz) and that of five different EEG bands of interest was computed for each EEG epoch analyzed (delta 0.5–2.5 Hz; theta 4.5–7.5; alpha 8–11.5 Hz; sigma 12–15 Hz; beta 15.5–30 Hz). Finally, relative power values were obtained by calculating the ratio of the absolute power of each band to the total power and then multiplying the value by 100. The relative power values obtained were then averaged for each sleep stage in all subjects. We chose to compare the relative spectral power obtained in the different groups because of the well-known important interindividual differences in EEG amplitude connected with several factors, such as skull thickness and sex, also previously reported in RBD patients.13 Analysis of REM Sleep Instability Subsequently, we carried out a careful normalization of data taking the average absolute power of the different EEG bands of interest during sleep stage 2 as the reference power, in each subject. Several alternatives were evaluated, but this was considered to be the best because sleep stage 2 is the sleep stage most represented in all recordings (accounting for approximately 50% of total sleep time) and it seems to be substantially unchanged in RBD.18,26 These average absolute power values were used as reference (Vref) for the subsequent, point-by-point calculation of the normalized values (Vnorm) of the observed absolute power values (Vobs) obtained in REM sleep, following the formula: Vnorm= (Vobs– Vref)/Vref In this way, if Vobs is higher than Vref, a positive Vnorm value is obtained, on the contrary, Vnorm is negative. The mean and standard deviation (SD) of the REM sleep Vnorm values for each EEG band were further used in this study for statistical analysis, with the first representing the normalized changes of REM sleep values versus sleep stage 2 and the latter representing the magnitude of their variability. We will refer to these new normalized EEG values as “EEG power ratio” in the following sections of this paper. Statistical Analysis Before running the final analyses, a power/sample size analysis was performed on the data obtained for the delta band during REM in all subjects recruited and a sample size of 11 subjects per group was found for a power 80% and alpha 0.05 or 13 subjects per group with power 80% and alpha 0.025. Between-group comparisons were performed by means of the analysis of variance, followed by the post hoc LSD test. The chi-square test was used for frequencies. Finally, the multiple regression analysis, with the calculation of the partial correlation coefficient was carried out between atonia index and the relative power of the different EEG bands during REM sleep. Differences were considered significant when they were below the p < .05 level. RESULTS Twenty-nine consecutive drug-naïve iRBD patients were retrospectively recruited (all men), as well as 14 iRBD patients under clonazepam (0.5–2 mg at bedtime) therapy (12 males and 2 females) and 21 normal controls (8 males and 13 females). The number of subjects in each group was higher than the sample size needed, according to the power analysis described above. The gender composition of the groups was evidently different (χ2 = 26.8, p < .0001); however, this was felt to be of low impact because of the normalization of data, and because no differences were found in relative power spectra between genders in normal controls. The comparison between age and the different polygraphic sleep parameters obtained in the three groups of subjects is reported in Table 1. Mean age at onset of RBD was 62.4 years (7.20 SD) in drug-naïve patients and 59.0 (4.01 SD) in treated patients (t = 1.68, NS). Only clonazepam was taken by the treated group, and none of the patients had comorbid conditions. Normal controls showed a higher number of awakenings, lower sleep efficiency, higher percentage of wakefulness after sleep onset, lower amount of slow-wave sleep and higher REM sleep atonia index than both untreated and treated iRBD groups, and a lower number of sleep stage shifts than iRBD patients. Drug-naïve iRBD only differed from treated iRBD because of a lower amount of sleep stage 2. Table 1 Comparison Between Age and Different Polygraphic Sleep Parameters Obtained in the Three Groups of Subjects. Variable 1. Controls (n=21) 2. Drug-naïve iRBD (n=29) 3. Treated iRBD (n=14) ANOVA Post hoc LSD tests Mean SD Mean SD Mean SD F(2,61) p< 1 versus 2 2 versus 3 1 versus 3 Age, years 66.8 7.24 68.2 6.46 66.3 4.88 0.520 NS Total sleep time, min 356.7 90.62 342.3 68.04 368.5 38.06 0.676 NS Sleep latency, min 23.6 25.90 28.5 48.65 16.0 8.06 0.564 NS REM sleep latency, min 94.0 84.75 88.7 46.81 106.0 60.02 0.341 NS Number of stage shifts/hour 12.2 3.93 16.6 6.52 14.3 5.07 3.923 .025 0.007 NS NS Number of awakenings/hour 7.2 2.90 5.1 2.96 4.0 2.32 5.946 .0044 0.012 NS 0.002 Sleep efficiency, % 68.3 12.55 76.5 13.62 83.5 7.85 6.744 .0023 0.022 NS 0.0006 Wakefulness after sleep onset, % 27.3 12.20 17.1 9.89 11.5 8.75 10.623 .0001 0.0012 NS 0.00005 Sleep stage 1, % 7.1 4.05 9.2 3.90 7.3 3.47 2.161 NS Sleep stage 2, % 41.3 10.89 38.5 8.53 46.6 9.14 3.393 .04 NS 0.012 NS Slow-wave sleep, % 10.4 7.82 17.5 7.45 17.5 6.75 6.434 .003 0.0014 0.0076 REM sleep, % 13.9 3.98 17.7 6.99 17.2 6.59 2.573 NS REM sleep atonia index 0.94 0.039 0.73 0.203 0.78 0.154 11.136 .000075 0.00002 NS 0.0042 PLMS index 11.0 14.33 27.2 31.60 14.9 19.66 2.895 NS Variable 1. Controls (n=21) 2. Drug-naïve iRBD (n=29) 3. Treated iRBD (n=14) ANOVA Post hoc LSD tests Mean SD Mean SD Mean SD F(2,61) p< 1 versus 2 2 versus 3 1 versus 3 Age, years 66.8 7.24 68.2 6.46 66.3 4.88 0.520 NS Total sleep time, min 356.7 90.62 342.3 68.04 368.5 38.06 0.676 NS Sleep latency, min 23.6 25.90 28.5 48.65 16.0 8.06 0.564 NS REM sleep latency, min 94.0 84.75 88.7 46.81 106.0 60.02 0.341 NS Number of stage shifts/hour 12.2 3.93 16.6 6.52 14.3 5.07 3.923 .025 0.007 NS NS Number of awakenings/hour 7.2 2.90 5.1 2.96 4.0 2.32 5.946 .0044 0.012 NS 0.002 Sleep efficiency, % 68.3 12.55 76.5 13.62 83.5 7.85 6.744 .0023 0.022 NS 0.0006 Wakefulness after sleep onset, % 27.3 12.20 17.1 9.89 11.5 8.75 10.623 .0001 0.0012 NS 0.00005 Sleep stage 1, % 7.1 4.05 9.2 3.90 7.3 3.47 2.161 NS Sleep stage 2, % 41.3 10.89 38.5 8.53 46.6 9.14 3.393 .04 NS 0.012 NS Slow-wave sleep, % 10.4 7.82 17.5 7.45 17.5 6.75 6.434 .003 0.0014 0.0076 REM sleep, % 13.9 3.98 17.7 6.99 17.2 6.59 2.573 NS REM sleep atonia index 0.94 0.039 0.73 0.203 0.78 0.154 11.136 .000075 0.00002 NS 0.0042 PLMS index 11.0 14.33 27.2 31.60 14.9 19.66 2.895 NS ANOVA = analysis of variance; iRBD = idiopathic REM sleep behavior disorder; LSD = least significant difference; NS = not significant; PLMS = periodic leg movements during sleep; REM = rapid-eye-movement; SD = standard deviation. Open in new tab Table 1 Comparison Between Age and Different Polygraphic Sleep Parameters Obtained in the Three Groups of Subjects. Variable 1. Controls (n=21) 2. Drug-naïve iRBD (n=29) 3. Treated iRBD (n=14) ANOVA Post hoc LSD tests Mean SD Mean SD Mean SD F(2,61) p< 1 versus 2 2 versus 3 1 versus 3 Age, years 66.8 7.24 68.2 6.46 66.3 4.88 0.520 NS Total sleep time, min 356.7 90.62 342.3 68.04 368.5 38.06 0.676 NS Sleep latency, min 23.6 25.90 28.5 48.65 16.0 8.06 0.564 NS REM sleep latency, min 94.0 84.75 88.7 46.81 106.0 60.02 0.341 NS Number of stage shifts/hour 12.2 3.93 16.6 6.52 14.3 5.07 3.923 .025 0.007 NS NS Number of awakenings/hour 7.2 2.90 5.1 2.96 4.0 2.32 5.946 .0044 0.012 NS 0.002 Sleep efficiency, % 68.3 12.55 76.5 13.62 83.5 7.85 6.744 .0023 0.022 NS 0.0006 Wakefulness after sleep onset, % 27.3 12.20 17.1 9.89 11.5 8.75 10.623 .0001 0.0012 NS 0.00005 Sleep stage 1, % 7.1 4.05 9.2 3.90 7.3 3.47 2.161 NS Sleep stage 2, % 41.3 10.89 38.5 8.53 46.6 9.14 3.393 .04 NS 0.012 NS Slow-wave sleep, % 10.4 7.82 17.5 7.45 17.5 6.75 6.434 .003 0.0014 0.0076 REM sleep, % 13.9 3.98 17.7 6.99 17.2 6.59 2.573 NS REM sleep atonia index 0.94 0.039 0.73 0.203 0.78 0.154 11.136 .000075 0.00002 NS 0.0042 PLMS index 11.0 14.33 27.2 31.60 14.9 19.66 2.895 NS Variable 1. Controls (n=21) 2. Drug-naïve iRBD (n=29) 3. Treated iRBD (n=14) ANOVA Post hoc LSD tests Mean SD Mean SD Mean SD F(2,61) p< 1 versus 2 2 versus 3 1 versus 3 Age, years 66.8 7.24 68.2 6.46 66.3 4.88 0.520 NS Total sleep time, min 356.7 90.62 342.3 68.04 368.5 38.06 0.676 NS Sleep latency, min 23.6 25.90 28.5 48.65 16.0 8.06 0.564 NS REM sleep latency, min 94.0 84.75 88.7 46.81 106.0 60.02 0.341 NS Number of stage shifts/hour 12.2 3.93 16.6 6.52 14.3 5.07 3.923 .025 0.007 NS NS Number of awakenings/hour 7.2 2.90 5.1 2.96 4.0 2.32 5.946 .0044 0.012 NS 0.002 Sleep efficiency, % 68.3 12.55 76.5 13.62 83.5 7.85 6.744 .0023 0.022 NS 0.0006 Wakefulness after sleep onset, % 27.3 12.20 17.1 9.89 11.5 8.75 10.623 .0001 0.0012 NS 0.00005 Sleep stage 1, % 7.1 4.05 9.2 3.90 7.3 3.47 2.161 NS Sleep stage 2, % 41.3 10.89 38.5 8.53 46.6 9.14 3.393 .04 NS 0.012 NS Slow-wave sleep, % 10.4 7.82 17.5 7.45 17.5 6.75 6.434 .003 0.0014 0.0076 REM sleep, % 13.9 3.98 17.7 6.99 17.2 6.59 2.573 NS REM sleep atonia index 0.94 0.039 0.73 0.203 0.78 0.154 11.136 .000075 0.00002 NS 0.0042 PLMS index 11.0 14.33 27.2 31.60 14.9 19.66 2.895 NS ANOVA = analysis of variance; iRBD = idiopathic REM sleep behavior disorder; LSD = least significant difference; NS = not significant; PLMS = periodic leg movements during sleep; REM = rapid-eye-movement; SD = standard deviation. Open in new tab None of the comparisons made between the three groups of subjects with the absolute EEG power spectra values, during all sleep stages, was significant (Supplemental Table e-1). The comparison between the relative EEG power obtained for the delta, theta, alpha, sigma, and beta bands in the three groups of subjects, during each sleep stage is depicted in Figure 1. Treated iRBD patients had lower delta band relative power than drug-naïve iRBD with statistical significance during REM sleep. The theta band was significantly higher in controls than in both patient groups only during slow-wave sleep. The alpha band tended to be lower in controls, but the difference reached statistical significance only versus treated iRBD in sleep stage 2 and slow-wave sleep. The sigma band was lowest in drug-naïve iRDB and statistically significant in sleep stages one, two, and slow-wave sleep. The beta band was higher in REM sleep in treated iRBD versus drug-naïve iRBD. A positive correlation was found between the beta band and the REM sleep atonia index (Supplemental Figure e-1). Figure 1 Open in new tabDownload slide Comparison between the relative electroencephalographic (EEG) power obtained for the delta, theta, alpha, sigma, and beta bands in the three groups of subjects, during each sleep stage. The numbers in the graph indicate the post hoc least significant difference (LSD) test p values obtained when the relative EEG band analysis of variance was statistically significant; data are shown as mean (bars) and standard deviation (SD; whiskers). Figure 1 Open in new tabDownload slide Comparison between the relative electroencephalographic (EEG) power obtained for the delta, theta, alpha, sigma, and beta bands in the three groups of subjects, during each sleep stage. The numbers in the graph indicate the post hoc least significant difference (LSD) test p values obtained when the relative EEG band analysis of variance was statistically significant; data are shown as mean (bars) and standard deviation (SD; whiskers). Figure 2 shows the comparison between the EEG power ratio obtained for all EEG bands in the three groups of subjects. It is possible to note that the clear decrease in delta, alfa, and sigma bands occurring in controls during REM sleep was significantly smaller in drug-naïve iRBD. Also, patients with treated iRBD showed a decrease in these bands smaller than that of controls but not versus untreated iRBD and the difference was statistically significant for the sigma band. The beta band EEG power ratio was, conversely, increased in both untreated and treated iRBD versus controls. Figure 2 Open in new tabDownload slide Comparison between the electroencephalographic (EEG) power ratio obtained for the delta, alpha, sigma, and beta bands in the three groups of subjects. All band analyses of variance were statistically significant; The numbers in the graph indicate the post hoc least significant difference (LSD) test p values. Data are shown as mean (bars) and standard error (SE; whiskers). Figure 2 Open in new tabDownload slide Comparison between the electroencephalographic (EEG) power ratio obtained for the delta, alpha, sigma, and beta bands in the three groups of subjects. All band analyses of variance were statistically significant; The numbers in the graph indicate the post hoc least significant difference (LSD) test p values. Data are shown as mean (bars) and standard error (SE; whiskers). The average SD of the EEG power ratio are shown in Figure 3. Drug-naïve iRBD patients show the highest values and controls the lowest; patients with treated iRBD having intermediate average values in all cases. In particular, the average values of the EEG power ratio SD obtained in patients taking clonazepam is significantly smaller than that of drug-naïve iRBD for the delta, sigma, and beta bands. Figure 3 Open in new tabDownload slide Comparison between the standard deviation of the electroencephalographic (EEG) power ratio (see Figure 2) obtained for the delta, alpha, sigma, and beta bands in the three groups of subjects. The numbers in the graph indicate the post hoc least significant difference (LSD) test p values obtained when the relative EEG band analysis of variance was statistically significant. Data are shown as mean (bars) and standard error (SE; whiskers). Figure 3 Open in new tabDownload slide Comparison between the standard deviation of the electroencephalographic (EEG) power ratio (see Figure 2) obtained for the delta, alpha, sigma, and beta bands in the three groups of subjects. The numbers in the graph indicate the post hoc least significant difference (LSD) test p values obtained when the relative EEG band analysis of variance was statistically significant. Data are shown as mean (bars) and standard error (SE; whiskers). DISCUSSION The present study is the first clear demonstration of EEG changes during REM sleep in iRBD patients in whom EEG power spectrum seems to be characterized by a lower degree of difference from that of stage 2 (vs. normal controls) and by a higher degree of instability. Moreover, chronic treatment with clonazepam seems to be able to revert only partially these changes. Fantini et al.13 were the first to report that, compared with controls, untreated iRBD patients had a lower spectral EEG power in the beta band, over the occipital regions during REM sleep. However, the observed effects were of marginal statistical significance and some methodological flaws affected their results because their beta1 band (13–22 Hz) included some sigma band frequencies, with the remaining sigma band frequencies included in the alpha band (8–13 Hz). Iranzo et al.14 tried to replicate the study by Fantini et al.13 and extended the analysis to untreated iRBD patients with mild cognitive impairment but used the same frequency bands. These authors only found, in REM sleep, small increases in power of the theta and beta2 bands in iRBD patients versus controls, in the C4 EEG channel, but a decrease in the beta1 band in the O1 lead. More significant changes were found for iRBD with mild cognitive impairment. Additionally, Sasai et al.27 found different results with iRBD patients (most of whom with mild cognitive impairment) showing decreased alpha and beta band power over the central and occipital regions in REM sleep. Massicotte-Marquez et al.26 did not analyze REM sleep but restricted their analysis to the whole NREM sleep and found an increase in the delta band (paralleled by the increase in the percentage of slow-wave sleep) in iRBD patients. Finally, O’Reilly et al.28 reported a lower spindle (automatically detected) density during NREM sleep in iRBD versus controls. Thus, the small differences reported, with marginal statistically significance, and the discrepancy between the studies indicate that a careful normalization of data is necessary when studying EEG signals as small in amplitude as those during REM sleep. Well-known sources of interindividual variability can disturb the analysis, such as those connected with the physical features of the volume conduction of EEG potentials (thickness of the skull and of other tissues), different between individuals and genders. A less important role can be foreseen for muscle potential contamination in normal controls, whereas it might assume a higher importance in iRBD patients with RSWA. For these reasons, we preferred to use relative spectral values and found that the drug-naïve iRBD patient delta band tended to be higher than that of controls in all sleep stages; clonazepam therapy was able to reduce significantly this band only during REM sleep (Figure 1). The increase in delta band in drug-naïve iRBD versus controls was paralleled by a decrease in the sigma band, statistically significant in all NREM sleep stages. Only during slow-wave sleep, also a significant decrease in the theta band was found in untreated iRBD patients, compared to controls. These data seem to confirm the already reported “slowing” of the EEG activity during wakefulness and sleep in RBD patients,13,14 and the relative decrease in sigma (or spindles)28 and theta bands might be considered, at least in part, the counterpart of the relative increase in delta band in drug-naïve iRBD. The differences between untreated iRBD and controls in the beta band were not statistically significant. Normal controls were also found to have a slightly less deep-sleep architecture than drug-free iRBD patients (higher number of awakenings, lower sleep efficiency, higher percentage of wakefulness after sleep onset, lower amount of slow-wave sleep but a lower number of sleep stage shifts than iRBD patients). This is in some agreement with previous reports of a relatively preserved sleep architecture in iRBD patients who have increased slow-wave sleep29 but also an increased sleep stage shift rate.30 Relative to drug-naïve iRBD subjects, patients taking clonazepam had significantly decreased delta in REM sleep, increased alpha in sleep stage 2 and slow-wave sleep, increased sigma in all NREM sleep stages, and increased beta in REM sleep only. These data seem to be in agreement with some previous findings indicating a major effect of this benzodiazepine on NREM sleep but only a minor effect on REM sleep in iRBD patients.17,18 Although specific data on the effects of clonazepam on the spectral content of the normal sleep EEG are lacking, it is known that almost all benzodiazepines have similar acute effects involving a reduction of EEG activities <10 Hz together with an enhancement of the sigma band during NREM sleep stage 2 and slow-wave sleep and of the beta band in REM sleep and stage 1.31,32 However, their chronic use, as in our case probably, is followed by less prominent changes that have been described to be detectable only in some part of the night (cycle).33 In normal controls, fast EEG activities during sleep (beta band) seem to show a behavior opposite to that of the delta band, reaching their maxima during REM sleep, when the slow-wave activity, as well as all other EEG bands with frequency below the beta range, show their minima.34–36 This effect of REM sleep on EEG was used in our study to achieve a reliable individual normalization of the highly interindividual variable power spectra, by calculating the degree by which REM sleep decreased or increased the power of each EEG band, taking the average value in sleep stage 2 as a reference. Sleep stage 2 has been found to be preserved in iRBD,18,26 (also in the present study). The normalized second-by-second REM EEG power spectra values obtained in this study allowed us to go further into a detailed analysis of the REM sleep EEG of drug-naïve iRBD patients and of the effects of clonazepam therapy. Overall, our results on the power ratio in REM sleep were well consistent with previous work on healthy volunteers.37 With our normalized data, we observed that the evident suppression of <15 Hz EEG bands seen in normal controls was clearly less pronounced in drug-naïve iRBD patients (Figure 2). The smaller EEG power REM suppression in the delta, alpha, and sigma bands in drug-naive iRBD patients was partially recovered in iRBD patients taking clonazepam. The beta band was increased by REM sleep in all groups, especially in both iRBD groups. This can be partially explained with muscle potential contamination by the RSWA in iRBD patients, as supported by the correlation between this band and the REM sleep atonia index (Supplemental Figure e-1). Moreover, it remained unchanged also in iRBD patients taking clonazepam that has been shown to have little, if any, effect on RSWA.17,18,22 These data seem to indicate that a smaller difference exists in iRBD patients between NREM and REM sleep which might indicate a generally weaker REM mechanism strength. Probably, our final observation of a higher variability of the normalized EEG power values in REM sleep of drug-naïve iRBD patients, which was clearly reduced in iRBD patients taking clonazepam (Figure 3), further supports the idea that weaker neurophysiological REM mechanisms are present in these patients that undergo fluctuations higher than those observed in controls and are reduced by clonazepam. Interestingly, a significant reduction in the variability measure was observed also for the beta band in patients taking clonazepam, suggesting that at least part of the beta band increase in REM sleep in drug-naïve iRBD might not be due only to the increased muscle activity—REM sleep atonia index has been reported to be unaffected by clonazepam,17,18 see also Table 1—but to a true brain activity. The REM sleep EEG instability reported here should not be confused with the decreased REM sleep stage stability in iRBD patients, indicated by an increased rate of sleep stage shifts, recently reported by Christensen et al.30 Descending REM-on glutamatergic neurons of the sublaterodorsal tegmental nucleus (SLD) exert a central role in generating atonia.38 Small bilateral lesions of the SLD result in RSWA and RBD-like behaviors in animals but do not seem to modify the amount of REM sleep; only larger lesions also affect the amount and duration of REM sleep.39,40 Two distinct populations of SLD neurons might control REM sleep atonia and REM sleep amount; however, REM sleep has usually been reported to be unaffected in iRBD, suggesting that only the SLD cells controlling REM sleep atonia might be involved. The results of the present study indicate, on the contrary, that even if the macroscopic aspects of REM sleep are grossly preserved, REM sleep EEG structure is indeed involved in iRBD. This involvement might be connected with the suggested activation of the cortex during REM sleep (including the motor cortex) exerted by excitatory intralaminar thalamocortical neurons, activated by ascending glutamatergic SLD neurons.38 The empirical notion that clonazepam is strongly effective in suppressing aggressive behaviors and oneiric contents, but not effective in restoring RSWA,17,18 is in line with our findings, suggesting that the therapeutic effect of clonazepam is probably exerted by acting on supratentorial rather than subtentorial networks, reducing the negative effects of the brainstem dysfunction on the supratentorial regions, without affecting the pathogenetic core of the disease. One limitation of this study should be mentioned because even if none of the subjects had an evident cognitive impairment, a mild cognitive impairment might have been present in some patients of both groups that might have affected, to some extent, some results. In conclusion, the REM sleep EEG structure changes found in this study disclose subtle but significant alterations in the cortical electrophysiology of iRBD that might represent the early expression of the supposed neurodegenerative processes already taking place at this stage of the disease6,41,42 and might be the target of better and effective future therapeutic strategies for this condition. SUPPLEMENTARY MATERIAL Supplementary material is available at SLEEP online. Funding This work (RF) was partially supported by a grant of the Italian Ministry of Health (“Ricerca Corrente”). MM is supported by the Swiss National Science Foundation (grant no. 320030_144007). DISCLOSURE STATEMENT OB reports consulting for Sapio Life. GP has consulted for UCB Pharma, Jazz, and Bioproject. MM has received a grant from Vifor. The remaining authors report no conflict of interest. REFERENCES 1. American Academy of Sleep Medicine , eds. International classification of sleep disorders , 3rd ed. Darien, IL : American Academy of Sleep Medicine , 2014 . WorldCat COPAC 2. Schenck CH Bundlie SR Ettinger MG Mahowald MW . Chronic behavioral disorders of human REM sleep: a new category of parasomnia . Sleep . 1986 ; 9 ( 2 ): 293 – 308 . Google Scholar Crossref Search ADS PubMed WorldCat 3. Schenck CH Mahowald MW . REM sleep behavior disorder: clinical, developmental, and neuroscience perspectives 16 years after its formal identification in SLEEP . Sleep . 2002 ; 25 ( 2 ): 120 – 138 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Boeve BF Saper CB . REM sleep behavior disorder: a possible early marker for synucleinopathies . Neurology . 2006 ; 66 ( 6 ): 796 – 797 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Schenck CH Boeve BF Mahowald MW . Delayed emergence of a parkinsonian disorder or dementia in 81% of older men initially diagnosed with idiopathic rapid eye movement sleep behavior disorder: a 16-year update on a previously reported series . Sleep Med . 2013 ; 14 ( 8 ): 744 – 748 . Google Scholar Crossref Search ADS PubMed WorldCat 6. Iranzo A Tolosa E Gelpi E et al. Neurodegenerative disease status and post-mortem pathology in idiopathic rapid-eye-movement sleep behaviour disorder: an observational cohort study . Lancet Neurol . 2013 ; 12 ( 5 ): 443 – 453 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Iranzo A Santamaría J Rye DB et al. Characteristics of idiopathic REM sleep behavior disorder and that associated with MSA and PD . Neurology . 2005 ; 65 ( 2 ): 247 – 252 . Google Scholar Crossref Search ADS PubMed WorldCat 8. Aurora RN Zak RS Maganti RK et al. ; Standards of Practice Committee; American Academy of Sleep Medicine . Best practice guide for the treatment of REM sleep behavior disorder (RBD) . J Clin Sleep Med . 2010 ; 6 ( 1 ): 85 – 95 . Google Scholar PubMed WorldCat 9. Schenck CH Mahowald MW . Long-term, nightly benzodiazepine treatment of injurious parasomnias and other disorders of disrupted nocturnal sleep in 170 adults . Am J Med . 1996 ; 100 ( 3 ): 333 – 337 . Google Scholar Crossref Search ADS PubMed WorldCat 10. Gagnon JF Postuma RB Montplaisir J . Update on the pharmacology of REM sleep behavior disorder . Neurology . 2006 ; 67 ( 5 ): 742 – 747 . Google Scholar Crossref Search ADS PubMed WorldCat 11. Gugger JJ Wagner ML . Rapid eye movement sleep behavior disorder . Ann Pharmacother . 2007 ; 41 ( 11 ): 1833 – 1841 . Google Scholar Crossref Search ADS PubMed WorldCat 12. Schenck CH Montplaisir JY Frauscher B et al. Rapid eye movement sleep behavior disorder: devising controlled active treatment studies for symptomatic and neuroprotective therapy—a consensus statement from the International Rapid Eye Movement Sleep Behavior Disorder Study Group . Sleep Med . 2013 ; 14 ( 8 ): 795 – 806 . Google Scholar Crossref Search ADS PubMed WorldCat 13. Fantini ML Gagnon JF Petit D et al. Slowing of electroencephalogram in rapid eye movement sleep behavior disorder . Ann Neurol . 2003 ; 53 ( 6 ): 774 – 780 . Google Scholar Crossref Search ADS PubMed WorldCat 14. Iranzo A Isetta V Molinuevo JL et al. Electroencephalographic slowing heralds mild cognitive impairment in idiopathic REM sleep behavior disorder . Sleep Med . 2010 ; 11 ( 6 ): 534 – 539 . Google Scholar Crossref Search ADS PubMed WorldCat 15. Ferri R Manconi M Plazzi G et al. A quantitative statistical analysis of the submentalis muscle EMG amplitude during sleep in normal controls and patients with REM sleep behavior disorder . J Sleep Res . 2008 ; 17 ( 1 ): 89 – 100 . Google Scholar Crossref Search ADS PubMed WorldCat 16. Ferri R Rundo F Manconi M et al. Improved computation of the atonia index in normal controls and patients with REM sleep behavior disorder . Sleep Med . 2010 ; 11 ( 9 ): 947 – 949 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Ferri R Zucconi M Marelli S Plazzi G Schenck CH Ferini-Strambi L . Effects of long-term use of clonazepam on nonrapid eye movement sleep patterns in rapid eye movement sleep behavior disorder . Sleep Med . 2013 ; 14 ( 5 ): 399 – 406 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Ferri R Marelli S Ferini-Strambi L et al. An observational clinical and video-polysomnographic study of the effects of clonazepam in REM sleep behavior disorder . Sleep Med . 2013 ; 14 ( 1 ): 24 – 29 . Google Scholar Crossref Search ADS PubMed WorldCat 19. Ferri R Marelli S Cosentino FI Rundo F Ferini-Strambi L Zucconi M . Night-to-night variability of automatic quantitative parameters of the chin EMG amplitude (Atonia Index) in REM sleep behavior disorder . J Clin Sleep Med . 2013 ; 9 ( 3 ): 253 – 258 . Google Scholar PubMed WorldCat 20. Ferri R Bruni O Fulda S Zucconi M Plazzi G . A quantitative analysis of the submentalis muscle electromyographic amplitude during rapid eye movement sleep across the lifespan . J Sleep Res . 2012 ; 21 ( 3 ): 257 – 263 . Google Scholar Crossref Search ADS PubMed WorldCat 21. Rechtschaffen A Kales A , eds. A Manual of Standardized Terminology, Techniques, and Scoring System for Sleep Stages of Human Subjects . Washington : Washington Public Health service; US Government Printing Office , 1968 . Google Preview WorldCat COPAC 22. Lapierre O Montplaisir J . Polysomnographic features of REM sleep behavior disorder: development of a scoring method . Neurology . 1992 ; 42 ( 7 ): 1371 – 1374 . Google Scholar Crossref Search ADS PubMed WorldCat 23. Dauvilliers Y Rompré S Gagnon JF Vendette M Petit D Montplaisir J . REM sleep characteristics in narcolepsy and REM sleep behavior disorder . Sleep . 2007 ; 30 ( 7 ): 844 – 849 . Google Scholar Crossref Search ADS PubMed WorldCat 24. Ferri R Gagnon JF Postuma RB Rundo F Montplaisir JY . Comparison between an automatic and a visual scoring method of the chin muscle tone during rapid eye movement sleep . Sleep Med . 2014 ; 15 ( 6 ): 661 – 665 . Google Scholar Crossref Search ADS PubMed WorldCat 25. Zucconi M Ferri R Allen R et al. ; International Restless Legs Syndrome Study Group (IRLSSG) . The official World Association of Sleep Medicine (WASM) standards for recording and scoring periodic leg movements in sleep (PLMS) and wakefulness (PLMW) developed in collaboration with a task force from the International Restless Legs Syndrome Study Group (IRLSSG) . Sleep Med . 2006 ; 7 ( 2 ): 175 – 183 . Google Scholar Crossref Search ADS PubMed WorldCat 26. Massicotte-Marquez J Carrier J Décary A et al. Slow-wave sleep and delta power in rapid eye movement sleep behavior disorder . Ann Neurol . 2005 ; 57 ( 2 ): 277 – 282 . Google Scholar Crossref Search ADS PubMed WorldCat 27. Sasai T Matsuura M Inoue Y . Electroencephalographic findings related with mild cognitive impairment in idiopathic rapid eye movement sleep behavior disorder . Sleep . 2013 ; 36 ( 12 ): 1893 – 1899 . Google Scholar Crossref Search ADS PubMed WorldCat 28. O’Reilly C Godin I Montplaisir J Nielsen T . REM sleep behaviour disorder is associated with lower fast and higher slow sleep spindle densities . J Sleep Res . 2015 ; 24 ( 6 ): 593 – 601 . Google Scholar Crossref Search ADS PubMed WorldCat 29. Schenck CH Callies AL Mahowald MW . Increased percentage of slow-wave sleep in REM sleep behavior disorder (RBD): a reanalysis of previously published data from a controlled study of RBD reported in SLEEP . Sleep . 2003 ; 26 ( 8 ): 1066; author reply 1067 . Google Scholar PubMed WorldCat 30. Christensen JA Jennum P Koch H et al. Sleep stability and transitions in patients with idiopathic REM sleep behavior disorder and patients with Parkinson’s disease . Clin Neurophysiol . 2016 ; 127 ( 1 ): 537 – 543 . Google Scholar Crossref Search ADS PubMed WorldCat 31. Borbély AA Mattmann P Loepfe M Strauch I Lehmann D . Effect of benzodiazepine hypnotics on all-night sleep EEG spectra . Hum Neurobiol . 1985 ; 4 ( 3 ): 189 – 194 . Google Scholar PubMed WorldCat 32. Tan X Uchida S Matsuura M Nishihara K Kojima T . Long-, intermediate- and short-acting benzodiazepine effects on human sleep EEG spectra . Psychiatry Clin Neurosci . 2003 ; 57 ( 1 ): 97 – 104 . Google Scholar Crossref Search ADS PubMed WorldCat 33. Bastien CH LeBlanc M Carrier J Morin CM . Sleep EEG power spectra, insomnia, and chronic use of benzodiazepines . Sleep . 2003 ; 26 ( 3 ): 313 – 317 . Google Scholar Crossref Search ADS PubMed WorldCat 34. Aeschbach D Borbély AA . All-night dynamics of the human sleep EEG . J Sleep Res . 1993 ; 2 ( 2 ): 70 – 81 . Google Scholar Crossref Search ADS PubMed WorldCat 35. Ferri R Cosentino FI Elia M Musumeci SA Marinig R Bergonzi P . Relationship between Delta, Sigma, Beta, and Gamma EEG bands at REM sleep onset and REM sleep end . Clin Neurophysiol . 2001 ; 112 ( 11 ): 2046 – 2052 . Google Scholar Crossref Search ADS PubMed WorldCat 36. Ferri R Elia M Musumeci SA Pettinato S . The time course of high-frequency bands (15–45 Hz) in all-night spectral analysis of sleep EEG . Clin Neurophysiol . 2000 ; 111 ( 7 ): 1258 – 1265 . Google Scholar Crossref Search ADS PubMed WorldCat 37. Tinguely G Finelli LA Landolt HP Borbély AA Achermann P . Functional EEG topography in sleep and waking: state-dependent and state-independent features . Neuroimage . 2006 ; 32 ( 1 ): 283 – 292 . Google Scholar Crossref Search ADS PubMed WorldCat 38. Peever J Luppi PH Montplaisir J . Breakdown in REM sleep circuitry underlies REM sleep behavior disorder . Trends Neurosci . 2014 ; 37 ( 5 ): 279 – 288 . Google Scholar Crossref Search ADS PubMed WorldCat 39. Boissard R Gervasoni D Schmidt MH Barbagli B Fort P Luppi PH . The rat ponto-medullary network responsible for paradoxical sleep onset and maintenance: a combined microinjection and functional neuroanatomical study . Eur J Neurosci . 2002 ; 16 ( 10 ): 1959 – 1973 . Google Scholar Crossref Search ADS PubMed WorldCat 40. Lu J Sherman D Devor M Saper CB . A putative flip-flop switch for control of REM sleep . Nature . 2006 ; 441 ( 7093 ): 589 – 594 . Google Scholar Crossref Search ADS PubMed WorldCat 41. Postuma RB Gagnon JF Montplaisir JY . REM sleep behavior disorder: from dreams to neurodegeneration . Neurobiol Dis . 2012 ; 46 ( 3 ): 553 – 558 . Google Scholar Crossref Search ADS PubMed WorldCat 42. Ferini-Strambi L . Does idiopathic REM sleep behavior disorder (iRBD) really exist? What are the potential markers of neurodegeneration in iRBD? Sleep Med . 2011 ; 12 Suppl 2 : S43 – S49 . Google Scholar Crossref Search ADS PubMed WorldCat Author notes Address correspondence to: Raffaele Ferri, MD, Sleep Research Centre, Department of Neurology I.C., Oasi Institute (IRCCS), Troina, Italy. Telephone: 39-0935-936111; Fax 39-0935-936694; Email: [email protected] © Sleep Research Society 2017. Published by Oxford University Press on behalf of the Sleep Research Society. All rights reserved. For permissions, please e-mail [email protected].

Palma,, Jose-Alberto;Norcliffe-Kaufmann,, Lucy;Perez, Miguel, A;Spalink, Christy, L;Kaufmann,, Horacio

doi: 10.1093/sleep/zsx083pmid: 28521050

Abstract Study Objectives Sudden unexpected death during sleep (SUDS) is the most common cause of death in patients with familial dysautonomia (FD), an autosomal recessive disease characterized by sensory and autonomic dysfunction. It remains unknown what causes SUDS in these patients and who is at highest risk. We tested the hypothesis that SUDS in FD is linked to sleep-disordered breathing. Methods We retrospectively identified patients with FD who died suddenly and unexpectedly during sleep and had undergone polysomnography within the 18-month period before death. For each case, we sampled one age-matched surviving subject with FD that had also undergone polysomnography within the 18-month period before study. Data on polysomnography, EKG, ambulatory blood pressure monitoring, arterial blood gases, blood count, and metabolic panel were analyzed. Results Thirty-two deceased cases and 31 surviving controls were included. Autopsy was available in six cases. Compared with controls, participants with SUDS were more likely to be receiving treatment with fludrocortisone (odds ratio [OR]; 95% confidence interval) (OR 29.7; 4.1–213.4), have untreated obstructive sleep apnea (OR 17.4; 1.5–193), and plasma potassium levels <4 mEq/L (OR 19.5; 2.36–161) but less likely to use noninvasive ventilation at night (OR 0.19; 0.06–0.61). Conclusions Initiation of noninvasive ventilation when required and discontinuation of fludrocortisone treatment may reduce the high incidence rate of SUDS in patients with FD. Our findings contribute to the understanding of the link between autonomic, cardiovascular, and respiratory risk factors in SUDS. autonomic nervous system, hereditary sensory and autonomic neuropathy, hypokalemia, sleep apnea, noninvasive ventilation Statement of Significance Sudden unexpected death during sleep (SUDS) is the most common cause of death in patients with familial dysautonomia (FD), although, until now, it was unknown who was at highest risk. This case–control study including polysomnography and arterial blood gases data shows that the presence of untreated obstructive sleep apnea, fludrocortisone treatment, and potassium levels in the low range of normality were independently associated with increased risk for sudden death during sleep in FD. In contrast, treatment with noninvasive ventilation at night was independently linked with reduced risk for SUDS. Initiation of noninvasive ventilation when required and discontinuation of fludrocortisone treatment may reduce the high incidence rate of SUDS in FD. INTRODUCTION Sudden unexpected death during sleep (SUDS) is the most common cause of death in patients with familial dysautonomia (FD, Riley–Day syndrome, hereditary sensory, and autonomic neuropathy type III), a rare autosomal recessive disorder first described in 1949 in children of Jewish Ashkenazi ancestry.1,2 The disease is due to a founder mutation in the IkB kinase-associated protein gene (IKBKAP)3–5 causing impaired development of sensory and afferent autonomic nerves.6 Hallmarks of FD include impaired pain and temperature sensation, absent deep tendon reflexes, gait ataxia,7 chronic lung disease,8 and afferent baroreflex failure leading to orthostatic hypotension and paroxysmal hypertension,9,10 all which contribute to morbidity and mortality.2,11 It remains unknown what causes SUDS in FD patients and who is at highest risk. One of the potential risk factors for SUDS is sleep-disordered breathing, present in most patients with FD.12–14 Ventilatory responses to hypercapnia are reduced and to hypoxia are almost absent in all FD patients. Thus, in response to hypoxia, patients develop paradoxical hypotension, hypoventilation, bradycardia, and potentially, death.12,15–18 We hypothesized that the high incidence of SUDS in patients with FD might be linked to respiratory abnormalities during sleep. To test this hypothesis, we analyzed the clinical features and polysomnography findings of patients with FD who died suddenly during sleep and compared them to age- and gender-matched FD patients who remained alive at the time of the study. To expand our results, we also included data from electrocardiograms, arterial blood gases, and blood metabolic panels obtained during daytime. METHODS Study Design We conducted a case–control study of patients with genetically confirmed FD. Cases were defined as patients with FD who died suddenly and unexpectedly during sleep. Controls were defined as patients with FD who remained alive at the time of the study. Cases and controls were drawn from the New York University (NYU) FD Registry. We followed the Strengthening the Reporting of Observational Studies in Epidemiology guidelines for case–control studies.19 NYU FD Dysautonomia Registry The NYU FD Registry is an ongoing, prospective registry of patients FD. The Registry started in 1970 and contains clinical and diagnostic data, including cause of death, on 670 patients at the time of the study. Of these, 327 (49%) remained alive at the time of the study. All patients have genetically confirmed FD; more than 99% are homozygous for the same mutation (6T>C change) in the IKBKAP gene. The majority of patients included in the Registry (52%) are from the United States. The remaining patients are from Israel, Canada, United Kingdom, South America, South Africa, Australia, and Belgium. Patients are followed closely and seen at least once a year. Cases Of all patients included in the NYU FD Registry during a 45-year period (1970–2015), we first extracted those individuals in whom SUDS was listed as the cause of death. SUDS was defined as a sudden, unexpected, witnessed or unwitnessed, nontraumatic, and nondrowning death occurring during sleep, with or without evidence of a seizure. Cases in which postmortem examinations were not performed—and therefore no toxicological or anatomical cause of death could be ruled out—were classified as “probable SUDS.”20 Patients with signs or symptoms of respiratory infection or fever at the time of death were excluded. Of all patients with probable SUDS, we then selected those who had a full hospital polysomnography performed in a period no longer than 18-months before death. Controls For each case of SUDS, we selected one age-matched control patient (age at the time of death). Controls were patients with FD included in the NYU FD Registry that were alive at the time of study and had undergone a polysomnography within the 18-month period before this study. Data Coll ection The following test results (continuous data) were obtained: Polysomnography variables comprised the apnea-hypopnea index (AHI) including obstructive and central apneas, the minimal (nadir) oxygen saturation during sleep, and the time of sleep with an oxygen saturation below 90%. Because all patients also had electrocardiogram (EKG), metabolic panel, complete blood count, and arterial blood gases performed in the same time period, the following parameters were extracted: heart rate, QRS interval, QTc interval, PR interval, hemoglobin, hematocrit, glucose, electrolytes, creatinine, blood urea nitrogen, pO2, pCO2, HCO3, and pH. The vast majority of participants also had ambulatory blood pressure (BP) monitoring, in which BP readings are recorded every 30 minutes during 24 hours;21 we obtained the standard deviation of the systolic BP, as a marker of cardiovascular autonomic variability.22 The dosages of the most frequent medications taken by these patients (fludrocortisone, midodrine, benzodiazepines, and clonidine) were also included in the analysis. The following clinical (categorical) data were obtained: presence of obstructive sleep apnea (OSA; defined as an AHI >5 events/hour); history of epileptic, hypoxic, or febrile seizures; presence of active epilepsy (defined as participants who were currently taking medication to control epilepsy, or had one or more epileptic seizures in the past year, or both); cardiovascular risk factors (hypertension, hypercholesterolemia, diabetes mellitus, stroke/transient ischemic attack) and heart disease (ischemic, valvular, or congenital/inherited heart disease and cardiac arrhythmia); presence of pacemaker; treatment history (midodrine, fludrocortisone, clonidine, benzodiazepines); and use of noninvasive ventilation such as continuous positive airway pressure or bilevel positive airway pressure. Because most patients with FD undergo Nissen fundoplication surgery and percutaneous gastrostomy tube placement for the management of vomiting episodes and neurogenic dysphagia, these were also included in the analysis. The frequency of nocturnal feedings (via the gastrostomy tube) was also analyzed. We made an effort to collect information on the circumstances of death by exhaustive medical record review, coroner’s office report, or by formal interview with the bereaved relatives. Autopsy findings were also documented, although postmortem examinations are seldom performed in this population due to religious reasons. Statistical Analysis Patient and event characteristics were described and compared between cases and controls using χ2 statistics (Pearson/Fisher exact test where appropriate) for categorical data and the Student t test/Mann–Whitney U test for quantitative continuous data. Missing quantitative data were handled with multiple imputation methods. If significant differences were found, these quantitative variables were transformed into categorical ones. Then, to identify categorical risk factors for SUDS (cases vs. controls), univariable and multivariable conditional logistic regression was employed, thereby accounting for matched data. The interaction term “with/without noninvasive ventilation” was used in the multivariable analysis when accounting for the risk factor OSA. p-Values <.05 were considered significant. The annual incidence rate was calculated according the formula d/y × 10,000, where d is the number of SUDS cases and y is the number of people at risk (number of subjects × observation period). Analyses were performed with SPSS version 18.0 (SPSS Inc., Chicago, USA). Protocol Approval and Patient Consents The Institutional Review Board of the NYU School of Medicine approved the NYU FD Registry. Written informed consent was obtained from all FD patients (or their guardians when appropriate) at the time of enrolment in the Registry. RESULTS Cases and Controls At the time of study, 669 patients with FD had been included in the NYU FD Registry. Of these, 343 (51%) had deceased. Of the deceased, 108 (31%) fulfilled criteria for probable SUDS. Therefore, the incidence rate of SUDS in FD is 3.4 per 1000 person-year. Of the 108 patients fulfilling criteria for probable SUDS, 32 (14 women) had undergone polysomnography in the 18-month period before death and were included in the final analysis (Table 1). SUDS occurred most frequently during the second and third decades of life (mean age at death was 29.3 ± 12.4 years old). Table 1 Characteristics of Cases With Sudden Unexpected Death During Sleep (n &#x003D; 32). Characteristics No (%) Women 14 (44) Age  0–10 2 (6)  11–17 7 (22)  18–30 9 (28)  31–50 14 (44)  >51 0 (0) History of seizures 26 (81)  Epileptic 11 (34)  Hypoxic 10 (31)  Hyponatremic 5 (16)  Febrile 1 (3) Active epilepsy 5 (16) Cardiac disease# 6 (19) Pacemaker 6 (19) Stroke/transient ischemic attack 0 (0) Diabetes mellitus 0 (0) Treatments  Benzodiazepines 24 (75)  Fludrocortisone 24 (75)   <0.1 mg/day 0   0.1–0.15 mg/day 7 (22)   0.2–0.3 mg/day 16 (50)   >0.3 mg/day 1 (3)  Clonidine 12 (38)  Midodrine 10 (31)  Fludrocortisone and midodrine 12 (38)  Beta-blockers 1 (3)  Noninvasive ventilation at night$ 3 (9) Autopsy 6 (19) Characteristics No (%) Women 14 (44) Age  0–10 2 (6)  11–17 7 (22)  18–30 9 (28)  31–50 14 (44)  >51 0 (0) History of seizures 26 (81)  Epileptic 11 (34)  Hypoxic 10 (31)  Hyponatremic 5 (16)  Febrile 1 (3) Active epilepsy 5 (16) Cardiac disease# 6 (19) Pacemaker 6 (19) Stroke/transient ischemic attack 0 (0) Diabetes mellitus 0 (0) Treatments  Benzodiazepines 24 (75)  Fludrocortisone 24 (75)   <0.1 mg/day 0   0.1–0.15 mg/day 7 (22)   0.2–0.3 mg/day 16 (50)   >0.3 mg/day 1 (3)  Clonidine 12 (38)  Midodrine 10 (31)  Fludrocortisone and midodrine 12 (38)  Beta-blockers 1 (3)  Noninvasive ventilation at night$ 3 (9) Autopsy 6 (19) #Cardiac disease includes ischemic, vascular, congenital/inherited disease, atrial fibrillation, and third-degree atrioventricular block. $Noninvasive ventilation at night includes continuous positive air pressure (CPAP) and bilevel positive airway pressure. Open in new tab Table 1 Characteristics of Cases With Sudden Unexpected Death During Sleep (n &#x003D; 32). Characteristics No (%) Women 14 (44) Age  0–10 2 (6)  11–17 7 (22)  18–30 9 (28)  31–50 14 (44)  >51 0 (0) History of seizures 26 (81)  Epileptic 11 (34)  Hypoxic 10 (31)  Hyponatremic 5 (16)  Febrile 1 (3) Active epilepsy 5 (16) Cardiac disease# 6 (19) Pacemaker 6 (19) Stroke/transient ischemic attack 0 (0) Diabetes mellitus 0 (0) Treatments  Benzodiazepines 24 (75)  Fludrocortisone 24 (75)   <0.1 mg/day 0   0.1–0.15 mg/day 7 (22)   0.2–0.3 mg/day 16 (50)   >0.3 mg/day 1 (3)  Clonidine 12 (38)  Midodrine 10 (31)  Fludrocortisone and midodrine 12 (38)  Beta-blockers 1 (3)  Noninvasive ventilation at night$ 3 (9) Autopsy 6 (19) Characteristics No (%) Women 14 (44) Age  0–10 2 (6)  11–17 7 (22)  18–30 9 (28)  31–50 14 (44)  >51 0 (0) History of seizures 26 (81)  Epileptic 11 (34)  Hypoxic 10 (31)  Hyponatremic 5 (16)  Febrile 1 (3) Active epilepsy 5 (16) Cardiac disease# 6 (19) Pacemaker 6 (19) Stroke/transient ischemic attack 0 (0) Diabetes mellitus 0 (0) Treatments  Benzodiazepines 24 (75)  Fludrocortisone 24 (75)   <0.1 mg/day 0   0.1–0.15 mg/day 7 (22)   0.2–0.3 mg/day 16 (50)   >0.3 mg/day 1 (3)  Clonidine 12 (38)  Midodrine 10 (31)  Fludrocortisone and midodrine 12 (38)  Beta-blockers 1 (3)  Noninvasive ventilation at night$ 3 (9) Autopsy 6 (19) #Cardiac disease includes ischemic, vascular, congenital/inherited disease, atrial fibrillation, and third-degree atrioventricular block. $Noninvasive ventilation at night includes continuous positive air pressure (CPAP) and bilevel positive airway pressure. Open in new tab Detailed circumstances of death were documented in 25 cases. Subjects were found dead at night (between 01.00 am and 05.00 am) in 14 cases, in the morning (between 08.00 am and 12.00 pm) in six cases, and in the afternoon or evening (between 05.00 pm and 11.00 pm) in five cases. All cases were presumed to be sleeping at the time of death. All cases (except one who was found sitting upright after a long car travel) were found lying in bed (15 in the supine position, seven in the prone position and two lying on their side). None was found partially or completely fallen off the bed. None of the cases were found with their head bent forward or other position suggesting compromised breathing. In none of the cases gagging or choking sounds were identified around the event. None of the cases had their tongue or lips bitten although there was urinary incontinence in one case. Postmortem studies were available in six cases. All of them showed brainstem, spinal cord, and dorsal root ganglia atrophy, which are pathological hallmarks of familial dysautonomia.23,24 These six cases showed no structural cardiac pathology and no structural brain lesions. Most of them had nonspecific pulmonary congestion or focal hemorrhage. These patients were therefore classified as definite SUDS (Supplementary Table 1). We included 31 age-matched surviving patients (15 women) with a recent polysomnography as controls. Characteristics of Cases and Controls Characteristics were analyzed in the 32 cases and 31 controls. AHI (p = .007) and the total sleep time spent with oxygen saturation below 90% were higher in cases than in controls (p = .014). In all cases and controls, apneas and hypopneas were predominantly obstructive. Minimum oxygen saturation during sleep (p = .029) and plasma potassium levels (p<.0001) were lower in cases than in controls. No patient had hypokalemia (<3.5 mEq/L) or hyperkalemia (>5.5 mEq/L). Cases were receiving a higher daily dose of fludrocortisone than controls (p<.0001). There were no differences in the EKG, arterial blood gases, BP variability, or complete blood count results (Table 2). Table 2 Comparison of Continuous Characteristics in Cases and Controls. Characteristic Cases (n = 32) Controls (n = 31) p-value Mean (SD) Range Mean (SD) Range Age, year 29.2 (12.5) 9–50 28.9 (11.7) 9–54 .99 Body mass index, kg/m2 17.9 (2.8) 13.7–23.9 17.2 (2.1) 14.4–22.2 .81 Polysomnography  Total sleep time, minutes 392 (145) 181–491 404 (101) 214–523 .78  REM latency, minutes 110 (59) 0–308 119 (61) 16–321 .81  Arousal index, events/hour 28.1 (16.1) 1.2–52.5 24.9 (18) 2–74.4 .38  Stage 1 sleep, % 7.5 (4.5) 1.8–24 5.28 (4.1) 0–17 .19  Stage 2 sleep, % 50.7 (13.2) 35.5–79.2 48.7 (25.1) 0–79 .55  Stage 3 sleep, % 19.1 (12.9) 0–38.3 25.1 (12.5) 0–90 .42  REM sleep, % 19.0 (11.4) 0–20.5 20.6 (10.1) 8.6–44 .63  Apnea-hypopnea index, events/hour 16.7 (14.4) 0–59.3 6.7 (4.2) 0–23.4 .007*  Apnea-hypopnea average duration, seconds 29.8 (7.2) 0–38 22.2 (11.8) 0–40 .73  Minimal O2 saturation, % 69 (18.9) 30–97 81 (8.4) 65–93 .029*  Total sleep time with SpO2 <90%, % 9.1 (15.4) 0.1–95 2.3 (4.1) 0–16.8 .014* EKG  Heart rate, bpm 83.5 (10) 68–107 82.5 (12) 59–105 .71  QRS interval, ms 80.6 (17.9) 48–100 89.2 (21.8) 50–160 .06  QT interval, ms 345 (27) 300–388 360 (42) 284–460 .38  QTc interval, ms 404 (11) 380–427 419 (29) 375–482 .12  PR interval, ms 121 (26) 80–160 133 (18) 96–162 .07 Standard deviation of 24-hour systolic blood pressure, mmHg 24.1 (6.5) 17.3–31 23.1 (5.1) 14.2–29.9 .60 Arterial blood gases  pH 7.41 (0.05) 7.34–7.51 7.40 (0.03) 7.33–7.46 .27  pO2, mmHg 78.5 (14.2) 52.5–107 83.5 (17.8) 50–105 .16  pCO2, mmHg 42.1 (7.65) 23–59 44.3 (5.8) 35–64 .25  pHCO3, mEq/L 25.1 (4.6) 15–31.1 27.3 (2.6) 24.3–35.6 .18 Hematological parameters  White blood cell count, ×109/L 7.55 (3.54) 4.3–10.2 7.32 (2.35) 4–13.7 .92  Red blood cell count, ×109/L 4.26 (0.47) 3.3–5.5 4.11 (0.6) 2.7–5.1 .40  Hemoglobin, mg/dl 12.75 (1.29) 10.3–15.2 15.68 (20.53) 9.8–14.8 .43  Hematocrit, % 37.61 (4.11) 29.8–46.6 36.41 (4.37) 26.7–44.3 .33 Metabolic parameters  Sodium, mEq/L 140 (4.3) 134–147 139 (3.3) 132–144 .31  Potassium, mEq/L 4.3 (0.6) 3.3–6.2 4.7 (0.5) 4.1–5.9 <.0001*  Chloride, mEq/L 102 (4.3) 88–109 100 (3.9) 91–107 .20  Creatinine, mg/dL 1 (0.43) 0.3–2.4 1.05 (1.08) 0.4–1.6 .43  Blood urea nitrogen, mg/dL 23.3 (9.5) 3–40 24.1 (11.5) 13–38 .51 Medications  Fludrocortisone, mg/day 0.14 (0.10) 0–0.4 0.02 (0.04) 0–0.1 <.0001*  Midodrine, mg/day 3 (4.7) 0–17.5 3.3 (4.9) 0–15 .66  Benzodiazepines, mg/day 4.8 (4.7) 0–17.5 4.9 (4.6) 0–16 .78  Clonidine, mg/day 0.04 (0.07) 0–0.2 0.06 (0.06) 0–0.2 .41 Characteristic Cases (n = 32) Controls (n = 31) p-value Mean (SD) Range Mean (SD) Range Age, year 29.2 (12.5) 9–50 28.9 (11.7) 9–54 .99 Body mass index, kg/m2 17.9 (2.8) 13.7–23.9 17.2 (2.1) 14.4–22.2 .81 Polysomnography  Total sleep time, minutes 392 (145) 181–491 404 (101) 214–523 .78  REM latency, minutes 110 (59) 0–308 119 (61) 16–321 .81  Arousal index, events/hour 28.1 (16.1) 1.2–52.5 24.9 (18) 2–74.4 .38  Stage 1 sleep, % 7.5 (4.5) 1.8–24 5.28 (4.1) 0–17 .19  Stage 2 sleep, % 50.7 (13.2) 35.5–79.2 48.7 (25.1) 0–79 .55  Stage 3 sleep, % 19.1 (12.9) 0–38.3 25.1 (12.5) 0–90 .42  REM sleep, % 19.0 (11.4) 0–20.5 20.6 (10.1) 8.6–44 .63  Apnea-hypopnea index, events/hour 16.7 (14.4) 0–59.3 6.7 (4.2) 0–23.4 .007*  Apnea-hypopnea average duration, seconds 29.8 (7.2) 0–38 22.2 (11.8) 0–40 .73  Minimal O2 saturation, % 69 (18.9) 30–97 81 (8.4) 65–93 .029*  Total sleep time with SpO2 <90%, % 9.1 (15.4) 0.1–95 2.3 (4.1) 0–16.8 .014* EKG  Heart rate, bpm 83.5 (10) 68–107 82.5 (12) 59–105 .71  QRS interval, ms 80.6 (17.9) 48–100 89.2 (21.8) 50–160 .06  QT interval, ms 345 (27) 300–388 360 (42) 284–460 .38  QTc interval, ms 404 (11) 380–427 419 (29) 375–482 .12  PR interval, ms 121 (26) 80–160 133 (18) 96–162 .07 Standard deviation of 24-hour systolic blood pressure, mmHg 24.1 (6.5) 17.3–31 23.1 (5.1) 14.2–29.9 .60 Arterial blood gases  pH 7.41 (0.05) 7.34–7.51 7.40 (0.03) 7.33–7.46 .27  pO2, mmHg 78.5 (14.2) 52.5–107 83.5 (17.8) 50–105 .16  pCO2, mmHg 42.1 (7.65) 23–59 44.3 (5.8) 35–64 .25  pHCO3, mEq/L 25.1 (4.6) 15–31.1 27.3 (2.6) 24.3–35.6 .18 Hematological parameters  White blood cell count, ×109/L 7.55 (3.54) 4.3–10.2 7.32 (2.35) 4–13.7 .92  Red blood cell count, ×109/L 4.26 (0.47) 3.3–5.5 4.11 (0.6) 2.7–5.1 .40  Hemoglobin, mg/dl 12.75 (1.29) 10.3–15.2 15.68 (20.53) 9.8–14.8 .43  Hematocrit, % 37.61 (4.11) 29.8–46.6 36.41 (4.37) 26.7–44.3 .33 Metabolic parameters  Sodium, mEq/L 140 (4.3) 134–147 139 (3.3) 132–144 .31  Potassium, mEq/L 4.3 (0.6) 3.3–6.2 4.7 (0.5) 4.1–5.9 <.0001*  Chloride, mEq/L 102 (4.3) 88–109 100 (3.9) 91–107 .20  Creatinine, mg/dL 1 (0.43) 0.3–2.4 1.05 (1.08) 0.4–1.6 .43  Blood urea nitrogen, mg/dL 23.3 (9.5) 3–40 24.1 (11.5) 13–38 .51 Medications  Fludrocortisone, mg/day 0.14 (0.10) 0–0.4 0.02 (0.04) 0–0.1 <.0001*  Midodrine, mg/day 3 (4.7) 0–17.5 3.3 (4.9) 0–15 .66  Benzodiazepines, mg/day 4.8 (4.7) 0–17.5 4.9 (4.6) 0–16 .78  Clonidine, mg/day 0.04 (0.07) 0–0.2 0.06 (0.06) 0–0.2 .41 *Statistically significant differences. REM = rapid eye movement; SD = standard deviation. Open in new tab Table 2 Comparison of Continuous Characteristics in Cases and Controls. Characteristic Cases (n = 32) Controls (n = 31) p-value Mean (SD) Range Mean (SD) Range Age, year 29.2 (12.5) 9–50 28.9 (11.7) 9–54 .99 Body mass index, kg/m2 17.9 (2.8) 13.7–23.9 17.2 (2.1) 14.4–22.2 .81 Polysomnography  Total sleep time, minutes 392 (145) 181–491 404 (101) 214–523 .78  REM latency, minutes 110 (59) 0–308 119 (61) 16–321 .81  Arousal index, events/hour 28.1 (16.1) 1.2–52.5 24.9 (18) 2–74.4 .38  Stage 1 sleep, % 7.5 (4.5) 1.8–24 5.28 (4.1) 0–17 .19  Stage 2 sleep, % 50.7 (13.2) 35.5–79.2 48.7 (25.1) 0–79 .55  Stage 3 sleep, % 19.1 (12.9) 0–38.3 25.1 (12.5) 0–90 .42  REM sleep, % 19.0 (11.4) 0–20.5 20.6 (10.1) 8.6–44 .63  Apnea-hypopnea index, events/hour 16.7 (14.4) 0–59.3 6.7 (4.2) 0–23.4 .007*  Apnea-hypopnea average duration, seconds 29.8 (7.2) 0–38 22.2 (11.8) 0–40 .73  Minimal O2 saturation, % 69 (18.9) 30–97 81 (8.4) 65–93 .029*  Total sleep time with SpO2 <90%, % 9.1 (15.4) 0.1–95 2.3 (4.1) 0–16.8 .014* EKG  Heart rate, bpm 83.5 (10) 68–107 82.5 (12) 59–105 .71  QRS interval, ms 80.6 (17.9) 48–100 89.2 (21.8) 50–160 .06  QT interval, ms 345 (27) 300–388 360 (42) 284–460 .38  QTc interval, ms 404 (11) 380–427 419 (29) 375–482 .12  PR interval, ms 121 (26) 80–160 133 (18) 96–162 .07 Standard deviation of 24-hour systolic blood pressure, mmHg 24.1 (6.5) 17.3–31 23.1 (5.1) 14.2–29.9 .60 Arterial blood gases  pH 7.41 (0.05) 7.34–7.51 7.40 (0.03) 7.33–7.46 .27  pO2, mmHg 78.5 (14.2) 52.5–107 83.5 (17.8) 50–105 .16  pCO2, mmHg 42.1 (7.65) 23–59 44.3 (5.8) 35–64 .25  pHCO3, mEq/L 25.1 (4.6) 15–31.1 27.3 (2.6) 24.3–35.6 .18 Hematological parameters  White blood cell count, ×109/L 7.55 (3.54) 4.3–10.2 7.32 (2.35) 4–13.7 .92  Red blood cell count, ×109/L 4.26 (0.47) 3.3–5.5 4.11 (0.6) 2.7–5.1 .40  Hemoglobin, mg/dl 12.75 (1.29) 10.3–15.2 15.68 (20.53) 9.8–14.8 .43  Hematocrit, % 37.61 (4.11) 29.8–46.6 36.41 (4.37) 26.7–44.3 .33 Metabolic parameters  Sodium, mEq/L 140 (4.3) 134–147 139 (3.3) 132–144 .31  Potassium, mEq/L 4.3 (0.6) 3.3–6.2 4.7 (0.5) 4.1–5.9 <.0001*  Chloride, mEq/L 102 (4.3) 88–109 100 (3.9) 91–107 .20  Creatinine, mg/dL 1 (0.43) 0.3–2.4 1.05 (1.08) 0.4–1.6 .43  Blood urea nitrogen, mg/dL 23.3 (9.5) 3–40 24.1 (11.5) 13–38 .51 Medications  Fludrocortisone, mg/day 0.14 (0.10) 0–0.4 0.02 (0.04) 0–0.1 <.0001*  Midodrine, mg/day 3 (4.7) 0–17.5 3.3 (4.9) 0–15 .66  Benzodiazepines, mg/day 4.8 (4.7) 0–17.5 4.9 (4.6) 0–16 .78  Clonidine, mg/day 0.04 (0.07) 0–0.2 0.06 (0.06) 0–0.2 .41 Characteristic Cases (n = 32) Controls (n = 31) p-value Mean (SD) Range Mean (SD) Range Age, year 29.2 (12.5) 9–50 28.9 (11.7) 9–54 .99 Body mass index, kg/m2 17.9 (2.8) 13.7–23.9 17.2 (2.1) 14.4–22.2 .81 Polysomnography  Total sleep time, minutes 392 (145) 181–491 404 (101) 214–523 .78  REM latency, minutes 110 (59) 0–308 119 (61) 16–321 .81  Arousal index, events/hour 28.1 (16.1) 1.2–52.5 24.9 (18) 2–74.4 .38  Stage 1 sleep, % 7.5 (4.5) 1.8–24 5.28 (4.1) 0–17 .19  Stage 2 sleep, % 50.7 (13.2) 35.5–79.2 48.7 (25.1) 0–79 .55  Stage 3 sleep, % 19.1 (12.9) 0–38.3 25.1 (12.5) 0–90 .42  REM sleep, % 19.0 (11.4) 0–20.5 20.6 (10.1) 8.6–44 .63  Apnea-hypopnea index, events/hour 16.7 (14.4) 0–59.3 6.7 (4.2) 0–23.4 .007*  Apnea-hypopnea average duration, seconds 29.8 (7.2) 0–38 22.2 (11.8) 0–40 .73  Minimal O2 saturation, % 69 (18.9) 30–97 81 (8.4) 65–93 .029*  Total sleep time with SpO2 <90%, % 9.1 (15.4) 0.1–95 2.3 (4.1) 0–16.8 .014* EKG  Heart rate, bpm 83.5 (10) 68–107 82.5 (12) 59–105 .71  QRS interval, ms 80.6 (17.9) 48–100 89.2 (21.8) 50–160 .06  QT interval, ms 345 (27) 300–388 360 (42) 284–460 .38  QTc interval, ms 404 (11) 380–427 419 (29) 375–482 .12  PR interval, ms 121 (26) 80–160 133 (18) 96–162 .07 Standard deviation of 24-hour systolic blood pressure, mmHg 24.1 (6.5) 17.3–31 23.1 (5.1) 14.2–29.9 .60 Arterial blood gases  pH 7.41 (0.05) 7.34–7.51 7.40 (0.03) 7.33–7.46 .27  pO2, mmHg 78.5 (14.2) 52.5–107 83.5 (17.8) 50–105 .16  pCO2, mmHg 42.1 (7.65) 23–59 44.3 (5.8) 35–64 .25  pHCO3, mEq/L 25.1 (4.6) 15–31.1 27.3 (2.6) 24.3–35.6 .18 Hematological parameters  White blood cell count, ×109/L 7.55 (3.54) 4.3–10.2 7.32 (2.35) 4–13.7 .92  Red blood cell count, ×109/L 4.26 (0.47) 3.3–5.5 4.11 (0.6) 2.7–5.1 .40  Hemoglobin, mg/dl 12.75 (1.29) 10.3–15.2 15.68 (20.53) 9.8–14.8 .43  Hematocrit, % 37.61 (4.11) 29.8–46.6 36.41 (4.37) 26.7–44.3 .33 Metabolic parameters  Sodium, mEq/L 140 (4.3) 134–147 139 (3.3) 132–144 .31  Potassium, mEq/L 4.3 (0.6) 3.3–6.2 4.7 (0.5) 4.1–5.9 <.0001*  Chloride, mEq/L 102 (4.3) 88–109 100 (3.9) 91–107 .20  Creatinine, mg/dL 1 (0.43) 0.3–2.4 1.05 (1.08) 0.4–1.6 .43  Blood urea nitrogen, mg/dL 23.3 (9.5) 3–40 24.1 (11.5) 13–38 .51 Medications  Fludrocortisone, mg/day 0.14 (0.10) 0–0.4 0.02 (0.04) 0–0.1 <.0001*  Midodrine, mg/day 3 (4.7) 0–17.5 3.3 (4.9) 0–15 .66  Benzodiazepines, mg/day 4.8 (4.7) 0–17.5 4.9 (4.6) 0–16 .78  Clonidine, mg/day 0.04 (0.07) 0–0.2 0.06 (0.06) 0–0.2 .41 *Statistically significant differences. REM = rapid eye movement; SD = standard deviation. Open in new tab To further understand the relationship between fludrocortisone and plasma potassium levels, we performed a linear regression including both cases and controls, finding that higher fludrocortisone daily dosage was associated with lower plasma potassium levels (r2 = 0.19; p = .0005). The prevalence of seizures of any kind, epileptic seizures, and the use of fludrocortisone was higher in cases than in controls. Conversely, the use of noninvasive ventilation was lower in cases than in controls. In multivariable analysis, of these, only fludrocortisone treatment (odds ratio [OR] 29.7; 95% confidence interval [CI]: 4.1–213.4), OSA (OR 17.4; 95% CI: 1.5–193.61) and potassium levels < 4 mEq/L (OR 33.4; 95% CI: 1.8–619.3) were independently associated with SUDS in FD patients, whereas the use of noninvasive ventilation at night (OR 0.19; 95% CI: 0.06–0.61) was a protective factor against SUDS (Table 3). When the interaction term “with or without noninvasive ventilation” was introduced in the multivariable analysis, untreated OSA was independently associated with SUDS (OR: 3.67; 95% CI: 1.26–12.16). There were no differences in the frequency of cardiac disease, Nissen fundoplication, percutaneous gastrostomy tube, or nocturnal feedings. Table 3 Comparison of Categorical Characteristics in Cases and Controls. Characteristic Cases (n = 32) Controls (n = 31) Univariable OR (95% CI) Multivariable OR (95% CI) Women, n (%) 14 (44) 15 (48) 0.82 (0.39–2.23) 0.16 (0.21–1.80) History of seizures, n (%) 26 (81) 14 (45) 5.26 (1.69–16.3)* 1.19 (0–50)  Epileptic 11 (34) 3 (10) 4.88 (1.21–19.7)* 1.26 (0–50)  Hypoxic 10 (31) 7 (23) 1.56 (0.50–4.80) 1.22 (0–50)  Hyponatremic 5 (16) 3 (10) 1.72 (0.38–7.95) 1.76 (0–50)  Febrile 1 (3) 1 (3) 0.97 (0.05–16.2) 9.21 (0–50) Active epilepsy, n (%) 5 (16) 3 (10) 1.72 (0.38–7.95) 1.74 (0–50) Cardiac disease#, n (%) 6 (19) 6 (19) 0.96 (0.27–3.38) 0.36 (0.07–17) Pacemaker, n (%) 6 (19) 7 (23) 0.80 (0.23–2.69) 0.25 (0.05–12) Stroke/TIA, n (%) 0 (0) 0 (0) 0.96 (0.01–50) — Diabetes mellitus, n (%) 0 (0) 0 (0) 0.96 (0.01–50) — Treatments, n (%)  Benzodiazepines 24 (75) 27 (87) 0.44 (0.12–1.67) 1.4 (0.14–13.5)  Fludrocortisone 22 (69) 6 (19) 9.17 (2.9–29.34)* 29.7 (4.1–213.4)*  Clonidine 12 (38) 19 (61) 0.38 (0.14–1.04) 0.53 (0.18–3.3)  Midodrine 10 (31) 13 (42) 0.63 (0.22–1.77) 0.07 (0.01–1.22)  Beta-blockers 1 (3) 2 (6) 0.47 (0.04–5.44) 0.02 (0.01–14)  Potassium supplements 6 (19) 8 (25) 0.66 (0.18–2.02) 0.71 (0.18–3.11)  Noninvasive ventilation at night$ 3 (9) 12 (39) 0.18 (0.04–0.72)* 0.19 (0.06–0.61)* Obstructive sleep apnea&, n (%) 18 (56) 9 (29) 2.93 (1.04–8.25)* 17.4 (1.5–193.6)*  With noninvasive ventilation 3 (9) 3 (10) — 0.96 (0.21–4.4)  Without noninvasive ventilation 15 (47) 6 (19) — 3.67 (1.26–12.16)* Minimum sleep SatO2 ≤ 88%, n (%) 22 (69) 16 (51) 1.27 (0.47–3.41) 0.32 (0.03–3.58) Potassium ≤ 4 mEq/L, n (%) 13 (41) 1 (3) 19.5 (2.36–161)* 33.4 (1.8–619.3)* Nissen fundoplication, n (%) 25 (78) 28 (80) 0.38 (0.10–1.69) 0.51 (0.21–2.11) Percutaneous gastrostomy tube, n (%) 25 (78) 28 (80) 0.38 (0.10–1.69) 0.51 (0.21–2.11) Nocturnal feedings, n (%) 8 (25) 12 (39) 0.52 (0.19–1.47) 0.33 (0.14–4.21) Characteristic Cases (n = 32) Controls (n = 31) Univariable OR (95% CI) Multivariable OR (95% CI) Women, n (%) 14 (44) 15 (48) 0.82 (0.39–2.23) 0.16 (0.21–1.80) History of seizures, n (%) 26 (81) 14 (45) 5.26 (1.69–16.3)* 1.19 (0–50)  Epileptic 11 (34) 3 (10) 4.88 (1.21–19.7)* 1.26 (0–50)  Hypoxic 10 (31) 7 (23) 1.56 (0.50–4.80) 1.22 (0–50)  Hyponatremic 5 (16) 3 (10) 1.72 (0.38–7.95) 1.76 (0–50)  Febrile 1 (3) 1 (3) 0.97 (0.05–16.2) 9.21 (0–50) Active epilepsy, n (%) 5 (16) 3 (10) 1.72 (0.38–7.95) 1.74 (0–50) Cardiac disease#, n (%) 6 (19) 6 (19) 0.96 (0.27–3.38) 0.36 (0.07–17) Pacemaker, n (%) 6 (19) 7 (23) 0.80 (0.23–2.69) 0.25 (0.05–12) Stroke/TIA, n (%) 0 (0) 0 (0) 0.96 (0.01–50) — Diabetes mellitus, n (%) 0 (0) 0 (0) 0.96 (0.01–50) — Treatments, n (%)  Benzodiazepines 24 (75) 27 (87) 0.44 (0.12–1.67) 1.4 (0.14–13.5)  Fludrocortisone 22 (69) 6 (19) 9.17 (2.9–29.34)* 29.7 (4.1–213.4)*  Clonidine 12 (38) 19 (61) 0.38 (0.14–1.04) 0.53 (0.18–3.3)  Midodrine 10 (31) 13 (42) 0.63 (0.22–1.77) 0.07 (0.01–1.22)  Beta-blockers 1 (3) 2 (6) 0.47 (0.04–5.44) 0.02 (0.01–14)  Potassium supplements 6 (19) 8 (25) 0.66 (0.18–2.02) 0.71 (0.18–3.11)  Noninvasive ventilation at night$ 3 (9) 12 (39) 0.18 (0.04–0.72)* 0.19 (0.06–0.61)* Obstructive sleep apnea&, n (%) 18 (56) 9 (29) 2.93 (1.04–8.25)* 17.4 (1.5–193.6)*  With noninvasive ventilation 3 (9) 3 (10) — 0.96 (0.21–4.4)  Without noninvasive ventilation 15 (47) 6 (19) — 3.67 (1.26–12.16)* Minimum sleep SatO2 ≤ 88%, n (%) 22 (69) 16 (51) 1.27 (0.47–3.41) 0.32 (0.03–3.58) Potassium ≤ 4 mEq/L, n (%) 13 (41) 1 (3) 19.5 (2.36–161)* 33.4 (1.8–619.3)* Nissen fundoplication, n (%) 25 (78) 28 (80) 0.38 (0.10–1.69) 0.51 (0.21–2.11) Percutaneous gastrostomy tube, n (%) 25 (78) 28 (80) 0.38 (0.10–1.69) 0.51 (0.21–2.11) Nocturnal feedings, n (%) 8 (25) 12 (39) 0.52 (0.19–1.47) 0.33 (0.14–4.21) *Significant associations. #Ischemic, vascular, and congenital/inherited disease, atrial fibrillation, and third-degree atrioventricular block. $Includes continuous positive air pressure (CPAP) and bilevel positive airway pressure. &Defined as an apnea-hypopnea index >5 events/hour. CI = confidence interval; OR = odds ratio; TIA = transient ischemic attack. Open in new tab Table 3 Comparison of Categorical Characteristics in Cases and Controls. Characteristic Cases (n = 32) Controls (n = 31) Univariable OR (95% CI) Multivariable OR (95% CI) Women, n (%) 14 (44) 15 (48) 0.82 (0.39–2.23) 0.16 (0.21–1.80) History of seizures, n (%) 26 (81) 14 (45) 5.26 (1.69–16.3)* 1.19 (0–50)  Epileptic 11 (34) 3 (10) 4.88 (1.21–19.7)* 1.26 (0–50)  Hypoxic 10 (31) 7 (23) 1.56 (0.50–4.80) 1.22 (0–50)  Hyponatremic 5 (16) 3 (10) 1.72 (0.38–7.95) 1.76 (0–50)  Febrile 1 (3) 1 (3) 0.97 (0.05–16.2) 9.21 (0–50) Active epilepsy, n (%) 5 (16) 3 (10) 1.72 (0.38–7.95) 1.74 (0–50) Cardiac disease#, n (%) 6 (19) 6 (19) 0.96 (0.27–3.38) 0.36 (0.07–17) Pacemaker, n (%) 6 (19) 7 (23) 0.80 (0.23–2.69) 0.25 (0.05–12) Stroke/TIA, n (%) 0 (0) 0 (0) 0.96 (0.01–50) — Diabetes mellitus, n (%) 0 (0) 0 (0) 0.96 (0.01–50) — Treatments, n (%)  Benzodiazepines 24 (75) 27 (87) 0.44 (0.12–1.67) 1.4 (0.14–13.5)  Fludrocortisone 22 (69) 6 (19) 9.17 (2.9–29.34)* 29.7 (4.1–213.4)*  Clonidine 12 (38) 19 (61) 0.38 (0.14–1.04) 0.53 (0.18–3.3)  Midodrine 10 (31) 13 (42) 0.63 (0.22–1.77) 0.07 (0.01–1.22)  Beta-blockers 1 (3) 2 (6) 0.47 (0.04–5.44) 0.02 (0.01–14)  Potassium supplements 6 (19) 8 (25) 0.66 (0.18–2.02) 0.71 (0.18–3.11)  Noninvasive ventilation at night$ 3 (9) 12 (39) 0.18 (0.04–0.72)* 0.19 (0.06–0.61)* Obstructive sleep apnea&, n (%) 18 (56) 9 (29) 2.93 (1.04–8.25)* 17.4 (1.5–193.6)*  With noninvasive ventilation 3 (9) 3 (10) — 0.96 (0.21–4.4)  Without noninvasive ventilation 15 (47) 6 (19) — 3.67 (1.26–12.16)* Minimum sleep SatO2 ≤ 88%, n (%) 22 (69) 16 (51) 1.27 (0.47–3.41) 0.32 (0.03–3.58) Potassium ≤ 4 mEq/L, n (%) 13 (41) 1 (3) 19.5 (2.36–161)* 33.4 (1.8–619.3)* Nissen fundoplication, n (%) 25 (78) 28 (80) 0.38 (0.10–1.69) 0.51 (0.21–2.11) Percutaneous gastrostomy tube, n (%) 25 (78) 28 (80) 0.38 (0.10–1.69) 0.51 (0.21–2.11) Nocturnal feedings, n (%) 8 (25) 12 (39) 0.52 (0.19–1.47) 0.33 (0.14–4.21) Characteristic Cases (n = 32) Controls (n = 31) Univariable OR (95% CI) Multivariable OR (95% CI) Women, n (%) 14 (44) 15 (48) 0.82 (0.39–2.23) 0.16 (0.21–1.80) History of seizures, n (%) 26 (81) 14 (45) 5.26 (1.69–16.3)* 1.19 (0–50)  Epileptic 11 (34) 3 (10) 4.88 (1.21–19.7)* 1.26 (0–50)  Hypoxic 10 (31) 7 (23) 1.56 (0.50–4.80) 1.22 (0–50)  Hyponatremic 5 (16) 3 (10) 1.72 (0.38–7.95) 1.76 (0–50)  Febrile 1 (3) 1 (3) 0.97 (0.05–16.2) 9.21 (0–50) Active epilepsy, n (%) 5 (16) 3 (10) 1.72 (0.38–7.95) 1.74 (0–50) Cardiac disease#, n (%) 6 (19) 6 (19) 0.96 (0.27–3.38) 0.36 (0.07–17) Pacemaker, n (%) 6 (19) 7 (23) 0.80 (0.23–2.69) 0.25 (0.05–12) Stroke/TIA, n (%) 0 (0) 0 (0) 0.96 (0.01–50) — Diabetes mellitus, n (%) 0 (0) 0 (0) 0.96 (0.01–50) — Treatments, n (%)  Benzodiazepines 24 (75) 27 (87) 0.44 (0.12–1.67) 1.4 (0.14–13.5)  Fludrocortisone 22 (69) 6 (19) 9.17 (2.9–29.34)* 29.7 (4.1–213.4)*  Clonidine 12 (38) 19 (61) 0.38 (0.14–1.04) 0.53 (0.18–3.3)  Midodrine 10 (31) 13 (42) 0.63 (0.22–1.77) 0.07 (0.01–1.22)  Beta-blockers 1 (3) 2 (6) 0.47 (0.04–5.44) 0.02 (0.01–14)  Potassium supplements 6 (19) 8 (25) 0.66 (0.18–2.02) 0.71 (0.18–3.11)  Noninvasive ventilation at night$ 3 (9) 12 (39) 0.18 (0.04–0.72)* 0.19 (0.06–0.61)* Obstructive sleep apnea&, n (%) 18 (56) 9 (29) 2.93 (1.04–8.25)* 17.4 (1.5–193.6)*  With noninvasive ventilation 3 (9) 3 (10) — 0.96 (0.21–4.4)  Without noninvasive ventilation 15 (47) 6 (19) — 3.67 (1.26–12.16)* Minimum sleep SatO2 ≤ 88%, n (%) 22 (69) 16 (51) 1.27 (0.47–3.41) 0.32 (0.03–3.58) Potassium ≤ 4 mEq/L, n (%) 13 (41) 1 (3) 19.5 (2.36–161)* 33.4 (1.8–619.3)* Nissen fundoplication, n (%) 25 (78) 28 (80) 0.38 (0.10–1.69) 0.51 (0.21–2.11) Percutaneous gastrostomy tube, n (%) 25 (78) 28 (80) 0.38 (0.10–1.69) 0.51 (0.21–2.11) Nocturnal feedings, n (%) 8 (25) 12 (39) 0.52 (0.19–1.47) 0.33 (0.14–4.21) *Significant associations. #Ischemic, vascular, and congenital/inherited disease, atrial fibrillation, and third-degree atrioventricular block. $Includes continuous positive air pressure (CPAP) and bilevel positive airway pressure. &Defined as an apnea-hypopnea index >5 events/hour. CI = confidence interval; OR = odds ratio; TIA = transient ischemic attack. Open in new tab DISCUSSION This is the first study analyzing possible risk factors associated with SUDS in patients with FD. Treatment with fludrocortisone, plasma potassium levels <4 mEq/L, and untreated OSA were independently associated with a greater likelihood of SUDS. Conversely, treatment with nocturnal noninvasive ventilation was associated with a reduced likelihood of SUDS. Fludrocortisone (9α-fluorocortisol) is a synthetic mineralocorticoid that increases renal sodium and water reabsorption, expands intravascular volume, and increases BP. Treatment with fludrocortisone for orthostatic hypotension in patients with FD became widespread in the 1990s, sometimes at very high dosages (up to 0.4 mg/day). Hypokalemia is a frequent side effect.25 Long-term use of fludrocortisone in FD exacerbates supine hypertension and accelerates renal damage.22 Creatinine and blood urea nitrogen levels were similar between cases and controls; it is reasonable to assume that lower serum potassium levels in cases with SUDS were the result of fludrocortisone treatment. Supporting this assumption, a linear regression showed that higher daily dosages of fludrocortisone were associated with lower plasma potassium levels. Cases and controls had similar degree of cardiovascular dysautonomia (as measured by ambulatory BP monitoring), indicating that the high frequency of fludrocortisone treatment in the cases was not an epiphenomenon of more severe cardiovascular autonomic dysfunction. In the general population, hypokalemia and plasma potassium levels in the lower range of normality are independent risk factors for life-threatening arrhythmias and sudden cardiac death.26 Interestingly, all drugs proven to reduce mortality and morbidity rates in patients with cardiovascular disease increase plasma potassium concentration.26 OSA is also a well-known risk factor for sudden cardiac death in the general population.27,28 The repetitive apneas and subsequent hypoxemia result in cardiac autonomic abnormalities leading to arrhythmias and death.29 Cardiac autonomic changes induced by sleep-disordered breathing are reversible by noninvasive ventilation with continuous positive airway pressure,30,31 which could potentially prevent sudden cardiac death. Our findings of reduced risk of SUDS in FD patients being treated with noninvasive ventilation are consistent with this. Moreover, in patients with FD, failure to receive inputs from peripheral chemoreceptors, carried by the IX (glossopharyngeal) and X (vagus) cranial nerves to the central nervous system, results in markedly blunted ventilatory responses to hypoxia. Ventilatory responses to hypercapnia are reduced but still present. Also, there is no compensatory increase in sympathetic outflow in response to hypoxia; instead, in FD patients, hypoxia results in bradycardia and hypotension.15,16,18 We found no evidence that epileptic seizures played a role in the pathogenesis of SUDS in FD. The presence of active epilepsy was similar in cases and controls, and the circumstances surrounding death were not indicative of seizures, except for one case with urinary incontinence. Therefore, SUDS in FD seems to result from a combination of respiratory and cardiac dysfunction, namely: (1) cardiac autonomic dysfunction in the setting of baroreflex failure aggravated by sleep-disordered breathing, (2) hypoxia-induced bradycardia, and (3) increased propensity toward arrhythmogenesis induced or aggravated by low plasma potassium levels. The incidence of SUDS in FD patients is one of the highest reported in any disorder. The annual incidence rate of SUDS in patients with FD is 3.4 per 1000 person-year, compared to 0.5–1 per 1000 person-year of sudden unexpected death in epilepsy.32 This high incidence is probably due to the unfortunate combination of sleep-disordered breathing and cardiac autonomic dysfunction. The interaction between cardiovascular autonomic and sleep dysfunctions also underlies the pathogenesis of SUDS in other conditions such as multiple system atrophy,33 congenital central hypoventilation syndrome,34 and long-QT syndrome.35 In our patients with FD, in addition, treatment with fludrocortisone leading to low plasma potassium concentrations likely increased the risk further. A major strength of our study is that we included only cases with probable or definite SUDS. This is more specific than previous studies that included all types of death (eg, pneumonia). Autopsy confirmation was only available in six cases, all of them showing no cardiac or brain structural abnormalities that could explain the death. In keeping with this, our study did not find any association between increased likelihood of SUDS and previous cardiac disease or decreased likelihood of SUDS and cardiac pacemaker implantation. This is consistent with a previous study showing that pacemakers might decrease the incidence of bradyarrhythmias but did not prevent sudden death in patients with FD.36 Another study found that increased QT variability was associated with all-cause death in FD, although this finding was probably due to ventilatory problems, such as sleep-disordered breathing, rather than to a primary arrhythmia.37 Five of our six cases with pathological confirmation had nonspecific pulmonary congestion with no specific signs of infection or disease. Pulmonary congestion and edema is a common and nonspecific phenomenon that may develop either as a consequence of hypoxia or a few hours after death as a time-dependent change due to pressure gradient between pulmonary vasculature and the alveolar spaces and an alteration of capillary permeability.38 Another potential limitation in our study is that we only selected one matched-control per case. Because we wanted to focus on sleep-related risk factors, we required all subjects to have a polysomnography performed recently. Inclusion of additional controls with incomplete information on sleep parameters would have resulted in biased results. Patients with FD were recommended to have a polysomnography at their annual visits at our center, regardless of the presence of symptoms of sleep-disordered breathing. In keeping with this, the percentage of patients with a polysomnography in the deceased patients group was similar to the percentage in the surviving patients group (~10% in both cases). However, as the percentage in both groups was low, selection bias is still a possibility. In conclusion, fludrocortisone treatment, plasma potassium levels <4 mEq/L, and OSA were independently associated with increased risk, whereas the use of noninvasive ventilation was associated with decreased risk of SUDS in patients with FD. Most importantly, both OSA and low serum potassium levels are treatable risk factors. Initiation of noninvasive ventilation when required and discontinuation of fludrocortisone treatment should reduce the high incidence rate of SUDS in patients with FD. SUPPLEMENTARY MATERIAL Supplementary material is available at SLEEP online. FUNDING This study was funded by National Institutes of Health (U54NS065736) and Dysautonomia Foundation, Inc. DISCLOSURE STATEMENT None declared. REFERENCES 1. Riley CM Day RL . Central autonomic dysfunction with defective lacrimation; report of five cases . Pediatrics . 1949 ; 3 ( 4 ): 468 – 478 . Google Scholar PubMed WorldCat 2. Norcliffe-Kaufmann L Slaugenhaupt SA Kaufmann H . Familial dysautonomia: History, genotype, phenotype and translational research . Prog Neurobiol . 2017 ; 152 : 131 – 148 . Google Scholar Crossref Search ADS PubMed WorldCat 3. Blumenfeld A Slaugenhaupt SA Axelrod FB et al. . Localization of the gene for familial dysautonomia on chromosome 9 and definition of DNA markers for genetic diagnosis . Nat Genet . 1993 ; 4 ( 2 ): 160 – 164 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Slaugenhaupt SA Blumenfeld A Gill SP et al. . Tissue-specific expression of a splicing mutation in the IKBKAP gene causes familial dysautonomia . Am J Hum Genet . 2001 ; 68 ( 3 ): 598 – 605 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Mezey E Parmalee A Szalayova I et al. . Of splice and men: what does the distribution of IKAP mRNA in the rat tell us about the pathogenesis of familial dysautonomia? Brain Res . 2003 ; 983 ( 1–2 ): 209 – 214 . Google Scholar Crossref Search ADS PubMed WorldCat 6. Hunnicutt BJ Chaverra M George L Lefcort F . IKAP/Elp1 is required in vivo for neurogenesis and neuronal survival, but not for neural crest migration . PLoS One . 2012 ; 7 ( 2 ): e32050 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Macefield VG Norcliffe-Kaufmann L Gutiérrez J Axelrod FB Kaufmann H . Can loss of muscle spindle afferents explain the ataxic gait in Riley-Day syndrome? Brain . 2011 ; 134 ( Pt 11 ): 3198 – 3208 . Google Scholar Crossref Search ADS PubMed WorldCat 8. Maayan HC . Respiratory aspects of Riley-Day Syndrome: familial dysautonomia . Paediatr Respir Rev . 2006 ; 7( Suppl 1) : S258 – S259 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Norcliffe-Kaufmann L Axelrod F Kaufmann H . Afferent baroreflex failure in familial dysautonomia . Neurology . 2010 ; 75 ( 21 ): 1904 – 1911 . Google Scholar Crossref Search ADS PubMed WorldCat 10. Norcliffe-Kaufmann L Palma JA Kaufmann H . Mother-induced hypertension in familial dysautonomia . Clin Auton Res . 2016 ; 26 ( 1 ): 79 – 81 . Google Scholar Crossref Search ADS PubMed WorldCat 11. Palma JA Norcliffe-Kaufmann L Fuente-Mora C Percival L Mendoza-Santiesteban C Kaufmann H . Current treatments in familial dysautonomia . Expert Opin Pharmacother . 2014 ; 15 ( 18 ): 2653 – 2671 . Google Scholar Crossref Search ADS PubMed WorldCat 12. McNicholas WT Rutherford R Grossman R Moldofsky H Zamel N Phillipson EA . Abnormal respiratory pattern generation during sleep in patients with autonomic dysfunction . Am Rev Respir Dis . 1983 ; 128 ( 3 ): 429 – 433 . Google Scholar Crossref Search ADS PubMed WorldCat 13. Gadoth N Sokol J Lavie P . Sleep structure and nocturnal disordered breathing in familial dysautonomia . J Neurol Sci . 1983 ; 60 ( 1 ): 117 – 125 . Google Scholar Crossref Search ADS PubMed WorldCat 14. Weese-Mayer DE Kenny AS Bennett HL Ramirez JM Leurgans SE . Familial dysautonomia: frequent, prolonged and severe hypoxemia during wakefulness and sleep . Pediatr Pulmonol . 2008 ; 43 ( 3 ): 251 – 260 . Google Scholar Crossref Search ADS PubMed WorldCat 15. Filler J Smith AA Stone S Dancis J . Respiratory control in familial dysautonomia . J Pediatr . 1965 ; 66 : 509 – 516 . Google Scholar Crossref Search ADS PubMed WorldCat 16. Edelman NH Cherniack NS Lahiri S Richards E Fishman AP . The effects of abnormal sympathetic nervous function upon the ventilatory response to hypoxia . J Clin Invest . 1970 ; 49 ( 6 ): 1153 – 1165 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Guilleminault C Mondini S Greenfield M . Abnormal respiratory pattern generation during sleep in patients with autonomic dysfunction . Am Rev Respir Dis . 1984 ; 129 ( 3 ): 512 – 513 . Google Scholar PubMed WorldCat 18. Bernardi L Hilz M Stemper B Passino C Welsch G Axelrod FB . Respiratory and cerebrovascular responses to hypoxia and hypercapnia in familial dysautonomia . Am J Respir Crit Care Med . 2003 ; 167 ( 2 ): 141 – 149 . Google Scholar Crossref Search ADS PubMed WorldCat 19. von Elm E Altman DG Egger M Pocock SJ Gøtzsche PC Vandenbroucke JP ; STROBE Initiative . The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies . PLoS Med . 2007 ; 4 ( 10 ): e296 . Google Scholar Crossref Search ADS PubMed WorldCat 20. Annegers JF . United States perspective on definitions and classifications . Epilepsia . 1997 ; 38 ( 11 Suppl ): S9 – 12 . Google Scholar Crossref Search ADS PubMed WorldCat 21. Pickering TG Shimbo D Haas D . Ambulatory blood-pressure monitoring . N Engl J Med . 2006 ; 354 ( 22 ): 2368 – 2374 . Google Scholar Crossref Search ADS PubMed WorldCat 22. Norcliffe-Kaufmann L Axelrod FB Kaufmann H . Developmental abnormalities, blood pressure variability and renal disease in Riley Day syndrome . J Hum Hypertens . 2013 ; 27 ( 1 ): 51 – 55 . Google Scholar Crossref Search ADS PubMed WorldCat 23. Cohen P Solomon NH . Familial dysautonomia; case report with autopsy . J Pediatr . 1955 ; 46 ( 6 ): 663 – 670 . Google Scholar Crossref Search ADS PubMed WorldCat 24. Pearson J Pytel BA Grover-Johnson N Axelrod F Dancis J . Quantitative studies of dorsal root ganglia and neuropathologic observations on spinal cords in familial dysautonomia . J Neurol Sci . 1978 ; 35 ( 1 ): 77 – 92 . Google Scholar Crossref Search ADS PubMed WorldCat 25. Chobanian AV Volicer L Tifft CP Gavras H Liang CS Faxon D . Mineralocorticoid-induced hypertension in patients with orthostatic hypotension . N Engl J Med . 1979 ; 301 ( 2 ): 68 – 73 . Google Scholar Crossref Search ADS PubMed WorldCat 26. Kjeldsen K . Hypokalemia and sudden cardiac death . Exp Clin Cardiol . 2010 ; 15 ( 4 ): e96 – e99 . Google Scholar PubMed WorldCat 27. Gami AS Howard DE Olson EJ Somers VK . Day-night pattern of sudden death in obstructive sleep apnea . N Engl J Med . 2005 ; 352 ( 12 ): 1206 – 1214 . Google Scholar Crossref Search ADS PubMed WorldCat 28. Gami AS Olson EJ Shen WK et al. . Obstructive sleep apnea and the risk of sudden cardiac death: a longitudinal study of 10,701 adults . J Am Coll Cardiol . 2013 ; 62 ( 7 ): 610 – 616 . Google Scholar Crossref Search ADS PubMed WorldCat 29. Palma JA Urrestarazu E Lopez-Azcarate J et al. . Increased sympathetic and decreased parasympathetic cardiac tone in patients with sleep related alveolar hypoventilation . Sleep . 2013 ; 36 ( 6 ): 933 – 940 . Google Scholar Crossref Search ADS PubMed WorldCat 30. Kufoy E Palma JA Lopez J et al. . Changes in the heart rate variability in patients with obstructive sleep apnea and its response to acute CPAP treatment . PLoS One . 2012 ; 7 ( 3 ): e33769 . Google Scholar Crossref Search ADS PubMed WorldCat 31. Palma JA Iriarte J Fernandez S et al. . Long-term continuous positive airway pressure therapy improves cardiac autonomic tone during sleep in patients with obstructive sleep apnea . Clin Auton Res . 2015 ; 25 ( 4 ): 225 – 232 . Google Scholar Crossref Search ADS PubMed WorldCat 32. Tomson T Nashef L Ryvlin P . Sudden unexpected death in epilepsy: current knowledge and future directions . Lancet Neurol . 2008 ; 7 ( 11 ): 1021 – 1031 . Google Scholar Crossref Search ADS PubMed WorldCat 33. Shimohata T Ozawa T Nakayama H Tomita M Shinoda H Nishizawa M . Frequency of nocturnal sudden death in patients with multiple system atrophy . J Neurol . 2008 ; 255 ( 10 ): 1483 – 1485 . Google Scholar Crossref Search ADS PubMed WorldCat 34. Rand CM Patwari PP Carroll MS Weese-Mayer DE . Congenital central hypoventilation syndrome and sudden infant death syndrome: disorders of autonomic regulation . Semin Pediatr Neurol . 2013 ; 20 ( 1 ): 44 – 55 . Google Scholar Crossref Search ADS PubMed WorldCat 35. Shamsuzzaman AS Somers VK Knilans TK Ackerman MJ Wang Y Amin RS . Obstructive sleep apnea in patients with congenital long QT syndrome: implications for increased risk of sudden cardiac death . Sleep . 2015 ; 38 ( 7 ): 1113 – 1119 . Google Scholar Crossref Search ADS PubMed WorldCat 36. Gold-von Simson G Rutkowski M Berlin D Axelrod FB . Pacemakers in patients with familial dysautonomia–a review of experience with 20 patients . Clin Auton Res . 2005 ; 15 ( 1 ): 15 – 20 . Google Scholar Crossref Search ADS PubMed WorldCat 37. Solaimanzadeh I Schlegel TT Feiveson AH et al. . Advanced electrocardiographic predictors of mortality in familial dysautonomia . Auton Neurosci . 2008 ; 144 ( 1–2 ): 76 – 82 . Google Scholar Crossref Search ADS PubMed WorldCat 38. Saukko P Knight B . The pathophysiology of death . In: Saukko P Knight B , eds. Forensic pathology . 3rd ed. London : Hodder Arnold , 2004 . Google Preview WorldCat COPAC © Sleep Research Society 2017. Published by Oxford University Press on behalf of the Sleep Research Society. All rights reserved. For permissions, please e-mail [email protected].

Stahl, Bethany, A;Slocumb, Melissa, E;Chaitin,, Hersh;DiAngelo, Justin, R;Keene, Alex, C

doi: 10.1093/sleep/zsx084pmid: 28541527

Abstract Study Objectives Dysregulation of sleep is associated with metabolic diseases, and metabolic rate (MR) is acutely regulated by sleep-wake behavior. In humans and rodent models, sleep loss is associated with obesity, reduced metabolic rate, and negative energy balance, yet little is known about the neural mechanisms governing interactions between sleep and metabolism. Methods We have developed a system to simultaneously measure sleep and MR in individual Drosophila, allowing for interrogation of neural systems governing interactions between sleep and metabolic rate. Results Like mammals, MR in flies is reduced during sleep and increased during sleep deprivation suggesting sleep-dependent regulation of MR is conserved across phyla. The reduction of MR during sleep is not simply a consequence of inactivity because MR is reduced ~30 minutes following the onset of sleep, raising the possibility that CO2 production provides a metric to distinguish different sleep states in the fruit fly. To examine the relationship between sleep and metabolism, we determined basal and sleep-dependent changes in MR is reduced in starved flies, suggesting that starvation inhibits normal sleep-associated effects on metabolic rate. Further, translin mutant flies that fail to suppress sleep during starvation demonstrate a lower basal metabolic rate, but this rate was further reduced in response to starvation, revealing that regulation of starvation-induced changes in MR and sleep duration are genetically distinct. Conclusions Therefore, this system provides the unique ability to simultaneously measure sleep and oxidative metabolism, providing novel insight into the physiological changes associated with sleep and wakefulness in the fruit fly. Drosophila, metabolism, respirometry, calorimetry, sleep Statement of Significance Metabolic disorders are associated with sleep disturbances, yet our understanding of the mechanisms underlying interactions between sleep and metabolism remains limited. Here, we describe a novel system to simultaneously record sleep and metabolic rate in single Drosophila. Our findings reveal that uninterrupted sleep bouts of 30 minutes or greater are associated with a reduction in metabolic rate providing a physiological readout of sleep. Use of this system, combined with existing genetic tools in Drosophila, will facilitate identification of novel sleep genes and neurons, with implications for understanding the relationship between sleep loss and metabolic disease. INTRODUCTION Dysregulation of sleep is strongly linked to metabolism-related pathologies, and reciprocal interactions between sleep and metabolism suggest these processes are integrated at the cellular and molecular levels.1,2 In mammals, metabolic rate (MR) is reduced during sleep raising the possibility that sleep provides a mechanism of energy conservation or partitioning.3 Although a reduction in MR and energy expenditure during sleep has been documented in mammalian and avian species,4,5 little is known about the genetic and neural mechanisms governing the effects of sleep on MR. The fruit fly, Drosophila melanogaster, displays all the behavioral characteristics of sleep and provides a powerful system for genetic investigation of interactions between sleep and diverse physiological processes.2,6,7 Here, we describe a novel single-fly respirometry assay in the fruit fly, designed to simultaneously measure sleep and whole-body MR that allows for genetic interrogation of the mechanisms regulating interactions between these processes. Sleep is characterized by physiological changes in brain activity or through the behavioral correlates that accompany these changes.8 Flies, like mammals, display distinct electrophysiological patterns that correlate with wake and rest.9,10 Additionally, flies display all the behavioral hallmarks of sleep including extended periods of behavioral quiescence, rebound following deprivation, increased arousal threshold, and species-specific posture.6,7 Sleep in Drosophila is typically defined by 5 minutes of behavioral quiescence because this correlates with other behavioral characteristics used to define sleep.7 Although these behavioral metrics of sleep have been studied extensively, significantly less is known about physiological changes associated with sleep in flies. In rodents and humans, MR is elevated in response to sleep deprivation and reduced during sleep, supporting the notion that metabolic processes are acutely regulated by sleep state.11–13 In flies and other small insects, stop-flow respirometry can be used to monitor CO2 production, a by-product of oxidative metabolism and a proxy for MR.14 Here, we describe a system to simultaneously measure sleep and MR in individual fruit flies. Our findings reveal that MR is reduced when flies sleep, and uninterrupted sleep bouts of ~30 minutes or greater are associated with an additional reduction in MR, indicating that flies exhibit sleep stages that are physiologically distinct. Further, we find that starvation inhibits sleep-associated reductions in MR, suggesting feeding state influence physiological changes associated with sleep. These findings suggest that sleep-dependent reductions in MR previously observed in mammals are conserved in the fruit fly and further support the notion that sleep provides a mechanism for energy conservation. MethodS Drosophila Maintenance and Fly Stocks Flies were grown and maintained on standard food (Bloomington Recipe, Genesee Scientific). Flies were maintained in incubators (Powers Scientific; Dros52) at 25°C on a 12:12 light/dark cycle, with humidity set to 55%–65%. The wild-type line used in this manuscript is the w1118 fly strain (Bloomington Stock #5905). The trsnnull allele is an excision of the trsnEY06981 locus derived from mobilizing the EPgy2 insertion.15 This allele removes the entire coding region of the gene and represents a null mutation that has been outcrossed to the w1118 background and has previously been described as Δtrsn.15 Unless noted in the figures, all experiments are performed in 3- to 5-day-old mated female flies. Measurement of MR and Locomotor Activity MR was measured at 25°C through indirect calorimetry, measuring the CO2 production of individual flies with a Li-7000 CO2 analyzer (LI-COR), which was calibrated with pure CO2 before each run. A stop-flow, push-through respirometry setup was constructed using Sable Systems equipment (Sable Systems International). The experimental setup included sampling CO2 from an empty chamber to assess baseline levels, alongside five behavioral chambers, each measuring CO2 production of a single fly. The weight of flies used for analysis were not taken into account because body size and energy stores are not perturbed in trsnnull flies and do not vary significantly in w1118 flies. Further, previous work using a comparable system suggests weight will have little effect on CO2 measurements unless there is an excess of >50% differences in size between individuals.14,16 To measure CO2 output, air was flushed from each chamber for 50 seconds providing a readout of CO2 accumulation over a 5-minute period. This 5-minute interval allows the coordinate and simultaneous activity-based assessment of sleep. The first 20 minutes of recordings were not included in analyses because this time was necessary to purge the system of ambient air and residual CO2 from the closed system. Dehumidified, CO2 free air was pumped through a mass flow control valve (Side-Trak 840 Series; Sierra Instruments, Inc.) to maintain the experimental flow rate of 100 mL/min. The air was then passed through water-permeable Nafion tubing (Perma Pure, LLC, Lakewood, New Jersey, USA, #TT-070) immersed in a reservoir containing deionized H2O to rehumidify the air before reaching the behavioral chambers. Nonpermeable Bev-A line tubing (United States Plastic Corp., Lima, Ohio, USA, #56280) was used throughout the rest of the system. Experiments were conducted by placing single flies in 70 mm × 20 mm glass tubes that fit a custom-built Drosophila Locomotor Activity Monitor (Trikinetics, Waltham, Massachusetts) with three sets of infrared (IR) beams for activity detection. The monitor was connected to a computer to record beam breaks every minute for each animal using standard Drosophila Activity Monitor (DAMS) activity software (Trikinetics, Waltham, Massachusetts) as previously described.17 These data were used to calculate sleep information by extracting immobility bouts of 5 minutes using a custom-generated python program. The total activity from all three beams was summed for each time point in order to determine overall activity. Video recordings for analysis of feeding activity were acquired using a handheld USB Digital microscope (Vivida, 2MP #eheV1-USBpro) camera at 12 fps with VirtualDub software (v.1.10.4). Each 60-minute video recording occurred between ZT01-04 to prevent circadian differences in sleep, feeding, and MR. During video recording, flies were simultaneously assayed for activity and MR, with the stop flow set to collect CO2 output every 2 minutes. Videos were manually scored for feeding activity in corresponding 2-minute intervals as a “feeding” or “nonfeeding” bin. Flies were briefly anesthetized using CO2 for sorting at least 24 hours before the start of an experiment to allow for metabolic recovery. For all experiments, flies were loaded into chambers by mouth pipette to avoid confounding effects of anesthesia and allowed to acclimate in the system with the air flowing for 12–24 hours before behavior experiments, unless otherwise specified. To control for effects of diet composition, all experimental flies were fed a consistent diet. Each chamber contained a single food vial containing 1% agar plus 5% sucrose (Sigma) with red food coloring (McCormick), which we have previously shown to result in sleep comparable to standard fly food.18 For starvation experiments, flies had access to 1% agar dissolved in dH2O and were acclimated for 12 hours during lights on with access to agar alone, with analyses beginning at ZT12 at lights off. All experimental runs included analysis of both experimental flies and relevant controls in a randomized order to account for any subtle variation between runs. Pharmacology Pharmacological-induced sleep was achieved through administration of gabaxodol (4,5,6,7-tetrahydroisoxazolo[5,4-c]pyridin-3-ol hydrochloride, THIP hydrochloride; Sigma Aldrich #85118-33-8) at the dosage of 0.1 mg/mL, as previously described.19 Gaboxadol was dissolved in dH2O with 1% agar and 5% sucrose. Flies were loaded into the respirometry system with the gaboxadol 2 hours before lights off (ZT10) and were maintained on the drug throughout the duration of the experiment, as described in the text. Sleep Deprivation Flies were acclimated to the respirometry system during the daytime (ZT0-12). For mechanical sleep deprivation, flies were shaken every 2–3 minutes for 12 hours in the modified DAMs monitor/respirometry system throughout the nighttime (ZT12-24) while simultaneously measuring MR. The mechanical stimulus was applied using a vortexer (Fisher Scientific, MultiTube Vortexer) and a repeat cycle relay switch (Macromatic, TR63122). Sleep rebound and corresponding MR was measured the following day from ZT0-ZT12. Sleep, Metabolic, and Statistical Analyses Respirometry recordings were analyzed using ExpeData PRO software (Sable Systems International, v1.8.4). The CO2 lag time from the chamber to the analyzer was corrected, the baseline was subtracted from each behavioral chamber, and the absolute CO2 levels (ppm) was converted to μL/hr using the recorded air flow rate. Integrating the CO2 trace revealed the total CO2 produced, or the average MR, per fly for each recording. These data were exported to Excel, where metabolic output was matched to activity, and sleep analyses were performed using a custom python program. Since individual flies were measured for either a 12-hour or 24-hour experimental duration (described in text), our raw data included resampling of MR or beam crosses for each hour. To account for these repeated hourly measures, we determined the mean of the hourly readings for each individual fly before our statistical analyses represented in the graphs, meaning that each fly is represented once and the “N” reported in each figure specifically refers to the number of individual flies assayed in the experiment. To detect significant differences for activity (number of beam crosses), mean VCO2 (μL/hour) or total sleep (minutes), we employed a Student t-test (day vs. night; untreated control vs. gaboxadol-treated; fed vs. starved), two-way analysis of variance (ANOVA) with Sidak’s multiple comparison correction (female, day vs. night; male, day vs. night), and a two-way ANOVA with Sidak’s multiple comparison correction (w1118, fed vs. starved; trsn, fed vs. starved), when appropriate using InStat software (GraphPad Software 6.0). The two-tailed p-value used to test significance is denoted as p < .05. To account for individual-specific differences in MR, we surveyed the MR throughout longer sleep bouts by calculating percent change in MR. This was determined by subtracting the MR during the first 5 minutes asleep from the MR during each of the subsequent 5 minutes asleep for the entire length of the sleep bout, divided by the MR during the first 5 minutes asleep, multiplied by 100 (eg, [{first 5 minutes MR} – {20 minutes MR}/{first 5 minutes MR}] × 100). We note some flies exhibited longer sleep bouts; however, this analysis was restricted to sleep bouts up to 60 minutes due to limited replicates with extended bout lengths. Moreover, a similar approach was employed as described above for analysis of MR for flies with repeated sleep bouts. If a single fly demonstrated multiple distinct sleep bouts, we determined the mean percent change in MR for each fly at each sleep bin. For these analyses, we performed a one-way ANOVA with Sidak correction comparing the initial percent change in MR (5-minute bin) to each of the subsequent sleep bins (15–60 minute bins at 5-minute intervals) using InStat software (GraphPad Software 6.0) with significance denoted as p < .05. We applied a linear regression model to characterize the relationship between both absolute vCO2 versus activity (number of beam crossings) and percent change in MR and sleep duration using InStat software (GraphPad Software 6.0) with significance denoted as p < .05. Comparison of slopes derived from regression lines in fed versus starved states was performed using analysis of covariance (F-statistic; GraphPad Software 6.0). Before modeling, we performed pretests, including: generation of residual versus fitted plots to determine homogeneity of variance, normal Q-Q plot, Pearson correlation table and linear model assumptions (B.L.U.E.). The culmination of these tests indicated that our data were both normally distributed and appropriate for linear regression modeling. RESULTS Long-Term Recordings of Sleep and MR To simultaneously measure the effects of sleep on MR, we designed a stop-flow respirometry system coupled to a custom-built DAM system (Figure 1A). Each DAM chamber contained three IR beams for precise detection of locomotor activity of a single fly.20 Humidified, fully oxygenated air was passed through each chamber, preventing desiccation and allowing for long-term recordings. After exiting the chamber, air was dehumidified and passed through a CO2 analyzer. The system was set to a stop-flow configuration, where the CO2 accumulation in each chamber was measured every 5 minutes, and these were matched to the corresponding locomotor activity of individual flies within this period (Figure 1B). Flies are diurnal with elevated locomotor levels during the day compared to night, and these activity patterns were maintained in the respirometry system in both male and female flies (Figure 1C and D), indicating that the moderate airflow used in this system does not disrupt sleep-wake behavior. In both male and female flies, the mean MR was elevated during the daytime compared to the night, supporting the notion that CO2 production is associated with periods of high activity (Figure 1E and F). Examination of CO2 levels in individual flies revealed a weak correlation in both females and males between total locomotor activity and CO2 levels (Figure 1G and H). However, vCO2 was significantly elevated in females with activity of >60 beam breaks and males >50 beam breaks per 5-minute bin compared to the 1–10 beam breaks bin, suggesting MR is elevated during periods of robust activity (Figure 1G and H). Therefore, this system effectively measures locomotor activity and MR simultaneously in individual Drosophila. Figure 1— Open in new tabDownload slide A system to measure MR in single flies. (A) MR was measured through indirect calorimetry. A stop-flow respirometry system measured the CO2 produced by single flies placed inside of a 70 mm long × 20 mm diameter glass tube. Each fly had access to 1% agar and 5% sucrose. Activity and sleep were measured simultaneously as MR using a Drosophila Locomotor Activity Monitor with three infrared beams running through each behavior chamber. The computer counted the number of beam breaks. (B) A representative 5-minute reading, with the activity in number of beam crosses and the amount of CO2 produced by each fly over time. (C) The MR and activity for one female fly. (D) The MR and activity for one male fly. (E) The activity of female (N = 24; p < .0001) and male (N = 35; p < .0001) flies in beam crosses per hour, over 12 hours of day and night (two-way ANOVA F(1,114) = 171.9, p < .0001). Condition-by-sex interaction is significant (two-way ANOVA F(1,114) = 15.30, p < .001). (F) The MR of female (N = 24; p < .01) and male (N = 35; p < .01) flies as CO2 produced per hour, over 12 hours of day and night (two-way ANOVA F(1,114) = 21.27, p < .0001). Condition-by-sex interaction is not significant (two-way ANOVA F(1,114) = 0.4137, p > .50). (G) Linear regression of absolute vCO2 readout versus activity of female flies (N = 24 each bin; R2 = 0.120) and (H) male flies (N = 35 each bin; R2 = 0.064). Gray dashed lines indicate 95% confidence interval. One-way ANOVA comparing the vCO2 at the 1–10 beam crossings bin to each subsequent beam crossing bin: females >60 crossings (N = 24 each bin; p < .05) and males >60 crossings (N = 35 each bin; p < .05). ANOVA = analysis of variance; IR = infrared. Figure 1— Open in new tabDownload slide A system to measure MR in single flies. (A) MR was measured through indirect calorimetry. A stop-flow respirometry system measured the CO2 produced by single flies placed inside of a 70 mm long × 20 mm diameter glass tube. Each fly had access to 1% agar and 5% sucrose. Activity and sleep were measured simultaneously as MR using a Drosophila Locomotor Activity Monitor with three infrared beams running through each behavior chamber. The computer counted the number of beam breaks. (B) A representative 5-minute reading, with the activity in number of beam crosses and the amount of CO2 produced by each fly over time. (C) The MR and activity for one female fly. (D) The MR and activity for one male fly. (E) The activity of female (N = 24; p < .0001) and male (N = 35; p < .0001) flies in beam crosses per hour, over 12 hours of day and night (two-way ANOVA F(1,114) = 171.9, p < .0001). Condition-by-sex interaction is significant (two-way ANOVA F(1,114) = 15.30, p < .001). (F) The MR of female (N = 24; p < .01) and male (N = 35; p < .01) flies as CO2 produced per hour, over 12 hours of day and night (two-way ANOVA F(1,114) = 21.27, p < .0001). Condition-by-sex interaction is not significant (two-way ANOVA F(1,114) = 0.4137, p > .50). (G) Linear regression of absolute vCO2 readout versus activity of female flies (N = 24 each bin; R2 = 0.120) and (H) male flies (N = 35 each bin; R2 = 0.064). Gray dashed lines indicate 95% confidence interval. One-way ANOVA comparing the vCO2 at the 1–10 beam crossings bin to each subsequent beam crossing bin: females >60 crossings (N = 24 each bin; p < .05) and males >60 crossings (N = 35 each bin; p < .05). ANOVA = analysis of variance; IR = infrared. MR Is Reduced in Sleeping Drosophila Five minutes of immobility in Drosophila associates with relevant behavioral and physiological sleep metrics, allowing for sleep duration to be inferred from periods of behavioral quiescence.7,10 To measure MR during sleep, female flies were acclimated in the locomotor chambers for 24 hours, followed by continuous measurements of sleep and MR for an additional 24 hours. Flies slept significantly more during the night (ZT12-24), which corresponded with a reduction in MR (Figure 2A–C). To determine whether changes in MR are associated with sleep bout duration, we investigated changes in CO2 production during sleep bouts. CO2 production during a single representative 60-minute sleep bout revealed a reduction in MR as sleep progressed (Figure 2D). To account for individual variation between replicates, change in MR was calculated as percent change for each 5-minute interval throughout the sleep bout compared to the first 5 minutes of sleep. To avoid confounds resulting from circadian differences in MR, analysis was limited to nighttime sleep. Regression analysis revealed a significant relationship between of vCO2 and sleep bout length (Figure 2E). Comparing the average percent change in MR during sleep for each individual bout revealed MR was significantly reduced following 35 minutes of sleep, indicating that longer periods of uninterrupted sleep are associated with reduced MR. Percent change in MR continued to decline as sleep progressed until reaching a maximum percent change in MR of ~−12% to 15% after 50 minutes of sleep. To confirm that reduction of MR during sleep is not simply due to lack of feeding activity, we compared MR during feeding and nonfeeding bins from ZT1-ZT3 and did not detect significant differences in MR between feeding and waking nonfeeding periods (Supplementary Figure S1). Moreover, we performed standard allometric analysis of body size versus MR to identify if weight variation among individual flies could function as a covariate affecting MR21,22 and determined that there is no effect of variation in body weight on MR (n = 34, R2 = 0.030). Therefore, reduced CO2 production is associated with consolidated sleep bouts, revealing that MR can be functionally separated from overall activity. Figure 2— Open in new tabDownload slide MR is reduced during sleep state. Female control flies (w1118) were allowed to acclimate in the system for 24 hours. (A) MR shows an inverse pattern to their sleep (N = 24). (B) Total minutes of sleep per 12 hours of day and night for B (N = 24; p < .001). (C) The MR of female flies as CO2 produced per hour, over 24 hours of day and night (N = 24; p < .002). (D) The MR throughout a single, representative sleep bout during the night. (E) Linear regression model comparing percent change in MR versus sleep duration, binned per 5 minutes (N = 24; R2 = 0.266). Gray dashed lines indicate 95% confidence interval. One-way ANOVA comparing the initial percent change in MR at the 10-minute sleep bin to each subsequent sleep bin reveals significant differences after 35 minutes asleep (N = 24 each sleep bin; p < .05). ANOVA = analysis of variance. Figure 2— Open in new tabDownload slide MR is reduced during sleep state. Female control flies (w1118) were allowed to acclimate in the system for 24 hours. (A) MR shows an inverse pattern to their sleep (N = 24). (B) Total minutes of sleep per 12 hours of day and night for B (N = 24; p < .001). (C) The MR of female flies as CO2 produced per hour, over 24 hours of day and night (N = 24; p < .002). (D) The MR throughout a single, representative sleep bout during the night. (E) Linear regression model comparing percent change in MR versus sleep duration, binned per 5 minutes (N = 24; R2 = 0.266). Gray dashed lines indicate 95% confidence interval. One-way ANOVA comparing the initial percent change in MR at the 10-minute sleep bin to each subsequent sleep bin reveals significant differences after 35 minutes asleep (N = 24 each sleep bin; p < .05). ANOVA = analysis of variance. MR During Sleep Deprivation and Rebound Sleep To further examine the relationship between sleep and MR, we sleep deprived flies during the night (ZT12-24) and measured vCO2 during deprivation and recovery (Figure 3A). Consistent with previous findings, sleep deprivation significantly increased sleep the following day (ZT0-6) compared to nonsleep deprived controls (Figure 3B–C). MR was elevated in sleep-deprived flies during deprivation (ZT12-24), and reduced during recovery (ZT0-6), fortifying the notion that reduced MR is associated with sleep. There was a significant correlation between MR and sleep bout duration, indicating that similar to nighttime sleep in undisturbed flies, prolonged bouts of daytime sleep are associated with reduced MR (Figure 3F). Rebound sleep demonstrated a significant reduction in metabolic as sleep duration progressed beyond 35 minutes (Figure 3F), further supporting the notion that daytime rebound recapitulates physiologically similar sleep-associated metabolic changes to nighttime sleep. Figure 3— Open in new tabDownload slide MR is elevated during sleep deprivation and reduced during rebound. (A) Female control flies (w1118) were acclimated during the day (ZT0-12). Mechanical sleep deprivation was applied during the 12-hour night (ZT12-24), and recovery was assessed the following day (ZT0-12). (B) Sleep-deprived flies (N = 15; purple) sleep more during the first 6 hours of daytime following deprivation (ZT0-6) relative to undisturbed controls (N = 15; black). (C) Quantification of total sleep shows that flies were sufficiently sleep deprived during nighttime (ZT18-24; p < .0001) and demonstrated increased sleep during the recovery period (ZT0-6; p < .0001). (D) Hourly profile of MR in sleep deprived and control flies. (E) Quantification MRs demonstrates elevated MR during sleep deprivation (ZT18-24; p < .0001) and reduced MR during recovery (ZT0-6; p < .0001). MR during recovery in sleep-deprived flies is comparable to levels of control flies during normal nighttime sleep (p > .327). (F) Regression analysis comparing percent change in MR versus sleep duration, binned per 5 minutes (N = 15; R2 = 0.174). Gray dashed lines indicate 95% confidence interval. One-way ANOVA comparing the initial percent change in MR at 10-minute sleep bin to each subsequent sleep bin reveals significant differences after 35 minutes asleep (N = 15 each sleep bin; p < .05). ANOVA = analysis of variance. Figure 3— Open in new tabDownload slide MR is elevated during sleep deprivation and reduced during rebound. (A) Female control flies (w1118) were acclimated during the day (ZT0-12). Mechanical sleep deprivation was applied during the 12-hour night (ZT12-24), and recovery was assessed the following day (ZT0-12). (B) Sleep-deprived flies (N = 15; purple) sleep more during the first 6 hours of daytime following deprivation (ZT0-6) relative to undisturbed controls (N = 15; black). (C) Quantification of total sleep shows that flies were sufficiently sleep deprived during nighttime (ZT18-24; p < .0001) and demonstrated increased sleep during the recovery period (ZT0-6; p < .0001). (D) Hourly profile of MR in sleep deprived and control flies. (E) Quantification MRs demonstrates elevated MR during sleep deprivation (ZT18-24; p < .0001) and reduced MR during recovery (ZT0-6; p < .0001). MR during recovery in sleep-deprived flies is comparable to levels of control flies during normal nighttime sleep (p > .327). (F) Regression analysis comparing percent change in MR versus sleep duration, binned per 5 minutes (N = 15; R2 = 0.174). Gray dashed lines indicate 95% confidence interval. One-way ANOVA comparing the initial percent change in MR at 10-minute sleep bin to each subsequent sleep bin reveals significant differences after 35 minutes asleep (N = 15 each sleep bin; p < .05). ANOVA = analysis of variance. MR Is Reduced During Pharmacologically Induced Sleep Gamma-amino butyric acid (GABA) signaling promotes sleep in diverse species, and the GABA-A receptor agonist gabaxodol potently induces sleep in Drosophila.19,23–25 To determine the effects of pharmacologically induced sleep on MR, we housed flies on agar containing 0.1 mg/mL gaboxadol and 5% sucrose in the respirometry chambers and measured the effects on sleep and MR (Figure 4A). Consistent with previous studies, sleep was elevated in gaboxadol-treated flies compared to controls throughout the 12-hour daytime recording19 (Figure 4B and C). Notably, MR was reduced in gaboxadol-treated flies during the daytime compared to controls, confirming that pharmacologically induced sleep lowers MR (Figure 4D and E). These experiments were limited to analysis of daytime sleep, therefore, we could not determine percent change in MR of untreated w1118 flies across sleep bouts, since control flies sleep very little during the day in this paradigm. Sleep bout length in gaboxadol-treated flies was associated with reduced MR (Figure 4F). Moreover, comparison of the of percent change in MR of each subsequent sleep bin relative to the first change at 10 minutes shows a robust reduction in MR in gaboxadol-treated flies after 30 minutes of sleep (Figure 4F). Because the percent change in MR is comparable to the MRs of wild-type flies during night sleep, it is possible that pharmacologically induced daytime sleep is physiologically comparable to nighttime sleep. Figure 4— Open in new tabDownload slide Reduced MR during pharmacologically induced sleep. (A) Female w1118 flies were loaded on sucrose or sucrose containing 0.1 mg/mL gaboxadol 2 hours before lights out (ZT10), acclimated to the system for 12 hours during the night phase and were measured for 12 hours (ZT0-12) during the following day. (B) Daytime sleep was significantly elevated in gaboxadol-treated flies (green) compared to flies fed sucrose alone (black). (C) Quantification of total sleep reveals gaboxadol-treated flies (N = 15) sleep significantly longer than untreated controls (N = 14; p < .0001). (D) MR was reduced throughout the 12-hour day. (E) Quantification of mean MR reveals a significant reduction in gaboxadol-treated flies (N = 15) compared to controls (N = 14; p < .0001). (F) Linear regression of percent change in MR versus sleep duration, binned per 5 minutes (N = 15; R2 = 0.200). Gray dashed lines indicate 95% confidence interval. One-way ANOVA comparing the initial percent change in MR at the 10-minute sleep bin to each subsequent sleep bin reveals significant differences after 30 minutes asleep. (N = 15 each sleep bin; p < .05). ANOVA = analysis of variance. Figure 4— Open in new tabDownload slide Reduced MR during pharmacologically induced sleep. (A) Female w1118 flies were loaded on sucrose or sucrose containing 0.1 mg/mL gaboxadol 2 hours before lights out (ZT10), acclimated to the system for 12 hours during the night phase and were measured for 12 hours (ZT0-12) during the following day. (B) Daytime sleep was significantly elevated in gaboxadol-treated flies (green) compared to flies fed sucrose alone (black). (C) Quantification of total sleep reveals gaboxadol-treated flies (N = 15) sleep significantly longer than untreated controls (N = 14; p < .0001). (D) MR was reduced throughout the 12-hour day. (E) Quantification of mean MR reveals a significant reduction in gaboxadol-treated flies (N = 15) compared to controls (N = 14; p < .0001). (F) Linear regression of percent change in MR versus sleep duration, binned per 5 minutes (N = 15; R2 = 0.200). Gray dashed lines indicate 95% confidence interval. One-way ANOVA comparing the initial percent change in MR at the 10-minute sleep bin to each subsequent sleep bin reveals significant differences after 30 minutes asleep. (N = 15 each sleep bin; p < .05). ANOVA = analysis of variance. The Effects of Starvation on Sleep and MR In mammals, starvation potently suppresses sleep and MR.26 Further, flies suppress sleep shortly after the onset of food deprivation, presumably to increase foraging behavior.18,27 To determine how starvation-induced sleep suppression impacts metabolic function, we compared the MR of fed and starved female w1118 flies. Flies were acclimated for 12 hours on food or agar, and MR was measured during the 12-hour night phase (ZT12-24; Figure 5A). In agreement with previous findings, sleep was reduced in starved flies throughout the 12-hour recording period28,29 (Figure 5B and C). Despite the loss of sleep in starved flies, MR was lower in starved animals compared to fed counterparts, providing further support that MR in Drosophila is modulated independently from locomotor activity (Figure 5D and E). There was a significantly stronger relationship between sleep bout length and MR in fed flies, providing evidence that starvation impairs sleep-associated physiological changes on MR (Figure 5F). To determine the effect of starvation on sleep-dependent regulation of MR, we compared the MR of each sleep bout in fed and starved animals. In fed flies, MR was reduced following 40 minutes of sleep compared to the first 5 minutes of sleep, yet when starved, MR is not significantly reduced as sleep progresses, further supporting the notion that starvation impedes sleep (Figure 5F). Taken together, these findings reveal that CO2 production is reduced in starved flies without affecting sleep-dependent changes in MR. Figure 5— Open in new tabDownload slide MR and sleep are reduced in starved flies. (A) Flies were fed or starved while acclimating to the system for 12 hours during the day (ZT0-12) before measurement throughout the night (ZT12-24). (B) Flies starved on agar (blue) slept less than flies housed on 5% sucrose (black) during the 12-hour night period. (C) Quantification of total sleep over the 12-hour night period reveals a significant reduction in starved flies (N = 30) compared to fed controls (N = 29; p < .0001). (D) MR is lower throughout the 12-hour nighttime period in starved flies. (E) Quantification of mean vCO2 production over this period reveals a signification reduction in starved animals (N = 30) relative to controls (N = 29; p < .01). (F) Regression analysis comparing percent change in MR versus sleep duration, binned per 5 minutes reveals a correlation in fed flies (N = 29; R2 = 0.201), but little effect in starved flies (N = 26 each sleep bin, four flies did not have any sleep bouts when starved; R2 = 0.067). Gray dashed lines indicate 95% confidence interval of each line. Comparison of the regression lines indicate that the slopes are different between the fed versus starved state (F = 5.319; p < .05). One-way ANOVA comparing the initial percent change in MR at the 10-minute sleep bin to each subsequent sleep bin within each group reveals significant differences after 40 minutes asleep in fed flies (N = 29 each sleep bin; p < .05) and no significant differences in starved flies (N = 26 each sleep bin). ANOVA = analysis of variance. Figure 5— Open in new tabDownload slide MR and sleep are reduced in starved flies. (A) Flies were fed or starved while acclimating to the system for 12 hours during the day (ZT0-12) before measurement throughout the night (ZT12-24). (B) Flies starved on agar (blue) slept less than flies housed on 5% sucrose (black) during the 12-hour night period. (C) Quantification of total sleep over the 12-hour night period reveals a significant reduction in starved flies (N = 30) compared to fed controls (N = 29; p < .0001). (D) MR is lower throughout the 12-hour nighttime period in starved flies. (E) Quantification of mean vCO2 production over this period reveals a signification reduction in starved animals (N = 30) relative to controls (N = 29; p < .01). (F) Regression analysis comparing percent change in MR versus sleep duration, binned per 5 minutes reveals a correlation in fed flies (N = 29; R2 = 0.201), but little effect in starved flies (N = 26 each sleep bin, four flies did not have any sleep bouts when starved; R2 = 0.067). Gray dashed lines indicate 95% confidence interval of each line. Comparison of the regression lines indicate that the slopes are different between the fed versus starved state (F = 5.319; p < .05). One-way ANOVA comparing the initial percent change in MR at the 10-minute sleep bin to each subsequent sleep bin within each group reveals significant differences after 40 minutes asleep in fed flies (N = 29 each sleep bin; p < .05) and no significant differences in starved flies (N = 26 each sleep bin). ANOVA = analysis of variance. Metabolic Changes During Sleep Are Intact in translin Mutant Flies It is possible that shared genes regulate starvation-induced reductions in sleep duration and sleep-dependent regulation of MR. We previously identified the RNA-binding protein translin (trsn), as essential for starvation-induced sleep suppression.29 Energy stores and feeding behavior are normal in trsn deficient flies, yet they fail to suppress sleep in response to starvation, suggesting trsn is required for the integration of sleep and metabolic state.29 To determine whether trsn affects MR, we measured sleep and MR in fed and starved trsn mutant flies. Flies were loaded into the respirometry system and allowed to acclimate for 12 hours during the day. Sleep and MR were then measured for the duration of the night phase (ZT12-24). In agreement with previous findings, control flies robustly suppressed sleep when starved on agar, while there was no significant effect of starvation on sleep duration in trsnnull flies (Figure 6A and B). In both control and trsnnull flies, CO2 production was reduced during starvation, suggesting trsn is not required for modulating MR in accordance with feeding state. These findings fortify the notion that MR can be regulated independently from both sleep and locomotor activity (Figure 6C and D). Interestingly, while MR was further reduced in trsnnull flies upon starvation, the basal MR of fed trsnnull flies was lower than w1118 controls (Figure 6C and D). For both w1118 and trsnnull flies, there was a stronger relationship between MR and sleep bout duration in fed flies than starved flies (Figure 6E and F), fortifying the notion that sleep-dependent changes in MR are not disrupted in trsnnull flies. Together, these results indicate that trsn is required for starvation-induced sleep suppression but is dispensable for sleep-induced modulation of MR. Figure 6— Open in new tabDownload slide Sleep-dependent changes in metabolism are intact in trsnnull flies. (A) Sleep did not significantly differ between trsnnull flies housed on sucrose or starved on agar alone. (B) Quantification of total nighttime sleep (ZT12-ZT24) revealed sleep is significantly lower in w1118 control flies (N = 30) housed on agar compared to fed (N = 28; p < .0001), while there is no significant difference between trsnnull flies (N = 28) housed on 5% sucrose or agar alone (N = 25; p > .05; two-way ANOVA F(1,107) = 42.52, p < .0001). Treatment-by-genotype interaction is significant (two-way ANOVA F(1,107) = 11.24), p < .01). (C) MR is lower in control and trsnnull flies housed on agar compared to flies housed on 5% sucrose. (D) Quantification revealed MR is lower in both starved trsnnull flies and controls (w1118, p < .0001; trsnnull, p < .01; two-way ANOVA F(1,107) = 28.22, p < .0001). There is no effect of treatment-by-genotype interaction (two-way ANOVA F(1,107) = 0.7162, p > .30). (E) Applied linear regression model comparing percent change in MR versus sleep duration, binned per 5 minutes reveals a correlation in w1118 fed flies (N = 28; R2 = 0.208), but only a weak effect in w1118 starved flies (N = 25, 5 flies did not sleep on agar; R2 = 0.097). Gray dashed lines indicate 95% confidence interval of each line. Comparison of the regression lines indicate that the slopes are different between the w1118 fed versus starved state (F = 7.09725, p< .01). One-way ANOVA comparing the initial percent change in MR at the 10-minute sleep bin to each subsequent sleep bin within each group reveals significant differences after 40 minutes asleep in fed flies (N = 28 each sleep bin; p < .05) and differences in starved flies beyond 55 minutes (N = 25 each sleep bin; p < .05). (F) Regression analysis model comparing percent change in MR versus sleep duration, binned per 5 minutes reveals a correlation in trsnnull fed flies (N = 28; R2 = 0.201), but only a weak effect in trsnnull starved flies (N = 25; R2 = 0.183). Gray dashed lines indicate 95% confidence interval of each line. Comparison of the regression lines indicates that the slopes do not differ between the trsnnull fed versus starved state (F = 5.0557, p < .05). One-way ANOVA comparing the initial percent change in MR at the 10-minute sleep bin to each subsequent sleep bin within each group reveals significant differences after 25 minutes asleep in trsnnull fed flies (N = 28 each sleep bin; p < .05) and differences in trsnnull starved flies beyond 30 minutes (N = 25 each sleep bin; p < .05). ANOVA = analysis of variance. Figure 6— Open in new tabDownload slide Sleep-dependent changes in metabolism are intact in trsnnull flies. (A) Sleep did not significantly differ between trsnnull flies housed on sucrose or starved on agar alone. (B) Quantification of total nighttime sleep (ZT12-ZT24) revealed sleep is significantly lower in w1118 control flies (N = 30) housed on agar compared to fed (N = 28; p < .0001), while there is no significant difference between trsnnull flies (N = 28) housed on 5% sucrose or agar alone (N = 25; p > .05; two-way ANOVA F(1,107) = 42.52, p < .0001). Treatment-by-genotype interaction is significant (two-way ANOVA F(1,107) = 11.24), p < .01). (C) MR is lower in control and trsnnull flies housed on agar compared to flies housed on 5% sucrose. (D) Quantification revealed MR is lower in both starved trsnnull flies and controls (w1118, p < .0001; trsnnull, p < .01; two-way ANOVA F(1,107) = 28.22, p < .0001). There is no effect of treatment-by-genotype interaction (two-way ANOVA F(1,107) = 0.7162, p > .30). (E) Applied linear regression model comparing percent change in MR versus sleep duration, binned per 5 minutes reveals a correlation in w1118 fed flies (N = 28; R2 = 0.208), but only a weak effect in w1118 starved flies (N = 25, 5 flies did not sleep on agar; R2 = 0.097). Gray dashed lines indicate 95% confidence interval of each line. Comparison of the regression lines indicate that the slopes are different between the w1118 fed versus starved state (F = 7.09725, p< .01). One-way ANOVA comparing the initial percent change in MR at the 10-minute sleep bin to each subsequent sleep bin within each group reveals significant differences after 40 minutes asleep in fed flies (N = 28 each sleep bin; p < .05) and differences in starved flies beyond 55 minutes (N = 25 each sleep bin; p < .05). (F) Regression analysis model comparing percent change in MR versus sleep duration, binned per 5 minutes reveals a correlation in trsnnull fed flies (N = 28; R2 = 0.201), but only a weak effect in trsnnull starved flies (N = 25; R2 = 0.183). Gray dashed lines indicate 95% confidence interval of each line. Comparison of the regression lines indicates that the slopes do not differ between the trsnnull fed versus starved state (F = 5.0557, p < .05). One-way ANOVA comparing the initial percent change in MR at the 10-minute sleep bin to each subsequent sleep bin within each group reveals significant differences after 25 minutes asleep in trsnnull fed flies (N = 28 each sleep bin; p < .05) and differences in trsnnull starved flies beyond 30 minutes (N = 25 each sleep bin; p < .05). ANOVA = analysis of variance. DISCUSSION MR is regulated in accordance with environmental changes and life history, providing a metric for whole-body metabolic function. While mammalian studies typically determine MR via O2 consumption or respiratory quotient (ratio of CO2 eliminated/O2 consumed), studies in Drosophila commonly measure CO2 production because it is directly correlated with O2 input and accurately reflects MR.14,16,30 Previous systems investigating MR in Drosophila have used single flies or populations to measure changes in CO2 in response to aging, temperature change, and dietary restriction.16,31,32 Here, we have modified a previously described single-fly respirometry system and DAM system to simultaneously measure MR and sleep. This system can measure CO2 production and locomotor activity over a 24-hour period, providing the ability to measure the relationship between sleep and metabolism, providing a system to investigate the complex relationships between diverse genetic and environmental factors with MR. Sleep-Metabolism Interactions in Mammals and Arthropods Regulation of sleep and metabolism is conserved at the molecular and physiological levels between Drosophila and mammals.33,34 Similar to mammals, flies modulate sleep and feeding in accordance with metabolic state, providing a system to investigate the genetic underpinnings of these behaviors.2 For instance, when starved, both flies and mammals suppress sleep presumably to forage for food.18,35,36 The finding that sleep-dependent reductions in MR are conserved in Drosophila supports the notion that an essential function of sleep is metabolic regulation. A number of previous studies suggest total sleep duration is positively correlated with basal MR, supporting the notion that sleep may be an adaptive mechanism of energy conservation.37,38 However, a meta-analysis study examining over 40 different mammalian species revealed a negative relationship between sleep and basal MR, opposing the energy conservation model of sleep.39 In humans, reduced MR during sleep accounts for as much as a 15% energy savings.40–42 Our findings reveal a similar reduction of MR during sleep in fruit flies, suggesting this may provide an evolutionarily adaptive mechanism to conserve energy. While our study is the first to examine the relationship between sleep and MR in Drosophila, previous studies implicated shared genetic or environmental factors in the regulation of sleep and metabolic function. For example, dopamine potently suppresses sleep in Drosophila, and flies harboring a mutation in the dopamine transporter gene fumin (fmn) exhibit reduced sleep and increased CO2 production,43,44 suggesting dopamine regulates both sleep and metabolic state. Importantly, MR remained elevated in fumin mutant flies when motor neurons were genetically silenced, indicating that the elevated MR does not result from differences in locomotor activity.44 Moreover, long-term sleep deprivation in Pacific beetle cockroach, Diploptera punctate, caused significant increases in O2 consumption and elevated basal MR compared to controls, indicating that sleep loss impacts metabolism.45 These data are in agreement with our findings, where we identify increased MR during sleep deprivation and reduced MR during sleep rebound or pharmacologically induced sleep, revealing a fundamental and direct relationship between sleep and lower MR, indicating that changes in CO2 production during sleep are due to changes in basal MR, rather than reduced locomotor activity. Environmental Factors Regulating Sleep and Metabolism In Drosophila, sleep and MR are influenced by diet, temperature, and age.46,47 Here, we discover that starvation conditions impede the physiological changes associated with normal sleep. This simultaneous assessment of sleep and metabolic state can be applied to determine how MR and sleep are related to starvation resistance. Selection for starvation-resistant Drosophila through experimental evolution results in flies that can survive over 2 weeks without food and exhibit a host of metabolic and developmental differences, thus providing a system to examine interactions between metabolism and behavior.48,49 The starvation-resistant flies exhibit increased body size, energy stores, and reduced MR, providing numerous mechanisms for energy conservation.49–51 Previously, we reported that sleep duration is increased in starvation-resistant flies and proposed that this provides an additional mechanism for energy conservation.50,52 Application of this approach measuring sleep and MR will provide the ability to determine MR in asleep and awake flies and identify whether reduced MR in starvation-resistant flies is a consequence of increased sleep or these traits have evolved in parallel. Evidence for Sleep-Associated Regulation of MRs in Drosophila In birds and mammals, sleep is associated with changes in cortical activity resulting in defined stages of sleep, such as rapid eye movement and nonrapid eye movement, which differ in physiology and function.53,54 In Drosophila, sleep studies have primarily used behavioral quiescence and body postures to denote sleep, and much less is known about how sleep impacts physiology. Recording of local field potentials in tethered animals reveals distinct differences between quiescent and active states, and sleep is associated with a reduction in 15–30 Hz local-field potentials. The reduction in 15–30 Hz oscillations is at its greatest following 15 minutes of immobility, suggesting that this physiological change in neuronal activity represents a deeper form of sleep, along with coordinate increases in arousal threshold.9,10 Evaluating sleep intensity in male and female Drosophila using an arousal-testing paradigm during extended nighttime sleep bouts identified a gradual decrease in responsiveness until a second, deeper sleep state was reached after ~30 minutes.10,55 Consistent with these findings, we report that MR decreases with sleep duration, reaching a significant reduction ~30 minutes following sleep onset. Taken together, these findings suggest MR may provide a physiological indicator of sleep intensity that compliments existing electrophysiology and behavioral methods of analysis to define a deeper sleep state in flies. A Role for translin in Regulating MR The RNA-binding protein trsn is a proposed integrator of sleep and metabolic state, and flies deficient for trsn fail to suppress sleep in response to starvation.29 Notably, the defect in trsn-mutant flies is specific to regulation sleep regulation because trsn-deficient flies have normal feeding and energy stores.29 Here, we find that starvation reduces MR in trsn mutant and wild-type flies. Even though trsn mutants demonstrate starvation-induced sleep suppression, our findings indicate that MR can still be modulated in trsn mutants in a starved state. We identify a lower basal MR in fed trsn mutants compared to controls, though this may be attributable to the trend toward increased nighttime sleep in trsn mutant flies. In addition to trsn, a number of additional genes and transmitters have been identified as regulating starvation-induced sleep suppression or hyperactivity, including Octopamine, clock, and the glucagon-like adipokinetic hormone.35,56,57 Therefore, this assay provides a direct readout of metabolic response to starvation and can be used for more detailed investigation of the mechanisms underlying the integration of sleep and metabolic state. Future Applications for Investigation of MR in Drosophila Beyond our initial analysis of the relationship between sleep and MR in Drosophila, this system allows for genetic screens or targeted genetic manipulations to identify novel genes regulating sleep, MR, and the integration of these processes. For example, the mushroom bodies, fan-shaped body, and circadian neurons modulate sleep and wakefulness in Drosophila,23,58–61 and the effects of manipulating these systems on sleep-dependent modulation of MR could be measured using this system. Similarly, application of this system could identify novel genes, neurons or environmental factors required for changes in MR during sleep. Recent studies in flies have identified neural circuits involved in sleep homeostasis,62–64 yet little is known about the physiological changes associated with rebound sleep in flies. In addition to sleep, the circadian system regulates metabolism in flies and mammals.65,66 In addition to measuring MR, respirometry can also be used to measure specific molecules being metabolized. The simultaneous measurement of CO2 and O2 enables the ability to identify the exchange ratios of O2 consumed and CO2 produced and ultimately infer the specific energy fuels utilized.67,68 More specifically, the ratio between the CO2 produced and O2 consumed at a steady state, also known as the respirometry quotient (RQ) or the respirometry exchange ratio which quantifies the same ratio but at any time point (eg, during exercise), can be used to identify food sources metabolized including fat (RQ = ~0.7), carbohydrates (RQ = ~1.0), or protein (RQ = ~0.8–0.9).14 Respirometry measurements have been used to identify substrates metabolized in both mammals69,70 and invertebrates.16 However, the sensitivity of O2 detection is lower than CO2, thus preventing detection of O2 changes in single flies, yet recent studies indicate that sleep can be measured in group-housed Drosophila.71 Therefore, it may be feasible for future studies utilizing groups of flies to determine metabolized energy stores using this system. CONCLUSIONS We describe a system for simultaneously measuring sleep and MR and, further, identify dynamic regulation of MR during individual sleep bouts. This system denotes MR as a readily identifiable marker of the physiological changes associated with sleep, which can be universally applied to examine the function of novel sleep genes and neurons in Drosophila. Ultimately, this unique system can be applied to examine precise interactions between numerous aspects of life history and circadian function coordinately with MR. SUPPLEMENTARY MATERIAL Supplementary material is available at SLEEP online. FUNDING This work was supported by NIH grants 1R01NS085252 to ACK and R15NS080155 to ACK and JRD. DISCLOSURE STATEMENT None declared. ACKNOWLEGMENTS The authors are grateful to Dr. Allen Gibbs (University of Nevada, Las Vegas) for initial guidance in establishing the respirometry system, as well as Mark Spencer (Trikinetics) and Dr. Thomas Foerster (Sable Systems), for technical advice and assistance designing this system. Dr. Paul Shaw (Washington University, St. Louis) provided critical experimental suggestions and feedback. References 1. Reutrakul S Van Cauter E . Interactions between sleep, circadian function, and glucose metabolism: implications for risk and severity of diabetes . Ann N Y Acad Sci . 2014 ; 1311 : 151 – 173 . Google Scholar Crossref Search ADS PubMed WorldCat 2. Yurgel M Masek P DiAngelo JR Keene A . Genetic dissection of sleep-metabolism interactions in the fruit fly . J Comp Physiol a Neuroethol Sens Neural Behav Physiol . 2015; 201 ( 9 ) : 869 – 877 . Google Scholar Crossref Search ADS PubMed WorldCat 3. Schmidt MH . The energy allocation function of sleep: a unifying theory of sleep, torpor, and continuous wakefulness . Neurosci Biobehav Rev . 2014 ; 47 : 122 – 153 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Berger RJ Phillips NH . Energy conservation and sleep . Behav Brain Res . 1995 ; 69 ( 1-2 ): 65 – 73 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Walker JM Berger RJ . Sleep as an adaptation for energy conservation functionally related to hibernation and shallow torpor . Prog Brain Res . 1980 ; 53 : 255 – 278 . Google Scholar Crossref Search ADS PubMed WorldCat 6. Hendricks JC Finn SM Panckeri KA et al. . Rest in Drosophila is a sleep-like state . Neuron . 2000 ; 25 ( 1 ): 129 – 138 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Shaw PJ Cirelli C Greenspan RJ Tononi G . Correlates of sleep and waking in Drosophila melanogaster . Science . 2000 ; 287 ( 5459 ): 1834 – 1837 . Google Scholar Crossref Search ADS PubMed WorldCat 8. Campbell SS Tobler I . Animal sleep: a review of sleep duration across phylogeny . Neurosci Biobehav Rev . 1984 ; 8 ( 3 ): 269 – 300 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Nitz DA van Swinderen B Tononi G Greenspan RJ . Electrophysiological correlates of rest and activity in Drosophila melanogaster . Curr Biol . 2002 ; 12 ( 22 ): 1934 – 1940 . Google Scholar Crossref Search ADS PubMed WorldCat 10. van Alphen B Yap MH Kirszenblat L Kottler B van Swinderen B . A dynamic deep sleep stage in Drosophila . J Neurosci . 2013 ; 33 ( 16 ): 6917 – 6927 . Google Scholar Crossref Search ADS PubMed WorldCat 11. Caron AM Stephenson R . Energy expenditure is affected by rate of accumulation of sleep deficit in rats . Sleep . 2010 ; 33 ( 9 ): 1226 – 1235 . Google Scholar Crossref Search ADS PubMed WorldCat 12. Valenti G Bonomi AG Westerterp KR . Quality sleep is associated with overnight metabolic rate in healthy older adults . J Gerontol A Biol Sci Med Sci . 2016 ; 72 ( 4 ): 567 – 571 . WorldCat 13. Spaeth AM Dinges DF Goel N . Resting metabolic rate varies by race and by sleep duration . Obesity . 2015 ; 23 ( 12 ): 2349 – 2356 . Google Scholar Crossref Search ADS PubMed WorldCat 14. Lighton J . Measuring Metabolic Rates: A Manual for Scientists . Oxford University Press; 2008 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC 15. Claussen M Koch R Jin ZY Suter B . Functional characterization of Drosophila Translin and Trax . Genetics . 2006 ; 174 ( 3 ): 1337 – 1347 . Google Scholar Crossref Search ADS PubMed WorldCat 16. Van Voorhies WA Khazaeli AA Curtsinger JW . Testing the “rate of living” model: further evidence that longevity and metabolic rate are not inversely correlated in Drosophila melanogaster . J Appl Physiol (1985) . 2004 ; 97 ( 5 ): 1915 – 1922 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Pfeiffenberger C Lear BC Keegan KP Allada R . Locomotor activity level monitoring using the Drosophila Activity Monitoring (DAM) system . Cold Spring Harb Protoc . 2010 ; 2010 ( 11 ): pdb.prot5518 . Google Scholar PubMed WorldCat 18. Keene AC Duboué ER McDonald DM et al. . Clock and cycle limit starvation-induced sleep loss in Drosophila . Curr Biol . 2010 ; 20 ( 13 ): 1209 – 1215 . Google Scholar Crossref Search ADS PubMed WorldCat 19. Dissel S Angadi V Kirszenblat L et al. . Sleep restores behavioral plasticity to Drosophila mutants . Curr Biol . 2015 ; 25 ( 10 ): 1270 – 1281 . Google Scholar Crossref Search ADS PubMed WorldCat 20. Garbe DS Bollinger WL Vigderman A et al. . Context-specific comparison of sleep acquisition systems in Drosophila . Biol Open . 2015 ; 4 ( 11 ): 1558 – 1568 . Google Scholar Crossref Search ADS PubMed WorldCat 21. Heusner AA . Energy metabolism and body size. I. Is the 0.75 mass exponent of Kleiber’s equation a statistical artifact? Respir Physiol . 1982 ; 48 ( 1 ): 1 – 12 . Google Scholar Crossref Search ADS PubMed WorldCat 22. Schilman PE Waters JS Harrison JF Lighton JR . Effects of temperature on responses to anoxia and oxygen reperfusion in Drosophila melanogaster . J Exp Biol . 2011 ; 214 ( Pt 8 ): 1271 – 1275 . Google Scholar Crossref Search ADS PubMed WorldCat 23. Parisky KM Agosto J Pulver SR et al. . PDF cells are a GABA-responsive wake-promoting component of the Drosophila sleep circuit . Neuron . 2008 ; 60 ( 4 ): 672 – 682 . Google Scholar Crossref Search ADS PubMed WorldCat 24. Agosto J Choi JC Parisky KM Stilwell G Rosbash M Griffith LC . Modulation of GABAA receptor desensitization uncouples sleep onset and maintenance in Drosophila . Nat Neurosci . 2008 ; 11 ( 3 ): 354 – 359 . Google Scholar Crossref Search ADS PubMed WorldCat 25. Wafford KA Ebert B . Gaboxadol—a new awakening in sleep . Curr Opin Pharmacol . 2006 ; 6 ( 1 ): 30 – 36 . Google Scholar Crossref Search ADS PubMed WorldCat 26. Adamantidis A de Lecea L . Sleep and metabolism: shared circuits, new connections . Trends Endocrinol Metab . 2008 ; 19 ( 10 ): 362 – 370 . Google Scholar Crossref Search ADS PubMed WorldCat 27. Linford NJ Chan TP Pletcher SD . Re-patterning sleep architecture in Drosophila through gustatory perception and nutritional quality . PLoS Genet . 2012 ; 8 ( 5 ): e1002668 . Google Scholar Crossref Search ADS PubMed WorldCat 28. McDonald DM Keene AC . The sleep-feeding conflict: understanding behavioral integration through genetic analysis in Drosophila . Aging (Albany NY) . 2010 ; 2 ( 8 ): 519 – 522 . Google Scholar Crossref Search ADS PubMed WorldCat 29. Murakami K Yurgel ME Stahl BA et al. . Translin is required for metabolic regulation of sleep . Curr Biol . 2016 ; 26 ( 7 ): 972 – 980 . Google Scholar Crossref Search ADS PubMed WorldCat 30. Yatsenko AS Marrone AK Kucherenko MM Shcherbata HR . Measurement of metabolic rate in Drosophila using respirometry . J. Vis. Exp . 2014 ;( 88 ): e51681 . doi: 10.3791/51681 . WorldCat 31. Mölich AB Förster TD Lighton JR . Hyperthermic overdrive: oxygen delivery does not limit thermal tolerance in Drosophila melanogaster . J Insect Sci . 2012 ; 12 : 109 . Google Scholar Crossref Search ADS PubMed WorldCat 32. Hulbert AJ Clancy DJ Mair W Braeckman BP Gems D Partridge L . Metabolic rate is not reduced by dietary-restriction or by lowered insulin/IGF-1 signalling and is not correlated with individual lifespan in Drosophila melanogaster . Exp Gerontol . 2004 ; 39 ( 8 ): 1137 – 1143 . Google Scholar Crossref Search ADS PubMed WorldCat 33. Allada R Siegel JM . Unearthing the phylogenetic roots of sleep . Curr Biol . 2008 ; 18 ( 15 ): R670 – R679 . Google Scholar Crossref Search ADS PubMed WorldCat 34. Padmanabha D Baker KD . Drosophila gains traction as a repurposed tool to investigate metabolism . Trends Endocrinol. Metab . 2014 : 1 – 10 . WorldCat 35. Lee G Park JH . Hemolymph sugar homeostasis and starvation-induced hyperactivity affected by genetic manipulations of the adipokinetic hormone-encoding gene in Drosophila melanogaster . Genetics . 2004 ; 167 ( 1 ): 311 – 323 . Google Scholar Crossref Search ADS PubMed WorldCat 36. Danguir J Nicolaidis S . Dependence of sleep on nutrients’ availability . Physiol Behav . 1979 ; 22 ( 4 ): 735 – 740 . Google Scholar Crossref Search ADS PubMed WorldCat 37. Elgar MA Pagel MD Harvey PH . Sleep in mammals . Anim. Behav . 1988 ; 36 ( 5 ): 1407 – 1419 . Google Scholar Crossref Search ADS WorldCat 38. Lesku JA Roth TC 2nd Amlaner CJ Lima SL . A phylogenetic analysis of sleep architecture in mammals: the integration of anatomy, physiology, and ecology . Am Nat . 2006 ; 168 ( 4 ): 441 – 453 . Google Scholar Crossref Search ADS PubMed WorldCat 39. Capellini I Nunn CL McNamara P Preston BT Barton RA . Energetic constraints, not predation, influence the evolution of sleep patterning in mammals . Funct Ecol . 2008 ; 22 ( 5 ): 847 – 853 . Google Scholar Crossref Search ADS PubMed WorldCat 40. Shapiro CM Goll CC Cohen GR Oswald I . Heat production during sleep . J Appl Physiol Respir Environ Exerc Physiol . 1984 ; 56 ( 3 ): 671 – 677 . Google Scholar PubMed WorldCat 41. Garby L Kurzer MS Lammert O Nielsen E . Energy expenditure during sleep in men and women: evaporative and sensible heat losses . Hum Nutr Clin Nutr . 1987 ; 41 ( 3 ): 225 – 233 . Google Scholar PubMed WorldCat 42. Goldberg GR, Prentice AM, Davies HL, Murgatroyd PR . Overnight and basal metabolic rates in men and women . Eur. J. Clin. Nutr . 1988 ; 42 ( 2 ): 137 – 144 . Google Scholar PubMed WorldCat 43. Kume K Kume S Park SK Hirsh J Jackson FR . Dopamine is a regulator of arousal in the fruit fly . J Neurosci . 2005 ; 25 ( 32 ): 7377 – 7384 . Google Scholar Crossref Search ADS PubMed WorldCat 44. Ueno T Tomita J Kume S Kume K . Dopamine modulates metabolic rate and temperature sensitivity in Drosophila melanogaster . PLoS One 2012 ; 7 ( 2 ). doi: 10.1371/journal.pone.0031513 . WorldCat 45. Stephenson R Chu KM Lee J . Prolonged deprivation of sleep-like rest raises metabolic rate in the Pacific beetle cockroach, Diploptera punctata (Eschscholtz) . J Exp Biol . 2007 ; 210 ( Pt 14 ): 2540 – 2547 . Google Scholar Crossref Search ADS PubMed WorldCat 46. Yurgel ME Masek P DiAngelo J Keene AC . Genetic dissection of sleep-metabolism interactions in the fruit fly . J Comp Physiol a Neuroethol Sens Neural Behav Physiol . 2015 ; 201 ( 9 ): 869 – 877 . Google Scholar Crossref Search ADS PubMed WorldCat 47. Griffith LC . Neuromodulatory control of sleep in Drosophila melanogaster: integration of competing and complementary behaviors . Curr Opin Neurobiol . 2013 ; 23 ( 5 ): 819 – 823 . Google Scholar Crossref Search ADS PubMed WorldCat 48. Schwasinger-Schmidt TE Kachman SD Harshman LG . Evolution of starvation resistance in Drosophila melanogaster: measurement of direct and correlated responses to artificial selection . J Evol Biol . 2012 ; 25 ( 2 ): 378 – 387 . Google Scholar Crossref Search ADS PubMed WorldCat 49. Masek P Reynolds LA Bollinger WL et al. . Altered regulation of sleep and feeding contributes to starvation resistance in Drosophila melanogaster . J Exp Biol . 2014 ; 217 ( Pt 17 ): 3122 – 3132 . Google Scholar Crossref Search ADS PubMed WorldCat 50. Slocumb ME Regalado JM Yoshizawa M et al. . Enhanced sleep is an evolutionarily adaptive response to starvation stress in Drosophila . PLoS One . 2015 ; 10 ( 7 ): e0131275 . Google Scholar Crossref Search ADS PubMed WorldCat 51. Harshman LG Hoffmann AA Clark AG . Selection for starvation resistance in Drosophila melanogaster: physiological correlates, enzyme activities and multiple stress responses . J. Evol. Biol . 1999 ; 12 : 370 – 379 . Google Scholar Crossref Search ADS WorldCat 52. Masek P Reynolds L a Bollinger WL et al. . Altered regulation of sleep and feeding contributes to starvation resistance in Drosophila melanogaster . J Exp Biol . 2014 ; 217 ( Pt 17 ): 3122 – 3132 . Google Scholar Crossref Search ADS PubMed WorldCat 53. Newman SM Paletz EM Rattenborg NC Obermeyer WH Benca RM . Sleep deprivation in the pigeon using the Disk-Over-Water method . Physiol Behav . 2008 ; 93 ( 1-2 ): 50 – 58 . Google Scholar Crossref Search ADS PubMed WorldCat 54. Siegel JM . Clues to the functions of mammalian sleep . Nature . 2005 ; 437 ( 7063 ): 1264 – 1271 . Google Scholar Crossref Search ADS PubMed WorldCat 55. Faville R Kottler B Goodhill GJ Shaw PJ van Swinderen B . How deeply does your mutant sleep? Probing arousal to better understand sleep defects in Drosophila . Sci Rep . 2015 ; 5 : 8454 . Google Scholar Crossref Search ADS PubMed WorldCat 56. Crocker A Shahidullah M Levitan IB Sehgal A . Identification of a neural circuit that underlies the effects of octopamine on sleep:wake behavior . Neuron . 2010 ; 65 ( 5 ): 670 – 681 . Google Scholar Crossref Search ADS PubMed WorldCat 57. Keene AC Duboué ER McDonald DM et al. . Clock and cycle limit starvation-induced sleep loss in Drosophila . Curr Biol . 2010 ; 20 ( 13 ): 1209 – 1215 . Google Scholar Crossref Search ADS PubMed WorldCat 58. Joiner WJ Crocker A White BH Sehgal A . Sleep in Drosophila is regulated by adult mushroom bodies . Nature . 2006 ; 441 ( 7094 ): 757 – 760 . Google Scholar Crossref Search ADS PubMed WorldCat 59. Pitman JL McGill JJ Keegan KP Allada R . A dynamic role for the mushroom bodies in promoting sleep in Drosophila . Nature . 2006 ; 441 ( 7094 ): 753 – 756 . Google Scholar Crossref Search ADS PubMed WorldCat 60. Donlea JM Thimgan MS Suzuki Y Gottschalk L Shaw PJ . Inducing sleep by remote control facilitates memory consolidation in Drosophila . Science . 2011 ; 332 ( 6037 ): 1571 – 1576 . Google Scholar Crossref Search ADS PubMed WorldCat 61. Guo F Yu J Jung HJ et al. . Circadian neuron feedback controls the Drosophila sleep–activity profile . Nature . 2016 ; 536 ( 7616 ): 292 – 297 . Google Scholar Crossref Search ADS PubMed WorldCat 62. Liu S Liu Q Tabuchi M Wu MN . Sleep drive is encoded by neural plastic changes in a dedicated circuit . Cell . 2016 ; 165 ( 6 ): 1347 – 1360 . Google Scholar Crossref Search ADS PubMed WorldCat 63. Seidner G Robinson JE Wu M et al. . Identification of neurons with a privileged role in sleep homeostasis in Drosophila melanogaster . Curr Biol . 2015 ; 25 ( 22 ): 2928 – 2938 . Google Scholar Crossref Search ADS PubMed WorldCat 64. Donlea JM Pimentel D Miesenböck G . Neuronal machinery of sleep homeostasis in Drosophila . Neuron . 2014 ; 81 ( 4 ): 860 – 872 . Google Scholar Crossref Search ADS PubMed WorldCat 65. Xu K DiAngelo JR Hughes ME Hogenesch JB Sehgal A . The circadian clock interacts with metabolic physiology to influence reproductive fitness . Cell Metab . 2011 ; 13 ( 6 ): 639 – 654 . Google Scholar Crossref Search ADS PubMed WorldCat 66. Lin JD Liu C Li S . Integration of energy metabolism and the mammalian clock . Cell Cycle . 2008 ; 7 ( 4 ): 453 – 457 . Google Scholar Crossref Search ADS PubMed WorldCat 67. Arrese E Soulages J . Insect fat body: energy, metabolism, and regulation . Annu. Rev. Entomol . 2010 ; 55 : 207 – 225 . Google Scholar Crossref Search ADS PubMed WorldCat 68. Marron MT Markow TA Kain KJ Gibbs AG . Effects of starvation and desiccation on energy metabolism in desert and mesic Drosophila . J Insect Physiol . 2003 ; 49 ( 3 ): 261 – 270 . Google Scholar Crossref Search ADS PubMed WorldCat 69. Melanson EL Ingebrigtsen JP Bergouignan A Ohkawara K Kohrt WM Lighton JR . A new approach for flow-through respirometry measurements in humans . Am J Physiol Regul Integr Comp Physiol . 2010 ; 298 ( 6 ): R1571 – R1579 . Google Scholar Crossref Search ADS PubMed WorldCat 70. Sinitskaya N Gourmelen S Schuster-Klein C Guardiola-Lemaitre B Pévet P Challet E . Increasing the fat-to-carbohydrate ratio in a high-fat diet prevents the development of obesity but not a prediabetic state in rats . Clin Sci (Lond) . 2007 ; 113 ( 10 ): 417 – 425 . Google Scholar Crossref Search ADS PubMed WorldCat 71. Liu C Haynes PR Donelson NC Aharon S Griffith LC . Sleep in populations of Drosophila melanogaster . eNeuro 2015 ; 2 ( 4 ): doi: 10.1523/ENEURO.0071-15.2015 . WorldCat Author notes * These authors contributed equally. Address correspondence to: Alex C. Keene, Department of Biological Sciences, Florida Atlantic University, 5353 Parkside Drive, Jupiter, FL 33458, USA. Tel: +561-799-8053; fax: +561-799-8061; Email: [email protected] © Sleep Research Society 2017. Published by Oxford University Press on behalf of the Sleep Research Society. All rights reserved. For permissions, please e-mail [email protected].

Jozwiak,, Natalia;Postuma, Ronald, B;Montplaisir,, Jacques;Latreille,, Véronique;Panisset,, Michel;Chouinard,, Sylvain;Bourgouin,, Pierre-Alexandre;Gagnon,, Jean-François

doi: 10.1093/sleep/zsx101pmid: 28645156

Abstract Study Objectives REM sleep behavior disorder (RBD) is a parasomnia affecting 33% to 46% of patients with Parkinson’s disease (PD). The existence of a unique and specific impaired cognitive profile in PD patients with RBD is still controversial. We extensively assessed cognitive functions to identify whether RBD is associated with more severe cognitive deficits in nondemented patients with PD. Methods One hundred sixty-two participants, including 53 PD patients with RBD, 40 PD patients without RBD, and 69 healthy subjects, underwent polysomnography, a neurological assessment and an extensive neuropsychological exam to assess attention, executive functions, episodic learning and memory, visuospatial abilities, and language. Results PD patients with RBD had poorer and clinically impaired performance in several cognitive tests compared to PD patients without RBD and healthy subjects. These two latter groups were similar on all cognitive measures. Mild cognitive impairment (MCI) diagnosis frequency was almost threefold higher in PD patients with RBD compared to PD patients without RBD (66% vs. 23%, p < .001). Moreover, subjective cognitive decline was reported in 89% of PD patients with RBD compared to 58% of PD patients without RBD (p = .024). Conclusions RBD in PD is associated with a more impaired cognitive profile and higher MCI diagnosis frequency, suggesting more severe and widespread neurodegeneration. This patient subgroup and their caregivers should receive targeted medical attention to better detect and monitor impairment and to enable the development of management interventions for cognitive decline and its consequences. Parkinson’s disease, REM sleep behavior disorder, Neuropsychology, Cognition, Mild cognitive impairment Statement of Significance This study shows that REM sleep behavior disorder is a major risk factor for poorer cognitive performance and mild cognitive impairment in Parkinson’s disease (PD). Therefore, this subgroup of patients should be targeted in future clinical trials on the progression of cognitive decline in PD. INTRODUCTION Parkinson’s disease (PD) is a neurodegenerative condition characterized by motor and nonmotor symptoms. Cognitive decline and sleep dysfunction are among the most common of these symptoms, with major consequences for patients as well as relatives and caregivers.1 The main cognitive domains affected in PD are attention, executive functions, episodic learning and memory, and visuospatial abilities.2,3 Cross-sectional studies have reported a 30% prevalence of dementia in PD,2–5 while longitudinal studies have shown that 75% to 80% of PD patients may develop cognitive impairment within 15 to 20 years of disease onset.4,6 Clinical risk factors for dementia and cognitive decline that have been identified in PD include age, disease duration, mild cognitive impairment (MCI), subjective cognitive decline (SCD), REM sleep behavior disorder (RBD), and orthostatic hypotension.2,7,8 RBD is a parasomnia characterized by loss of REM sleep muscle atonia, resulting in undesirable motor activity during REM sleep as people “act out their dreams.” RBD prevalence in PD ranges from 33% to 46% when diagnosed with polysomnography (PSG).9,10 Although RBD is a risk factor for dementia in PD,7,11 there is a lack of consensus in the literature on the existence of a distinct cognitive profile in nondemented PD patients based on the presence of RBD.1 Indeed, some studies have associated the presence of RBD in PD with more severe cognitive impairment and higher MCI frequency,12–19 whereas others have not.20–26 However, most of these studies have methodological limitations that could explain the divergent results, including the use of screening tests with poor sensitivity to measure cognition,17,18,21–23,25 absence of PSG to diagnose RBD,12,14,16,18,19,21–23,25 small sample size,13–15,18,20,24 absence of MCI diagnosis,12,15–18,20–26 or absence of a healthy control group to better interpret the results.12,14,16–18,20–25 We previously reported neuropsychological findings in a small sample cohort of PD patients.13 In the present study, we used an extensive neuropsychological assessment to investigate cognition and PSG exams to confirm RBD in three PD cohorts (original, replication, and combined), and we included a healthy control group without cognitive impairment for comparison. We also investigated MCI diagnosis frequency in the combined PD cohort using the MCI diagnostic criteria proposed by the Movement Disorder Society Task Force.27 Moreover, we performed additional analyses to determine the frequency of SCD in PD-RBD compared to PD-nRBD and to explore the association between cognition, gender, and RBD onset in PD patients with RBD. METHODS Participants One hundred eighty-six subjects participated in the study. PD patients were recruited at the Department of Neurology of the Montreal General Hospital and the Unité des troubles du mouvement André Barbeau of the Centre Hospitalier de l’Université de Montréal to participate in a study on sleep and cognition in PD. All PD patients but five (all idiopathic RBD patients seen in our sleep clinic who developed PD during the follow-up) were consecutive patients seen at their annual assessment in a movement disorder clinic and were referred by a neurologist (RBP, SC, or MP) for this study regardless of the patient’s sleep and cognitive complaints. Inclusion criteria for PD patients were: (1) a diagnosis of probable idiopathic PD confirmed by a neurologist specialized in movement disorders,28 (2) age from 40 to 80 years, and (3) at least 6 years of schooling (completed elementary school). Exclusion criteria were: (1) parkinsonism of other cause than PD; (2) presence of dementia according to the neuropsychological assessment and neurological exam; (3) a major psychiatric disorder (including major depression, schizophrenia, bipolar disorder) according to the criteria of the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision;29 (4) a respiratory event index (apneas plus hypopneas) ≥ 20; (5) history of head injury, brain tumor, encephalitis, stroke, unstable hypertension, diabetes, or chronic obstructive pulmonary disease; and (6) abnormal electroencephalography (EEG) suggesting epilepsy. Dopaminergic medication was converted to levodopa dose equivalents. Details on medication are presented in Table 1. Control subjects without PD or cognitive impairment were recruited through a newspaper advertisement or by word of mouth and were subject to the same inclusion and exclusion criteria. The protocol was approved by a hospital–university ethics committee and participants gave their written informed consent to participate. Table 1 Demographic, Clinical, and Mood Characteristics: Combined PD Cohorts and Controls. Variables Combined PD-RBD (A) Combined PD-nRBD (B) Controls (C) p value; post hoc (n = 53) (n = 40) (n = 69) Sex men, n (%) 40 (75) 21 (53) 38 (55) .03; A > B*, A > C* Age, years 68.0 ± 8.4 63.2 ± 8.5 63.3 ± 10.3 .01; A > B*, A > C** Education, years 14.5 ± 3.9 15.1 ± 3.0 14.5 ± 2.8 ns RBD duration, years 4.1 ± 3.5 — — — PD duration since diagnosis, years 6.1 ± 4.5 6.1 ± 4.3 — ns Hoehn and Yahr stage 2.5 ± 0.8 2.2 ± 0.9 — ns UPDRS part III “on” 23.1 ± 9.5 20.3 ± 9.8 — ns REM tonic EMG, % 61.3 ± 34.3 17.1 ± 23.9 — .000 REM phasic EMG, % 31.6 ± 20.9 14.3 ± 12.1 — .000 Levodopa equivalent dosage, mg 492.2 ± 402.3 383.7 ± 283.3 — ns  Levodopa use, n (%) 42 (79) 32 (80) — ns  Dopamine agonist use, n (%) 25 (46) 28 (78) — .004 Antidepressant use, n (%) 10 (19) 7 (19) — ns Antianxiolitic use, n (%) 19 (35) 7 (19) — ns ISI scores 11.6 ± 6.2 10.8 ± 7.8 6.6 ± 4.4 .001; A > C***, B > C** ESS scores 9.3 ± 4.9 9.5 ± 4.9 6.9 ± 4.0 .008; A > C*, B > C* BDI-II scores 10.9 ± 5.8 10.7 ± 7.5 5.9 ± 5.7 .000; A > C***, B > C*** BAI scores 11.4 ± 8.7 9.6 ± 6.3 5.4 ± 5.4 .001; A > C***, B > C* MMSE scores 28.0 ± 1.9 29.2 ± 0.9 29.3 ± 1.1 .000; A < B***, A < C*** Variables Combined PD-RBD (A) Combined PD-nRBD (B) Controls (C) p value; post hoc (n = 53) (n = 40) (n = 69) Sex men, n (%) 40 (75) 21 (53) 38 (55) .03; A > B*, A > C* Age, years 68.0 ± 8.4 63.2 ± 8.5 63.3 ± 10.3 .01; A > B*, A > C** Education, years 14.5 ± 3.9 15.1 ± 3.0 14.5 ± 2.8 ns RBD duration, years 4.1 ± 3.5 — — — PD duration since diagnosis, years 6.1 ± 4.5 6.1 ± 4.3 — ns Hoehn and Yahr stage 2.5 ± 0.8 2.2 ± 0.9 — ns UPDRS part III “on” 23.1 ± 9.5 20.3 ± 9.8 — ns REM tonic EMG, % 61.3 ± 34.3 17.1 ± 23.9 — .000 REM phasic EMG, % 31.6 ± 20.9 14.3 ± 12.1 — .000 Levodopa equivalent dosage, mg 492.2 ± 402.3 383.7 ± 283.3 — ns  Levodopa use, n (%) 42 (79) 32 (80) — ns  Dopamine agonist use, n (%) 25 (46) 28 (78) — .004 Antidepressant use, n (%) 10 (19) 7 (19) — ns Antianxiolitic use, n (%) 19 (35) 7 (19) — ns ISI scores 11.6 ± 6.2 10.8 ± 7.8 6.6 ± 4.4 .001; A > C***, B > C** ESS scores 9.3 ± 4.9 9.5 ± 4.9 6.9 ± 4.0 .008; A > C*, B > C* BDI-II scores 10.9 ± 5.8 10.7 ± 7.5 5.9 ± 5.7 .000; A > C***, B > C*** BAI scores 11.4 ± 8.7 9.6 ± 6.3 5.4 ± 5.4 .001; A > C***, B > C* MMSE scores 28.0 ± 1.9 29.2 ± 0.9 29.3 ± 1.1 .000; A < B***, A < C*** Results are expressed as mean ± standard deviation. BAI = Beck Anxiety Inventory; BDI-II = Beck Depression Inventory second edition; EMG = electromyography; ESS = Epworth Sleepiness Scale; ISI = Insomnia Severity Index; MMSE = Mini-Mental State Examination; ns = not significant; PD = Parkinson’s disease; PD-RBD = PD with RBD; PD-nRBD = PD without RBD; RBD = REM sleep behavior disorder; UPDRS = Unified Parkinson’s Disease Rating Scale. ***p < .001; ** = p < .01; * = p < .05. Open in new tab Table 1 Demographic, Clinical, and Mood Characteristics: Combined PD Cohorts and Controls. Variables Combined PD-RBD (A) Combined PD-nRBD (B) Controls (C) p value; post hoc (n = 53) (n = 40) (n = 69) Sex men, n (%) 40 (75) 21 (53) 38 (55) .03; A > B*, A > C* Age, years 68.0 ± 8.4 63.2 ± 8.5 63.3 ± 10.3 .01; A > B*, A > C** Education, years 14.5 ± 3.9 15.1 ± 3.0 14.5 ± 2.8 ns RBD duration, years 4.1 ± 3.5 — — — PD duration since diagnosis, years 6.1 ± 4.5 6.1 ± 4.3 — ns Hoehn and Yahr stage 2.5 ± 0.8 2.2 ± 0.9 — ns UPDRS part III “on” 23.1 ± 9.5 20.3 ± 9.8 — ns REM tonic EMG, % 61.3 ± 34.3 17.1 ± 23.9 — .000 REM phasic EMG, % 31.6 ± 20.9 14.3 ± 12.1 — .000 Levodopa equivalent dosage, mg 492.2 ± 402.3 383.7 ± 283.3 — ns  Levodopa use, n (%) 42 (79) 32 (80) — ns  Dopamine agonist use, n (%) 25 (46) 28 (78) — .004 Antidepressant use, n (%) 10 (19) 7 (19) — ns Antianxiolitic use, n (%) 19 (35) 7 (19) — ns ISI scores 11.6 ± 6.2 10.8 ± 7.8 6.6 ± 4.4 .001; A > C***, B > C** ESS scores 9.3 ± 4.9 9.5 ± 4.9 6.9 ± 4.0 .008; A > C*, B > C* BDI-II scores 10.9 ± 5.8 10.7 ± 7.5 5.9 ± 5.7 .000; A > C***, B > C*** BAI scores 11.4 ± 8.7 9.6 ± 6.3 5.4 ± 5.4 .001; A > C***, B > C* MMSE scores 28.0 ± 1.9 29.2 ± 0.9 29.3 ± 1.1 .000; A < B***, A < C*** Variables Combined PD-RBD (A) Combined PD-nRBD (B) Controls (C) p value; post hoc (n = 53) (n = 40) (n = 69) Sex men, n (%) 40 (75) 21 (53) 38 (55) .03; A > B*, A > C* Age, years 68.0 ± 8.4 63.2 ± 8.5 63.3 ± 10.3 .01; A > B*, A > C** Education, years 14.5 ± 3.9 15.1 ± 3.0 14.5 ± 2.8 ns RBD duration, years 4.1 ± 3.5 — — — PD duration since diagnosis, years 6.1 ± 4.5 6.1 ± 4.3 — ns Hoehn and Yahr stage 2.5 ± 0.8 2.2 ± 0.9 — ns UPDRS part III “on” 23.1 ± 9.5 20.3 ± 9.8 — ns REM tonic EMG, % 61.3 ± 34.3 17.1 ± 23.9 — .000 REM phasic EMG, % 31.6 ± 20.9 14.3 ± 12.1 — .000 Levodopa equivalent dosage, mg 492.2 ± 402.3 383.7 ± 283.3 — ns  Levodopa use, n (%) 42 (79) 32 (80) — ns  Dopamine agonist use, n (%) 25 (46) 28 (78) — .004 Antidepressant use, n (%) 10 (19) 7 (19) — ns Antianxiolitic use, n (%) 19 (35) 7 (19) — ns ISI scores 11.6 ± 6.2 10.8 ± 7.8 6.6 ± 4.4 .001; A > C***, B > C** ESS scores 9.3 ± 4.9 9.5 ± 4.9 6.9 ± 4.0 .008; A > C*, B > C* BDI-II scores 10.9 ± 5.8 10.7 ± 7.5 5.9 ± 5.7 .000; A > C***, B > C*** BAI scores 11.4 ± 8.7 9.6 ± 6.3 5.4 ± 5.4 .001; A > C***, B > C* MMSE scores 28.0 ± 1.9 29.2 ± 0.9 29.3 ± 1.1 .000; A < B***, A < C*** Results are expressed as mean ± standard deviation. BAI = Beck Anxiety Inventory; BDI-II = Beck Depression Inventory second edition; EMG = electromyography; ESS = Epworth Sleepiness Scale; ISI = Insomnia Severity Index; MMSE = Mini-Mental State Examination; ns = not significant; PD = Parkinson’s disease; PD-RBD = PD with RBD; PD-nRBD = PD without RBD; RBD = REM sleep behavior disorder; UPDRS = Unified Parkinson’s Disease Rating Scale. ***p < .001; ** = p < .01; * = p < .05. Open in new tab Procedure All participants underwent one-night PSG recordings in the sleep laboratory. Sleep was recorded using a polygraph composed of two EEG electrodes: a central (C3/A2) and an occipital (O2/A1). A left and right electro-oculogram were used to measure eye movements and a submental electromyography (EMG) to measure muscle activity. Oral and nasal airflow and thoracic and abdominal movements were recorded, and oximetry was performed to exclude sleep apnea and hypopnea syndrome. Sleep stages were recorded according to a method developed for RBD patients and described in detail elsewhere.30 RBD was diagnosed by a sleep specialist (JM) according to the criteria of the International Classification of Sleep Disorders, Second Edition and PSG criteria.30,31 Percentages of REM tonic and phasic EMG activity were calculated using a previously published method.30 All patients underwent a detailed neurological examination (RBP), including administration of the Unified Parkinson’s Disease Rating Scale.32 The Beck Depressive Inventory, Second Edition,33 and Beck Anxiety Inventory34 were administrated to quantify depressive and anxiety symptom severity. The Epworth Sleepiness Scale (ESS)35 and Insomnia Severity Index36 were used to assess daytime sleepiness and insomnia symptom severity. The neuropsychological assessment was divided into two 90-minute sessions and included measures of five cognitive domains: attention, executive functions, episodic verbal and nonverbal learning and memory, visuospatial abilities, and language (neuropsychological tests, variables, and normative data are presented in Supplementary Table 1). Test administration and scoring followed standard procedures.37 All patients took their usual medications prior to neuropsychological assessment. MCI diagnosis was established by a consensus between the neurologist and neuropsychologist according to the Movement Disorder Society Task Force criteria27 as: (1) a subjective cognitive complaint during the interview with the participant or spouse/caregiver, or a score > 25, or the response 3 (quite often) or 4 (very often) on at least one item on the Cognitive Failure Questionnaire (CFQ);38 (2) objective evidence of cognitive decline defined by performance at 1.5 standard deviations below the standardized mean on at least two variables in the same cognitive domain (Supplementary Table 1); and (3) the cognitive impairment does not significantly alter daily living activities and functioning. Impaired daily functioning was determined during the interview with patients and their relatives. Impairment was assessed in terms of decline in several activities, including managing finances, performing chores, preparing meals, shopping, driving, and using public transportation. MCI subtypes were defined as: amnestic MCI–single domain, nonamnestic MCI–single domain, amnestic MCI–multiple domain, or nonamnestic MCI–multiple domain. SCD was identified in PD patients without MCI by a subjective cognitive complaint during the interview with the participant or spouse/caregiver, or a score > 25, or the response 3 (quite often) or 4 (very often) on at least one item on the CFQ.38 Statistical Analysis We first divided PD patients with and without RBD (PD-RBD and PD-nRBD) into the original cohort (previously published study),13 a replication cohort (new patients), and a combined cohort. Independent sample t-tests or nonparametric U Mann-Whitney tests (for not normally distributed variables) were applied to compare demographic, clinical, and mood variables between PD-RBD and PD-nRBD participants in the original, replication, and combined cohort, and univariate analyses of variance were applied to compare demographic, clinical, and mood variables between PD-RBD, PD-nRBD, and healthy controls. Subsequent univariate analyses of covariance with Bonferroni post hoc set at p < .05 were performed to assess cognitive performance between PD-RBD and PD-nRBD patients (original and replication cohorts) and between patients and controls in the combined cohort. Univariate analyses of covariance were also performed for additional analyses in the combined PD-RBD group to compare cognitive performance between gender (men vs. women), and RBD onset (prior vs. same time/after PD diagnosis), and in the combined cohorts to compare cognition between men only (PD-RBD vs. PD-nRBD vs. controls) and between women only (PD-RBD vs. PD-nRBD vs. controls). Certain cognitive variables were log transformed to ensure normality of distribution. Age and sex were included as covariates in the analysis of cognitive performance, for all three-cohort comparisons, given that they differed significantly between PD groups and were related to cognitive performance. Pearson’s or the nonparametric Spearman’s correlations were conducted between the percentages of REM sleep EMG activity (tonic or phasic) and cognitive variables in all PD patients. Pearson χ2 tests were used to compare proportions of men, medication users (levodopa, dopamine agonist, anxiolytic, and antidepressant), patients with clinically impaired cognitive performance (≥ 1.5 standard deviations below the standardized mean in PD-RBD vs. PD-nRBD), MCI diagnosis (PD-RBD vs. PD-nRBD), and SCD (PD-RBD without MCI vs. PD-nRBD without MCI). The association between RBD and each cognitive variable was defined as highly probable if all three cohorts (original, replication, and combined) showed a significant association with p < .05 for each cohort, probable if in two of three cohorts p < .05, possible if one cohort showed a significant association with p < .05, and probably not associated if all p values were p > .05. All analyses were computed using Statistical Package for the Social Sciences version 21 (Chicago, Illinois). RESULTS From the initial sample, 24 participants were excluded: 4 PD-RBD patients, 8 PD-nRBD patients, and 12 controls. Reasons for exclusion included sleep apnea (4 PD-RBD, 4 PD-nRBD, 3 controls), lower education (4 controls), atypical PD (4 PD-nRBD), and presence of MCI or dementia (5 controls). Of the remaining 162 participants, 53 were classified as PD-RBD, 40 were PD-nRBD, and 69 were healthy controls. Demographic, Clinical, and Mood Characteristics Results for the combined cohort are presented in Table 1, and results for the original and the replication cohorts are presented in the Supplementary Table 2. In both the original and combined cohort, the proportion of men was significantly higher in PD-RBD patients compared to PD-nRBD patients and controls. PD-RBD patients were also significantly older and had lower MMSE scores than PD-nRBD patients in the replication cohort, the combined cohort, and versus controls. In both the replication and combined cohorts, the proportion of PD-nRBD patients using a dopaminergic agonist was higher compared to PD-RBD patients. Of the three PD cohorts, PD-RBD patients had the highest REM tonic and phasic EMG activity. In the replication cohort, PD-RBD patients had a more advanced stage of PD. No significant between-group differences were found for education, PD duration (diagnosis), motor symptom severity, levodopa use and dosage, or antidepressant and anxiolytic use. No significant differences were found in the sleep and mood questionnaires between the two PD groups. However, both PD groups scored higher than controls on all questionnaires. No correlation was observed between questionnaire scores and cognitive test performance. The results were similar between the original and replication PD cohorts except for a younger age for the replication PD-nRBD cohort. Neuropsychological Assessment Cognitive Performance Results for the three PD cohorts are presented in Table 2. PD-RBD patients performed worse than PD-nRBD in at least three cohorts (highly probable impaired) on tests assessing attention (Stroop Color-Word Test [III-II, time]), executive functions (Trail Making Test [part B, time and B–A, time]), and episodic verbal learning and memory (Rey Auditory-Verbal Learning Test [immediate recall]). PD-RBD patients performed worse than PD-nRBD in at least two cohorts (probable impaired) on tests assessing attention (Digit Span subtest [scaled score] and Stroop Color-Word Test [III-II, errors]), executive functions (Stroop Color-Word Test [IV-III, errors] and Verbal Fluency [semantic]), episodic verbal learning and memory (Rey Auditory-Verbal Learning Test [sum of trials 1 to 5]), and visuospatial abilities (Rey-O Complex Figure Test [copy] and Block Design subtest [scaled score]). PD-RBD patients also performed worse than PD-nRBD patients in only one cohort (possible impaired) on tests assessing executive functions (Verbal Fluency [letter]) and episodic verbal learning and memory (Rey Auditory-Verbal Learning Test [delayed recall]). No significant association was found between RBD and impaired cognitive performance for episodic nonverbal learning and memory and language. The results were similar between the original and replication PD cohorts except for poorer performance on visuospatial abilities for the original PD-RBD cohort. Table 2 Cognitive Performance: All PD Cohorts and Controls. Cognitive domains and tests Original PD-RBD Original PD-nRBD p value Replication PD-RBD Replication PD-nRBD p value Combined PD-RBD (A) Combined PD-nRBD (B) Controls (C) p value; post hoc (n = 21) (n = 16) (n = 32) (n = 24) (n = 53) (n = 40) (n = 69) Attention  Digit Span, scaled score 9.7 ± 2.3 11.8 ± 2.4 .01 9.6 ± 2.3 11.0 ± 3.0 ns 9.6 ± 2.3 11.3 ± 2.7 11.4 ± 3.2 .003; A < B**, A < Cw**  Trail Making Test part A, time, sec 57.3 ± 25.2 42.8 ± 16.8 nsª 54.8 ± 24.0 46.1 ± 20.0 nsª 55.8 ± 24.3 44.8 ± 18.6 37.4 ± 14.6 .001ª; A > Cm*** Stroop Color-Word Test  III-II time, sec 74.8 ± 38.2 45.8 ± 14.2 .03ª 82.8 ± 38.7 49.5 ± 19.4 .003ª 79.5 ± 38.2 48.0 ± 17.4 47.4 ± 38.7 .000ª; A > Bm,w***, A > Cm***  III-II errors 3.4 ± 4.2 0.8 ± 2.3 .04ª 3.0 ± 5.1 1.1 ± 2.9 nsª 3.2 ± 4.7 1.0 ± 2.7 0.6 ± 2.1 .007ª; A > B*, A > Cm** Executive functions  Trail Making Test part B, time, sec 157.6 ± 92.0 97.0 ± 59.3 .01ª 170.4 ± 89.2 99.3 ± 87.9 .04ª 165.3 ± 89.6 98.4 ± 76.8 95.9 ± 43.8 .000ª; A > Bm,w***, A > Cm***   Part B–part A, time, sec 100.3 ± 79.0 54.2 ± 49.0 .02ª 115.4 ± 72.4 53.2 ± 77.9 .001ª 109.4 ± 74.7 53.6 ± 67.1 58.5 ± 37.8 .000ª; A > Bm,w***, A > Cm** Stroop Color-Word Test  IV-III time, sec 31.2 ± 10.9 22.3 ± 16.7 ns 31.9 ± 33.7 22.1 ± 23.3 ns 31.8 ± 29.3 22.2 ± 21.7 14.2 ± 36.8 nsª  IV-III, errors 4.0 ± 4.3 2.9 ± 4.8 nsª 6.2 ± 5.9 2.1 ± 4.0 .03ª 5.7 ± 5.6 2.3 ± 4.2 1.9 ± 3.7 .003ª; A > Bw*, A > Cw** Verbal fluency  Semantic 26.7 ± 6.8 35.3 ± 6.9 .000 28.8 ± 7.8 36.9 ± 10.1 ns 27.9 ± 7.4 36.3 ± 8.9 36.5 ± 6.8 .000; A < Bm***, A < Cm,w***  Letter 32.0 ± 12.4 38.3 ± 9.4 ns 30.9 ± 11.0 39.4 ± 10.7 ns 31.3 ± 11.5 38.9 ± 10.0 37.2 ± 11.1 .008; A < B*, A < C* Episodic learning and memory Verbal RAVLT  Sum of trials 1 to 5 38.0 ± 8.6 45.5 ± 5.9 .04 37.3 ± 12.5 47.3 ± 12.8 ns 37.5 ± 11.1 46.6 ± 10.5 49.3 ± 8.6 .000; A < Bm**, A < Cm***  List B 4.8 ± 1.5 5.1 ± 1.4 ns 3.8 ± 1.9 5.1 ± 2.0 ns 4.2 ± 1.8 5.1 ± 1.8 5.6 ± 1.9 .002; A < Cw**  Immediate recall 6.4 ± 3.0 9.1 ± 2.6 .03 7.0 ± 3.5 10.3 ± 4.1 .04 6.8 ± 3.3 9.8 ± 3.6 10.4 ± 2.7 .000; A < Bm**, A < Cm,w***  Delayed recall 7.1 ± 3.1 9.8 ± 2.1 ns 7.4 ± 3.5 10.1 ± 4.0 ns 7.3 ± 3.3 10.0 ± 3.4 10.0 ± 3.1 .008; A < B*, A < C**  Recognition 12.8 ± 2.6 13.9 ± 1.2 ns 12.7 ± 2.3 14.0 ± 1.3 ns 12.7 ± 2.4 13.9 ± 1.3 14.1 ± 1.2 .009; A < Cm** Nonverbal Rey-O Complex Figure Test  Immediate recall 11.6 ± 6.0 14.3 ± 6.1 nsª 14.2 ± 6.9 17.4 ± 7.1 nsª 13.2 ± 6.6 16.1 ± 6.8 15.3 ± 6.1 ns  Delayed recall 11.4 ± 5.2 12.9 ± 6.4 nsª 12.8 ± 7.0 16.4 ± 7.9 nsª 12.2 ± 6.3 15.0 ± 7.4 15.5 ± 6.0 ns Visuospatial  Rey-O Complex Figure Test, copy 26.5 ± 7.1 30.1 ± 3.4 .04 28.2 ± 5.0 31.0 ± 4.0 ns 27.5 ± 5.9 30.6 ± 3.7 30.3 ± 4.0 .02; A < B*  Block Design, scaled score 8.1 ± 1.9 12.0 ± 2.4 .000 9.5 ± 3.2 10.2 ± 3.8 ns 8.9 ± 2.8 10.9 ± 3.4 11.5 ± 3.2 .000; A < B*, A < Cm***  Bells test, omissions 4.1 ± 2.8 3.2 ± 4.7 nsª 2.8 ± 2.9 2.7 ± 3.7 nsª 3.4 ± 2.9 2.9 ± 4.1 2.2 ± 2.8 nsª Language  Boston Naming Test — — — 28.7 ± 1.8 29.4 ± 0.8 ns 28.7 ± 1.8 29.4 ± 0.8 — ns  Vocabulary, scaled score — — — 10.2 ± 2.0 11.6 ± 2.4 ns 10.2 ± 2.0 11.6 ± 2.4 — ns  MMSE language score 7.8 ± 0.5 7.9 ± 0.5 ns 7.7 ± 0.7 8.0 ± 0.2 ns 7.7 ± 0.6 7.9 ± 0.4 8.0 ± 0.1 .01; A < C* Cognitive domains and tests Original PD-RBD Original PD-nRBD p value Replication PD-RBD Replication PD-nRBD p value Combined PD-RBD (A) Combined PD-nRBD (B) Controls (C) p value; post hoc (n = 21) (n = 16) (n = 32) (n = 24) (n = 53) (n = 40) (n = 69) Attention  Digit Span, scaled score 9.7 ± 2.3 11.8 ± 2.4 .01 9.6 ± 2.3 11.0 ± 3.0 ns 9.6 ± 2.3 11.3 ± 2.7 11.4 ± 3.2 .003; A < B**, A < Cw**  Trail Making Test part A, time, sec 57.3 ± 25.2 42.8 ± 16.8 nsª 54.8 ± 24.0 46.1 ± 20.0 nsª 55.8 ± 24.3 44.8 ± 18.6 37.4 ± 14.6 .001ª; A > Cm*** Stroop Color-Word Test  III-II time, sec 74.8 ± 38.2 45.8 ± 14.2 .03ª 82.8 ± 38.7 49.5 ± 19.4 .003ª 79.5 ± 38.2 48.0 ± 17.4 47.4 ± 38.7 .000ª; A > Bm,w***, A > Cm***  III-II errors 3.4 ± 4.2 0.8 ± 2.3 .04ª 3.0 ± 5.1 1.1 ± 2.9 nsª 3.2 ± 4.7 1.0 ± 2.7 0.6 ± 2.1 .007ª; A > B*, A > Cm** Executive functions  Trail Making Test part B, time, sec 157.6 ± 92.0 97.0 ± 59.3 .01ª 170.4 ± 89.2 99.3 ± 87.9 .04ª 165.3 ± 89.6 98.4 ± 76.8 95.9 ± 43.8 .000ª; A > Bm,w***, A > Cm***   Part B–part A, time, sec 100.3 ± 79.0 54.2 ± 49.0 .02ª 115.4 ± 72.4 53.2 ± 77.9 .001ª 109.4 ± 74.7 53.6 ± 67.1 58.5 ± 37.8 .000ª; A > Bm,w***, A > Cm** Stroop Color-Word Test  IV-III time, sec 31.2 ± 10.9 22.3 ± 16.7 ns 31.9 ± 33.7 22.1 ± 23.3 ns 31.8 ± 29.3 22.2 ± 21.7 14.2 ± 36.8 nsª  IV-III, errors 4.0 ± 4.3 2.9 ± 4.8 nsª 6.2 ± 5.9 2.1 ± 4.0 .03ª 5.7 ± 5.6 2.3 ± 4.2 1.9 ± 3.7 .003ª; A > Bw*, A > Cw** Verbal fluency  Semantic 26.7 ± 6.8 35.3 ± 6.9 .000 28.8 ± 7.8 36.9 ± 10.1 ns 27.9 ± 7.4 36.3 ± 8.9 36.5 ± 6.8 .000; A < Bm***, A < Cm,w***  Letter 32.0 ± 12.4 38.3 ± 9.4 ns 30.9 ± 11.0 39.4 ± 10.7 ns 31.3 ± 11.5 38.9 ± 10.0 37.2 ± 11.1 .008; A < B*, A < C* Episodic learning and memory Verbal RAVLT  Sum of trials 1 to 5 38.0 ± 8.6 45.5 ± 5.9 .04 37.3 ± 12.5 47.3 ± 12.8 ns 37.5 ± 11.1 46.6 ± 10.5 49.3 ± 8.6 .000; A < Bm**, A < Cm***  List B 4.8 ± 1.5 5.1 ± 1.4 ns 3.8 ± 1.9 5.1 ± 2.0 ns 4.2 ± 1.8 5.1 ± 1.8 5.6 ± 1.9 .002; A < Cw**  Immediate recall 6.4 ± 3.0 9.1 ± 2.6 .03 7.0 ± 3.5 10.3 ± 4.1 .04 6.8 ± 3.3 9.8 ± 3.6 10.4 ± 2.7 .000; A < Bm**, A < Cm,w***  Delayed recall 7.1 ± 3.1 9.8 ± 2.1 ns 7.4 ± 3.5 10.1 ± 4.0 ns 7.3 ± 3.3 10.0 ± 3.4 10.0 ± 3.1 .008; A < B*, A < C**  Recognition 12.8 ± 2.6 13.9 ± 1.2 ns 12.7 ± 2.3 14.0 ± 1.3 ns 12.7 ± 2.4 13.9 ± 1.3 14.1 ± 1.2 .009; A < Cm** Nonverbal Rey-O Complex Figure Test  Immediate recall 11.6 ± 6.0 14.3 ± 6.1 nsª 14.2 ± 6.9 17.4 ± 7.1 nsª 13.2 ± 6.6 16.1 ± 6.8 15.3 ± 6.1 ns  Delayed recall 11.4 ± 5.2 12.9 ± 6.4 nsª 12.8 ± 7.0 16.4 ± 7.9 nsª 12.2 ± 6.3 15.0 ± 7.4 15.5 ± 6.0 ns Visuospatial  Rey-O Complex Figure Test, copy 26.5 ± 7.1 30.1 ± 3.4 .04 28.2 ± 5.0 31.0 ± 4.0 ns 27.5 ± 5.9 30.6 ± 3.7 30.3 ± 4.0 .02; A < B*  Block Design, scaled score 8.1 ± 1.9 12.0 ± 2.4 .000 9.5 ± 3.2 10.2 ± 3.8 ns 8.9 ± 2.8 10.9 ± 3.4 11.5 ± 3.2 .000; A < B*, A < Cm***  Bells test, omissions 4.1 ± 2.8 3.2 ± 4.7 nsª 2.8 ± 2.9 2.7 ± 3.7 nsª 3.4 ± 2.9 2.9 ± 4.1 2.2 ± 2.8 nsª Language  Boston Naming Test — — — 28.7 ± 1.8 29.4 ± 0.8 ns 28.7 ± 1.8 29.4 ± 0.8 — ns  Vocabulary, scaled score — — — 10.2 ± 2.0 11.6 ± 2.4 ns 10.2 ± 2.0 11.6 ± 2.4 — ns  MMSE language score 7.8 ± 0.5 7.9 ± 0.5 ns 7.7 ± 0.7 8.0 ± 0.2 ns 7.7 ± 0.6 7.9 ± 0.4 8.0 ± 0.1 .01; A < C* Results are expressed as mean ± standard deviation. ªlog transformations. Also statistically significant when comparing Menm and Womenw only in the combined cohorts. MMSE = Mini-Mental State Examination; ns = not significant; PD = Parkinson’s disease; PD-RBD = PD with REM sleep behavior disorder; PD-nRBD = PD without RBD; RAVLT = Rey Auditory-Verbal Learning Test. ***p < .001; ** = p < .01; * = p < .05. Open in new tab Table 2 Cognitive Performance: All PD Cohorts and Controls. Cognitive domains and tests Original PD-RBD Original PD-nRBD p value Replication PD-RBD Replication PD-nRBD p value Combined PD-RBD (A) Combined PD-nRBD (B) Controls (C) p value; post hoc (n = 21) (n = 16) (n = 32) (n = 24) (n = 53) (n = 40) (n = 69) Attention  Digit Span, scaled score 9.7 ± 2.3 11.8 ± 2.4 .01 9.6 ± 2.3 11.0 ± 3.0 ns 9.6 ± 2.3 11.3 ± 2.7 11.4 ± 3.2 .003; A < B**, A < Cw**  Trail Making Test part A, time, sec 57.3 ± 25.2 42.8 ± 16.8 nsª 54.8 ± 24.0 46.1 ± 20.0 nsª 55.8 ± 24.3 44.8 ± 18.6 37.4 ± 14.6 .001ª; A > Cm*** Stroop Color-Word Test  III-II time, sec 74.8 ± 38.2 45.8 ± 14.2 .03ª 82.8 ± 38.7 49.5 ± 19.4 .003ª 79.5 ± 38.2 48.0 ± 17.4 47.4 ± 38.7 .000ª; A > Bm,w***, A > Cm***  III-II errors 3.4 ± 4.2 0.8 ± 2.3 .04ª 3.0 ± 5.1 1.1 ± 2.9 nsª 3.2 ± 4.7 1.0 ± 2.7 0.6 ± 2.1 .007ª; A > B*, A > Cm** Executive functions  Trail Making Test part B, time, sec 157.6 ± 92.0 97.0 ± 59.3 .01ª 170.4 ± 89.2 99.3 ± 87.9 .04ª 165.3 ± 89.6 98.4 ± 76.8 95.9 ± 43.8 .000ª; A > Bm,w***, A > Cm***   Part B–part A, time, sec 100.3 ± 79.0 54.2 ± 49.0 .02ª 115.4 ± 72.4 53.2 ± 77.9 .001ª 109.4 ± 74.7 53.6 ± 67.1 58.5 ± 37.8 .000ª; A > Bm,w***, A > Cm** Stroop Color-Word Test  IV-III time, sec 31.2 ± 10.9 22.3 ± 16.7 ns 31.9 ± 33.7 22.1 ± 23.3 ns 31.8 ± 29.3 22.2 ± 21.7 14.2 ± 36.8 nsª  IV-III, errors 4.0 ± 4.3 2.9 ± 4.8 nsª 6.2 ± 5.9 2.1 ± 4.0 .03ª 5.7 ± 5.6 2.3 ± 4.2 1.9 ± 3.7 .003ª; A > Bw*, A > Cw** Verbal fluency  Semantic 26.7 ± 6.8 35.3 ± 6.9 .000 28.8 ± 7.8 36.9 ± 10.1 ns 27.9 ± 7.4 36.3 ± 8.9 36.5 ± 6.8 .000; A < Bm***, A < Cm,w***  Letter 32.0 ± 12.4 38.3 ± 9.4 ns 30.9 ± 11.0 39.4 ± 10.7 ns 31.3 ± 11.5 38.9 ± 10.0 37.2 ± 11.1 .008; A < B*, A < C* Episodic learning and memory Verbal RAVLT  Sum of trials 1 to 5 38.0 ± 8.6 45.5 ± 5.9 .04 37.3 ± 12.5 47.3 ± 12.8 ns 37.5 ± 11.1 46.6 ± 10.5 49.3 ± 8.6 .000; A < Bm**, A < Cm***  List B 4.8 ± 1.5 5.1 ± 1.4 ns 3.8 ± 1.9 5.1 ± 2.0 ns 4.2 ± 1.8 5.1 ± 1.8 5.6 ± 1.9 .002; A < Cw**  Immediate recall 6.4 ± 3.0 9.1 ± 2.6 .03 7.0 ± 3.5 10.3 ± 4.1 .04 6.8 ± 3.3 9.8 ± 3.6 10.4 ± 2.7 .000; A < Bm**, A < Cm,w***  Delayed recall 7.1 ± 3.1 9.8 ± 2.1 ns 7.4 ± 3.5 10.1 ± 4.0 ns 7.3 ± 3.3 10.0 ± 3.4 10.0 ± 3.1 .008; A < B*, A < C**  Recognition 12.8 ± 2.6 13.9 ± 1.2 ns 12.7 ± 2.3 14.0 ± 1.3 ns 12.7 ± 2.4 13.9 ± 1.3 14.1 ± 1.2 .009; A < Cm** Nonverbal Rey-O Complex Figure Test  Immediate recall 11.6 ± 6.0 14.3 ± 6.1 nsª 14.2 ± 6.9 17.4 ± 7.1 nsª 13.2 ± 6.6 16.1 ± 6.8 15.3 ± 6.1 ns  Delayed recall 11.4 ± 5.2 12.9 ± 6.4 nsª 12.8 ± 7.0 16.4 ± 7.9 nsª 12.2 ± 6.3 15.0 ± 7.4 15.5 ± 6.0 ns Visuospatial  Rey-O Complex Figure Test, copy 26.5 ± 7.1 30.1 ± 3.4 .04 28.2 ± 5.0 31.0 ± 4.0 ns 27.5 ± 5.9 30.6 ± 3.7 30.3 ± 4.0 .02; A < B*  Block Design, scaled score 8.1 ± 1.9 12.0 ± 2.4 .000 9.5 ± 3.2 10.2 ± 3.8 ns 8.9 ± 2.8 10.9 ± 3.4 11.5 ± 3.2 .000; A < B*, A < Cm***  Bells test, omissions 4.1 ± 2.8 3.2 ± 4.7 nsª 2.8 ± 2.9 2.7 ± 3.7 nsª 3.4 ± 2.9 2.9 ± 4.1 2.2 ± 2.8 nsª Language  Boston Naming Test — — — 28.7 ± 1.8 29.4 ± 0.8 ns 28.7 ± 1.8 29.4 ± 0.8 — ns  Vocabulary, scaled score — — — 10.2 ± 2.0 11.6 ± 2.4 ns 10.2 ± 2.0 11.6 ± 2.4 — ns  MMSE language score 7.8 ± 0.5 7.9 ± 0.5 ns 7.7 ± 0.7 8.0 ± 0.2 ns 7.7 ± 0.6 7.9 ± 0.4 8.0 ± 0.1 .01; A < C* Cognitive domains and tests Original PD-RBD Original PD-nRBD p value Replication PD-RBD Replication PD-nRBD p value Combined PD-RBD (A) Combined PD-nRBD (B) Controls (C) p value; post hoc (n = 21) (n = 16) (n = 32) (n = 24) (n = 53) (n = 40) (n = 69) Attention  Digit Span, scaled score 9.7 ± 2.3 11.8 ± 2.4 .01 9.6 ± 2.3 11.0 ± 3.0 ns 9.6 ± 2.3 11.3 ± 2.7 11.4 ± 3.2 .003; A < B**, A < Cw**  Trail Making Test part A, time, sec 57.3 ± 25.2 42.8 ± 16.8 nsª 54.8 ± 24.0 46.1 ± 20.0 nsª 55.8 ± 24.3 44.8 ± 18.6 37.4 ± 14.6 .001ª; A > Cm*** Stroop Color-Word Test  III-II time, sec 74.8 ± 38.2 45.8 ± 14.2 .03ª 82.8 ± 38.7 49.5 ± 19.4 .003ª 79.5 ± 38.2 48.0 ± 17.4 47.4 ± 38.7 .000ª; A > Bm,w***, A > Cm***  III-II errors 3.4 ± 4.2 0.8 ± 2.3 .04ª 3.0 ± 5.1 1.1 ± 2.9 nsª 3.2 ± 4.7 1.0 ± 2.7 0.6 ± 2.1 .007ª; A > B*, A > Cm** Executive functions  Trail Making Test part B, time, sec 157.6 ± 92.0 97.0 ± 59.3 .01ª 170.4 ± 89.2 99.3 ± 87.9 .04ª 165.3 ± 89.6 98.4 ± 76.8 95.9 ± 43.8 .000ª; A > Bm,w***, A > Cm***   Part B–part A, time, sec 100.3 ± 79.0 54.2 ± 49.0 .02ª 115.4 ± 72.4 53.2 ± 77.9 .001ª 109.4 ± 74.7 53.6 ± 67.1 58.5 ± 37.8 .000ª; A > Bm,w***, A > Cm** Stroop Color-Word Test  IV-III time, sec 31.2 ± 10.9 22.3 ± 16.7 ns 31.9 ± 33.7 22.1 ± 23.3 ns 31.8 ± 29.3 22.2 ± 21.7 14.2 ± 36.8 nsª  IV-III, errors 4.0 ± 4.3 2.9 ± 4.8 nsª 6.2 ± 5.9 2.1 ± 4.0 .03ª 5.7 ± 5.6 2.3 ± 4.2 1.9 ± 3.7 .003ª; A > Bw*, A > Cw** Verbal fluency  Semantic 26.7 ± 6.8 35.3 ± 6.9 .000 28.8 ± 7.8 36.9 ± 10.1 ns 27.9 ± 7.4 36.3 ± 8.9 36.5 ± 6.8 .000; A < Bm***, A < Cm,w***  Letter 32.0 ± 12.4 38.3 ± 9.4 ns 30.9 ± 11.0 39.4 ± 10.7 ns 31.3 ± 11.5 38.9 ± 10.0 37.2 ± 11.1 .008; A < B*, A < C* Episodic learning and memory Verbal RAVLT  Sum of trials 1 to 5 38.0 ± 8.6 45.5 ± 5.9 .04 37.3 ± 12.5 47.3 ± 12.8 ns 37.5 ± 11.1 46.6 ± 10.5 49.3 ± 8.6 .000; A < Bm**, A < Cm***  List B 4.8 ± 1.5 5.1 ± 1.4 ns 3.8 ± 1.9 5.1 ± 2.0 ns 4.2 ± 1.8 5.1 ± 1.8 5.6 ± 1.9 .002; A < Cw**  Immediate recall 6.4 ± 3.0 9.1 ± 2.6 .03 7.0 ± 3.5 10.3 ± 4.1 .04 6.8 ± 3.3 9.8 ± 3.6 10.4 ± 2.7 .000; A < Bm**, A < Cm,w***  Delayed recall 7.1 ± 3.1 9.8 ± 2.1 ns 7.4 ± 3.5 10.1 ± 4.0 ns 7.3 ± 3.3 10.0 ± 3.4 10.0 ± 3.1 .008; A < B*, A < C**  Recognition 12.8 ± 2.6 13.9 ± 1.2 ns 12.7 ± 2.3 14.0 ± 1.3 ns 12.7 ± 2.4 13.9 ± 1.3 14.1 ± 1.2 .009; A < Cm** Nonverbal Rey-O Complex Figure Test  Immediate recall 11.6 ± 6.0 14.3 ± 6.1 nsª 14.2 ± 6.9 17.4 ± 7.1 nsª 13.2 ± 6.6 16.1 ± 6.8 15.3 ± 6.1 ns  Delayed recall 11.4 ± 5.2 12.9 ± 6.4 nsª 12.8 ± 7.0 16.4 ± 7.9 nsª 12.2 ± 6.3 15.0 ± 7.4 15.5 ± 6.0 ns Visuospatial  Rey-O Complex Figure Test, copy 26.5 ± 7.1 30.1 ± 3.4 .04 28.2 ± 5.0 31.0 ± 4.0 ns 27.5 ± 5.9 30.6 ± 3.7 30.3 ± 4.0 .02; A < B*  Block Design, scaled score 8.1 ± 1.9 12.0 ± 2.4 .000 9.5 ± 3.2 10.2 ± 3.8 ns 8.9 ± 2.8 10.9 ± 3.4 11.5 ± 3.2 .000; A < B*, A < Cm***  Bells test, omissions 4.1 ± 2.8 3.2 ± 4.7 nsª 2.8 ± 2.9 2.7 ± 3.7 nsª 3.4 ± 2.9 2.9 ± 4.1 2.2 ± 2.8 nsª Language  Boston Naming Test — — — 28.7 ± 1.8 29.4 ± 0.8 ns 28.7 ± 1.8 29.4 ± 0.8 — ns  Vocabulary, scaled score — — — 10.2 ± 2.0 11.6 ± 2.4 ns 10.2 ± 2.0 11.6 ± 2.4 — ns  MMSE language score 7.8 ± 0.5 7.9 ± 0.5 ns 7.7 ± 0.7 8.0 ± 0.2 ns 7.7 ± 0.6 7.9 ± 0.4 8.0 ± 0.1 .01; A < C* Results are expressed as mean ± standard deviation. ªlog transformations. Also statistically significant when comparing Menm and Womenw only in the combined cohorts. MMSE = Mini-Mental State Examination; ns = not significant; PD = Parkinson’s disease; PD-RBD = PD with REM sleep behavior disorder; PD-nRBD = PD without RBD; RAVLT = Rey Auditory-Verbal Learning Test. ***p < .001; ** = p < .01; * = p < .05. Open in new tab Additional analyses on the PD-RBD group revealed that men performed worse than women on the Rey Auditory-Verbal Learning Test (sum of trials 1 to 5, delayed recall, and recognition) and Bells test (Supplementary Table 3). Moreover, PD-RBD patients with RBD onset prior to PD diagnosis performed worse than PD-RBD patients with RBD onset at the same time/after PD diagnosis on the Verbal Fluency (letter) and Boston Naming Test (Supplementary Table 4). In addition, a higher proportion of RBD-PD patients had clinically impaired performance compared to PD-nRBD patients on the following cognitive tests: the Stroop Color-Word Test (III–II, time and errors; IV–III, time and errors), Trail Making Test (part B, time), Verbal Fluency (semantic), Rey Auditory-Verbal Learning Test (sum of trials 1 to 5, immediate recall), and Rey–O Complex Figure Test (copy) (Figure 1). Figure 1 Open in new tabDownload slide Percentage of patients with clinically impaired performance on neuropsychological tests. ***p < .001; **p < .01; *p < .05. MMSE = Mini-Mental State Examination; PD-RBD = Parkinson’s disease with REM sleep behavior disorder; PD-nRBD = PD without RBD; RAVLT = Rey Auditory-Verbal Learning Test; ROCF = Rey-O Complex Figure; Stroop components: II = Naming, III = Interference, IV = Flexibility. Figure 1 Open in new tabDownload slide Percentage of patients with clinically impaired performance on neuropsychological tests. ***p < .001; **p < .01; *p < .05. MMSE = Mini-Mental State Examination; PD-RBD = Parkinson’s disease with REM sleep behavior disorder; PD-nRBD = PD without RBD; RAVLT = Rey Auditory-Verbal Learning Test; ROCF = Rey-O Complex Figure; Stroop components: II = Naming, III = Interference, IV = Flexibility. Comparing combined PD cohort to healthy subjects (Table 2), PD-RBD patients performed worse than controls on tests assessing attention (all tests), executive functions (Trail Making Test [part B, time], Stroop Color-Word Test [IV-III, errors], and Verbal Fluency [semantic and letter]), episodic verbal learning and memory (Rey Auditory-Verbal Learning Test [all variables]), visuospatial abilities (Block Design subtest [scaled score]), and language (Mini-Mental State Examination [language score]). No significant differences were found between PD-nRBD patients and controls for all cognitive tests. Overall, most of the previous results remained significant when comparing men and women only but with slightly lower effects when comparing women, probably due to the smaller sample size (Table 2). Correlations were performed for PD patients as one group between the percentage of REM sleep EMG activity and the cognitive variables (Table 3). Higher tonic EMG activity was associated with poorer performance on the Stroop Color-Word Test (III-II, time), Trail Making Test (part B, time and B–A, time), Rey Auditory-Verbal Learning Test (sum of trials 1 to 5), and Block Design subtest (scaled score). On the other hand, higher phasic EMG activity was related to poorer performance on the Stroop Color-Word Test (III-II, time), Trail Making Test (part B, time), Verbal Fluency (semantic), Rey Auditory-Verbal Learning Test (List B), and Block Design subtest (scaled score). Table 3 Correlations Between Cognitive Measures and REM Sleep EMG Features in all PD Patients. Neuropsychological tests EMG tonic EMG phasic Attention  Digit Span, scaled score 0.027 −0.138  Trail Making Test part A, time, secª 0.169 0.213 Stroop Color-Word Test  III-II, time, secª 0.366*** 0.288**  III-II, errorsª 0.176 0.113 Executive functions  Trail Making Test part B, time, secª 0.277** 0.222*  Part B–part A, time, secª 0.286** 0.205 Stroop Color-Word Test  IV-III time, secª 0.098 −0.108  IV-III, errorsª 0.156 0.134 Verbal fluency  Semantic −0.217 −0.279*  Letter −0.181 −0.115 Episodic learning and memory Verbal RAVLT  Sum of trials 1 to 5 −0.226* −0.183  List B −0.107 −0.272*  Immediate recall −0.174 −0.209  Delayed recall −0.161 −0.184 Recognition −0.118 0.028 Nonverbal Rey-O Complex Figure Test  Immediate recall −0.167 −0.197  Delayed recall −0.130 −0.184 Visuospatial  Rey-O Complex Figure Test, copy −0.194 −0.163  Block Design, scaled score −0.324** −0.251*  Bells test, omissionsª 0.120 0.029 Language  Boston Naming Test −0.123 −0.037  Vocabulary, scaled score 0.054 −0.115  MMSE language score −0.096 −0.087 Neuropsychological tests EMG tonic EMG phasic Attention  Digit Span, scaled score 0.027 −0.138  Trail Making Test part A, time, secª 0.169 0.213 Stroop Color-Word Test  III-II, time, secª 0.366*** 0.288**  III-II, errorsª 0.176 0.113 Executive functions  Trail Making Test part B, time, secª 0.277** 0.222*  Part B–part A, time, secª 0.286** 0.205 Stroop Color-Word Test  IV-III time, secª 0.098 −0.108  IV-III, errorsª 0.156 0.134 Verbal fluency  Semantic −0.217 −0.279*  Letter −0.181 −0.115 Episodic learning and memory Verbal RAVLT  Sum of trials 1 to 5 −0.226* −0.183  List B −0.107 −0.272*  Immediate recall −0.174 −0.209  Delayed recall −0.161 −0.184 Recognition −0.118 0.028 Nonverbal Rey-O Complex Figure Test  Immediate recall −0.167 −0.197  Delayed recall −0.130 −0.184 Visuospatial  Rey-O Complex Figure Test, copy −0.194 −0.163  Block Design, scaled score −0.324** −0.251*  Bells test, omissionsª 0.120 0.029 Language  Boston Naming Test −0.123 −0.037  Vocabulary, scaled score 0.054 −0.115  MMSE language score −0.096 −0.087 ªSpearman correlations. EMG = electromyography; MMSE = Mini-Mental State Examination; PD = Parkinson’s disease; RAVLT = Rey Auditory-Verbal Learning Test. ***p < .001; ** = p < .01; * = p < .05. Open in new tab Table 3 Correlations Between Cognitive Measures and REM Sleep EMG Features in all PD Patients. Neuropsychological tests EMG tonic EMG phasic Attention  Digit Span, scaled score 0.027 −0.138  Trail Making Test part A, time, secª 0.169 0.213 Stroop Color-Word Test  III-II, time, secª 0.366*** 0.288**  III-II, errorsª 0.176 0.113 Executive functions  Trail Making Test part B, time, secª 0.277** 0.222*  Part B–part A, time, secª 0.286** 0.205 Stroop Color-Word Test  IV-III time, secª 0.098 −0.108  IV-III, errorsª 0.156 0.134 Verbal fluency  Semantic −0.217 −0.279*  Letter −0.181 −0.115 Episodic learning and memory Verbal RAVLT  Sum of trials 1 to 5 −0.226* −0.183  List B −0.107 −0.272*  Immediate recall −0.174 −0.209  Delayed recall −0.161 −0.184 Recognition −0.118 0.028 Nonverbal Rey-O Complex Figure Test  Immediate recall −0.167 −0.197  Delayed recall −0.130 −0.184 Visuospatial  Rey-O Complex Figure Test, copy −0.194 −0.163  Block Design, scaled score −0.324** −0.251*  Bells test, omissionsª 0.120 0.029 Language  Boston Naming Test −0.123 −0.037  Vocabulary, scaled score 0.054 −0.115  MMSE language score −0.096 −0.087 Neuropsychological tests EMG tonic EMG phasic Attention  Digit Span, scaled score 0.027 −0.138  Trail Making Test part A, time, secª 0.169 0.213 Stroop Color-Word Test  III-II, time, secª 0.366*** 0.288**  III-II, errorsª 0.176 0.113 Executive functions  Trail Making Test part B, time, secª 0.277** 0.222*  Part B–part A, time, secª 0.286** 0.205 Stroop Color-Word Test  IV-III time, secª 0.098 −0.108  IV-III, errorsª 0.156 0.134 Verbal fluency  Semantic −0.217 −0.279*  Letter −0.181 −0.115 Episodic learning and memory Verbal RAVLT  Sum of trials 1 to 5 −0.226* −0.183  List B −0.107 −0.272*  Immediate recall −0.174 −0.209  Delayed recall −0.161 −0.184 Recognition −0.118 0.028 Nonverbal Rey-O Complex Figure Test  Immediate recall −0.167 −0.197  Delayed recall −0.130 −0.184 Visuospatial  Rey-O Complex Figure Test, copy −0.194 −0.163  Block Design, scaled score −0.324** −0.251*  Bells test, omissionsª 0.120 0.029 Language  Boston Naming Test −0.123 −0.037  Vocabulary, scaled score 0.054 −0.115  MMSE language score −0.096 −0.087 ªSpearman correlations. EMG = electromyography; MMSE = Mini-Mental State Examination; PD = Parkinson’s disease; RAVLT = Rey Auditory-Verbal Learning Test. ***p < .001; ** = p < .01; * = p < .05. Open in new tab Mild Cognitive Impairment and Subjective Cognitive Decline MCI diagnosis frequency and subtypes are presented in Table 4. MCI diagnosis was more frequent in PD-RBD patients (66%) compared to PD-nRBD patients (23%) (p < .001). Of the PD-RBD patients, 24% had MCI–single domain (9% amnestic MCI, 15% nonamnestic MCI), and 42% had MCI–multiple domain (23% amnestic MCI, 19% nonamnestic MCI). Of the PD-nRBD patients, 8% had the MCI–single domain subtype (3% amnestic MCI, 5% nonamnestic MCI) and 15% had MCI–multiple domain (7.5% amnestic MCI, 7.5% nonamnestic MCI). Three (33%) of the 9 PD-nRBD patients with MCI diagnosis had excessive REM sleep muscle tone without history or presence of movements during REM sleep during PSG (prodromal RBD). Moreover, SCD frequency was higher in PD-RBD patients without MCI (16 out of 18, 89%) compared to PD-nRBD patients without MCI (18 out of 31, 58%) (p = .024). Table 4 Mild Cognitive Impairment Diagnosis Frequency in PD Patients. MCI subtypes Combined PD-RBD Combined PD-nRBD n = 53 n = 40 MCI total, n 35 9 MCI–single domain, n 13 3  Amnesic MCI–single domain, n 5 1  Nonamnesic MCI–single domain, n 8 2   Executive, n 8 2 MCI–multiple domain, n 22 6  Nonamnesic MCI–multiple domain, n 10 3   Attention and executive, n 3 1   Attention and visuospatial, n 0 0   Executive and visuospatial, n 2 2   Attention, executive and visuospatial, n 4 0   Attention, executive and language, n 1 0  Amnesic MCI–multiple domain, n 12 3   + Attention, n 1 0   + Executive, n 3 0   + Attention and executive, n 4 0   + Attention and visuospatial, n 0 1   + Executive and visuospatial, n 4 2 MCI subtypes Combined PD-RBD Combined PD-nRBD n = 53 n = 40 MCI total, n 35 9 MCI–single domain, n 13 3  Amnesic MCI–single domain, n 5 1  Nonamnesic MCI–single domain, n 8 2   Executive, n 8 2 MCI–multiple domain, n 22 6  Nonamnesic MCI–multiple domain, n 10 3   Attention and executive, n 3 1   Attention and visuospatial, n 0 0   Executive and visuospatial, n 2 2   Attention, executive and visuospatial, n 4 0   Attention, executive and language, n 1 0  Amnesic MCI–multiple domain, n 12 3   + Attention, n 1 0   + Executive, n 3 0   + Attention and executive, n 4 0   + Attention and visuospatial, n 0 1   + Executive and visuospatial, n 4 2 MCI = Mild cognitive impairment; PD-RBD = Parkinson’s disease with REM sleep behavior disorder; PD-nRBD = PD without RBD. Open in new tab Table 4 Mild Cognitive Impairment Diagnosis Frequency in PD Patients. MCI subtypes Combined PD-RBD Combined PD-nRBD n = 53 n = 40 MCI total, n 35 9 MCI–single domain, n 13 3  Amnesic MCI–single domain, n 5 1  Nonamnesic MCI–single domain, n 8 2   Executive, n 8 2 MCI–multiple domain, n 22 6  Nonamnesic MCI–multiple domain, n 10 3   Attention and executive, n 3 1   Attention and visuospatial, n 0 0   Executive and visuospatial, n 2 2   Attention, executive and visuospatial, n 4 0   Attention, executive and language, n 1 0  Amnesic MCI–multiple domain, n 12 3   + Attention, n 1 0   + Executive, n 3 0   + Attention and executive, n 4 0   + Attention and visuospatial, n 0 1   + Executive and visuospatial, n 4 2 MCI subtypes Combined PD-RBD Combined PD-nRBD n = 53 n = 40 MCI total, n 35 9 MCI–single domain, n 13 3  Amnesic MCI–single domain, n 5 1  Nonamnesic MCI–single domain, n 8 2   Executive, n 8 2 MCI–multiple domain, n 22 6  Nonamnesic MCI–multiple domain, n 10 3   Attention and executive, n 3 1   Attention and visuospatial, n 0 0   Executive and visuospatial, n 2 2   Attention, executive and visuospatial, n 4 0   Attention, executive and language, n 1 0  Amnesic MCI–multiple domain, n 12 3   + Attention, n 1 0   + Executive, n 3 0   + Attention and executive, n 4 0   + Attention and visuospatial, n 0 1   + Executive and visuospatial, n 4 2 MCI = Mild cognitive impairment; PD-RBD = Parkinson’s disease with REM sleep behavior disorder; PD-nRBD = PD without RBD. Open in new tab DISCUSSION There is a lack of consensus in the literature on a cross-sectional association between RBD and poorer cognitive performance in nondemented individuals with PD.1 This is mainly due to methodological differences and limits between the studies. In the present study, we compared performance on a broad range of cognitive tests in two different cohorts of nondemented PD patients with and without RBD confirmed by PSG. We found highly probable or probable associations between the presence of RBD in PD and poorer performance on cognitive tests measuring attention, executive functions, episodic verbal learning and memory, and visuospatial abilities. Preliminary results in the PD-RBD group also show that men and patients with RBD onset prior to PD diagnosis are at higher risk of poorer cognitive performance. In the analyses using the combined PD cohort, we included a healthy control group to better interpret the results. We found that, in addition to the above-mentioned cognitive deficits, compared to controls, PD-RBD patients had poorer performance on cognitive test measuring delayed recall and recognition of verbal information, and language. PD patients without RBD had similar cognitive performance to controls on all cognitive tests. Two PSG manifestations of RBD, that is, excessive tonic and phasic EMG activity during REM sleep, were associated with poorer performance on several cognitive tests in PD. Moreover, using the proposed criteria for MCI diagnosis in PD,27 we found almost threefold higher frequency of MCI diagnosis in PD-RBD patients compared to PD patients without RBD. Taken together, these results indicate that RBD is strongly associated with cognitive impairment in PD. Our results are consistent with other longitudinal studies suggesting that the presence of RBD in individuals with PD is an important clinical risk factor for the development of dementia.7,11,39,40 Another clinical risk factor for cognitive decline and dementia in the general population is SCD, which is characterized by a self-perception of a decline in cognitive performance in daily life in the absence of objective cognitive impairment measured by neuropsychological assessment.41 In PD, the presence of SCD has been poorly studied. One study reported that SCD in PD with normal cognition predicts future cognitive decline.8 We found a higher frequency of SCD in PD-RBD patients. Although this result should be validated in a larger cohort and its predictive value determined in a longitudinal design, our results strengthen the link between RBD and the risk of cognitive decline in PD, even in patients with normal cognition. Consequently, PD patients with cognitive impairment should be carefully screened clinically for the presence of RBD. This subgroup of patients should also be targeted in future clinical trials on the progression of cognitive decline in PD. Cognitive impairment, determined by poorer cognitive performance or by the presence of MCI or dementia, is a well-known nonmotor feature in PD.2,27 Two distinct patterns of cognitive deficits have been identified in PD.27,42 The first pattern is more related to posterior cortically based cognitive deficits, probably reflecting nondopaminergic cortical dysfunction associated with explicit memory and visuospatial deficits. The second pattern is related more to frontostriatal cognitive deficits, reflecting dopaminergic dysfunctions associated with attention and executive deficits. It has been suggested that the dementia incidence in patients with PD was associated with posterior cortical deficits, whereas frontostriatal deficits were not.27,42 Nevertheless, patients with PD are commonly characterized by a range of heterogeneous cognitive impairments, even in early stage, suggesting that 15% of patients with PD can present with both patterns of cognitive deficits.42 In the present study, PD-RBD patients had poorer performance on cognitive tasks measuring working memory, visual search, mental flexibility, processing speed, cognitive inhibition, word retrieval and delayed recall of verbal information, and visuospatial organization compared to PD patients without RBD. Their cognitive profile reflects both frontostriatal and posterior cortical deficits. The mechanisms underlying the association between cognitive impairment and RBD in PD remain to be determined. RBD has been associated with lesions of the brainstem regions involved in muscle atonia and motor control during REM sleep.43 However, much evidence now suggests that RBD is more than a simple sleep disorder, and that patients having RBD are at high risk for developing dementia.44 Moreover, cortical dysfunctions such as cortical hypometabolism, EEG slowing, and cortical thickness have also been identified in RBD patients without PD.45–47 Individuals with concomitant PD and RBD have more specific brain anatomical and functional changes compared to PD patients without RBD. In deed, several studies using quantitative EEG, event-related potentials, magnetic resonance imaging (voxel-based morphometry), and positron emission tomography ([11C] methylpiperidyl propionate acetylcholinesterase) have reported brain dysfunctions associated with the presence of RBD in PD.48–52 Ford et al.50 reported subtle changes in white matter integrity (widespread) and reductions in gray matter volume (posterior areas) in PD with clinical RBD. Recently, a study also found smaller volumes in the pontomesencephalic tegmentum, medullary reticular formation, hypothalamus, thalamus, putamen, amygdala, and anterior cingulate cortex in PD patient with clinical RBD.51 In a functional neuroimaging study, Kotagal et al.52 showed that the presence of clinical RBD in PD could be accounted for by progressive neocortical, limbic, cortical, and thalamic cholinergic denervation. Dysfunctions of cholinergic systems, namely the nucleus basalis of Meynert and pedunculopontine nucleus, and their projections to subcortical and cortical regions have been related to cognitive impairment in PD.53,54 Other studies have identified a distinct clinical subtype in PD related to the presence of RBD, with higher risk for dysautonomia, hallucinations, freezing and falls, symmetric disease, and a nontremor dominant subtype.55–57 Some of these nonmotor features have also been related to cholinergic impairment in PD.54 This suggests a common cholinergic mechanism between RBD, cognitive decline, and other nonmotor features in this population. Interestingly, some of these previous studies reported that PD patients without RBD were similar to healthy controls except for motor impairment.48,55,56 This is similar to our results showing normal cognitive functioning in PD patients without RBD, suggesting that the presence of RBD in individuals with PD is associated with a more severe subtype of the disease, with a strong negative prognostic.40 Therefore, PD-RBD patients should be considered as a distinct PD subgroup, not only for neuroprotective trials or for examinations of the underlying pathophysiological mechanisms but also for the development of animal models of PD. It remains to determine whether RBD onset in the course of PD is the starting point for more rapid and pronounced neurodegeneration. Follow-up of these patients will be important to detect the emergence of cognitive decline, dysautonomia, hallucinations, and gait and balance difficulties and to provide medical attention and management to the patients and their caregivers. This study includes certain limitations. First, the PD-RBD group was older and contained a higher proportion of men compared to the PD-nRBD and control groups. Although we controlled for age and gender and performed additional analyses with gender as the independent variable, we cannot discount potential age or gender effects on some of our results. Moreover, the proportion of patients taking dopaminergic agonists was also higher in the PD-nRBD compared to the PD-RBD group. The chronic cognitive effects of dopaminergic agonists in PD remain controversial,58 and further studies are needed to better understand their impacts on PD in terms of cognition and RBD status. Second, there is neither a systematic definition of SCD nor a standardized method to measure SCD in PD populations. Our results should therefore be confirmed in larger PD populations using appropriate SCD assessment methods. Finally, although cognitive performance was similar between PD-nRBD and healthy subjects, future studies may find differences in larger-sized samples, or differences in cognitive functions that were not deeply examined in the present study (eg, language, planning, working memory, and procedural learning). In conclusion, RBD in PD is associated with higher risks for cognitive deficits. In addition, the presence of RBD in PD increases the risk of a MCI diagnosis and SCD, which are associated with cognitive decline and dementia development in PD. Future studies attempting to identify a distinct cognitive profile in PD with RBD should use a greater variety of tests to more deeply assess all the language components (ie, naming, reading, writing, understanding, and pragmatism) and higher executive functions (ie, planning, problem solving) as well as procedural learning. SUPPLEMENTARY MATERIAL Supplementary material is available at SLEEP online. ETHICS COMMITTEE APPROVAL All procedures were performed at the Centre for Advanced Research in Sleep Medicine at the Hôpital du Sacré-Coeur de Montréal (Quebec, Canada). The hospital’s ethical committee approved the study, and all subjects signed a written consent before participating. FUNDING RB Postuma received personal compensation for travel and speaker fees from Biotie, Biospective, Boehringer, Roche, Novartis Canada and Teva Neurosciences, and is funded by grants from the Canadian Institutes of Health Research (CIHR), Parkinson Society of Canada, W. Garfield Weston Foundation, Webster Foundation, and Fonds de Recherche du Québec - Santé (FRQ-S). J Montplaisir receives research support from GlaxoSmithKline, Merck Pharma, and the CIHR, Parkinson Society of Canada, and W. Garfield Weston Foundation. He holds a Canada Research Chair in Sleep Medicine. M. Panisset receives research support from the NIH, W. Garfield Weston Foundation, and Medtronic. He serves on advisory boards for Allergan and Merz. S Chouinard receives research support from Abbvie. JF Gagnon receives research support from the CIHR, W. Garfield Weston Foundation, Parkinson Society of Canada, and FRQ-S. He holds a Canada Research Chair in Cognitive Decline in Pathological Aging. N Jozwiak, V Latreille, and PA Bourgouin have nothing to disclose. DISCLOSURE STATEMENT None declared. ACKNOWLEDGMENTS This study was supported by grants from the Canadian Institutes of Health Research and Fonds de Recherche du Québec—Santé. REFERENCES 1. Gagnon JF , Postuma RB , Lyonnais-Lafond G . Cognition and the sleep–wake cycle in Parkinson’s disease . In: Videnovic A , & Högl B , eds. Disorders of Sleep and Circadian Rhythms in Parkinson’s Disease . Vienna, Austria : Springer-Verlag ; 2015 : 183 – 194 . Google Scholar Crossref Search ADS Google Preview WorldCat COPAC 2. Dubois B , Burn D , Goetz C et al. Diagnostic procedures for Parkinson’s disease dementia: recommendations from the movement disorder society task force . Mov Disord . 2007 ; 22 ( 16 ): 2314 – 2324 . Google Scholar Crossref Search ADS PubMed WorldCat 3. Aarsland D , Brønnick K , Fladby T . Mild cognitive impairment in Parkinson’s disease . Curr Neurol Neurosci Rep . 2011 ; 11 ( 4 ): 371 – 378 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Aarsland D , Kurz MW . The epidemiology of dementia associated with Parkinson’s disease . Brain Pathol . 2010 ; 20 ( 3 ): 633 – 639 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Riedel O , Klotsche J , Spottke A et al. Frequency of dementia, depression, and other neuropsychiatric symptoms in 1,449 outpatients with Parkinson’s disease . J Neurol . 2010 ; 257 ( 7 ): 1073 – 1082 . Google Scholar Crossref Search ADS PubMed WorldCat 6. Hely MA , Reid WG , Adena MA , Halliday GM , Morris JG . The Sydney multicenter study of Parkinson’s disease: the inevitability of dementia at 20 years . Mov Disord . 2008 ; 23 ( 6 ): 837 – 844 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Anang JB , Gagnon JF , Bertrand JA et al. Predictors of dementia in Parkinson disease: a prospective cohort study . Neurology . 2014 ; 83 ( 14 ): 1253 – 1260 . Google Scholar Crossref Search ADS PubMed WorldCat 8. Hong JY , Sunwoo MK , Chung SJ et al. Subjective cognitive decline predicts future deterioration in cognitively normal patients with Parkinson’s disease . Neurobiol Aging . 2014 ; 35 ( 7 ): 1739 – 1743 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Sixel-Döring F , Trautmann E , Mollenhauer B , Trenkwalder C . Associated factors for REM sleep behavior disorder in Parkinson disease . Neurology . 2011 ; 77 ( 11 ): 1048 – 1054 . Google Scholar Crossref Search ADS PubMed WorldCat 10. Gagnon JF , Bédard MA , Fantini ML et al. REM sleep behavior disorder and REM sleep without atonia in Parkinson’s disease . Neurology . 2002 ; 59 ( 4 ): 585 – 589 . Google Scholar Crossref Search ADS PubMed WorldCat 11. Nomura T , Inoue Y , Kagimura T , Nakashima K . Clinical significance of REM sleep behavior disorder in Parkinson’s disease . Sleep Med . 2013 ; 14 ( 2 ): 131 – 135 . Google Scholar Crossref Search ADS PubMed WorldCat 12. Sinforiani E , Zangaglia R , Manni R et al. REM sleep behavior disorder, hallucinations, and cognitive impairment in Parkinson’s disease . Mov Disord . 2006 ; 21 ( 4 ): 462 – 466 . Google Scholar Crossref Search ADS PubMed WorldCat 13. Gagnon JF , Vendette M , Postuma RB et al. Mild cognitive impairment in rapid eye movement sleep behavior disorder and Parkinson’s disease . Ann Neurol . 2009 ; 66 ( 1 ): 39 – 47 . Google Scholar Crossref Search ADS PubMed WorldCat 14. Erro R , Santangelo G , Picillo M et al. Link between non-motor symptoms and cognitive dysfunctions in de novo, drug-naive PD patients . J Neurol . 2012 ; 259 ( 9 ): 1808 – 1813 . Google Scholar Crossref Search ADS PubMed WorldCat 15. Marques A , Dujardin K , Boucart M et al. REM sleep behaviour disorder and visuoperceptive dysfunction: a disorder of the ventral visual stream? J Neurol . 2010 ; 257 ( 3 ): 383 – 391 . Google Scholar Crossref Search ADS PubMed WorldCat 16. Chahine LM , Xie SX , Simuni T et al. Longitudinal changes in cognition in early Parkinson’s disease patients with REM sleep behavior disorder . Parkinsonism Relat Disord . 2016 ; 27 : 102 – 106 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Gong Y , Xiong KP , Mao CJ et al. Clinical manifestations of Parkinson disease and the onset of rapid eye movement sleep behavior disorder . Sleep Med . 2014 ; 15 ( 6 ): 647 – 653 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Wang G , Wan Y , Wang Y et al. Visual hallucinations and associated factors in Chinese patients with Parkinson’s disease: roles of RBD and visual pathway deficit . Parkinsonism Relat Disord . 2010 ; 16 ( 10 ): 695 – 696 . Google Scholar Crossref Search ADS PubMed WorldCat 19. Zhang JR , Chen J , Yang ZJ et al. Rapid eye movement sleep behavior disorder symptoms correlate with domains of cognitive impairment in parkinson’s disease . Chin Med J (Engl) . 2016 ; 129 ( 4 ): 379 – 385 . Google Scholar Crossref Search ADS PubMed WorldCat 20. Plomhause L , Dujardin K , Duhamel A et al. Rapid eye movement sleep behavior disorder in treatment-naïve Parkinson disease patients . Sleep Med . 2013 ; 14 ( 10 ): 1035 – 1037 . Google Scholar Crossref Search ADS PubMed WorldCat 21. Bugalho P , da Silva JA , Neto B . Clinical features associated with REM sleep behavior disorder symptoms in the early stages of Parkinson’s disease . J Neurol . 2011 ; 258 ( 1 ): 50 – 55 . Google Scholar Crossref Search ADS PubMed WorldCat 22. Lavault S , Leu-Semenescu S , Tezenas du Montcel S , Cochen de Cock V , Vidailhet M , Arnulf I . Does clinical rapid eye movement behavior disorder predict worse outcomes in Parkinson’s disease? J Neurol . 2010 ; 257 ( 7 ): 1154 – 1159 . Google Scholar Crossref Search ADS PubMed WorldCat 23. Yoritaka A , Ohizumi H , Tanaka S , Hattori N . Parkinson’s disease with and without REM sleep behaviour disorder: are there any clinical differences? Eur Neurol . 2009 ; 61 ( 3 ): 164 – 170 . Google Scholar Crossref Search ADS PubMed WorldCat 24. Benninger D , Waldvogel D , Bassetti CL . REM sleep behavior disorder predicts cognitive impairment in Parkinson disease without dementia . Neurology . 2008 ; 71 ( 12 ): 955 – 956 . Google Scholar Crossref Search ADS PubMed WorldCat 25. Lee JE , Kim KS , Shin HW , Sohn YH . Factors related to clinically probable REM sleep behavior disorder in Parkinson disease . Parkinsonism Relat Disord . 2010 ; 16 ( 2 ): 105 – 108 . Google Scholar Crossref Search ADS PubMed WorldCat 26. Sixel-Döring F , Trautmann E , Mollenhauer B , Trenkwalder C . Rapid eye movement sleep behavioral events: a new marker for neurodegeneration in early Parkinson disease? Sleep . 2014 ; 37 ( 3 ): 431 – 438 . Google Scholar Crossref Search ADS PubMed WorldCat 27. Litvan I , Goldman JG , Tröster AI et al. Diagnostic criteria for mild cognitive impairment in Parkinson’s disease: Movement Disorder Society Task Force guidelines . Mov Disord . 2012 ; 27 ( 3 ): 349 – 356 . Google Scholar Crossref Search ADS PubMed WorldCat 28. National Collaborating Centre for Chronic Conditions . Parkinson’s Disease: National Clinical Guideline for Diagnosis and Management in Primary and Secondary Care . London : Royal College of Physicians of London ; 2006 : 22 – 47 . WorldCat COPAC 29. American Psychiatric Association . Diagnostic and Statistical Manual of Mental Disorders .4th ed. Text revision (DSM-IV-TR). Washington, DC : American Psychiatric Association ; 2000 . WorldCat COPAC 30. Montplaisir J , Gagnon JF , Fantini ML et al. Polysomnographic diagnosis of idiopathic REM sleep behavior disorder . Mov Disord . 2010 ; 25 ( 13 ): 2044 – 2051 . Google Scholar Crossref Search ADS PubMed WorldCat 31. American Academy of Sleep Medicine . The International Classification of Sleep Disorders: Diagnostic and Coding Manual .2nd ed. Westchester, IL : American Academy of Sleep Medicine ; 2005 . WorldCat COPAC 32. The Unified Parkinson’s Disease Rating Scale (UPDRS): status and recommendations . Mov Disord . 2003 ; 18 ( 7 ): 738 – 750 . Crossref Search ADS PubMed WorldCat 33. Beck AT , Steer RA , Brown GK. The Beck Depression Inventory .2nd ed. San Antonio : Psychological Corporation ; 1996 . Google Preview WorldCat COPAC 34. Beck AT , Epstein N , Brown G , Steer RA . An inventory for measuring clinical anxiety: psychometric properties . J Consult Clin Psychol . 1988 ; 56 ( 6 ): 893 – 897 . Google Scholar Crossref Search ADS PubMed WorldCat 35. Johns MW . A new method for measuring daytime sleepiness: the Epworth sleepiness scale . Sleep . 1991 ; 14 ( 6 ): 540 – 545 . Google Scholar Crossref Search ADS PubMed WorldCat 36. Morin CM , Belleville G , Bélanger L , Ivers H . The Insomnia Severity Index: psychometric indicators to detect insomnia cases and evaluate treatment response . Sleep . 2011 ; 34 ( 5 ): 601 – 608 . Google Scholar Crossref Search ADS PubMed WorldCat 37. Strauss E , Sherman EM , Spreen O. A Compendium of Neuropsychological Tests: Administration, Norms, and Commentary .3rd ed. New York : Oxford University Press ; 2006 . Google Preview WorldCat COPAC 38. Broadbent DE , Cooper PF , FitzGerald P , Parkes KR . The Cognitive Failures Questionnaire (CFQ) and its correlates . Br J Clin Psychol . 1982 ; 21 ( 1 ): 1 – 16 . Google Scholar Crossref Search ADS PubMed WorldCat 39. Anang JB , Nomura T , Romenets SR , Nakashima K , Gagnon JF , Postuma RB . Dementia predictors in Parkinson disease: a validation study . J Parkinsons Dis . 2017 ; 7 ( 1 ): 159 – 162 . Google Scholar Crossref Search ADS PubMed WorldCat 40. Fereshtehnejad SM , Romenets SR , Anang JB , Latreille V , Gagnon JF , Postuma RB . New clinical subtypes of parkinson disease and their longitudinal progression: a prospective cohort comparison with other phenotypes . JAMA Neurol . 2015 ; 72 ( 8 ): 863 – 873 . Google Scholar Crossref Search ADS PubMed WorldCat 41. Molinuevo JL , Rabin LA , Amariglio R et al. ; Subjective Cognitive Decline Initiative (SCD-I) Working Group . Implementation of subjective cognitive decline criteria in research studies . Alzheimers Dement . 2017 ; 13 ( 3 ): 296 – 311 . Google Scholar Crossref Search ADS PubMed WorldCat 42. Williams-Gray CH , Evans JR , Goris A et al. The distinct cognitive syndromes of Parkinson’s disease: 5 year follow-up of the CamPaIGN cohort . Brain . 2009 ; 132 ( 11 ): 2958 – 2969 . Google Scholar Crossref Search ADS PubMed WorldCat 43. Jennum P , Christensen JA , Zoetmulder M . Neurophysiological basis of rapid eye movement sleep behavior disorder: informing future drug development . Nat Sci Sleep . 2016 ; 8 : 107 – 120 . Google Scholar Crossref Search ADS PubMed WorldCat 44. Iranzo A , Fernández-Arcos A , Tolosa E et al. Neurodegenerative disorder risk in idiopathic REM sleep behavior disorder: study in 174 patients . PLoS One . 2014 ; 9 ( 2 ): e89741 . Google Scholar Crossref Search ADS PubMed WorldCat 45. Vendette M , Montplaisir J , Gosselin N et al. Brain perfusion anomalies in rapid eye movement sleep behavior disorder with mild cognitive impairment . Mov Disord . 2012 ; 27 ( 10 ): 1255 – 1261 . Google Scholar Crossref Search ADS PubMed WorldCat 46. Rahayel S , Montplaisir J , Monchi O et al. Patterns of cortical thinning in idiopathic rapid eye movement sleep behavior disorder . Mov Disord . 2015 ; 30 ( 5 ): 680 – 687 . Google Scholar Crossref Search ADS PubMed WorldCat 47. Rodrigues Brazète J , Montplaisir J , Petit D et al. Electroencephalogram slowing in rapid eye movement sleep behavior disorder is associated with mild cognitive impairment . Sleep Med . 2013 ; 14 ( 11 ): 1059 – 1063 . Google Scholar Crossref Search ADS PubMed WorldCat 48. Gagnon JF , Fantini ML , Bédard MA et al. Association between waking EEG slowing and REM sleep behavior disorder in PD without dementia . Neurology . 2004 ; 62 ( 3 ): 401 – 406 . Google Scholar Crossref Search ADS PubMed WorldCat 49. Gaudreault PO , Gagnon JF , Montplaisir J et al. Abnormal occipital event-related potentials in Parkinson’s disease with concomitant REM sleep behavior disorder . Parkinsonism Relat Disord . 2013 ; 19 ( 2 ): 212 – 217 . Google Scholar Crossref Search ADS PubMed WorldCat 50. Ford AH , Duncan GW , Firbank MJ et al. Rapid eye movement sleep behavior disorder in Parkinson’s disease: magnetic resonance imaging study . Mov Disord . 2013 ; 28 ( 6 ): 832 – 836 . Google Scholar Crossref Search ADS PubMed WorldCat 51. Boucetta S , Salimi A , Dadar M , Jones BE , Collins DL , Dang-Vu TT . Structural brain alterations associated with rapid eye movement sleep behavior disorder in Parkinson’s disease . Sci Rep . 2016 ; 6 : 26782 . Google Scholar Crossref Search ADS PubMed WorldCat 52. Kotagal V , Albin RL , Müller ML et al. Symptoms of rapid eye movement sleep behavior disorder are associated with cholinergic denervation in Parkinson disease . Ann Neurol . 2012 ; 71 ( 4 ): 560 – 568 . Google Scholar Crossref Search ADS PubMed WorldCat 53. Müller ML , Bohnen NI . Cholinergic dysfunction in Parkinson’s disease . Curr Neurol Neurosci Rep . 2013 ; 13 ( 9 ): 377 . Google Scholar Crossref Search ADS PubMed WorldCat 54. Yarnall A , Rochester L , Burn DJ . The interplay of cholinergic function, attention, and falls in Parkinson’s disease . Mov Disord . 2011 ; 26 ( 14 ): 2496 – 2503 . Google Scholar Crossref Search ADS PubMed WorldCat 55. Postuma RB , Montplaisir J , Lanfranchi P et al. Cardiac autonomic denervation in Parkinson’s disease is linked to REM sleep behavior disorder . Mov Disord . 2011 ; 26 ( 8 ): 1529 – 1533 . Google Scholar Crossref Search ADS PubMed WorldCat 56. Romenets SR , Gagnon JF , Latreille V et al. Rapid eye movement sleep behavior disorder and subtypes of Parkinson’s disease . Mov Disord . 2012 ; 27 ( 8 ): 996 – 1003 . Google Scholar Crossref Search ADS PubMed WorldCat 57. Videnovic A , Marlin C , Alibiglou L , Planetta PJ , Vaillancourt DE , Mackinnon CD . Increased REM sleep without atonia in Parkinson disease with freezing of gait . Neurology . 2013 ; 81 ( 12 ): 1030 – 1035 . Google Scholar Crossref Search ADS PubMed WorldCat 58. Poletti M , Bonuccelli U . Acute and chronic cognitive effects of levodopa and dopamine agonists on patients with Parkinson’s disease: a review . Ther Adv Psychopharmacol . 2013 ; 3 ( 2 ): 101 – 113 . Google Scholar Crossref Search ADS PubMed WorldCat Author notes Address correspondence to: Jean-Francois Gagnon, PhD, Département de psychologie, Université du Québec à Montréal, C.P. 8888 succ. Centre-ville, Montréal (Québec), Canada, H3C 3P8. Telephone: (514) 987 4184; Fax: (514) 987 7953; Email: [email protected] © Sleep Research Society 2017. Published by Oxford University Press on behalf of the Sleep Research Society. All rights reserved. For permissions, please e-mail [email protected].

Denis, Dan; Akhtar, Reece; Holding, Benjamin C; Murray, Christina; Panatti, Jennifer; Claridge, Gordon; Sadeh, Avi; Barclay, Nicola L; O’Leary, Rachael; Maughan, Barbara; McAdams, Tom A; Rowe, Richard; Eley, Thalia C; Viding, Essi; Gregory, Alice M

doi: 10.1093/sleep/zsx070pmid: 28575510

Blom,, Kerstin;Jernelöv,, Susanna;Rück,, Christian;Lindefors,, Nils;Kaldo,, Viktor

doi: 10.1093/sleep/zsx108pmid: 28655183

Abstract This 3-year follow-up compared insomnia treatment to depression treatment for patients with both diagnoses. Forty-three participants were randomized to either treatment, in the form of Internet-delivered therapist-guided cognitive behavior therapy (CBT), and 37 (86%) participants provided primary outcome data at the 3-year follow-up. After 3 years, reductions on depression severity were similar in both groups (between-group effect size, d = 0.33, p = .45), while the insomnia treatment had superior effects on insomnia severity (d = 0.66, p < .05). Overall, insomnia treatment was thus more beneficial than depression treatment. The implication for practitioners, supported by previous research, is that patients with co-occurring depression and insomnia should be offered CBT for insomnia, in addition to medication or psychological treatment for depression. Insomnia, Depression, Cognitive behavioral therapy, Insomnia–comorbid, Long-term follow-up Statement of Significance The treatment paradigm for patients with co-occurring depression and insomnia has mostly been to focus on the depression. Research shows that this is not sufficient—untreated insomnia hinders recovery and increases risk for relapse into depression. This randomized controlled trial, with a 3-year follow-up, directly compares insomnia treatment to depression treatment (both cognitive behavior therapy, CBT) for 43 patients with insomnia and depression. At the 3-year follow-up, we conclude that treating insomnia was overall more beneficial than treating depression. Further research is needed to explore the mechanisms leading to this result. The implication for practitioners, supported by previous research, is that patients with co-occurring depression and insomnia should be offered CBT for insomnia. INTRODUCTION Major depression is well known as a highly prevalent, serious, and costly disorder for individuals and society. Sleep problems in the form of insomnia disorder are even more common but have historically been seen as symptoms secondary to other disorders. However, insomnia leads to great suffering and large societal costs.1 A majority of depressed individuals have co-occurring insomnia, but the view that sleep problems are secondary to depression is challenged by accumulating evidence: insomnia is a risk factor for depression,2,3 depression treatment is not a sufficient remedy for insomnia,4,5 insomnia hinders recovery from depression,6,7 and residual insomnia increases the risk for depression relapse.8,9 These and other findings have resulted in updates of diagnosis manuals,10,11 to reflect that insomnia disorder can and should be diagnosed as a comorbid, rather than secondary disorder when co-occurring with other disorders. Still, treatment practices have not yet shifted accordingly, and comorbid insomnia is thus often not adequately treated. National treatment guidelines for psychiatric disorders (eg, NICE, APA) generally focus on depression, and guidelines for insomnia are often missing entirely. Cognitive behavior therapy (CBT) for insomnia (CBT-i) has superior long-term effects compared to sleep medication.12 It is also what most patients prefer but rarely get.13,14 A way of increasing access to CBT-i is to provide it via the Internet (ICBT-i), which in many studies has been proven effective,15,16 as has CBT for depression (ICBT-d).17 There are, however, very few randomized controlled follow-ups of more than 6–12 months for depression and insomnia treatments. In order to challenge the treatment paradigm for co-occurring insomnia and depression, we compared ICBT-i to ICBT-d for patients with both diagnoses. Our previously published results from this randomized controlled trial (RCT), with a 12-month follow-up, showed that ICBT-i was overall more beneficial than ICBT-d: its effects on insomnia were larger, while the effects on depression were similar in both groups. Participants in ICBT-i also reduced their sleep medication use more and had lower self-rated need for further treatment.18 This brief report presents a 36-month follow-up of our previous RCT, to explore whether the superiority of the insomnia treatment is maintained. METHOD This was a 36-month follow-up (FU36) of an RCT conducted at the Internet Psychiatry Clinic in Stockholm, Sweden. The study is described in more detail in the original report.18 Participants were 43 adults (53% women, mean age 47 years, 63% used sleep medication and 30% antidepressants) diagnosed with major depression and insomnia by psychiatrists using the M.I.N.I. interview.19 Other comorbidities were allowed, except sleep apnea, narcolepsy, and bipolar disorder. Primary outcome measures were the Self-Rated Montgomery Åsberg Depression Rating Scale (MADRS-S)20–22 and the Insomnia Severity Index (ISI).23–27 Secondary measures at the 36-month follow-up were remission rates and use of sleep and depression medication. See Figure 1 for a study flow chart. Figure 1 Open in new tabDownload slide Participant flowchart. ICBT = therapist-guided Internet-based cognitive behavioral therapy; ISI = Insomnia Severity Index; MADRS-S = Self-Rated Montgomery Åsberg Depression Rating Scale. Figure 1 Open in new tabDownload slide Participant flowchart. ICBT = therapist-guided Internet-based cognitive behavioral therapy; ISI = Insomnia Severity Index; MADRS-S = Self-Rated Montgomery Åsberg Depression Rating Scale. Patients were randomized to 9 weeks of therapist-guided Internet-delivered CBT, either for insomnia (ICBT-i) or for depression (ICBT-d). ICBT-d has been tested in an RCT, using a recommended treatment length of 8 weeks and ending up with an average treatment time of 10 weeks.28 It has also been used in regular care at the Internet Psychiatry Clinic since 2007, with a treatment time of 12 weeks to allow for 2 weeks’ absence. Data from the first 1203 patients have been reported in an effectiveness study.29 The original treatment included a module on sleep, which was removed for the purpose of this trial and allowed for a 9-week treatment window, with one module per week. The insomnia treatment, ICBT-i, has been tested in two RCTs,30,31 including one 3-year follow-up,32 using a treatment time of 8 weeks. More detailed information about the treatments can be found in the original article. The follow-up data presented here were collected using self-rated online questionnaires and blinded telephone interviews, used for diagnosing and to impute primary data for those with missing online data, according to the method described by Hedman et al.33 Statistics used were hierarchical linear mixed effect modeling, to handle missing data and ensure an intent-to-treat analysis.34,35 The data were best modeled by using two periods; time-piece 1 from pre to post, including eight weekly measures, and time-piece 2 from post to FU36, including follow-ups after 6 and 12 months (FU6 and FU12). The final model had a random intercept, a time × group interaction for time-piece 1, and no interaction for time-piece 2, that is, the groups had parallel slopes during the follow-up period. Since MADRS-S has an item measuring sleep, a sensitivity analysis was done for MADRS-S without the sleep item. Chi-square or Fisher exact test was used for dichotomous data. Cohen’s d for effect size was based on observed data. The trial was registered at clinicaltrials.gov (NCT01256099) and approved by the Regional Ethical Review Board in Stockholm (2009/1810–31/3). RESULTS The follow-up response rate at FU36 was high, (86%, see Figure 1). Table 1 presents detailed outcome data. Between-group effects sizes at FU36 were: Cohen’s d (ISI) = 0.66 (95% confidence interval [CI] = 0.1–1.22) and Cohen’s d (MADRS-S) = 0.33 (95% CI = −0.33 to 0.99), both in favor of ICBT-i. Both groups showed statistically significant improvements from pretreatment to posttreatment on both ISI and MADRS-S. There was a significant interaction effect pre to post on ISI, (F = 6.64, df = 458, p = .010). From post to FU36, the group difference on ISI was maintained and significant (p = .047). For MADRS-S, there was no interaction from pre to post (F = 0.29, df = 456, p = .59) nor a significant group difference during the follow-up period (p = .45). The sensitivity analysis of MADRS-S without the sleep item did not affect the results in any significant way. Figure 2 shows observed and estimated means for ISI and MADRS-S. Table 1 Primary Outcome Measures, Descriptives of Observed Data, and Statistical Tests. Measure Group Pre Post FU6 FU12 FU36 Effect size within group (Cohen’s d) Mixed models (Scale range) M (SD) M (SD) M (SD) M (SD) M (SD) Pre-Post Pre-FU6 Pre-FU12 Pre-FU36 Test p n n n n n ITT (n = 43) ISI ICBT-i 18.6 (4.0) 22 13.0 (6.4) 21 10.9 (5.9) 21 10.2 (5.0) 20 11.9 (4.9) 20 1.06 1.54 1.84 1.42 ICBT-i pre-post <.001 (0–28) ICBT-d pre and post <.001 ICBT-d 20.0 (4.1) 21 17.0 (7.0) 21 15.6 (6.7) 20 14.6 (7.3) 19 15.5 (6.8) 17 0.54 0.81 0.95 0.75 Interaction pre and post .01 Group diff post-FU36 .047 MADRS-S ICBT-i 25.1 (5.9) 22 18.7 (11.4) 21 15.7 (8.6) 21 16.7 (9.2) 20 16.4 (7.4) 20 0.74 1.30 1.12 1.23 ICBT-i pre-FU36 <.001 (0–54) ICBT-d pre-FU36 <.001 ICBT-d 26.0 (6.9) 21 20.5 (9.8) 21 18.0 (7.1) 20 18.5 (9.1) 19 19.0 (9.0) 17 0.66 1.14 0.94 1.01 Interaction pre and post .59 Group diff post-FU36 .45 Measure Group Pre Post FU6 FU12 FU36 Effect size within group (Cohen’s d) Mixed models (Scale range) M (SD) M (SD) M (SD) M (SD) M (SD) Pre-Post Pre-FU6 Pre-FU12 Pre-FU36 Test p n n n n n ITT (n = 43) ISI ICBT-i 18.6 (4.0) 22 13.0 (6.4) 21 10.9 (5.9) 21 10.2 (5.0) 20 11.9 (4.9) 20 1.06 1.54 1.84 1.42 ICBT-i pre-post <.001 (0–28) ICBT-d pre and post <.001 ICBT-d 20.0 (4.1) 21 17.0 (7.0) 21 15.6 (6.7) 20 14.6 (7.3) 19 15.5 (6.8) 17 0.54 0.81 0.95 0.75 Interaction pre and post .01 Group diff post-FU36 .047 MADRS-S ICBT-i 25.1 (5.9) 22 18.7 (11.4) 21 15.7 (8.6) 21 16.7 (9.2) 20 16.4 (7.4) 20 0.74 1.30 1.12 1.23 ICBT-i pre-FU36 <.001 (0–54) ICBT-d pre-FU36 <.001 ICBT-d 26.0 (6.9) 21 20.5 (9.8) 21 18.0 (7.1) 20 18.5 (9.1) 19 19.0 (9.0) 17 0.66 1.14 0.94 1.01 Interaction pre and post .59 Group diff post-FU36 .45 Group diff post-FU36 = test of difference between groups during the period from post to FU36; FU6 = 6-month follow-up; FU12 = 12-month follow-up; FU36 = 36-month follow-up; ICBT-I = group receiving Internet-based cognitive behavior therapy for insomnia; ICBT-i pre-post = test of change within the ICBT-i group pre-post; ICBT-d = group receiving Internet-based cognitive behavior therapy for depression; ICBT-d pre-post = test of change within the ICBT-d group pre-post; Interaction pre-post = test of interaction effect time × group pre-post; ISI = Insomnia Severity Index; ITT = intent-to-treat-analysis, that is, all 43 participants included; MADRS-S = Montgomery Åsberg Depression Rating Scale—Self rating; n, number of participants providing primary outcome data; Pre = before treatment; Post = after treatment. Open in new tab Table 1 Primary Outcome Measures, Descriptives of Observed Data, and Statistical Tests. Measure Group Pre Post FU6 FU12 FU36 Effect size within group (Cohen’s d) Mixed models (Scale range) M (SD) M (SD) M (SD) M (SD) M (SD) Pre-Post Pre-FU6 Pre-FU12 Pre-FU36 Test p n n n n n ITT (n = 43) ISI ICBT-i 18.6 (4.0) 22 13.0 (6.4) 21 10.9 (5.9) 21 10.2 (5.0) 20 11.9 (4.9) 20 1.06 1.54 1.84 1.42 ICBT-i pre-post <.001 (0–28) ICBT-d pre and post <.001 ICBT-d 20.0 (4.1) 21 17.0 (7.0) 21 15.6 (6.7) 20 14.6 (7.3) 19 15.5 (6.8) 17 0.54 0.81 0.95 0.75 Interaction pre and post .01 Group diff post-FU36 .047 MADRS-S ICBT-i 25.1 (5.9) 22 18.7 (11.4) 21 15.7 (8.6) 21 16.7 (9.2) 20 16.4 (7.4) 20 0.74 1.30 1.12 1.23 ICBT-i pre-FU36 <.001 (0–54) ICBT-d pre-FU36 <.001 ICBT-d 26.0 (6.9) 21 20.5 (9.8) 21 18.0 (7.1) 20 18.5 (9.1) 19 19.0 (9.0) 17 0.66 1.14 0.94 1.01 Interaction pre and post .59 Group diff post-FU36 .45 Measure Group Pre Post FU6 FU12 FU36 Effect size within group (Cohen’s d) Mixed models (Scale range) M (SD) M (SD) M (SD) M (SD) M (SD) Pre-Post Pre-FU6 Pre-FU12 Pre-FU36 Test p n n n n n ITT (n = 43) ISI ICBT-i 18.6 (4.0) 22 13.0 (6.4) 21 10.9 (5.9) 21 10.2 (5.0) 20 11.9 (4.9) 20 1.06 1.54 1.84 1.42 ICBT-i pre-post <.001 (0–28) ICBT-d pre and post <.001 ICBT-d 20.0 (4.1) 21 17.0 (7.0) 21 15.6 (6.7) 20 14.6 (7.3) 19 15.5 (6.8) 17 0.54 0.81 0.95 0.75 Interaction pre and post .01 Group diff post-FU36 .047 MADRS-S ICBT-i 25.1 (5.9) 22 18.7 (11.4) 21 15.7 (8.6) 21 16.7 (9.2) 20 16.4 (7.4) 20 0.74 1.30 1.12 1.23 ICBT-i pre-FU36 <.001 (0–54) ICBT-d pre-FU36 <.001 ICBT-d 26.0 (6.9) 21 20.5 (9.8) 21 18.0 (7.1) 20 18.5 (9.1) 19 19.0 (9.0) 17 0.66 1.14 0.94 1.01 Interaction pre and post .59 Group diff post-FU36 .45 Group diff post-FU36 = test of difference between groups during the period from post to FU36; FU6 = 6-month follow-up; FU12 = 12-month follow-up; FU36 = 36-month follow-up; ICBT-I = group receiving Internet-based cognitive behavior therapy for insomnia; ICBT-i pre-post = test of change within the ICBT-i group pre-post; ICBT-d = group receiving Internet-based cognitive behavior therapy for depression; ICBT-d pre-post = test of change within the ICBT-d group pre-post; Interaction pre-post = test of interaction effect time × group pre-post; ISI = Insomnia Severity Index; ITT = intent-to-treat-analysis, that is, all 43 participants included; MADRS-S = Montgomery Åsberg Depression Rating Scale—Self rating; n, number of participants providing primary outcome data; Pre = before treatment; Post = after treatment. Open in new tab Figure 2 Open in new tabDownload slide Change in insomnia and depression severity—comparison between treatments, estimated and observed means. ISI = Insomnia Severity Index, mean values; MADRS-S = Montgomery Åsberg Depression Rating Scale—Self rating, mean values; w = weeks; mo = months; y = years; ICBT-i = group receiving Internet-based cognitive behavior therapy for insomnia; ICBT-d = group receiving Internet-based cognitive behavior therapy for depression; Est. = estimated means from hierarchical mixed modeling, Obs. = observed means. Figure 2 Open in new tabDownload slide Change in insomnia and depression severity—comparison between treatments, estimated and observed means. ISI = Insomnia Severity Index, mean values; MADRS-S = Montgomery Åsberg Depression Rating Scale—Self rating, mean values; w = weeks; mo = months; y = years; ICBT-i = group receiving Internet-based cognitive behavior therapy for insomnia; ICBT-d = group receiving Internet-based cognitive behavior therapy for depression; Est. = estimated means from hierarchical mixed modeling, Obs. = observed means. Blinded telephone assessments at FU36 diagnosed major depression in six, that is, 35%, of the interviewed participants in ICBT-i, and eight, that is, 57%, of the interviewed participants in ICBT-d, and insomnia in six from each group (33% and 43%, respectively). At FU36, three participants in ICBT-i used sleep medication daily and four used it irregularly, while six in ICBT-d used sleep medication daily. Eight participants in ICBT-i versus five in ICBT-d used antidepressant medication. None of these differences were statistically significant. DISCUSSION After 3 years, the improvements in both symptom areas were maintained. The insomnia treatment was superior to the depression treatment with regard to insomnia symptoms, while there was no significant difference between treatment groups in symptoms of depression. Fifty-six and 62% of the interviewed participants no longer had a diagnosis of depression and insomnia, respectively. Contrary to the findings up to the 12-month follow-up,18 there were no significant differences between groups regarding the number of remitters or sleep medication users, the latter being reduced by 50% compared to baseline in both groups. Previously, only one study of CBT-i has had an equally long follow-up period, also reporting maintained improvements.32 Another recent study, comparing antidepressants plus CBT-i to antidepressants plus a control treatment, had similar results: effects on depression were similar in both groups and effects on insomnia were larger in the CBT-i group. They also found that improvements in insomnia mediated depression outcomes.36 A possible explanation to our findings, is that insomnia needs to be treated with specific methods that have specific effects,37,38 while many different types of depression treatment have been found effective, as long as they were structured and delivered with quality.39 In fact, several previous studies of CBT-i have reported reduced levels of co-occurring depressive symptoms after treatment.40–42 In a qualitative study of the sample in our study, we found that participants in ICBT-i were more positive about the treatment than those in ICBT-d, and that higher symptom levels and additional comorbidities were more hindering in the work with ICBT-d.43 Even though symptom reductions were large for ICBT-i and moderate to large for ICBT-d, the symptom levels after 30 years were less than satisfactory. At baseline, the participants had unusually high symptom levels, compared to other studies, and 51% reported sleeping less than 6 hours on average. The aforementioned qualitative study indicated that this is likely to have hindered their work with the treatments, especially in ICBT-d, and another recent study found that insomniacs with short objective sleep duration were less responsive to CBT-i.44 Our results indicate that while ICBT-i was superior, it is likely not a sufficient treatment for a majority of these patients. A limitation of this study is the small sample size and thereby low statistical power. Even though we found significant differences for insomnia severity, the nonsignificant results, particularly on depression severity, may be the result of low power. The small sample size also makes the issue with missing data at the 3-year follow-up extra important, even though the relative number of missing data is rather small for a long-term follow-up. We handle attrition by using the recommended way to control for missing data, hierarchical linear mixed effect modeling,34,35 to ensure an intent-to-treat-analysis that is state-of-the-art. Another limitation is the depression treatment length. Though supported by the previous evidence for the same treatment program,28,29 it was shorter than many face-to-face treatments for depression, and outcomes might have been different if we had used another mode of delivery or treatment length. Also, this study used guided Internet treatments, which may impact the characteristics of the participants and their involvement in the therapeutic techniques presented. It is thus possible that the results could differ from a similar study made in a face-to-face context. A replication in such a setting is recommended. Due to the above limitations, the results should be interpreted with caution and viewed in the context of other studies on the topic of comorbid insomnia and depression. Taken together, the evidence to date points to the treatment of insomnia as being important for this patient group, both as a way of reducing insomnia severity long term, as a mediator in the recovery from depression, and possibly as prevention of relapse into depression. This calls for an increased and more specific focus on treating insomnia when it is presented together with depression. CONCLUSION CBT-i was effective in treating both insomnia and depression, and the results were maintained 3 years after treatment. According to this and other studies on this patient group, patients with co-occurring insomnia and depression should ideally be offered CBT-i in addition to medication or psychological treatment for depression. FUNDING This project was funded by the regional agreement on medical training and clinical research (ALF) between Stockholm County Council and Karolinska Institutet, Söderström-Königska Foundation, KI funds and AFA Sickness Insurance Research Fund. CLINICAL TRIAL REGISTRATION clinicaltrials.gov no. NCT01256099 DISCLOSURE STATEMENT No conflicts of interests. ACKNOWLEDGMENTS The authors would like to thank Annette Skeppling and Daniel Björkander for their work with collecting the data for this paper. REFERENCES 1. Daley M , Morin CM , LeBlanc M , Gregoire JP , Savard J . The economic burden of insomnia: direct and indirect costs for individuals with insomnia syndrome, insomnia symptoms, and good sleepers . Sleep . 2009 ; 32 : 55 – 64 . Google Scholar Crossref Search ADS PubMed WorldCat 2. Baglioni C , Battagliese G , Feige B et al. Insomnia as a predictor of depression: a meta-analytic evaluation of longitudinal epidemiological studies . J Affect Disord . 2011 ; 135 : 10 – 9 . Google Scholar Crossref Search ADS PubMed WorldCat 3. Ford DE , Kamerow DB . Epidemiologic study of sleep disturbances and psychiatric disorders. An opportunity for prevention? JAMA . 1989 ; 262 : 1479 – 1484 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Hauri P , Chernik D , Hawkins D , Mendels J . Sleep of depressed patients in remission . Arch Gen Psychiatry . 1974 ; 31 : 386 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Carney CE , Segal ZV , Edinger JD , Krystal AD . A comparison of rates of residual insomnia symptoms following pharmacotherapy or cognitive-behavioral therapy for major depressive disorder . J Clin Psychiatry . 2007 ; 68 : 254 – 60 . Google Scholar Crossref Search ADS PubMed WorldCat 6. Kennedy GJ , Kelman HR , Thomas C . Persistence and remission of depressive symptoms in late life . Am J Psychiatry . 1991 ; 148 : 174 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Pigeon WR , Hegel M , Unutzer J et al. Is insomnia a perpetuating factor for late-life depression in the IMPACT cohort? Sleep . 2008 ; 31 : 481 – 8 . Google Scholar Crossref Search ADS PubMed WorldCat 8. Franzen PL , Buysse DJ . Sleep disturbances and depression: risk relationships for subsequent depression and therapeutic implications . Dialogues Clin Neurosci . 2008 ; 10 : 473 – 81 . Google Scholar PubMed WorldCat 9. Dombrovski AY , Cyranowski JM , Mulsant BH et al. Which symptoms predict recurrence of depression in women treated with maintenance interpersonal psychotherapy? Depress Anxiety . 2008 ; 25 : 1060 – 6 . Google Scholar Crossref Search ADS PubMed WorldCat 10. American Psychiatric Association . Diagnostic and Statistical Manual of Mental Disorders (DSM-5®) . American Psychiatric Pub ; 2013 . WorldCat COPAC 11. Medicine AAoS . International Classification of Sleep Disorders—Third Edition (ICSD-3) . AASM Resource Library ; 2014 . WorldCat COPAC 12. Riemann D , Perlis ML . The treatments of chronic insomnia: a review of benzodiazepine receptor agonists and psychological and behavioral therapies . Sleep Med Rev . 2009 ; 13 : 205 – 14 . Google Scholar Crossref Search ADS PubMed WorldCat 13. McHugh RK , Whitton SW , Peckham AD , Welge JA , Otto MW . Patient preference for psychological vs pharmacologic treatment of psychiatric disorders: a meta-analytic review . J Clin Psychiatry . 2013 ; 74 : 1478 – 602 . Google Scholar Crossref Search ADS WorldCat 14. Kwan BM , Dimidjian S , Rizvi SL . Treatment preference, engagement, and clinical improvement in pharmacotherapy versus psychotherapy for depression . Behav Res Ther . 2010 ; 48 : 799 – 804 . Google Scholar Crossref Search ADS PubMed WorldCat 15. Ye YY , Chen NK , Chen J et al. Internet-based cognitive-behavioural therapy for insomnia (ICBT-i): a meta-analysis of randomised controlled trials . BMJ Open . 2016 ; 6 : e010707 . Google Scholar Crossref Search ADS PubMed WorldCat 16. Zachariae R , Lyby MS , Ritterband LM , O’Toole MS . Efficacy of internet-delivered cognitive-behavioral therapy for insomnia–a systematic review and meta-analysis of randomized controlled trials . Sleep Med Rev . 2016 ; 30 : 1 – 10 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Andrews G , Cuijpers P , Craske MG , McEvoy P , Titov N . Computer therapy for the anxiety and depressive disorders is effective, acceptable and practical health care: a meta-analysis . PLoS One . 2010 ; 5 : e13196 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Blom K , Jernelöv S , Kraepelien M et al. Internet treatment addressing either insomnia or depression, for patients with both diagnoses: a randomized trial . Sleep . 2015 ; 38 : 267 – 77 . Google Scholar Crossref Search ADS PubMed WorldCat 19. Sheehan DV , Lecrubier Y , Sheehan KH et al. The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10 . J Clin Psychiatry . 1998 ; 59 ( Suppl 20 ): 22 – 33 ; quiz 4–57. Google Scholar PubMed WorldCat 20. Svanborg P , Åsberg M . A new self-rating scale for depression and anxiety states based on the Comprehensive Psychopathological Rating Scale . Acta Psychiatr Scand . 1994 ; 89 : 21 – 8 . Google Scholar Crossref Search ADS PubMed WorldCat 21. Montgomery SA , Asberg M . A new depression scale designed to be sensitive to change . Br J Psychiatry . 1979 ; 134 : 382 – 9 . Google Scholar Crossref Search ADS PubMed WorldCat 22. Svanborg P , Ekselius L . Self-assessment of DSM-IV criteria for major depression in psychiatric out- and inpatients . Nord J Psychiatry . 2003 ; 57 : 291 – 6 . Google Scholar Crossref Search ADS PubMed WorldCat 23. Bastien CH , Vallieres A , Morin CM . Validation of the Insomnia Severity Index as an outcome measure for insomnia research . Sleep Med . 2001 ; 2 : 297 – 307 . Google Scholar Crossref Search ADS PubMed WorldCat 24. Morin CM , Belleville G , Belanger L , Ivers H . The Insomnia Severity Index: psychometric indicators to detect insomnia cases and evaluate treatment response . Sleep . 2011 ; 34 : 601 – 8 . Google Scholar Crossref Search ADS PubMed WorldCat 25. Morin CM , Espie CA. Insomnia: A Clinician’s Guide to Assessment and Treatment . Springer ; 2003 . Google Preview WorldCat COPAC 26. Thorndike FP , Ritterband LM , Saylor DK , Magee JC , Gonder-Frederick LA , Morin CM . Validation of the insomnia severity index as a web-based measure . Behav Sleep Med . 2011 ; 9 : 216 – 23 . Google Scholar Crossref Search ADS PubMed WorldCat 27. Chen PY , Yang CM , Morin CM . Validating the cross-cultural factor structure and invariance property of the Insomnia Severity Index: evidence based on ordinal EFA and CFA . Sleep Med . 2015 ; 16 : 598 – 603 . Google Scholar Crossref Search ADS PubMed WorldCat 28. Andersson G , Bergström J , Holländare F , Carlbring P , Kaldo V , Ekselius L . Internet-based self-help for depression: randomised controlled trial . Br J Psychiatry . 2005 ; 187 : 456 – 61 . Google Scholar Crossref Search ADS PubMed WorldCat 29. Hedman E , Ljótsson B , Kaldo V et al. Effectiveness of Internet-based cognitive behaviour therapy for depression in routine psychiatric care . J Affect Disord . 2013 . WorldCat 30. Kaldo V , Jernelöv S , Blom K et al. Guided internet cognitive behavioral therapy for insomnia compared to a control treatment — A randomized trial . Behav Res Ther . 2015 ; 71 : 90 – 100 . Google Scholar Crossref Search ADS PubMed WorldCat 31. Blom K , Tarkian Tillgren H , Wiklund T et al. Internet- vs. group-delivered cognitive behavior therapy for insomnia: a randomized controlled non-inferiority trial . Behav Res Ther . 2015 ; 70 : 47 – 55 . Google Scholar Crossref Search ADS PubMed WorldCat 32. Blom K , Jernelöv S , Rück C , Lindefors N , Kaldo V . Three-year follow-up of insomnia and hypnotics after controlled internet treatment for insomnia . Sleep . 2016 ; 39 : 1267 – 74 . Google Scholar Crossref Search ADS PubMed WorldCat 33. Hedman E , Ljótsson B , Blom K et al. Telephone versus internet administration of self-report measures of social anxiety, depressive symptoms, and insomnia: psychometric evaluation of a method to reduce the impact of missing data . J Med Internet Res . 2013 ; 15 : e229 . Google Scholar Crossref Search ADS PubMed WorldCat 34. Hesser H . Modeling individual differences in randomized experiments using growth models: recommendations for design, statistical analysis and reporting of results of internet interventions . Internet Interventions . 2015 ; 2 : 110 – 20 . Google Scholar Crossref Search ADS WorldCat 35. Gueorguieva R , Krystal JH . Move over ANOVA: progress in analyzing repeated-measures data and its reflection in papers published in the Archives of General Psychiatry . Arch Gen Psychiatry . 2004 ; 61 : 310 – 7 . Google Scholar Crossref Search ADS PubMed WorldCat 36. Manber R , Buysse DJ , Edinger J et al. Efficacy of cognitive-behavioral therapy for insomnia combined with antidepressant pharmacotherapy in patients with comorbid depression and insomnia: a randomized controlled trial . J Clin Psychiatry . 2016 ; 77 : e1316 – e23 . Google Scholar Crossref Search ADS PubMed WorldCat 37. Kaldo V , Ramnerö J , Jernelöv S . Involving clients in treatment methods: a neglected interaction in the therapeutic relationship . J Consult Clin Psychol . 2015 ; 83 : 1136 – 41 . Google Scholar Crossref Search ADS PubMed WorldCat 38. Harvey L , Inglis SJ , Espie CA . Insomniacs’ reported use of CBT components and relationship to long-term clinical outcome . Behav Res Ther . 2002 ; 40 : 75 – 83 . Google Scholar Crossref Search ADS PubMed WorldCat 39. Cuijpers P , Andersson G , Donker T , van Straten A . Psychological treatment of depression: results of a series of meta-analyses . Nord J Psychiatry . 2011 ; 65 : 354 – 64 . Google Scholar Crossref Search ADS PubMed WorldCat 40. Riemann D . Does effective management of sleep disorders reduce depressive symptoms and the risk of depression? Drugs . 2009 ; 69 ( Suppl 2 ): 43 – 64 . Google Scholar Crossref Search ADS PubMed WorldCat 41. Lancee J , van den Bout J , van Straten A , Spoormaker VI . Baseline depression levels do not affect efficacy of cognitive-behavioral self-help treatment for insomnia . Depress Anxiety . 2013 ; 30 : 149 – 156 . Google Scholar Crossref Search ADS PubMed WorldCat 42. Christensen H , Batterham PJ , Gosling JA et al. Effectiveness of an online insomnia program (SHUTi) for prevention of depressive episodes (the GoodNight Study): a randomised controlled trial . The Lancet Psychiatry . 2016 ; 3 : 333 – 41 . Google Scholar Crossref Search ADS PubMed WorldCat 43. Blom K , Jernelöv S , Lindefors N , Kaldo V . Facilitating and hindering factors in Internet-delivered treatment for insomnia and depression . Internet Interventions . 2016 ; 4 : 51 – 60 . Google Scholar Crossref Search ADS WorldCat 44. Bathgate CJ , Edinger JD , Krystal AD . Insomnia patients with objective short sleep duration have a blunted response to cognitive behavioral therapy for insomnia . Sleep . 2017 ; 40 : zsw012 – zsw . WorldCat Author notes Address correspondence to: Kerstin Blom, LP, PhD, Centre for Psychiatry Research, Department of Clinical Neuroscience, Karolinska Institutet, M46, Huddinge hospital, SE-141 86, Stockholm, Sweden. Telephone: 46 70 1655228; Email: [email protected] © Sleep Research Society 2017. Published by Oxford University Press on behalf of the Sleep Research Society. All rights reserved. For permissions, please e-mail [email protected].

Spira, Adam P; An, Yang; Peng, Yu; Wu, Mark N; Simonsick, Eleanor M; Ferrucci, Luigi; Resnick, Susan M

doi: 10.1093/sleep/zsx076pmid: 28482100

Abstract Study Objectives The apolipoprotein E (APOE) Ɛ4 allele increases Alzheimer’s disease (AD) risk and has been linked to a greater risk of sleep-disordered breathing. We investigated the association of APOE genotype with nonrespiratory sleep parameters. Methods We studied 1264 cognitively normal participants in the Baltimore Longitudinal Study of Aging (mean = 57.5 ± 16.1 years, range 19.9–92.0, 48.2% women, 19.8% African American) with APOE genotyping and self-reported sleep duration (≥9, 7 or 8, ≤6 hours), difficulty falling/staying asleep, and napping. We compared Ɛ4 carriers with all noncarriers and compared persons at reduced (Ɛ2/Ɛ2 or Ɛ2/Ɛ3) or elevated AD risk (≥1 Ɛ4 allele) with those neutral for AD risk (Ɛ3/Ɛ3). Results In fully adjusted models, those with ≥1 Ɛ4 allele had a greater odds of being in a shorter sleep duration category compared to all noncarriers (odds ratio [OR] = 1.41, 95% confidence interval [CI] 1.06, 1.88) and Ɛ3/Ɛ3 carriers (OR = 1.43, 95% CI 1.06, 1.92). Compared to Ɛ3/Ɛ3 carriers, Ɛ2/Ɛ2 or Ɛ2/Ɛ3 carriers had a lower odds of reporting napping (OR = 0.64, 95% CI 0.43, 0.96). Among participants aged ≥50 years, sleep duration findings remained and Ɛ4 carriers had a greater odds of trouble falling/staying asleep than noncarriers (OR = 1.49, 95% CI 1.02, 2.17). We found some evidence for stronger associations of Ɛ4 with sleep duration among African Americans. Conclusions Self-reported sleep duration, napping, and trouble falling/staying asleep differ by APOE genotype. Studies are needed to examine whether APOE promotes AD by degrading sleep and to clarify the role of race in these associations. older adults, APOE, genotype, sleep, napping Statement of Significance Multiple studies suggest an association between the apolipoprotein E (APOE) Ɛ4 allele and risk of sleep-disordered breathing, but little is known regarding links between APOE genotype and nonrespiratory sleep parameters. Given accumulating evidence for a role of disturbed sleep in the development of Alzheimer’s disease (AD) and the increased risk of AD among Ɛ4 carriers, studies of APOE and nonrespiratory sleep parameters are needed. To our knowledge, this is the first study in a cognitively intact sample to link the Ɛ4 allele to shorter self-reported sleep duration and difficulty falling/staying asleep and to show reduced napping among those with Ɛ2/Ɛ2 or Ɛ2/Ɛ3 genotypes (protective against AD). Given the need to better understand whether sleep alters AD risk, these findings warrant further examination. INTRODUCTION Sleep disturbance has recently gained attention as a risk factor for Alzheimer’s disease (AD). Studies in mice and Drosophila indicate that sleep deprivation promotes β-amyloid (Aβ) deposition in the brain.1,2 Further, initial cross-sectional studies in humans demonstrate associations of shorter self-reported sleep duration and poorer sleep quality with greater brain Aβ, measured by positron emission tomography scans.3,4 Lower sleep efficiency and greater sleep fragmentation, measured by wrist actigraphy, also have been linked to greater brain Aβ burden, measured in cerebrospinal fluid.5 Apolipoprotein E (APOE) genotype is associated with risk for late onset AD. The APOE Ɛ4 allele has been linked to a 2- to 3-fold increased risk of AD in heterozygotes (ie, Ɛ2/Ɛ4 or Ɛ3/Ɛ4 genotypes) and 10 times the risk in Ɛ4/Ɛ4 homozygotes, compared to carriers of the most common APOE genotype, Ɛ3/Ɛ3.6 Conversely, the Ɛ2 allele appears to confer some protection against AD.7 Approximately 20%–30% of the population has one or more Ɛ4 alleles, placing a substantial proportion of the population at elevated AD risk.8 Importantly, the APOE Ɛ4 allele has also been linked to sleep disturbances. APOE Ɛ4 carriers have been shown to have an increased risk of sleep-disordered breathing (SDB),9,10 and studies have found that APOE Ɛ4 and SDB interact, such that SDB severity is more strongly associated with cognitive impairment among Ɛ4 carriers than in noncarriers.11–13 Studies have also investigated links between APOE and nonrespiratory sleep parameters but primarily in persons with cognitive impairment.14–18 Surprisingly little has been published, however, regarding associations between the Ɛ4 allele and nonrespiratory sleep parameters, such as sleep duration, difficulty falling or staying asleep, and napping, in cognitively intact samples. In a study of 122 participants in the Baltimore Longitudinal Study of Aging (BLSA), we reported that 50% of those with sleep duration <7 hours had at least one Ɛ4 allele, compared to only 21% of those reporting 7 hours sleep, and 24% of those reporting >7 hours (Table 2 of prior paper).19 Here, we follow-up our initial observation in a sample undergoing neuroimaging assessments and conduct an in-depth study of the association of APOE genotype with self-reported sleep duration, difficulty falling or remaining asleep, and napping in a much larger sample of cognitively unimpaired adults in the BLSA. We hypothesized that those with one or more Ɛ4 alleles would have shorter sleep duration and greater difficulty falling or staying asleep than those without an Ɛ4 allele. Because the Ɛ2 allele is associated with a decreased risk of AD and better sleep may protect against AD, we also investigated whether those with Ɛ2/Ɛ2 or Ɛ2/Ɛ3 genotypes reported longer sleep duration, better quality sleep, and distinct napping habits. METHODS Participants We studied individuals who participated in the BLSA between 1991 and 2000 (the years when the sleep measures of interest were administered). The BLSA is a cohort study of aging that began in 1958 and is ongoing.20 BLSA participants are exceptionally healthy on enrollment; eligibility criteria require them to be free of major diseases with the exception of controlled hypertension and to have no cognitive impairment, physical disability, mobility limitations, or conditions that impair functioning or limit life expectancy. Further, they cannot be taking ongoing antibiotics, chronic pain medication, immunosuppressants, corticosteroids, or histamine H2 blockers. Of 1742 participants, we excluded 369 who were missing APOE genotype data and two who were missing self-report sleep data (see below). Further, race/ethnic differences exist in links between Ɛ4 genotype and AD risk, with weaker associations between Ɛ4 and AD in African Americans than in whites.21 The substantial majority of BLSA participants are either white or African American, affording little statistical power to investigate the associations under study in participants who were neither white nor African American. Accordingly, we excluded 64 participants who reported another race/ethnicity. Finally, we excluded 43 participants who were cognitively impaired at the time of sleep assessment (see cognitive assessment details below), leaving a total analytic sample of 1264. APOE Genotype Participants underwent blood draws, and APOE genotype was ascertained using standard procedures. Determinations of APOE genotypes were performed during two different epochs over the course of the BLSA. The earlier assays followed the approach of Hixon and Vernier22 and were based on polymerase chain reaction (PCR) amplification with HhaI restriction isotyping to determine APOE genotype. Sequences that encompass amino acid positions 112 and 158 are amplified by PCR, and an HhaI restriction endonuclease is used to cut the DNA. This results in unique permutations of sizes of HhaI which correspond to specific APOE genotypes.22 More recent assays used the TaqMan method which is a PCR-based system using oligonucleotide probes specific for particular alleles that have been labeled using fluorogenic reporter dyes.23 Sleep Variables Our primary sleep measures were self-report indices of sleep duration, sleep quality, and napping habits. As part of the BLSA interview between 1991 and 2000, participants were asked to report the average number of hours of sleep they obtained at night. Responses were recorded as integers and were used as our measure of sleep duration. Participants were also asked whether they “often have trouble falling asleep at night or awaken in the middle of the night and can’t go back to sleep,” (yes or no). They were also asked if they nap, with response options “rarely or never,” “3–5 times/week,” “1–2 times/week,” or “daily”; we recoded responses of “rarely or never” as 0 and others as 1. In addition, participants were asked about common symptoms of SDB, which may confound the associations of interest (see Statistical Analyses). Specifically, they reported whether they “snore often and loudly” (yes or no) and if they “often become drowsy or fall asleep during the daytime when you wish to be awake? (e.g. falling asleep watching TV or reading)” (yes or no). Other Measures At each BLSA visit, participants’ height and weight were measured and body mass index (BMI) was calculated in kg/m2. Blood pressure was measured by a sphygmomanometer, and fasting blood glucose and cholesterol were measured. Participants also completed a neuropsychological test battery at routine visits, and those screening positive for cognitive impairment (ie, ≥4 errors on the Blessed Information Memory Concentration Test24 or Clinical Dementia Rating Scale25 scores ≥0.5) had their test data evaluated at a consensus diagnostic conference, where cognitive status is determined to be normal, mild cognitive impairment (MCI) according to the Petersen criteria,26 or dementia according to Diagnostic and Statistical Manual of Mental Disorders, Third Edition-Revised criteria.27 Those who screened negative for potential impairment or screened positive but were determined to be cognitively normal through a consensus conference were considered cognitively intact.28 Depressive symptomatology was measured by the Center for Epidemiologic Studies-Depression Scale (CES-D).29 A vascular burden score was calculated by summing the number of the following conditions participants reported: smoking status, hypercholesterolemia; hypertension; diabetes mellitus; myocardial infarction; angina; and stroke or transient ischemic attack. In addition to self-report, diabetes contributed to the vascular burden score if fasting blood glucose was ≥126 mg dl–1, hypertension contributed if systolic blood pressure was ≥140 mm Hg or diastolic blood pressure was ≥90 mm Hg, and hypercholesterolemia contributed if recorded cholesterol was ≥200 mg dl–1. Statistical Analyses We fit two sets of ordinal logistic regression models to determine the association between APOE genotype, our primary predictor, and reported sleep duration, which we categorized as ≥9 hours, 7 or 8 hours, and ≤6 hours. Ordinal logistic regression produces odds ratios (ORs), which represent the odds of being in a shorter sleep duration category. The proportional odds assumption was met for these sleep duration analyses. Our selection of sleep duration categories was based on the National Sleep Foundation’s recent recommendation of 7–8 hours sleep duration for adults aged 65 and older.30 We used these cutpoints rather than the recent American Academy of Sleep Medicine and Sleep Research Society recommendations because the latter are restricted to persons aged 18–60 years31 and would therefore be inappropriate for much of our sample. Next, we fit two sets of logistic regression models to determine the association of APOE genotype with reports of problems falling or staying asleep and reports of napping. For all analyses, we first compared Ɛ4 allele carriers (Ɛ2/Ɛ4, Ɛ3/Ɛ4, or Ɛ4/Ɛ4 genotypes), who are at elevated risk for AD, with noncarriers. Next, we compared participants at lower AD risk (Ɛ2/Ɛ2 or Ɛ2/Ɛ3 genotype) or greater AD risk (Ɛ4 carriers) with those neutral for AD risk (Ɛ3/Ɛ3 genotype). Minimally adjusted models (Model 1) were adjusted for age, age squared, race (African American or white), and sex. Fully adjusted models (Model 2) were further adjusted for years of education, BMI, CES-D scores, vascular burden score, and—because APOE genotype has been linked to SDB which often manifests in snoring and daytime sleepiness—we also adjusted for responses to questions about snoring and excessive daytime sleepiness. To examine whether this association was stronger in older adults, we repeated analyses in participants aged ≥50 years. Importantly, race/ethnic differences appear to exist in the association between Ɛ4 genotype and AD risk, with a weaker link among African Americans than in whites.21,32 Thus, we explored potential interactions of APOE Ɛ4 status (≥1 vs. 0 Ɛ4 alleles) with race on sleep variables by adding an interaction term (APOE Ɛ4 status × race) to fully adjusted models. All analyses were conducted using SAS 9.4 (Cary, North Carolina). RESULTS On average, participants were 57.5 ± 16.1 years of age (range 20–92) (Table 1). Approximately 48% were women and 20% were African American. They had a mean educational attainment of 16.7 ± 2.6 years (range 8–21), mean BMI of 26.1 ± 4.3 kg/m2, and mean CES-D scores of 6.6 ± 6.7. Participants’ mean sleep duration was 7.1 ± 1.0 hours (range 4–10); 321 (25.6%) reported <7 hours sleep, 874 (69.6%) reported 7–8 hours sleep, and 61 (4.9%) reported >8 hours sleep. Almost 25% reported snoring often or loudly and 31% reported often feeling drowsy or unintentionally falling asleep during the day. Overall, 352 participants (27.8%) were Ɛ4 carriers (Ɛ2/Ɛ4, Ɛ3/Ɛ4, or Ɛ4/Ɛ4 genotype) and 912 (72.2%) were noncarriers (Table 2). Of the 912 noncarriers, 750 participants (82.2%) had the Ɛ3/Ɛ3 genotype and 162 (17.8%) had either the Ɛ2/Ɛ2 or Ɛ2/Ɛ3 genotype. Table 1 Participant Characteristics (Mean ± Standard Deviation or n (%)) by Sleep Duration (N = 1264 Unless Otherwise Noted). Characteristic . Full sample . Ɛ2/Ɛ2, Ɛ2/Ɛ3 . Ɛ3/Ɛ3 . Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 . p-value* . Age, years 57.5 ± 16.1 59.4 ± 15.2 57.5 ± 16.5 56.3 ± 15.5 .133 range 19.9–92.0 range 27.0–89.2 range 20.9–92.0 range 19.9–88.2 Women 609 (48.2) 68 (42.0) 373 (50.0) 168 (48.4) .207 African American 250 (19.8) 39 (24.1) 119 (15.7) 92 (26.5) <.0001 Education, years 16.7 ± 2.6 16.7 ± 2.7 16.8 ± 2.5 16.6 ± 2.6 .669 range 8.0–21.0 range 8.0–21.0 range 8.0–21.0 range 8.0–21.0 Body mass index (kg/m2) 26.1 ± 4.3 26.1 ± 4.2 25.9 ± 4.2 26.4 ± 4.6 .204 range 17.0–46.4 range 18.8–42.1 range 17.6–46.4 range 17.0–45.5 CES-D (n = 1225) 6.6 ± 6.7 6.9 ± 6.9 6.6 ± 6.8 6.5 ± 6.4 .814 range 0–50.0 range 0–42 range 0–50.0 range 0–37.0 Vascular burden score 1.0 (1.0) 1.2 (1.1) 1.0 (1.0) 1.0 (0.9) .245 range 0–6 range 0–6 range 0–6 range 0–4 Snores often/ loudly (n = 1225) 303 (24.7) 48 (31.0) 179 (24.5) 76 (22.4) .120 Daytime sleepiness (n = 1,243) 387 (31.1) 46 (28.8) 237 (31.7) 104 (30.4) .732 Sleep duration (n = 1256) .007 ≤6 hours 321 (25.6) 44 (27.2) 168 (22.5) 109 (31.4) 7 or 8 hours 874 (69.6) 106 (65.4) 541 (72.4) 227 (65.4) ≥9 hours 61 (4.9) 12 (7.4) 38 (5.1) 11 (3.2) Trouble falling asleep/ returning to sleep (n = 1249) 299 (23.9) 36 (22.2) 176 (23.8) 87 (25.0) .785 Takes naps (n = 1257) 512 (40.7) 55 (34.2) 325 (43.4) 132 (37.9) .043 Characteristic . Full sample . Ɛ2/Ɛ2, Ɛ2/Ɛ3 . Ɛ3/Ɛ3 . Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 . p-value* . Age, years 57.5 ± 16.1 59.4 ± 15.2 57.5 ± 16.5 56.3 ± 15.5 .133 range 19.9–92.0 range 27.0–89.2 range 20.9–92.0 range 19.9–88.2 Women 609 (48.2) 68 (42.0) 373 (50.0) 168 (48.4) .207 African American 250 (19.8) 39 (24.1) 119 (15.7) 92 (26.5) <.0001 Education, years 16.7 ± 2.6 16.7 ± 2.7 16.8 ± 2.5 16.6 ± 2.6 .669 range 8.0–21.0 range 8.0–21.0 range 8.0–21.0 range 8.0–21.0 Body mass index (kg/m2) 26.1 ± 4.3 26.1 ± 4.2 25.9 ± 4.2 26.4 ± 4.6 .204 range 17.0–46.4 range 18.8–42.1 range 17.6–46.4 range 17.0–45.5 CES-D (n = 1225) 6.6 ± 6.7 6.9 ± 6.9 6.6 ± 6.8 6.5 ± 6.4 .814 range 0–50.0 range 0–42 range 0–50.0 range 0–37.0 Vascular burden score 1.0 (1.0) 1.2 (1.1) 1.0 (1.0) 1.0 (0.9) .245 range 0–6 range 0–6 range 0–6 range 0–4 Snores often/ loudly (n = 1225) 303 (24.7) 48 (31.0) 179 (24.5) 76 (22.4) .120 Daytime sleepiness (n = 1,243) 387 (31.1) 46 (28.8) 237 (31.7) 104 (30.4) .732 Sleep duration (n = 1256) .007 ≤6 hours 321 (25.6) 44 (27.2) 168 (22.5) 109 (31.4) 7 or 8 hours 874 (69.6) 106 (65.4) 541 (72.4) 227 (65.4) ≥9 hours 61 (4.9) 12 (7.4) 38 (5.1) 11 (3.2) Trouble falling asleep/ returning to sleep (n = 1249) 299 (23.9) 36 (22.2) 176 (23.8) 87 (25.0) .785 Takes naps (n = 1257) 512 (40.7) 55 (34.2) 325 (43.4) 132 (37.9) .043 *p-values from chi-square tests for categorical variables and analysis of variance F-tests for continuous variables. CES-D = Center for Epidemiologic Studies-Depression Scale. Open in new tab Table 1 Participant Characteristics (Mean ± Standard Deviation or n (%)) by Sleep Duration (N = 1264 Unless Otherwise Noted). Characteristic . Full sample . Ɛ2/Ɛ2, Ɛ2/Ɛ3 . Ɛ3/Ɛ3 . Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 . p-value* . Age, years 57.5 ± 16.1 59.4 ± 15.2 57.5 ± 16.5 56.3 ± 15.5 .133 range 19.9–92.0 range 27.0–89.2 range 20.9–92.0 range 19.9–88.2 Women 609 (48.2) 68 (42.0) 373 (50.0) 168 (48.4) .207 African American 250 (19.8) 39 (24.1) 119 (15.7) 92 (26.5) <.0001 Education, years 16.7 ± 2.6 16.7 ± 2.7 16.8 ± 2.5 16.6 ± 2.6 .669 range 8.0–21.0 range 8.0–21.0 range 8.0–21.0 range 8.0–21.0 Body mass index (kg/m2) 26.1 ± 4.3 26.1 ± 4.2 25.9 ± 4.2 26.4 ± 4.6 .204 range 17.0–46.4 range 18.8–42.1 range 17.6–46.4 range 17.0–45.5 CES-D (n = 1225) 6.6 ± 6.7 6.9 ± 6.9 6.6 ± 6.8 6.5 ± 6.4 .814 range 0–50.0 range 0–42 range 0–50.0 range 0–37.0 Vascular burden score 1.0 (1.0) 1.2 (1.1) 1.0 (1.0) 1.0 (0.9) .245 range 0–6 range 0–6 range 0–6 range 0–4 Snores often/ loudly (n = 1225) 303 (24.7) 48 (31.0) 179 (24.5) 76 (22.4) .120 Daytime sleepiness (n = 1,243) 387 (31.1) 46 (28.8) 237 (31.7) 104 (30.4) .732 Sleep duration (n = 1256) .007 ≤6 hours 321 (25.6) 44 (27.2) 168 (22.5) 109 (31.4) 7 or 8 hours 874 (69.6) 106 (65.4) 541 (72.4) 227 (65.4) ≥9 hours 61 (4.9) 12 (7.4) 38 (5.1) 11 (3.2) Trouble falling asleep/ returning to sleep (n = 1249) 299 (23.9) 36 (22.2) 176 (23.8) 87 (25.0) .785 Takes naps (n = 1257) 512 (40.7) 55 (34.2) 325 (43.4) 132 (37.9) .043 Characteristic . Full sample . Ɛ2/Ɛ2, Ɛ2/Ɛ3 . Ɛ3/Ɛ3 . Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 . p-value* . Age, years 57.5 ± 16.1 59.4 ± 15.2 57.5 ± 16.5 56.3 ± 15.5 .133 range 19.9–92.0 range 27.0–89.2 range 20.9–92.0 range 19.9–88.2 Women 609 (48.2) 68 (42.0) 373 (50.0) 168 (48.4) .207 African American 250 (19.8) 39 (24.1) 119 (15.7) 92 (26.5) <.0001 Education, years 16.7 ± 2.6 16.7 ± 2.7 16.8 ± 2.5 16.6 ± 2.6 .669 range 8.0–21.0 range 8.0–21.0 range 8.0–21.0 range 8.0–21.0 Body mass index (kg/m2) 26.1 ± 4.3 26.1 ± 4.2 25.9 ± 4.2 26.4 ± 4.6 .204 range 17.0–46.4 range 18.8–42.1 range 17.6–46.4 range 17.0–45.5 CES-D (n = 1225) 6.6 ± 6.7 6.9 ± 6.9 6.6 ± 6.8 6.5 ± 6.4 .814 range 0–50.0 range 0–42 range 0–50.0 range 0–37.0 Vascular burden score 1.0 (1.0) 1.2 (1.1) 1.0 (1.0) 1.0 (0.9) .245 range 0–6 range 0–6 range 0–6 range 0–4 Snores often/ loudly (n = 1225) 303 (24.7) 48 (31.0) 179 (24.5) 76 (22.4) .120 Daytime sleepiness (n = 1,243) 387 (31.1) 46 (28.8) 237 (31.7) 104 (30.4) .732 Sleep duration (n = 1256) .007 ≤6 hours 321 (25.6) 44 (27.2) 168 (22.5) 109 (31.4) 7 or 8 hours 874 (69.6) 106 (65.4) 541 (72.4) 227 (65.4) ≥9 hours 61 (4.9) 12 (7.4) 38 (5.1) 11 (3.2) Trouble falling asleep/ returning to sleep (n = 1249) 299 (23.9) 36 (22.2) 176 (23.8) 87 (25.0) .785 Takes naps (n = 1257) 512 (40.7) 55 (34.2) 325 (43.4) 132 (37.9) .043 *p-values from chi-square tests for categorical variables and analysis of variance F-tests for continuous variables. CES-D = Center for Epidemiologic Studies-Depression Scale. Open in new tab Table 2 Distribution of APOE Genotype, n (%). Ɛ2/Ɛ2, Ɛ2/Ɛ3 . Ɛ3/Ɛ3 . Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 . 162 (12.8) 750 (59.3) 352 (27.8) Ɛ2/Ɛ2, Ɛ2/Ɛ3 . Ɛ3/Ɛ3 . Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 . 162 (12.8) 750 (59.3) 352 (27.8) APOE = apolipoprotein E. Open in new tab Table 2 Distribution of APOE Genotype, n (%). Ɛ2/Ɛ2, Ɛ2/Ɛ3 . Ɛ3/Ɛ3 . Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 . 162 (12.8) 750 (59.3) 352 (27.8) Ɛ2/Ɛ2, Ɛ2/Ɛ3 . Ɛ3/Ɛ3 . Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 . 162 (12.8) 750 (59.3) 352 (27.8) APOE = apolipoprotein E. Open in new tab All Participants In the full sample, compared to participants with no APOE Ɛ4 alleles, those with ≥1 Ɛ4 allele had a 38% greater odds of being in a shorter sleep duration category (OR = 1.38, 95% confidence interval [CI] 1.05, 1.80); this association was similar in the fully adjusted model (OR = 1.41, 95% CI 1.06, 1.88) (Table 3). Similarly, in analyses that used the Ɛ3/Ɛ3 genotype as the reference group, those with ≥1 Ɛ4 allele had a 38% greater odds of being in a shorter sleep duration category in Model 1 (OR = 1.38, 95% CI 1.04, 1.82) and a 43% greater odds in Model 2 (OR = 1.43, 95% CI 1.06, 1.92). Table 3 Association Between APOE Genotype and Shorter Sleep Duration*. . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Shorter sleep duration n = 1256 n = 1166 ≥1 Ɛ4 alleles 1.38 (1.05, 1.80), p = .021 1.41 (1.06, 1.88), p = .018 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.99 (0.68, 1.45), p = .969 1.05 (0.70, 1.57), p = .812 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.38 (1.04, 1.82), p = .026 1.43 (1.06, 1.92), p = .019 . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Shorter sleep duration n = 1256 n = 1166 ≥1 Ɛ4 alleles 1.38 (1.05, 1.80), p = .021 1.41 (1.06, 1.88), p = .018 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.99 (0.68, 1.45), p = .969 1.05 (0.70, 1.57), p = .812 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.38 (1.04, 1.82), p = .026 1.43 (1.06, 1.92), p = .019 *Results of ordinal logistic regression analyses with sleep duration as the outcome (categorized as ≥9 hours, 7 or 8 hours, ≤6 hours), reflecting odds of being in a shorter sleep duration category. Model 1 adjusted for age, age2, sex, race; Model 2 adjusted for age, age2, sex, race, education, BMI, CES-D, vascular burden score, snoring, and daytime sleepiness. Bold values indicate p <.05. APOE = apolipoprotein; BMI = body mass index; CES-D = Center for Epidemiologic Studies-Depression Scale; CI = confidence interval; OR = odds ratio. Open in new tab Table 3 Association Between APOE Genotype and Shorter Sleep Duration*. . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Shorter sleep duration n = 1256 n = 1166 ≥1 Ɛ4 alleles 1.38 (1.05, 1.80), p = .021 1.41 (1.06, 1.88), p = .018 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.99 (0.68, 1.45), p = .969 1.05 (0.70, 1.57), p = .812 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.38 (1.04, 1.82), p = .026 1.43 (1.06, 1.92), p = .019 . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Shorter sleep duration n = 1256 n = 1166 ≥1 Ɛ4 alleles 1.38 (1.05, 1.80), p = .021 1.41 (1.06, 1.88), p = .018 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.99 (0.68, 1.45), p = .969 1.05 (0.70, 1.57), p = .812 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.38 (1.04, 1.82), p = .026 1.43 (1.06, 1.92), p = .019 *Results of ordinal logistic regression analyses with sleep duration as the outcome (categorized as ≥9 hours, 7 or 8 hours, ≤6 hours), reflecting odds of being in a shorter sleep duration category. Model 1 adjusted for age, age2, sex, race; Model 2 adjusted for age, age2, sex, race, education, BMI, CES-D, vascular burden score, snoring, and daytime sleepiness. Bold values indicate p <.05. APOE = apolipoprotein; BMI = body mass index; CES-D = Center for Epidemiologic Studies-Depression Scale; CI = confidence interval; OR = odds ratio. Open in new tab However, there was no statistically significant difference in the odds of reporting trouble falling or staying asleep between those with ≥1 Ɛ4 allele and other participants in the full sample, whether the reference group consisted of all other participants or only those with the Ɛ3/Ɛ3 genotype (Table 4). Similarly, there was no difference in this domain between those with Ɛ2/Ɛ2 or Ɛ2/Ɛ3 genotypes and those with the Ɛ3/Ɛ3 genotype. Table 4 Association of APOE Genotype With Trouble Falling or Staying Asleep and Napping. . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Trouble falling or staying asleep n = 1249 n = 1160 ≥1 Ɛ4 alleles 1.15 (0.86, 1.55), p = .351 1.21 (0.87, 1.67), p = .258 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.91 (0.60, 1.39), p = .681 0.84 (0.53, 1.33), p = .451 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.13 (0.83, 1.54), p = .422 1.17 (0.84, 1.63), p = .363 Taking naps n = 1257 n = 1167 ≥1 Ɛ4 alleles 0.90 (0.69, 1.18), p = .450 0.90 (0.68, 1.19), p = .455 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.62 (0.42, 0.90), p = .012 0.64 (0.43, 0.96), p = .028 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 0.83 (0.63, 1.09), p = .178 0.83 (0.62, 1.10), p = .209 . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Trouble falling or staying asleep n = 1249 n = 1160 ≥1 Ɛ4 alleles 1.15 (0.86, 1.55), p = .351 1.21 (0.87, 1.67), p = .258 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.91 (0.60, 1.39), p = .681 0.84 (0.53, 1.33), p = .451 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.13 (0.83, 1.54), p = .422 1.17 (0.84, 1.63), p = .363 Taking naps n = 1257 n = 1167 ≥1 Ɛ4 alleles 0.90 (0.69, 1.18), p = .450 0.90 (0.68, 1.19), p = .455 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.62 (0.42, 0.90), p = .012 0.64 (0.43, 0.96), p = .028 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 0.83 (0.63, 1.09), p = .178 0.83 (0.62, 1.10), p = .209 Model 1 adjusted for age, age2, sex, race; Model 2 adjusted for age, age2, sex, race, education, BMI, CES-D, vascular burden score, snoring, and daytime sleepiness. Bold values indicate p <.05. APOE = apolipoprotein; BMI = body mass index; CES-D = Center for Epidemiologic Studies-Depression Scale; CI = confidence interval; OR = odds ratio. Open in new tab Table 4 Association of APOE Genotype With Trouble Falling or Staying Asleep and Napping. . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Trouble falling or staying asleep n = 1249 n = 1160 ≥1 Ɛ4 alleles 1.15 (0.86, 1.55), p = .351 1.21 (0.87, 1.67), p = .258 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.91 (0.60, 1.39), p = .681 0.84 (0.53, 1.33), p = .451 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.13 (0.83, 1.54), p = .422 1.17 (0.84, 1.63), p = .363 Taking naps n = 1257 n = 1167 ≥1 Ɛ4 alleles 0.90 (0.69, 1.18), p = .450 0.90 (0.68, 1.19), p = .455 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.62 (0.42, 0.90), p = .012 0.64 (0.43, 0.96), p = .028 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 0.83 (0.63, 1.09), p = .178 0.83 (0.62, 1.10), p = .209 . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Trouble falling or staying asleep n = 1249 n = 1160 ≥1 Ɛ4 alleles 1.15 (0.86, 1.55), p = .351 1.21 (0.87, 1.67), p = .258 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.91 (0.60, 1.39), p = .681 0.84 (0.53, 1.33), p = .451 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.13 (0.83, 1.54), p = .422 1.17 (0.84, 1.63), p = .363 Taking naps n = 1257 n = 1167 ≥1 Ɛ4 alleles 0.90 (0.69, 1.18), p = .450 0.90 (0.68, 1.19), p = .455 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.62 (0.42, 0.90), p = .012 0.64 (0.43, 0.96), p = .028 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 0.83 (0.63, 1.09), p = .178 0.83 (0.62, 1.10), p = .209 Model 1 adjusted for age, age2, sex, race; Model 2 adjusted for age, age2, sex, race, education, BMI, CES-D, vascular burden score, snoring, and daytime sleepiness. Bold values indicate p <.05. APOE = apolipoprotein; BMI = body mass index; CES-D = Center for Epidemiologic Studies-Depression Scale; CI = confidence interval; OR = odds ratio. Open in new tab With respect to napping, there was no difference when those with ≥1 Ɛ4 allele were compared to all other participants. Compared to participants with the Ɛ3/Ɛ3 genotype, those with either the Ɛ2/Ɛ2 or Ɛ2/Ɛ3 genotype were 38% less likely to report taking naps in minimally adjusted analyses (OR = 0.62, 95% CI 0.42, 0.90) and 36% less likely in fully adjusted analyses (OR = 0.64, 95% CI 0.43, 0.96) (Table 4). Participants Aged ≥50 Years We repeated analyses in the subset of 804 participants aged ≥50 years. They had a mean age of 67.1 ± 10.3 and 16.6 ± 2.7 years of education; 43.3% were women and 18.0% African American. In minimally adjusted analyses, compared to participants without any APOE Ɛ4 alleles, those with ≥1 Ɛ4 allele had a 49% greater odds of being in a shorter sleep duration category (OR = 1.49, 95% CI 1.07, 2.09); this association increased in the fully adjusted model (OR = 1.58, 95% 1.10, 1.25) (Table 5). The same pattern emerged when participants with ≥1 Ɛ4 allele were compared to those with the Ɛ3/Ɛ3 genotype, but there was no significant difference in sleep duration between Ɛ2/Ɛ2 or Ɛ2/Ɛ3 carriers and Ɛ3/Ɛ3 carriers. Table 5 Participants Aged ≥50 Only: Association Between APOE Genotype and Shorter Sleep Duration*. . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Shorter sleep duration n = 804 n = 742 ≥1 Ɛ4 alleles 1.49 (1.07, 2.09), p = .020 1.58 (1.10, 1.25), p = .013 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.98 (0.62, 1.56), p = .944 1.05 (0.64, 1.70), p = .855 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.49 (1.05, 2.11), p = .025 1.59 (1.10, 2.30), p = .014 . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Shorter sleep duration n = 804 n = 742 ≥1 Ɛ4 alleles 1.49 (1.07, 2.09), p = .020 1.58 (1.10, 1.25), p = .013 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.98 (0.62, 1.56), p = .944 1.05 (0.64, 1.70), p = .855 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.49 (1.05, 2.11), p = .025 1.59 (1.10, 2.30), p = .014 *Results of ordinal logistic regression analyses with sleep duration as the outcome (categorized as ≥9 hours, 7 or 8 hours, ≤6 hours), reflecting odds of being in a shorter sleep duration category. Model 1 adjusted for age, age2, sex, race; Model 2 adjusted for age, age2, sex, race, education, BMI, CES-D, vascular burden score, snoring, and daytime sleepiness. Bold values indicate p <.05. APOE = apolipoprotein; BMI = body mass index; CES-D = Center for Epidemiologic Studies-Depression Scale; CI = confidence interval; OR = odds ratio. Open in new tab Table 5 Participants Aged ≥50 Only: Association Between APOE Genotype and Shorter Sleep Duration*. . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Shorter sleep duration n = 804 n = 742 ≥1 Ɛ4 alleles 1.49 (1.07, 2.09), p = .020 1.58 (1.10, 1.25), p = .013 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.98 (0.62, 1.56), p = .944 1.05 (0.64, 1.70), p = .855 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.49 (1.05, 2.11), p = .025 1.59 (1.10, 2.30), p = .014 . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Shorter sleep duration n = 804 n = 742 ≥1 Ɛ4 alleles 1.49 (1.07, 2.09), p = .020 1.58 (1.10, 1.25), p = .013 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.98 (0.62, 1.56), p = .944 1.05 (0.64, 1.70), p = .855 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.49 (1.05, 2.11), p = .025 1.59 (1.10, 2.30), p = .014 *Results of ordinal logistic regression analyses with sleep duration as the outcome (categorized as ≥9 hours, 7 or 8 hours, ≤6 hours), reflecting odds of being in a shorter sleep duration category. Model 1 adjusted for age, age2, sex, race; Model 2 adjusted for age, age2, sex, race, education, BMI, CES-D, vascular burden score, snoring, and daytime sleepiness. Bold values indicate p <.05. APOE = apolipoprotein; BMI = body mass index; CES-D = Center for Epidemiologic Studies-Depression Scale; CI = confidence interval; OR = odds ratio. Open in new tab In addition, in Model 1, participants aged ≥50 years with ≥1 Ɛ4 allele had a 38% greater odds of reporting difficulty falling or staying asleep compared to those with 0 Ɛ4 alleles, but this was at the trend level (Table 6). After further adjustment in Model 2, however, a significant association emerged, such that those with ≥1 Ɛ4 allele had almost a 50% greater odds of trouble falling or staying asleep (OR = 1.49, 95% CI 1.02, 2.17). There was a trend-level association in Model 1 between Ɛ2/Ɛ2 or Ɛ2/Ɛ3 genotype and a reduced odds of napping in this older subsample (OR = 0.67, 95% CI 0.43, 1.04) but not after further adjustment in Model 2 (OR = 0.72, 95% CI 0.45, 1.15). Table 6 Participants Aged ≥50 Only: Association of APOE Genotype With Trouble Falling or Staying Asleep and Napping. . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Trouble falling or staying asleep n = 793 n = 734 ≥1 Ɛ4 alleles 1.38 (0.97, 1.96), p = .070 1.49 (1.02, 2.17), p = .042 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.94 (0.58, 1.54), p = .815 0.82 (0.48, 1.41), p = .471 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.37 (0.95, 1.96), p = .090 1.43 (0.97, 2.12), p = .073 Taking naps n = 801 n = 740 ≥1 Ɛ4 alleles 0.83 (0.60, 1.15), p = .264 0.82 (0.57, 1.16), p = .252 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.67 (0.43, 1.04), p = .073 0.72 (0.45, 1.15), p = .166 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 0.77 (0.55, 1.08), p = .128 0.77 (0.53, 1.10), p = .148 . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Trouble falling or staying asleep n = 793 n = 734 ≥1 Ɛ4 alleles 1.38 (0.97, 1.96), p = .070 1.49 (1.02, 2.17), p = .042 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.94 (0.58, 1.54), p = .815 0.82 (0.48, 1.41), p = .471 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.37 (0.95, 1.96), p = .090 1.43 (0.97, 2.12), p = .073 Taking naps n = 801 n = 740 ≥1 Ɛ4 alleles 0.83 (0.60, 1.15), p = .264 0.82 (0.57, 1.16), p = .252 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.67 (0.43, 1.04), p = .073 0.72 (0.45, 1.15), p = .166 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 0.77 (0.55, 1.08), p = .128 0.77 (0.53, 1.10), p = .148 Model 1 adjusted for age, age2, sex, race; Model 2 adjusted for age, age2, sex, race, education, BMI, CES-D, vascular burden score, snoring, and daytime sleepiness. Bold values indicate p <.05. APOE = apolipoprotein; BMI = body mass index; CES-D = Center for Epidemiologic Studies-Depression Scale; CI = confidence interval; OR = odds ratio. Open in new tab Table 6 Participants Aged ≥50 Only: Association of APOE Genotype With Trouble Falling or Staying Asleep and Napping. . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Trouble falling or staying asleep n = 793 n = 734 ≥1 Ɛ4 alleles 1.38 (0.97, 1.96), p = .070 1.49 (1.02, 2.17), p = .042 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.94 (0.58, 1.54), p = .815 0.82 (0.48, 1.41), p = .471 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.37 (0.95, 1.96), p = .090 1.43 (0.97, 2.12), p = .073 Taking naps n = 801 n = 740 ≥1 Ɛ4 alleles 0.83 (0.60, 1.15), p = .264 0.82 (0.57, 1.16), p = .252 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.67 (0.43, 1.04), p = .073 0.72 (0.45, 1.15), p = .166 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 0.77 (0.55, 1.08), p = .128 0.77 (0.53, 1.10), p = .148 . Model 1 . Model 2 . OR (95% CI), p-value . OR (95% CI), p-value . Trouble falling or staying asleep n = 793 n = 734 ≥1 Ɛ4 alleles 1.38 (0.97, 1.96), p = .070 1.49 (1.02, 2.17), p = .042 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.94 (0.58, 1.54), p = .815 0.82 (0.48, 1.41), p = .471 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 1.37 (0.95, 1.96), p = .090 1.43 (0.97, 2.12), p = .073 Taking naps n = 801 n = 740 ≥1 Ɛ4 alleles 0.83 (0.60, 1.15), p = .264 0.82 (0.57, 1.16), p = .252 0 Ɛ4 alleles 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ2, Ɛ2/Ɛ3 0.67 (0.43, 1.04), p = .073 0.72 (0.45, 1.15), p = .166 Ɛ3/Ɛ3 1.00 (ref) 1.00 (ref) Ɛ2/Ɛ4, Ɛ3/Ɛ4, Ɛ4/Ɛ4 0.77 (0.55, 1.08), p = .128 0.77 (0.53, 1.10), p = .148 Model 1 adjusted for age, age2, sex, race; Model 2 adjusted for age, age2, sex, race, education, BMI, CES-D, vascular burden score, snoring, and daytime sleepiness. Bold values indicate p <.05. APOE = apolipoprotein; BMI = body mass index; CES-D = Center for Epidemiologic Studies-Depression Scale; CI = confidence interval; OR = odds ratio. Open in new tab APOE and Race Interactions When we examined the APOE Ɛ4 (≥1 vs. 0 Ɛ4 alleles) × race interaction in fully adjusted models, we observed an interaction term p-value = .879 for trouble falling or staying asleep, p = .851 for napping, and p = .155 for sleep duration. Because the interaction for sleep duration approached a statistical trend level, we explored this by stratifying by race. We did so in the minimally adjusted model to avoid overfitting the data and found among the 248 African Americans that those with ≥1 Ɛ4 allele had a 65% greater odds of being in a shorter sleep duration category, but this was at the trend level (OR = 1.65, 95% CI 0.98, 2.78) (Supplemental Table 1). Compared to African American participants with the Ɛ3/Ɛ3 genotype, however, those with ≥1 Ɛ4 allele had a statistically significant 83% greater odds of shorter sleep (OR = 1.83, 95% CI 1.05, 3.19). This association did not reach significance in minimally adjusted models, however, among the 1008 white participants (Supplemental Table 1). DISCUSSION We examined the association of APOE genotype with self-reported nonrespiratory sleep parameters in 1264 cognitively intact adults aged 20 to 92 years (mean = 58). After adjustment for potential confounders, we found that those with one or more Ɛ4 alleles—and therefore at elevated risk for AD—had a greater odds of reporting shorter sleep duration than those with no Ɛ4 alleles (Ɛ2/Ɛ2, Ɛ2/Ɛ3 and Ɛ3/Ɛ3 genotypes) and those with the Ɛ3/Ɛ3 genotype, which is neutral for AD risk. We also found that those with the Ɛ2/Ɛ2 or Ɛ2/Ɛ3 genotype and thus at reduced risk of AD, had a reduced odds of reporting napping, compared to those with the neutral Ɛ3/Ɛ3 genotype. When we limited analyses to adults aged ≥50 years, the findings concerning sleep duration were stronger, and we found that those with an Ɛ4 allele had a greater odds of difficulty falling or staying asleep than those without; however, the associations with napping were no longer significant. We also found some evidence for an interaction of APOE Ɛ4 with race, with trend-level and significant associations of the Ɛ4 allele with sleep duration among African American but not white participants. To our knowledge, this is the first study to identify associations of APOE genotype with self-reported sleep duration, difficulty initiating or maintaining sleep, and napping habits in cognitively intact individuals. A few prior studies have investigated associations between APOE genotype and nonrespiratory sleep/wake parameters in persons without dementia, although most have not focused on a potential effect of APOE on these sleep parameters per se. For example, Lim et al.33 showed that greater consolidation of objectively measured rest/activity rhythms buffered against the risk of AD conferred by the Ɛ4 allele and reduced the effect of Ɛ4 on density of neurofibrillary tangles. Also, Asada et al.34 showed in a case-control study that napping duration interacted with APOE Ɛ4 genotype. Naps longer than 60 minutes were associated with a greater AD risk among Ɛ4 carriers but not noncarriers, and there was evidence that short naps were associated with reductions in risk of incident AD in both groups.34 Further, Tsapanou et al.35 reported that Ɛ4 genotype was associated with self-report measures of SDB-related phenomena (snoring, shortness of breath during sleep/ headache on waking) but found no associations with measures of sleep quality or adequacy or of daytime sleepiness. In a particularly relevant recent pilot study (n = 31), Drogos et al.36 found that Ɛ4 carriers had poorer sleep efficiency, as measured by actigraphy and polysomnography (PSG) and shorter sleep duration and greater wake after sleep onset and REM sleep, measured by PSG, than noncarriers; they detected no associations, however, of APOE Ɛ4 status with self-reported sleep measures. In addition, it should be noted that a number of genome-wide association studies have been performed examining variants associated with sleep duration.37–41 Polymorphisms in or near APOE were not identified in these studies; however, it is worth noting that only some polymorphisms associated with specific genes (eg, Pax8 and ABCC9) were replicated between studies. Our results speak to the intriguing possibility that APOE Ɛ4 truncates sleep duration and increases sleep onset latency and fragmentation. In light of prior support for a causal link between sleep deprivation and AD progression,1,2 if APOE Ɛ4 shortens sleep, it would be plausible that APOE exerts its effect on AD risk at least in part by shortening sleep duration, complementing the aforementioned findings concerning rest/activity consolidation as a moderator of the effect of Ɛ4 on AD risk.33 It is also plausible that, rather than directly affecting sleep, APOE genotype modulates vulnerability to sleep disturbance. Indeed, findings that anticholinergic medication use is associated with greater sleep disturbance in cognitively normal Ɛ4 carriers than in noncarriers support the notion that the sleep of Ɛ4 carriers is more easily perturbed.18 If disturbed sleep does promote amyloid burden and AD risk, our findings might suggest that APOE Ɛ4-positive individuals with sleep disturbance receive special clinical attention with the goal of optimizing nighttime sleep duration and thereby reducing the risk of AD.33 Our findings in analyses that included both younger people and those aged 50 years and older also raise the possibility that the APOE Ɛ2 allele reduces the likelihood of regular napping. If napping has a negative impact on brain health, our findings that the Ɛ2 allele is associated with reduced napping may suggest that reduced napping is a mechanism through which Ɛ2 protects against AD. Although self-report of napping has been linked to a reduced risk of cognitive decline,42 when naps have been measured by actigraphy, longer naps have been tied to poorer performance on neuropsychological tests,43,44 and napping 3 days per week or more has been linked to amyloid deposition.5 When we limited our analyses to participants aged 50 years and older, however, the association between APOE Ɛ2 and napping was no longer significant. Whether this was due to reduced statistical power or a specific effect of Ɛ2 on napping in younger people only is unclear and could be especially important if an effect of naps on cognition and brain health varies by age. Further research is needed to clarify the role of the Ɛ2 allele in these associations. Importantly, we identified a potential interaction between the APOE Ɛ4 allele and race with regard to sleep duration. Although this should be interpreted cautiously given the nonsignificant interaction term in the model, between-race differences have been noted in associations between APOE Ɛ4 genotype and cognitive outcomes,21 and it is unclear whether sleep disturbance may play a role in these associations. In light of the growing interest in sleep as a potential risk factor for AD and in race/ethnic disparities in AD, further studies in this domain are warranted. This study’s strength is its large community-dwelling sample with well-characterized cognitive status and APOE genotyping. However, several limitations warrant consideration. First, our sleep measures were limited to self-report. Studies of APOE and sleep in large samples that include actigraphy or PSG would clarify whether the associations we observed between APOE and nonrespiratory sleep parameters also arise when sleep is measured objectively. Also, although our participants were considered to be free of MCI or dementia, we are unable to rule out the possibility that Ɛ4 carriers had, on average, a higher level of Aβ deposition in their brains. Because Aβ deposition leads to sleep loss in AD mice45 and Drosophila,2 observed differences in sleep duration, onset and maintenance difficulties, and napping between Ɛ4 carriers and noncarriers may be explained at least in part by preclinical AD. Moreover, reports of reduced nighttime sleep duration and greater daytime napping may be attributable to SDB. We adjusted for SDB risk factors and markers (BMI, snoring, and excessive daytime sleepiness), but residual confounding by SDB may explain our results. If napping does in fact reflect SDB, this would still be of interest: although the Ɛ4 allele has been linked to a greater risk of SDB,9,10 it has not yet been established that Ɛ2/Ɛ2 and Ɛ2/Ɛ3 genotypes protect against SDB. Studies with PSG in large samples with APOE genotyping are needed to further examine this possibility. Additional studies are also needed to clarify the nonrespiratory polysomnographic correlates of APOE genotype in cognitively normal populations. Although a study of persons with MCI found that Ɛ4 carriers obtain less REM sleep than noncarriers,16 little is known about whether APOE genotype affects REM or slow-wave sleep in those without impairment. Such findings would be important as they might indicate sleep-specific mechanisms by which APOE affects AD risk. In summary, our results indicate an association of APOE genotype with nonrespiratory sleep parameters: sleep duration, difficulty falling or staying asleep, and napping, and suggest that the association with sleep duration may differ between African Americans and whites. Findings provide further support for an important link between sleep and AD risk that warrants further study using objective sleep measures. SUPPLEMENTARY MATERIAL Supplementary material is available at SLEEP online. FUNDING This study was supported in part by the Intramural Research Program (IRP), National Institute on Aging (NIA), National Institutes of Health (NIH), and by Research and Development Contract HHSN-260-2004-00012C. Dr. Spira was supported in part by R01AG050507 and other grants from the NIA and has received funding from the William and Ella Owens Medical Research Foundation. Investigators from the NIH NIA IRP were involved in all aspects of this manuscript, including the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. DISCLOSURE STATEMENT APS has agreed to serve as a consultant to Awarables, Inc. in support of an NIH grant. SMR has an immediate family member who has received grant/research support unrelated to this study from Takeda, Roche, Pfizer, Lundbeck, Johnson & Johnson, Intracellular, GE Healthcare, DART Neuroscience, Avid/Lilly, and Piramal Imaging. The other authors report no conflicts of interest. ACKNOWLEDGMENTS Results related to this study were presented at the 2016 Meeting of the American Association for Geriatric Psychiatry, Washington, DC. REFERENCES 1. Kang JE , Lim MM, Bateman RJet al. Amyloid-beta dynamics are regulated by orexin and the sleep-wake cycle . Science . 2009 ; 326 ( 5955 ): 1005 – 1007 . Google Scholar Crossref Search ADS PubMed WorldCat 2. Tabuchi M , Lone SR, Liu Set al. Sleep interacts with aβ to modulate intrinsic neuronal excitability . Curr Biol . 2015 ; 25 ( 6 ): 702 – 712 . Google Scholar Crossref Search ADS PubMed WorldCat 3. Spira AP , Gamaldo AA, An Yet al. Self-reported sleep and β-amyloid deposition in community-dwelling older adults . JAMA Neurol . 2013 ; 70 ( 12 ): 1537 – 1543 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 4. Sprecher KE , Bendlin BB, Racine AMet al. Amyloid burden is associated with self-reported sleep in nondemented late middle-aged adults . Neurobiol Aging . 2015 ; 36 ( 9 ): 2568 – 2576 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Ju YE , McLeland JS, Toedebusch CDet al. Sleep quality and preclinical Alzheimer disease . JAMA Neurol . 2013 ; 70 ( 5 ): 587 – 593 . Google Scholar Crossref Search ADS PubMed WorldCat 6. Corder EH , Saunders AM, Strittmatter WJet al. Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer’s disease in late onset families . Science . 1993 ; 261 ( 5123 ): 921 – 923 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Conejero-Goldberg C , Gomar JJ, Bobes-Bascaran Tet al. APOE2 enhances neuroprotection against Alzheimer’s disease through multiple molecular mechanisms . Mol Psychiatry . 2014 ; 19 ( 11 ): 1243 – 1250 . Google Scholar Crossref Search ADS PubMed WorldCat 8. Alzheimer’s Association . 2015 Alzheimer’s disease facts and figures . Alzheimers Dement . 2015 ; 11 ( 3 ): 332 – 384 . Crossref Search ADS PubMed WorldCat 9. Kadotani H , Kadotani T, Young Tet al. Association between apolipoprotein E epsilon4 and sleep-disordered breathing in adults . JAMA . 2001 ; 285 ( 22 ): 2888 – 2890 . Google Scholar Crossref Search ADS PubMed WorldCat 10. Gottlieb DJ , DeStefano AL, Foley DJet al. APOE epsilon4 is associated with obstructive sleep apnea/hypopnea: the Sleep Heart Health Study . Neurology . 2004 ; 63 ( 4 ): 664 – 668 . Google Scholar Crossref Search ADS PubMed WorldCat 11. O’Hara R , Schröder CM, Kraemer HCet al. Nocturnal sleep apnea/hypopnea is associated with lower memory performance in APOE epsilon4 carriers . Neurology . 2005 ; 65 ( 4 ): 642 – 644 . Google Scholar Crossref Search ADS PubMed WorldCat 12. Spira AP , Blackwell T, Stone KLet al. Sleep disordered breathing and cognition in community-dwelling older women . J Am Geriatr Soc . 2008 ; 56 ( 1 ): 45 – 50 . Google Scholar Crossref Search ADS PubMed WorldCat 13. Nikodemova M , Finn L, Mignot E, Salzieder N, Peppard PE. Association of sleep disordered breathing and cognitive deficit in APOE ε4 carriers . Sleep . 2013 ; 36 ( 6 ): 873 – 880 . Google Scholar Crossref Search ADS PubMed WorldCat 14. de Oliveira FF , Bertolucci PH, Chen ES, Smith Mde A. Assessment of sleep satisfaction in patients with dementia due to Alzheimer’s disease . J Clin Neurosci . 2014 ; 21 ( 12 ): 2112 – 2117 . Google Scholar Crossref Search ADS PubMed WorldCat 15. Craig D , Hart DJ, Passmore AP. Genetically increased risk of sleep disruption in Alzheimer’s disease . Sleep . 2006 ; 29 ( 8 ): 1003 – 1007 . Google Scholar Crossref Search ADS PubMed WorldCat 16. Hita-Yanez E , Atienza M, Gil-Neciga E, Cantero JL. Disturbed sleep patterns in elders with mild cognitive impairment: the role of memory decline and ApoE varepsilon4 genotype . Curr Alzheimer Res . 2012 ; 9 ( 3 ): 290 – 297 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Hita-Yañez E , Atienza M, Cantero JL. Polysomnographic and subjective sleep markers of mild cognitive impairment . Sleep . 2013 ; 36 ( 9 ): 1327 – 1334 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Nebes RD , Pollock BG, Perera S, Halligan EM, Saxton JA. The greater sensitivity of elderly APOE ε4 carriers to anticholinergic medications is independent of cerebrovascular disease risk . Am J Geriatr Pharmacother . 2012 ; 10 ( 3 ): 185 – 192 . Google Scholar Crossref Search ADS PubMed WorldCat 19. Spira AP , Gonzalez CE, Venkatraman VKet al. Sleep duration and subsequent cortical thinning in cognitively normal older adults . Sleep . 2016 ; 39 ( 5 ): 1121 – 1128 . Google Scholar Crossref Search ADS PubMed WorldCat 20. Shock N , Greulich R, Andres Ret al. Normal human aging: The Baltimore Longitudinal Study of Aging . Washington, DC : NIH , 1984 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 21. Manly JJ , Mayeux R. Ethnic differences in dementia and Alzheimer’s disease . In: Anderson NB, Bulatao RA, Cohen B, eds. Critical Perspectives on Racial and Ethnic Differences in Health in Late Life . Washington, D.C .: National Academies Press , 2004 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 22. Hixson JE , Vernier DT. Restriction isotyping of human apolipoprotein E by gene amplification and cleavage with HhaI . J Lipid Res . 1990 ; 31 ( 3 ): 545 – 548 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 23. Koch W , Ehrenhaft A, Griesser Ket al. TaqMan systems for genotyping of disease-related polymorphisms present in the gene encoding apolipoprotein E . Clin Chem Lab Med . 2002 ; 40 ( 11 ): 1123 – 1131 . Google Scholar Crossref Search ADS PubMed WorldCat 24. Blessed G , Tomlinson BE, Roth M. The association between quantitative measures of dementia and of senile change in the cerebral grey matter of elderly subjects . Br J Psychiatry . 1968 ; 114 ( 512 ): 797 – 811 . Google Scholar Crossref Search ADS PubMed WorldCat 25. Morris JC . The Clinical Dementia Rating (CDR): current version and scoring rules . Neurology . 1993 ; 43 ( 11 ): 2412 – 2414 . Google Scholar Crossref Search ADS PubMed WorldCat 26. Petersen RC , Smith GE, Waring SC, Ivnik RJ, Tangalos EG, Kokmen E. Mild cognitive impairment: clinical characterization and outcome . Arch Neurol . 1999 ; 56 ( 3 ): 303 – 308 . Google Scholar Crossref Search ADS PubMed WorldCat 27. American Psychiatric Association . Diagnostic and statistical manual of mental disorders: DSM-III .3rd ed. Washington, DC : American Psychiatric Association , 1980 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 28. McCarrey AC , An Y, Kitner-Triolo MH, Ferrucci L, Resnick SM. Sex differences in cognitive trajectories in clinically normal older adults . Psychol Aging . 2016 ; 31 ( 2 ): 166 – 175 . Google Scholar Crossref Search ADS PubMed WorldCat 29. Radloff L . The CES-D Scale: a self report depression scale for research in the general population . Applied Psychological Measurement . 1977 ; 1 ( 3 ): 385 – 401 . Google Scholar Crossref Search ADS WorldCat 30. Hirshkowitz M , Whiton K, Albert SMet al. National Sleep Foundation’s sleep time duration recommendations: methodology and results summary . Sleep Health . 2015 ; 1 ( 1 ): 40 – 43 . Google Scholar Crossref Search ADS PubMed WorldCat 31. Watson NF , Badr MS, Belenky Get al. Joint consensus statement of the American Academy of Sleep Medicine and Sleep Research Society on the recommended amount of sleep for a healthy adult: methodology and discussion . Sleep . 2015 ; 38 ( 8 ): 1161 – 1183 . Google Scholar Crossref Search ADS PubMed WorldCat 32. Farrer LA , Cupples LA, Haines JLet al. Effects of age, sex, and ethnicity on the association between apolipoprotein E genotype and Alzheimer disease. A meta-analysis. APOE and Alzheimer Disease Meta Analysis Consortium . JAMA . 1997 ; 278 ( 16 ): 1349 – 1356 . Google Scholar Crossref Search ADS PubMed WorldCat 33. Lim AS , Yu L, Kowgier M, Schneider JA, Buchman AS, Bennett DA. Modification of the relationship of the apolipoprotein E ε4 allele to the risk of Alzheimer disease and neurofibrillary tangle density by sleep . JAMA Neurol . 2013 ; 70 ( 12 ): 1544 – 1551 . Google Scholar Crossref Search ADS PubMed WorldCat 34. Asada T , Motonaga T, Yamagata Z, Uno M, Takahashi K. Associations between retrospectively recalled napping behavior and later development of Alzheimer’s disease: association with APOE genotypes . Sleep . 2000 ; 23 ( 5 ): 629 – 634 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 35. Tsapanou A , Scarmeas N, Gu Yet al. Examining the association between apolipoprotein E (APOE) and self-reported sleep disturbances in non-demented older adults . Neurosci Lett . 2015 ; 606 : 72 – 76 . Google Scholar Crossref Search ADS PubMed WorldCat 36. Drogos LL , Gill SJ, Tyndall AVet al. Evidence of association between sleep quality and APOE ε4 in healthy older adults: A pilot study . Neurology . 2016 ; 87 ( 17 ): 1836 – 1842 . Google Scholar Crossref Search ADS PubMed WorldCat 37. Allebrandt KV , Amin N, Müller-Myhsok Bet al. A K(ATP) channel gene effect on sleep duration: from genome-wide association studies to function in Drosophila . Mol Psychiatry . 2013 ; 18 ( 1 ): 122 – 132 . Google Scholar Crossref Search ADS PubMed WorldCat 38. Gottlieb DJ , Hek K, Chen THet al. Novel loci associated with usual sleep duration: the CHARGE Consortium Genome-Wide Association Study . Mol Psychiatry . 2015 ; 20 ( 10 ): 1232 – 1239 . Google Scholar Crossref Search ADS PubMed WorldCat 39. Scheinfeldt LB , Gharani N, Kasper RSet al. Using the Coriell Personalized Medicine Collaborative Data to conduct a genome-wide association study of sleep duration . Am J Med Genet B Neuropsychiatr Genet . 2015 ; 168 ( 8 ): 697 – 705 . Google Scholar Crossref Search ADS PubMed WorldCat 40. Cade BE , Gottlieb DJ, Lauderdale DSet al. Common variants in DRD2 are associated with sleep duration: the CARe consortium . Hum Mol Genet . 2016 ; 25 ( 1 ): 167 – 179 . Google Scholar Crossref Search ADS PubMed WorldCat 41. Jones SE , Tyrrell J, Wood ARet al. Genome-wide association analyses in 128,266 individuals identifies new morningness and sleep duration loci . PLoS Genet . 2016 ; 12 ( 8 ): e1006125 . Google Scholar Crossref Search ADS PubMed WorldCat 42. Keage HA , Banks S, Yang KL, Morgan K, Brayne C, Matthews FE. What sleep characteristics predict cognitive decline in the elderly? Sleep Med . 2012 ; 13 ( 7 ): 886 – 892 . Google Scholar Crossref Search ADS PubMed WorldCat 43. Blackwell T , Yaffe K, Ancoli-Israel Set al. ; Study of Osteoporotic Fractures Group. Poor sleep is associated with impaired cognitive function in older women: the study of osteoporotic fractures . J Gerontol A Biol Sci Med Sci . 2006 ; 61 ( 4 ): 405 – 410 . Google Scholar Crossref Search ADS PubMed WorldCat 44. Cross N , Terpening Z, Rogers NLet al. Napping in older people ‘at risk’ of dementia: relationships with depression, cognition, medical burden and sleep quality . J Sleep Res . 2015 ; 24 ( 5 ): 494 – 502 . Google Scholar Crossref Search ADS PubMed WorldCat 45. Roh JH , Huang Y, Bero AWet al. Disruption of the sleep-wake cycle and diurnal fluctuation of β-amyloid in mice with Alzheimer’s disease pathology . Sci Transl Med . 2012 ; 4 ( 150 ): 150ra122 . Google Scholar Crossref Search ADS PubMed WorldCat Author notes Address correspondence to: Adam P. Spira, PhD, 624 N. Broadway, Hampton House, Rm. 794 Baltimore, MD 21205, USA. Telephone: 410-614-9498; Fax: 410-614-7469; Email: [email protected] This work was performed at the National Institute on Aging Intramural Research Program and at Johns Hopkins University. Published by Oxford University Press on behalf of Sleep Research Society (SRS) 2017. This work is written by (a) US Government employee(s) and is in the public domain in the US. Published by Oxford University Press on behalf of Sleep Research Society (SRS) 2017. This work is written by (a) US Government employee(s) and is in the public domain in the US.

Muzalev,, Konstantin;Lobbezoo,, Frank;Janal, Malvin, N;Raphael, Karen, G

doi: 10.1093/sleep/zsx078pmid: 28482089

Abstract Study Objectives Sleep bruxism (SB) is considered as a possible etiological factor for temporomandibular disorder (TMD) pain. However, polysomnographic (PSG) studies, which are current “gold standard” diagnostic approach to SB, failed to prove an association between SB and TMD. A possible explanation could be that PSG studies have considered only limited characteristics of SB activity: the number of SB events per hour and, sometimes, the total duration of SB per night. According to the sports sciences literature, lack of adequate rest time between muscle activities leads to muscle overloading and pain. Therefore, the aim of this study was to determine whether the intervals between bruxism events differ between patients with and without TMD pain. Methods Two groups of female volunteers were recruited: myofascial TMD pain group (n=124) and non-TMD control group (n=46). From these groups, we selected 86 (69%) case participants and 37 (80%) controls who had at least two SB episodes per night based on PSG recordings. A linear mixed model was used to compare case and control groups over the repeated observations of interepisode intervals. Results The duration of interepisode intervals was statistically similar in the case (mean [standard deviation {SD}] 1137.7 [1975.8] seconds)] and control (mean [SD] 1192.0 [1972.0] seconds) groups. There were also a similar number of SB episodes per hour and a total duration of SB episodes in both groups. Conclusions The current data fail to support the idea that TMD pain can be explained by increasing number of SB episodes per hour of sleep or decreasing the time between SB events. facial pain, myalgia, muscle contraction, sleep, polysomnography Statement of Significance The role of sleep bruxism (SB) in the etiology of jaw-muscle pain remains controversial. In this paper, we aimed to explain developing of jaw-muscle pain by taking into account the duration of “rest-time” intervals between SB events. Because no significant difference between case group with jaw-muscle pain and pain-free control group was detected, we conclude that the duration of “rest-time” intervals is not the key factor. Further research is needed to clarify the possible etiological factors for jaw-muscle pain. INTRODUCTION Sleep bruxism (SB) is a repetitive jaw-muscle activity characterized by clenching or grinding of the teeth and/or by bracing or thrusting of the mandible during sleep.1 It is a rather common condition. Reported prevalence is up to 41% (range: 3.5%–40.6%) in the general adult population,2 with the highest estimates coming from studies that used self-report to diagnose SB. Currently, polysomnography (PSG) with audio-video recordings is considered the most accurate method to diagnose SB.3,4 Based on PSG studies,5 the reported prevalence of SB in the general population is around 7%. SB has long been considered as a possible etiological factor for temporomandibular disorders (TMDs).6–8 TMD is a collective term embracing a number of clinical problems of the musculoskeletal structures of the masticatory system. The most frequently reported symptom is pain originating from the masticatory muscles, which is often aggravated during function.9–12 To diagnose jaw-muscle pain (myofascial TMD pain) as well as other TMDs, structured self-report and clinical instruments like the Research Diagnostic Criteria for TMD10 (RDC/TMD) and the Diagnostic Criteria for TMD12 (DC/TMD) are recommended. In the dental office, jaw-muscle pain can be found in many patients who are assumed to engage in SB. It can be speculated that patients with developed jaw-muscle pain are presumed to have “overloaded” their muscles, which is not unlikely to occur due to the unconscious nature of SB. Several hypotheses have been proposed to explain the development of overuse-related muscle pain.13 The most important of these are the duration of the muscle activity,14 the type of muscle contraction15 (ie, concentric or eccentric), and the rest intervals between subsequent muscle contractions.16 It is widely accepted that the balance between muscle work and recovery plays a crucial role in preventing overloading.17 A lack of adequate recovery might thus lead to muscle overloading and pain. Previous studies about the association between SB and jaw-muscle pain, which used PSG and the RDC/TMD as the respective diagnostics approaches, have used only few characteristics of jaw-muscle activity to describe SB, viz, the number of bruxism episodes per hour of sleep and, in some cases, the total duration of bruxism episodes.18,19 These studies yielded contradictory results: some found an association between TMD pain and SB20, whereas others detected either no association21 or even a negative one.22,23 There are no studies so far describing associations between the duration of interepisode intervals (IEI) and jaw-muscle pain. However, this characteristic of jaw-muscle activity could be important in producing muscle overloading that in turn may cause jaw-muscle pain. Therefore, the aim of this study was to determine, based on sleep-laboratory PSG recordings, whether the intervals between SB events (IEI) differ between patients with and without myofascial face pain. We hypothesized shorter IEI in TMD patients than non-TMD control participants. MATERIAL AND METHODS The study was approved by The Institutional Review Board at the New York University (NYU) School of Medicine (New York, New York) (study # I07-303). Participants Participants were recruited from among patients seeking treatment at the NYU College of Dentistry. Before entering the study, all participants completed a full informed consent process and signed an informed consent form. For the present study, only female volunteers were recruited, given the higher prevalence of TMD pain in women.24 The participants were allocated in the case group or in the control group based on the presence or absence of myofascial TMD pain, independent of their own beliefs regarding the presence or absence of SB. The control sample was a demographic match to the case participants regarding age, socioeconomic status, and race. The RDC/TMD was used to establish the diagnosis of myofascial TMD pain. Two raters used RDC/TMD training tapes and materials and were initially calibrated to high levels of diagnostic concordance, with repeat periodic reliability testing throughout the study. One of the two raters was the “main” rater providing the study data, whereas the other rater (who was experienced in using the RDC) served as gold standard. Potential participants were excluded from either group if they reported a history of trauma to the face, acute dental problems, or recent extensive dental treatment. At least 48 hours should have passed between the latest dental treatment and the RDC/TMD examination. Also, persons were excluded from participation if they were pregnant, habitually smoked after bedtime, habitually slept less than 4 hours per night, had a neuropathic facial pain condition, or had been diagnosed with severe obstructive sleep apnea (ie, an apnea-hypopnea index 30 or more events/hour of sleep) requiring continuous positive airway pressure, which would have interfered with SB measurement (see below). Moreover, studying the IEI between SB events requires the presence of at least two SB episodes per night. Thus, participant who presented less than two SB episodes per night based on PSG registration (see below) were excluded from the final analysis. In all, 124 women with a diagnosis of myofascial TMD pain and 46 pain-free control participants completed the sleep laboratory studies. From them, 86 case participants (69%) and 37 controls (80%) had at least two SB episodes. Thus, the final analysis of IEI between SB events was based on data from 123 participants. Polysomnography The PSG registrations were performed at a sleep laboratory affiliated with the NYU School of Medicine. Participants were studied in the sleep laboratory for two consecutive nights. The first night allowed for adaptation to the sleep laboratory environment. The second night was used for the registration of jaw-muscle activity and sleep architecture. Data from the first night were, however, used for the statistical analysis in 10 instances: three cases failed to return for the second night and six cases and one control were missing data during the second night due to technical problems. The onset and offset times of the nocturnal PSG recordings were determined from each participant’s habitual sleep times, with the recordings running approximately from 10:30 pm to 07:00 am. The PSG record consisted of a six-channel electroencephalogram, a bilateral electrooculogram, a bilateral submental (chin) and anterior tibialis electromyogram (EMG), a right and left masseter and temporalis EMG, an electrocardiogram, chest and abdominal motion (by means of belts with piezoelectric sensors), body position, airflow by nasal pressure transducer and nasal-oral thermistor, and oximetry. Outcome Measures PSG data were exported to Stellate Harmonie software (Natus, San Carlos, California) for analysis. Two raters independently scored sleep stages, arousals, apneas, and periodic limb movements. Inter-rater reliability for identification of SB episodes by the two scorers was excellent (κ = 0.89). For reasons unrelated to this study, one of the sleep scorers was not blinded to the participants’ case-control status. Jaw-muscle activity was analyzed using the Research Diagnostic Criteria for SB (RDC/SB) for the right masseter.3 Masseter activity that exceeded twice the amplitude of the relaxed waking EMG level before sleep was used as the threshold. Against this reference, phasic episodes were defined by three or more brief (> 0.25 seconds and < 2.0 seconds) EMG bursts. Tonic episodes were scored if the burst duration was longer than 2 seconds. A mixed episode corresponds to phasic and tonic bursts, separated by an interval lasting less than 3 seconds. EMG episodes that were separated by an IEI of at least 3 seconds were scored as different episodes. Data Analysis As a first step, the frequency of SB episodes per hour of sleep was calculated and participants who had two or more SB episodes per night were selected. Then, the frequency of SB episodes per hour of sleep, the duration of SB episodes, and the duration of the intervals between SB episodes were calculated. The time between sleep onset and the first SB episode, as well as the time between the last SB episode and the end of sleep, were excluded from the last calculation. Independent samples t-tests were used to compare age and educational levels between the two groups, and a chi-square test was used to compare groups on race, ethnicity, marital, and work status. A median test was used to compare groups on the average number and average duration of the SB episodes. A linear mixed model was used to compare case and control groups over the repeated observations of IEI, following rank transformation of those durations to correct skew in their distribution. The choice of rank transformation, instead of commonly used logarithm or square root one, was based on our experience that suggests that rank transformation deals better with outliers.25 All tests employed a significance level of five percent. Statistical analysis was performed using IBM SPSS Statistics 23 software (IBM Corp., Armonk, New York, USA). RESULTS The control and case groups did not differ in terms of measured demographic characteristics. Controls and cases were, respectively, of similar age (mean [standard deviation {SD}] = 34.7 [12.9] vs. 37.8 [13.4] years; p = .23), educational level (mean [SD] = 15.6 [2.1] vs. 15.9 [2.2] years; p = .57), race (61.1 vs. 70% white; p = .23), Hispanic ethnicity (26.7 vs. 17.3%; p = .77), currently employed for pay (69.4 vs. 59.5%; p = 0.31), and never married (66.7 vs. 65.8%; χ2 = 1.5, p = .92). Cases reported a mean (SD) characteristic pain intensity of 5.1 (1.75) and median pain chronicity of 84 months. Polysomnography. The number of SB episodes per hour of sleep ranged widely in this sample, from 0 to 11, with a median of 1.4 episodes. There were a similar number of SB episodes per hour in control and case groups (mean [SD] = 2.2 [1.9] and 2.1 [2.0]), respectively (p = .77). The duration of SB episodes per night also varied widely, with durations ranging from 0.9 to 402.6 seconds and a median duration of 45 seconds. Groups had a similar total duration of SB episodes (mean [SD] = 72.2 [71.4] seconds for the control group and 66.7 [74.3] seconds for the case group; p = .3). Figure 1 shows the distribution of IEI before and after rank transformation was performed. As shown, there was considerable skew in the crude distribution (A), so that while median intervals were between 5 and 6 minutes (281 seconds in cases and 393 seconds in controls), those intervals could range above 4 hours. Rank transformation of these data greatly improved the symmetry of these distributions (B), and those data were used in further analysis. Although mean rank was lower among cases than controls, consistent with the hypothesized shorter latency to next event, the linear mixed model analysis indicated statistically similar levels in the two groups. PSG results are summarized in Table 1. Figure 1 Open in new tabDownload slide Box plot showing the distribution of the duration of the intervals between sleep bruxism (SB) episodes in control and case participants before (A) and after (B) rank transformation. Figure 1 Open in new tabDownload slide Box plot showing the distribution of the duration of the intervals between sleep bruxism (SB) episodes in control and case participants before (A) and after (B) rank transformation. Table 1 Polysomnograpy (PSG): Comparison of Sleep Bruxism (SB) Measures in Myofascial Temporomandibular Pain Disorder (TMD) Cases and Controls. PSG Measure Controls (n=37) Cases (n=86) p value Mean (SD) Median Mean (SD) Median RMMA episodes (per hour) 2.2 (1.9) 1.4 2.1 (2.0) 1.4 .77 Duration (all episodes, seconds) 72.2 (71.4) 48.9 66.7 (74.3) 41.2 .3 Duration of interepisode intervals (seconds) 1192.0 (1972.0) 393 1137.7 (1975.8) 281 .51 Duration of interepisode intervals (rank) 827.4 (482.4) 864 806.7 (463.5) 793 .44 PSG Measure Controls (n=37) Cases (n=86) p value Mean (SD) Median Mean (SD) Median RMMA episodes (per hour) 2.2 (1.9) 1.4 2.1 (2.0) 1.4 .77 Duration (all episodes, seconds) 72.2 (71.4) 48.9 66.7 (74.3) 41.2 .3 Duration of interepisode intervals (seconds) 1192.0 (1972.0) 393 1137.7 (1975.8) 281 .51 Duration of interepisode intervals (rank) 827.4 (482.4) 864 806.7 (463.5) 793 .44 SD = standard deviation. Open in new tab Table 1 Polysomnograpy (PSG): Comparison of Sleep Bruxism (SB) Measures in Myofascial Temporomandibular Pain Disorder (TMD) Cases and Controls. PSG Measure Controls (n=37) Cases (n=86) p value Mean (SD) Median Mean (SD) Median RMMA episodes (per hour) 2.2 (1.9) 1.4 2.1 (2.0) 1.4 .77 Duration (all episodes, seconds) 72.2 (71.4) 48.9 66.7 (74.3) 41.2 .3 Duration of interepisode intervals (seconds) 1192.0 (1972.0) 393 1137.7 (1975.8) 281 .51 Duration of interepisode intervals (rank) 827.4 (482.4) 864 806.7 (463.5) 793 .44 PSG Measure Controls (n=37) Cases (n=86) p value Mean (SD) Median Mean (SD) Median RMMA episodes (per hour) 2.2 (1.9) 1.4 2.1 (2.0) 1.4 .77 Duration (all episodes, seconds) 72.2 (71.4) 48.9 66.7 (74.3) 41.2 .3 Duration of interepisode intervals (seconds) 1192.0 (1972.0) 393 1137.7 (1975.8) 281 .51 Duration of interepisode intervals (rank) 827.4 (482.4) 864 806.7 (463.5) 793 .44 SD = standard deviation. Open in new tab DISCUSSION The aim of this study was to test, using sleep-laboratory PSG recordings, whether the IEI between SB events differ between patients with and without myofascial pain. According to the sports sciences literature, lack of adequate rest-time between muscle activities leads to muscle overloading and pain.26 However, the recommended RDC/SB mainly focus on two characteristics of jaw-muscle contraction: the number of bruxism episodes and bursts per hour of sleep. In this paper, we hypothesized that the same SB activity (ie, with the same number of episodes and bursts) can have a different effect on the jaw muscles depending on its distribution over the night. For example, patients with jaw-muscle pain can have their SB episodes distributed over the night in a skewed fashion, with the episodes mainly being present in a short period of time. This can lead to much more load and possible muscle injury than in cases where the SB episodes are more or less equally distributed over the night. To test this hypothesis, we compared the duration of IEI between SB events in individuals with jaw-muscle pain and pain-free controls. The most important finding of this study is that there is no significant difference between the case and control groups in the duration of IEI. The two groups also did not differ in the number of SB episodes per hour of sleep or in the duration of the SB episodes. The last findings echo the results reported by Raphael et al.,21 from whose data set we selected current participants with two or more SB episodes per hour of sleep. It is noteworthy that more control than case participants (80% vs. 69%; p = .045, two-tailed Fisher’s Exact test) were included in the present study from the data set of Raphael et. al.21 This is consistent with the earlier reported observation that TMD-pain patients have fewer SB episodes than non-TMD controls.21–23 The relationship between SB and TMD pain has been the topic of several studies over the last years.27 The exact nature of the relationship between these conditions, however, is still unknown, with two major but conflicting theories aiming to explain the association: the Vicious Cycle Theory and the Pain Adaptation Model.28,29 The Vicious Cycle Theory suggests that an initiating factor, such as SB, results in pain that reflexively leads to muscle spasm. In turn, this spasm leads to further pain and dysfunction, thus completing the loop. However, the evidence that supports the Vicious Cycle Theory is limited.30–32 The Pain Adaptation Model, on the other hand, suggests that muscle pain leads to a reduction in muscle activity, which protects the muscle system from further injury and promotes healing.33,34 This model is commonly considered as the most appropriate explanation for the effects of pain on muscle performance.35 Although neither the Vicious Cycle Theory nor the Pain Adaptation Model considers the temporal delays in “causality” that might exist between pain and muscle function, most of the studies on SB-TMD pain association that used PSG-based diagnostics also support the Pain Adaptation Model, providing evidence of negative or no association between the two. One PSG-based paper that supports the Vicious Cycle Theory21 used a single-night PSG, which is not in line with the RDC/SB that recommends two nights of PSG registrations and reported very high rates of SB episodes in both case and control groups, suggesting a unique sample. Interestingly, the results of our research are consistent neither with the Vicious Cycle Theory nor with the Pain Adaptation Model. Based on the Pain Adaptation Model, one could expect that patients with jaw-muscle pain would have less SB activity than the control group, whereas based on the Vicious Cycle Theory, a decrease in jaw-muscle activity is to be expected. However, in this research, we did not find any difference in SB activity between groups. This suggests that the association between pain and muscle function is not hardwired and that other factors than pain alone determine the motor outcome.34 Indeed, pain has a multidimensional nature with several characteristics: duration, intensity, location in the specific part of motor unit, and individual response to pain.36–39 Moreover, the jaw-muscle system has a functional and structural heterogeneity.40,41 It is likely that this complex biomechanical system adapts to pain in different ways in order to maintain required functional integrity and to protect itself from further injury. This adaptation may not necessarily lead to muscle performance decrement but rather to a redistribution of function to other uninjured units.42 The above reasoning suggests that other factors than SB may play a role in maintaining jaw-muscle pain. In 2006, a large-scale project entitled Orofacial Pain: Prospective Evaluation and Risk Assessment (OPPERA) started to identify the risk factors for TMD. This prospective cohort study evaluated 202 phenotypic risk factors from six domains: sociodemographic, general health status, pain sensitivity, cardiac autonomic function, and psychological and clinical orofacial characteristics.43 From these, the frequency of somatic symptoms, for example, a running nose, fatigue, and dizziness, was the strongest psychosocial predictor of TMD incidence.44 Smaller contributions were found for measures of psychological stress, anxiety, obsessive-compulsive feelings, pain-coping strategies, and sleep quality. Moreover, genetic associations were found, implicating six single-nucleotide polymorphisms as risk factors for chronic TMD. This emphasizes that TMD is a complex disorder which is caused by interplay of multiple genetic and environmental factors, and that a univariate association between SB and jaw-muscle pain does not represent the actual, far more complex situation. As suggested by the sports science literature, successful training must involve load but also must include adequate recovery periods.17 As a consequence of load, the athlete may experience acute feelings of fatigue or even pain. Followed by an adequate rest period, the acute fatigue results in a positive adaptation or improvement in performance. This is the basis of effective training programs. However, in case a disruption occurs between appropriate training load and adequate recovery, the athletes may develop a so-called nonfunctional overtraining which will lead to a decrease in performance, depression, and pain that may last for several weeks or months. Several confounding factors have been reported to contribute to nonfunctional overloading, such as inadequate diet, somatic symptoms (eg, upper respiratory tract infections), psychosocial distress (family or work related), and sleep disorders.17 This multifactorial explanatory model of overtraining syndrome seems to have much in common with the results of OPPERA study, which also suggests an interplay between similar factors in the etiology of TMD pain. The noticeable difference between the overloading in sports and the data we presented in the current study is the time scale at which the loading and recovery occurred. In sports science literature, the training which includes both loading and recovery periods takes days to weeks (if not months). However, when SB IEI were considered in the present research, the time intervals lasted only seconds to minutes. The load and recovery model from sports literature could be applied to SB in case variable activity would be present for days to weeks. This would probably provide a different outcome than reported in the present research. Unfortunately, the current study design, with two nights of SB registration, did not allow to study the “rest” intervals at a longer time scale. On the other hand, although the sport science literature does not provide evidence regarding the effect of short-lasting (eg, seconds) rest-time intervals on muscle injury, this evidence is available from experimental animal studies. Numerous experimental studies have shown that shorter rest-time interval (eg, seconds to minutes) between muscle contraction indeed lead to more muscle injury and dysfunction.16,45 Most participants evidenced a low rate of SB events. From our sample of 123 participants with two and more SB episodes per night, only 53 case participants (61.2%) and 23 controls (62.2 %) fulfilled minimal criteria for Sleep Bruxism Diagnosis46 but 18 (14.6% of the studied sample) had more than four SB episodes per hour of sleep, fulfilling the criteria for the high SB intensity group (Table 2). As the best test of the muscle overloading theory would be conducted among those with some critically high rate of SB, we compared inter-event intervals between cases and controls within each of the three SB intensity groups. Results were similar in each stratum (analysis not shown), but further work with extreme subjects may be necessary. Table 2 Distribution of participants Based on Sleep Bruxism (SB) Diagnostic Cutoff Criteria. SB intensity Controls Count (%) Cases Count (%) Total Count (%) Low (>1 and ≤2 episodes/hour) 9 (24.3) 20 (23.3) 29 (23.6) Moderate (>2 and ≤4 episodes/hour) 8 (21.6) 21 (24.4) 29 (23.6) High (>4 episodes/hour) 6 (16.2) 12 (14) 18 (14.6) SB intensity Controls Count (%) Cases Count (%) Total Count (%) Low (>1 and ≤2 episodes/hour) 9 (24.3) 20 (23.3) 29 (23.6) Moderate (>2 and ≤4 episodes/hour) 8 (21.6) 21 (24.4) 29 (23.6) High (>4 episodes/hour) 6 (16.2) 12 (14) 18 (14.6) Open in new tab Table 2 Distribution of participants Based on Sleep Bruxism (SB) Diagnostic Cutoff Criteria. SB intensity Controls Count (%) Cases Count (%) Total Count (%) Low (>1 and ≤2 episodes/hour) 9 (24.3) 20 (23.3) 29 (23.6) Moderate (>2 and ≤4 episodes/hour) 8 (21.6) 21 (24.4) 29 (23.6) High (>4 episodes/hour) 6 (16.2) 12 (14) 18 (14.6) SB intensity Controls Count (%) Cases Count (%) Total Count (%) Low (>1 and ≤2 episodes/hour) 9 (24.3) 20 (23.3) 29 (23.6) Moderate (>2 and ≤4 episodes/hour) 8 (21.6) 21 (24.4) 29 (23.6) High (>4 episodes/hour) 6 (16.2) 12 (14) 18 (14.6) Open in new tab Although the results of our study suggest that the duration of interepisode SB intervals is not the “key factor” in the explanatory model of TMD pain, the quality of “rest intervals” still may play a role in the etiology of myofascial TMD pain. Previously, the analysis of EMG activity occurring outside of defined SB and other motor events showed that the levels were significantly higher in myofascial TMD patients compared to non-TMD controls.47 These long-lasting periods of elevated EMG activity between SB episodes could play a role in inadequate muscle recovery and eventually lead to persistent jaw-muscle pain. This suggestion is in line with the modified stress-hyperactivity-pain theory proposed by Ohrbach and McCall,48 which focused on chronic low-grade hyperactivity. The fact that sleep background EMG activity was significantly higher for woman with myofascial TMD pain than for control woman rises a question about the thresholds used to define SB events. The background EMG activity is routinely used as the threshold to identify SB activity. One can speculate that using that threshold, which is significantly higher for pain participants than for controls, could introduce a bias. The higher threshold could lead to inclusion less SB episodes in the case group than in the control group, which may have compromised the outcomes of the study. The alternative for the threshold based on background EMG activity would be the one based on percentages from the maximum voluntary contractions (MVC), as proposed by Lavigne et al.3 However, using this threshold in myofascial TMD-pain participants may also introduce a bias. Participants with pain in their jaw muscles could try to avoid more pain during function and therefore do not express maximum bite force during the MVC recording. Further research is needed to establish the most reliable threshold for EMG activity during sleep, when participants with jaw-muscle pain are investigated. Given these limits, current data fail to support the idea that TMD pain can be explained by increasing number of SB episodes per hour of sleep or decreasing time between SB events. DISCLOSURE STATEMENT Dr. Lobbezoo received funding from the University of Amsterdam/New York University Visiting Guest Faculty Program 2014–2015, for his Visiting Professorship at New York University College of Dentistry. The original study (Raphael et al. 21) was funded in part by grant R01 DE018569 from the National Institutes of Health, Bethesda, MD. The authors declare no potential conflicts of interest with respect to the authorship and/or publication of this article. REFERENCES 1. Lobbezoo F , Ahlberg J , Glaros AG et al. Bruxism defined and graded: an international consensus . J Oral Rehabil . 2013 ; 40 ( 1 ): 2 – 4 . Google Scholar Crossref Search ADS PubMed WorldCat 2. Manfredini D , Restrepo C , Diaz-Serrano K , Winocur E , Lobbezoo F . Prevalence of sleep bruxism in children: a systematic review of the literature . J Oral Rehabil . 2013 ; 40 ( 8 ): 631 – 642 . Google Scholar Crossref Search ADS PubMed WorldCat 3. Lavigne GJ , Rompré PH , Montplaisir JY . Sleep bruxism: validity of clinical research diagnostic criteria in a controlled polysomnographic study . J Dent Res . 1996 ; 75 ( 1 ): 546 – 552 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Raphael KG , Santiago V , Lobbezoo F . Is bruxism a disorder or a behaviour? Rethinking the international consensus on defining and grading of bruxism . J Oral Rehabil . 2016 ; 43 ( 10 ): 791 – 798 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Maluly M , Andersen ML , Dal-Fabbro C et al. Polysomnographic study of the prevalence of sleep bruxism in a population sample . J Dent Res . 2013 ; 92 ( 7 Suppl ): 97S – 103S . Google Scholar Crossref Search ADS PubMed WorldCat 6. Glaros AG , Glass EG , McLaughlin L . Knowledge and beliefs of dentists regarding temporomandibular disorders and chronic pain . J Orofac Pain . 1994 ; 8 ( 2 ): 216 – 222 . Google Scholar PubMed WorldCat 7. Velly AM , Philippe P , Gornitsky M . Heterogeneity of temporomandibular disorders: cluster and case-control analyses . J Oral Rehabil . 2002 ; 29 ( 10 ): 969 – 979 . Google Scholar Crossref Search ADS PubMed WorldCat 8. Osterberg T , Carlsson GE . Relationship between symptoms of temporomandibular disorders and dental status, general health and psychosomatic factors in two cohorts of 70-year-old subjects . Gerodontology . 2007 ; 24 ( 3 ): 129 – 135 . Google Scholar Crossref Search ADS PubMed WorldCat 9. McNeill C , Mohl ND , Rugh JD , Tanaka TT . Temporomandibular disorders: diagnosis, management, education, and research . J Am Dent Assoc . 1990 ; 120 ( 3 ): 253, 255, 257 passim . Google Scholar Crossref Search ADS PubMed WorldCat 10. Dworkin SF , LeResche L . Research diagnostic criteria for temporomandibular disorders: review, criteria, examinations and specifications, critique . J Craniomandib Disord . 1992 ; 6 ( 4 ): 301 – 355 . Google Scholar PubMed WorldCat 11. Okeson J. Orofacial pain: guidelines for classification, assessment, and management . 3rd ed. Chicago : Quintessence Publishing ; 1996 Google Preview WorldCat COPAC 12. Schiffman E , Ohrbach R , Truelove E et al. ; International RDC/TMD Consortium Network, International association for Dental Research; Orofacial Pain Special Interest Group, International Association for the Study of Pain . Diagnostic Criteria for Temporomandibular Disorders (DC/TMD) for Clinical and Research Applications: recommendations of the International RDC/TMD Consortium Network and Orofacial Pain Special Interest Group . J Oral Facial Pain Headache . 2014 ; 28 ( 1 ): 6 – 27 . Google Scholar Crossref Search ADS PubMed WorldCat 13. Kreher JB , Schwartz JB . Overtraining syndrome: a practical guide . Sports Health . 2012 ; 4 ( 2 ): 128 – 138 . Google Scholar Crossref Search ADS PubMed WorldCat 14. Rashedi E , Nussbaum MA . Cycle time influences the development of muscle fatigue at low to moderate levels of intermittent muscle contraction . J Electromyogr Kinesiol . 2016 ; 28 : 37 – 45 . Google Scholar Crossref Search ADS PubMed WorldCat 15. Lieber RL , Fridén J . Muscle damage is not a function of muscle force but active muscle strain . J Appl Physiol (1985) . 1993 ; 74 ( 2 ): 520 – 526 . Google Scholar PubMed WorldCat 16. Cutlip RG , Geronilla KB , Baker BA et al. Impact of stretch-shortening cycle rest interval on in vivo muscle performance . Med Sci Sports Exerc . 2005 ; 37 ( 8 ): 1345 – 1355 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Meeusen R , Duclos M , Foster C et al. ; European College of Sport Science; American College of Sports Medicine . Prevention, diagnosis, and treatment of the overtraining syndrome: joint consensus statement of the European College of Sport Science and the American College of Sports Medicine . Med Sci Sports Exerc . 2013 ; 45 ( 1 ): 186 – 205 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Camparis CM , Formigoni G , Teixeira MJ , Bittencourt LR , Tufik S , de Siqueira JT . Sleep bruxism and temporomandibular disorder: Clinical and polysomnographic evaluation . Arch Oral Biol . 2006 ; 51 ( 9 ): 721 – 728 . Google Scholar Crossref Search ADS PubMed WorldCat 19. Smith MT , Wickwire EM , Grace EG et al. Sleep disorders and their association with laboratory pain sensitivity in temporomandibular joint disorder . Sleep . 2009 ; 32 ( 6 ): 779 – 790 . Google Scholar Crossref Search ADS PubMed WorldCat 20. Rossetti LM , Pereira de Araujo Cdos R , Rossetti PH , Conti PC . Association between rhythmic masticatory muscle activity during sleep and masticatory myofascial pain: a polysomnographic study . J Orofac Pain . 2008 ; 22 ( 3 ): 190 – 200 . Google Scholar PubMed WorldCat 21. Raphael KG , Sirois DA , Janal MN et al. Sleep bruxism and myofascial temporomandibular disorders: a laboratory-based polysomnographic investigation . J Am Dent Assoc . 2012 ; 143 ( 11 ): 1223 – 1231 . Google Scholar Crossref Search ADS PubMed WorldCat 22. Lavigne GJ , Rompré PH , Montplaisir JY , Lobbezoo F . Motor activity in sleep bruxism with concomitant jaw muscle pain. A retrospective pilot study . Eur J Oral Sci . 1997 ; 105 ( 1 ): 92 – 95 . Google Scholar Crossref Search ADS PubMed WorldCat 23. Rompré PH , Daigle-Landry D , Guitard F , Montplaisir JY , Lavigne GJ . Identification of a sleep bruxism subgroup with a higher risk of pain . J Dent Res . 2007 ; 86 ( 9 ): 837 – 842 . Google Scholar Crossref Search ADS PubMed WorldCat 24. LeResche L . Epidemiology of temporomandibular disorders: implications for the investigation of etiologic factors . Crit Rev Oral Biol Med . 1997 ; 8 ( 3 ): 291 – 305 . Google Scholar Crossref Search ADS PubMed WorldCat 25. Conover WJ , Iman RL . Rank transformations as a bridge between parametric and nonparametric statistic . Am Stat . 1981 ; 35 ( 3 ) 124 – 129 . WorldCat 26. Meeusen R , Duclos M , Gleeson M , Rietjens G , Steinacker J , Urhausen A . Prevention, diagnosis and treatment of the overtraining syndrome—ECSS position statement “task force” . Eur J Sport Sci . 2006 ; 6 ( 1 ): 1 – 14 . Google Scholar Crossref Search ADS WorldCat 27. Manfredini D , Lobbezoo F . Relationship between bruxism and temporomandibular disorders: a systematic review of literature from 1998 to 2008 . Oral Surg Oral Med Oral Pathol Oral Radiol Endod . 2010 ; 109 ( 6 ): e26 – e50 . Google Scholar Crossref Search ADS PubMed WorldCat 28. Travell JG , Rinzler S , Herman M . Pain and disability of the shoulder and arm. Treatment by intramuscular infiltration with procaine hydrochloride . J Am Med Assoc . 1942 ; 120 : 417 – 422 . Google Scholar Crossref Search ADS WorldCat 29. Lund JP , Donga R , Widmer CG , Stohler CS . The pain-adaptation model: a discussion of the relationship between chronic musculoskeletal pain and motor activity . Can J Physiol Pharmacol . 1991 ; 69 ( 5 ): 683 – 694 . Google Scholar Crossref Search ADS PubMed WorldCat 30. Stohler CS , Zhang X , Lund JP . The effect of experimental jaw muscle pain on postural muscle activity . Pain . 1996 ; 66 ( 2-3 ): 215 – 221 . Google Scholar Crossref Search ADS PubMed WorldCat 31. Matre DA , Sinkjaer T , Svensson P , Arendt-Nielsen L . Experimental muscle pain increases the human stretch reflex . Pain . 1998 ; 75 ( 2-3 ): 331 – 339 . Google Scholar Crossref Search ADS PubMed WorldCat 32. Svensson P , De Laat A , Graven-Nielsen T , Arendt-Nielsen L . Experimental jaw-muscle pain does not change heteronymous H-reflexes in the human temporalis muscle . Exp Brain Res . 1998 ; 121 ( 3 ): 311 – 318 . Google Scholar Crossref Search ADS PubMed WorldCat 33. Lund JP , Lavigne G , Dubner R , Sessle B. Orofacial Pain: From Basic Science to Clinical Management . Chicago : Quintessence ; 2001 . Google Preview WorldCat COPAC 34. Janal MN , Raphael KG , Klausner J , Teaford M . The role of tooth-grinding in the maintenance of myofascial face pain: a test of alternate models . Pain Med . 2007 ; 8 ( 6 ): 486 – 496 . Google Scholar Crossref Search ADS PubMed WorldCat 35. Murray GM , Peck CC . Orofacial pain and jaw muscle activity: a new model . J Orofac Pain . 2007 ; 21 ( 4 ): 263 – 78; discussion 279 . Google Scholar PubMed WorldCat 36. Svensson P , Macaluso GM , De Laat A , Wang K . Effects of local and remote muscle pain on human jaw reflexes evoked by fast stretches at different clenching levels . Exp Brain Res . 2001 ; 139 ( 4 ): 495 – 502 . Google Scholar Crossref Search ADS PubMed WorldCat 37. Jensen MP , Nielson WR , Kerns RD . Toward the development of a motivational model of pain self-management . J Pain . 2003 ; 4 ( 9 ): 477 – 492 . Google Scholar Crossref Search ADS PubMed WorldCat 38. Viane I , Crombez G , Eccleston C , Devulder J , De Corte W . Acceptance of the unpleasant reality of chronic pain: effects upon attention to pain and engagement with daily activities . Pain . 2004 ; 112 ( 3 ): 282 – 288 . Google Scholar Crossref Search ADS PubMed WorldCat 39. Bodéré C , Téa SH , Giroux-Metges MA , Woda A . Activity of masticatory muscles in subjects with different orofacial pain conditions . Pain . 2005 ; 116 ( 1-2 ): 33 – 41 . Google Scholar Crossref Search ADS PubMed WorldCat 40. Hannam AG , McMillan AS . Internal organization in the human jaw muscles . Crit Rev Oral Biol Med . 1994 ; 5 ( 1 ): 55 – 89 . Google Scholar Crossref Search ADS PubMed WorldCat 41. van Eijden TM , Turkawski SJ . Morphology and physiology of masticatory muscle motor units . Crit Rev Oral Biol Med . 2001 ; 12 ( 1 ): 76 – 91 . Google Scholar Crossref Search ADS PubMed WorldCat 42. Westgaard RH , de Luca CJ . Motor unit substitution in long-duration contractions of the human trapezius muscle . J Neurophysiol . 1999 ; 82 ( 1 ): 501 – 504 . Google Scholar PubMed WorldCat 43. Slade GD , Ohrbach R , Greenspan JD et al. Painful temporomandibular disorder: decade of discovery from OPPERA studies . J Dent Res . 2016 ; 95 ( 10 ): 1084 – 1092 . Google Scholar Crossref Search ADS PubMed WorldCat 44. Fillingim RB , Ohrbach R , Greenspan JD et al. Psychological factors associated with development of TMD: the OPPERA prospective cohort study . J Pain . 2013 ; 14 ( 12 Suppl): 75 – 90 . Google Scholar Crossref Search ADS WorldCat 45. Stauber WT , Willems ME . Prevention of histopathologic changes from 30 repeated stretches of active rat skeletal muscles by long inter-stretch rest times . Eur J Appl Physiol . 2002 ; 88 ( 1-2 ): 94 – 99 . Google Scholar Crossref Search ADS PubMed WorldCat 46. Rompré PH , Daigle-Landry D , Guitard F , Montplaisir JY , Lavigne GJ . Identification of a sleep bruxism subgroup with a higher risk of pain . J Dent Res . 2007 ; 86 ( 9 ): 837 – 842 . Google Scholar Crossref Search ADS PubMed WorldCat 47. Raphael KG , Janal MN , Sirois DA et al. Masticatory muscle sleep background electromyographic activity is elevated in myofascial temporomandibular disorder patients . J Oral Rehabil . 2013 ; 40 ( 12 ): 883 – 891 . Google Scholar Crossref Search ADS PubMed WorldCat 48. Ohrbach R , McCall WD Jr. The stress-hyperactivity-pain theory of myogenic pain. Proposal for a revised theory . Pain Forum . 1996 ; 5 : 51 – 66 . Google Scholar Crossref Search ADS WorldCat Author notes Address correspondence to: K. Muzalev, DDS, MSc, Department of Oral Kinesiology, Academic Centre for Dentistry Amsterdam (ACTA), Gustav Mahlerlaan 3004, 1081 LA Amsterdam, The Netherlands. Telephone: 31 20 59 80378; Fax: 31 20 59 80333; Email: [email protected] © Sleep Research Society 2017. Published by Oxford University Press on behalf of the Sleep Research Society. All rights reserved. For permissions, please e-mail [email protected].

Hennies,, Nora;Lambon Ralph, Matthew, A;Durrant, Simon, J;Cousins, James, N;Lewis, Penelope, A

doi: 10.1093/sleep/zsx102pmid: 28821209

Abstract Study Objectives Extracting regularities from stimuli in our environment and generalizing these to new situations are fundamental processes in human cognition. Sleep has been shown to enhance these processes, possibly by facilitating reactivation-triggered memory reorganization. Here, we assessed whether cued reactivation during slow wave sleep (SWS) promotes the beneficial effect of sleep on abstraction of statistical regularities. Methods We used an auditory statistical learning task, in which the benefit of sleep has been firmly established. Participants were exposed to a probabilistically determined sequence of tones and subsequently tested for recognition of novel short sequences adhering to this same statistical pattern in both immediate and delayed recall sessions. In different groups, the exposure stream was replayed during SWS in the night between the recall sessions (SWS-replay group), in wake just before sleep (presleep replay group), or not at all (control group). Results Surprisingly, participants who received replay in sleep performed worse in the delayed recall session than the control and the presleep replay group. They also failed to show the association between SWS and task performance that has been observed in previous studies and was present in the controls. Importantly, sleep structure and sleep quality did not differ between groups, suggesting that replay during SWS did not impair sleep but rather disrupted or interfered with sleep-dependent mechanisms that underlie the extraction of the statistical pattern. Conclusions These findings raise important questions about the scope of cued memory reactivation and the mechanisms that underlie sleep-related generalization. sleep and memory, abstraction, memory reactivation, statistical learning Statement of Significance We demonstrate that experimental reactivation of memories during sleep can interfere with memory consolidation, leading to lower level of statistical knowledge the next day. Furthermore, such reactivation disrupts the otherwise systematic relationship between time spent in slow wave sleep and the degree of statistical abstraction. These findings are significant in that they provide an initial exploration of how memory replay in sleep interacts with abstraction processes, suggesting that mere reactivation may not always be the most useful way to process memories. INTRODUCTION Extracting statistical regularities from our environment, integrating them across different modalities, and generalizing to new exemplars or situations are fundamental processes in the formation of knowledge.1,2 Accumulating evidence suggests that these processes of statistical learning and integrative processing are facilitated by sleep.3–10 Specifically, sleep has been shown to promote the emergence of hidden rules or underlying patterns,3–5,7,8 to transfer them across modality boundaries,11 to strengthen connections between distinct elements,6,9,12 and to facilitate the integration of new information with pre-existing knowledge.13–15 The underlying mechanisms, however, are still unclear. One hypothesis is that abstraction is facilitated by sleep through temporally overlapping reactivation of individual memories that share common elements, leading to a strengthening of the overlapping parts.16 During sleep, memories are spontaneously reactivated,17–21 leading to memory improvement.22 Hippocampal reactivations are thought to be orchestrated at a neocortical level by slow neural oscillations during slow wave sleep (SWS).23–26 According to this idea, the depolarizing up-phases of the slow oscillations drive the formation of spindle-ripple events. Spindle-ripple events grouped during slow oscillations may play a key role in hippocampal-neocortical dialog during sleep,23,26–28 which may also underlie processes of abstraction and generalization.16 Spontaneously occurring reactivations during sleep can be manipulated by targeted memory reactivation (TMR).29 During TMR, memory cues that are associated with a prior learning episode are represented during post-learning sleep and thought to bias or trigger spontaneously occurring reactivations, thereby manipulating the consolidation process.29,30 In a series of recent studies, TMR has been successfully applied during SWS to increase the benefit of sleep on procedural and declarative memory.31–36 Whether TMR can be used in a similar way to promote memory reorganization and processes of abstraction and generalization is largely unknown. The first evidence in support of this hypothesis was provided in a recent study by Batterink et al.37 who showed that auditory cueing during sleep can influence grammatical rule learning and generalization. In this study, participants who were re-exposed to the language during sleep showed larger gains in grammatical generalization. These results are a first indication that memory reactivation during sleep may underlie abstraction and generalizing processes. The current study aimed to further explore whether cued memory reactivation during SWS can be used to enhance the beneficial effect of sleep on abstraction of statistical regularities. We used an auditory statistical learning task, for which the benefit of sleep and the association between abstraction performance and SWS is well established.7,8,11 In this task, participants are exposed to a sequence of tones that is probabilistically determined. Subsequently, participants are tested for recognition of short new sequences adhering to this same statistical structure. Performance in the recognition task therefore reflects whether the underlying statistical structure has been abstracted and can be applied to new exemplars. Additionally, a visual version of the recall test exists, which allows to test whether regularities have generalized beyond the modality of learning.11 Durrant et al.16 showed that abstraction of the underlying statistical structure was enhanced after consolidation across sleep and predicted by the time spent in SWS. SWS also predicted a trade-off between recruitment of medial temporal lobe and striatum during subsequent use of this knowledge8 and knowledge transfer to the visual modality.11 In short, abstraction and generalization performance in this task clearly benefits from sleep, particularly SWS, and therefore presents a good target for cued memory reactivation. To examine the effect of TMR during sleep in the present study, chunks of the probabilistic auditory sequence were represented during SWS in one group of participants. Another group, in which no TMR was applied, served as a control. To assess whether the effect of TMR was specific to sleep, the auditory sequence was represented during wakefulness directly before sleep in a third group of participants. Based on previous findings by Durrant et al.,8,11,16 we expected to find general overnight performance improvement, further enhanced by TMR during SWS. METHODS Participants Forty-two right-handed healthy volunteers participated in this experiment after informed consent was obtained, approved by the University of Manchester Research Ethics Committee. All had normal or corrected-to-normal vision, no hearing problems, and no history of neurological, psychiatric, or sleep disorders. Participants reported a regular sleep pattern over the month preceding the experiment and followed a standardized sleep schedule (11.00 pm–7.00 am) for 4 days prior to study begin. Participants were randomly assigned to one of three experimental groups (n = 14): SWS-replay (SWS-R) group (mean age: 22.86, standard deviation [SD]: 3.48, five females), control group (mean age: 22.50, SD: 3.44, six females) and presleep-replay (PS-R) group (mean age: 20.46, SD: 2.99, six females), which differed in terms of the replay. Stimuli The same stimuli were used in the current experiment as were used by Durrant et al.7,8 The stimuli were made up of sequences of pure tones (lasting 200 ms each) with seven different frequencies (261.63, 288.86, 318.93, 352.12, 388.77, 429.24, and 473.92 Hz), which were obtained by dividing an octave into seven equal intervals in pitch space. These intervals are not heard in Western tonal music and were used in order to avoid creating melodic fragments familiar to Western listeners. Tones were sampled with a frequency of 44.1 Hz, had a fixed amplitude, and were Gaussian modulated to avoid aliasing edge effects. There was a 20-ms gap between tones in a sequence. The stimuli involved one exposure stream and 168 short test streams. The exposure stream consisted of 1818 tones and lasted 6 minutes and 40 seconds. The test streams consisted of 18 tones lasting 3.96 seconds each. In addition to the auditory test streams, the stimuli also involved 84 visual test streams, in which a yellow circle moved from left to right across a black background on the computer screen along 18 defined locations. On a computer screen with a resolution of 1024 × 768 pixels, the circle started in a location 62 pixels from the left edge of the screen where it remained for 200 ms. It then disappeared for 20 ms and appeared in its next location 53 pixels to the right, where it again remained for 200 ms. This process continued for 18 horizontal locations, giving the appearance of a circle moving across the screen in a series of discrete events. The vertical position for each event could take one of seven evenly spaced vertical locations (−250 pixels, −166.67 pixels, −83.333 pixels, 0 pixels, 83.333 pixels, 166.67 pixels, 250 pixels, relative to the center of the screen). The seven vertical locations were chosen in analogy with the seven possible pitch height locations in the auditory sequence. Auditory and visual sequences both consisted of discrete events over time, varying equally in height and following the same timings. Participants were not informed of this analogy. The exposure stream, 42 of the auditory test streams and 42 of the visual test streams followed a probabilistic structure (structured condition). The probability for each potential transition between the current item (tone for the auditory stream and screen height for the visual stream) and the next item was determined by a transition matrix, forming a first-order Markov chain (Figure 1). In the transition matrix, each row contained one likely transition (p = .9, shown in white in Figure 1) and six unlikely transitions (p = .0167, shown in black in Figure 1). This ensured that a given item was followed by a particular item 90% of the time and by any of the other six items 10% of the time, making the sequences probabilistic. All seven items occurred overall with an equal probability (uniform zero-order transition), assuring that participants had to acquire sequence knowledge rather than just information on how frequently individual items occurred. The other half of the auditory and visual test streams (42 each) were generated randomly, with an equal probability for each tone/height at every position in the sequence (unstructured condition). As all test streams of the structured condition had the same probabilistic structure as the exposure stream, those test streams were considered as similar to the exposure stream, whereas the test streams of the unstructured condition were not similar to the exposure stream. Figure 1 Open in new tabDownload slide Generation of structured and unstructured sequences. (A) Transition matrix for the exposure stream and structured test sequences. Rows index the last tone that has occurred and columns tones that could occur next. The probability for each transition is reflected in the color with white = 0.9 and black = 0.0167. Tones occur overall with an equal frequency, ensuring that this cannot provide additional structural information. (B) Examples of a structured test stream and an unstructured test stream. Each test stream consisted of 18 tones. The sequence of structured test streams was determined based on the transition matrix, with the constraint that the number of high probability transitions was between 10 and 16. The sequence of unstructured test streams was generated randomly with an equal probability of 0.143 for each transition, resulting in four high probability transitions in this particular case. Low probability transitions are indicated in red. (C) Analogous to the auditory test streams, the visual test streams were sequences of a yellow circle moving from left to right across a black background. The circle started in the left edge of the screen and appeared at 18 distinct horizontal locations (indicated by vertical lines), giving the impression that it was moving from left to right. The vertical position could take one of seven evenly spaced locations, analogous to the seven pitch heights of the auditory streams. Structured and unstructured visual test streams were created analogous to auditory test streams. Figure 1 Open in new tabDownload slide Generation of structured and unstructured sequences. (A) Transition matrix for the exposure stream and structured test sequences. Rows index the last tone that has occurred and columns tones that could occur next. The probability for each transition is reflected in the color with white = 0.9 and black = 0.0167. Tones occur overall with an equal frequency, ensuring that this cannot provide additional structural information. (B) Examples of a structured test stream and an unstructured test stream. Each test stream consisted of 18 tones. The sequence of structured test streams was determined based on the transition matrix, with the constraint that the number of high probability transitions was between 10 and 16. The sequence of unstructured test streams was generated randomly with an equal probability of 0.143 for each transition, resulting in four high probability transitions in this particular case. Low probability transitions are indicated in red. (C) Analogous to the auditory test streams, the visual test streams were sequences of a yellow circle moving from left to right across a black background. The circle started in the left edge of the screen and appeared at 18 distinct horizontal locations (indicated by vertical lines), giving the impression that it was moving from left to right. The vertical position could take one of seven evenly spaced locations, analogous to the seven pitch heights of the auditory streams. Structured and unstructured visual test streams were created analogous to auditory test streams. For the replay, the exposure stream was divided into six fragments, each 66-second long. All six fragments were played twice in randomized order, with a 10-second gap between the fragments. This replay stream had a length of approximately 15 minutes. Experimental Task and Design All three experimental groups followed the same basic protocol (shown in Figure 2), which involved two experimental sessions, one in the evening at 9.00 pm ± 1 hour and one the following morning at 8.00 am ± 1 hour. All participants slept the night (from 11 pm to 7.00 am) between the two sessions in a bedroom in the Sleep Research Laboratory at the University of Manchester, and their sleep was monitored using polysomnography (PSG). Before each session, alertness was measured using the Stanford Sleepiness Scale38 (SSS) and the Karolinska Sleepiness Scale39 (KSS). Figure 2 Open in new tabDownload slide Experimental design. All groups encoded the exposure stream at 9.00 pm, followed by an immediate recall test session. Subsequently, participants completed a two-back working memory task. Overnight sleep was monitored with polysomnography. At 8.00 am, participants undertook a delayed recall session, followed by the visual recall task. In the SWS-replay (SWS-R) group the replay stream was presented during SWS and in the presleep-replay (PS-R) group during the two-back task. In the control group, no replay was done. Figure 2 Open in new tabDownload slide Experimental design. All groups encoded the exposure stream at 9.00 pm, followed by an immediate recall test session. Subsequently, participants completed a two-back working memory task. Overnight sleep was monitored with polysomnography. At 8.00 am, participants undertook a delayed recall session, followed by the visual recall task. In the SWS-replay (SWS-R) group the replay stream was presented during SWS and in the presleep-replay (PS-R) group during the two-back task. In the control group, no replay was done. Session 1 started with a learning phase lasting 8 minutes in which the auditory exposure stream was presented in order to familiarize participants with the transition probabilities. While the exposure stream was played, the term “Tone Stream” was presented in the middle of the computer screen, in order to focus participants’ attention toward the auditory stream. Participants were informed that an immediate and a delayed recall task would follow, but they were not informed about the underlying probabilistic structure of the auditory stream.8 Directly after the learning phase, participants conducted the immediate recall task which lasted 15 minutes. In this task, 42 structured and 42 unstructured auditory test streams were presented in randomized order. While a test stream was played, the instruction “Listen” as well as the number of the current trial out of the total number was presented in the middle of the computer screen (“Trial 18 of 84: Listen”). Subsequently, a 5-second response period (indicated by the phrase “Trial 18 of 84: Respond now”) followed, in which participants indicated whether or not the test stream sounded similar to the exposure stream, by pressing correspondingly labeled buttons (“familiar” or “unfamiliar”) on the computer keyboard. Participants were instructed to give their response as soon as they were sure. They were also informed in advance that half of the test streams were similar to the exposure stream and that the other half was unfamiliar. The immediate recall task was followed by a two-back task (adapted from the study by Kane et al.40) which lasted 15 minutes. This task is a demanding working memory task, and it was presented here to provide an opportunity to replay the auditory stream to the PS-R group outside of the focus of attention. To avoid differences in the experimental design, which could potentially influence memory performance, the two-back task was conducted in all groups. In each trial of the two-back task, one of eight phonologically distinct letters (B, F, K, H, M, Q, R, X) was displayed in the middle of the screen for 500 ms, followed by a blank 2500 ms interstimulus interval. For each trial, starting from trial number three, participants attempted to press one button (“Yes”) if the current letter matched the letter that appeared two items ago (B-f-b) and another button (“No”) if the two letters did not match. Participants were instructed to respond as quickly and accurately as possible. Participants could give their response as soon as the letter appeared on the screen and until the end of the interstimulus interval. Following a short practice run, participants performed six blocks of 48 trials each. Within each block each letter appeared six times, once as a target and five times as a foil. To prevent recognition based on perceptual features only, letters appeared randomly in either upper or lower case. Participants were asked to maintain their focus on the task. After each block, feedback was given on accuracy and participants were encouraged to try to improve their performance in the following block. While participants completed this task, brown noise was played. Afterward electrodes were attached for the purpose of overnight monitoring and participants went to bed at 11.00 pm. The following morning participants were woken up at 7.00 am. Session 2 started 30 minutes later with the delayed recall task. Trial structure and instructions of the delayed recall task were analogous to the immediate recall task, but 42 novel structured and 42 novel unstructured test streams were used. In the visual recall task, participants were presented with the 84 visual test sequences, in randomized order, and asked to indicate within the subsequent 5-second response period, whether or not the visual test streams were similar to the auditory exposure stream. Written instructions and the trial number out of the total number of trials were presented prior to each trial. In order to prevent participants from imagining the auditory analogue to the visual sequence, the seven different auditory tones were randomly played while the visual sequence was presented. Participants were instructed to ignore those tones and to use the visual information only in their judgment. Session 2 lasted 30 minutes (15 minutes for the auditory recall task and 15 minutes for the visual recall task). The three experimental groups differed in the presentation of the replay stream. In the SWS-R group, the replay stream was presented softly during the first two to three cycles of SWS. The replay started in the first extended period of SWS and was stopped immediately upon arousal or leaving SWS. In the PS-R group, the replay stream was presented during the two-back task. In both groups, the replay stream was played on PC speakers, with an approximate intensity of 48 dB, embedded in brown noise. In the control group no replay was done. In the PS-R group, the replay stream was presented during the two-back task, which served as distractor task and aimed to prevent rehearsal or active listening for a better comparison with the covert replay in the SWS-R group. Equipment The experimental tasks were realized using Cogent 2000 developed at the Functional Imaging Laboratory (University College, London), implemented using MATLAB© 7.5. Sound was generated using the onboard SoundMAX© digital audio chip and heard through a pair of Sennheiser © HD207 noise-cancelling headphones. Behavioral Data Analysis Data were analyzed with SPSS© 20.0 and MATLAB© 7.5. The sensitivity index d-prime (d’ = z-score[hits]—z-score[false alarms]) was calculated for the detection of the structured sequences for each session from the number of hits (correct identification of structured sequences) and the number of false alarms (incorrect identification of unstructured sequences as being structured). In cases where maximum hits or no false alarms occurred, half a trial was added or subtracted from the proportion correct when considering all test trials of the session (eg, 0.5/84) in order to avoid an infinite z-score.41 The difference between the performance on the delayed and the immediate recall session gave a measure of consolidation. A 2 × 3 mixed measures analysis of variance (ANOVA), with within-subject factor session (levels: immediate, delayed) and between-subject factor group (levels: SWS-R group, PS-R group, control group) was used on the d’ measures to investigate performance differences between groups. A one-way ANOVA on the d’ measures of the visual recall task was used to assess differences between groups. In all our results, we considered p < .05 as significant and all tests were two tailed. Significant effects were further explored with Bonferroni-corrected t tests. One-sample t tests were used to test if performance was above chance. For each round of the two-back task, the sensitivity index d-prime was calculated from the number of hits and false alarms. Differences between groups were assessed using a 3 × 6 mixed measures ANOVA with factors group and round (levels: round 1 to round 6). PSG Data Acquisition and Analysis An Embla© N7000 system was used for the EEG recording (200 Hz sampling rate). Six scalp electrodes were positioned using the international 10–20 system (F3, F4, C3, C4, O1, O2) with contralateral mastoid references. Two electrooculographic channels monitored eye movements and three chin electromyographic channels monitored muscle tone; a ground electrode was also attached. NuPrep© exfoliating agent was used to prepare the scalp, and electrodes were attached using EC2© electrogel. Impedance of less than 5 kΩ was verified at each electrode. Sleep data were visually scored using RemLogic© 1.1 software, in 30-second epochs, bandpass filtered between 0.3 and 35 Hz, by two trained sleep researchers according to the AASM Manual (American Academy of Sleep Medicine, Westchester, Illinois). Following the AASM recommendations, arousals were scored if there was an abrupt shift of EEG frequency including alpha, theta, and/or frequencies greater than 16 Hz (but not spindles) that lasted at least 3 seconds, with at least 10 seconds of stable sleep preceding the change, incorporating information from occipital and central derivations. The proportion of time spent in each sleep stage (stage 1, stage 2, SWS, REM) and the overall sleep duration were calculated. As previous studies showed that the amount of SWS predicted a performance increase from the immediate to the delayed recall session,7,8 we measured the correlation between SWS and overnight performance change for all experimental groups. A multivariate ANOVA (MANOVA) on the time spent in each sleep stage was used to examine group differences in the sleep structure. For spindle detection, raw EEG data of nonrapid eye movement sleep (non-REM: including stage 2 and SWS) were cleaned of artifacts and bandpass filtered (12–15 Hz) using a linear finite impulse response filter in EEG lab v.9.0. An automated detection algorithm,42 which counts amplitude fluctuations in the filtered time series, which exceed a predetermined threshold, as spindles, was used to determine the number of spindle events at each electrode. Reported results are averaged across frontal and central channels. Group differences were assessed using a one-way ANOVA. Power spectral density during SWS was analyzed on central (averaged across C3 and C4) and frontal (averaged across F3 and F4) channels using Welch’s method. This utilized a 4-second Hamming window length with 50 % overlap, focussing on frequency bands that are prominent during SWS, that is, slow oscillation (0.3–1 Hz), delta (1–4 Hz), and sigma/spindle (12-15Hz) bands. A MANOVA was used to assess group differences. As replay during sleep could have potentially caused sleep disruptions, resulting in reduced sleep quality, sleep quality was examined and compared between the three experimental groups. The following measures regarding sleep quality, which have been used in the literature43,44 were considered: time awake after sleep onset (in minutes), sleep efficiency (TST in percentage of the time from sleep onset until the last wake event), the number of transitions from one sleep stage to another, the transition index (number of transitions per hour of sleep), the number of awakenings (> 15 seconds), the awakening index (number of awakenings per hour of sleep), the number of arousals, and the arousal index (number of arousals per hour of sleep). A MANOVA examined group differences on these variables. To assess more subtle changes in sleep quality related to SWS, the sleep stage in which the replay was presented, the arousal index, the transition index, and the awakening index (number of events per time spent in SWS) for SWS only, were calculated and analyzed with a MANOVA. RESULTS Abstraction Performance and Association with Sleep Parameters Auditory Recall Task Performance in the auditory recall task served as measure for abstraction, as participants needed some knowledge of the auditory probabilistic pattern to correctly identify new sequences that followed the same pattern. To assess the effect of replay on abstraction performance, we were particularly interested in how the overnight change in performance differed between groups. The results are shown in Figure 3A. A 2 × 3 mixed measures ANOVA with factors session and group revealed no significant main effect of session, F(1,39) = 0.01, p = .93, no difference between groups, F(2,39) = 2.21, p = .12, but importantly a significant session × group interaction, F(2,39) = 5.42, p = .01. While, surprisingly, the SWS-R group showed a significant decrease in performance, at a Bonferroni-corrected α-level of 0.017, (mean S1: 1.13, standard error [SE]: 0.14, mean S2: 0.77, SE: 0.12), t(13) = 2.80, p = .015, the performance in the control group (mean S1: 1.25, SE: 0.11, mean S2: 1.34, SE: 0.15), t(13) = 0.74, p = .47, and the PS-R group (mean S1: 1.16, SE: 0.14, mean S2: 1.45, SE: 0.23), t(13) = 1.68, p = .12, did not change across sessions. As expected, performance between groups did not differ in the immediate recall session, t(26) ≤ 0.65, p ≥ .52, indicating that all groups had a comparable performance prior to the consolidation interval. After sleep, in the delayed recall session, the SWS-R group performed significantly worse than the control group, t(26) = 2.94, p = .007, and the PS-R group, t(26) = 2.66, p = .013, at a Bonferroni-corrected α-level of 0.017. Performance between control and PS-R group did not differ, t(26) = 0.41, p = .69. Importantly, performances in each session in each group were significantly greater than chance, t(13)≥ 6.27, p < .001, demonstrating that participants in all conditions were successful in conducting the task. In summary, against our expectations, these results showed selective overnight performance impairment for the SWS-R group. Figure 3 Open in new tabDownload slide Behavioral results. (A) Auditory recall task. While in the immediate recall session, there was no difference in the performance between groups, in the delayed recall session a difference between the groups emerged (assessed by one-way analysis of variance [ANOVA]). The SWS-replay group showed a significant decrease in correct recognition of structured and unstructured sequences after consolidation, whereas the control group and the presleep-replay group showed no change in performance. This group difference in the performance change across consolidation was significant in a 2 × 3 ANOVA with factors session and group. (B) Visual recall task. The control group and the presleep-replay group exhibited strong performance, while the SWS-replay group performed at chance. The difference in performance between the SWS-replay and the two other groups was significant. **p < .001, *p < .05, n.s.: p > .1. Figure 3 Open in new tabDownload slide Behavioral results. (A) Auditory recall task. While in the immediate recall session, there was no difference in the performance between groups, in the delayed recall session a difference between the groups emerged (assessed by one-way analysis of variance [ANOVA]). The SWS-replay group showed a significant decrease in correct recognition of structured and unstructured sequences after consolidation, whereas the control group and the presleep-replay group showed no change in performance. This group difference in the performance change across consolidation was significant in a 2 × 3 ANOVA with factors session and group. (B) Visual recall task. The control group and the presleep-replay group exhibited strong performance, while the SWS-replay group performed at chance. The difference in performance between the SWS-replay and the two other groups was significant. **p < .001, *p < .05, n.s.: p > .1. To investigate whether the performance impairment in the SWS-R group was driven by over-generalization or lack of abstraction, we also assessed the raw scores of hits, misses, false alarms, and correct rejections with 2 × 3 mixed measures ANOVAs with factors session and group (see Table 1 for the raw performance scores). Over-generalization would be reflected in an increase of false alarms, an increase in misses would suggest impairment in the abstraction of the statistical pattern. However, the results did not show any significant effects on the interaction between session and group (hits: F = 1.65, p = .21, correct rejections: F = 0.45, p = .64, misses: F = 1.00, p = .38, false alarms: F = 0.76, p = .48) suggesting that over-generalization as well as impaired abstraction may have contributed to the impaired performance in the SWS-R group. Table 1 Raw Performance Scores of the Immediate and Delayed Auditory Statistical Learning Task. Group Measurement Session 1 Session 2 p-value SWS-R Hits 25.5 ± 0.4 22.1 ± 0.6 .06 Correct rejections 30.2 ± 0.6 29.6 ± 0.6 .43 False alarms 10.0 ± 0.4 11.4 ± 0.4 .11 Misses 18.3 ± 0.4 20.4 ± 0.5 .20 Control Hits 24.4 ± 0.5 24.1 ± 0.5 .81 Correct rejections 33.3 ± 0.4 33.5 ± 0.7 .87 False alarms 8.5 ± 0.4 8.7 ± 0.7 .87 Misses 17.8 ± 0.5 17.7 ± 0.5 .94 PS-R Hits 27.0 ± 0.3 25.6 ± 0.5 .23 Correct rejections 32.6 ± 0.4 33.5 ± 0.6 .51 False alarms 8.9 ± 0.4 8.1 ± 0.6 .61 Misses 15.7 ± 0.3 16.0 ± 0.4 .73 Group Measurement Session 1 Session 2 p-value SWS-R Hits 25.5 ± 0.4 22.1 ± 0.6 .06 Correct rejections 30.2 ± 0.6 29.6 ± 0.6 .43 False alarms 10.0 ± 0.4 11.4 ± 0.4 .11 Misses 18.3 ± 0.4 20.4 ± 0.5 .20 Control Hits 24.4 ± 0.5 24.1 ± 0.5 .81 Correct rejections 33.3 ± 0.4 33.5 ± 0.7 .87 False alarms 8.5 ± 0.4 8.7 ± 0.7 .87 Misses 17.8 ± 0.5 17.7 ± 0.5 .94 PS-R Hits 27.0 ± 0.3 25.6 ± 0.5 .23 Correct rejections 32.6 ± 0.4 33.5 ± 0.6 .51 False alarms 8.9 ± 0.4 8.1 ± 0.6 .61 Misses 15.7 ± 0.3 16.0 ± 0.4 .73 Data are means ± SE; 84 items (42 old and 42 new) were presented in each session. p-values are from independent samples t-tests on the performance score of the two sessions. Open in new tab Table 1 Raw Performance Scores of the Immediate and Delayed Auditory Statistical Learning Task. Group Measurement Session 1 Session 2 p-value SWS-R Hits 25.5 ± 0.4 22.1 ± 0.6 .06 Correct rejections 30.2 ± 0.6 29.6 ± 0.6 .43 False alarms 10.0 ± 0.4 11.4 ± 0.4 .11 Misses 18.3 ± 0.4 20.4 ± 0.5 .20 Control Hits 24.4 ± 0.5 24.1 ± 0.5 .81 Correct rejections 33.3 ± 0.4 33.5 ± 0.7 .87 False alarms 8.5 ± 0.4 8.7 ± 0.7 .87 Misses 17.8 ± 0.5 17.7 ± 0.5 .94 PS-R Hits 27.0 ± 0.3 25.6 ± 0.5 .23 Correct rejections 32.6 ± 0.4 33.5 ± 0.6 .51 False alarms 8.9 ± 0.4 8.1 ± 0.6 .61 Misses 15.7 ± 0.3 16.0 ± 0.4 .73 Group Measurement Session 1 Session 2 p-value SWS-R Hits 25.5 ± 0.4 22.1 ± 0.6 .06 Correct rejections 30.2 ± 0.6 29.6 ± 0.6 .43 False alarms 10.0 ± 0.4 11.4 ± 0.4 .11 Misses 18.3 ± 0.4 20.4 ± 0.5 .20 Control Hits 24.4 ± 0.5 24.1 ± 0.5 .81 Correct rejections 33.3 ± 0.4 33.5 ± 0.7 .87 False alarms 8.5 ± 0.4 8.7 ± 0.7 .87 Misses 17.8 ± 0.5 17.7 ± 0.5 .94 PS-R Hits 27.0 ± 0.3 25.6 ± 0.5 .23 Correct rejections 32.6 ± 0.4 33.5 ± 0.6 .51 False alarms 8.9 ± 0.4 8.1 ± 0.6 .61 Misses 15.7 ± 0.3 16.0 ± 0.4 .73 Data are means ± SE; 84 items (42 old and 42 new) were presented in each session. p-values are from independent samples t-tests on the performance score of the two sessions. Open in new tab Visual Recall Task As the visual stimuli in the visual recall task shared no superficial characteristics with the auditory exposure stream, but only coincided with the underlying statistical pattern,8,11 this task could only be solved by abstracting the probabilistic pattern and generalizing it to the visual modality. Therefore, this task allowed us to measure the effect of replay on cross-modal generalization. The results are shown in Figure 3B. One participant of the SWS-R group did not complete the visual recall task due to a technical failure. A one-way ANOVA revealed a marginally significant difference in performance between the three experimental groups, F(2,38) = 3.10, p = .057. Planned post hoc comparisons, using a Bonferroni-corrected α-level of 0.025, showed that the SWS-R group performed significantly worse than the control group, t(25) = 2.49, p = .020, and marginally worse than the PS-R group, t(25) = 2.25, p = .033. At the group level, performance was above chance for the control group and the PS-R group (Bonferroni-corrected α-level of 0.017), t(13) ≥ 2.88, p ≤ .013, demonstrating that they were successful in learning the task. At the individual level, eight participants of the control group and seven participants of the PS-R group performed above chance. Performance of the SWS-R group did not exceed chance level at the group level, t(12) = 0.79, p = .45, only two participants performed above chance at the individual level. These results are in line with the results of the auditory recall task and showed selective failure of the SWS-R group to transfer knowledge about the statistical structure to the visual domain. Association Between Overnight Performance Change and SWS Previous studies using this paradigm showed that the amount of SWS predicted the behavioral performance change from the immediate to the delayed recall session.7,8 This association was also assessed in the current study. For one participant of the control group, no PSG data were available, due to technical difficulties during the sleep monitoring. Results of the PSG analysis of the remaining participants are presented in Table 2. The control group showed, as expected, a positive correlation between the proportion spend in SWS and the behavioral performance change from immediate to delayed recall, r(13) = 0.59, p = .035), shown in Figure 4. Participants with a large proportion of SWS showed an overnight improvement in performance, while participants with little SWS showed performance impairment. This correlation was specific to SWS; no other sleep stage (S1, S2, REM, TST) showed a significant correlation, r(14) ≤ 0.11, p ≥ 0.73. In the SWS-R group, we found no association between the change in performance and SWS, r(14) = −0.12, p = .67 or any other sleep stage, r(14) ≤ 0.32, p ≥ .27. While sleep-related processing seemed to facilitate the behavioral performance in the control group, our results suggest that TMR disturbed these mechanisms and thereby abolished the beneficial effect of sleep. Table 2 Polysomnography Results. Parameter SWS-R group Control group PS-R group p-value Total sleep time (minutes) 459 ± 19 433 ± 13 428 ± 14 .34 REM (%) 18.7 ± 1.3 18.7 ± 1.6 19.7 ± 1.1 .83 Stage 1 (%) 6.0 ± 1.1 5.7 ± 0.9 4.5 ± 0.8 .52 Stage 2 (%) 63.7 ± 1.4 61.8 ± 1.5 58.4 ± 1.8 .07 SWS (%) 11.6 ± 1.5 13.8 ± 1.5 17.4 ± 1.4 .03* Non-REM spindle density 0.77 ± 0.09 0.83 ± 0.10 0.91 ± 0.10 .62 Parameter SWS-R group Control group PS-R group p-value Total sleep time (minutes) 459 ± 19 433 ± 13 428 ± 14 .34 REM (%) 18.7 ± 1.3 18.7 ± 1.6 19.7 ± 1.1 .83 Stage 1 (%) 6.0 ± 1.1 5.7 ± 0.9 4.5 ± 0.8 .52 Stage 2 (%) 63.7 ± 1.4 61.8 ± 1.5 58.4 ± 1.8 .07 SWS (%) 11.6 ± 1.5 13.8 ± 1.5 17.4 ± 1.4 .03* Non-REM spindle density 0.77 ± 0.09 0.83 ± 0.10 0.91 ± 0.10 .62 non-REM = nonrapid eye movement sleep (including SWS and stage 2 sleep); PS-R = presleep replay; SWS-R = SWS replay. Spindle density is measured as number per minute. Data are means ± SE, p-values are from one-way ANOVAs. *Significance at p = .05 level. Open in new tab Table 2 Polysomnography Results. Parameter SWS-R group Control group PS-R group p-value Total sleep time (minutes) 459 ± 19 433 ± 13 428 ± 14 .34 REM (%) 18.7 ± 1.3 18.7 ± 1.6 19.7 ± 1.1 .83 Stage 1 (%) 6.0 ± 1.1 5.7 ± 0.9 4.5 ± 0.8 .52 Stage 2 (%) 63.7 ± 1.4 61.8 ± 1.5 58.4 ± 1.8 .07 SWS (%) 11.6 ± 1.5 13.8 ± 1.5 17.4 ± 1.4 .03* Non-REM spindle density 0.77 ± 0.09 0.83 ± 0.10 0.91 ± 0.10 .62 Parameter SWS-R group Control group PS-R group p-value Total sleep time (minutes) 459 ± 19 433 ± 13 428 ± 14 .34 REM (%) 18.7 ± 1.3 18.7 ± 1.6 19.7 ± 1.1 .83 Stage 1 (%) 6.0 ± 1.1 5.7 ± 0.9 4.5 ± 0.8 .52 Stage 2 (%) 63.7 ± 1.4 61.8 ± 1.5 58.4 ± 1.8 .07 SWS (%) 11.6 ± 1.5 13.8 ± 1.5 17.4 ± 1.4 .03* Non-REM spindle density 0.77 ± 0.09 0.83 ± 0.10 0.91 ± 0.10 .62 non-REM = nonrapid eye movement sleep (including SWS and stage 2 sleep); PS-R = presleep replay; SWS-R = SWS replay. Spindle density is measured as number per minute. Data are means ± SE, p-values are from one-way ANOVAs. *Significance at p = .05 level. Open in new tab Figure 4 Open in new tabDownload slide Relationship between slow wave sleep (SWS) and behavioral performance. The SWS-replay (SWS-R) group showed no association between SWS and the overnight performance change. In the control group, the improvement in task performance from the immediate- to the delayed-recall session was significantly correlated with the amount of SWS obtained. The presleep-replay (PS-R) group showed a significant but negative association. Participants with a high proportion of SWS showed a decrease in performance from the immediate- to the delayed-recall session while participants with a low proportion of SWS showed an improvement. Figure 4 Open in new tabDownload slide Relationship between slow wave sleep (SWS) and behavioral performance. The SWS-replay (SWS-R) group showed no association between SWS and the overnight performance change. In the control group, the improvement in task performance from the immediate- to the delayed-recall session was significantly correlated with the amount of SWS obtained. The presleep-replay (PS-R) group showed a significant but negative association. Participants with a high proportion of SWS showed a decrease in performance from the immediate- to the delayed-recall session while participants with a low proportion of SWS showed an improvement. The PS-R group showed a strong correlation between SWS and the behavioral performance change, but surprisingly, this association was negative, r(14) = −0.70, p = .005. Participants with a high proportion of SWS showed a decrease in performance from immediate to delayed recall, while participants with a low proportion of SWS showed an improvement. This correlation was specific to SWS; no other sleep stage showed a significant correlation, r(14) ≤ 0.43, p ≥ .12. No Difference in Sleep Quality Between Groups One explanation for the observed behavioral interference effect could be that the replay—independent of the cues themselves—disrupted sleep and thereby disrupted sleep-related consolidation processes or impaired sleep quality resulting in a less restorative function of sleep. To investigate this, we compared alertness, sleep structure, and sleep quality between groups. Alertness Differences in alertness between the three experimental groups were assessed with one-way ANOVAs on the average scores of the KSS and the SSS for each session. Groups did not differ in both sessions on both scales, F(2,39) ≤ 1.87, p ≥ .17, suggesting that there was no difference in alertness between groups in either session. The change in alertness between S1 and S2 did not differ between groups for either the KSS, F(2,39) = 0.90, p = .42, or the SSS, F(2,39) = 0.99, p = .38. As differences in alertness could also be reflected by differences in responses times, response times were assessed between the three experimental groups. To account for the fact, that response time is a sensitive measure on the statistical learning paradigm, that is, participants are faster on correct than incorrect trials,8 a 2 × 3 ANOVA with factors trial accuracy (levels: correct, false) and group was conducted on response times, separately for each session. Importantly, there was no significant main effect of group (immediate recall: F(2,39) = 0.21, p = .82; delayed recall: F(2,39) = 0.64, p = .54) and no significant interaction between group and accuracy (immediate recall: F(2,39) = 1.40, p = .26; delayed recall: F(2,39) = 1.82, p = .18) in either session, confirming that all groups had a similar pattern of response times. Together, these results suggest the observed group differences in performance were not due to differences in alertness. Sleep Structure A MANOVA on the proportions spent in each sleep stage was used to investigate differences between groups in sleep structure. The analysis showed no significant multivariate effect of group, F(8,72) = 1.42, p = .20. To assess more subtle differences between groups, univariate F tests were examined for each variable. The results are presented in Table 2. These analyses revealed a significant effect for SWS, F(2,38) = 4.05, p = .025. This effect was driven by significantly more SWS in the PS-R group compared to the SWS-R group, t(26) = 2.82, p = .009 and a trend toward more SWS of the PS-R group compared to the control group, t(25) = 1.78, p = .088. Importantly, however, there was no difference between the SWS-R group and the control group, t(25) = 1.02, p = .32, suggesting that these two groups were comparable in terms of the amount of SWS they obtained. Univariate analyses also revealed a marginal significant effect for stage 2 sleep (S2), F(2,38) = 2.92, p = .066. This was driven by significantly more S2 in the SWS-R group compared to the PS-R group, t(26) = 2.32, p = .029, but again there was no difference between the other groups t ≤ 1.44, p ≥ .16. To investigate more subtle differences specific to SWS, the sleep stage in which the replay took place, we assessed with a MANOVA differences in the power spectral density of SWS, in slow oscillation, delta and spindle frequency bands of central and frontal electrodes. The results are presented in Table 3. The multivariate group effect was not significant, F(12,68) = 1.09, p = .38. Planned univariate F tests on all dependent variables also showed no differences between groups, F(2,38) ≤ 2.05, p ≥ .14. In summary, these results suggest that the observed differences in abstraction performance between groups were not due to either differences in the overall sleep structure or SWS-specific structural changes caused by the replay. Table 3 Power Spectral Density (µV2/Hz) During Slow Wave Sleep. Frequency band SWS-R group Control group PS-R group p-value Central Slow oscillation (0.3–1 Hz) 989 ± 113 871 ± 101 1139 ± 128 .28 Delta (1–4 Hz) 201 ± 25 228 ± 26 272 ± 37 .24 Sigma (12–15 Hz) 2.46 ± 0.42 3.13 ± 0.52 3.07 ± 0.46 .55 Frontal Slow oscillation (0.3–1 Hz) 1043 ± 104 952 ± 107 1276 ± 133 .14 Delta (1–4 Hz) 265 ± 34 301 ± 35 386 ± 66 .20 Sigma (12–15 Hz) 2.59 ± 0.54 3.57 ± 0.61 3.45 ± 0.57 .43 Frequency band SWS-R group Control group PS-R group p-value Central Slow oscillation (0.3–1 Hz) 989 ± 113 871 ± 101 1139 ± 128 .28 Delta (1–4 Hz) 201 ± 25 228 ± 26 272 ± 37 .24 Sigma (12–15 Hz) 2.46 ± 0.42 3.13 ± 0.52 3.07 ± 0.46 .55 Frontal Slow oscillation (0.3–1 Hz) 1043 ± 104 952 ± 107 1276 ± 133 .14 Delta (1–4 Hz) 265 ± 34 301 ± 35 386 ± 66 .20 Sigma (12–15 Hz) 2.59 ± 0.54 3.57 ± 0.61 3.45 ± 0.57 .43 PS-R = presleep replay; SWS-R = SWS replay. Data are means ± SE. p-values are from one-way ANOVAs. Open in new tab Table 3 Power Spectral Density (µV2/Hz) During Slow Wave Sleep. Frequency band SWS-R group Control group PS-R group p-value Central Slow oscillation (0.3–1 Hz) 989 ± 113 871 ± 101 1139 ± 128 .28 Delta (1–4 Hz) 201 ± 25 228 ± 26 272 ± 37 .24 Sigma (12–15 Hz) 2.46 ± 0.42 3.13 ± 0.52 3.07 ± 0.46 .55 Frontal Slow oscillation (0.3–1 Hz) 1043 ± 104 952 ± 107 1276 ± 133 .14 Delta (1–4 Hz) 265 ± 34 301 ± 35 386 ± 66 .20 Sigma (12–15 Hz) 2.59 ± 0.54 3.57 ± 0.61 3.45 ± 0.57 .43 Frequency band SWS-R group Control group PS-R group p-value Central Slow oscillation (0.3–1 Hz) 989 ± 113 871 ± 101 1139 ± 128 .28 Delta (1–4 Hz) 201 ± 25 228 ± 26 272 ± 37 .24 Sigma (12–15 Hz) 2.46 ± 0.42 3.13 ± 0.52 3.07 ± 0.46 .55 Frontal Slow oscillation (0.3–1 Hz) 1043 ± 104 952 ± 107 1276 ± 133 .14 Delta (1–4 Hz) 265 ± 34 301 ± 35 386 ± 66 .20 Sigma (12–15 Hz) 2.59 ± 0.54 3.57 ± 0.61 3.45 ± 0.57 .43 PS-R = presleep replay; SWS-R = SWS replay. Data are means ± SE. p-values are from one-way ANOVAs. Open in new tab Sleep Quality As replay during sleep could potentially disrupt sleep and impair sleep quality, which could explain the observed performance decrease in the SWS-R group, sleep quality was assessed with respect to awakenings, arousals, and sleep-stage transitions. The results are summarized in Table 4. A MANOVA was used to examine group differences. The multivariate effect of group was not significant, F(16,64) = 0.99, p = .48. To assess more subtle differences between groups, univariate F tests were examined for each variable, however, none of the variables showed a significant effect, F ≤ 1.86, p ≥ .17 Furthermore, mean occipital alpha power during non-REM sleep, which can be an indicator of arousal or brief awakenings,32 was also assessed but did not differ between groups (SWS-R: 3.43 ± 0.42 µV2/Hz, PS-R: 3.74 ± 0.31 µV2/Hz, control: 4.73 ± 0.55 µV2/Hz; F(2,38) = 2.31, p = .11). In summary, sleep quality did not differ between groups, and it is therefore unlikely differences in sleep quality explain the performance impairment of the SWS-R group. Since replay was presented during SWS in the SWS-R group and hence might have affected this sleep stage in particular, we also examined sleep quality of SWS only. A MANOVA was used on the following dependent variables: SWS transition index, SWS awakening index, and SWS arousal index. Results are presented in Table 5. The multivariate group effect was not significant, F(6,74) = 1.75, p = .12. Planned univariate F tests on the dependent variables revealed a significant group difference in the SWS arousal index, F(2,38) = 0.64, p = .039, driven by a marginal significantly higher arousal index for the control group compared to the SWS-R group (t(25) = 2.00, p = .056) and the PS-R group (t = 2.01, p = .055). The SWS arousal index did not differ between the SWS-R group and the PS-R group, t(26) = 0.19, p = .85. Table 4 Overall Sleep Quality. Parameter SWS-R group Control group PS-R group p-value Time awake after sleep onset (minutes) 28.2 ± 4.2 33.1 ± 6.0 40.9 ± 10.7 .49 Sleep efficiency (%) 94.3 ± 0.7 92.9 ± 1.3 91.6 ± 2.1 .44 No. transitions 94.1 ± 8.0 100.9 ± 12.9 75.9 ± 6.9 .17 Transition index 11.7 ± 1.0 12.9 ± 1.6 9.8 ± 0.9 .17 No. awakenings 12.9 ± 1.7 17.2 ± 4.1 11.6 ± 1.5 .30 Awakening index 1.67 ± 0.19 2.42 ± 0.59 1.63 ± 0.22 .26 No. of arousals 81.1 ± 12.5 91.5 ± 8.2 76.9 ± 5.8 .53 Arousal index 10.7 ± 1.8 12.7 ± 1.1 10.8 ± 0.8 .48 Parameter SWS-R group Control group PS-R group p-value Time awake after sleep onset (minutes) 28.2 ± 4.2 33.1 ± 6.0 40.9 ± 10.7 .49 Sleep efficiency (%) 94.3 ± 0.7 92.9 ± 1.3 91.6 ± 2.1 .44 No. transitions 94.1 ± 8.0 100.9 ± 12.9 75.9 ± 6.9 .17 Transition index 11.7 ± 1.0 12.9 ± 1.6 9.8 ± 0.9 .17 No. awakenings 12.9 ± 1.7 17.2 ± 4.1 11.6 ± 1.5 .30 Awakening index 1.67 ± 0.19 2.42 ± 0.59 1.63 ± 0.22 .26 No. of arousals 81.1 ± 12.5 91.5 ± 8.2 76.9 ± 5.8 .53 Arousal index 10.7 ± 1.8 12.7 ± 1.1 10.8 ± 0.8 .48 Index = number of events per hour of sleep; PS-R = presleep replay; SWS-R = SWS replay; sleep efficiency = (time awake)/(sleep period) × 100. Data are means ± SE. p-values are from one-way ANOVAs. Open in new tab Table 4 Overall Sleep Quality. Parameter SWS-R group Control group PS-R group p-value Time awake after sleep onset (minutes) 28.2 ± 4.2 33.1 ± 6.0 40.9 ± 10.7 .49 Sleep efficiency (%) 94.3 ± 0.7 92.9 ± 1.3 91.6 ± 2.1 .44 No. transitions 94.1 ± 8.0 100.9 ± 12.9 75.9 ± 6.9 .17 Transition index 11.7 ± 1.0 12.9 ± 1.6 9.8 ± 0.9 .17 No. awakenings 12.9 ± 1.7 17.2 ± 4.1 11.6 ± 1.5 .30 Awakening index 1.67 ± 0.19 2.42 ± 0.59 1.63 ± 0.22 .26 No. of arousals 81.1 ± 12.5 91.5 ± 8.2 76.9 ± 5.8 .53 Arousal index 10.7 ± 1.8 12.7 ± 1.1 10.8 ± 0.8 .48 Parameter SWS-R group Control group PS-R group p-value Time awake after sleep onset (minutes) 28.2 ± 4.2 33.1 ± 6.0 40.9 ± 10.7 .49 Sleep efficiency (%) 94.3 ± 0.7 92.9 ± 1.3 91.6 ± 2.1 .44 No. transitions 94.1 ± 8.0 100.9 ± 12.9 75.9 ± 6.9 .17 Transition index 11.7 ± 1.0 12.9 ± 1.6 9.8 ± 0.9 .17 No. awakenings 12.9 ± 1.7 17.2 ± 4.1 11.6 ± 1.5 .30 Awakening index 1.67 ± 0.19 2.42 ± 0.59 1.63 ± 0.22 .26 No. of arousals 81.1 ± 12.5 91.5 ± 8.2 76.9 ± 5.8 .53 Arousal index 10.7 ± 1.8 12.7 ± 1.1 10.8 ± 0.8 .48 Index = number of events per hour of sleep; PS-R = presleep replay; SWS-R = SWS replay; sleep efficiency = (time awake)/(sleep period) × 100. Data are means ± SE. p-values are from one-way ANOVAs. Open in new tab Table 5 Sleep Quality of Slow Wave Sleep. Parameter SWS-R group Control group PS-R group p-value SWS transition index 0.35 ± 0.08 0.41 ± 0.10 0.22 ± 0.04 .17 SWS awakening index 0.012 ± 0.004 0.010 ± 0.007 0.010 ± 0.004 .96 SWS arousal index 0.054 ± 0.011 0.104 ± 0.025 0.053 ± 0.008 .04* Parameter SWS-R group Control group PS-R group p-value SWS transition index 0.35 ± 0.08 0.41 ± 0.10 0.22 ± 0.04 .17 SWS awakening index 0.012 ± 0.004 0.010 ± 0.007 0.010 ± 0.004 .96 SWS arousal index 0.054 ± 0.011 0.104 ± 0.025 0.053 ± 0.008 .04* PS-R = presleep-replay; SWS index = number of events during SWS divided by the total amount of SWS (in min); SWS-R = SWS replay. Data are means ± SE. p-values are from one-way ANOVAs. *Significance at p = .05 level. Open in new tab Table 5 Sleep Quality of Slow Wave Sleep. Parameter SWS-R group Control group PS-R group p-value SWS transition index 0.35 ± 0.08 0.41 ± 0.10 0.22 ± 0.04 .17 SWS awakening index 0.012 ± 0.004 0.010 ± 0.007 0.010 ± 0.004 .96 SWS arousal index 0.054 ± 0.011 0.104 ± 0.025 0.053 ± 0.008 .04* Parameter SWS-R group Control group PS-R group p-value SWS transition index 0.35 ± 0.08 0.41 ± 0.10 0.22 ± 0.04 .17 SWS awakening index 0.012 ± 0.004 0.010 ± 0.007 0.010 ± 0.004 .96 SWS arousal index 0.054 ± 0.011 0.104 ± 0.025 0.053 ± 0.008 .04* PS-R = presleep-replay; SWS index = number of events during SWS divided by the total amount of SWS (in min); SWS-R = SWS replay. Data are means ± SE. p-values are from one-way ANOVAs. *Significance at p = .05 level. Open in new tab Overall, none of the measures that we used to explore differences in alertness, sleep structure, and sleep quality showed a negative effect of replay. Hence, our findings suggest that the abstraction impairment of the SWS-R group was unlikely to be due to mechanical disruption of sleep but due to mechanistic interference of the replay cues with sleep-dependent memory processing. Details of the Replay In the SWS-R group, the replay was stopped immediately when participants left SWS or showed signs of an arousal. On average, the replay was stopped 1.7 ± 0.1 (SE) times per participant (minimum: 0 times, maximum: 4 times). For one participant of the SWS-R group, the six replay fragments were only played once. N-Back Task Data from one participant of the control group were lost due to a technical failure. A 3 × 6 mixed measures ANOVA was used to assess differences in performance on the two-back task between experimental groups. Importantly, we observed no difference between groups, F(2,38) = 1.80, p = .18 or group × round interaction, F(10,190) = 1.35, p = .21. The main effect of round was significant, F(5,190) = 6.70, p ≤ .001, with increasing performance across rounds. These results suggest that the PS-R group was not distracted by the presentation of the replay stream and focused on the two-back task. DISCUSSION In the current study, we explored whether TMR during SWS further enhances the beneficial effect of sleep on the extraction of statistical regularities in an auditory statistical learning paradigm. Surprisingly, we found that the beneficial effect of sleep on abstraction was abolished when the probabilistic auditory sequence was replayed during SWS. While the overnight performance change in the detection of structured and unstructured auditory sequences was positively correlated with the amount of SWS in the control group, the group that received the replay during SWS showed no such association and performance was impaired after sleep. This negative effect on task performance was specific to the replay during SWS. Therefore, these results suggest that sleep-dependent mechanisms, which mediate the abstraction of the underlying statistical pattern, were disrupted. That sleep facilitates processes of abstraction and generalization has been observed in a range of different tasks involving verbal concepts,6 probabilistic rules,3 number sequences,4 and grammar learning in infants.45 The beneficial effect of sleep on the extraction of auditory probabilistic sequences used in the current task was established by Durrant et al.7,8,11 They consistently reported that abstraction performance in this task improved after sleep and that the level of improvement was predicted by the amount of SWS obtained.7,8,11 Although we did not observe a performance improvement across sleep in any group, we replicated the association between SWS and performance change in the control group. In line with the previous studies, participants with a high proportion of SWS showed an improvement in performance across sleep, while participants with a low proportion of SWS showed impairment. These findings support the hypothesis that processes during SWS mediate the abstraction of statistical regularities. The benefit of sleep on memory consolidation and reorganization has been attributed to repeated reactivation of memory traces during SWS, driven by the hippocampus.19,21,46–49 Memory reactivations during SWS occur in temporal relationship with hippocampal sharp wave ripples and are thought to be orchestrated with the occurrence of thalamocortical spindles by the up-phases of slow oscillations.17,46,50 This temporal synchrony between memory reactivation and increased neocortical receptivity, induced by spindles, may enable hippocampal-neocortical information exchange and memory reorganization.23 Functional imaging results from Durrant et al.8 suggest that the underlying mechanisms of abstraction in the current task indeed involve SWS-mediated reorganization of the brain circuits that support memory. Specifically, Durrant et al.8 demonstrated that the overnight performance change was associated with a gradual shift from the hippocampal to the striatal memory system and that this change in the underlying neural substrates was predicted by the amount of nocturnal SWS. One hypothesis is that such sleep-related enhancements of abstraction and generalization result from the recurrent reactivation of memory elements that are shared between individual item memories.16,51 By strengthening the overlapping connections between separate memories, the “gist” emerges that enables generalization to new stimuli or situations.16,51 Based on this theory, we hypothesized that TMR would, by manipulating the occurrence of spontaneous reactivations,28 promote the abstraction process through a selective enhancement of the highly likely transitions. Surprisingly, however, we observed the opposite: TMR impaired task performance for both auditory and visual versions of the task and abolished the association with SWS. These findings suggest that representing the probabilistic sequence during SWS interfered with sleep-dependent memory processing. The auditory recall task can be solved either by abstracting transition statistics (ie, the probabilistic sequence) or by using episodic memory of concrete fragments of the exposure sequence.8,52,53 The visual recall task, however, can only be solved by applying knowledge about the probabilistic structure as the visual stimuli share no superficial characteristics with the auditory exposure stream. They only coincide with the underlying statistical pattern.8,11 Thus, presenting the stimuli in a different modality at testing ensured that explicit episodic memory for fragments could not aid test performance and extraction of the underlying probabilistic pattern was necessary. Importantly, we found that the replay-driven overnight impairment was also present in the visual recall task. While participants of the control and PS-R group performed well above chance in this task, clearly demonstrating some knowledge of the statistical structure, performance of the replay group was at chance level. Hence, the current findings suggest that the replay of the exposure stream interfered with the abstraction process and did not just impair episodic memory. From our study, no conclusion can be drawn in terms of the underlying mechanisms. We can only speculate that the probabilistic nature of the replay cue caused this interference. As we used the probabilistic sequence as memory cue during SWS, which is highly variable, the cue and the stored memory representations of the concrete fragments did not perfectly overlap. Therefore, the spontaneous reactivation of the concrete fragments might have been disrupted by the presentation of the partially overlapping memory cues, resulting in impaired performance and lack of association with SWS. Even though this is speculative, it raises important questions about the nature of TMR-cued reactivation. So far, in all studies that successfully applied TMR to enhance the sleep benefit on memory stabilization, only perfectly matching stimuli were used as reactivation cues. Cousins et al.,31,35 for example, used a fixed (not probabilistically determined) auditory sequence as memory cue during SWS (in the same sleep laboratory with the same set up in terms of volume, background noise, and replay procedure as the current study) and found enhanced sequence knowledge after sleep. A recent study by Batterink et al.,37 which provided the first evidence that TMR can manipulate grammatical generalization during sleep, also used memory cues that did not vary probabilistically. Based on these results, the question arises whether partially overlapping memory cues might have the potential to disrupt the beneficial effect of sleep on memory reorganization. Another explanation might be that due to the length of the replay fragments, processes necessary for stabilizing reactivated memory traces were disrupted. A recent study by Schreiner et al.54 showed that presenting additional auditory input within a period of 1500 ms after a memory cue completely blocked the beneficial effect of cueing during sleep on later recall. They suggested that during this sensitive period additional input disrupted neural and oscillatory processes critical for memory stabilization after reactivation during sleep. As in the current study, cues were presented in long blocks interfering effects might have occurred during these sensitive phases and blocked sleep-dependent processing as suggested by Schreiner et al.54 In the current study, the replay-related performance impairment was specific to the replay during sleep. The group who received the replay during wakefulness, directly before sleep, showed comparable performance to the control group. Surprisingly, however, the replay before sleep also had an interfering effect and reversed the association with SWS. Participants with a low proportion of SWS showed an overnight improvement in performance while participants with a high proportion of SWS got worse. These results may indicate that SWS in this group was associated with the consolidation of the wrong thing, such as for example unnecessary details like low probability transitions. As participants in this group received another presentation round of the exposure stream after the test phase, it is likely that they perceived the exposure stream differently and focused on other aspects compared to the initial encoding. Why this causes impairment, remains unclear. However, it is possible that it disrupted an abstraction process which had already begun after the initial training session. This group also showed a slight difference in the sleep structure with a higher proportion of SWS than the two other groups, which might also suggest that some different processes occurred. Importantly, however, participants performed well above chance in the visual recall task and showed, unlike the group who received the replay during SWS, no impairment in the auditory recall task, indicating that this group had abstracted the underlying statistical structure. Therefore, the catastrophic interference with the abstraction of the underlying statistical structure was specific to the replay during SWS. In terms of limitations, we cannot exclude the possibility that the two-back task caused a nonspecific interference effect, which might explain the lack of overnight performance improvement in the control group that was reported in previous studies.7,8 As the two-back task was performed by all experimental groups, it is, however, unlikely that this led to the observed differences between the groups or the unexpected behavior of the PS-R group. We should also note that, while the sample size of N = 14 is in line with prior studies of how sleep interacts with this statistical learning paradigm7,8,11 as well as the bulk of the literature on TMR using auditory cues,32,55–58 a greater sample size may nevertheless have provided additional confidence in the results. Hence, due to the small sample size of the experimental groups the results have to be interpreted with caution. Another limitation is the between-subject design. The observed negative effect of the replay on task performance could also be explained by a mechanical disruption of sleep independent of the memory cue itself, leading to a disruption of sleep-dependent consolidation processes. However, this is unlikely, since general sleep and SWS parameters, such as SWS amount, spectral power of frequency bands that are dominant during SWS, and SWS quality measures did not differ between the SWS-R group and the control group. Another possibility is that the replay caused a general impairment of sleep quality resulting in a less restorative function of sleep and increased tiredness, which could theoretically explain the impaired performance of the replay group. However, comparable sleep quality and alertness measures between groups suggest that it is highly unlikely that the replay impaired sleep quality and that differences in performance were due to differences in alertness. Overall, our findings suggest that the replay-related impairment was caused through a specific interference with the consolidation process and not through a mechanical disruption of sleep. In conclusion, the current results suggest that representing the probabilistic auditory sequence during SWS interfered with the abstraction process and therefore impaired subsequent performance in both auditory and visual recall tasks. These findings raise important questions about the scope and the underlying mechanisms of cued memory reactivation, which need to be addressed in further studies. FUNDING This work was supported by the Biotechnology and Biological Sciences Research Council: grant [BB/F01760X/1] and by the University of Manchester. DISCLOSURE STATEMENT None declared. REFERENCES 1. Lambon Ralph MA , Sage K , Jones RW , Mayberry EJ . Coherent concepts are computed in the anterior temporal lobes . Proc Natl Acad Sci U S A . 2010 ; 107 ( 6 ): 2717 – 2722 . Google Scholar Crossref Search ADS PubMed WorldCat 2. Rogers TT , McClelland JL. Semantic Cognition: A Parallel Distributed Processing Approach . Cambridge, MA : MIT Press ; 2004 . Google Preview WorldCat COPAC 3. Djonlagic I , Rosenfeld A , Shohamy D , Myers C , Gluck M , Stickgold R . Sleep enhances category learning . Learn Mem . 2009 ; 16 ( 12 ): 751 – 755 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Wagner U , Gais S , Haider H , Verleger R , Born J . Sleep inspires insight . Nature . 2004 ; 427 ( 6972 ): 352 – 355 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Fischer S , Drosopoulos S , Tsen J , Born J . Implicit learning—explicit knowing: a role for sleep in memory system interaction . J Cogn Neurosci . 2006 ; 18 ( 3 ): 311 – 319 . http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=pubmed&cmd=Retrieve&dopt=AbstractPlus&list_uids=16602193. 6. Lau H , Tucker MA , Fishbein W . Daytime napping: effects on human direct associative and relational memory . Neurobiol Learn Mem . 2010 ; 93 ( 4 ): 554 – 560 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Durrant SJ , Taylor C , Cairney S , Lewis PA . Sleep-dependent consolidation of statistical learning . Neuropsychologia . 2011 ; 49 ( 5 ): 1322 – 1331 . Google Scholar Crossref Search ADS PubMed WorldCat 8. Durrant SJ , Cairney SA , Lewis PA . Overnight consolidation aids the transfer of statistical knowledge from the medial temporal lobe to the striatum . Cereb Cortex . 2013 ; 23 ( 10 ): 2467 – 2478 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Ellenbogen JM , Hu PT , Payne JD , Titone D , Walker MP . Human relational memory requires time and sleep . Proc Natl Acad Sci U S A . 2007 ; 104 ( 18 ): 7723 – 7728 . Google Scholar Crossref Search ADS PubMed WorldCat 10. Gais S , Born J . Declarative memory consolidation: mechanisms acting during human sleep . Learn Mem . 2004 ; 11 ( 6 ): 679 – 685 . Google Scholar Crossref Search ADS PubMed WorldCat 11. Durrant SJ , Cairney SA , Lewis PA . Cross-modal transfer of statistical information benefits from sleep . Cortex . 2016 ; 78 : 85 – 99 . Google Scholar Crossref Search ADS PubMed WorldCat 12. Lau H , Alger SE , Fishbein W . Relational memory: a daytime nap facilitates the abstraction of general concepts . PLoS One . 2011 ; 6 ( 11 ): e27139 . Google Scholar Crossref Search ADS PubMed WorldCat 13. Tamminen J , Payne JD , Stickgold R , Wamsley EJ , Gaskell MG . Sleep spindle activity is associated with the integration of new memories and existing knowledge . J Neurosci . 2010 ; 30 ( 43 ): 14356 – 14360 . Google Scholar Crossref Search ADS PubMed WorldCat 14. Tamminen J , Lambon Ralph MA , Lewis PA . The role of sleep spindles and slow-wave activity in integrating new information in semantic memory . J Neurosci . 2013 ; 33 ( 39 ): 15376 – 15381 . Google Scholar Crossref Search ADS PubMed WorldCat 15. Cai DJ , Mednick SA , Harrison EM , Kanady JC , Mednick SC . REM, not incubation, improves creativity by priming associative networks . Proc Natl Acad Sci U S A . 2009 ; 106 ( 25 ): 10130 – 10134 . Google Scholar Crossref Search ADS PubMed WorldCat 16. Lewis PA , Durrant SJ . Overlapping memory replay during sleep builds cognitive schemata . Trends Cogn Sci . 2011 ; 15 ( 8 ): 343 – 351 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Kudrimoti HS , Barnes CA , McNaughton BL . Reactivation of hippocampal cell assemblies: effects of behavioral state, experience, and EEG dynamics . J Neurosci . 1999 ; 19 ( 10 ): 4090 – 4101 . http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=pubmed&cmd=Retrieve&dopt=AbstractPlus&list_uids=10234037. 18. Pavlides C , Winson J . Influences of hippocampal place cell firing in the awake state on the activity of these cells during subsequent sleep episodes . J Neurosci . 1989 ; 9 ( 8 ): 2907 – 2918 . http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=pubmed&cmd=Retrieve&dopt=AbstractPlus&list_uids=2769370. 19. Qin YL , McNaughton BL , Skaggs WE , Barnes CA . Memory reprocessing in corticocortical and hippocampocortical neuronal ensembles . Philos Trans R Soc Lond B Biol Sci . 1997 ; 352 ( 1360 ): 1525 – 1533 . Google Scholar Crossref Search ADS PubMed WorldCat 20. Sutherland GR , McNaughton B . Memory trace reactivation in hippocampal and neocortical neuronal ensembles . Curr Opin Neurobiol . 2000 ; 10 ( 2 ): 180 – 186 . Google Scholar Crossref Search ADS PubMed WorldCat 21. Wilson MA , McNaughton BL . Reactivation of hippocampal ensemble memories during sleep . Science . 1994 ; 265 ( 5172 ): 676 – 679 . Google Scholar Crossref Search ADS PubMed WorldCat 22. Dupret D , O’Neill J , Pleydell-Bouverie B , Csicsvari J . The reorganization and reactivation of hippocampal maps predict spatial memory performance . Nat Neurosci . 2010 ; 13 ( 8 ): 995 – 1002 . Google Scholar Crossref Search ADS PubMed WorldCat 23. Sirota A , Csicsvari J , Buhl D , Buzsáki G . Communication between neocortex and hippocampus during sleep in rodents . Proc Natl Acad Sci U S A . 2003 ; 100 ( 4 ): 2065 – 2069 . Google Scholar Crossref Search ADS PubMed WorldCat 24. Battaglia FP , Benchenane K , Sirota A , Pennartz CMA , Wiener SI . The hippocampus: hub of brain network communication for memory . Trends Cogn Sci (Regul Ed) . 2011 ; 15 ( 7 ): 310 – 318 . Google Scholar PubMed WorldCat 25. Contreras D , Destexhe A , Sejnowski TJ , Steriade M . Control of spatiotemporal coherence of a thalamic oscillation by corticothalamic feedback . Science . 1996 ; 274 ( 5288 ): 771 – 774 . http://www.ncbi.nlm.nih.gov/pubmed/8864114. Accessed October 18, 2014 . 26. Steriade M . Grouping of brain rhythms in corticothalamic systems . Neuroscience . 2006 ; 137 ( 4 ): 1087 – 1106 . Google Scholar Crossref Search ADS PubMed WorldCat 27. Mölle M , Born J . Slow oscillations orchestrating fast oscillations and memory consolidation . Prog Brain Res . 2011 ; 193 : 93 – 110 . Google Scholar Crossref Search ADS PubMed WorldCat 28. Oudiette D , Paller KA . Upgrading the sleeping brain with targeted memory reactivation . Trends Cogn Sci . 2013 ; 17 ( 3 ): 142 – 149 . Google Scholar Crossref Search ADS PubMed WorldCat 29. Bendor D , Wilson MA . Biasing the content of hippocampal replay during sleep . Nat Neurosci . 2012 ; 15 ( 10 ): 1439 – 1444 . Google Scholar Crossref Search ADS PubMed WorldCat 30. Dave AS , Margoliash D . Song replay during sleep and computational rules for sensorimotor vocal learning . Science . 2000 ; 290 ( 5492 ): 812 – 816 . http://www.ncbi.nlm.nih.gov/pubmed/11052946. Accessed November 15, 2014 . 31. Cousins JN , El-Deredy W , Parkes LM , Hennies N , Lewis PA . Cued memory reactivation during slow-wave sleep promotes explicit knowledge of a motor sequence . J Neurosci . 2014 ; 34 ( 48 ): 15870 – 15876 . Google Scholar Crossref Search ADS PubMed WorldCat 32. Rudoy JD , Voss JL , Westerberg CE , Paller KA . Strengthening individual memories by reactivating them during sleep . Science . 2009 ; 326 ( 5956 ): 1079 . Google Scholar Crossref Search ADS PubMed WorldCat 33. Rihm J , Diekelmann S , Born J , Rasch B . Reactivating memories during sleep by odors: odor specificity and associated changes in sleep oscillations . J Cogn Neurosci . 2014 : 1 – 13 . WorldCat 34. Antony JJWJ , Gobel EEW , O’Hare JJK , Reber PJ , Paller KA . Cued memory reactivation during sleep influences skill learning . Nat Neurosci . 2012 ; 15 ( 8 ): 1114 – 1116 . Google Scholar Crossref Search ADS PubMed WorldCat 35. Cousins JN , El-Deredy W , Parkes LM , Hennies N , Lewis PA . Cued reactivation of motor learning during sleep leads to overnight changes in functional brain activity and connectivity . PLoS Biol . 2016 ; 14 ( 5 ): e1002451 . Google Scholar Crossref Search ADS PubMed WorldCat 36. Schreiner T , Rasch B . Boosting vocabulary learning by verbal cueing during sleep . Cereb Cortex . 2014 : 1 – 11 . WorldCat 37. Batterink LJ , Paller KA . Sleep-based memory processing facilitates grammatical generalization: evidence from targeted memory reactivation . Brain Lang . 2017 ; 167 : 83 – 93 . Google Scholar Crossref Search ADS PubMed WorldCat 38. Hoddes E , Zarcone V , Smythe H , Phillips R , Dement WC . Quantification of sleepiness: a new approach . Psychophysiology . 1973 ; 10 ( 4 ): 431 – 436 . Google Scholar Crossref Search ADS PubMed WorldCat 39. Glenville M , Broughton R . Reliability of the Stanford sleepiness scale compared to short duration performance tests and the Wilkinson auditory vigilance task . Adv Biosci . 21 : 235 – 244 . http://www.ncbi.nlm.nih.gov/pubmed/755721. Accessed November 28, 2014 . 40. Kane MJ, Conway ARA, Miura TK, Colflesh, GJH. Working memory, attention control, and the n-back task: a question of construct validity. J Exp Psychol-Learn Mem Cogn , 2007, 33(3), 615–622. doi: 10.1037/0278-7393.33.3.615 41. Stanislaw H , Todorov N . Calculation of signal detection theory measures . Behav Res Methods Instrum Comput . 1999 ; 31 ( 1 ): 137 – 149 . http://www.ncbi.nlm.nih.gov/pubmed/10495845. Accessed October 22, 2016 . 42. Ferrarelli F , Huber R , Peterson MJ et al. Reduced sleep spindle activity in schizophrenia patients . Am J Psychiatry . 2007 ; 164 ( 3 ): 483 – 492 . Google Scholar Crossref Search ADS PubMed WorldCat 43. Buysse DJ , Reynolds CF , Monk TH , Berman SR , Kupfer DJ . The Pittsburgh sleep quality index: a new instrument for psychiatric practice and research . Psychiatry Res . 1989 ; 28 ( 2 ): 193 – 213 . http://www.ncbi.nlm.nih.gov/pubmed/2748771. Accessed November 17, 2014 . 44. Conway MA , Gardiner JM , Perfect TJ , Anderson SJ , Cohen G . Changes in memory awareness during learning: the acquisition of knowledge by psychology undergraduates . J Exp Psychol Gen . 1997 ; 126 ( 4 ): 393 – 413 . http://psycnet.apa.org/journals/xge/126/4/393/. 45. Gómez RL , Bootzin RR , Nadel L . Naps promote abstraction in language-learning infants . Psychol Sci . 2006 ; 17 ( 8 ): 670 – 674 . Google Scholar Crossref Search ADS PubMed WorldCat 46. Buzsáki G . The hippocampo-neocortical dialogue . Cereb Cortex . 1996 ; 6 ( 2 ): 81 – 92 . Google Scholar Crossref Search ADS PubMed WorldCat 47. Maquet P , Laureys S , Peigneux P et al. Experience-dependent changes in cerebral activation during human REM sleep . Nat Neurosci . 2000 ; 3 ( 8 ): 831 – 836 . Google Scholar Crossref Search ADS PubMed WorldCat 48. Diekelmann S , Born J . The memory function of sleep . Nat Rev Neurosci . 2010 ; 11 ( 2 ): 114 – 126 . Google Scholar PubMed WorldCat 49. Rasch B , Born J . About sleep’s role in memory . Physiol Rev . 2013 ; 93 ( 2 ): 681 – 766 . Google Scholar Crossref Search ADS PubMed WorldCat 50. Axmacher N , Elger CE , Fell J . Ripples in the medial temporal lobe are relevant for human memory consolidation . Brain . 2008 ; 131 ( Pt 7 ): 1806 – 1817 . Google Scholar Crossref Search ADS PubMed WorldCat 51. Stickgold R , Walker MP . Sleep-dependent memory triage: evolving generalization through selective processing . Nat Neurosci . 2013 ; 16 ( 2 ): 139 – 145 . Google Scholar Crossref Search ADS PubMed WorldCat 52. Perruchet P , Pacton S . Implicit learning and statistical learning: one phenomenon, two approaches . Trends Cogn Sci . 2006 ; 10 ( 5 ): 233 – 238 . Google Scholar Crossref Search ADS PubMed WorldCat 53. Tunney RJ , Altmann GT . Two modes of transfer in artificial grammar learning . J Exp Psychol Learn Mem Cogn . 2001 ; 27 ( 3 ): 614 – 639 . http://www.ncbi.nlm.nih.gov/pubmed/11394670. Accessed December 1, 2014 . 54. Schreiner T , Lehmann M , Rasch B . Auditory feedback blocks memory benefits of cueing during sleep . Nat Commun . 2015 ; 6 : 8729 . Google Scholar Crossref Search ADS PubMed WorldCat 55. Antony JW , Gobel EW , O’Hare JK , Reber PJ , Paller KA . Cued memory reactivation during sleep influences skill learning . Nat Neurosci . 2012 ; 15 ( 8 ): 1114 – 1116 . Google Scholar Crossref Search ADS PubMed WorldCat 56. Oudiette D , Antony JW , Creery JD , Paller KA . The role of memory reactivation during wakefulness and sleep in determining which memories endure . J Neurosci . 2013 ; 33 ( 15 ): 6672 – 6678 . Google Scholar Crossref Search ADS PubMed WorldCat 57. Fuentemilla L , Miró J , Ripollés P et al. Hippocampus-dependent strengthening of targeted memories via reactivation during sleep in humans . Curr Biol . 2013 ; 23 ( 18 ): 1769 – 1775 . Google Scholar Crossref Search ADS PubMed WorldCat 58. Cairney SA , Durrant SJ , Hulleman J , Lewis PA . Targeted memory reactivation during slow wave sleep facilitates emotional memory consolidation . Sleep . 2014 ; 37 ( 4 ): 701 – 7, 707A . Google Scholar Crossref Search ADS PubMed WorldCat Author notes Address correspondence to: Nora Hennies, Neuroscience and Aphasia Research Unit, School of Psychological Sciences, University of Manchester, Brunswick Street, M13 9PL Manchester, United Kingdom. Telephone: + 44 029 208 70467; Fax: +44 (0)29 2087 4679; Email: [email protected] © Sleep Research Society 2017. Published by Oxford University Press on behalf of the Sleep Research Society. All rights reserved. For permissions, please e-mail [email protected].