SLEEP

, Van Ryswyk, Emer;Anderson, Craig, S;Antic, Nicholas, A;Barbe,, Ferran;Bittencourt,, Lia;Freed,, Ruth;Heeley,, Emma;Liu,, Zhihong;Loffler, Kelly, A;Lorenzi-Filho,, Geraldo;Luo,, Yuanming;Margalef, Maria J, Masdeu;McEvoy, R, Doug;Mediano,, Olga;Mukherjee,, Sutapa;Ou,, Qiong;Woodman,, Richard;Zhang,, Xilong;Chai-Coetzer, Ching, Li

doi: 10.1093/sleep/zsz152pmid: 31587046

Abstract Study Objectives Poor adherence to continuous positive airway pressure (CPAP) commonly affects therapeutic response in obstructive sleep apnea (OSA). We aimed to determine predictors of adherence to CPAP among participants of the Sleep Apnea and cardioVascular Endpoints (SAVE) trial. Methods SAVE was an international, randomized, open trial of CPAP plus usual care versus usual care (UC) alone in participants (45–75 years) with co-occurring moderate-to-severe OSA (≥12 episodes/h of ≥4% oxygen desaturation) and established cardiovascular (CV) disease. Baseline sociodemographic, health and lifestyle factors, OSA symptoms, and 1-month change in daytime sleepiness, as well as CPAP side effects and adherence (during sham screening, titration week, and in the first month), were entered in univariate linear regression analyses to identify predictors of CPAP adherence at 24 months. Variables with p <0.2 were assessed for inclusion in a multivariate linear mixed model with country, age, and sex included a priori and site as a random effect. Results Significant univariate predictors of adherence at 24 months in 1,121 participants included: early adherence measures, improvement in daytime sleepiness at 1 month, fixed CPAP pressure, some measures of OSA severity, cardiovascular disease history, breathing pauses, and very loud snoring. While observed adherence varied between countries, adherence during sham screening, initial titration, and the first month of treatment retained independent predictive value in the multivariate model along with fixed CPAP pressure and very loud snoring. Conclusions Early CPAP adherence had the greatest predictive value for identifying those at highest risk of non-adherence to long-term CPAP therapy. Clinical Trial Registration SAVE is registered with clinicaltrials.gov (NCT00738179). sleep apnea, obstructive, continuous positive airway pressure, patient compliance Statement of Significance Poor adherence to CPAP commonly affects therapeutic response in those affected by obstructive sleep apnea (OSA). Among 1,121 participants in the SAVE study, aged 45–75 years with co-occurring moderate-to-severe OSA and cardiovascular disease, who were randomized to CPAP therapy, greater early adherence was most predictive of higher long-term adherence at 24 months. Other positive predictor variables were higher CPAP pressure, very loud snoring and country. While other variables were significant univariate predictors, they did not retain overall significance and explained only a small component of the variation in adherence. These data may be utilized to better target efforts to improve CPAP adherence. Introduction Obstructive sleep apnea (OSA) is a common chronic condition characterized by cyclical reductions in airflow and oxygen saturation, causing sleep fragmentation and daytime sleepiness, and is associated with motor vehicle accidents and cardiometabolic disease [1, 2]. Continuous positive airway pressure (CPAP) is the main treatment of OSA in people with moderate-to-severe OSA [3]. Adherence is often challenging, but is critical to improving daytime sleepiness, quality of life, and control of elevated blood pressure [4–6]. A review of 82 studies found CPAP nonadherence to affect at least one third of treated patients [3]. Adherence is particularly problematic among those diagnosed via screening in cardiovascular clinics [7–9], where OSA causes minimal symptoms and patients have other health concerns. Greater severity of OSA and high early CPAP use have consistently been shown to predict greater long-term use, whereas the influence of sleepiness, age, sex, and lifestyle on CPAP use appear to be inconsistent [7, 10–16]. Use of psychoactive medications, depression, and lower CPAP pressures may predict lower use [17, 18]. While studies designed to test ways of enhancing CPAP adherence are limited in number, and results sometimes disappointing, they suggest that improvements are possible using behavioural therapy and/or educational interventions [19]; identifying those at highest risk of nonadherence may facilitate better targeting of such interventions to those most in need and therefore result in greater cost-efficacy. The Sleep Apnea and cardioVascular Endpoints (SAVE) study (NCT00738179) was an international, multicenter, randomized controlled trial designed to determine whether CPAP can reduce the risk of major cardiovascular (CV) events in people with coexisting CV disease and moderate-to-severe OSA [8]. A preliminary evaluation of SAVE participants (n = 275) from Australia, New Zealand (NZ), and China found that the only independent predictors of CPAP adherence at 12 months were two variables: adherence to CPAP (positive) and side effects (negative) at 1 month [20]. Given the importance of CPAP adherence, we wished to extend our earlier study by examining a broader range of variables in the full population of SAVE participants. Herein, we report the role of demographic, lifestyle, and clinical variables, as predictors of CPAP adherence at 24 months among SAVE participants randomly allocated to CPAP treatment. We also detail the frequency and progression of side effects during follow-up to gain insight into the efficacy of our side effect minimization strategies. Methods Participants Participants included in these analyses were those randomized to the CPAP treatment arm of the SAVE study and included in the primary analysis [8], who had CPAP adherence data available in the first 2 years of follow-up. In brief, SAVE participants were aged 45–75 years with established coronary or cerebrovascular disease and coexisting moderate-to-severe OSA (≥12 oxygen desaturations of ≥4% per hour during overnight monitoring with an ApneaLink device [ResMed, Sydney, Australia]) [21, 22] recruited from 89 centers in China, Australia, NZ, Brazil, Spain, India, and the United States. Patients were excluded from SAVE if they had used CPAP previously, had another household member enrolled in the study, were at increased risk of motor vehicle accidents because of their occupation (e.g. professional drivers), had severe excessive daytime sleepiness (Epworth Sleepiness Scale [ESS] score >15) or a self-reported “fall-asleep” or “near miss” accident in the previous 12 months. In Spain, only patients with ESS scores of ≤10 were included as their national practice guidelines recommend that OSA patients with ESS score ≥11 be treated with CPAP. Patients were also excluded if they had severe nocturnal oxygen desaturations (>10% overnight recording time with SpO2 ≤80%); if >50% of events in their diagnostic sleep recording showed a pattern of nasal pressure fluctuations consistent with Cheyne–Stokes respiration; if there was a history of heart failure (New York Heart Association categories III–IV); or evidence of severe respiratory disease (FEV1/FVC <70% and FEV1 <50% predicted and/or resting, awake SpO2 <90%); or had another serious condition that could affect participation. The study was approved by all relevant human research and ethics committees for each participating site, and all participants provided written informed consent. CPAP treatment and follow-up Patients who met the eligibility criteria and gave written, informed consent, were provided with a sham CPAP machine to use for 1 week, to assess their willingness and likely ability to adhere to the treatment. They were phoned by the site coordinator 3 days later to assess their progress. If average use of the sham device was ≥3 h per night during this 1-week screening period, they were randomized (using a web-based program with stratification of treatment allocation by site, type of CV disease, and OSA severity) to receive either CPAP treatment in addition to usual care, or usual care (UC) alone. CPAP group participants were provided with auto-titrating CPAP devices (REMstar Auto M Series, Philips Respironics, Inc.), and after 1 week, a fixed pressure was determined for subsequent use, based on the 90th percentile pressure determined by the auto-titrating device. In a small number of cases, mainly in which technical issues were experienced in the first week, the auto-titration was repeated for a second week before conversion to fixed pressure. To optimize CPAP use, site coordinators were provided with face-to-face education and training on CPAP at regional or country-wide start-up investigator meetings and workshops, supplemented by written instructions for ensuring suitable fitting of the mask and instructions on how to troubleshoot common side-effects (Supplementary Figure S1). Trial staff had the option of using heated humidification if participants complained of nasal congestion or mouth/airway dryness. Participants had one-to-one training on the use of CPAP, viewed a patient education DVD, and were provided with additional written information on CPAP to take home. Follow-up was undertaken in person by research staff at 1 week, when the switch was made from auto to fixed CPAP, and at 1, 3, and 6 months, and each subsequent 6 months (alternating between telephone and in-office reviews) until trial completion. Adherence to CPAP was recorded using inbuilt compliance meters based on breathing detection, with usage data for the preceding period downloaded at each in-office appointment for review and trouble-shooting for side-effects. In addition, the core laboratory at the Adelaide Institute for Sleep Health (AISH) held monthly meetings to review all sleep quality data throughout the study. If average CPAP adherence at a site had fallen below 3 h per night, the country research manager was advised to investigate reasons and if necessary, initiate a program of re-education and training at the site. Advice on how to assist participants experiencing persistent difficulties with CPAP was also given by core laboratory staff upon request. Predictor variables Variables assessed as potential predictors of CPAP adherence were baseline demographics (including country, age, and sex), medical history (coronary and/or cerebrovascular disease, diabetes), anthropometric (body mass index [BMI]), and lifestyle factors such as self-reported smoking, alcohol consumption, and exercise (measured using Godin Leisure-Time Exercise Questionnaire) [23]. Severity of OSA was quantified using the frequency of ≥4% oxygen desaturation episodes and respiratory events, and the recording time spent with SpO2 <90%. Severity of daytime sleepiness (ESS score), the frequency and loudness of snoring, and frequency of witnessed apneas, were also assessed. Mood and health-related quality of life were measured using the Hospital Anxiety and Depression Scale (HADS) [24] and 36 item Short Form Health Survey (SF-36) [25], respectively. Adherence to sham CPAP during the 1-week screening period before randomization, adherence to auto-CPAP during titration, and adherence during the first month of CPAP therapy after randomization were also assessed as potential predictors, as well as side effects reported at 1 month. CPAP side effects were categorized into seven main types, with a patient receiving a score of one point for the presence of any complaints within each category, producing a total possible side effect score of seven points. The categories were (1) mouth dryness, (2) nasal symptoms, (3) eye problems, (4) claustrophobia, (5) noise problems, (6) facial soreness or skin irritation from the mask, and (7) mask fit or leak problems (i.e. trouble keeping the mask in place, air leaks from the mask or difficulty putting the mask on). Data analysis Baseline characteristics were assessed according to country of recruitment, and for all participants. Australian and NZ data were combined because of similar medical and sleep training schemes, clinical guidelines, and health care systems; and data for the sole participant in the United States were combined with that of participants in Brazil to facilitate analysis. Average CPAP adherence and the percentage of participants with “good” adherence (defined as ≥4 h per night), was determined at 1, 3, 6, 12, and 24 months. Side-effects reported at the 1-month visit, and across the 24 months of participation, were also assessed. The primary dependent outcome was average daily CPAP use in the period downloaded at the designated 24-month visit, or from an end of study visit near 24 months (between 547 and 913 days postrandomization). This download data corresponded to average usage in the preceding time interval, from 18 to 24 months postrandomization, as 6 months of data could be stored on the machines’ inbuilt recording devices. In the SAVE study, recruitment took place over a period of years. End-of-study visits were completed over a limited period of time, which meant that some participants had the opportunity to participate for longer than others. Participants without adherence data at the 24-month appointment and whose end of study appointment fell outside the specified time window were excluded from these analyses. Univariate linear regression analyses were used to identify potential predictors of long-term (24-month) adherence (utilizing all baseline and 1-month predictor variables described above); the percentage of variation in long-term adherence explained by each variable was inferred from R-squared. Variables with univariate significance at p <0.2 were assessed for inclusion in the multivariate analysis, which consisted of a linear mixed effects model with site included as a random effect, and age, sex, and country included as fixed effects a priori. Collinearity was assessed using normal linear regression collinearity diagnostic testing; where variance inflation factor was ≥5, the variable that explained the greatest variance was retained. To improve the model fit and also as an aid in clinical interpretation, some continuous variables were divided into categorical indicators: age (≤ or >60 years at randomization); BMI (<25, 25–<30, 30–<35, and ≥35 kg/m2); ODI (≤20, >20–30, and >30 events/h); HADS anxiety and depression (< or ≥8 points on the relevant subscale); time spent below 90% SpO2 (< or ≥40 min); and baseline ESS (≤10 vs. ≥11). The model was iteratively refined, first excluding variables where p >0.2 and then p >0.05, to retain only the a priori factors and variables that were significant independent predictors of long-term CPAP adherence. The total variance explained by this final multivariate model was also calculated, by comparing the residual error of the model to residual error of the base model with only the random effect for site. Interaction terms between each remaining significant factor and country were also assessed. An alternative model was also evaluated, including only information that would have been available at baseline prior to the instigation of treatment. All analyses were conducted with IBM SPSS Statistics version 23. Results Data were available for quantification of adherence at 24-month appointments (i.e. average nightly adherence between 18 and 24 months) in 1,121 CPAP group participants (participant flow diagram, Supplementary Figure S2). A CONSORT diagram of recruitment and follow-up of SAVE participants is outlined elsewhere [8]. Table 1 shows there were significant inter-country differences in the baseline characteristics of included participants. However, within each country, participants were on average, overweight to obese (BMI range 27.3–31.5 kg/m2), predominantly male (range 73%–89%), with moderate-to-severe OSA (apnea-hypopnea index [AHI] range 24.1–31.8 events/h) and not pathologically sleepy (ESS range 5.1–8.8). Overall, more participants had a history of coronary artery disease (54%) than cerebrovascular disease (49%) (categories not mutually exclusive), with the highest proportion of cerebrovascular disease among the Chinese (63%). Table 1. Summary of baseline characteristics of participants allocated to CPAP therapy Total India Australia/NZ Brazil† Spain China P (n = 1,121) (n = 42) (n = 144) (n = 93) (n = 117) (n = 725) Characteristic Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Age (years) 61.34 7.58 56.71A,B,S,C 8.82 63.92I,C 7.25 61.33I 6.27 62.62I 7.94 60.89I,A 7.48 <0.001 BMI (kg/m2) 28.67 4.46 29.61B,C 5.54 31.54C 4.86 31.93I,C 5.7 30.63C 4.27 27.31I,A,B,S 3.43 <0.001 ESS (score) 7 4 8S 3 8S,C 4 9S,C 4 5I,A,B,C 3 7A,B,S 4 <0.001 ODI (4%, events/h) 28.3 14.1 26.7 15.6 23.8C 11.4 26.4 14.8 27.1 14.1 29.7A 14.2 <0.001 AHI (events/h) 29.4 16 26.1 18.6 24.4C 14.1 24.1C 14.5 25.9C 16 31.8A,B,S 15.9 <0.001 Count % Count % Count % Count % Count % Count % Sex Female 208 18.6 6 14.3 16 11.1 25 26.9 19 16.2 142 19.6 0.027 Disease history Coronary 610 54.4 16 38.1 129 89.6 83 89.2 85 72.6 297 41.0 <0.001 Cerebrovascular 552 49.2 26 61.9 21 14.6 13 14.0 34 29.1 458 63.2 <0.001 Very loud snorer‡ 470 41.9 4 9.5 46 31.9 48 51.6 26 22.2 346 47.7 <0.001 Regular drinker 284 25.3 4 9.5 85 59.0 21 22.6 45 38.5 129 17.8 <0.001 Smoking Never or past 945 84.3 40 95.2 132 91.7 86 92.5 97 82.9 590 81.4 0.001 Current 176 15.7 2 4.8 12 8.3 7 7.5 20 17.1 135 18.6 Exercise Sedentary 285 25.4 16 38.1 47 32.6 73 78.5 35 29.9 114 15.7 <0.001 Moderate/highly active 836 74.6 26 61.9 97 67.4 20 21.5 82 70.1 611 84.3 Total India Australia/NZ Brazil† Spain China P (n = 1,121) (n = 42) (n = 144) (n = 93) (n = 117) (n = 725) Characteristic Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Age (years) 61.34 7.58 56.71A,B,S,C 8.82 63.92I,C 7.25 61.33I 6.27 62.62I 7.94 60.89I,A 7.48 <0.001 BMI (kg/m2) 28.67 4.46 29.61B,C 5.54 31.54C 4.86 31.93I,C 5.7 30.63C 4.27 27.31I,A,B,S 3.43 <0.001 ESS (score) 7 4 8S 3 8S,C 4 9S,C 4 5I,A,B,C 3 7A,B,S 4 <0.001 ODI (4%, events/h) 28.3 14.1 26.7 15.6 23.8C 11.4 26.4 14.8 27.1 14.1 29.7A 14.2 <0.001 AHI (events/h) 29.4 16 26.1 18.6 24.4C 14.1 24.1C 14.5 25.9C 16 31.8A,B,S 15.9 <0.001 Count % Count % Count % Count % Count % Count % Sex Female 208 18.6 6 14.3 16 11.1 25 26.9 19 16.2 142 19.6 0.027 Disease history Coronary 610 54.4 16 38.1 129 89.6 83 89.2 85 72.6 297 41.0 <0.001 Cerebrovascular 552 49.2 26 61.9 21 14.6 13 14.0 34 29.1 458 63.2 <0.001 Very loud snorer‡ 470 41.9 4 9.5 46 31.9 48 51.6 26 22.2 346 47.7 <0.001 Regular drinker 284 25.3 4 9.5 85 59.0 21 22.6 45 38.5 129 17.8 <0.001 Smoking Never or past 945 84.3 40 95.2 132 91.7 86 92.5 97 82.9 590 81.4 0.001 Current 176 15.7 2 4.8 12 8.3 7 7.5 20 17.1 135 18.6 Exercise Sedentary 285 25.4 16 38.1 47 32.6 73 78.5 35 29.9 114 15.7 <0.001 Moderate/highly active 836 74.6 26 61.9 97 67.4 20 21.5 82 70.1 611 84.3 AHI, apnea–hypopnea index; BMI, body mass index; CPAP, continuous positive airway pressure; ESS, Epworth Sleepiness Scale; NZ, New Zealand; ODI, oxygen desaturation index. Exercise levels correspond to Godin Leisure-Time Exercise Questionnaire (LTEQ) result [23]. Regular drinker = drinks at least one alcoholic drink per week. Significance testing comparing data by country used ANOVA for continuous variables and χ 2 for categorical variables; *Significant at the p <0.05 level. Superscript letters with continuous variables indicate significant Bonferroni-corrected post hoc comparisons between countries, as follows: A = significantly different from Australia/NZ, B = significantly different from Brazil/United States, C = significantly different from China, I = significantly different from India, S = significantly different from Spain. †“Brazil” includes 1 participant from United States; data merged due to low cell numbers. ‡Self-reported snoring loudness, response options were: as loud as breathing, as loud as talking, louder than talking, or very loud. For analysis, this was reduced to a categorical variable where “very loud” was contrasted with any other answer. Open in new tab Table 1. Summary of baseline characteristics of participants allocated to CPAP therapy Total India Australia/NZ Brazil† Spain China P (n = 1,121) (n = 42) (n = 144) (n = 93) (n = 117) (n = 725) Characteristic Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Age (years) 61.34 7.58 56.71A,B,S,C 8.82 63.92I,C 7.25 61.33I 6.27 62.62I 7.94 60.89I,A 7.48 <0.001 BMI (kg/m2) 28.67 4.46 29.61B,C 5.54 31.54C 4.86 31.93I,C 5.7 30.63C 4.27 27.31I,A,B,S 3.43 <0.001 ESS (score) 7 4 8S 3 8S,C 4 9S,C 4 5I,A,B,C 3 7A,B,S 4 <0.001 ODI (4%, events/h) 28.3 14.1 26.7 15.6 23.8C 11.4 26.4 14.8 27.1 14.1 29.7A 14.2 <0.001 AHI (events/h) 29.4 16 26.1 18.6 24.4C 14.1 24.1C 14.5 25.9C 16 31.8A,B,S 15.9 <0.001 Count % Count % Count % Count % Count % Count % Sex Female 208 18.6 6 14.3 16 11.1 25 26.9 19 16.2 142 19.6 0.027 Disease history Coronary 610 54.4 16 38.1 129 89.6 83 89.2 85 72.6 297 41.0 <0.001 Cerebrovascular 552 49.2 26 61.9 21 14.6 13 14.0 34 29.1 458 63.2 <0.001 Very loud snorer‡ 470 41.9 4 9.5 46 31.9 48 51.6 26 22.2 346 47.7 <0.001 Regular drinker 284 25.3 4 9.5 85 59.0 21 22.6 45 38.5 129 17.8 <0.001 Smoking Never or past 945 84.3 40 95.2 132 91.7 86 92.5 97 82.9 590 81.4 0.001 Current 176 15.7 2 4.8 12 8.3 7 7.5 20 17.1 135 18.6 Exercise Sedentary 285 25.4 16 38.1 47 32.6 73 78.5 35 29.9 114 15.7 <0.001 Moderate/highly active 836 74.6 26 61.9 97 67.4 20 21.5 82 70.1 611 84.3 Total India Australia/NZ Brazil† Spain China P (n = 1,121) (n = 42) (n = 144) (n = 93) (n = 117) (n = 725) Characteristic Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Age (years) 61.34 7.58 56.71A,B,S,C 8.82 63.92I,C 7.25 61.33I 6.27 62.62I 7.94 60.89I,A 7.48 <0.001 BMI (kg/m2) 28.67 4.46 29.61B,C 5.54 31.54C 4.86 31.93I,C 5.7 30.63C 4.27 27.31I,A,B,S 3.43 <0.001 ESS (score) 7 4 8S 3 8S,C 4 9S,C 4 5I,A,B,C 3 7A,B,S 4 <0.001 ODI (4%, events/h) 28.3 14.1 26.7 15.6 23.8C 11.4 26.4 14.8 27.1 14.1 29.7A 14.2 <0.001 AHI (events/h) 29.4 16 26.1 18.6 24.4C 14.1 24.1C 14.5 25.9C 16 31.8A,B,S 15.9 <0.001 Count % Count % Count % Count % Count % Count % Sex Female 208 18.6 6 14.3 16 11.1 25 26.9 19 16.2 142 19.6 0.027 Disease history Coronary 610 54.4 16 38.1 129 89.6 83 89.2 85 72.6 297 41.0 <0.001 Cerebrovascular 552 49.2 26 61.9 21 14.6 13 14.0 34 29.1 458 63.2 <0.001 Very loud snorer‡ 470 41.9 4 9.5 46 31.9 48 51.6 26 22.2 346 47.7 <0.001 Regular drinker 284 25.3 4 9.5 85 59.0 21 22.6 45 38.5 129 17.8 <0.001 Smoking Never or past 945 84.3 40 95.2 132 91.7 86 92.5 97 82.9 590 81.4 0.001 Current 176 15.7 2 4.8 12 8.3 7 7.5 20 17.1 135 18.6 Exercise Sedentary 285 25.4 16 38.1 47 32.6 73 78.5 35 29.9 114 15.7 <0.001 Moderate/highly active 836 74.6 26 61.9 97 67.4 20 21.5 82 70.1 611 84.3 AHI, apnea–hypopnea index; BMI, body mass index; CPAP, continuous positive airway pressure; ESS, Epworth Sleepiness Scale; NZ, New Zealand; ODI, oxygen desaturation index. Exercise levels correspond to Godin Leisure-Time Exercise Questionnaire (LTEQ) result [23]. Regular drinker = drinks at least one alcoholic drink per week. Significance testing comparing data by country used ANOVA for continuous variables and χ 2 for categorical variables; *Significant at the p <0.05 level. Superscript letters with continuous variables indicate significant Bonferroni-corrected post hoc comparisons between countries, as follows: A = significantly different from Australia/NZ, B = significantly different from Brazil/United States, C = significantly different from China, I = significantly different from India, S = significantly different from Spain. †“Brazil” includes 1 participant from United States; data merged due to low cell numbers. ‡Self-reported snoring loudness, response options were: as loud as breathing, as loud as talking, louder than talking, or very loud. For analysis, this was reduced to a categorical variable where “very loud” was contrasted with any other answer. Open in new tab A country-by-country comparison of average CPAP adherence per night, recorded at each follow-up appointment is provided in Table 2. At 24 months, adherence was significantly lower in China compared with Australia/NZ, but there were no other significant between-country differences at that timepoint. The proportions of participants with good CPAP adherence, defined as ≥4 h per night, are shown in Supplementary Figure S3. Most SAVE participants had good adherence in the first 6 months (78% for sham-CPAP, decreasing to 56% for CPAP at 6 months). While most Australian/NZ participants maintained a good level of adherence for the duration of the study, those from India were poorly adherent to CPAP by 3 months. Table 2. Comparison of CPAP adherence data by country (average hours of use/night) Total India Australia/NZ Brazil* Spain China P Timepoint Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Sham (run-in screening) 5.2 1.4 5.1 1.5 5.1B 1.4 5.6A,S 1.3 5.0B 1.5 5.2 1.4 0.014 Titration (autoCPAP, 1 week) 5.3 1.7 4.9 2.2 5.1 1.8 5.4 1.5 4.8C 2.0 5.4S 1.6 0.004 1 month 4.5 2.1 3.8 1.9 4.7 2.3 4.8 1.9 4.2 2.3 4.5 2.1 0.046 3 months 4.2 2.2 3.4A,B 2.1 4.5I 2.4 4.6I 2.0 3.9 2.4 4.2 2.2 0.009 6 months 4.0 2.3 3.2A,B 2.1 4.5I 2.5 4.5I 1.9 3.8 2.6 3.9 2.3 0.005 12 months 3.6 2.4 3.0 2.3 4.1 2.7 4.0 2.1 3.7 2.6 3.5 2.4 0.011 24 months† 3.4 2.6 3.0 2.6 3.9C 2.9 3.9 2.2 3.4 2.7 3.2A 2.5 0.008 Total India Australia/NZ Brazil* Spain China P Timepoint Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Sham (run-in screening) 5.2 1.4 5.1 1.5 5.1B 1.4 5.6A,S 1.3 5.0B 1.5 5.2 1.4 0.014 Titration (autoCPAP, 1 week) 5.3 1.7 4.9 2.2 5.1 1.8 5.4 1.5 4.8C 2.0 5.4S 1.6 0.004 1 month 4.5 2.1 3.8 1.9 4.7 2.3 4.8 1.9 4.2 2.3 4.5 2.1 0.046 3 months 4.2 2.2 3.4A,B 2.1 4.5I 2.4 4.6I 2.0 3.9 2.4 4.2 2.2 0.009 6 months 4.0 2.3 3.2A,B 2.1 4.5I 2.5 4.5I 1.9 3.8 2.6 3.9 2.3 0.005 12 months 3.6 2.4 3.0 2.3 4.1 2.7 4.0 2.1 3.7 2.6 3.5 2.4 0.011 24 months† 3.4 2.6 3.0 2.6 3.9C 2.9 3.9 2.2 3.4 2.7 3.2A 2.5 0.008 CPAP, continuous positive airway pressure; NZ, New Zealand. Values indicate adherence at the indicated appointment, averaged from data collected over the interval since the preceding appointment or up to 6 months. Significance testing comparing data by country by ANOVA. Superscript letters indicate significant Bonferroni-corrected post hoc comparisons between countries, as follows: A = significantly different from Australia/NZ, B = significantly different from Brazil/United States, C = significantly different from China, I = significantly different from India, S = significantly different from Spain. *“Brazil” includes 1 participant from United States; data merged due to low cell numbers. †24 months includes values downloaded at designated 24 month appointments for 1,035 participants and at end of study appointments close to 2 years for a further 86 participants. Open in new tab Table 2. Comparison of CPAP adherence data by country (average hours of use/night) Total India Australia/NZ Brazil* Spain China P Timepoint Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Sham (run-in screening) 5.2 1.4 5.1 1.5 5.1B 1.4 5.6A,S 1.3 5.0B 1.5 5.2 1.4 0.014 Titration (autoCPAP, 1 week) 5.3 1.7 4.9 2.2 5.1 1.8 5.4 1.5 4.8C 2.0 5.4S 1.6 0.004 1 month 4.5 2.1 3.8 1.9 4.7 2.3 4.8 1.9 4.2 2.3 4.5 2.1 0.046 3 months 4.2 2.2 3.4A,B 2.1 4.5I 2.4 4.6I 2.0 3.9 2.4 4.2 2.2 0.009 6 months 4.0 2.3 3.2A,B 2.1 4.5I 2.5 4.5I 1.9 3.8 2.6 3.9 2.3 0.005 12 months 3.6 2.4 3.0 2.3 4.1 2.7 4.0 2.1 3.7 2.6 3.5 2.4 0.011 24 months† 3.4 2.6 3.0 2.6 3.9C 2.9 3.9 2.2 3.4 2.7 3.2A 2.5 0.008 Total India Australia/NZ Brazil* Spain China P Timepoint Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Sham (run-in screening) 5.2 1.4 5.1 1.5 5.1B 1.4 5.6A,S 1.3 5.0B 1.5 5.2 1.4 0.014 Titration (autoCPAP, 1 week) 5.3 1.7 4.9 2.2 5.1 1.8 5.4 1.5 4.8C 2.0 5.4S 1.6 0.004 1 month 4.5 2.1 3.8 1.9 4.7 2.3 4.8 1.9 4.2 2.3 4.5 2.1 0.046 3 months 4.2 2.2 3.4A,B 2.1 4.5I 2.4 4.6I 2.0 3.9 2.4 4.2 2.2 0.009 6 months 4.0 2.3 3.2A,B 2.1 4.5I 2.5 4.5I 1.9 3.8 2.6 3.9 2.3 0.005 12 months 3.6 2.4 3.0 2.3 4.1 2.7 4.0 2.1 3.7 2.6 3.5 2.4 0.011 24 months† 3.4 2.6 3.0 2.6 3.9C 2.9 3.9 2.2 3.4 2.7 3.2A 2.5 0.008 CPAP, continuous positive airway pressure; NZ, New Zealand. Values indicate adherence at the indicated appointment, averaged from data collected over the interval since the preceding appointment or up to 6 months. Significance testing comparing data by country by ANOVA. Superscript letters indicate significant Bonferroni-corrected post hoc comparisons between countries, as follows: A = significantly different from Australia/NZ, B = significantly different from Brazil/United States, C = significantly different from China, I = significantly different from India, S = significantly different from Spain. *“Brazil” includes 1 participant from United States; data merged due to low cell numbers. †24 months includes values downloaded at designated 24 month appointments for 1,035 participants and at end of study appointments close to 2 years for a further 86 participants. Open in new tab Supplementary Table S1 outlines the side-effects to CPAP reported at the 1-month visit, where the most common were dry mouth/throat (30%) and nasal symptoms (26%), and these and all others were more often reported by participants from Australia/NZ compared with those from other countries. Supplementary Figure S4 outlines side-effects to CPAP over the first 24 months of the study. The two most common side-effects were nasal symptoms and dry mouth/throat, affecting about one third, whereas mask fit/leak problems showed the greatest temporal variation, from 12% at 3 months to 7% at 24 months. Soreness/skin irritation also decreased over time, from 7% to 3% between 1 and 24 months, whereas claustrophobia was low (1%) throughout the trial. Results of the univariate analysis are presented in Table 3. Adherence at 1 month explained 27% of the observed variance in adherence at the 24 month appointment. Other early adherence variables also individually explained considerable proportions of the variance: the titration week adherence explained 17% and sham CPAP adherence explained 8%. All other predictor variables each explained <2% of the variance in adherence at 24 months. Table 3. Univariate analysis of predictors of CPAP adherence at 24 months* (h/night) Parameter Unstandardized regression coefficient SE P R2: change from null model Age > 60 years (ref. ≤ 60) 0.484 0.155 0.002† 0.009 Sex (female) −0.167 0.198 0.400 0.001 BMI (<25, 25–<30, 30–<35, ≥35 kg/m2) −0.190 0.090 0.035† 0.004 Country −0.160 0.062 0.010† 0.006 Diabetes 0.019 0.168 0.910 <0.001 Cardiovascular disease history  Coronary 0.502 0.154 0.001† 0.009  Cerebrovascular −0.516 0.153 0.001† 0.010 Baseline ESS score (≤5, 6–10, ≥11) 0.195 0.106 0.066 0.003 Change in ESS at 1 month 0.083 0.021 <0.001† 0.014 Snoring loudness (very loud) 0.565 0.155 <0.001† 0.012 Snoring frequency (almost every day) 0.369 0.197 0.062 0.003 Breathing pauses (never/almost never) −0.413 0.206 0.045† 0.004 Early CPAP adherence (h/night)  Sham screening 0.520 0.052 <0.001† 0.081  Titration week 0.633 0.041 <0.001† 0.173  1 month 0.622 0.031 <0.001† 0.269 HADS anxiety subscale ≥ 8 0.089 0.021 0.644 <0.001 HADS depression subscale ≥ 8 −0.378 0.175 0.031† 0.004 SF-36  Physical component score‡ −0.002 0.010 0.879 <0.001  Mental component score‡ 0.007 0.009 0.423 0.001  Health transition§ −0.142 0.079 0.073 0.003 Exercise score (LTEQ)  Score (continuous) −0.006 0.003 0.029† 0.004  Moderate/highly active −0.319 0.177 0.071 0.003 Regular drinker 0.271 0.177 0.126 0.002 Current smoker −0.587 0.221 0.005† 0.007 Side effects  Score¶ −0.064 0.068 0.348 0.001  Dry mouth/throat −0.202 0.168 0.230 0.001  Nasal symptoms −0.126 0.176 0.476 <0.001  Eye problems −0.368 0.270 0.173 0.002  Claustrophobia −0.828 0.914 0.365 0.001  Noise problems −0.123 0.405 0.761 <0.001  Soreness/skin irritation 0.513 0.295 0.083 0.003  Mask fit/leak problems 0.161 0.243 0.509 <0.001 ODI (4%; ≤20, >20–30, >30 events/h) 0.236 0.088 0.008† 0.006 AHI (events/h) 0.015 0.005 0.002† 0.008 TST (evaluation, h) 0.109 0.040 0.007† 0.006 T90 ≥40 min (ref. <40) 0.412 0.154 0.007† 0.006 T90 (% of evaluation time) 0.014 0.004 0.002† 0.009 Cheyne–Stokes respiration (% events) 0.013 0.016 0.397 0.001 CPAP fixed pressure (cm H2O) 0.111 0.030 <0.001† 0.013 Parameter Unstandardized regression coefficient SE P R2: change from null model Age > 60 years (ref. ≤ 60) 0.484 0.155 0.002† 0.009 Sex (female) −0.167 0.198 0.400 0.001 BMI (<25, 25–<30, 30–<35, ≥35 kg/m2) −0.190 0.090 0.035† 0.004 Country −0.160 0.062 0.010† 0.006 Diabetes 0.019 0.168 0.910 <0.001 Cardiovascular disease history  Coronary 0.502 0.154 0.001† 0.009  Cerebrovascular −0.516 0.153 0.001† 0.010 Baseline ESS score (≤5, 6–10, ≥11) 0.195 0.106 0.066 0.003 Change in ESS at 1 month 0.083 0.021 <0.001† 0.014 Snoring loudness (very loud) 0.565 0.155 <0.001† 0.012 Snoring frequency (almost every day) 0.369 0.197 0.062 0.003 Breathing pauses (never/almost never) −0.413 0.206 0.045† 0.004 Early CPAP adherence (h/night)  Sham screening 0.520 0.052 <0.001† 0.081  Titration week 0.633 0.041 <0.001† 0.173  1 month 0.622 0.031 <0.001† 0.269 HADS anxiety subscale ≥ 8 0.089 0.021 0.644 <0.001 HADS depression subscale ≥ 8 −0.378 0.175 0.031† 0.004 SF-36  Physical component score‡ −0.002 0.010 0.879 <0.001  Mental component score‡ 0.007 0.009 0.423 0.001  Health transition§ −0.142 0.079 0.073 0.003 Exercise score (LTEQ)  Score (continuous) −0.006 0.003 0.029† 0.004  Moderate/highly active −0.319 0.177 0.071 0.003 Regular drinker 0.271 0.177 0.126 0.002 Current smoker −0.587 0.221 0.005† 0.007 Side effects  Score¶ −0.064 0.068 0.348 0.001  Dry mouth/throat −0.202 0.168 0.230 0.001  Nasal symptoms −0.126 0.176 0.476 <0.001  Eye problems −0.368 0.270 0.173 0.002  Claustrophobia −0.828 0.914 0.365 0.001  Noise problems −0.123 0.405 0.761 <0.001  Soreness/skin irritation 0.513 0.295 0.083 0.003  Mask fit/leak problems 0.161 0.243 0.509 <0.001 ODI (4%; ≤20, >20–30, >30 events/h) 0.236 0.088 0.008† 0.006 AHI (events/h) 0.015 0.005 0.002† 0.008 TST (evaluation, h) 0.109 0.040 0.007† 0.006 T90 ≥40 min (ref. <40) 0.412 0.154 0.007† 0.006 T90 (% of evaluation time) 0.014 0.004 0.002† 0.009 Cheyne–Stokes respiration (% events) 0.013 0.016 0.397 0.001 CPAP fixed pressure (cm H2O) 0.111 0.030 <0.001† 0.013 Unless otherwise specified, values are baseline measures. *Adherence at 24 months refers to an average adherence for the 6-month period downloaded at the designated 24-month visit, or from an end of study visit near 24 months. †Significant at p < 0.05 level using linear regression. Regression coefficient indicates hours of change in CPAP adherence per unit increase in the indicated parameter, or presence versus absence for categorical variables. Variables with p values <0.2 as indicated in italics were assessed for inclusion in multivariate analysis. ‡SF36 physical and mental component scores: the higher the score the less disability, that is a score of zero is equivalent to maximum disability and a score of 100 is equivalent to no disability. §SF36 Health Transition: participants rated their health compared to 1 year ago as: (1) much better, (2) somewhat better, (3) about the same, (4) somewhat worse, or (5) much worse. ||Total sleep time is the total ApneaLink recording time with any unreliable sections excluded (e.g. initialization period just before going to sleep and immediately thereafter, and any identified artefacts removed). ¶For the CPAP side effects score, a patient received one point in their score for the presence of any complaints in each of the following seven categories (highest possible score of seven); score is sum of (1) mouth dryness, (2) nasal symptoms, (3) eye problems, (4) claustrophobia, (5) noise problems, (6) facial soreness or skin irritation from the mask, and (7) mask fit or leak problems (i.e. trouble keeping the mask in place, air leaks from the mask or difficulty putting the mask on). Open in new tab Table 3. Univariate analysis of predictors of CPAP adherence at 24 months* (h/night) Parameter Unstandardized regression coefficient SE P R2: change from null model Age > 60 years (ref. ≤ 60) 0.484 0.155 0.002† 0.009 Sex (female) −0.167 0.198 0.400 0.001 BMI (<25, 25–<30, 30–<35, ≥35 kg/m2) −0.190 0.090 0.035† 0.004 Country −0.160 0.062 0.010† 0.006 Diabetes 0.019 0.168 0.910 <0.001 Cardiovascular disease history  Coronary 0.502 0.154 0.001† 0.009  Cerebrovascular −0.516 0.153 0.001† 0.010 Baseline ESS score (≤5, 6–10, ≥11) 0.195 0.106 0.066 0.003 Change in ESS at 1 month 0.083 0.021 <0.001† 0.014 Snoring loudness (very loud) 0.565 0.155 <0.001† 0.012 Snoring frequency (almost every day) 0.369 0.197 0.062 0.003 Breathing pauses (never/almost never) −0.413 0.206 0.045† 0.004 Early CPAP adherence (h/night)  Sham screening 0.520 0.052 <0.001† 0.081  Titration week 0.633 0.041 <0.001† 0.173  1 month 0.622 0.031 <0.001† 0.269 HADS anxiety subscale ≥ 8 0.089 0.021 0.644 <0.001 HADS depression subscale ≥ 8 −0.378 0.175 0.031† 0.004 SF-36  Physical component score‡ −0.002 0.010 0.879 <0.001  Mental component score‡ 0.007 0.009 0.423 0.001  Health transition§ −0.142 0.079 0.073 0.003 Exercise score (LTEQ)  Score (continuous) −0.006 0.003 0.029† 0.004  Moderate/highly active −0.319 0.177 0.071 0.003 Regular drinker 0.271 0.177 0.126 0.002 Current smoker −0.587 0.221 0.005† 0.007 Side effects  Score¶ −0.064 0.068 0.348 0.001  Dry mouth/throat −0.202 0.168 0.230 0.001  Nasal symptoms −0.126 0.176 0.476 <0.001  Eye problems −0.368 0.270 0.173 0.002  Claustrophobia −0.828 0.914 0.365 0.001  Noise problems −0.123 0.405 0.761 <0.001  Soreness/skin irritation 0.513 0.295 0.083 0.003  Mask fit/leak problems 0.161 0.243 0.509 <0.001 ODI (4%; ≤20, >20–30, >30 events/h) 0.236 0.088 0.008† 0.006 AHI (events/h) 0.015 0.005 0.002† 0.008 TST (evaluation, h) 0.109 0.040 0.007† 0.006 T90 ≥40 min (ref. <40) 0.412 0.154 0.007† 0.006 T90 (% of evaluation time) 0.014 0.004 0.002† 0.009 Cheyne–Stokes respiration (% events) 0.013 0.016 0.397 0.001 CPAP fixed pressure (cm H2O) 0.111 0.030 <0.001† 0.013 Parameter Unstandardized regression coefficient SE P R2: change from null model Age > 60 years (ref. ≤ 60) 0.484 0.155 0.002† 0.009 Sex (female) −0.167 0.198 0.400 0.001 BMI (<25, 25–<30, 30–<35, ≥35 kg/m2) −0.190 0.090 0.035† 0.004 Country −0.160 0.062 0.010† 0.006 Diabetes 0.019 0.168 0.910 <0.001 Cardiovascular disease history  Coronary 0.502 0.154 0.001† 0.009  Cerebrovascular −0.516 0.153 0.001† 0.010 Baseline ESS score (≤5, 6–10, ≥11) 0.195 0.106 0.066 0.003 Change in ESS at 1 month 0.083 0.021 <0.001† 0.014 Snoring loudness (very loud) 0.565 0.155 <0.001† 0.012 Snoring frequency (almost every day) 0.369 0.197 0.062 0.003 Breathing pauses (never/almost never) −0.413 0.206 0.045† 0.004 Early CPAP adherence (h/night)  Sham screening 0.520 0.052 <0.001† 0.081  Titration week 0.633 0.041 <0.001† 0.173  1 month 0.622 0.031 <0.001† 0.269 HADS anxiety subscale ≥ 8 0.089 0.021 0.644 <0.001 HADS depression subscale ≥ 8 −0.378 0.175 0.031† 0.004 SF-36  Physical component score‡ −0.002 0.010 0.879 <0.001  Mental component score‡ 0.007 0.009 0.423 0.001  Health transition§ −0.142 0.079 0.073 0.003 Exercise score (LTEQ)  Score (continuous) −0.006 0.003 0.029† 0.004  Moderate/highly active −0.319 0.177 0.071 0.003 Regular drinker 0.271 0.177 0.126 0.002 Current smoker −0.587 0.221 0.005† 0.007 Side effects  Score¶ −0.064 0.068 0.348 0.001  Dry mouth/throat −0.202 0.168 0.230 0.001  Nasal symptoms −0.126 0.176 0.476 <0.001  Eye problems −0.368 0.270 0.173 0.002  Claustrophobia −0.828 0.914 0.365 0.001  Noise problems −0.123 0.405 0.761 <0.001  Soreness/skin irritation 0.513 0.295 0.083 0.003  Mask fit/leak problems 0.161 0.243 0.509 <0.001 ODI (4%; ≤20, >20–30, >30 events/h) 0.236 0.088 0.008† 0.006 AHI (events/h) 0.015 0.005 0.002† 0.008 TST (evaluation, h) 0.109 0.040 0.007† 0.006 T90 ≥40 min (ref. <40) 0.412 0.154 0.007† 0.006 T90 (% of evaluation time) 0.014 0.004 0.002† 0.009 Cheyne–Stokes respiration (% events) 0.013 0.016 0.397 0.001 CPAP fixed pressure (cm H2O) 0.111 0.030 <0.001† 0.013 Unless otherwise specified, values are baseline measures. *Adherence at 24 months refers to an average adherence for the 6-month period downloaded at the designated 24-month visit, or from an end of study visit near 24 months. †Significant at p < 0.05 level using linear regression. Regression coefficient indicates hours of change in CPAP adherence per unit increase in the indicated parameter, or presence versus absence for categorical variables. Variables with p values <0.2 as indicated in italics were assessed for inclusion in multivariate analysis. ‡SF36 physical and mental component scores: the higher the score the less disability, that is a score of zero is equivalent to maximum disability and a score of 100 is equivalent to no disability. §SF36 Health Transition: participants rated their health compared to 1 year ago as: (1) much better, (2) somewhat better, (3) about the same, (4) somewhat worse, or (5) much worse. ||Total sleep time is the total ApneaLink recording time with any unreliable sections excluded (e.g. initialization period just before going to sleep and immediately thereafter, and any identified artefacts removed). ¶For the CPAP side effects score, a patient received one point in their score for the presence of any complaints in each of the following seven categories (highest possible score of seven); score is sum of (1) mouth dryness, (2) nasal symptoms, (3) eye problems, (4) claustrophobia, (5) noise problems, (6) facial soreness or skin irritation from the mask, and (7) mask fit or leak problems (i.e. trouble keeping the mask in place, air leaks from the mask or difficulty putting the mask on). Open in new tab Table 4 shows independent predictors of adherence at the 24-month appointment in multivariate mixed effects regression analysis. Many variables that were significant univariate predictors were not retained as independent predictors. Since the early adherence measures (i.e. sham CPAP, titration week, and 1-month adherence) were all significant independent predictors and not collinear, each was retained in the final multivariate model, together with CPAP fixed pressure and self-reported very loud snoring. Early adherence, whether during the sham screening phase, initial titration, or the first month of treatment, was a strong independent predictor of long-term adherence. Country was also significant, and age and sex were included in the model a priori although were not significant. Overall, the final model explained 26.2% of the variance in long-term adherence, when compared with the base model with site alone. In a sensitivity analysis, since there were significant differences between observed adherence values and baseline participant characteristics between countries, we also assessed a further model in which interaction terms with country were added for all other factors in the final model; this did not significantly improve the overall predictive value (likelihood ratio test p = 0.423). In addition, none of the individual interaction terms reached significance. We also assessed a multivariable model including only variables with univariate p <0.2 and that would have been known prior to the intervention, with no measures of early adherence or side effects (Supplementary Table S2); this baseline data only model explained only 6.6% of the observed variance at 24 months, compared with the base model with site as a random intercept. Table 4. Predictors of CPAP adherence at 24 months (h/night)—multivariate analysis Estimates of fixed effects Type III test of fixed effect Parameter Estimate 95% CI Equivalent minutes difference per unit increase P P Lower Upper Intercept −1.083 −1.884 −0.282 −65.0 0.008 0.016 Age ≤60 years (ref. >60) −0.198 −0.475 0.079 −11.9 0.160 0.160 Country Ref. China 0.011  India 0.329 −0.443 1.102 19.8 0.401  Australia/New Zealand 0.817 0.343 1.291 49.0 0.001  Brazil* 0.473 −0.098 1.043 28.4 0.102  Spain 0.593 −0.028 1.214 35.6 0.060 Sex (female, ref. male) −0.039 −0.380 0.302 −2.4 0.821 0.821 Very loud snoring† 0.373 0.094 0.652 22.4 0.009 0.009 CPAP fixed pressure (cm H20) 0.069 0.015 0.122 4.1 0.012 0.012 Early CPAP adherence (h/night)  Sham screening 0.193 0.089 0.296 11.6 <0.001 0.0003  Titration week 0.164 0.052 0.276 9.8 0.004 0.004  1 month 0.456 0.374 0.539 27.4 <0.001 3.27E-26 Estimates of fixed effects Type III test of fixed effect Parameter Estimate 95% CI Equivalent minutes difference per unit increase P P Lower Upper Intercept −1.083 −1.884 −0.282 −65.0 0.008 0.016 Age ≤60 years (ref. >60) −0.198 −0.475 0.079 −11.9 0.160 0.160 Country Ref. China 0.011  India 0.329 −0.443 1.102 19.8 0.401  Australia/New Zealand 0.817 0.343 1.291 49.0 0.001  Brazil* 0.473 −0.098 1.043 28.4 0.102  Spain 0.593 −0.028 1.214 35.6 0.060 Sex (female, ref. male) −0.039 −0.380 0.302 −2.4 0.821 0.821 Very loud snoring† 0.373 0.094 0.652 22.4 0.009 0.009 CPAP fixed pressure (cm H20) 0.069 0.015 0.122 4.1 0.012 0.012 Early CPAP adherence (h/night)  Sham screening 0.193 0.089 0.296 11.6 <0.001 0.0003  Titration week 0.164 0.052 0.276 9.8 0.004 0.004  1 month 0.456 0.374 0.539 27.4 <0.001 3.27E-26 CPAP, continuous positive airways pressure; CI, confidence interval. Multivariate analysis using a linear mixed model with sites as a random effect and use of CPAP at 24 months as the dependent outcome variable (age, sex, and country included a priori). *Brazil includes one participant from United States; data merged due to low cell numbers. †Self-reported snoring loudness, response options were: as loud as breathing, as loud as talking, louder than talking, or very loud. For analysis, this was reduced to a categorical variable where “very loud” was contrasted with any other answer. Open in new tab Table 4. Predictors of CPAP adherence at 24 months (h/night)—multivariate analysis Estimates of fixed effects Type III test of fixed effect Parameter Estimate 95% CI Equivalent minutes difference per unit increase P P Lower Upper Intercept −1.083 −1.884 −0.282 −65.0 0.008 0.016 Age ≤60 years (ref. >60) −0.198 −0.475 0.079 −11.9 0.160 0.160 Country Ref. China 0.011  India 0.329 −0.443 1.102 19.8 0.401  Australia/New Zealand 0.817 0.343 1.291 49.0 0.001  Brazil* 0.473 −0.098 1.043 28.4 0.102  Spain 0.593 −0.028 1.214 35.6 0.060 Sex (female, ref. male) −0.039 −0.380 0.302 −2.4 0.821 0.821 Very loud snoring† 0.373 0.094 0.652 22.4 0.009 0.009 CPAP fixed pressure (cm H20) 0.069 0.015 0.122 4.1 0.012 0.012 Early CPAP adherence (h/night)  Sham screening 0.193 0.089 0.296 11.6 <0.001 0.0003  Titration week 0.164 0.052 0.276 9.8 0.004 0.004  1 month 0.456 0.374 0.539 27.4 <0.001 3.27E-26 Estimates of fixed effects Type III test of fixed effect Parameter Estimate 95% CI Equivalent minutes difference per unit increase P P Lower Upper Intercept −1.083 −1.884 −0.282 −65.0 0.008 0.016 Age ≤60 years (ref. >60) −0.198 −0.475 0.079 −11.9 0.160 0.160 Country Ref. China 0.011  India 0.329 −0.443 1.102 19.8 0.401  Australia/New Zealand 0.817 0.343 1.291 49.0 0.001  Brazil* 0.473 −0.098 1.043 28.4 0.102  Spain 0.593 −0.028 1.214 35.6 0.060 Sex (female, ref. male) −0.039 −0.380 0.302 −2.4 0.821 0.821 Very loud snoring† 0.373 0.094 0.652 22.4 0.009 0.009 CPAP fixed pressure (cm H20) 0.069 0.015 0.122 4.1 0.012 0.012 Early CPAP adherence (h/night)  Sham screening 0.193 0.089 0.296 11.6 <0.001 0.0003  Titration week 0.164 0.052 0.276 9.8 0.004 0.004  1 month 0.456 0.374 0.539 27.4 <0.001 3.27E-26 CPAP, continuous positive airways pressure; CI, confidence interval. Multivariate analysis using a linear mixed model with sites as a random effect and use of CPAP at 24 months as the dependent outcome variable (age, sex, and country included a priori). *Brazil includes one participant from United States; data merged due to low cell numbers. †Self-reported snoring loudness, response options were: as loud as breathing, as loud as talking, louder than talking, or very loud. For analysis, this was reduced to a categorical variable where “very loud” was contrasted with any other answer. Open in new tab Discussion Among participants with CV disease and moderate-to-severe OSA in the treatment arm of the SAVE study, mean CPAP usage assessed at the 24-month review for the preceding period was 3.4 ± 2.6 h/night. Raw values for CPAP adherence varied between countries across time, with generally lower adherence in India, and Spain at early time points, and higher in Australia/NZ and Brazil. The strongest predictors of adherence at 24 months were early adherence measures. Adherence in the first month was the most predictive of adherence at 24 months, followed by sham screening adherence, then titration week(s) adherence. CPAP pressure level, very loud snoring, and location were also important. Restricting the analysis to use only information available before treatment resulted in a model that was poorly predictive of future use, explaining less than 7% of the observed variance. This perhaps serves to emphasize how poorly the determinants of good long-term compliance in treatment-naïve individuals are understood. Early adherence to CPAP therapy was an independent predictor of CPAP use at 24 months in this analysis, and agrees with previous findings [13, 26–29], including our 2013 analysis of adherence to CPAP therapy in a subset of participants at 12 months in SAVE [20]. It was not expected that the 1-week pretrial screening of sham CPAP with subtherapeutic airflow pressure would have significantly reduced either night-time OSA severity or OSA-related symptoms such as snoring and daytime sleepiness. Nonetheless, it was retained in the model as an independent predictor of long-term CPAP adherence, suggesting that psychological factors may influence willingness to engage with the treatment, in addition to perceived symptomatic benefit. As previously reported, higher CPAP pressure was associated with greater compliance [17, 18], perhaps reflecting greater OSA symptom relief in those with more severe OSA, however this accounted for only a small proportion of the long-term values. The fact that OSA severity alone (as measured by AHI, ODI or absolute/relative time with oxygen saturation <90%) was not found to be an independent predictor of adherence in the present study, but has been reported in several previous studies [11, 12, 15, 16, 30, 31], may be attributable to inclusion of only those with moderate-to-severe OSA in SAVE. Snoring loudness was independently associated with increased usage, while witnessed apneas and improvements in daytime sleepiness were significant in univariate but not multivariate analysis, suggesting that in this clinical trial population these symptoms had little influence on whether participants continued with the therapy. It is noteworthy that early side effects did not predict 24-month adherence in our study, but did predict adherence at 12 months in the 2013 SAVE trial analysis (275 participants from Australia, NZ, and China) [20]. Effort was made throughout the SAVE trial to address side effects where indicated (Supplementary Figure S2), including the addition of humidification; use of nasal corticosteroid spray; checking/changing mask fit/type including the use of nasal pillows; and hypnotic/sedative for short-term relief of claustrophobia. Over the duration of the trial, certain side effects reduced in frequency (mask fit/leak problems, soreness/skin irritation and noise problems) whilst others varied little or remained at the same frequency and eye problems slightly increased in frequency; side effect severity was not quantitatively measured. The interventions undertaken within our trial to reduce side effects may explain why side effects did not consistently predict adherence. Baratta et al. [14] reported that the main reasons for poor adherence reported by participants in their study were self-reported mask-related and pressure-related side effects, nasal symptoms, and psychological and social factors. However, they did not specifically and independently assess the influence of side effects. Surprisingly, other recent studies have not systematically assessed the impact of side effects on compliance to CPAP therapy [7, 11, 13]. More research is needed to clarify the predictive nature of side effects. In keeping with the results of several previous studies [13, 15, 16, 32] and our 2013 preliminary analysis of SAVE trial data [20], self-reported daytime sleepiness (ESS) was a significant univariate but not independent predictor of long-term CPAP adherence, as was the change in ESS score from screening to 1-month follow-up. However, other studies have found a relationship between level of reported sleepiness and adherence [33, 34]. The lack of an association between ESS score and adherence in our study may be because participants were generally minimally symptomatic (mean ESS 7.3). Schoch et al. [35] found that the correlation between ESS score and adherence rate was nonlinear, with the best adherence for ESS scores between 11 and 14 (full range being 0–24). The relationship between sleepiness and adherence requires further clarification. This is one of the largest studies investigating predictors of CPAP adherence. It fills many of the gaps left by previous research in this area by sytematically examining the impact of side effects, lifestyle factors and underlying depression and anxiety. Multiple types of variables (lifestyle, medical history, sleepiness, OSA severity, and mood) were taken into consideration in the multivariate analysis, providing accurate estimates of the predictive value of each. Detailed and standardized prospective assessment of subjects [36], and the completeness of follow-up as part of a clinical trial were other strong aspects of this study. There were however some limitations. While multiple variable types were assessed, potential psychological predictors of behavior were not assessed (e.g. self-efficacy and behavior-change constructs). Secondly, participants were enrolled in the trial only if they used sham CPAP for an average of ≥3 h/night during the 1-week sham screening period; given that sham CPAP use was an independent predictor of long-term use, this means that participants at highest risk of poor long-term adherence were excluded. Also, while our country-by-country analysis demonstrated that adherence was low in India, our Indian sample size was small. Additionally, the daytime sleepiness levels of participants in this trial were generally lower than those who are treated at sleep clinics, and this may have impacted the observed adherence. Finally, the participants were patients with established cardiovascular disease who consented to be screened for OSA in order to be enrolled in a clinical trial, which may indicate some degree of altruism, a tendency to comply with medical recommendations and/or increased motivation toward improving their own health. Furthermore, patients with very severe sleepiness and very severe oxygen desaturation were excluded. Thus, caution needs to be exercised when extrapolating the results to a general sleep apnea clinic population. A relatively small proportion of the total variance in CPAP adherence was explained in this and previous studies [18]. New analytical approaches describing specific clinical phenotypes or clusterings of a broad range of variable types within OSA groups might provide further insights into why some patients accept and adhere well to CPAP while others do not. Types of variables important to consider include not just biomedical but also social, psychological, location-related, and/or health care-delivery-related variables [37]. Without this information it may be difficult to design effective interventions to improve CPAP treatment. According to the most recent Cochrane review assessing the efficacy of educational and/or behavioral interventions to improve CPAP use in adults with OSA (2014), there is low quality evidence indicating that behavioral therapy results in a large increase in CPAP machine use, and moderate-quality evidence that short-term educational interventions result in modest increases in CPAP use [19]. Thus, the quality of future interventional studies must be a priority. Also, most research has been conducted in CPAP-naïve people, and more research is needed into interventions for those struggling to persist with therapy and optimal timing and duration of such interventions [19]. Given that early adherence was a strong predictor of long-term compliance, for now, a short-term trial of CPAP therapy (of approximately 1 month) may be the best practical means to identify those at highest risk of future nonadherence to treatment, allowing clinical and research resources to be targeted to those who most need it. Further investigation and efforts to improve CPAP adherence in people from emerging economies (in this case, China and India) is also warranted. Funding The SAVE trial was supported by project grants (1006501 [2011–2015] and 1060078 [2014–2016]) from the National Health and Medical Research Council (NHMRC) of Australia and by Respironics Sleep and Respiratory Research Foundation and Philips Respironics. Supplementary trial funding was provided by Fisher & Paykel Healthcare, the Australasian Sleep Trials Network (enabling grant 343020 from the NHMRC), the Spanish Respiratory Society (grant 105-2011 to Drs. Barbe and Mediano), and Fondo de Investigaciones Sanitarias (grant 13/02053 to Drs. Barbe and Mediano). In-kind donations were provided by Respironics for CPAP equipment and by ResMed for sleep apnea diagnostic devices. Conflict of interest statement. Professor McEvoy reports receiving research grants from Philips Respironics, Fisher & Paykel, and the National Health and Medical Research Council of Australia; research equipment grants from Philips Respironics, ResMed, and Airliquide; and speaker fees from ResMed. Professor Anderson reports receiving advisory committee member sitting fees from AMGEN and speaker fees as well as travel reimbursement from Takeda. All other authors have no conflicts to disclose. Acknowledgments We thank the individuals who participated in this trial and their families; the clinical and research teams of the various sleep, cardiology, and neurology departments; and the teams at The George Institute and the Adelaide Institute for Sleep Health, for their work on the SAVE study. We are also grateful to Dr Qiang Li for his provision of information relating to analyses undertaken in the 2013 study utilizing early SAVE trial data [20] to allow us to undertake comparable analyses. A complete list of sites and trial investigators and coordinators in the Sleep Apnea Cardiovascular Endpoints (SAVE) study is provided in a Supplementary Appendix to the main trial report, available at NEJM.org. References 1. Ward KL , et al. Excessive daytime sleepiness increases the risk of motor vehicle crash in obstructive sleep apnea . J Clin Sleep Med. 2013 ; 9 ( 10 ): 1013 – 1021 . Google Scholar PubMed WorldCat 2. Drager LF , et al. Obstructive sleep apnea: a cardiometabolic risk in obesity and the metabolic syndrome . J Am Coll Cardiol. 2013 ; 62 ( 7 ): 569 – 576 . Google Scholar Crossref Search ADS PubMed WorldCat 3. Rotenberg BW , et al. Trends in CPAP adherence over twenty years of data collection: a flattened curve . J Otolaryngol Head Neck Surg. 2016 ; 45 ( 1 ): 43 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Weaver TE , et al. Relationship between hours of CPAP use and achieving normal levels of sleepiness and daily functioning . Sleep. 2007 ; 30 ( 6 ): 711 – 719 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Antic NA , et al. The effect of CPAP in normalizing daytime sleepiness, quality of life, and neurocognitive function in patients with moderate to severe OSA . Sleep. 2011 ; 34 ( 1 ): 111 – 119 . Google Scholar Crossref Search ADS PubMed WorldCat 6. Bratton DJ , et al. CPAP vs mandibular advancement devices and blood pressure in patients with obstructive sleep apnea: a systematic review and meta-analysis . JAMA. 2015 ; 314 ( 21 ): 2280 – 2293 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Florés M , et al. Predictors of long-term adherence to continuous positive airway pressure in patients with obstructive sleep apnoea and acute coronary syndrome . J Thorac Dis. 2018 ; 10 ( suppl 1 ): S124 – S134 . Google Scholar Crossref Search ADS PubMed WorldCat 8. McEvoy RD , et al. ; SAVE Investigators and Coordinators . CPAP for prevention of cardiovascular events in obstructive sleep apnea . N Engl J Med. 2016 ; 375 ( 10 ): 919 – 931 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Peker Y , et al. Effect of positive airway pressure on cardiovascular outcomes in coronary artery disease patients with nonsleepy obstructive sleep apnea. The RICCADSA randomized controlled trial . Am J Respir Crit Care Med. 2016 ; 194 ( 5 ): 613 – 620 . Google Scholar Crossref Search ADS PubMed WorldCat 10. Craig SE , et al. Continuous positive airway pressure improves sleepiness but not calculated vascular risk in patients with minimally symptomatic obstructive sleep apnoea: the MOSAIC randomised controlled trial . Thorax. 2012 ; 67 ( 12 ): 1090 – 1096 . Google Scholar Crossref Search ADS PubMed WorldCat 11. Campos-Rodriguez F , et al. ; Spanish Sleep Network . Long-term adherence to continuous positive airway pressure therapy in non-sleepy sleep apnea patients . Sleep Med. 2016 ; 17 : 1 – 6 . Google Scholar Crossref Search ADS PubMed WorldCat 12. Gagnadoux F , et al. ; IRSR Sleep Cohort Group . Adherence to positive airway pressure in non-sleepy patients with obstructive sleep apnoea . Eur Respir J. 2013 ; 42 ( 3 ): 863 – 866 . Google Scholar Crossref Search ADS PubMed WorldCat 13. Luyster FS , et al. Long-term use of continuous positive airway pressure therapy in coronary artery disease patients with nonsleepy obstructive sleep apnea . Clin Cardiol. 2017 ; 40 ( 12 ): 1297 – 1302 . Google Scholar Crossref Search ADS PubMed WorldCat 14. Baratta F , et al. Long-term prediction of adherence to continuous positive air pressure therapy for the treatment of moderate/severe obstructive sleep apnea syndrome . Sleep Med. 2018 ; 43 : 66 – 70 . Google Scholar Crossref Search ADS PubMed WorldCat 15. Kohler M , et al. Predictors of long-term compliance with continuous positive airway pressure . Thorax. 2010 ; 65 ( 9 ): 829 – 832 . Google Scholar Crossref Search ADS PubMed WorldCat 16. Balakrishnan K , et al. Predicting CPAP use and treatment outcomes using composite indices of sleep apnea severity . J Clin Sleep Med. 2016 ; 12 ( 6 ): 849 – 854 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Campos-Rodriguez F , et al. Long-term continuous positive airway pressure compliance in females with obstructive sleep apnoea . Eur Respir J. 2013 ; 42 ( 5 ): 1255 – 1262 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Law M , et al. Depression may reduce adherence during CPAP titration trial . J Clin Sleep Med. 2014 ; 10 ( 2 ): 163 – 169 . Google Scholar PubMed WorldCat 19. Wozniak DR , Lasserson TJ , Smith I . Educational, supportive and behavioural interventions to improve usage of continuous positive airway pressure machines in adults with obstructive sleep apnoea . Cochrane Database Syst Rev . 2014 ;( 1 ): CD007736 . doi:10.1002/14651858.CD007736.pub2 WorldCat 20. Chai-Coetzer CL , et al. Predictors of long-term adherence to continuous positive airway pressure therapy in patients with obstructive sleep apnea and cardiovascular disease in the SAVE study . Sleep. 2013 ; 36 ( 12 ): 1929 – 1937 . Google Scholar Crossref Search ADS PubMed WorldCat 21. American Academy of Sleep Medicine Task Force . Sleep-related breathing disorders in adults: recommendations for syndrome definition and measurement techniques in clinical research. The Report of an American Academy of Sleep Medicine Task Force . Sleep . 1999 ; 22 ( 5 ): 667 – 689 . Crossref Search ADS PubMed WorldCat 22. Gantner D , et al. Diagnostic accuracy of a questionnaire and simple home monitoring device in detecting obstructive sleep apnoea in a Chinese population at high cardiovascular risk . Respirology. 2010 ; 15 ( 6 ): 952 – 960 . Google Scholar Crossref Search ADS PubMed WorldCat 23. Godin G , et al. A simple method to assess exercise behavior in the community . Can J Appl Sport Sci. 1985 ; 10 ( 3 ): 141 – 146 . Google Scholar PubMed WorldCat 24. Snaith RP . The hospital anxiety and depression scale . Health Qual Life Outcomes. 2003 ; 1 : 29 . Google Scholar Crossref Search ADS PubMed WorldCat 25. Ware JE , Jr , et al. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection . Med Care. 1992 ; 30 ( 6 ): 473 – 483 . Google Scholar Crossref Search ADS PubMed WorldCat 26. Weaver TE , et al. Night-to-night variability in CPAP use over the first three months of treatment . Sleep. 1997 ; 20 ( 4 ): 278 – 283 . Google Scholar Crossref Search ADS PubMed WorldCat 27. Budhiraja R , et al. Early CPAP use identifies subsequent adherence to CPAP therapy . Sleep. 2007 ; 30 ( 3 ): 320 – 324 . Google Scholar PubMed WorldCat 28. McArdle N , et al. Long-term use of CPAP therapy for sleep apnea/hypopnea syndrome . Am J Respir Crit Care Med. 1999 ; 159 ( 4 Pt 1 ): 1108 – 1114 . Google Scholar Crossref Search ADS PubMed WorldCat 29. Popescu G , et al. Continuous positive airway pressure for sleep apnoea/hypopnoea syndrome: usefulness of a 2 week trial to identify factors associated with long term use . Thorax. 2001 ; 56 ( 9 ): 727 – 733 . Google Scholar Crossref Search ADS PubMed WorldCat 30. Somers ML , et al. Continuous positive airway pressure adherence for obstructive sleep apnea . ISRN Otolaryngol. 2011 ; 2011 : 943586 . Google Scholar Crossref Search ADS PubMed WorldCat 31. Wohlgemuth WK , et al. Attempters, adherers, and non-adherers: latent profile analysis of CPAP use with correlates . Sleep Med. 2015 ; 16 ( 3 ): 336 – 342 . Google Scholar Crossref Search ADS PubMed WorldCat 32. Gagnadoux F , et al. ; IRSR sleep cohort group . Influence of marital status and employment status on long-term adherence with continuous positive airway pressure in sleep apnea patients . PLoS One. 2011 ; 6 ( 8 ): e22503 . Google Scholar Crossref Search ADS PubMed WorldCat 33. Nadal N , et al. Predictors of CPAP compliance in different clinical settings: primary care versus sleep unit . Sleep Breath. 2018 ; 22 ( 1 ): 157 – 163 . Google Scholar Crossref Search ADS PubMed WorldCat 34. Jacobsen AR , et al. Determinants for adherence to continuous positive airway pressure therapy in obstructive sleep apnea . PLoS One. 2017 ; 12 ( 12 ): e0189614 . Google Scholar Crossref Search ADS PubMed WorldCat 35. Schoch OD , et al. Baseline predictors of adherence to positive airway pressure therapy for sleep apnea: a 10-year single-center observational cohort study . Respiration. 2014 ; 87 ( 2 ): 121 – 128 . Google Scholar Crossref Search ADS PubMed WorldCat 36. Antic NA , et al. The sleep apnea cardiovascular endpoints (SAVE) trial: rationale, ethics, design, and progress . Sleep. 2015 ; 38 ( 8 ): 1247 – 1257 . Google Scholar Crossref Search ADS PubMed WorldCat 37. Gordon A , et al. Untreated sleep apnea: an analysis of administrative data to identify risk factors for early nonadherence . J Clin Sleep Med. 2018 ; 14 ( 8 ): 1303 – 1313 . Google Scholar Crossref Search ADS PubMed WorldCat © Sleep Research Society 2019. Published by Oxford University Press on behalf of the Sleep Research Society. All rights reserved. For permissions, please e-mail [email protected]. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Zielinski, Mark R; Atochin, Dmitriy N; McNally, James M; McKenna, James T; Huang, Paul L; Strecker, Robert E; Gerashchenko, Dmitry

doi: 10.1093/sleep/zsz143pmid: 31328777

Abstract Slow-wave activity (SWA) is an oscillatory neocortical activity occurring in the electroencephalogram delta (δ) frequency range (~0.5–4 Hz) during nonrapid eye movement sleep. SWA is a reliable indicator of sleep homeostasis after acute sleep loss and is involved in memory processes. Evidence suggests that cortical neuronal nitric oxide synthase (nNOS) expressing neurons that coexpress somatostatin (SST) play a key role in regulating SWA. However, previous studies lacked selectivity in targeting specific types of neurons that coexpress nNOS—cells which are activated in the cortex after sleep loss. We produced a mouse model that knocks out nNOS expression in neurons that coexpress SST throughout the cortex. Mice lacking nNOS expression in SST positive neurons exhibited significant impairments in both homeostatic low-δ frequency range SWA production and a recognition memory task that relies on cortical input. These results highlight that SST+/nNOS+ neurons are involved in the SWA homeostatic response and cortex-dependent recognition memory. cortex, delta power, neuronal nitric oxide synthase, recognition memory, slow-wave activity, somatostatin Statement of Significance Electrical brain activity in the 0.5–4 Hz delta frequency range occurring during nonrapid eye movement sleep is known as slow-wave activity (SWA). SWA is an indicator of sleep need after acute sleep loss and is associated with memory functioning. A unique population of neurons in the cortex that express somatostatin (SST) and neuronal nitric oxide (nNOS) are active during sleep, although their function is unknown. Using a strategy to abolish expression of nNOS in SST–nNOS cells localized within the cortex and caudate-putamen, we demonstrate that these cells are critical to the generation of SWA in the low frequency range and cortical-dependent memory. These findings support the idea that SST–nNOS neurons are involved in SWA sleep responses to sleep loss. Introduction Electroencephalographic (EEG) slow-wave activity (SWA) is a slow synchronized oscillatory neocortical activity occurring in the delta (δ) frequency range (~0.5–4 Hz) occurring during nonrapid eye movement (NREM) sleep [1]. SWA pressure typically increases with time spent awake and decreases during subsequent sleep, although there are numerous incidences when this relationship does not exist [2]. Nevertheless, SWA is considered a reliable indicator of sleep homeostasis after acute sleep loss [1]. Evidence using cortical EEG electrodes or local field potentials indicate that increased cortical neuronal activity, which ensues after whisker stimulation or sleep deprivation, takes place, in part, through cortical networks to increase SWA [3]. Additionally, local changes in SWA are an indicator of learning-induced plasticity related to complex cognitive functions such as working memory [4]. Neurons expressing neuronal nitric oxide synthase (nNOS) within the cerebral cortex are classified into type I and type II based on their morphological features and gene expression profile [5]. Type I nNOS neurons are more intensely stained in nNOS immunohistochemical or NADPH diaphorase staining procedures and have larger somata than type II cells [5]. Type I nNOS neurons originate from the medial ganglionic eminence (MGE), possess long-range projections, and coexpress gamma-aminobutyric acid (GABA), neuropeptide Y (NPY), somatostatin (SST), neurokinin 1 (NK1) receptor, and N-methyl-D-aspartate (NMDA) glutamate receptor [5–9]. We previously demonstrated that type I nNOS neurons are activated during episodes of NREM sleep associated with increased SWA [10]. We also showed that optogenetically evoked responses in nNOS-positive cells of the cerebral cortex are consistent with their role in slow-wave sleep physiology [11]. These sleep-active type I nNOS neurons coexpress SST and NK1 receptors [12]. We performed injections of an NK1 receptor agonist (substance P-fragment 1,7) and an antagonist (CP96345) into the cerebral cortex and found that SWA was locally enhanced by the NK1 receptor agonist and reduced by the NK1 receptor antagonist [13]. Because NK1 receptors are expressed in type I nNOS neurons but not in other neurons in the cerebral cortex [12], the NK1 receptor pharmacology findings implicate type I nNOS neurons in the modulation of SWA. A recent study involving the chemogenetic activation of SST-positive cells in the cerebral cortex showed increased SWA, slope of individual slow waves, and NREM sleep duration compared with control conditions, whereas chemogenetic inhibition of these cells reduced SWA and slow-wave incidence without changing time spent in NREM sleep [14]. Furthermore, the excitatory drive during the slow oscillation appears to be mainly countered by inhibitory activity of SST-positive interneurons [15]. Consequently, it is plausible that these changes in SWA resulted from the enhanced activation of the type I nNOS neurons—cells that are all SST-positive [5]. The activation of nNOS neurons can lead to an increased production of nitric oxide (NO) [16, 17], which may mediate homeostatic SWA enhancements. Consistent with this mechanism, nNOS knockout (KO) mice (B6.129S4-Nos1tm1Plh/J) have a profound deficiency in the production of SWA during recovery sleep that follows sleep deprivation [18]. Interestingly, NK1 receptor neurons located within the cerebral cortex that express nNOS continue to be activated but do not show enhanced SWA during recovery sleep after sleep loss, which supports the conclusion that the production of NO by nNOS contributes to these SWA enhancements [18]. However, we cannot conclude whether these changes in SWA in nNOS KO mice are caused by the nNOS deficiency in type I nNOS neurons in the cerebral cortex or by cells producing nNOS in subcortical regions because nNOS is deficient in the whole brain of nNOS KO mice. Herein, we produced a novel mouse model that specifically knocks out nNOS expression in type I nNOS neurons in the brain using a cross-sectional strategy targeting SST+/nNOS+ colocalized cells in the cerebral cortex and caudate-putamen. We show that nNOS deficiency in type I nNOS neurons leads to impairments in both SWA production and recognition memory. SWA was diminished during spontaneous sleep in these mice and they exhibited reduced homeostatic response to sleep deprivation specifically in the low-δ frequency range of SWA (less than 1.5 Hz). Memory performances in the cortex-dependent floor texture recognition (FTR) test was also impaired in mice lacking SST+/nNOS+ cells, but not in the hippocampal-dependent novel object recognition (NOR) test. Materials and Methods Animals All experimental protocols were approved by the VA Boston Healthcare System Institutional Animal Care and Use Committee and were in compliance with the National Institutes of Health guidelines. At all times, care was taken to minimize animal discomfort and avoid pain. Mice were maintained under 12 hr: 12 hr light/dark cycle with light onset at Zeitgeber 0 (ZT0) at 22 ± 3°C. Water and food were available ad libitum. Generation of nNOS knockout mice that are selective for type I nNOS neurons nNOS floxed mice (nNOSflox/flox) mice were generated by gene targeting in mouse embryonic stem cells and successfully used to produce cell-specific nNOS knockout in the renal-collecting duct principal cells, vas deferens, and seminiferous tubules within testis [19]. Mice that are homozygous for this allele are viable, fertile, normal in size and do not display any gross physical or behavioral abnormalities [19]. Homozygous nNOS floxed mice were backcrossed on the C57BL/6J background for three generations and then mated with hemizygous, Ssttm2.1(cre)Zjh/J mice (B6N.Cg-Ssttm2.1(cre)Zjh/J; the Jackson Laboratory, Stock No. 018973; Supplementary Figure S1). Female offspring that were SST-Cre positive and heterozygous for nNOS floxed transgene were then mated with homozygous nNOS floxed male mice. The offspring from this mating, which were homozygous for the floxed nNOS transgene and either SST-Cre positive (Sst-Cre+/−; Nos1flox/flox) or negative (Sst-Cre−/−; Nos1flox/flox), were perfused with formalin. Brains of these mice were processed for either nNOS staining or SST/nNOS double staining (see Immunohistochemistry section). Furthermore, Ssttm2.1(cre)Zjh/J mice were crossed with Ai14 (B6;129S6-Gt(ROSA)26Sortm14(CAG-tdTomato)Hze/J) reporter mice. The offspring from this mating (Sst-Cre+/−; Ai14 mice) were analyzed for the presence of nNOS/RFP double-labeled cells. Upon weaning, mice were genotyped by PCR using the following SST primers: #11224, common reverse, GGG CCA GGA GTT AAG GAA GA; #11225, wild type forward, TCT GAA AGA CTT GCG TTT GG; #9989, mutant forward, TGG TTT GTC CAA ACT CAT CAA; the Jackson Laboratory; stock #018973. The size of PCR-amplified products was 465 bp in wild-type mice, 200 bp in mutant mice, and both 200 and 465 bp in heterozygote mice. The following nNOSflox primers were used: TGT TCC ATG CAC TGT GTT AGC (forward) and GAT ACG TGT AGA GGG CAA ATG (reverse). The size of PCR-amplified products was 421 bp in wild-type mice, 484 bp in mutant mice, and both 421 and 484 bp in heterozygote mice. Polysomnography recording Sst-Cre+/−; Nos1flox/flox (N = 10) and Sst-Cre−/−; Nos1flox/flox (N = 9) mice used for polysomnography surgery were anesthetized with isoflurane (1%–2%) and surgically implanted with an EEG electrode in the somatosensory cortex (1 mm posterior to bregma and 1 mm lateral to the midline) and a ground electrode over the cerebellum (0.5 mm posterior to lambda placed centrally) as described previously [13]. Mice were also implanted with two electromyogram (EMG) electrodes in the nuchal muscles. The electrodes were attached to a pedestal and formed into a head mount with dental cement. Mice were tethered to wireless transponders (F20-EET transponders; Data Sciences International, St. Paul, MN) on a swivel mounted system (Neurotargeting Systems, Inc., Chestnut Hill, MA), allowing the mice to move freely in their cage as described previously [13]. The cages were positioned on top of receiver plates that functioned to detect potential data from FM signals from the transponders to a data exchange matrix (Data Sciences International, St. Paul, MN). The signals were transferred to a computer using the Dataquest A.R.T system. EEG and EMG signals were amplified, analog-to-digital–converted, and stored at 500 Hz. After at least 2 week postsurgical recovery, EEG and EMG were recorded via telemetry using DQ ART 4.1 software (Data Sciences Inc.). Recordings consisted of a “baseline” day, followed by a 6 hr sleep deprivation initiated at light onset (ZT0) and performed by “gentle handling” [20], followed by a 6 hr recovery period. Sleep deprivation procedures Sleep deprivation procedures were similar to those previously published [20]. Sst-Cre+/−; Nos1flox/flox (N = 10) and Sst-Cre−/−; Nos1flox/flox (N = 9) mice were continuously observed and, when inactive and appeared to be entering sleep, cage tapping occurred. As sleep deprivation progressed and it became more difficult to sustain wakefulness, novel objects were introduced into the home cage, and, when necessary, an artist’s brush was used to stroke the fur or vibrissae. Multiple sleep latency test To objectively measure sleepiness, we performed a variation of the murine Multiple Sleep Latency Test [18] with 10 Sst-Cre+/−; Nos1flox/flox and 9 Sst-Cre−/−; Nos1flox/flox mice. Starting 3 hr into the light period, mice were subjected to five 20 min sessions of forced waking, each followed by a 20 min nap opportunity within a continuous 200 min time interval. The latency to the onset of NREM and rapid eye movement (REM) sleep as well as the percentage of time spent in both states was determined for each nap opportunity. Identification of sleep/wake states and sleep/wake data analyses After completion of the data collection, expert scorers determined states of sleep and wakefulness in 10 s epochs by examining the recordings visually using SleepSign software. The EEG and EMG data were scored for waking, REM sleep, and NREM sleep. Vigilance states—including NREM sleep, REM sleep, and waking—were determined in 2 hr time blocks as described previously [20]. As indices of the consolidation of behavioral states, the average duration and number of bouts for each state were calculated. Fast Fourier transformations of EEG signals (μV2) using a Hanning window were made within each 10 s epoch for each sleep state. To account for interindividual differences in absolute EEG power spectra, power spectra density in each frequency bin and for each state was expressed as a percentage of a reference value, calculated across the baseline day for each individual mouse as the mean total EEG power spectra across all frequency bins (0.5–30 Hz) and all behavioral states. This reference value was weighted so that for each animal each state contributed equally to the total EEG power [21]. Thus, EEG power spectra were normalized, expressed as a fractional percentage of total power spectra within the 0.5–30 Hz range, and presented in 0.5 Hz bins in the frequency range of 0.5–30 Hz for 24 hr periods. Additionally, NREM sleep δ power spectra in the EEG (0.5–4 Hz) were determined as a fractional percentage of total EEG power spectra across a 24 hr period in 2 hr time bins for each individual mouse. EEG power spectra analysis was performed for light and dark periods during for each sleep state during spontaneous sleep and sleep after experimental treatments. Because one mouse had corrupted baseline recordings during wakefulness, these data were excluded from analyses of normalized spectra. Memory testing Sst-Cre+/−; Nos1flox/flox (N = 8) and Sst-Cre−/−; Nos1flox/flox (N = 8) littermate controls were tested in the FTR and NOR tasks. Both tasks exploit the innate preference of mice to interact with a novel sensory stimulus over a familiar one. The preference for novel stimuli is quantified as the preference ratio, which increases as a function of time spent exploring the novel stimulus relative to the familiar one in wild-type mice. To be recognized as familiar upon repeat exposure, a stimulus must be encoded in memory. These assays are thus memory-dependent, and memory deficits are manifested as a loss of preference for novelty (i.e., a reduction in the preference ratio). We assessed performance on both tasks to discern genotype-specific deficits in cortex-dependent somatosensory function (FTR) and both cortex- and hippocampus-dependent visuospatial (NOR) modalities. Before the start of these behavioral tasks, we performed 3 day successive handling and 2 day successive 10 min habituation in the behavioral arena (black-walled open field; 25 × 25 × 50 cm). On day 1, mice were subjected to a sampling period in which they experienced a single texture (smooth floor) in an arena for 10 min. The mice then remained in their home cage for 24 hr, and on day 2 were subjected for 10 min to a testing period, during which the mice experienced two textures on opposing halves of the floor: smooth floor (familiar stimulus) and grooved floor (novel stimulus). This procedure was done according to the FTR task, which is the cortical-dependent task that does not depend on the hippocampus [22]. We then proceeded with the NOR task on day 3. The cage contained a uniformly familiar floor texture (smooth floor), but one of the objects made familiar on days 1–2 was replaced with a novel one. The mouse was allowed to explore both novel and familiar objects for 10 min. The sampling and testing periods started between ZT 4 and 6. Video recordings were obtained during both training and test sessions. The light intensity in the cage was maintained at 60–70 lux. Interactions with objects were quantified by visual off-line scoring of the video recordings by a trained observer using EthoVision XT 13.0 (Noldus Information Technology, Leesburg, VA). The exploratory behavior towards objects was defined as each instance in which a mouse’s nose touched the object or was oriented toward the object and came within 2 cm of it. Sitting on the objects was not considered. Preference ratio was calculated as the time spent on exploration of the novel object or texture during first 4 min of the sampling period, divided by the time spent exploring the novel object or texture plus the time exploring familiar object or texture. Immunohistochemistry As previously described [10], mice were anesthetized with an intraperitoneal injection of sodium pentobarbital and perfused with 10% formalin in phosphate-buffered saline (PBS). Mouse brains were placed in 10% formalin overnight and then placed in a 30% sucrose solution at 4°C. Coronal sections (40 μm) were made on a freezing microtome. nNOS immunostaining. nNOS immunohistochemical analysis was performed in Sst-Cre+/−; Nos1flox/flox mice (N = 5) and their littermate control mice (N = 5). The coronal sections were washed in PBS, treated with 1% hydrogen peroxide in PBS, washed in PBS, and blocked in 5% donkey serum and 1% trition X-100 in PBS. The brain sections were incubated overnight in mouse anti-nNOS antibodies (Sigma Aldrich, St. Louis, MO) in PBS (1:2000) at room temperature. Thereafter, sections were washed in PBS and incubated in biotinylated donkey anti-mouse IgG (Jackson Immuno Research, Westgrove, PA) in blocking solution (1:500) for 2 hr. Then, sections were washed in PBS and incubated in ABC (1:200; Vectastain ABC kit; Vector Laboratories, Burlingame, CA) for 2 hr. Sections were then washed in PBS and placed in a diaminobenzidine tetrahydrochloride and nickel chloride solution. The sections were then washed in PBS to stop the reaction. The tissue sections were mounted onto gelatin-coated slides and coverslips were applied using Vector Mount (Vector Laboratories, Burlingame, CA). nNOS/SST double staining. To determine nNOS/SST coexpression in the brain, brain sections were incubated in the blocking buffer for 2 hr and then overnight in a combination of mouse anti-nNOS (1:1000; Sigma–Aldrich) and rat anti-SST (1:1000; MAB354, Chemicon) antibody. The sections were then rinsed in PBS, incubated in a mixture of donkey DyLight 488 anti-mouse IgG (1:500; Jackson ImmunoResearch, West Grove, PA) and biotinylated donkey anti-rat IgG (1:500; Jackson ImmunoResearch, West Grove, PA), again rinsed in PBS, and incubated in DyLight 594 streptavidin conjugate (1:500; Jackson ImmunoResearch, West Grove, PA). After rinsing in PBS, all sections were mounted on gelatin-coated slides and cover-slipped using Fluoromount mounting media (Electron Microscopy Sciences, Hatfield, PA). nNOS/RFP immunofluorescent double staining. For the double-label study in Sst-Cre; Ai14 mice (N = 5), brain sections were processed for nNOS/RFP double-fluorescence labeling. After 2 hr of incubation in the blocking buffer, the sections were incubated overnight in a combination of mouse anti-nNOS (1:1000; Sigma–Aldrich) and rabbit anti-red fluorescent protein (RFP; 1:1500; Rockland Immunochemicals, Gilbertsville, PA) antibody. The sections were then rinsed in PBS, incubated in a mixture of donkey DyLight 488 anti-mouse IgG (1:500; Jackson ImmunoResearch, West Grove, PA) and biotinylated donkey anti-rabbit IgG (1:500; Jackson ImmunoResearch, West Grove, PA), again rinsed in PBS, and incubated in DyLight 594 streptavidin conjugate (1:500; Jackson ImmunoResearch, West Grove, PA). After rinsing in PBS, all sections were mounted on gelatin-coated slides and cover-slipped using Fluoromount mounting media (Electron Microscopy Sciences, Hatfield, PA). NREM sleep slow-wave detection and analysis NREM slow-wave detection was performed in Sst-Cre+/−; Nos1flox/flox (N = 10) and Sst-Cre−/−; Nos1flox/flox (N = 9) as described previously [23]. Detection and analysis was performed using a custom-written MATLAB script (The Math Works, Inc., Natick, MA) which is available upon request. First, EEG data were bandpass filtered between 0.2 and 4.5 Hz. Then the timing of all positive-to-negative and negative-to-positive signal zero crossings was determined, as well as the negative and positive peaks between successive crossings. The intervals between consecutive positive-to-negative crossings were then calculated, with putative slow-wave events defined as cases where the length of the interval was between 0.5 and 2 s (0.5–4 Hz). Additionally, the minimum amplitude and minimum-to-maximum amplitude ratio was required to be greater than 66.6% of the average of the respective amplitude values across the whole recording epoch for inclusion. Analysis of detected slow waves was based on that previously described [24]. The total duration of detected slow waves, along with the duration of both the negative and positive phases of each event, was calculated. The difference in μV between the negative and positive peaks was determined as the amplitude of the slow waves. The slope of each slow wave was determined as the velocity of the change between the negative and positive peaks. The amplitudes and slopes of the slow waves were normalized as a percentage change of 24 hr ad libitum baseline sleep or time-of-day matched ad libitum sleep for recovery sleep responses. Sleep spindle analysis Sleep spindles were detected in Sst-Cre+/−; Nos1flox/flox (N = 10) and Sst-Cre−/−; Nos1flox/flox (N = 9) using an automated algorithm developed in house using MATLAB (Math Works, Inc., Natick, MA). Briefly, EEG data were band-pass filtered across the sigma/spindle frequency range (10–15 Hz; Butterworth filter). Next, the root-mean-square (RMS) power was calculated to provide an upper envelope of the band-pass signal, which was then exponentially transformed to accentuate spindle-generated peaks above background. We employed an adaptive threshold approach where putative spindle peaks were identified in RMS transformed data via crossing of an upper threshold value, set as 2.5× the mean RMS power of EEG across all NREM epochs. Additional inclusion criteria included a minimum spindle duration of 0.5 s, based on crossing of a lower threshold set at 1× mean RMS power, and a minimum interevent interval of 0.5 s. This algorithm was used to determine the total number of spindles across all sleep/wake states and the spindle density during NREM sleep (i.e., the number of NREM spindles/NREM min), the duration of each individual spindle during NREM sleep (seconds), the amplitude of the spindle, and the median peak of the greatest value of each individual spindle during NREM sleep that were normalized to 24 hr ad libitum NREM sleep or to time-of-day matched ad libitum NREM sleep. The MATLAB script that we developed is available upon request. Statistical analyses Groups were compared by two-tailed t-tests or two-way analysis of variance (ANOVA) followed by Tukey’s honestly significant difference test using SPSS software (IBM). For comparison of EEG power spectra, a mixed repeated-measures ANOVA was performed first as an omnibus test. When significant interactions were found, single bins were compared between genotypes using t-tests [25]. Statistical significance was set at p < 0.05, and results are reported as mean ± SEM. Results Mouse model of selective ablation of nNOS expression in type I nNOS neurons in the cerebral cortex and caudate-putamen To produce mice with a cell-specific knockout of nNOS in type I neurons, we crossed nNOSflox/flox mice with Ssttm2.1(cre)Zjh/J mice. Sst-Cre−/−; Nos1flox/flox mice that were homozygous for floxed nNOS but not expressing Cre recombinase served as controls for homozygous floxed nNOS and expressing Cre recombinase under the control of an SST promoter (Sst-Cre+/−; Nos1flox/flox mice). In Sst-Cre−/−; Nos1flox/flox mice, the pattern of nNOS immunostaining in cortex (Figure 1A and Supplementary Figure S2B) was consistent with nNOS expression reported in wild-type animals in other studies [26–28]. Similar to the reported studies of brain areas expressing nNOS-positive neurons [29], intensely stained immuno-positive nNOS neurons were observed in the cerebral cortex, caudate-putamen, hippocampus, pedunculopontine tegmental nucleus, locus coeruleus, laterodorsal tegmental nucleus, interpeduncular nucleus, basal forebrain, and hypothalamus of Sst-Cre−/−; Nos1flox/flox control mice (data not shown). Figure 1. Open in new tabDownload slide Selective ablation of nNOS expression in type I nNOS neurons in Sst-Cre+/−; Nos1flox/flox mice. (A, B) Photomicrographs of a brain section of Sst-Cre−/−; Nos1flox/flox mouse (A) and its littermate Sst-Cre+/−; Nos1flox/flox control mouse immunostaining by DAB method (black color). In the cerebral cortex, positive type I nNOS cells are absent in the cell-specific knockout mice (B) but present in the control mice (A). (C, C′, and C″) nNOS immunofluorescence (green) and RFP fluorescence (red; SST containing cells) in the brain of Sst-Cre; Ai14 mice. Both nNOS-positive cells (C) and RFP-positive cells (C′) are present in the cerebral cortex (Cx) and caudate-putamen (CPu), but RFP-positive cells are more numerous. Merged image (C″) demonstrates that all nNOS-positive cells co-label with RFP (yellow). (D and E) The presence of nNOS-positive neurons in the dorsolateral periaqueductal gray (DLPAG) and pedunculopontine tegmental nucleus (PPTg) of Sst-Cre+/−; Nos1flox/flox mice identified by DAB immunostaining (D). nNOS immunofluorescence (E) and RFP fluorescence (E′) do not co-localize (E″) in the PPTg of Sst-Cre; Ai14 mice. Aq, aqueduct. (F andG) The presence of nNOS-positive neurons in the locus coeruleus (LC) and dorsal tegmental nucleus, central (DTG) of Sst-Cre+/−; Nos1flox/flox mice identified by DAB immunostaining (F). nNOS immunofluorescence (G) and RFP fluorescence (G′) do not co-localize (G″) in the LC of Sst-Cre; Ai14 mice. Scale bar = 100 µm. Figure 1. Open in new tabDownload slide Selective ablation of nNOS expression in type I nNOS neurons in Sst-Cre+/−; Nos1flox/flox mice. (A, B) Photomicrographs of a brain section of Sst-Cre−/−; Nos1flox/flox mouse (A) and its littermate Sst-Cre+/−; Nos1flox/flox control mouse immunostaining by DAB method (black color). In the cerebral cortex, positive type I nNOS cells are absent in the cell-specific knockout mice (B) but present in the control mice (A). (C, C′, and C″) nNOS immunofluorescence (green) and RFP fluorescence (red; SST containing cells) in the brain of Sst-Cre; Ai14 mice. Both nNOS-positive cells (C) and RFP-positive cells (C′) are present in the cerebral cortex (Cx) and caudate-putamen (CPu), but RFP-positive cells are more numerous. Merged image (C″) demonstrates that all nNOS-positive cells co-label with RFP (yellow). (D and E) The presence of nNOS-positive neurons in the dorsolateral periaqueductal gray (DLPAG) and pedunculopontine tegmental nucleus (PPTg) of Sst-Cre+/−; Nos1flox/flox mice identified by DAB immunostaining (D). nNOS immunofluorescence (E) and RFP fluorescence (E′) do not co-localize (E″) in the PPTg of Sst-Cre; Ai14 mice. Aq, aqueduct. (F andG) The presence of nNOS-positive neurons in the locus coeruleus (LC) and dorsal tegmental nucleus, central (DTG) of Sst-Cre+/−; Nos1flox/flox mice identified by DAB immunostaining (F). nNOS immunofluorescence (G) and RFP fluorescence (G′) do not co-localize (G″) in the LC of Sst-Cre; Ai14 mice. Scale bar = 100 µm. In Sst-Cre+/−; Nos1flox/flox mice, intensely stained nNOS neurons (i.e., type I nNOS cells) were absent throughout the cerebral cortex but SST immuno-positive neurons were present in that brain area (Supplementary Figure S2). Therefore, the loss of nNOS immuno-positive cells was apparent within the cerebral cortex in these mice (Figure 1B). Loss of nNOS-positive cells was also seen in the caudate-putamen of these mice. However, there were several examples of nNOS immunostaining in additional brain areas of Sst-Cre+/−; Nos1flox/flox mice including the hippocampus, pedunculopontine tegmental nucleus, locus coeruleus, laterodorsal tegmental nucleus, interpeduncular nucleus, basal forebrain, and hypothalamus (Figure 1, D and F and Supplementary Figure S3). Intensely stained immuno-positive nNOS neurons were observed in all these brain regions, and the pattern of nNOS immunostaining was similar to that of the control Sst-Cre−/−; Nos1flox/flox mice. Thus, apparent differences in the number of nNOS-immunostained cells between Sst-Cre+/−; Nos1flox/flox mice and Sst-Cre−/−; Nos1flox/flox mice appear to be localized to the cerebral cortex and caudate-putamen—areas where type I nNOS cells are activated during recovery sleep after sleep loss [10]. To determine which SST-containing cells express nNOS throughout the brain, we analyzed the co-expression pattern of RFP and nNOS in the offspring of Ssttm2.1(cre)Zjh/J mice crossed with Ai14 (B6;129S6-Gt(ROSA)26Sortm14(CAG-tdTomato)Hze/J) reporter mice (i.e., Sst-Cre+/−; Ai14 mice). Since RFP was expected to be expressed in all SST-Cre containing cells in Sst-Cre+/−; Ai14 mice, we looked for the presence of nNOS/RFP double-labeled cells in these mice. nNOS was co-localized with RFP in all type I nNOS neurons in the cerebral cortex and caudate-putamen (Figure 1, C, C′, and C″). nNOS was also colocalized with RFP in the neurons in the olfactory bulb, basolateral amygdala, corpus callosum, and a few neurons in the basal forebrain but not in any other brain areas (Figure 1, E, E′, E″, G, G′, and G″). These data demonstrate that Sst-Cre+/−; Nos1flox/flox mice represent a model of selective ablation of nNOS expression in type I nNOS neurons in the brain. SST is reported to be expressed in about 18% of type II nNOS neurons in the cerebral cortex [5] and 4%–7% of all nNOS-immunoreactive neurons in the hippocampus [30]. Therefore, nNOS expression is expected to be reduced in these neurons by crossing Sst-Cre+/− mice with nNOSflox/flox mice. nNOS expression distributed in the cytoplasm is very low in type II nNOS cells [5]. Consequently, our immunostaining procedures were not sensitive enough to detect positively stained type II nNOS neurons. Thus, it is likely that the SST+/type I nNOS neurons drive the observed effects, and the SST+/type II nNOS neuron influence is minimal. Mice lacking nNOS in SST cells exhibit similar spontaneous sleep state characteristics as wild-type control mice Male Sst-Cre+/−; Nos1flox/flox mice (9.4 ± 0.3 weeks old, n = 10) and Sst-Cre−/−; Nos1flox/flox control mice (9.4 ± 0.2 weeks old, n = 9) were used for the polysomnography studies. Spontaneous sleep/wake state amounts and the characteristics of sleep/wake bout episode durations and episode frequencies were similar between Sst-Cre+/−; Nos1flox/flox and Sst-Cre−/−; Nos1flox/flox control mice (Figure 2, A–D). When examined across the 24 hr period, Sst-Cre+/−; Nos1flox/flox mice and Sst-Cre−/−; Nos1flox/flox mice exhibited similar durations of wakefulness [(Sst-Cre+/−; Nos1flox/flox: 52.06 ± 0.70%); (Sst-Cre−/−; Nos1flox/flox: 50.96 ± 1.10%)], NREM sleep [(Sst-Cre+/−; Nos1flox/flox: 42.49 ± 0.79%); (Sst-Cre−/−; Nos1flox/flox: 42.80% ± 1.03)], and REM sleep [(Sst-Cre+/−; Nos1flox/flox: 5.36 ± 0.35%); (Sst-Cre−/−; Nos1flox/flox: 6.13 ± 0.19%)], t-test, p > 0.05. NREM sleep and REM sleep amounts displayed diurnal variations, with significantly greater amounts occurring during the light-on than the light-off periods (Figure 2, B and C). However, the profile of these changes was nearly identical between the genotypes during NREM sleep [(Sst-Cre+/−; Nos1flox/flox: Light period: 58.71 ± 1.41%, Dark period: 26.28 ± 1.57%); (Sst-Cre−/−; Nos1flox/flox: Light period: 58.62 ± 0.95%, Dark period: 26.98 ± 1.81%)] and REM sleep [(Sst-Cre+/−; Nos1flox/flox: Light period: 7.54 ± 0.40%, Dark period: 3.19 ± 0.34%); (Sst-Cre−/−; Nos1flox/flox: Light period: 8.34 ± 0.26%, Dark period: 3.91 ± 0.35%)] (Figure 2, A–C). We did not observe any significant differences in either the number or duration of sleep bouts between genotypes (Figure 2, D and E). Since a previous study showed that the average NREM sleep bout episode duration was approximately one-third shorter in nNOS KO mice than in wild-type control mice [18], we plotted NREM sleep bout episode duration and episode frequency in histograms for Sst-Cre+/−; Nos1flox/flox mice and Sst-Cre−/−; Nos1flox/flox mice. Our analysis determined that there were no differences in bout episode duration or episode frequency between the genotypes across wakefulness, NREM sleep, or REM sleep states (Supplementary Figure S4A–C). Figure 2. Open in new tabDownload slide Total sleep amounts, the length and duration of sleep bouts, are similar in Sst-Cre+/−; Nos1flox/flox mice and their littermate control mice. (A–C) Percentage of time spent in wakefulness, NREM sleep, and REM sleep (mean ± SEM) during the light–dark phase is similar in the cell-specific nNOS knockout mice and control mice. (D) Number of sleep bouts calculated during 24 hr of spontaneous sleep is similar in the cell-specific nNOS knockout mice and control mice. (E) Average bout duration calculated during 24 hr of spontaneous sleep is similar in the cell-specific nNOS knockout mice and control mice. Figure 2. Open in new tabDownload slide Total sleep amounts, the length and duration of sleep bouts, are similar in Sst-Cre+/−; Nos1flox/flox mice and their littermate control mice. (A–C) Percentage of time spent in wakefulness, NREM sleep, and REM sleep (mean ± SEM) during the light–dark phase is similar in the cell-specific nNOS knockout mice and control mice. (D) Number of sleep bouts calculated during 24 hr of spontaneous sleep is similar in the cell-specific nNOS knockout mice and control mice. (E) Average bout duration calculated during 24 hr of spontaneous sleep is similar in the cell-specific nNOS knockout mice and control mice. Mice lacking nNOS in SST cells show specific defects in the EEG low-δ power frequency range during NREM sleep EEG spectral power is expressed as a proportion of total EEG spectral power over all frequencies (0–30 Hz) in 0.5 Hz frequency bins and was analyzed using two-way ANOVAs. During spontaneous wakefulness and REM sleep (Figure 3, A and C), the profile of EEG spectral power was nearly identical between Sst-Cre+/−; Nos1flox/flox mice and Sst-Cre−/−; Nos1flox/flox control mice. A main effect was revealed for both genotype and EEG spectral power frequency during spontaneous NREM sleep (F1,1054 = 41.928, p < 0.0001 and F61,1054 = 279.905, p < 0.0001, respectively; Figure 3B). Additionally, a frequency × genotype interaction was observed (F61,1054 = 5.946, p < 0.0001). Sst-Cre+/−; Nos1flox/flox mice showed significantly lower spectral power in the low-δ frequency range (0.5–3.0 Hz) during NREM sleep than Sst-Cre−/−; Nos1flox/flox control mice (27.8% reduction, Figure 3B). The greatest difference observed between these genotypes was in the frequency range between 1 and 2 Hz (33.7% reduction, Figure 3D). The typically found circadian variation of enhanced NREM sleep δ power (0.5–4.0 Hz) during the dark period in mice was blunted in Sst-Cre+/−; Nos1flox/flox mice compared with Sst-Cre−/−; Nos1flox/flox mice (Figure 3E). Consequently, differences between the light-on and light-off periods were found between the genotypes (light-on/light-off: 120.4 ± 2.3 in Sst-Cre+/−; Nos1flox/flox mice vs 107.3 ± 3.1 in Sst-Cre−/−; Nos1flox/flox mice; p = 0.01 in t-test). EEG power spectra during wakefulness and REM sleep, total amounts of sleep and wakefulness, and number and duration of sleep episode bouts were not significantly different between Sst-Cre+/−; Nos1flox/flox mice and Sst-Cre−/−; Nos1flox/flox control mice. Figure 3. Open in new tabDownload slide Sst-Cre+/−; Nos1flox/flox mice show specific defects in the low δ range of EEG power during spontaneous sleep. (A–C) Spectral power was expressed as a proportion of spectral power of all frequencies (0–30 Hz) during each 0.5 Hz frequency bin and analyzed using two-way ANOVA. During wakefulness (A) and REM sleep (C), the profile of spectral power was nearly identical between Sst-Cre+/−; Nos1flox/flox mice and their littermate control mice (Sst-Cre−/−; Nos1flox/flox mice). During NREM sleep (B), Sst-Cre+/−; Nos1flox/flox mice showed significantly less spectral power in the low δ range than the littermate control mice. (D) The highest difference in power ratio between Sst-Cre+/−; Nos1flox/flox mice and their littermate control mice was found to be around 1.5 Hz. (E) Circadian variation of the δ power was blunted during NREM sleep in Sst-Cre+/−; Nos1flox/flox mice compared with Sst-Cre−/−; Nos1flox/flox control mice. *p < 0.05 between genotypes. Figure 3. Open in new tabDownload slide Sst-Cre+/−; Nos1flox/flox mice show specific defects in the low δ range of EEG power during spontaneous sleep. (A–C) Spectral power was expressed as a proportion of spectral power of all frequencies (0–30 Hz) during each 0.5 Hz frequency bin and analyzed using two-way ANOVA. During wakefulness (A) and REM sleep (C), the profile of spectral power was nearly identical between Sst-Cre+/−; Nos1flox/flox mice and their littermate control mice (Sst-Cre−/−; Nos1flox/flox mice). During NREM sleep (B), Sst-Cre+/−; Nos1flox/flox mice showed significantly less spectral power in the low δ range than the littermate control mice. (D) The highest difference in power ratio between Sst-Cre+/−; Nos1flox/flox mice and their littermate control mice was found to be around 1.5 Hz. (E) Circadian variation of the δ power was blunted during NREM sleep in Sst-Cre+/−; Nos1flox/flox mice compared with Sst-Cre−/−; Nos1flox/flox control mice. *p < 0.05 between genotypes. Mice lacking nNOS in SST cells exhibit similar EEG theta (θ) power as control mice Previous studies in mice and rats suggest that EEG θ power during wakefulness is an indicator of the homeostatic sleep pressure because it positively correlated with EEG δ power during subsequent NREM sleep [31–33]. Therefore, we compared the waking EEG during the last 3 hr of the sleep deprivation (i.e., during the 3 hr immediately preceding sleep onset) with time-of-day matched ad libitum baseline sleep as well as during the last 3 hr of the dark period of ad libitum sleep (i.e., a time before the highest sleep propensity time of the day) with the first 3 hr of the light period (ZT 0–3) of ad libitum sleep in order to identify frequency components that could have contributed to the EEG δ power differences observed during NREM sleep between Sst-Cre+/−; Nos1flox/flox mice and Sst-Cre−/−; Nos1flox/flox control mice. In 3 hr of the baseline dark periods (ZT 21–24) vs. the beginning of the light period (ZT 0–3) as well as the sleep deprivation (ZT 3–6) vs. baseline (ZT 3–6), the waking EEG showed clear enhanced θ activity in both strains of mice (Supplementary Figure S5). The EEG power in θ frequency range of 7–9 Hz was higher during sleep deprivation than baseline dark periods in both Sst-Cre+/−; Nos1flox/flox mice and Sst-Cre−/−; Nos1flox/flox mice (sleep deprivation/baseline ratio, paired t-test: Sst-Cre+/−; Nos1flox/flox mice, 1.216 ± 0.055, df = 7, t = 4.731, p = 0.002; Sst-Cre−/−; Nos1flox/flox mice, 1.202 ± 0.049, t = 4.270, df = 8, p = 0.003). However, no differences were observed in the waking EEG between Sst-Cre+/−; Nos1flox/flox mice and their littermate control mice during the last 3 hr of the sleep deprivation. There was also no difference in the fractional % of θ power in the frequency range of 5–9 Hz between the genotypes during the last 3 hr of the baseline dark periods (unpaired t-test: Sst-Cre+/−; Nos1flox/flox mice, 0.0216 ± 0.0008; Sst-Cre−/−; Nos1flox/flox mice, 0.0218 ± 0.0005; t = 0.138, df = 16, p = 0.892), although the peak of the θ frequency band was found to be different between Sst-Cre+/−; Nos1flox/flox mice and Sst-Cre−/−; Nos1flox/flox mice (7.5 and 6.5 Hz, respectively). In addition, we did not observe any significant differences in EEG theta power during spontaneous REM sleep (data not shown). Homeostatic enhancements in sleep amounts after sleep deprivation are preserved in mice lacking nNOS in SST cells The homeostatic increase in NREM and REM sleep amounts and reductions in wakefulness typically observed after acute sleep deprivation was found in the first 2 hr of recovery sleep (ZT 6–8) in both Sst-Cre+/−; Nos1flox/flox and Sst-Cre−/−; Nos1flox/flox control mice (Figure 4, A and B). A two-way ANOVA of the effect of genotype (Sst-Cre+/−; Nos1flox/flox and Sst-Cre−/−; Nos1flox/flox) and homeostatic pressure (baseline and recovery sleep after sleep deprivation) on sleep amounts was conducted. A main effect was found for sleep deprivation enhancing NREM sleep amounts (F1,34 = 25.32, p < 0.0001). A main effect of sleep deprivation enhancing REM sleep amounts was also found (F1,34 = 39.87, p < 0.0001). However, there was no significant main effect of genotype for either NREM or REM sleep amounts or the interaction of genotype with sleep deprivation procedures. Collectively, these data indicate that enhanced sleep amounts after sleep deprivation are preserved in Sst-Cre+/−; Nos1flox/flox mice. The Multiple Sleep Latency Test (MSLT) is used to quantify sleepiness used in humans that has been adapted for use in rodents [18, 34]. Since we did not observe significant differences in sleep amounts between the genotypes during recovery after sleep deprivation, we employed the MSLT as an additional homeostatic measure. Nevertheless, Sst-Cre+/−; Nos1flox/flox mice and Sst-Cre−/−; Nos1flox/flox control mice exhibited similar amounts of NREM and REM sleep (Figure 4, C, D, and F) and latency to reach NREM sleep (Figure 4E) during the 20 min nap opportunities of the test. Accordingly, these data suggest that Sst-Cre+/−; Nos1flox/flox mice are not more or less sleepy than controls after sleep deprivation. Figure 4. Open in new tabDownload slide The homeostatic response of total sleep amounts and the length and duration of sleep bouts is similar in Sst-Cre+/−; Nos1flox/flox mice and their littermate control mice. (A and B) Percentage of time spent in wakefulness, NREM sleep, and REM sleep during 120 min of recovery sleep after sleep deprivation (ZT6–ZT8) did not differ between Sst-Cre−/−; Nos1flox/flox mice (A) and Sst-Cre+/−; Nos1flox/flox mice (B). (C and D) Percentage of time spent in NREM sleep (C) and REM sleep (D) during MSLT did not differ between Sst-Cre−/−; Nos1flox/flox mice and Sst-Cre+/−; Nos1flox/flox mice. (E and F) NREM sleep latency (E) and percentage of time spent in NREM sleep (F) calculated during the entire 200 min period of MSLT did not differ between Sst-Cre−/−; Nos1flox/flox mice and Sst-Cre+/−; Nos1flox/flox mice. The light and dark bars indicate period of sleep deprivation and sleep opportunities of MSLT. *p < 0.05 between recovery and baseline. Figure 4. Open in new tabDownload slide The homeostatic response of total sleep amounts and the length and duration of sleep bouts is similar in Sst-Cre+/−; Nos1flox/flox mice and their littermate control mice. (A and B) Percentage of time spent in wakefulness, NREM sleep, and REM sleep during 120 min of recovery sleep after sleep deprivation (ZT6–ZT8) did not differ between Sst-Cre−/−; Nos1flox/flox mice (A) and Sst-Cre+/−; Nos1flox/flox mice (B). (C and D) Percentage of time spent in NREM sleep (C) and REM sleep (D) during MSLT did not differ between Sst-Cre−/−; Nos1flox/flox mice and Sst-Cre+/−; Nos1flox/flox mice. (E and F) NREM sleep latency (E) and percentage of time spent in NREM sleep (F) calculated during the entire 200 min period of MSLT did not differ between Sst-Cre−/−; Nos1flox/flox mice and Sst-Cre+/−; Nos1flox/flox mice. The light and dark bars indicate period of sleep deprivation and sleep opportunities of MSLT. *p < 0.05 between recovery and baseline. The enhanced homeostatic EEG low-δ power during NREM sleep after sleep loss response is selectively impaired in mice lacking nNOS in SST cells EEG spectral power and sleep/wake state amounts were calculated in mice during subsequent sleep (i.e., recovery sleep; ZT 6–12) after 6 hr of sleep deprivation (ZT 0–6; Figures 5 and 6). A repeated measures ANOVA with a Greenhouse-Geisser correction determined that mean EEG δ power during NREM sleep differed over time (F3.297,46.162 = 36.198, p < 0.0001). A time × genotype interaction was also found (F3.297,46.162 = 6.194, p = 0.001). Post hoc analysis between the groups at each level of each factor was then conducted. When expressed as a proportion of total EEG spectral power of all frequencies (0–30 Hz), an increase in EEG spectral power in the δ power range (0.5–4.0 Hz) of the EEG was evident during the first 2 hr of NREM sleep immediately after sleep deprivation for Sst-Cre−/−; Nos1flox/flox mice (p < 0.05, Bonferroni-adjusted for multiple comparisons). However, this enhancement was present only during the first 30 min of recovery sleep after sleep deprivation in Sst-Cre+/−; Nos1flox/flox mice (Figure 5B). Because the slow oscillation component of SWA may be particularly important for the consolidation of memories [35, 36], NREM sleep EEG δ power during recovery sleep was also normalized by the baseline values of the 2 last hr of the light period in the frequency range of 0–1.5 Hz (low δ power), 1.5–3.0 Hz (medium δ power), and 3.0–4.5 Hz (high δ power; Figure 5, C–E). The low, medium, and high EEG δ power measures were then compared between genotypes by mixed-design repeated measures ANOVA. Mean EEG δ power in the frequency range of 0–1.5 Hz differed over time (F4.516,63.226 = 20.780, p < 0.0001). A significant time × genotype interaction was also found (F4.516,63.226 = 16.118, p < 0.0001). Mean EEG δ power in the frequency range of 0–1.5 Hz increased during NREM sleep in the first available 90 min immediately after sleep deprivation in both Sst-Cre+/−; Nos1flox/flox mice and Sst-Cre−/−; Nos1flox/flox control mice, although Sst-Cre−/−; Nos1flox/flox control mice had significantly greater enhancements (unpaired t-test: Sst-Cre+/−; Nos1flox/flox mice, 15.01 ± 4.06%; Sst-Cre−/−; Nos1flox/flox mice, 63.98 ± 8.76%; t = 5.25, df = 17, p < 0.0001). A main effect for time was also revealed for the medium (1.5–3.0 Hz) and high (3.0–4.5 Hz) EEG δ power frequency ranges (F11,154 = 32.161, p < 0.0001 and F11,154 = 38.145, p < 0.0001, respectively). However, there was no significant difference in the profile of the EEG δ power changes between the genotypes in the frequency range groupings above 1.5 Hz (Figure 5, D and E). Thus, the diminution of homeostatic response to sleep deprivation of SWA in the low EEG δ power range was found to be very selective in Sst-Cre+/−; Nos1flox/flox mice. Figure 5. Open in new tabDownload slide The homeostatic response of SWA to sleep loss is impaired in the low δ range in Sst-Cre+/-; Nos1flox/flox mice. (A) Percentage of time spent in wakefulness of Sst-Cre+/−; Nos1flox/flox mice and control mice during the 6 hr sleep deprivation period and subsequent recovery sleep. The graph shows the efficacy of our sleep deprivation procedures. (B) Spectral power was calculated in mice during recovery sleep after sleep deprivation. When expressed as a proportion of spectral power of all frequencies (0–30 Hz), an increase in spectral power in the low δ power range (0.5–4.0 Hz) of the EEG was clearly evident during the period of about 120 min after termination of sleep deprivation in Sst-Cre−/−; Nos1flox/flox mice but greatly diminished in Sst-Cre+/−; Nos1flox/flox mice. (C–E) When expressed as percent from baseline, NREM sleep δ power during recovery sleep was reduced in the frequency range of <1.5 Hz (C), but was similar to controls in the range of 1.5–3.0 Hz (D) and 3.0–4.5 Hz (E). *p < 0.05 between genotypes; #p < 0.05 between time groups. Figure 5. Open in new tabDownload slide The homeostatic response of SWA to sleep loss is impaired in the low δ range in Sst-Cre+/-; Nos1flox/flox mice. (A) Percentage of time spent in wakefulness of Sst-Cre+/−; Nos1flox/flox mice and control mice during the 6 hr sleep deprivation period and subsequent recovery sleep. The graph shows the efficacy of our sleep deprivation procedures. (B) Spectral power was calculated in mice during recovery sleep after sleep deprivation. When expressed as a proportion of spectral power of all frequencies (0–30 Hz), an increase in spectral power in the low δ power range (0.5–4.0 Hz) of the EEG was clearly evident during the period of about 120 min after termination of sleep deprivation in Sst-Cre−/−; Nos1flox/flox mice but greatly diminished in Sst-Cre+/−; Nos1flox/flox mice. (C–E) When expressed as percent from baseline, NREM sleep δ power during recovery sleep was reduced in the frequency range of <1.5 Hz (C), but was similar to controls in the range of 1.5–3.0 Hz (D) and 3.0–4.5 Hz (E). *p < 0.05 between genotypes; #p < 0.05 between time groups. Figure 6. Open in new tabDownload slide Sst-Cre+/−; Nos1flox/flox mice show impairment in memory in floor texture recognition (FTR) test but not in novel object recognition (NOR) test. (A–D) The procedure consisted of habituation (A), exploration of two identical objects on a smooth floor (B), exploration of two identical objects on a half smooth/half grooved floor (C), and exploration of a familiar object plus novel object (D). The preference ratio was significantly lower in Sst-Cre+/−; Nos1flox/flox in FTR test but was not significantly different between Sst-Cre+/−; Nos1flox/flox mice and Sst-Cre−/−; Nos1flox/flox mice in NOR test (E). *p < 0.05 between genotypes. Figure 6. Open in new tabDownload slide Sst-Cre+/−; Nos1flox/flox mice show impairment in memory in floor texture recognition (FTR) test but not in novel object recognition (NOR) test. (A–D) The procedure consisted of habituation (A), exploration of two identical objects on a smooth floor (B), exploration of two identical objects on a half smooth/half grooved floor (C), and exploration of a familiar object plus novel object (D). The preference ratio was significantly lower in Sst-Cre+/−; Nos1flox/flox in FTR test but was not significantly different between Sst-Cre+/−; Nos1flox/flox mice and Sst-Cre−/−; Nos1flox/flox mice in NOR test (E). *p < 0.05 between genotypes. Mice lacking nNOS exhibit reduced slow-wave negative duration-associated hyperpolarization and slow-wave slope homeostatic responses In Sst-Cre+/−; Nos1flox/flox and Sst-Cre−/−; Nos1flox/flox control mice, main effects were found for the amplitudes and slopes of the slow waves, which were greater during the dark period compared with the light period (F1,34 = 16.449, p < 0.001 and F1,34 = 6.809, p = 0.013, respectively; Table 1 and Supplementary Figure S6). An interaction was observed between genotypes and light and dark periods for the amplitudes of individual slow waves (F1,34 = 6.354, p = 0.017). Sst-Cre−/−; Nos1flox/flox control mice exhibited greater amplitudes and slopes for individual slow waves during the dark period vs. light period (F1,16 = 60.575, p < 0.001 and F1,16 = 27.625, p < 0.001), although no significant time-of-day differences were observed in Sst-Cre+/−; Nos1flox/flox mice. Additionally, a main effect of genotype was observed for lower slow-wave durations comprised largely of lower amounts in the negative duration hyperpolarization phase in Sst-Cre+/−; Nos1flox/flox mice compared with control mice (F1,34 = 8.334, p = 0.007 and F1,34 = 14.638, p = 0.001, respectively). The reduced slow-wave duration occurred during the dark period only (F1,34 = 5.691, p = 0.029), although the reduced negative duration in the slow waves was seen in both the dark and light periods (dark period: F1,34 = 8.271, p = 0.010; light period: F1,34 = 6.367, p = 0.022). Table 1. Characteristics of rhythmic events occurring in the EEG during light and dark periods of spontaneous sleep and during the first 6 hr of recovery sleep after sleep deprivation in Sst-Cre−/−; Nos1flox/flox and Sst-Cre+/−; Nos1flox/flox mice: individual NREM sleep slow-wave wavelets Genotype . Duration (ms) . Positive duration (ms) . Negative duration (ms) . Amplitude (% normalized to 24 hr values) . Slope (% normalized to 24 hr values) . Sst-Cre−/−; Nos1flox/flox control mice . . . . . . Dark period 360.155 ± 7.166 165.909 ± 2.467 194.246 ± 4.873 103.615 ± 0.644 102.422 ± 0.647 Light period 352.502 ± 7.038 162.533 ± 3.052 189.969 ± 4.375 98.116 ± 0.290* 98.735 ± 0.270* Baseline 349.716 ± 6.945 161.548 ± 3.276 188.167 ± 4.215 95.904 ± 0.890 97.169 ± 0.777 Recovery sleep 363.914 ± 6.608 165.952 ± 2.126 197.962 ± 4.839 103.514 ± 2.171* 101.688 ± 1.951* Genotype Duration (ms) Positive duration (ms) Negative duration (ms) Amplitude (% normalized to 24 hr values) Slope (% normalized to 24 hs values) Sst-Cre+/−; Nos1flox/flox mice Dark period 341.098 ± 4.002† 162.023 ± 1.799 179.075 ± 2.429† 100.824 ± 1.350 100.484 ± 1.378 Light period 339.267 ± 3.903 160.831 ± 2.269 178.436 ± 1.837† 99.561 ± 0.552 99.703 ± 0.574 Baseline 338.238 ± 3.651 160.340 ± 2.354 177.897 ± 1.661† 97.834 ± 0.607 98.317 ± 0.568 Recovery sleep 340.043 ± 4.426† 161.309 ± 1.777 178.734 ± 2.857† 107.238 ± 4.776 107.447 ± 5.350 Genotype . Duration (ms) . Positive duration (ms) . Negative duration (ms) . Amplitude (% normalized to 24 hr values) . Slope (% normalized to 24 hr values) . Sst-Cre−/−; Nos1flox/flox control mice . . . . . . Dark period 360.155 ± 7.166 165.909 ± 2.467 194.246 ± 4.873 103.615 ± 0.644 102.422 ± 0.647 Light period 352.502 ± 7.038 162.533 ± 3.052 189.969 ± 4.375 98.116 ± 0.290* 98.735 ± 0.270* Baseline 349.716 ± 6.945 161.548 ± 3.276 188.167 ± 4.215 95.904 ± 0.890 97.169 ± 0.777 Recovery sleep 363.914 ± 6.608 165.952 ± 2.126 197.962 ± 4.839 103.514 ± 2.171* 101.688 ± 1.951* Genotype Duration (ms) Positive duration (ms) Negative duration (ms) Amplitude (% normalized to 24 hr values) Slope (% normalized to 24 hs values) Sst-Cre+/−; Nos1flox/flox mice Dark period 341.098 ± 4.002† 162.023 ± 1.799 179.075 ± 2.429† 100.824 ± 1.350 100.484 ± 1.378 Light period 339.267 ± 3.903 160.831 ± 2.269 178.436 ± 1.837† 99.561 ± 0.552 99.703 ± 0.574 Baseline 338.238 ± 3.651 160.340 ± 2.354 177.897 ± 1.661† 97.834 ± 0.607 98.317 ± 0.568 Recovery sleep 340.043 ± 4.426† 161.309 ± 1.777 178.734 ± 2.857† 107.238 ± 4.776 107.447 ± 5.350 *p < 0.05 between light and dark periods or baseline and recovery sleep treatments. †p < 0.05 between genotypes. Open in new tab Table 1. Characteristics of rhythmic events occurring in the EEG during light and dark periods of spontaneous sleep and during the first 6 hr of recovery sleep after sleep deprivation in Sst-Cre−/−; Nos1flox/flox and Sst-Cre+/−; Nos1flox/flox mice: individual NREM sleep slow-wave wavelets Genotype . Duration (ms) . Positive duration (ms) . Negative duration (ms) . Amplitude (% normalized to 24 hr values) . Slope (% normalized to 24 hr values) . Sst-Cre−/−; Nos1flox/flox control mice . . . . . . Dark period 360.155 ± 7.166 165.909 ± 2.467 194.246 ± 4.873 103.615 ± 0.644 102.422 ± 0.647 Light period 352.502 ± 7.038 162.533 ± 3.052 189.969 ± 4.375 98.116 ± 0.290* 98.735 ± 0.270* Baseline 349.716 ± 6.945 161.548 ± 3.276 188.167 ± 4.215 95.904 ± 0.890 97.169 ± 0.777 Recovery sleep 363.914 ± 6.608 165.952 ± 2.126 197.962 ± 4.839 103.514 ± 2.171* 101.688 ± 1.951* Genotype Duration (ms) Positive duration (ms) Negative duration (ms) Amplitude (% normalized to 24 hr values) Slope (% normalized to 24 hs values) Sst-Cre+/−; Nos1flox/flox mice Dark period 341.098 ± 4.002† 162.023 ± 1.799 179.075 ± 2.429† 100.824 ± 1.350 100.484 ± 1.378 Light period 339.267 ± 3.903 160.831 ± 2.269 178.436 ± 1.837† 99.561 ± 0.552 99.703 ± 0.574 Baseline 338.238 ± 3.651 160.340 ± 2.354 177.897 ± 1.661† 97.834 ± 0.607 98.317 ± 0.568 Recovery sleep 340.043 ± 4.426† 161.309 ± 1.777 178.734 ± 2.857† 107.238 ± 4.776 107.447 ± 5.350 Genotype . Duration (ms) . Positive duration (ms) . Negative duration (ms) . Amplitude (% normalized to 24 hr values) . Slope (% normalized to 24 hr values) . Sst-Cre−/−; Nos1flox/flox control mice . . . . . . Dark period 360.155 ± 7.166 165.909 ± 2.467 194.246 ± 4.873 103.615 ± 0.644 102.422 ± 0.647 Light period 352.502 ± 7.038 162.533 ± 3.052 189.969 ± 4.375 98.116 ± 0.290* 98.735 ± 0.270* Baseline 349.716 ± 6.945 161.548 ± 3.276 188.167 ± 4.215 95.904 ± 0.890 97.169 ± 0.777 Recovery sleep 363.914 ± 6.608 165.952 ± 2.126 197.962 ± 4.839 103.514 ± 2.171* 101.688 ± 1.951* Genotype Duration (ms) Positive duration (ms) Negative duration (ms) Amplitude (% normalized to 24 hr values) Slope (% normalized to 24 hs values) Sst-Cre+/−; Nos1flox/flox mice Dark period 341.098 ± 4.002† 162.023 ± 1.799 179.075 ± 2.429† 100.824 ± 1.350 100.484 ± 1.378 Light period 339.267 ± 3.903 160.831 ± 2.269 178.436 ± 1.837† 99.561 ± 0.552 99.703 ± 0.574 Baseline 338.238 ± 3.651 160.340 ± 2.354 177.897 ± 1.661† 97.834 ± 0.607 98.317 ± 0.568 Recovery sleep 340.043 ± 4.426† 161.309 ± 1.777 178.734 ± 2.857† 107.238 ± 4.776 107.447 ± 5.350 *p < 0.05 between light and dark periods or baseline and recovery sleep treatments. †p < 0.05 between genotypes. Open in new tab During the first 2 hr of recovery sleep after sleep deprivation when SWA (0.5–4 Hz frequency range) was significantly different between genotypes, a main effect was found for an enhancement in amplitude and the slope of slow waves compared with time-of-day spontaneous sleep values when analyzing both genotypes (F1,34 = 9.393, p = 0.004 and F1,34 = 5.133, p = 0.030, respectively). Post hoc analysis indicated that Sst-Cre−/−; Nos1flox/flox control mice showed enhanced slow-wave amplitudes and slopes during the first 2 hr of recovery sleep vs. baseline sleep (F1,16 = 10.525, p = 0.005 and F1,16 = 4.629, p = 0.047, respectively), although Sst-Cre+/−; Nos1flox/flox mice did not show any significant differences in these measures. A main effect was found for lower overall duration and negative duration of individual slow waves in Sst-Cre+/−; Nos1flox/flox mice vs. control mice (F1,34 = 10.504, p = 0.003 and F1,34 = 17.747, p < 0.001, respectively). This effect occurred, in part, due to the overall reduced slow-wave duration happening during the baseline period (F1,34 = 9.351, p = 0.007) and reduced negative durations of the slow waves found during the baseline and recovery sleep periods of Sst-Cre+/−; Nos1flox/flox mice compared with control mice (baseline: F1,34 = 5.561, p = 0.031; recovery sleep: F1,34 = 12.298, p = 0.003). Mice lacking nNOS in SST cells exhibit normal NREM sleep spindle responses during spontaneous sleep and recovery sleep Sst-Cre+/−; Nos1flox/flox mice largely showed similar NREM sleep spindle responses during spontaneous sleep and recovery sleep postsleep deprivation as Sst-Cre−/−; Nos1flox/flox control mice (Table 2 and Supplementary Figure S7). In both genotypes, a main effect was found for the reduced total number of spindles regardless of sleep/wake state occurring during the dark period vs. light period (F1,34 = 221.268, p < 0.001). There were fewer total number of spindles regardless of sleep/wake state during the dark period vs. light period for both genotypes (Sst-Cre−/−; Nos1flox/flox: F1,16 = 149.596, p < 0.001; Sst-Cre+/−; Nos1flox/flox: F1,18 = 96.669, p < 0.001). However, greater number of sleep spindles (F1,34 = 450.141, p < 0.001) and sleep spindle density (F1,34 = 10.937, p = 0.002) occurred during the more NREM sleep latent light period than the dark period. Post hoc analysis determined that the number of spindles (Sst-Cre-/−; Nos1flox/flox: F1,16 = 215.525, p < 0.001; Sst-Cre+/−; Nos1flox/flox: F1,18 = 236.694, p < 0.001) and spindle density (Sst-Cre−/−; Nos1flox/flox: F1,16 = 8.619, p = 0.010; Sst-Cre+/−; Nos1flox/flox: F1,18 = 4.694, p = 0.044) taking place during NREM sleep was similarly enhanced during the light period compared with the dark period (Sst-Cre−/−; Nos1flox/flox: F1,16 = 12.602, p = 0.003; Sst-Cre+/−; Nos1flox/flox: F1,18 = 6.541, p = 0.020). A main effect was also found for Sst-Cre+/−; Nos1flox/flox mice having greater sleep spindle durations compared with Sst-Cre−/−; Nos1flox/flox mice (F1,34 = 5.303, p = 0.028). Time-of-day differences in the median of the peak amplitude (Hz) of the sleep spindles were only ~0.1 Hz greater during the dark period compared with the light period. No other significant differences in sleep spindle variables were observed during recovery sleep or between genotypes. Table 2. Characteristics of rhythmic events occurring in the EEG during light and dark periods of spontaneous sleep and during the first 6 hr of recovery sleep after sleep deprivation in Sst-Cre−/−; Nos1flox/flox and Sst-Cre+/−; Nos1flox/flox mice: sleep spindles Genotype . Total spindles (number) . NREM sleep spindles (number) . NREM sleep spindle density (number of spindles /min) . Spindle duration (s) . Spindle amplitude (% normalized to 24 hr values) . Spindle median peak (Hz) . Sst-Cre−/−; Nos1flox/flox control mice . . . . . . . Dark WT 122.472 ± 9.401 89.815 ± 6.867 5.504 ± 0.161 1.450 ± 0.033 98.454 ± 0.431 12.083 ± 0.045 Light WT 247.167 ± 3.944* 213.083 ± 4.833* 6.099 ± 0.123* 1.484 ± 0.028 101.546 ± 0.431 12.019 ± 0.032 Baseline 244.667 ± 6.940 208.315 ± 7.429 6.347 ± 0.111 1.553 ± 0.035 102.672 ± 0.550 11.983 ± 0.045 Recovery sleep 256.463 ± 5.182 225.519 ± 4.812 5.952 ± 0.091* 1.356 ± 0.024* 95.150 ± 4.791* 12.102 ± 0.031* Genotype Total spindles (number) NREM sleep spindles (number) NREM sleep spindle density (number of spindles /min) Spindle duration (s) Spindle amplitude (% normalized to 24 hr values) Spindle median peak (Hz) Sst-Cre+/−; Nos1flox/flox mice Dark KO 118.808 ± 10.800 80.700 ± 5.851 5.230 ± 0.286 1.598 ± 0.077 110.309 ± 9.407 12.071 ± 0.050 Light KO 255.275 ± 8.719* 207.075 ± 5.765* 5.929 ± 0.150* 1.562 ± 0.034 89.691 ± 9.407 11.946 ± 0.044 Baseline 256.467 ± 9.713 205.133 ± 6.818 6.140 ± 0.143 1.588 ± 0.038 89.968 ± 9.428 11.942 ± 0.046 Recovery sleep 258.183 ± 7.560 220.033 ± 5.755 5.681 ± 0.077* 1.411 ± 0.015* 107.208 ± 19.882* 12.061 ± 0.030* Genotype . Total spindles (number) . NREM sleep spindles (number) . NREM sleep spindle density (number of spindles /min) . Spindle duration (s) . Spindle amplitude (% normalized to 24 hr values) . Spindle median peak (Hz) . Sst-Cre−/−; Nos1flox/flox control mice . . . . . . . Dark WT 122.472 ± 9.401 89.815 ± 6.867 5.504 ± 0.161 1.450 ± 0.033 98.454 ± 0.431 12.083 ± 0.045 Light WT 247.167 ± 3.944* 213.083 ± 4.833* 6.099 ± 0.123* 1.484 ± 0.028 101.546 ± 0.431 12.019 ± 0.032 Baseline 244.667 ± 6.940 208.315 ± 7.429 6.347 ± 0.111 1.553 ± 0.035 102.672 ± 0.550 11.983 ± 0.045 Recovery sleep 256.463 ± 5.182 225.519 ± 4.812 5.952 ± 0.091* 1.356 ± 0.024* 95.150 ± 4.791* 12.102 ± 0.031* Genotype Total spindles (number) NREM sleep spindles (number) NREM sleep spindle density (number of spindles /min) Spindle duration (s) Spindle amplitude (% normalized to 24 hr values) Spindle median peak (Hz) Sst-Cre+/−; Nos1flox/flox mice Dark KO 118.808 ± 10.800 80.700 ± 5.851 5.230 ± 0.286 1.598 ± 0.077 110.309 ± 9.407 12.071 ± 0.050 Light KO 255.275 ± 8.719* 207.075 ± 5.765* 5.929 ± 0.150* 1.562 ± 0.034 89.691 ± 9.407 11.946 ± 0.044 Baseline 256.467 ± 9.713 205.133 ± 6.818 6.140 ± 0.143 1.588 ± 0.038 89.968 ± 9.428 11.942 ± 0.046 Recovery sleep 258.183 ± 7.560 220.033 ± 5.755 5.681 ± 0.077* 1.411 ± 0.015* 107.208 ± 19.882* 12.061 ± 0.030* *p < 0.05 between light and dark periods or baseline and recovery sleep treatments. Open in new tab Table 2. Characteristics of rhythmic events occurring in the EEG during light and dark periods of spontaneous sleep and during the first 6 hr of recovery sleep after sleep deprivation in Sst-Cre−/−; Nos1flox/flox and Sst-Cre+/−; Nos1flox/flox mice: sleep spindles Genotype . Total spindles (number) . NREM sleep spindles (number) . NREM sleep spindle density (number of spindles /min) . Spindle duration (s) . Spindle amplitude (% normalized to 24 hr values) . Spindle median peak (Hz) . Sst-Cre−/−; Nos1flox/flox control mice . . . . . . . Dark WT 122.472 ± 9.401 89.815 ± 6.867 5.504 ± 0.161 1.450 ± 0.033 98.454 ± 0.431 12.083 ± 0.045 Light WT 247.167 ± 3.944* 213.083 ± 4.833* 6.099 ± 0.123* 1.484 ± 0.028 101.546 ± 0.431 12.019 ± 0.032 Baseline 244.667 ± 6.940 208.315 ± 7.429 6.347 ± 0.111 1.553 ± 0.035 102.672 ± 0.550 11.983 ± 0.045 Recovery sleep 256.463 ± 5.182 225.519 ± 4.812 5.952 ± 0.091* 1.356 ± 0.024* 95.150 ± 4.791* 12.102 ± 0.031* Genotype Total spindles (number) NREM sleep spindles (number) NREM sleep spindle density (number of spindles /min) Spindle duration (s) Spindle amplitude (% normalized to 24 hr values) Spindle median peak (Hz) Sst-Cre+/−; Nos1flox/flox mice Dark KO 118.808 ± 10.800 80.700 ± 5.851 5.230 ± 0.286 1.598 ± 0.077 110.309 ± 9.407 12.071 ± 0.050 Light KO 255.275 ± 8.719* 207.075 ± 5.765* 5.929 ± 0.150* 1.562 ± 0.034 89.691 ± 9.407 11.946 ± 0.044 Baseline 256.467 ± 9.713 205.133 ± 6.818 6.140 ± 0.143 1.588 ± 0.038 89.968 ± 9.428 11.942 ± 0.046 Recovery sleep 258.183 ± 7.560 220.033 ± 5.755 5.681 ± 0.077* 1.411 ± 0.015* 107.208 ± 19.882* 12.061 ± 0.030* Genotype . Total spindles (number) . NREM sleep spindles (number) . NREM sleep spindle density (number of spindles /min) . Spindle duration (s) . Spindle amplitude (% normalized to 24 hr values) . Spindle median peak (Hz) . Sst-Cre−/−; Nos1flox/flox control mice . . . . . . . Dark WT 122.472 ± 9.401 89.815 ± 6.867 5.504 ± 0.161 1.450 ± 0.033 98.454 ± 0.431 12.083 ± 0.045 Light WT 247.167 ± 3.944* 213.083 ± 4.833* 6.099 ± 0.123* 1.484 ± 0.028 101.546 ± 0.431 12.019 ± 0.032 Baseline 244.667 ± 6.940 208.315 ± 7.429 6.347 ± 0.111 1.553 ± 0.035 102.672 ± 0.550 11.983 ± 0.045 Recovery sleep 256.463 ± 5.182 225.519 ± 4.812 5.952 ± 0.091* 1.356 ± 0.024* 95.150 ± 4.791* 12.102 ± 0.031* Genotype Total spindles (number) NREM sleep spindles (number) NREM sleep spindle density (number of spindles /min) Spindle duration (s) Spindle amplitude (% normalized to 24 hr values) Spindle median peak (Hz) Sst-Cre+/−; Nos1flox/flox mice Dark KO 118.808 ± 10.800 80.700 ± 5.851 5.230 ± 0.286 1.598 ± 0.077 110.309 ± 9.407 12.071 ± 0.050 Light KO 255.275 ± 8.719* 207.075 ± 5.765* 5.929 ± 0.150* 1.562 ± 0.034 89.691 ± 9.407 11.946 ± 0.044 Baseline 256.467 ± 9.713 205.133 ± 6.818 6.140 ± 0.143 1.588 ± 0.038 89.968 ± 9.428 11.942 ± 0.046 Recovery sleep 258.183 ± 7.560 220.033 ± 5.755 5.681 ± 0.077* 1.411 ± 0.015* 107.208 ± 19.882* 12.061 ± 0.030* *p < 0.05 between light and dark periods or baseline and recovery sleep treatments. Open in new tab During the first 6 hr of recovery sleep postsleep deprivation, a main effect was found for enhanced spindle numbers during sleep/wake states in both genotypes (F1,34 = 6.481, p = 0.016). During recovery sleep, main effects were also found for reduced sleep spindle density (F1,34 = 15.188, p < 0.001) and sleep spindle duration (F1,34 = 15.628, p < 0.001) during NREM sleep. A main effect of reduced NREM sleep spindle density was observed in Sst-Cre+/−; Nos1flox/flox mice compared with Sst-Cre−/−; Nos1flox/flox control mice (F1,34 = 4.765, p = 0.036). Post hoc analysis found that NREM sleep spindle density (Sst-Cre−/−; Nos1flox/flox: F1,16 = 7.55, p = 0.014; Sst-Cre+/−; Nos1flox/flox: F1,18 = 7.970, p = 0.011) and sleep spindle duration (Sst-Cre−/−; Nos1flox/flox: F1,16 = 21.994, p < 0.001; Sst-Cre+/−; Nos1flox/flox: F1,18 = 18.290, p < 0.001) were similarly reduced during recovery NREM sleep vs. baseline NREM sleep in both genotypes, and the median of the peak amplitude (Hz) of sleep spindles was only ~0.1 Hz greater during recovery sleep compared with baseline values. No other significant differences in sleep spindle variables were observed during recovery sleep or between genotypes. Mice lacking nNOS in SST cells have impaired memory in the cortical-dependent floor-texture recognition task Male Sst-Cre+/−; Nos1flox/flox and Sst-Cre−/−; Nos1flox/flox littermate control mice were administered the floor-texture recognition (FTR) task (Figure 6 and Supplementary Figure S8)—a cognitive test used in mice that was developed to assess cortical-dependent memory consolidation [22]. Mice given two identical objects on smooth floor surfaces demonstrated similar amounts of time exploring either object. Twenty-four hours later, the mice were given the FTR task and a main effect was found for preference for the time exploring the object placed on top of the textured floor vs. smooth surface (i.e., treatment) and between genotypes. A treatment x genotype interaction was observed. Sst-Cre−/−; Nos1flox/flox control mice exhibited an enhanced preference for exploring the object on the floor respective to the opposing smooth surfaced side (63.91 ± 3.01%; t = 2.880, df = 13, p = 0.013), whereas Sst-Cre+/−; Nos1flox/flox mice did not show any preference for the textured floor compared with the nontextured floor (51.11 ± 3.21%). The day after the FTR task was given, we exposed animals to the NOR to test hippocampal-dependent memory recognition. However, no significant differences were found between genotypes in the NOR task. Both genotypes exhibited enhanced preference for the percentage of time spent exploring the novel object in the NOR task (Sst-Cre+/−; Nos1flox/flox mice, 73.63 ± 4.64; Sst-Cre−/−; Nos1flox/flox mice, 76.69 ± 5.33; df = 14, t = 0.432, p = 0.672). Additionally, there were no significant differences in the distance traveled or speed in these tasks which suggests that movement activity did not mediate the cognitive findings. Discussion Herein, we report evidence that SST+/nNOS+ neurons are involved in low-delta electroencephalogram activity and cortical-dependent recognition memory. Our previous studies implicated sleep active GABAergic nNOS neurons in the regulation of SWA [10, 13, 18, 29, 37]. We found that the expression of the neuronal activity marker Fos in these neurons was markedly enhanced during sleep after a period of sleep loss and varied in parallel with the amount of SWA in three mammalian species including mice, rats, and hamsters [10]. Increased cFos expression is associated with enhanced Ca2+ influx, which is expected to increase NO generation in nNOS-containing neurons via enhanced nNOS production in response to NMDA receptor activation [16, 17]. GABA is enhanced in the somatosensory cortex during NREM sleep vs. wakefulness [38], and GABAergic inputs have a localized regulation of neuronal Ca2+ signaling [39]. nNOS synthesis would be expected to increase in SST+/nNOS+ neurons because Ca2+ influx arising through L-type voltage-sensitive Ca2+ channels is capable of stimulating transcription from the nNOS promoter [40]. Thus, nNOS neurons in the cerebral cortex are likely to generate increased amounts of both NO and nNOS during sleep after extended periods of sleep loss. Since recent studies demonstrated that the loss of nNOS expression in nNOS KO mice was associated with reduced production of SWA during NREM sleep [18], it is plausible that increased generation of nNOS and/or NO in nNOS neurons is a vital mechanism for enhancing SWA production. However, the nNOS KO mouse model is not specific for cortical neurons because expression of nNOS is reduced in the whole animal. In addition, the NOS1 gene was targeted in nNOS KO mice for deletion in exon 2—an exon only expressed in nNOSα. Therefore, the nNOS deficiency in nNOS KO mice may be partly compensated by the presence of truncated nNOS isoforms (nNOSβ and nNOSγ). In the present study, we used nNOSflox/flox mice that possess loxP sites flanking exon 6 of NOS1 gene which reduce the expression of all nNOS isoforms. Furthermore, we used an intersectional strategy targeting SST to restrict nNOS deficiency to type I nNOS neurons that express SST in the cortex and caudate-putamen. Type I nNOS neurons are found not only in the cerebral cortex, but also in the caudate-putamen, olfactory bulb, corpus callosum, and amygdala [41–43]. Indeed, analysis of nNOS distribution in the brains of offspring of Sst-Cre+/− mice crossed with Ai14 reporter mice indicated that these cells are contained within these brain regions, although noncortical brain regions were found to contain nNOS expressing cells that do not express SST (Figure 1). Thus, both our current findings and other published results suggest that SST+/nNOS+ cells are a unique cell type of common origin that are located in the cerebral cortex, caudate-putamen, olfactory bulb, corpus callosum, and amygdala. We restricted nNOS deficiency to SST-expressing cells by crossing nNOSflox/flox mice with Ssttm2.1(cre)Zjh mice. Our histological analysis showed that Cre is expressed in type I nNOS neurons selectively, and that ablation of nNOS expression was almost exclusively limited to type I nNOS neurons of the Sst-Cre+/−; Nos1flox/flox mice (Figure 1). Thus, we produced a highly selective model of nNOS knockout in type I Sst+ neurons. The homeostatic drive associated with sleep loss is usually indexed by one or more of the following independent measures: (1) increased rebound of SWA, (2) increased rebound of sleep amount, (3) increased consolidation of sleep episodes, and/or (4) reduced latency to sleep onset [1, 44, 45]. In Sst-Cre+/−; Nos1flox/flox mice, we observed a significant diminution in enhanced SWA response to sleep deprivation within the 0.5–4.0 Hz frequency range (Figure 5B). Further analysis of the frequency bins revealed that the SWA rebound was selectively reduced during recovery sleep only in the frequency range below 1.5 Hz, whereas there was no significant difference in the magnitude of SWA rebound in the medium and high δ frequency range (1.5–3.0 Hz and 3.0–4.5 Hz; Figure 5, C–E). We did not observe any differences in other measures of homeostatic drive between Sst-Cre+/−; Nos1flox/flox mice and their Sst-Cre−/−; Nos1flox/flox littermate control mice including a rebound of REM sleep and NREM sleep (Figure 4, A and B), bout duration of REM sleep and NREM sleep (Supplementary Figure S4), latency to sleep onset in the MSLT test (Figure 4E), and EEG power in the θ frequency range either during sleep deprivation procedures or baseline dark periods (Supplementary Figure S5). There are additional measures of homeostatic drive that are less frequently used, such as an increase in EEG power in the θ frequency band during waking [32, 46], although we did not find differences in EEG θ power between Sst-Cre+/−; Nos1flox/flox mice and controls in the present study (Supplementary Material). Collectively, these results suggest that the homeostatic response of enhanced low-δ-range SWA after sleep loss is regulated independently of other measures of homeostatic drive and is, in part, dependent on nNOS expression in SST+ neurons located in the cortex and caudate-putamen. Nevertheless, we cannot conclude the contribution of cortical and subcortical Sst+/nNOS+ neurons with our findings and future brain area–targeting approaches are needed. The homeostatic drive involving sleep and SWA is regulated via different molecular, immune, and neuronal and nonneuronal mechanisms [31, 47–49]. Nevertheless, instances exist where sleep and SWA are regulated by independent mechanisms [2]. Similar to Sst-Cre+/−; Nos1flox/flox mice, SWA is diminished during spontaneous NREM sleep in the low-δ power range (up to 2.5 Hz) in nNOS KO mice [18]. Several changes in spontaneous NREM sleep patterns are described in nNOS KO mice [18], such as increased SWA during wakefulness, reduced NREM sleep amounts, and shortened bouts of NREM sleep, although these changes were not observed in Sst-Cre+/−; Nos1flox/flox mice studied herein. Interestingly, the homeostatic increase in SWA following sleep deprivation was attenuated in the low δ range of both nNOS KO mice and Sst-Cre+/−; Nos1flox/flox mice [18]. However, shortened latency to sleep onset in MSLT was observed in nNOS KO mice [18], but not in Sst-Cre+/−; Nos1flox/floxmice. Thus, we observed both similarities and differences in sleep pattern between Sst-Cre+/−; Nos1flox/flox mice and nNOS KO mice (Table 3). It is likely that the similar changes, such as reduction of SWA in the low δ range, depend on the deficiency of nNOS expression in type I nNOS neurons. Other changes, such as elevated levels of SWA during wakefulness, fragmented NREM sleep, reduced total amounts of NREM sleep, and shortened latency to sleep onset in MSLT [18], are likely modulated by the deficiency of nNOS expression in other cells types. Overall, the results of our current study implicate type I nNOS neurons within the cortex and caudate-putamen in the homeostatic regulation of SWA specifically in the low δ power range but not in regulating enhanced sleep amounts as there were no significant differences in sleep duration or in the MSLT between genotypes. Table 3. Summary of comparison between whole-body nNOS KO and Sst-Cre+/−; Nos1flox/flox mice compared with controls Sleep parameter . Whole-body nNOS KO 18 . SST+-nNOS KO . SWA during NREM sleep Reduced in low delta (0.5–2.5 Hz) Reduced in low delta (0–3.0 Hz) SWA response to sleep deprivation Reduced (0.5–4 Hz) Reduced (0.5–4 Hz) SWA during wakefulness Increased (0.5–4 Hz) No change Amounts of NREM sleep Reduced No change Duration of NREM sleep episode bouts 1/3 shorter No change Latency to sleep onset in MSLT Reduced No change NREM and REM sleep amounts in MSLT Increased No change Sleep parameter . Whole-body nNOS KO 18 . SST+-nNOS KO . SWA during NREM sleep Reduced in low delta (0.5–2.5 Hz) Reduced in low delta (0–3.0 Hz) SWA response to sleep deprivation Reduced (0.5–4 Hz) Reduced (0.5–4 Hz) SWA during wakefulness Increased (0.5–4 Hz) No change Amounts of NREM sleep Reduced No change Duration of NREM sleep episode bouts 1/3 shorter No change Latency to sleep onset in MSLT Reduced No change NREM and REM sleep amounts in MSLT Increased No change Open in new tab Table 3. Summary of comparison between whole-body nNOS KO and Sst-Cre+/−; Nos1flox/flox mice compared with controls Sleep parameter . Whole-body nNOS KO 18 . SST+-nNOS KO . SWA during NREM sleep Reduced in low delta (0.5–2.5 Hz) Reduced in low delta (0–3.0 Hz) SWA response to sleep deprivation Reduced (0.5–4 Hz) Reduced (0.5–4 Hz) SWA during wakefulness Increased (0.5–4 Hz) No change Amounts of NREM sleep Reduced No change Duration of NREM sleep episode bouts 1/3 shorter No change Latency to sleep onset in MSLT Reduced No change NREM and REM sleep amounts in MSLT Increased No change Sleep parameter . Whole-body nNOS KO 18 . SST+-nNOS KO . SWA during NREM sleep Reduced in low delta (0.5–2.5 Hz) Reduced in low delta (0–3.0 Hz) SWA response to sleep deprivation Reduced (0.5–4 Hz) Reduced (0.5–4 Hz) SWA during wakefulness Increased (0.5–4 Hz) No change Amounts of NREM sleep Reduced No change Duration of NREM sleep episode bouts 1/3 shorter No change Latency to sleep onset in MSLT Reduced No change NREM and REM sleep amounts in MSLT Increased No change Open in new tab SWA has been linked to the induction of cortical plastic changes because it increases locally after a learning task and is positively correlated with postsleep performance improvement [50]. It has been suggested that slow oscillations, by inducing depolarized up- and hyperpolarized down-states of neuronal activity, can provide a supra-ordinate temporal frame for the dialogue between the neocortex and the subcortical structures [51]. A shift in membrane potential of up- and down-states between 0.5 and 1.5 Hz occurring during NREM sleep may facilitate the redistribution of memories for long-term storage [52–54], although a selective weakening of synaptic inputs may also occur [55, 56]. Based on these observations, we expected that memory would be impaired in Sst-Cre−/−; Nos1flox/flox mice since SWA was diminished. We chose to test memory in FTR test because it is a cognitive task that specifically relies on the cerebral cortex [22]. Our results indicated that memory was impaired in the FTR test in Sst-Cre+/−; Nos1flox/flox mice (Figure 6). A recent study demonstrated that M2 secondary cortex to S1 primary somatosensory cortex top-down input, and not vice versa, regulated memory consolidation specifically during NREM sleep in the FTR task [22]. The authors of this study suggested that the directional effect of the M2 to S1 activity is explained by the fact that SWA during NREM sleep propagates in an anteroposterior direction [57–59]. That SST+/nNOS+ cells are active throughout the cortex, it is likely, although speculative, that these cells release nNOS in an anteroposterior direction to alter SWA. We did not find any significant difference in the NOR task in nNOSflox/flox mice—a memory task that is partly dependent on the hippocampus [22]. Our findings therefore suggest that sleep active cortical and caudate-putaman SST+/nNOS+ cells are not involved in hippocampal-dependent memory recognition. Nevertheless, these cognitive testing results do not permit the conclusion that the impairment of memory in these mice is directly related to the reduced SWA production. However, that local administration of the nNOS inhibitor Nω-propyl-L-arginine into the perirhinal cortex impairs long-term visual recognition memory [60] and spatial recognition memory is impaired in nNOS KO mice [61], it is possible that memory impairments observed in the present study were related to nNOS deficiency rather than to SWA reduction. Cognition could also be altered indirectly in Sst-Cre+/−; Nos1flox/flox mice via changes in other physiological parameters, such as attention or mood, especially taken into account that a subset of SST+/nNOS+ neurons is located in the basolateral amygdala [43]. Additional studies will be needed to thoroughly assess mechanisms of SST+/nNOS+ neurons in cognition. Our findings that Sst-Cre+/−; Nos1flox/flox had reduced negative durations in slow-wave wavelets compared with control mice and the slopes of the slow waves were not altered in their homeostatic SWA responses to sleep deprivation like the control mice suggests that type I nNOS neurons are specifically involved with the hyperpolarization of the NREM sleep slow waves. These data are consistent with a study in mice lacking SST that do not have changes in the slope of the slow waves after sleep deprivation [14], although the current findings suggest that SST+ nNOS+ cells are a major contributor to the enhanced slope of slow waves occurring after sleep deprivation. It has been hypothesized that reductions in the hyperpolarization phase of slow waves inhibits the ability of neurons to reestablish synaptic homeostasis from enhanced neuronal activity [62]. Consequently, the inability of Sst-Cre+/−; Nos1flox/flox mice to reestablish synaptic homeostasis via reduced hyperpolarization of the NREM sleep slow waves could be, in part, related to impairments in cognition observed in the FRT. Sleep spindles are waxing and waning EEG burst phenomenon occurring in the 10–15 EEG frequency range [63]. Although they occur during waking and sleep states, a growing literature indicates that NREM sleep spindles aid in facilitating memory formation and learning, in part, through altering synaptic plasticity [64]. Since evidence indicates that sleep spindles can be regulated by large cortical networks and that they are associated with SWA [65], it was plausible that sleep spindles dynamics could be altered in Sst-Cre+/−; Nos1flox/flox mice that exhibited altered spontaneous and homeostatic responses to sleep deprivation. Nevertheless, Sst-Cre+/−; Nos1flox/flox mice showed similar total numbers of sleep spindle numbers during wake/sleep and sleep spindle numbers, density, duration, amplitude, and median peak frequency (Hz) during NREM sleep as control mice suggesting that the cognitive impairments were not due to altered sleep spindles (Table 2 and Supplementary Figure S6). In conclusion, we demonstrate for the first time that SST+/nNOS+ neurons function to enhance the homeostatic response of SWA in the low δ power range and aid in proper cortical-dependent memory functioning. Consequently, SST+/nNOS+ neuron dysfunction could contribute to pathophysiological changes observed in various diseases associated with both disturbed SWA production and cognitive impairments including Alzheimer’s disease [66], schizophrenia [67], epilepsy [68], and traumatic brain injury [69, 70]. Moreover, future studies will be needed to examine how SST+/nNOS+ neurons affect individuals with sleep apnea and insomnia, especially veterans who exhibit greater incidence of these disorders and often exhibit impairments in SWA and cognition [71, 72]. Funding This work was supported by the Department of Veterans Affairs Medical Research Service Award, Department of Veterans Affairs Career Award IBX002823A (MRZ), and National Institutes of Health (NINDS-NS064193 [to DG], NINDS-NS106406 [to DG], and NIA-AG061774). MRZ and DG conceived and planned the experiments. MRZ, JMM, JTM, and DG performed the experiments and calculations; MRZ, DNA, JMM, JTM, PLH, RES, and DG contributed to the interpretation of the results. All authors provided critical feedback and helped shaped the research, analysis, and manuscript. Conflict of interest statement. JTM received partial salary compensation and funding from Merck Investigator Sponsored Programs, but has no conflict of interest with this work. References 1. Achermann P , et al. Dynamics of EEG slow wave activity during physiological sleep and after administration of benzodiazepine hypnotics . Hum Neurobiol. 1987 ; 6 ( 3 ): 203 – 210 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 2. Davis CJ , et al. Delta wave power: an independent sleep phenotype or epiphenomenon? J Clin Sleep Med. 2011 ; 7 ( 5 Suppl ): S16 – S18 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 3. Krueger JM , et al. Sleep’s Kernel: Surprisingly small sections of brain, and even neuronal and glial networks in a dish, display many electrical indicators of sleep . Scientist. 2016 ; 30 ( 3 ): 36 – 41 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 4. Pugin F , et al. Local increase of sleep slow wave activity after three weeks of working memory training in children and adolescents . Sleep. 2015 ; 38 ( 4 ): 607 – 614 . doi: 10.5665/sleep.4580 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Perrenoud Q , et al. Characterization of type I and type II nNOS-expressing interneurons in the barrel cortex of mouse . Front Neural Circuits. 2012 ; 6 : 36 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 6. Marin O , et al. Origin and molecular specification of striatal interneurons . J Neurosci. 2000 ; 20 ( 16 ): 6063 – 6076 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Tricoire L , et al. Neuronal nitric oxide synthase expressing neurons: a journey from birth to neuronal circuits . Front Neural Circuits. 2012 ; 6 : 82 . Google Scholar Crossref Search ADS PubMed WorldCat 8. Tamamaki N , et al. Long-range GABAergic connections distributed throughout the neocortex and their possible function . Front Neurosci. 2010 ; 4 : 202 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Weiss SW , et al. NMDAR1 glutamate receptor subunit isoforms in neostriatal, neocortical, and hippocampal nitric oxide synthase neurons . J Neurosci. 1998 ; 18 ( 5 ): 1725 – 1734 . Google Scholar Crossref Search ADS PubMed WorldCat 10. Gerashchenko D , et al. Identification of a population of sleep-active cerebral cortex neurons . Proc Natl Acad Sci U S A. 2008 ; 105 ( 29 ): 10227 – 10232 . Google Scholar Crossref Search ADS PubMed WorldCat 11. Gerashchenko D , et al. Sleep state dependence of optogenetically evoked responses in neuronal nitric oxide synthase-positive cells of the cerebral cortex . Neuroscience. 2018 ; 379 : 189 – 201 . Google Scholar Crossref Search ADS PubMed WorldCat 12. Dittrich L , et al. Cortical nNOS neurons co-express the NK1 receptor and are depolarized by Substance P in multiple mammalian species . Front Neural Circuits. 2012 ; 6 : 31 . Google Scholar Crossref Search ADS PubMed WorldCat 13. Zielinski MR , et al. Substance P and the neurokinin-1 receptor regulate electroencephalogram non-rapid eye movement sleep slow-wave activity locally . Neuroscience. 2015 ; 284 : 260 – 272 . Google Scholar Crossref Search ADS PubMed WorldCat 14. Funk CM , et al. Role of somatostatin-positive cortical interneurons in the generation of sleep slow waves . J Neurosci. 2017 ; 37 ( 38 ): 9132 – 9148 . Google Scholar Crossref Search ADS PubMed WorldCat 15. Niethard N , et al. Cortical circuit activity underlying sleep slow oscillations and spindles . Proc Natl Acad Sci U S A. 2018 ; 115 ( 39 ): E9220 – E9229 . Google Scholar Crossref Search ADS PubMed WorldCat 16. Garthwaite J . Glutamate, nitric oxide and cell-cell signalling in the nervous system . Trends Neurosci. 1991 ; 14 ( 2 ): 60 – 67 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Garthwaite J , et al. Endothelium-derived relaxing factor release on activation of NMDA receptors suggests role as intercellular messenger in the brain . Nature. 1988 ; 336 ( 6197 ): 385 – 388 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Morairty SR , et al. A role for cortical nNOS/NK1 neurons in coupling homeostatic sleep drive to EEG slow wave activity . Proc Natl Acad Sci U S A. 2013 ; 110 ( 50 ): 20272 – 20277 . Google Scholar Crossref Search ADS PubMed WorldCat 19. Hyndman KA , et al. Renal collecting duct NOS1 maintains fluid-electrolyte homeostasis and blood pressure . Hypertension. 2013 ; 62 ( 1 ): 91 – 98 . Google Scholar Crossref Search ADS PubMed WorldCat 20. Zielinski MR , et al. A novel telemetric system to measure polysomnographic biopotentials in freely moving animals . J Neurosci Methods. 2013 ; 216 ( 2 ): 79 – 86 . Google Scholar Crossref Search ADS PubMed WorldCat 21. Franken P , et al. Genetic variation in EEG activity during sleep in inbred mice . Am J Physiol. 1998 ; 275 ( 4 ): R1127 – R1137 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 22. Miyamoto D , et al. Top-down cortical input during NREM sleep consolidates perceptual memory . Science. 2016 ; 352 ( 6291 ): 1315 – 1318 . Google Scholar Crossref Search ADS PubMed WorldCat 23. Latchoumane CV , et al. Thalamic spindles promote memory formation during sleep through triple phase-locking of cortical, thalamic, and hippocampal rhythms . Neuron. 2017 ; 95 ( 2 ): 424 – 435.e6 . Google Scholar Crossref Search ADS PubMed WorldCat 24. Massart R , et al. The genome-wide landscape of DNA methylation and hydroxymethylation in response to sleep deprivation impacts on synaptic plasticity genes . Transl Psychiatry. 2014 ; 4 : e347 . Google Scholar Crossref Search ADS PubMed WorldCat 25. Vassalli A , et al. Electroencephalogram paroxysmal θ characterizes cataplexy in mice and children . Brain. 2013 ; 136 ( Pt 5 ): 1592 – 1608 . Google Scholar Crossref Search ADS PubMed WorldCat 26. Gotti S , et al. Distribution of nitric oxide synthase immunoreactivity in the mouse brain . Microsc Res Tech. 2005 ; 68 ( 1 ): 13 – 35 . Google Scholar Crossref Search ADS PubMed WorldCat 27. Ling S , et al. The expression of neuronal nitric oxide synthase in the brain of the mouse during embryogenesis . Anat Rec (Hoboken). 2012 ; 295 ( 3 ): 504 – 514 . Google Scholar Crossref Search ADS PubMed WorldCat 28. Rodrigo J , et al. Localization of nitric oxide synthase in the adult rat brain . Philos Trans R Soc Lond B Biol Sci. 1994 ; 345 ( 1312 ): 175 – 221 . Google Scholar Crossref Search ADS PubMed WorldCat 29. Pasumarthi RK , et al. Further characterization of sleep-active neuronal nitric oxide synthase neurons in the mouse brain . Neuroscience. 2010 ; 169 ( 1 ): 149 – 157 . Google Scholar Crossref Search ADS PubMed WorldCat 30. Jinno S , et al. Patterns of colocalization of neuronal nitric oxide synthase and somatostatin-like immunoreactivity in the mouse hippocampus: quantitative analysis with optical disector . Neuroscience. 2004 ; 124 ( 4 ): 797 – 808 . Google Scholar Crossref Search ADS PubMed WorldCat 31. Kalinchuk AV , et al. Cholinergic neurons of the basal forebrain mediate biochemical and electrophysiological mechanisms underlying sleep homeostasis . Eur J Neurosci. 2015 ; 41 ( 2 ): 182 – 195 . Google Scholar Crossref Search ADS PubMed WorldCat 32. Vyazovskiy VV , et al. Theta activity in the waking EEG is a marker of sleep propensity in the rat . Brain Res. 2005 ; 1050 ( 1-2 ): 64 – 71 . Google Scholar Crossref Search ADS PubMed WorldCat 33. Mang GM , et al. Altered sleep homeostasis in rev-erbα knockout mice . Sleep. 2016 ; 39 ( 3 ): 589 – 601 . doi: 10.5665/sleep.5534 Google Scholar Crossref Search ADS PubMed WorldCat 34. Veasey SC , et al. Murine multiple sleep latency test: phenotyping sleep propensity in mice . Sleep. 2004 ; 27 ( 3 ): 388 – 393 . doi: 10.1093/sleep/27.3.388 Google Scholar Crossref Search ADS PubMed WorldCat 35. Binder S , et al. Transcranial slow oscillation stimulation during NREM sleep enhances acquisition of the radial maze task and modulates cortical network activity in rats . Front Behav Neurosci. 2013 ; 7 : 220 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 36. González-Rueda A , et al. Activity-dependent downscaling of subthreshold synaptic inputs during slow-wave-sleep-like activity in vivo . Neuron. 2018 ; 97 ( 6 ): 1244 – 1252.e5 . Google Scholar Crossref Search ADS PubMed WorldCat 37. Kilduff TS , et al. Activation of cortical interneurons during sleep: an anatomical link to homeostatic sleep regulation? Trends Neurosci. 2011 ; 34 ( 1 ): 10 – 19 . Google Scholar Crossref Search ADS PubMed WorldCat 38. Vanini G , et al. GABA-to-ACh ratio in basal forebrain and cerebral cortex varies significantly during sleep . Sleep. 2012 ; 35 ( 10 ): 1325 – 1334 . doi: 10.5665/sleep.2106 Google Scholar Crossref Search ADS PubMed WorldCat 39. Higley MJ . Localized GABAergic inhibition of dendritic ca(2+) signalling . Nat Rev Neurosci. 2014 ; 15 ( 9 ): 567 – 572 . Google Scholar Crossref Search ADS PubMed WorldCat 40. Sasaki M , et al. Dynamic regulation of neuronal NO synthase transcription by calcium influx through a CREB family transcription factor-dependent mechanism . Proc Natl Acad Sci U S A. 2000 ; 97 ( 15 ): 8617 – 8622 . Google Scholar Crossref Search ADS PubMed WorldCat 41. Barbaresi P , et al. Intracallosal neuronal nitric oxide synthase neurons colocalize with neurokinin 1 substance P receptor in the rat . J Comp Neurol. 2015 ; 523 ( 4 ): 589 – 607 . Google Scholar Crossref Search ADS PubMed WorldCat 42. Wang X , et al. Density and neurochemical profiles of neuronal nitric oxide synthase-expressing interneuron in the mouse basolateral amygdala . Brain Res. 2017 ; 1663 : 106 – 113 . Google Scholar Crossref Search ADS PubMed WorldCat 43. Bocchio M , Fisher SP, Unal G, Ellender TJ, Vyazovskiy VV, Capogna M. Sleep and serotonin modulate paracapsular nitric oxide synthase expressing neurons of the amygdala . eNeuro . 2016 ; 3 ( 5 ). pii: ENEURO.0177-16.2016. Google Scholar OpenURL Placeholder Text WorldCat 44. Davis CJ , et al. The neuron-specific interleukin-1 receptor accessory protein is required for homeostatic sleep and sleep responses to influenza viral challenge in mice . Brain Behav Immun. 2015 ; 47 : 35 – 43 . Google Scholar Crossref Search ADS PubMed WorldCat 45. Bjorness TE , et al. An adenosine-mediated glial-neuronal circuit for homeostatic sleep . J Neurosci. 2016 ; 36 ( 13 ): 3709 – 3721 . Google Scholar Crossref Search ADS PubMed WorldCat 46. Ehlen JC , et al. Period-amplitude analysis reveals wake-dependent changes in the electroencephalogram during sleep deprivation . Sleep. 2013 ; 36 ( 11 ): 1723 – 1735 . doi: 10.5665/sleep.3138 Google Scholar Crossref Search ADS PubMed WorldCat 47. Zielinski MR , et al. The NLRP3 inflammasome modulates sleep and NREM sleep delta power induced by spontaneous wakefulness, sleep deprivation and lipopolysaccharide . Brain Behav Immun. 2017 ; 62 : 137 – 150 . Google Scholar Crossref Search ADS PubMed WorldCat 48. Porkka-Heiskanen T , et al. Adenosine: a mediator of the sleep-inducing effects of prolonged wakefulness . Science. 1997 ; 276 ( 5316 ): 1265 – 1268 . Google Scholar Crossref Search ADS PubMed WorldCat 49. Oishi Y , et al. Slow-wave sleep is controlled by a subset of nucleus accumbens core neurons in mice . Nat Commun. 2017 ; 8 ( 1 ): 734 . Google Scholar Crossref Search ADS PubMed WorldCat 50. Huber R , et al. Local sleep and learning . Nature. 2004 ; 430 ( 6995 ): 78 – 81 . Google Scholar Crossref Search ADS PubMed WorldCat 51. Marshall L , et al. The contribution of sleep to hippocampus-dependent memory consolidation . Trends Cogn Sci. 2007 ; 11 ( 10 ): 442 – 450 . Google Scholar Crossref Search ADS PubMed WorldCat 52. Sirota A , et al. 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 53. Sirota A , et al. Interaction between neocortical and hippocampal networks via slow oscillations . Thalamus Relat Syst. 2005 ; 3 ( 4 ): 245 – 259 . Google Scholar Crossref Search ADS PubMed WorldCat 54. Harmony T . The functional significance of delta oscillations in cognitive processing . Front Integr Neurosci. 2013 ; 7 : 83 . Google Scholar Crossref Search ADS PubMed WorldCat 55. Bartram J , et al. Cortical Up states induce the selective weakening of subthreshold synaptic inputs . Nat Commun. 2017 ; 8 ( 1 ): 665 . Google Scholar Crossref Search ADS PubMed WorldCat 56. Puentes-Mestril C , et al. Linking network activity to synaptic plasticity during sleep: hypotheses and recent data . Front Neural Circuits. 2017 ; 11 : 61 . Google Scholar Crossref Search ADS PubMed WorldCat 57. Phillips KG , et al. Decoupling of sleep-dependent cortical and hippocampal interactions in a neurodevelopmental model of schizophrenia . Neuron. 2012 ; 76 ( 3 ): 526 – 533 . Google Scholar Crossref Search ADS PubMed WorldCat 58. Nir Y , et al. Regional slow waves and spindles in human sleep . Neuron. 2011 ; 70 ( 1 ): 153 – 169 . Google Scholar Crossref Search ADS PubMed WorldCat 59. Massimini M , et al. The sleep slow oscillation as a traveling wave . J Neurosci. 2004 ; 24 ( 31 ): 6862 – 6870 . Google Scholar Crossref Search ADS PubMed WorldCat 60. Tamagnini F , et al. Nitric oxide-dependent long-term depression but not endocannabinoid-mediated long-term potentiation is crucial for visual recognition memory . J Physiol. 2013 ; 591 ( 16 ): 3963 – 3979 . Google Scholar Crossref Search ADS PubMed WorldCat 61. Zoubovsky SP , et al. Working memory deficits in neuronal nitric oxide synthase knockout mice: potential impairments in prefrontal cortex mediated cognitive function . Biochem Biophys Res Commun. 2011 ; 408 ( 4 ): 707 – 712 . Google Scholar Crossref Search ADS PubMed WorldCat 62. Tononi G , et al. Time to be SHY? Some comments on sleep and synaptic homeostasis . Neural Plast. 2012 ; 2012 : 415250 . Google Scholar Crossref Search ADS PubMed WorldCat 63. Uygun DS , et al. Validation of an automated sleep spindle detection method for mouse EEG . Sleep . 2018 ; in press. Google Scholar OpenURL Placeholder Text WorldCat 64. Ulrich D . Sleep spindles as facilitators of memory formation and learning . Neural Plast. 2016 ; 2016 : 1796715 . Google Scholar Crossref Search ADS PubMed WorldCat 65. McVea DA , et al. Large scale cortical functional networks associated with slow-wave and spindle-burst-related spontaneous activity . Front Neural Circuits. 2016 ; 10 : 103 . Google Scholar Crossref Search ADS PubMed WorldCat 66. Mander BA , et al. β-amyloid disrupts human NREM slow waves and related hippocampus-dependent memory consolidation . Nat Neurosci. 2015 ; 18 ( 7 ): 1051 – 1057 . Google Scholar Crossref Search ADS PubMed WorldCat 67. Hiatt JF , et al. Further evidence of abnormal non-rapid-eye-movement sleep in schizophrenia . Arch Gen Psychiatry. 1985 ; 42 ( 8 ): 797 – 802 . Google Scholar Crossref Search ADS PubMed WorldCat 68. Höller Y , et al. Is There a relation between EEG-slow waves and memory dysfunction in epilepsy? A critical appraisal . Front Hum Neurosci. 2015 ; 9 : 341 . Google Scholar Crossref Search ADS PubMed WorldCat 69. Modarres M , Kuzma NN, Kretzmer T, Pack AI, Lim MM. EEG slow waves in traumatic brain injury: Convergent findings in mouse and man . Neurobiol Sleep Circadian Rhythms . 2016 ; 1 . Google Scholar OpenURL Placeholder Text WorldCat 70. Paterno R , et al. Pathophysiology and treatment of memory dysfunction after traumatic brain injury . Curr Neurol Neurosci Rep. 2017 ; 17 ( 7 ): 52 . Google Scholar Crossref Search ADS PubMed WorldCat 71. Capaldi VF 2nd, et al. Sleep disruptions among returning combat veterans from Iraq and Afghanistan . Mil Med. 2011 ; 176 ( 8 ): 879 – 888 . Google Scholar Crossref Search ADS PubMed WorldCat 72. Germain A . Sleep disturbances as the hallmark of PTSD: where are we now? Am J Psychiatry. 2013 ; 170 ( 4 ): 372 – 382 . Google Scholar Crossref Search ADS PubMed WorldCat 73. Knoblauch V , et al. Human sleep spindle characteristics after sleep deprivation . Clin Neurophysiol. 2003 ; 114 ( 12 ): 2258 – 2267 . Google Scholar Crossref Search ADS PubMed WorldCat 74. Rosinvil T , et al. Age-related changes in sleep spindles characteristics during daytime recovery following a 25-hour sleep deprivation . Front Hum Neurosci. 2015 ; 9 : 323 . Google Scholar Crossref Search ADS PubMed WorldCat Published by Oxford University Press on behalf of Sleep Research Society (SRS) 2019. 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) 2019.

Cabeza, de Baca, Tomás;Chayama, Koharu, Loulou;Redline,, Susan;Slopen,, Natalie;Matsushita,, Fumika;Prather, Aric, A;Williams, David, R;Buring, Julie, E;Zaslavsky, Alan, M;Albert, Michelle, A

doi: 10.1093/sleep/zsz149pmid:

Burgdorf, Jeffrey S; Vitaterna, Martha H; Olker, Christopher J; Song, Eun Joo; Christian, Edward P; Sørensen, Laurits; Turek, Fred W; Madsen, Torsten M; Khan, M Amin; Kroes, Roger A; Moskal, Joseph R

doi: 10.1093/sleep/zsz135pmid: 31504971

Abstract Study Objectives The present studies examine the effects of NMDAR activation by NYX-2925 diurnal rhythmicity of both sleep and wake as well as emotion. Methods Twenty-four-hour sleep EEG recordings were obtained in sleep-deprived and non-sleep-deprived rats. In addition, the day–night cycle of both activity and mood was measured using home cage ultrasonic-vocalization recordings. Results NYX-2925 significantly facilitated non-REM (NREM) sleep during the lights-on (sleep) period, and this effect persisted for 3 days following a single dose in sleep-deprived rats. Sleep-bout duration and REM latencies were increased without affecting total REM sleep, suggesting better sleep quality. In addition, delta power during wake was decreased, suggesting less drowsiness. NYX-2925 also rescued learning and memory deficits induced by sleep deprivation, measured using an NMDAR-dependent learning task. Additionally, NYX-2925 increased positive affect and decreased negative affect, primarily by facilitating the transitions from sleep to rough-and-tumble play and back to sleep. In contrast to NYX-2925, the NMDAR antagonist ketamine acutely (1–4 hours post-dosing) suppressed REM and non-REM sleep, increased delta power during wake, and blunted the amplitude of the sleep-wake activity rhythm. Discussion These data suggest that NYX-2925 could enhance behavioral plasticity via improved sleep quality as well as vigilance during wake. As such, the facilitation of sleep by NYX-2925 has the potential to both reduce symptom burden on neurological and psychiatric disorders as well as serve as a biomarker for drug effects through restoration of sleep architecture. sleep, EEG, ultrasonic vocalizations, NMDA receptors, sleep deprivation, ketamine Statement of Significance This manuscript details the development of a novel method for measuring the diurnal rhythm of emotion in rats using a novel NMDAR modulator, NYX-2925, that facilitates synaptic plasticity and is in Phase II trials for neuropathic pain disorders in which sleep disturbances are a core symptom. NYX-2925 produces a robust and long-lasting facilitation of the diurnal rhythm of affect and enhances measures of NREM sleep during the light phase. A parallel clinical trial with NYX-2925 in sleep is currently being conducted. Introduction NMDA receptor (NMDAR) activity is critical for diurnal rhythmicity, sleep, and memory consolidation. The NMDAR antagonist MK-801 completely eliminates both REM and NREM sleep for approximately the first 2 hours after systemic administration [1]. Interestingly, rebound increases in both REM and NREM sleep are seen approximately 2–12 hours after MK-801 administration [1]. MK-801 also disrupts consolidation of learning and memory across a wide range of tasks [2], whereas the NMDAR glycine site partial agonist d-cycloserine facilitates learning and memory consolidation [3, 4]. Recently, it has been shown that sleep drive is controlled in part via phosphorylation of the NR2B subtype of the NDMAR receptor [5]. Similarly, sleep deprivation decreases hippocampal cell surface expression of the obligatory NMDAR subunit GRIN1, and reduces NMDAR current and disrupts NMDAR-dependent LTP and LTD [6]. In comparison, pharmacological manipulations of other neurotransmitter systems (e.g., GABA, 5-HT, histamine, melatonin) have more moderate effects on both sleep-EEG and memory consolidation [7, 8]. Thus, facilitation of NMDAR activity likely enhances memory consolidation via facilitation of REM and NREM sleep, whereas inhibiting NMDAR activity has opposite effects. Sleep is causally linked to the retention of recently learned memories. Sleep deprivation disrupts a wide variety of learning and memory tasks including contextual fear extinction [9], spatial learning in both the Morris water maze and radial arm maze [10], motor learning consolidation [11, 12], and declarative memory [13]. At the cellular and molecular levels, sleep induces dendritic spine pruning and reduces the expression of cell surface AMPA receptors [14, 15]. In this way, sleep promotes mature dendritic spine formation and consolidation via an NMDAR-dependent process. NYX-2925 is a novel, orally available, small-molecule NMDAR modulator and is distinct from known NMDAR agonists or antagonists such as D-cycloserine, ketamine, MK-801, kynurenic acid, or ifenprodil [16]. NYX-2925 robustly facilitates NMDAR current and NMDAR-dependent long-term potentiation (LTP) in brain slices, and NMDAR-dependent learning and memory in vivo by activating the NMDAR [16]. NYX-2925 is in development as a therapy for chronic pain and is currently under evaluation in two phase 2 clinical studies, one in subjects with painful diabetic peripheral neuropathy and the other in subjects with fibromyalgia (ClinicalTrials.gov identifiers: NCT03219320; NCT03249103). Both indications being studied are chronic pain conditions in which sleep disruption is often a core symptom reported by those affected [17, 18]. This study was designed to evaluate whether NMDAR activation, with a single dose of NYX-2925, enhances the amplitude of the day–night cycle of sleep and mood, and enhances learning and memory consolidation in a sleep-deprivation paradigm in rats. Materials and Methods Animals Male 2- to 3-month-old Sprague-Dawley (SD) rats from Charles River (United States) were used. Rats were housed in Lucite cages with aspen-wood chip bedding, maintained on a 12:12 light:dark cycle (lights-on at 6 am), and given ad libitum access to Teklad lab chow (Envigo, United States) and tap water throughout the study. All experiments were approved by the Northwestern University Animal Care and Use Committee. Sleep EEG studies Rats were anesthetized using isoflurane and implanted with cortical EEG skull screws and EMG neck muscle wires (Pinnacle, USA). EEG/EMG signals were captured via a tethered system (Pinnacle, USA), and sleep/wake states were scored in 10-second epochs with a combination of manual scoring and machine learning [19] in a treatment-blinded manner. One week after surgery, rats were placed in sleep-recording cages (35.6 cm diameter × 30.5 cm high; Pinnacle, USA) within sound-attenuating chambers that also blocked ambient outside light. After 24 hours of habituation, rats were dosed with drug or vehicle as described below at zeitgeber (ZT) 5 (11 am) and EEG/EMG was recorded for 24–72 hours post-dosing. Additional groups of rats received sleep deprivation using the Pinnacle (USA) sleep deprivation system from ZT0-6. Rats were dosed with NYX-2925 (0.1, 1, 10 mg/kg PO; Sai Life Sciences, India) formulated in 0.5% carboxymethylcellulose/0.9% sterile saline (CMC) or CMC vehicle. A separate cohort of rats were dosed with ketamine (10 mg/kg IV) or sterile saline vehicle via the lateral-tail vein. Sleep-deprivation learning studies Heterospecific rough-and-tumble play was conducted as previously described [16, 20]. Animals received three consecutive days of light-touch habituation, which does not induce ultrasonic vocalizations (USVs), before testing [20]. These animals did not receive sleep-EEG surgeries or testing. Briefly, heterospecific rough-and-tumble play stimulation was administered by the experimenter’s right hand. Rats received 3 minutes of play consisting of alternating 15-second blocks of play and 15 seconds of no stimulation. The experimenter was blind to the treatment condition of the animals. At the end of the 3-minute session, running speed (cm/second) for the animal to traverse a 57-centimeter arena and touch the experimenters’ hand to self-administer play was measured manually with a digital stopwatch. High-frequency USVs were recorded (see below; Avisoft UltraSoundGate, Germany) during the 6 × 15-second no-stimulation blocks and analyzed by sonogram (Avisoft SASlab Pro, Germany) in a blinded manner, as described previously [18]. Rats were dosed with NYX-2925 (1 mg/kg PO; Sai Advantium, India) in CMC or CMC vehicle 1 hour before testing. Previously it has been shown that positive modulation of the NMDAR with d-cycloserine has been shown to rescue sleep deprivation-induced deficits in learning [21]. Home cage activity and USV recordings Rats were housed three per cage by experimental condition in circular acrylic home cages (35.6 cm diameter × 30.5 cm high; Pinnacle, USA) with aspen-wood chip bedding and a Plexiglas lid with 9 × 50 cm holes and a microphone (Avisoft, Germany) suspended from the center hole. These animals did not receive sleep-EEG surgeries or testing. Rats were maintained on a 12:12 light:dark cycle (lights-on at 6 am), and given ad libitum access to lab chow and tap water throughout the study. USVs were recorded (Avisoft UltraSoundGate, Germany) in 15-minute bins for 24 hours and analyzed via sonogram (Avisoft SASlab Pro, Germany) with high (R > .90) blinded inter-rater reliability. Rats were dosed with NYX-2925 (10 mg/kg PO), CMC vehicle PO, ketamine (10 mg/kg IV), or sterile saline vehicle (IV) at ZT5, and did not previously receive EEG surgery or testing. Frequency-modulated 50-kHz USVs and 20-kHz USVs are validated measures of positive and negative emotion in rats, respectively [22]. A wide range of hedonic stimuli (social interaction, food, drugs of abuse) increase rates of frequency-modulated 50-kHz USVs [23–26]. Aversive stimuli, on the other hand, uniformly decrease rates of frequency-modulated 50-kHz USVs and increase rates of 20-kHz USVs [23, 26, 27]. Additionally, the neural circuit of rat 50-kHz USVs and 20-kHz is essentially the same as the human positive and negative affect circuit [22, 23]. Hedonic frequency-modulated 50-kHz USVs, neutral flat 50-kHz USVs, and aversive 20-kHz USVs were classified as previously described [23]. Behavioral activation (e.g., locomotor activity, sniffing, eating, and drinking) was quantified by the total sound output from the recording microphone which captures both sonic and ultrasonic sound. Inactivity (% of total time) was defined by the amount of time in which the sound intensity was similar to levels recorded from an empty cage. Activity was defined as 100− (minus) inactivity. Statistical analysis Sleep, activity, and USV data were analyzed by analysis of variance (ANOVA), followed by Fisher’s PLSD post hoc test (Statview, USA). The accuracy of the transition from light:dark for locomotor activity was calculated via a three parameter time-response curve (GraphPad Prism, USA), with accuracy being measured by the time to a half-maximal increase in activity relative to lights-off. The level of statistical significance was set at p < .05. Results NYX-2925 (1 mg/kg PO), as compared to vehicle, increased NREM sleep time (Figure 1, A and B; F(1, 31) = 15.8, p < .05; Fishers PLSD post hoc test NYX-2925 vs. vehicle for non-sleep-deprived rats p < .05 and sleep-deprived rats p < .05) and total sleep time (F(1, 31) = 12.6, p < .05; Fishers PLSD post hoc test NYX-2925 vs. vehicle for non-sleep-deprived rats p < .05 and sleep-deprived rats p < .05) during the sleep phase (lights-on) in both non-deprived and sleep-deprived rats without affecting REM sleep (Figure 1, D and E; F(1, 31) = 0.01, p > .05). In contrast, ketamine (10 mg/kg IV), as compared to vehicle, acutely suppressed NREM and REM sleep followed by a rebound increase in NREM (Figure 1C) and REM (Figure 1F) in non-deprived rats (NREM – Drug × Time [F(9, 12) = 3.4, p < .05]; Fishers PLSD post hoc test ketamine vs. vehicle 1 and 2 hours post dose [decrease] and 4 hours post dose [increase] p < .05; REM – Drug × Time [F(9, 12) = 2.8, p < .05]; Fishers PLSD post hoc test ketamine vs. vehicle 1 + 2 hours post dose [decrease] and 5 + 6, 7 + 8 hours post dose [increase] p < .05). Figure 1. Open in new tabDownload slide NYX-2925 facilitates NREM sleep during the light phase, whereas the NMDAR antagonist ketamine inhibits NREM. Non-deprived or sleep-deprived (ZT0-6) rats received either NYX-2925 (1 mg/kg PO), ketamine (10 mg/kg IV; non-deprived only) or vehicle at ZT5 and sleep EEG and EMG was recorded for 24 hours post-dosing. (A) NREM and (D) REM sleep diurnal rhythm before (baseline) and after NYX-2925 or vehicle administration. NYX-2925 increased (B) NREM sleep but did not alter (E) REM sleep during the light phase in both non-deprived and sleep-deprived rats. In contrast, ketamine acutely reduced both (C) NREM and (F) REM sleep in non-deprived rats followed by rebound increase in both sleep states. *p < .05 Fisher’s PLSD post hoc test vs. vehicle. Data are reported as mean ± SEM. Sleep deprived (vehicle n = 7, NYX-2925 n = 9, non-deprived controls n = 14), non-sleep deprived (vehicle n = 8, NYX-2925 n = 11), ketamine (vehicle n = 11, ketamine n = 4). NYX-2925 (0.1, 1, 10 mg/kg PO) facilitated NREM sleep 24 hours post-dosing in sleep-deprived rats (Figure 2A, F(3, 45) = 4.3, p < .05, Fishers PLSD post hoc 0.1, 1, 10 vs. vehicle p < .05). The 10 mg/kg PO dose showed a persistent effect in continuing to facilitate NREM at 48, and 72 hours post-dosing (Figure 2B, F(1, 21) = 6.5, F(1, 14) 5.8, F(1, 13) 6.7, p < .05). In addition, NYX-2925 (10 mg/kg PO) increased sleep-bout duration, increased NREM to REM latency, decreased delta power in wake, and increased both delta power in NREM and theta power in REM (Figure 2C, F(1, 14) = 10.3, 9.9, 9.4, 7.8, 13.2, p < .05) as compared to vehicle across all three testing days. Figure 2. Open in new tabDownload slide Dose response and time course for NYX-2925 facilitation of NREM in sleep-deprived rats. Sleep-deprived (ZT0-6; day 1–2) rats received either NYX-2925 (0.1, 1, 10 mg/kg PO) or vehicle at ZT5 and sleep EEG and EMG were recorded for 3 days post-dosing. (A) NYX-2925 (0.1, 1, 10 mg/kg PO) increased NREM sleep during the light phase as compared to vehicle during the first 24 hours after dosing, whereas (B) only the 10 mg/kg PO dose facilitated NREM across all 3 days. (C) NYX-2925 (10 mg/kg PO) across all 3 days significantly increased the duration of individual sleep bouts, the amount of time spent in NREM before transitioning to REM in a sleep bout, delta power during NREM, and theta power during REM. NYX-2925 also decreased delta power during wake. p < .05 Fisher’s PLSD post hoc test vs. vehicle for all doses and timepoints. Data are reported as mean ± SEM. (A) Vehicle n = 15, NYX-2925 0.1 mg/kg n = 9, 1 mg/kg n = 17, 10 mg/kg n = 8; (B) vehicle N = 15 (d1), 8 (d2), 7 (d1), NYX-2925 n = 8; (C) n = 8 per group. Twenty-four hours of sleep deprivation resulted in a suppression of positive emotional learning (PEL), play reward as measured by running speed to self-administer play, and hedonic USVs, and an increase in aversive USVs, all of which were rescued by 1 mg/kg NYX-2925 (Figure 3A–C, F(2, 15) = 40.5, 26.4, 17.2, 12.0, p < .05; Fishers PLSD post hoc test p < .05 non-deprived vs. deprived-vehicle and deprived-NYX-2925 vs. deprived-vehicle separately for each dependent measure). NYX-2925 (1 mg/kg; 1 hour post-dosing) was chosen for the PEL study given a previous report showing an increase in PEL at this dose and time-point in non-deprived rats [16]. Twenty-four hours of sleep deprivation also lead to a complete elimination of REM (mean ± SEM 100 ± 0.0 reduction, within subjects t(5) = 16.9, p < .05), and a reduction in NREM (mean ± SEM 70.1 ± 2.6 reduction, within subjects t(5) = 24.4, p < .05), and total sleep time (mean ± SEM 73.7 ± 2.5 reduction, within subjects t(5) = 29.1 p < .05) as compared to non-deprived baseline values (data not shown), thus validating the 24-hour sleep-deprivation protocol. Figure 3. Open in new tabDownload slide Sleep deprivation inhibits NMDAR-dependent positive emotional learning which is rescued with pretreatment with NYX-2925. (A) NYX-2925 (1 mg/kg PO) increased the rates of hedonic 50-kHz USVs in response to a temporal conditioned stimulus (CS) that predicated heterospecific rough-and-tumble play in rats that had received 23 hours of sleep deprivation as well as non-dosed/non-deprived naive controls. (B) Rates of hedonic 50-kHz USVs or aversive 20-kHz USVs in response to unconditioned heterospecific rough-and-tumble play. (C) Approach latency (cm/second) for the rats to approach the experimenter’s hand in order to self-administer heterospecific rough and tumble play. *p < .05 NYX-2925 vs. vehicle, #p < .05 naive vs. vehicle Fisher’s PLSD post hoc test. Data are reported as mean ± SEM. N = 6 per group. In the 24-hour home cage USV recording study, NYX-2925 (1 mg/kg) also facilitated the amplitude of the diurnal rhythm of locomotor behavior as well as positive affect, and enhanced positive affect amplitude by suppressing wrong-time activity, whereas ketamine inhibited the amplitude of locomotor behavior by increasing wrong-time activity. As shown in Figure 4A, NYX-2925 enhanced the amplitude of the positive affect as measured by vector amplitude on activity during the dark phase (F(2, 13) = 16.5, 21.3, p < .05; Fishers PLSD post hoc test p < .05 NYX-2925 vs. vehicle), without altering phase angle or activity during the dark phase (F(2, 13) = 1.3, 0.2, p > .05). A clear peak in positive affect was seen across all dosing groups during the lights-off period as measured by the standard deviation of the phase angle vs. a random distribution (F test; p <.0001). As shown in Figure 4B, NYX-2925 suppressed negative affect across the 24-hour period measured by dark-phase and light-phase activity (F(2, 13) = 4.2, 4.7, p < .05; Fishers PLSD post hoc test p < .05 NYX-2925 vs. vehicle), without altering phase amplitude or angle (F(2, 13) = 3.6, 0.2, p > .05). A clear peak in negative affect was seen across all dosing groups during the lights-off period as measured by the standard deviation of the phase angle vs. a random distribution (F test [p <.0001]). As shown in Figure 4C, NYX-2925 enhanced, and ketamine inhibited, the amplitude of the locomotor activity rhythm as measured by the percent of total activity that occurred during the dark phase (F(2, 13) = 13.6, p < .05; Fishers PLSD post hoc test p < .05 NYX-2925 vs. vehicle or ketamine vs. vehicle), which was primarily due to decreased and increased activity during the light phase for NYX-2925 and ketamine, respectively (F(2, 13) = 122.8, p < .05; Fishers PLSD post hoc test p < .05 NYX-2925 vs. vehicle or ketamine vs. vehicle), with now change seen in activity during the dark (F(2, 13) = 0.4, p > .05). Using circular statistics, ketamine decreased the vector amplitude (F(2, 13) = 15.4, p < .05; Fishers PLSD post hoc test p < .05 ketamine vs. vehicle) and shifted the phase angle (F(2, 13) = 31.1, p < .05; Fishers PLSD post hoc test p < .05 ketamine vs. vehicle). A clear peak in locomotor behavior was seen across all dosing groups during the lights-off period as measured by the standard deviation of the phase angle vs. a random distribution as measured by an F test (p < .0001). The accuracy of the transition from light:dark for locomotor was facilitated by NYX-2925 as compared to vehicle (F(1,10) = 6.1, p < .05), whereas the ketamine group did not show a clear activity transition (i.e., they did not adequately fit the three-phase model). The mean ± SEM absolute error (in min) vs. lights-on time values were: vehicle 34.0 ± 5.6; NYX-2925 15.9 ± 4.7; ketamine 128.3 ± 41.5. Fragmentation was indexed as the percent of the time in which phase inappropriate behavior was exhibited, namely locomotor behavior and/or ultrasonic calling during the lights-on period, and lack of locomotor behavior and/or ultrasonic calling during the lights-off period. NYX-2925 decreased and ketamine increased fragmentation using this measure (F(2,13) = 19.0, p <.05; Fishers PLSD post hoc test p < .05 for each pairwise comparison); the mean ± SEM% fragmentation values were: vehicle 34.0 ± 5.8; NYX-2925 14.6 ± 3.5; ketamine 59.4 ± 4.6. As shown in Figure 4D, ketamine eliminated rates of positive affect and total affect for the first 4 hours post-dosing (F(2, 13) = 7.9, 6.4, p < .05; Fishers PLSD post hoc test p <.05 ketamine vs. vehicle) without affecting negative affect (F(2, 13) = 2.5, p > .05). Figure 4. Open in new tabDownload slide NYX-2925 facilitates the amplitude and phase transition timing of diurnal rhythm of both activity and emotional expression whereas ketamine does the opposite. Non-deprived rats received either NYX-2925 (10 mg/kg PO), ketamine (10 mg/kg IV) or vehicle at ZT5 and both sound levels (activity) and USVs (hedonic and aversive calls) were recorded in sound-attenuated chambers housing three rats per cage. (A–C) Diurnal rhythm of locomotor activity as well as positive and negative affect. (D) Acutely, ketamine suppressed hedonic and aversive USVs for the first 4 hours post-dosing. *p < .05 ANOVA NYX-2925 vs. vehicle. Data are reported as mean ± SEM. Vehicle n = 6, NYX-2925 n = 6, ketamine n = 4. Discussion The data presented here demonstrate that NMDAR activation enhances the amplitude of the day–night cycle of NREM sleep and emotion. The transition from sleep and non-affect to wake and positive affect occurs at lights-off. The quality of this transition, namely accuracy of the transition time, the amplitude of the change, and the lack of fragmentation in the activity cycle requires NMDAR activation. Positive emotional expression also follows the day–night cycle. Negative affect appears to occur when out-of-phase activity is exhibited. NMDAR activation improves sleep and mood by facilitating these behaviors at the appropriate times of day. The ability of the lights-off stimulus to trigger a rapid and long-lasting increase in activity and positive affect appears to be a naturalistic form of NMDAR-dependent behavioral plasticity. The long-lasting effects of NYX-2925 are likely due to enhancement of structural plasticity, as evidenced by trafficking of NMDA and AMPA receptors into the synapse, as well as mature dendritic spine formation [16]. The homeostatic regulation of sleep and affective states during wake appear to be regulated by a form of NMDAR-dependent behavioral plasticity. In rats, lights-off is a strong signal for the induction of locomotor activity and pro-social behavior resulting in hedonic 50-kHz calls (Figure 4). The temporal precision (phase angle) as well as the amplitude and the fidelity of the response (fragmentation) is facilitated by NMDAR activation and inhibited by NMDAR antagonism (Figure 4). Thus, NMDAR activation appears to alter primarily the activity transitions from lights-on/off as depicted by the model in Figure 5. Figure 5. Open in new tabDownload slide Sleep-affect model. (top panel) NYX-2925 treatment synchronized day–night activity patterns results in limited wake/affect during the day and robust positive affect during the night. (bottom panel) NMDAR antagonists or sleep deprivation disrupts the rhythm. Desynchronized activity patterns lead to increased wake and negative affect during the day and mixed affect during the night. The synchronization of activity and affect appear to be a form of NMDAR-dependent behavioral plasticity which is regulated by the lights-on and lights-off stimuli. Ketamine may produce its antidepressant effects through an acute suppression of emotion followed by a rebound increase in positive emotion and sleep. In the studies reported here, ketamine produces an immediate and robust suppression of positive and negative affect followed by a rebound increase in positive affect (Figure 4), which is consistent with the acute antidepressant effect of ketamine [28]. In addition, the rebound increase in NREM and REM after ketamine treatment are consistent with the rat and human literature on ketamine and sleep [29, 30], and may contribute to the long-lasting therapeutic effects of the drug [29]. In rats, the daily rhythm of positive affect appears to be maximal during the active lights-off period (Figure 4A) and is driven by an NMDAR-dependent process. Descriptively, there appears to be early and late activity cycle peaks in positive affect, especially in the NYX-2925 group, which may help explain conflicting reports in the human literature showing either the strongest peak in positive affect occurring in the morning or in the night, before bed [31, 32]. This may translate to an ability to coordinate pro-social play behavior right after the start of the wake cycle and to complete pro-social behavior right before the end of the wake cycle, which is likely NMDAR-dependent. Future studies should further examine the potential biphasic function of positive affect. Interestingly, error in this diurnal pattern leads to increased negative affect, especially during wake activity in the dark phase. This observation mirrors the increased negative affect seen in humans following sleep deprivation [33]. NMDAR activation facilitates the synchronization of these two bouts of play with lights-on and lights-off, respectively (Figure 4A). Negative affect appears to represent an error or termination signal during 24-hour home cage recordings. Positive affect is associated with the coordination of pro-social play behavior whereas negative affect occurs when these behaviors are not coordinated between individual rats. In particular, negative affect is evident during arousal that occurs during sleep in the light phase, and during the transition from play to sleep during lights-on. It is hypothesized that NMDAR activation reduces negative affect and facilitates the transition from inactivity to activity during lights-off and activity to inactivity during lights-on with negative affect being an index of error in these states. The facilitation of sleep and the daily rhythm of mood by NYX-2925 may serve as a biomarker for brain NMDAR activation and contribute to its therapeutic effects. Disrupted sleep is a common symptom of chronic pain, thus improving sleep quality may prove to be beneficial when treating these disorders. This hypothesis is supported by the finding that sleep deprivation has been shown to increase pain sensitivity [34]. Patients with chronic and/or neuropathic pain show a marked reduction in sleep quality; this is especially evident in increased sleep fragmentation and feeling “unrefreshed” in the morning, and these sleep disturbances are positively correlated to pain [17, 18]. Interestingly pregabalin, one of the most widely used therapies in chronic pain, both reduces pain scores and improves sleep quality in post-herpetic neuralgia patients [35] but comes with substantial side effects and a burdensome dosing regimen. Facilitation of NMDAR activity with NYX-2925 has the potential to address the symptoms of chronic pain while also treating the associated deficits in sleep, which are essential for restorative homeostatic plasticity in centrally mediated pain circuits. Funding J.R.M. was supported by National Institutes of Health (NIH) grant NS100173. J.S.B. was supported by NIH grants MH094835 and CA199928. Conflict of interest statement. J.S.B., T.M.M., M.A.K., R.A.K., and J.R.M. are full-time employees of Aptinyx. L.C. is an Atpyinx consultant. E.P.C. was an Aptinyx consultant. References 1. Campbell IG , et al. NREM delta stimulation following MK-801 is a response of sleep systems . J Neurophysiol. 1996 ; 76 ( 6 ): 3714 – 3720 . Google Scholar Crossref Search ADS PubMed WorldCat 2. van der Staay FJ , et al. Effects of the cognition impairer MK-801 on learning and memory in mice and rats . Behav Brain Res. 2011 ; 220 ( 1 ): 215 – 229 . Google Scholar Crossref Search ADS PubMed WorldCat 3. Ledgerwood L , et al. Effects of D-cycloserine on extinction of conditioned freezing . Behav Neurosci. 2003 ; 117 ( 2 ): 341 – 349 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Monahan JB , et al. D-cycloserine, a positive modulator of the N-methyl-D-aspartate receptor, enhances performance of learning tasks in rats . Pharmacol Biochem Behav. 1989 ; 34 ( 3 ): 649 – 653 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Wang Z , et al. Quantitative phosphoproteomic analysis of the molecular substrates of sleep need . Nature . 2018 ; 558 ( 7710 ): 435 – 439 . Google Scholar Crossref Search ADS PubMed WorldCat 6. Chen C , et al. Altered NMDA receptor trafficking contributes to sleep deprivation-induced hippocampal synaptic and cognitive impairments . Biochem Biophys Res Commun. 2006 ; 340 ( 2 ): 435 – 440 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Idzikowski C . The pharmacology of human sleep, a work in progress? Curr Opin Pharmacol. 2014 ; 14 : 90 – 96 . Google Scholar Crossref Search ADS PubMed WorldCat 8. Izquierdo I , et al. Behavioural pharmacology and its contribution to the molecular basis of memory consolidation . Behav Pharmacol. 2000 ; 11 ( 7-8 ): 517 – 534 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Melo I , et al. Sleep supports cued fear extinction memory consolidation independent of circadian phase . Neurobiol Learn Mem. 2016 ; 132 : 9 – 17 . Google Scholar Crossref Search ADS PubMed WorldCat 10. Smith C . Sleep states and memory processes . Behav Brain Res. 1995 ; 69 ( 1-2 ): 137 – 145 . Google Scholar Crossref Search ADS PubMed WorldCat 11. Yang G , et al. Sleep promotes branch-specific formation of dendritic spines after learning . Science. 2014 ; 344 ( 6188 ): 1173 – 1178 . Google Scholar Crossref Search ADS PubMed WorldCat 12. Walker MP , et al. Practice with sleep makes perfect: sleep-dependent motor skill learning . Neuron. 2002 ; 35 ( 1 ): 205 – 211 . Google Scholar Crossref Search ADS PubMed WorldCat 13. Stickgold R , et al. Sleep, learning, and dreams: off-line memory reprocessing . Science. 2001 ; 294 ( 5544 ): 1052 – 1057 . Google Scholar Crossref Search ADS PubMed WorldCat 14. Vyazovskiy VV , et al. Molecular and electrophysiological evidence for net synaptic potentiation in wake and depression in sleep . Nat Neurosci. 2008 ; 11 ( 2 ): 200 – 208 . Google Scholar Crossref Search ADS PubMed WorldCat 15. Maret S , et al. Sleep and waking modulate spine turnover in the adolescent mouse cortex . Nat Neurosci. 2011 ; 14 ( 11 ): 1418 – 1420 . Google Scholar Crossref Search ADS PubMed WorldCat 16. Khan MA , et al. NYX-2925 Is a Novel NMDA receptor-specific spirocyclic-β-lactam that modulates synaptic plasticity processes associated with learning and memory . Int J Neuropsychopharmacol. 2018 ; 21 ( 3 ): 242 – 254 . Google Scholar Crossref Search ADS PubMed WorldCat 17. McCracken LM , et al. Disrupted sleep patterns and daily functioning in patients with chronic pain . Pain Res Manag. 2002 ; 7 ( 2 ): 75 – 79 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Foley D , et al. Sleep disturbances and chronic disease in older adults: results of the 2003 National Sleep Foundation Sleep in America Survey . J Psychosom Res . 2004 ; 56 ( 5 ): 497 – 502 . doi: 10.1016/j.jpsychores.2004.02.010 Google Scholar Crossref Search ADS PubMed WorldCat 19. Gao V , et al. Multiple classifier systems for automatic sleep scoring in mice . J Neurosci Methods. 2016 ; 264 : 33 – 39 . Google Scholar Crossref Search ADS PubMed WorldCat 20. Ghoreishi-Haack N , et al. NYX-2925 Is a Novel N-methyl-d-aspartate receptor modulator that induces rapid and long-lasting analgesia in rat models of neuropathic pain . J Pharmacol Exp Ther. 2018 ; 366 ( 3 ): 485 – 497 . Google Scholar Crossref Search ADS PubMed WorldCat 21. Silvestri AJ , et al. Effects of REM deprivation and an NMDA agonist on the extinction of conditioned fear . Physiol Behav. 2008 ; 93 ( 1-2 ): 274 – 281 . Google Scholar Crossref Search ADS PubMed WorldCat 22. Burgdorf J , et al. Frequency-modulated 50 kHz ultrasonic vocalizations: a tool for uncovering the molecular substrates of positive affect . Neurosci Biobehav Rev. 2011 ; 35 ( 9 ): 1831 – 1836 . Google Scholar Crossref Search ADS PubMed WorldCat 23. Burgdorf J , et al. Ultrasonic vocalizations of rats (Rattus norvegicus) during mating, play, and aggression: behavioral concomitants, relationship to reward, and self-administration of playback . J Comp Psychol. 2008 ; 122 ( 4 ): 357 – 367 . Google Scholar Crossref Search ADS PubMed WorldCat 24. Burgdorf J , et al. Nucleus accumbens amphetamine microinjections unconditionally elicit 50-kHz ultrasonic vocalizations in rats . Behav Neurosci. 2001 ; 115 ( 4 ): 940 – 944 . Google Scholar Crossref Search ADS PubMed WorldCat 25. Panksepp J , et al. 50-kHz chirping (laughter?) in response to conditioned and unconditioned tickle-induced reward in rats: effects of social housing and genetic variables . Behav Brain Res. 2000 ; 115 ( 1 ): 25 – 38 . Google Scholar Crossref Search ADS PubMed WorldCat 26. Burgdorf J , et al. Anticipation of rewarding electrical brain stimulation evokes ultrasonic vocalization in rats . Behav Neurosci. 2000 ; 114 ( 2 ): 320 – 327 . Google Scholar Crossref Search ADS PubMed WorldCat 27. Burgdorf J , et al. Evaluation of rat ultrasonic vocalizations as predictors of the conditioned aversive effects of drugs . Psychopharmacology (Berl). 2001 ; 155 ( 1 ): 35 – 42 . Google Scholar Crossref Search ADS PubMed WorldCat 28. Zarate CA Jr, et al. A randomized trial of an N-methyl-D-aspartate antagonist in treatment-resistant major depression . Arch Gen Psychiatry. 2006 ; 63 ( 8 ): 856 – 864 . Google Scholar Crossref Search ADS PubMed WorldCat 29. Duncan WC , et al. Concomitant BDNF and sleep slow wave changes indicate ketamine-induced plasticity in major depressive disorder . Int J Neuropsychopharmacol. 2013 ; 16 ( 2 ): 301 – 311 . Google Scholar Crossref Search ADS PubMed WorldCat 30. Feinberg I , et al. Stimulation of NREM delta EEG by ketamine administration during waking: demonstration of dose dependence . Neuropsychopharmacology. 1995 ; 12 ( 1 ): 89 – 90 . Google Scholar Crossref Search ADS PubMed WorldCat 31. Stone AA , et al. A population approach to the study of emotion: diurnal rhythms of a working day examined with the Day Reconstruction Method . Emotion. 2006 ; 6 ( 1 ): 139 – 149 . Google Scholar Crossref Search ADS PubMed WorldCat 32. Boivin DB , et al. Complex interaction of the sleep-wake cycle and circadian phase modulates mood in healthy subjects . Arch Gen Psychiatry. 1997 ; 54 ( 2 ): 145 – 152 . Google Scholar Crossref Search ADS PubMed WorldCat 33. Scott JP , et al. Effects of sleep deprivation and exercise on cognitive, motor performance and mood . Physiol Behav. 2006 ; 87 ( 2 ): 396 – 408 . Google Scholar Crossref Search ADS PubMed WorldCat 34. Roehrs T , et al. Sleep loss and REM sleep loss are hyperalgesic . Sleep. 2006 ; 29 ( 2 ): 145 – 151 . Google Scholar Crossref Search ADS PubMed WorldCat 35. Sabatowski R , et al. ; 1008-045 Study Group . Pregabalin reduces pain and improves sleep and mood disturbances in patients with post-herpetic neuralgia: results of a randomised, placebo-controlled clinical trial . Pain. 2004 ; 109 ( 1-2 ): 26 – 35 . Google Scholar Crossref Search ADS PubMed WorldCat © Sleep Research Society 2019. Published by Oxford University Press [on behalf of the Sleep Research Society]. All rights reserved. This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial reproduction and distribution of the work, in any medium, provided the original work is not altered or transformed in any way, and that the work is properly cited. For commercial re-use, please contact [email protected] © Sleep Research Society 2019. Published by Oxford University Press [on behalf of the Sleep Research Society]. All rights reserved.

Feemster, John, C;Jung,, Youngsin;Timm, Paul, C;Westerland, Sarah, M;Gossard, Thomas, R;Teigen, Luke, N;Buchal, Lauren, A;Cattaneo, Elena F, D;Imlach, Charlotte, A;Mccarter, Stuart, J;Smith, Kevin, L;Boeve, Bradley, F;Silber, Michael, H;St Louis, Erik, K

Gandotra, Kamal; Jaskiw, George E; Wilson, Brigid; Konicki, P Eric; Rosenberg, Carl E; Strohl, Kingman P

doi: 10.1093/sleep/zsz144pmid: 31595968

Nightmares are an intrinsic component of several disorders including post-traumatic stress disorder (PTSD), rapid eye movement (REM) sleep behavior disorder and night terrors [1] and afflict some 2%–6% of adults [2–4]. The prevalence is higher in veterans who have suffered combat-related trauma [5]. Pickworth et al. [6] suggested that amplified central nervous system α1-adrenergic receptor stimulation heightened reactivity during sleep and thus constituted the basis of nightmares in PTSD. We observed patients on risperidone had fewer complaints about nightmares. We report a series of seven patients in whom low-dose risperidone proved effective in attenuating nightmares that emerged as part of their primary PTSD [1]. The pharmacological rationale for trying risperidone was based on prior case series illustrating its success in treating nightmares. Pharmacological agents affecting the neurotransmitters norepinephrine, serotonin, and dopamine are associated with patient reports of nightmares. A possible association exists between reports of nightmares and agents affecting the neurotransmitters acetylcholine, GABA, and histamine. Given its multiple receptor binding profile, risperidone acts as an antagonist at the following receptors: D2, 5-HT2A, Alpha 1, Alpha 2, H1 (moderate affinity). Thus, its action to alleviate nightmares may be multifactorial regarding receptors. In this case, series veterans treated with risperidone filled out the Nightmare Distress Questionnaire (NDQ) at each visit; one of the physicians who treated a veteran with risperidone successfully ablated his nightmares subsequently leading to the case series. We monitored veterans treated with risperidone for nightmares, the sample only includes veterans prescribed risperidone to treat nightmare disorder, no other medications were utilized to treat nightmares. All veterans met the criteria for nightmare disorder as defined by the ICSD [7]; remission was designated as a score of 0 on the nightmare distress questionnaire; as veterans who scored zero were no longer reporting anxiety dreams and were deemed in remission. Individuals who scored 21 met the ICSD criteria for nightmare disorder. Unfortunately, we have no data regarding the frequency of nightmares as no question in the Nightmare Distress Questionnaire quantifies the frequency of nightmares. A linear mixed-effects model was estimated predicting nightmare distress questionnaire scores with time, dose, the interaction of time and dose, and a random intercept for each subject. This modeling approach was selected given the unbalanced groups and missing observations. Differences in questionnaire scores over time were assessed as model contrasts, averaging over levels of dose. All analyses were performed in R 3.5.1 using the lme4 and emmeans packages. There was an initial prompt decrease in symptoms as the NDQ scores at 4 weeks exhibited a trend toward baseline. At 12 weeks, 30% of the sample had an NDQ score of 0 (Figure 1). Figure 1. Open in new tabDownload slide Legends shown are a mean ± SD of NDQ values at baseline, 4, 8, and 12 weeks. Test for trends (p < 0.01). Figure 1. Open in new tabDownload slide Legends shown are a mean ± SD of NDQ values at baseline, 4, 8, and 12 weeks. Test for trends (p < 0.01). In 30% of the veterans in this case series, nightmares resolved over 12 weeks of risperidone in the low-dose range (0.5–2 mg HS). In an earlier case series (n = 4), risperidone doses 1–3 mg/d were reported to attenuate nightmares [8]. Risperidone improves REM density acutely [9], and this may explain is transient short-term therapeutic benefit for nightmares. However, more than 12 weeks, not all improved. This could be tachyphylaxis. Thus, long-term efficacy would require confirmation by a randomized controlled trial. Low-dose risperidone was generally well-tolerated and effective in resolving PTSD-related nightmares more than 12 weeks. There is increasing awareness that nightmares per se increase distress and elevate suicide risk [10, 11] Conflict of interest statement. None declared. Work Performed: Louis Stokes VA Medical Center, 10701 East Blvd, Cleveland, Ohio 44106. References 1. American Psychiatric Association . Diagnostic and Statistical Manual of Mental Disorders DSM-5 . Washington, DC : American Psychiatric Publishing, Inc. ; 2013 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 2. Belicki K , et al. Predisposition for nightmares: a study of hypnotic ability, vividness of imagery, and absorption . J Clin Psychol. 1986 ; 42 ( 5 ): 714 – 718 . Google Scholar Crossref Search ADS PubMed WorldCat 3. Bixler EO , et al. Prevalence of sleep disorders in the Los Angeles metropolitan area . Am J Psychiatry. 1979 ; 136 ( 10 ): 1257 – 1262 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Ohayon MM , et al. Prevalence of nightmares and their relationship to psychopathology and daytime functioning in insomnia subjects . Sleep. 1997 ; 20 ( 5 ): 340 – 348 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Neylan TC , et al. Sleep disturbances in the Vietnam generation: findings from a nationally representative sample of male Vietnam veterans . Am J Psychiatry. 1998 ; 155 ( 7 ): 929 – 933 . Google Scholar Crossref Search ADS PubMed WorldCat 6. Pickworth WB , et al. Sleep suppression induced by intravenous and intraventricular infusions of methoxamine in the dog . Exp Neurol. 1977 ; 57 ( 3 ): 999 – 1011 . Google Scholar Crossref Search ADS PubMed WorldCat 7. American Academy of Sleep Medicine . International Classification of Sleep Disorders . 3rd ed. Darien : American Academy of Sleep Medicine ; 2014 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC 8. Detweiler MB , et al. Risperidone for post-traumatic combat nightmares: a report of four cases . Consult Pharm. 2011 ; 26 ( 12 ): 920 – 928 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Giménez S , et al. Effects of olanzapine, risperidone and haloperidol on sleep after a single oral morning dose in healthy volunteers . Psychopharmacology (Berl). 2007 ; 190 ( 4 ): 507 – 516 . Google Scholar Crossref Search ADS PubMed WorldCat 10. Sandman N , et al. Nightmares as predictors of suicide: an extension study including war veterans . Sci Rep. 2017 ; 7 : 44756 . Google Scholar Crossref Search ADS PubMed WorldCat 11. Perlis ML , et al. Nocturnal wakefulness as a previously unrecognized risk factor for suicide . J Clin Psychiatry. 2016 ; 77 ( 6 ): e726 – e733 . Google Scholar Crossref Search ADS PubMed WorldCat Published by Oxford University Press on behalf of Sleep Research Society (SRS) 2019. 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) 2019.

Melpignano,, Andrea;Parrino,, Liborio;Santamaria,, Joan;Gaig,, Carles;Trippi,, Irene;Serradell,, Monica;Mutti,, Carlotta;Riccò,, Matteo;Iranzo,, Alex

doi: 10.1093/sleep/zsz142pmid: 31323084

Abstract Objective To evaluate the role of sleep cyclic alternating pattern (CAP) in patients with isolated REM sleep behavior disorder (IRBD) and ascertain whether CAP metrics might represent a marker of phenoconversion to a defined neurodegenerative condition. Methods Sixty-seven IRBD patients were included and classified into patients who phenoconverted to a neurodegenerative disease (RBD converters: converter REM sleep behavior disorder [cRBD]; n = 34) and remained disease-free (RBD non-converters: non-converter REM sleep behavior disorder [ncRBD]; n = 33) having a similar follow-up duration. Fourteen age- and gender-balanced healthy controls were included for comparisons. Results Compared to controls, CAP rate and CAP index were significantly decreased in IRBD mainly due to a decrease of A1 phase subtypes (A1 index) despite an increase in duration of both CAP A and B phases. The cRBD group had significantly lower values of CAP rate and CAP index when compared with the ncRBD group and controls. A1 index was significantly reduced in both ncRBD and cRBD groups compared to controls. When compared to the ncRBD group, A3 index was significantly decreased in the cRBD group. The Kaplan-Meier curve applied to cRBD estimated that a value of CAP rate below 32.9% was related to an average risk of conversion of 9.2 years after baseline polysomnography. Conclusion IRBD is not exclusively a rapid eye movement (REM) sleep parasomnia, as non-rapid eye movement (non-REM) sleep microstructure can also be affected by CAP changes. Further studies are necessary to confirm that a reduction of specific CAP metrics is a marker of neurodegeneration in IRBD. isolated REM sleep behavior disorder, cyclic alternating pattern, neurodegeneration, sleep microstructure Statement of Significance Isolated REM behavior disorder (IRBD) is a sleep condition that often precedes Parkinson disease, dementia with Lewy bodies, and multiple system atrophy. Cyclic alternating pattern (CAP) is a microstructural age-dependent component of non-rapid eye movement (non-REM) sleep and translates a condition of arousal, autonomic, and behavioral resilience. We found that CAP rate and CAP index were decreased in IRBD mainly due to a decrease of A1 phase subtype (A1 index). This finding suggests that non-REM sleep microstructure is involved in IRBD and that CAP metrics might provide information on neurodegeneration in this disorder. Whether the described alterations of CAP parameters are specific markers of early IRBD conversion or sensitive but unspecific measures of potential neurodegeneration will be clarified in future studies. Introduction The loss of physiological muscle atonia during rapid eye movement (REM) sleep and consequent dream-enacting behavior are the most characterizing features of REM sleep behavior disorder (RBD) [1]. Several studies have demonstrated that the majority of patients with the isolated form of RBD (IRBD) eventually develop a neurodegenerative disease, typically the synucleinopathies Parkinson disease (PD), dementia with Lewy bodies (DLB) and multiple system atrophy (MSA) with a risk of conversion of 90.9% in 14 years from the diagnosis of IRBD [2–5]. Markers of neurodegeneration in IRBD patients have been analyzed in several studies [6]. An increased short-term risk for developing a synucleinopathy has been linked to an early onset of olfactory damage, color discrimination loss [7], impaired striatal dopamine transporter uptake and substantia nigra hyperechogenicity [8]. Other works have focused on electroencephalography (EEG) [9], sleep spindle density [10, 11], slow wave characteristics [12], the severity of REM sleep atonia loss [13], heart-rate variability [14] non-motor manifestations [15], gray and white matter anomalies [16], and initial cognitive decline [17, 18]. Polysomnographic (PSG) studies in neurodegenerative disorders are not conclusive and are limited only to traditional sleep parameters of macrostructure and electromyographic (EMG) activity. Cyclic alternating pattern (CAP) is a spontaneous and physiological rhythm of non-rapid eye movement (non-REM) sleep, characterized by EEG oscillations arranged in sequences of cycles composed of transient electro-cortical events (phase A) and by the following retrieval to background EEG activity (phase B). CAP sequence is made of at least two succeeding CAP cycles. Phase A pattern can be divided into three subtypes, on the base of the reciprocal proportion of EEG synchrony and EEG desynchronization: subtype A1 where EEG synchrony of high-amplitude is the main activity, subtype A2 characterized by a combination of slow and fast rhythms with 20%–50% of phase A occupied by EEG desynchronization, and subtype A3 with a predominant EEG activity of rapid low-voltage rhythms with more than 50% of phase A occupied by EEG desynchronization [19]. CAP is a natural component of the sleep texture and a marker of sleep instability [19]. This double-nature allows CAP to exert both a direct involvement in the functional mechanisms of sleep disturbances and a sleep-protecting action, intervening in the homeostatic regulation of sleep and contributing to the mechanisms of self-maintenance against external or internal perturbations [20, 21]. The basic guidelines for scoring CAP are restricted to non-REM sleep where the duration of phases A and B range between 2 and 60 seconds. In REM sleep, EEG arousals are generally separated by intervals longer than 60 seconds, a time-frame which exceeds the scoring criteria of CAP guidelines [22]. So far, several studies have revealed the presence of relevant CAP alterations despite normal sleep macrostructure [19]. However, to the best of our knowledge, only few works have examined CAP in neurodegenerative disorders. In MSA, there is a difficulty in differentiating the CAP A phases A1, A2, and A3 subtypes and there is a significantly lower CAP rate [23]. In subjects with mild cognitive impairment (MCI), CAP rate and CAP A1 phases are decreased, with a relative increase in A2 and A3 subtypes [24]. Decreased duration and reduced number of CAP sequences are reported in Alzheimer disease [24]. RBD is the result of a dysfunction in the brainstem nuclei that modulate REM sleep and their anatomical inputs from other regions [25]. However, some studies have shown that the macrostructure and microstructure of non-REM sleep can also be impaired in IRBD. In particular, increased percentage of N3 [26], increased slow transient EEG events and decreased fast transient events [27] have been described in IRBD. To the best of our knowledge EEG microstructure by CAP analysis has not been examined in IRBD. The aim of the present study was to evaluate the role of CAP in patients with IRBD and to evaluate whether CAP metrics represent an early marker of neurodegeneration. Methods Participants The study comprises 67 consecutive IRBD patients diagnosed by clinical history and by one night of PSG between 1999 and 2016 at the Neurology Service of Hospital Clinic of Barcelona, Barcelona, Spain. At the time of PSG, patients reported the absence of motor symptoms and lacking cognitive complaints, and the neurological examination was normal ruling out parkinsonism, cerebellar signs and other clinical features indicative of neurological dysfunction. After the diagnosis of IRBD was performed, patients were systematically followed at our institutions every 3–6 months, as described in detail elsewhere [2]. At the end of the current study (April 2018), we reviewed our databases to identify those patients from this study who were diagnosed with a defined neurodegenerative syndrome (PD, DLB, MSA, and MCI), according to accepted criteria [28–31]. For the purpose of this study, IRBD patients were classified into (1) patients who developed a defined neurodegenerative syndrome (n = 34; converter group: converter REM sleep behavior disorder [cRBD]) and (2) patients who remained disease-free (n = 33; non-converter group, non-converter REM sleep behavior disorder [ncRBD]) during the follow-up period from baseline (the day when PSG confirmed RBD) until the time of this study (April 2018). The diagnosis of IRBD was performed according to International Classification of Sleep Disorders-3 criteria [1]. We excluded IRBD patients with a diagnosis of other coexistent sleep disorders, an apnea-hypopnea index (AHI) > 15 per hour, a periodic limb movement index during sleep (PLMI) > 15 per hour and those taking psychoactive drugs (e.g., antidepressants, benzodiazepines, clonazepam, melatonin) at the time of PSG. Patients with previous diagnosis of obstructive sleep apnea and treated successfully with CPAP were not excluded. For comparisons, 14 age- and gender-matched healthy controls were included. These controls were selected from our databases of healthy volunteers in whom PSG ruled out RBD, and had AHI ≤ 15 per hour and PLMI ≤ 15 per hour. Controls were non-blood relatives of the IRBD patients included in the current study who were free of neurological diseases and reported lack of motor, cognitive, or sleep complaints such as dream-enacting behaviors and recurrent nightmares. Polysomnography and CAP analysis In PSG we used four EEG channel (C3-A2, C4-A1, O1-A2, O2-A2) plus bipolar montage (C3-01 and C3-O2) for studies recorded before 2007, and six EEG channels (F4-A1, C4-A1, O2-A1, F3-A2, C3-A2, O1-A2) plus bipolar channel (C3-01 and C3-O2) for studies recorded after 2006 (all placed in standard scalp locations using the international 10/20 system), electrooculography for left and right eyes (E1-M2, E2-M2), submental surface EMG of the mentalis and the four limbs (afterwards removed for CAP scoring), electrocardiography, thermistor for nasal and oral airflow, thoracic and abdominal respiratory effort strain gauges, finger pulse oximetry, and synchronized audio-visual recording. Visual sleep-stage scoring was performed according to standardized criteria on a 30-second epochs and sleep efficiency, total sleep time, stage 1 (N1) sleep time, stage 2 (N2) sleep time, stage 3 (N3) sleep time, REM sleep time, arousal index (number of arousals per hour of sleep), AHI (number of apneas plus hypopneas per hour of sleep) and PLMI (number of periodic leg movements per hour of sleep) were calculated according to standard criteria [32, 33]. CAP scoring was performed by a Sleep Medicine expert neurologist (AM). In Barcelona who was previously trained at the Sleep Disorders Center of Parma University. We used RemLogic as software for CAP scoring (Embla RemLogic). The CAP scorer was blind regarding the clinical features of patients including the diagnosis (IRBD and control) and the clinical evolution (converter and nonconverter). Scorer was also unaware of the sample size of each group of participants. Audio-visual recording and EMG in the mentalis and limb traces were removed during CAP assessment to maintain the scorer blindness. CAP was scored using the CAP Atlas criteria by Terzano et al. [22]. CAP periods, with A phase subtypes (A1, A2, A3) and B phases, as well as non-CAP periods were identified. The following CAP parameters were acquired: CAP rate (CAP time/total non-REM time × 100), CAP index, subtypes of A phases indexes and mean duration of A and B phases. Statistical analysis Continuous variables are reported as mean ± standard deviation, being preliminarily evaluated through Shapiro-Wilk W test in order to assess whether their values had a Gaussian distribution. As all variables had p values < 0.100, a non-Gaussian distribution was assumed, and their comparisons were performed through Mann-Whitney test for unpaired data, or Kruskal-Wallis and Dunn’s multiple comparison tests when appropriate. Categorical variables were reported as percent values, and correlates of IRBD groups with demographic and clinical data were initially assessed through the Chi-squared test with Yates correction. The relationship between CAP rate and time of conversion to a defined neurodegenerative syndrome (PD, DLB, MSA, and MCI) was examined with a Kaplan-Meier curve that was used to define the risk for conversion according to CAP parameters at the baseline PSG, based on the median values of CAP rate, and data were compared with a Mantel-Cox log-rank test. Results Demographic and clinical data The cRBD group comprised 27 males and seven females with a mean age of 66.7 ± 5.2 years. The ncRBD group was formed by 24 males and nine females with a mean age of 67.3 ± 5.5 years, while the control group comprised 13 males and one female with a mean age of 63.8 ± 6.6 years. Both IRBD groups and controls were balanced for age and gender. The mean follow-up was similar between the cRBD and the ncRBD groups (7.6 ± 4.1 versus 6.9 ± 3.7 years; p = 0.573). Converters developed PD in 14 cases, DLB in 17 cases, and MCI in three instances. Sleep macrostructure Eventually, no differences in sleep macrostructure were found between controls and the total sample of IRBD individuals (ncRBD + cRBD) except for slight lower sleep efficiency and total sleep time and higher percentage of N1 sleep stage in IRBD (Table 1). There were no differences among the three groups (ncRBD, cRBD, and controls) in the sleep architecture parameters except for lower percentage of N2 sleep stage in the cRBD group (Table 2). Table 1. Conventional polysomnographic parameters in controls and patients with IRBD Controls (n = 14) IRBD (n = 67) P value SE (%) 82.1 ± 7.0 74.3 ± 13.2 0.043 TST (minutes) 378.3 ± 40.5 342.4 ± 65.3 0.039 N1 (%) 13.3 ± 5.4 19.7 ± 10.9 0.034 N2 (%) 51.0 ± 10.0 46.8 ± 11.1 NS N3 (%) 17.1 ± 8.5 15.1 ± 9.6 NS REM sleep (%) 18.6 ± 3.9 18.4 ± 7.4 NS AI/hour (n) 14.1 ± 6.6 18.3 ± 9.2 NS Controls (n = 14) IRBD (n = 67) P value SE (%) 82.1 ± 7.0 74.3 ± 13.2 0.043 TST (minutes) 378.3 ± 40.5 342.4 ± 65.3 0.039 N1 (%) 13.3 ± 5.4 19.7 ± 10.9 0.034 N2 (%) 51.0 ± 10.0 46.8 ± 11.1 NS N3 (%) 17.1 ± 8.5 15.1 ± 9.6 NS REM sleep (%) 18.6 ± 3.9 18.4 ± 7.4 NS AI/hour (n) 14.1 ± 6.6 18.3 ± 9.2 NS SE: Sleep efficiency; TST: total sleep time; N1: sleep stage 1; N2: sleep stage 2; N3: sleep stage 3; REM: rapid eye movement sleep; AI: arousal index. Data are presented as mean ± standard deviation; NS: no significant. Open in new tab Table 1. Conventional polysomnographic parameters in controls and patients with IRBD Controls (n = 14) IRBD (n = 67) P value SE (%) 82.1 ± 7.0 74.3 ± 13.2 0.043 TST (minutes) 378.3 ± 40.5 342.4 ± 65.3 0.039 N1 (%) 13.3 ± 5.4 19.7 ± 10.9 0.034 N2 (%) 51.0 ± 10.0 46.8 ± 11.1 NS N3 (%) 17.1 ± 8.5 15.1 ± 9.6 NS REM sleep (%) 18.6 ± 3.9 18.4 ± 7.4 NS AI/hour (n) 14.1 ± 6.6 18.3 ± 9.2 NS Controls (n = 14) IRBD (n = 67) P value SE (%) 82.1 ± 7.0 74.3 ± 13.2 0.043 TST (minutes) 378.3 ± 40.5 342.4 ± 65.3 0.039 N1 (%) 13.3 ± 5.4 19.7 ± 10.9 0.034 N2 (%) 51.0 ± 10.0 46.8 ± 11.1 NS N3 (%) 17.1 ± 8.5 15.1 ± 9.6 NS REM sleep (%) 18.6 ± 3.9 18.4 ± 7.4 NS AI/hour (n) 14.1 ± 6.6 18.3 ± 9.2 NS SE: Sleep efficiency; TST: total sleep time; N1: sleep stage 1; N2: sleep stage 2; N3: sleep stage 3; REM: rapid eye movement sleep; AI: arousal index. Data are presented as mean ± standard deviation; NS: no significant. Open in new tab Table 2. Conventional polysomnographic parameters between controls and patients with IRBD who remained disease-free and patients who converted to a defined neurodegenerative disorder Controls (n = 14) ncRBD (n = 33) cRBD (n = 34) P value SE (%) 82.1 ± 7.0 75.5 ± 11.9 73.0 ± 14.4 NS TST (minutes) 378.3 ± 40.5 347.1 ± 58.9 338.0 ± 71.7 NS N1 (%) 13.3 ± 5.4 17.7 ± 7.9 21.6 ± 13.1 NS N2 (%) 51.0 ± 9.7 52.2 ± 10.5 41.6 ± 9.0 <0.001 N3 (%) 17.1 ± 8.5 13.1 ± 9.4 17.0 ± 9.6 NS REM sleep (%) 18.6 ± 3.9 16.9 ± 5.9 18.4 ± 6.9 NS AI/hour (n) 14.1 ± 6.6 18.0 ± 7.6 18.5 ± 10.7 NS Controls (n = 14) ncRBD (n = 33) cRBD (n = 34) P value SE (%) 82.1 ± 7.0 75.5 ± 11.9 73.0 ± 14.4 NS TST (minutes) 378.3 ± 40.5 347.1 ± 58.9 338.0 ± 71.7 NS N1 (%) 13.3 ± 5.4 17.7 ± 7.9 21.6 ± 13.1 NS N2 (%) 51.0 ± 9.7 52.2 ± 10.5 41.6 ± 9.0 <0.001 N3 (%) 17.1 ± 8.5 13.1 ± 9.4 17.0 ± 9.6 NS REM sleep (%) 18.6 ± 3.9 16.9 ± 5.9 18.4 ± 6.9 NS AI/hour (n) 14.1 ± 6.6 18.0 ± 7.6 18.5 ± 10.7 NS ncRBD: non-converters REM sleep behavior disorder; cRBD: converters REM sleep behavior disorder; SE: Sleep efficiency; TST: total sleep time; N1: sleep stage 1; N2: sleep stage 2; N3: sleep stage 3; REM: rapid eye movement sleep; AI: arousal index. Data are presented as mean ± standard deviation; NS: no significant. Open in new tab Table 2. Conventional polysomnographic parameters between controls and patients with IRBD who remained disease-free and patients who converted to a defined neurodegenerative disorder Controls (n = 14) ncRBD (n = 33) cRBD (n = 34) P value SE (%) 82.1 ± 7.0 75.5 ± 11.9 73.0 ± 14.4 NS TST (minutes) 378.3 ± 40.5 347.1 ± 58.9 338.0 ± 71.7 NS N1 (%) 13.3 ± 5.4 17.7 ± 7.9 21.6 ± 13.1 NS N2 (%) 51.0 ± 9.7 52.2 ± 10.5 41.6 ± 9.0 <0.001 N3 (%) 17.1 ± 8.5 13.1 ± 9.4 17.0 ± 9.6 NS REM sleep (%) 18.6 ± 3.9 16.9 ± 5.9 18.4 ± 6.9 NS AI/hour (n) 14.1 ± 6.6 18.0 ± 7.6 18.5 ± 10.7 NS Controls (n = 14) ncRBD (n = 33) cRBD (n = 34) P value SE (%) 82.1 ± 7.0 75.5 ± 11.9 73.0 ± 14.4 NS TST (minutes) 378.3 ± 40.5 347.1 ± 58.9 338.0 ± 71.7 NS N1 (%) 13.3 ± 5.4 17.7 ± 7.9 21.6 ± 13.1 NS N2 (%) 51.0 ± 9.7 52.2 ± 10.5 41.6 ± 9.0 <0.001 N3 (%) 17.1 ± 8.5 13.1 ± 9.4 17.0 ± 9.6 NS REM sleep (%) 18.6 ± 3.9 16.9 ± 5.9 18.4 ± 6.9 NS AI/hour (n) 14.1 ± 6.6 18.0 ± 7.6 18.5 ± 10.7 NS ncRBD: non-converters REM sleep behavior disorder; cRBD: converters REM sleep behavior disorder; SE: Sleep efficiency; TST: total sleep time; N1: sleep stage 1; N2: sleep stage 2; N3: sleep stage 3; REM: rapid eye movement sleep; AI: arousal index. Data are presented as mean ± standard deviation; NS: no significant. Open in new tab Sleep microstructure and CAP analysis Compared to controls, CAP rate and CAP index were decreased in IRBD. Despite the decreased amount of CAP rate, the IRBD group showed increased duration in both CAP A and B phases. This reduction of CAP rate was mainly due to a decrease of the subtype A1 index, whereas A2 and A3 indexes remained stable (Table 3). Table 3. CAP parameters in controls and patients with IRBD Controls (n = 14) IRBD (n = 67) P value CAP rate (%) 42.7 ± 7.5 30.0 ± 10.3 <0.001 CAP index (n) 52.1 ± 9.7 39.6 ± 24.6 <0.001 A1 index (n) 38.1 ± 11.1 22.7 ± 10.5 <0.001 A2 index (n) 4.0 ± 2.4 2.9 ± 2.2 NS A3 index (n) 10.1 ± 4.5 9.4 ± 5.7 NS Duration A (seconds) 9.8 ± 1.0 11.0 ± 2.0 0.019 Duration B (seconds) 21.2 ± 1.6 24.0 ± 3.0 <0.001 Controls (n = 14) IRBD (n = 67) P value CAP rate (%) 42.7 ± 7.5 30.0 ± 10.3 <0.001 CAP index (n) 52.1 ± 9.7 39.6 ± 24.6 <0.001 A1 index (n) 38.1 ± 11.1 22.7 ± 10.5 <0.001 A2 index (n) 4.0 ± 2.4 2.9 ± 2.2 NS A3 index (n) 10.1 ± 4.5 9.4 ± 5.7 NS Duration A (seconds) 9.8 ± 1.0 11.0 ± 2.0 0.019 Duration B (seconds) 21.2 ± 1.6 24.0 ± 3.0 <0.001 CAP: cyclic alternating pattern. Data are presented as mean ± standard deviation; NS: no significant. Open in new tab Table 3. CAP parameters in controls and patients with IRBD Controls (n = 14) IRBD (n = 67) P value CAP rate (%) 42.7 ± 7.5 30.0 ± 10.3 <0.001 CAP index (n) 52.1 ± 9.7 39.6 ± 24.6 <0.001 A1 index (n) 38.1 ± 11.1 22.7 ± 10.5 <0.001 A2 index (n) 4.0 ± 2.4 2.9 ± 2.2 NS A3 index (n) 10.1 ± 4.5 9.4 ± 5.7 NS Duration A (seconds) 9.8 ± 1.0 11.0 ± 2.0 0.019 Duration B (seconds) 21.2 ± 1.6 24.0 ± 3.0 <0.001 Controls (n = 14) IRBD (n = 67) P value CAP rate (%) 42.7 ± 7.5 30.0 ± 10.3 <0.001 CAP index (n) 52.1 ± 9.7 39.6 ± 24.6 <0.001 A1 index (n) 38.1 ± 11.1 22.7 ± 10.5 <0.001 A2 index (n) 4.0 ± 2.4 2.9 ± 2.2 NS A3 index (n) 10.1 ± 4.5 9.4 ± 5.7 NS Duration A (seconds) 9.8 ± 1.0 11.0 ± 2.0 0.019 Duration B (seconds) 21.2 ± 1.6 24.0 ± 3.0 <0.001 CAP: cyclic alternating pattern. Data are presented as mean ± standard deviation; NS: no significant. Open in new tab Table 4 details the three-group microstructural data. The cRBD group had lower values of CAP rate and CAP index when compared with the ncRBD group and the control group. Compared to controls, a reduction of the slow components of CAP (A1 index) was observed in both ncRBD and cRBD groups, but the difference between converters and non-converters was not significant. Phase B duration showed similar findings. When compared to the ncRBD group, subtype A3 index was decreased in the cRBD group (p < 0.010), while subtype A2 index differed only between converters and controls (p < 0.05). Table 4. Comparisons in CAP parameters between controls and in patients with IRBD who remained disease-free and converted to a neurodegenerative disorder Control (n = 14) ncRBD (n = 33) cRBD (n = 34) P value Post hoc P value [1] [2] [3] [1] versus [2] [1] versus [3] [2] versus [3] CAP rate (%) 42.7 ± 7.5 33.7 ± 9.4 26.5 ± 9.9 <0.001 <0.05 <0.0001 <0.05 CAP index (n) 52.1 ± 9.7 39.2 ± 11.9 31.0 ± 10.8 <0.001 <0.010 0.0001 <0.05 A1 index (n) 38.1 ± 11.1 24.1 ± 12.0 21.4 ± 8.7 <0.001 <0.010 <0.0001 NS A2 index (n) 4.0 ± 2.4 3.4 ± 2.1 2.4 ± 2.1 0.016 NS <0.05 NS A3 index (n) 10.1 ± 4.5 11.7 ± 6.0 7.1 ± 4.4 0.004 NS NS <0.010 Duration A (seconds) 9.8 ± 1.0 11.0 ± 1.8 11.0 ± 2.2 NS NS NS NS Duration B (seconds) 21.2 ± 1.6 23.7 ± 2.9 24.4 ± 3.1 <0.001 <0.05 <0.001 NS Control (n = 14) ncRBD (n = 33) cRBD (n = 34) P value Post hoc P value [1] [2] [3] [1] versus [2] [1] versus [3] [2] versus [3] CAP rate (%) 42.7 ± 7.5 33.7 ± 9.4 26.5 ± 9.9 <0.001 <0.05 <0.0001 <0.05 CAP index (n) 52.1 ± 9.7 39.2 ± 11.9 31.0 ± 10.8 <0.001 <0.010 0.0001 <0.05 A1 index (n) 38.1 ± 11.1 24.1 ± 12.0 21.4 ± 8.7 <0.001 <0.010 <0.0001 NS A2 index (n) 4.0 ± 2.4 3.4 ± 2.1 2.4 ± 2.1 0.016 NS <0.05 NS A3 index (n) 10.1 ± 4.5 11.7 ± 6.0 7.1 ± 4.4 0.004 NS NS <0.010 Duration A (seconds) 9.8 ± 1.0 11.0 ± 1.8 11.0 ± 2.2 NS NS NS NS Duration B (seconds) 21.2 ± 1.6 23.7 ± 2.9 24.4 ± 3.1 <0.001 <0.05 <0.001 NS ncRBD: non-converters REM sleep behavior disorder; cRBD: converters REM sleep behavior disorder; CAP: cyclic alternating pattern. Data are presented as mean ± standard deviation; NS: no significant. Open in new tab Table 4. Comparisons in CAP parameters between controls and in patients with IRBD who remained disease-free and converted to a neurodegenerative disorder Control (n = 14) ncRBD (n = 33) cRBD (n = 34) P value Post hoc P value [1] [2] [3] [1] versus [2] [1] versus [3] [2] versus [3] CAP rate (%) 42.7 ± 7.5 33.7 ± 9.4 26.5 ± 9.9 <0.001 <0.05 <0.0001 <0.05 CAP index (n) 52.1 ± 9.7 39.2 ± 11.9 31.0 ± 10.8 <0.001 <0.010 0.0001 <0.05 A1 index (n) 38.1 ± 11.1 24.1 ± 12.0 21.4 ± 8.7 <0.001 <0.010 <0.0001 NS A2 index (n) 4.0 ± 2.4 3.4 ± 2.1 2.4 ± 2.1 0.016 NS <0.05 NS A3 index (n) 10.1 ± 4.5 11.7 ± 6.0 7.1 ± 4.4 0.004 NS NS <0.010 Duration A (seconds) 9.8 ± 1.0 11.0 ± 1.8 11.0 ± 2.2 NS NS NS NS Duration B (seconds) 21.2 ± 1.6 23.7 ± 2.9 24.4 ± 3.1 <0.001 <0.05 <0.001 NS Control (n = 14) ncRBD (n = 33) cRBD (n = 34) P value Post hoc P value [1] [2] [3] [1] versus [2] [1] versus [3] [2] versus [3] CAP rate (%) 42.7 ± 7.5 33.7 ± 9.4 26.5 ± 9.9 <0.001 <0.05 <0.0001 <0.05 CAP index (n) 52.1 ± 9.7 39.2 ± 11.9 31.0 ± 10.8 <0.001 <0.010 0.0001 <0.05 A1 index (n) 38.1 ± 11.1 24.1 ± 12.0 21.4 ± 8.7 <0.001 <0.010 <0.0001 NS A2 index (n) 4.0 ± 2.4 3.4 ± 2.1 2.4 ± 2.1 0.016 NS <0.05 NS A3 index (n) 10.1 ± 4.5 11.7 ± 6.0 7.1 ± 4.4 0.004 NS NS <0.010 Duration A (seconds) 9.8 ± 1.0 11.0 ± 1.8 11.0 ± 2.2 NS NS NS NS Duration B (seconds) 21.2 ± 1.6 23.7 ± 2.9 24.4 ± 3.1 <0.001 <0.05 <0.001 NS ncRBD: non-converters REM sleep behavior disorder; cRBD: converters REM sleep behavior disorder; CAP: cyclic alternating pattern. Data are presented as mean ± standard deviation; NS: no significant. Open in new tab No difference was found in the CAP rate between the 14 patients who developed PD and the 17 patients who manifested DLB (27.9 ± 11.0 versus 25.8 ± 9.5; p = 0.586). The Kaplan-Meier curve applied to cRBD distinguished two subgroups based on the median cutoff value of CAP rate (32.9%). Below this CAP value, the average risk of conversion was 9.2 years after baseline diagnostic PSG. Patients with a CAP rate value >32.9% converted after a mean of 12.6 years (Figure 1). The p value of Log-rank Mantel-Cox was 0.037. Figure 1. Open in new tabDownload slide Kaplan-Meier curve in IRBD patients with a CAP rate values lower and higher than 32.9%. Below this value the average risk of conversion was 9.2 years after baseline diagnostic PSG. Patients with a CAP rate values greater than 32.9% underwent conversion after a mean of 12.6 years. Figure 1. Open in new tabDownload slide Kaplan-Meier curve in IRBD patients with a CAP rate values lower and higher than 32.9%. Below this value the average risk of conversion was 9.2 years after baseline diagnostic PSG. Patients with a CAP rate values greater than 32.9% underwent conversion after a mean of 12.6 years. Discussion The present study showed that CAP rate in IRBD is significantly reduced and this feature was mostly pronounced in those patients who converted to a defined neurodegenerative syndrome. In particular, the lowest values of CAP rate were found in earlier RBD converters. To avoid potential biases, IRBD recordings were free from psychoactive medication and were not affected by other sleep pathologies, in particular periodic limb movements and untreated sleep-related breathing disorders. The CAP system acts as an adaptive mechanism in the maintenance of the arousal balance and therefore represents a topical guardian of sleep resilience [34]. Sleep architecture and the shifts between sleep and wakefulness are controlled by switching mechanisms regulated by several nuclei mainly located in the brainstem, midbrain areas, hypothalamus, and motor cortex [35–39]. Proposed models describe the transitions between sleep and wakefulness and those between REM sleep and non-REM sleep as flip-flop switches [40, 41]. These switches are mutually dependent and determine the wake–sleep cyclic rhythm. Instability is a basic feature of all complex systems. Within certain ranges, instability supports the flow of biological variability and warrants a flexible and adaptive behavior. Moreover, CAP maintains both EEG and vegetative functions in-phase through regular fluctuations [42]. Besides the acknowledged motor dysfunction of REM sleep, our IRBD patients showed alterations also in non-REM sleep microstructure. Compared to controls, IRBD patients showed longer CAP cycles, which corresponds to a non-physiological feature as CAP cycles represent time-constant structures of normal non-REM sleep [43]. Despite the increased length of CAP cycles (supported by both more extended phases A and B), sleep recordings in IRBD patients showed impaired arousal modulation (lower CAP rate values) accompanied by a decrease of A1 subtypes (A1 index). Other non-REM sleep alterations have also been described in IRBD. Regarding microstructure, one study described increased slow transient EEG events and decreased fast transient events [27]. Interestingly, non-REM sleep instability remains statistically unmodified even after long-term use of clonazepam in patients with IRBD. Some studies have shown increased N3 sleep percentage [26] while others have shown normal amount [26, 44, 45]. CAP is endowed within the physiological texture of the sleep process [20] and reflects the interaction between the REM-off and REM-on mechanisms [21]. Accordingly, CAP sequences commonly precede the transition from non-REM to REM sleep [19] and alterations of CAP parameters characterize both non-REM and REM sleep disorders. Drops of CAP rate and A1 subtypes have been reported in narcolepsy [46], MSA [23], MCI and Alzheimer disease [24]. Of note, narcolepsy and MSA can be associated with RBD [25]. Non-REM sleep microstructural dysfunction was found in both RBD converters and non-converters. However, while the majority of macrostructural sleep parameters did not differ between ncRBD and cRBD groups, significant findings characterized most CAP parameters. The loss of A1 phases, observed in both RBD groups, could be directly linked to an impairment of non-REM sleep regulation mechanisms and GABAergic systems [47], while the reduction of A3, detected only in cRBD could indicate an involvement of the REM-on cholinergic arousal system [21], a dysfunction which is often described in PD and DLB [48, 49]. In the absence of relevant sleep macrostructural differences between ncRBD and cRBD, our data suggest an important role in the maintenance of the slow oscillations, that is, the A1 phases CAP, which actively participate in the buildup and consolidation of slow-wave sleep [21, 50]. These non-REM sleep microstructural alterations could be associated with the Braak staging system in PD which postulates that pathological abnormalities start in the medulla (and olfactory structures) followed by a gradual ascension to more rostral structures [51, 52]. In this perspective, low amounts of CAP rate and A1 subtypes already in the initial phases of the disease could reflect the anatomical changes of Braak stage 2 (pontine damage), while later sleep microstructure alterations (loss of A3 subtype) could be associated to Braak stage 3 (midbrain damage). In cRBD the neurodegenerative process could also involve other areas including brainstem pre-thalamic fibers with alterations in EEG rhythms or elements of thalamic origin such as K-complexes, sleep spindles and alpha-activity in the occipital cortex [53, 54]. Overall, the attenuated EEG reactivity in sleep, more evident in cRBD, is in line with pathological involvement of arousal generating structures in PD and DLB [48, 49]. Potential signs of neurodegeneration in IRBD have been investigated in several studies [6–18, 55–59]. The present study indicates that CAP parameters can be exploited as a marker of conversion in IRBD. Values of CAP rate <32.9% are predictive of a manifest synucleinopathy within 9 years from IRBD diagnosis whereas CAP values > 32.9% are related to a later onset of conversion to PD and DLB (median of 12.8 years). The enhanced alterations of CAP metrics in cRBD likely reflect a major remodeling of structural networks [60] and a more disrupted functional connectivity during sleep [61]. Some limitations of our study have to be acknowledged. They include a relatively small sample size, a lower number of controls compared with the two IRBD groups, and the lack of neuropathological data confirming the clinical diagnosis of PD and DLB. Another limitation is that at baseline (when the PSG was performed) most of the patients were not tested to detect other markers of neurodegeneration such as abnormal neuropsychology tests, hyposmia, constipation, depression, and neuroimaging abnormalities. Strengths of our study include the diagnosis of RBD based on both clinical and PSG data, and the total blindness of CAP scoring to the subject condition (patient or control). In summary, our data indicate that IRBD is not exclusively a REM sleep parasomnia, as non-REM sleep is also affected at the microstructural level. Further studies are necessary to confirm that global impairment of arousal mechanisms expressed by a significant reduction of specific CAP metrics (CAP rate and phase A subtypes) can be a marker of neurodegeneration in IRBD. Funding None declared. Conflict of interest statement None declared. References 1. American Academy of Sleep Medicine . International Classification of Sleep Disorders . 3rd ed. Darien, IL : American Academy of Sleep Medicine ; 2014 . WorldCat COPAC 2. Iranzo A , 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 3. Postuma RB , et al. Quantifying the risk of neurodegenerative disease in idiopathic REM sleep behavior disorder . Neurology. 2009 ; 72 ( 15 ): 1296 – 1300 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Schenck CH , et al. 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 5. Iranzo A , et al. Rapid-eye-movement sleep behaviour disorder as an early marker for a neurodegenerative disorder: a descriptive study . Lancet Neurol. 2006 ; 5 ( 7 ): 572 – 577 . Google Scholar Crossref Search ADS PubMed WorldCat 6. dos Santos AB , et al. Are sleep disturbances preclinical markers of Parkinson’s disease? Neurochem Res. 2015 ; 40 ( 3 ): 421 – 427 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Postuma RB , et al. Olfaction and color vision identify impending neurodegeneration in rapid eye movement sleep behavior disorder . Ann Neurol. 2011 ; 69 ( 5 ): 811 – 818 . Google Scholar Crossref Search ADS PubMed WorldCat 8. Iranzo A , et al. ; Sleep Innsbruck Barcelona (SINBAR) group . Decreased striatal dopamine transporter uptake and substantia nigra hyperechogenicity as risk markers of synucleinopathy in patients with idiopathic rapid-eye-movement sleep behaviour disorder: a prospective study [corrected] . Lancet Neurol. 2010 ; 9 ( 11 ): 1070 – 1077 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Fantini ML , 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 10. Latreille V , et al. Sleep spindles in Parkinson’s disease may predict the development of dementia . Neurobiol Aging. 2015 ; 36 ( 2 ): 1083 – 1090 . Google Scholar Crossref Search ADS PubMed WorldCat 11. Christensen JA , et al. Decreased sleep spindle density in patients with idiopathic REM sleep behavior disorder and patients with Parkinson’s disease . Clin Neurophysiol. 2014 ; 125 ( 3 ): 512 – 519 . Google Scholar Crossref Search ADS PubMed WorldCat 12. Christensen JA , et al. Data-driven modeling of sleep EEG and EOG reveals characteristics indicative of pre-Parkinson’s and Parkinson’s disease . J Neurosci Methods. 2014 ; 235 : 262 – 276 . Google Scholar Crossref Search ADS PubMed WorldCat 13. Postuma RB , et al. Severity of REM atonia loss in idiopathic REM sleep behavior disorder predicts Parkinson disease . Neurology. 2010 ; 74 ( 3 ): 239 – 244 . Google Scholar Crossref Search ADS PubMed WorldCat 14. Sorensen GL , et al. Reduced sympathetic activity in idiopathic rapid-eye-movement sleep behavior disorder and Parkinson’s disease . Auton Neurosci. 2013 ; 179 ( 1-2 ): 138 – 141 . Google Scholar Crossref Search ADS PubMed WorldCat 15. Garcia-Ruiz PJ , et al. Non-motor symptoms of Parkinson’s disease A review…from the past . J Neurol Sci. 2014 ; 338 ( 1-2 ): 30 – 33 . Google Scholar Crossref Search ADS PubMed WorldCat 16. Scherfler C , et al. ; SINBAR (Sleep Innsbruck Barcelona) Group . White and gray matter abnormalities in idiopathic rapid eye movement sleep behavior disorder: a diffusion-tensor imaging and voxel-based morphometry study . Ann Neurol. 2011 ; 69 ( 2 ): 400 – 407 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Gagnon JF , et al. Cognition in rapid eye movement sleep behavior disorder . Front Neurol. 2012 ; 3 : 82 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Manni R , et al. Cognitive dysfunction and REM sleep behavior disorder: key findings in the literature and preliminary longitudinal findings . Int J Psychophysiol. 2013 ; 89 ( 2 ): 213 – 217 . Google Scholar Crossref Search ADS PubMed WorldCat 19. Parrino L , et al. Cyclic alternating pattern (CAP): the marker of sleep instability . Sleep Med Rev. 2012 ; 16 ( 1 ): 27 – 45 . Google Scholar Crossref Search ADS PubMed WorldCat 20. Halász P , et al. Two features of sleep slow waves: homeostatic and reactive aspects–from long term to instant sleep homeostasis . Sleep Med. 2014 ; 15 ( 10 ): 1184 – 1195 . Google Scholar Crossref Search ADS PubMed WorldCat 21. Terzano MG , et al. CAP and arousals are involved in the homeostatic and ultradian sleep processes . J Sleep Res. 2005 ; 14 ( 4 ): 359 – 368 . Google Scholar Crossref Search ADS PubMed WorldCat 22. Terzano MG , et al. Atlas, rules, and recording techniques for the scoring of cyclic alternating pattern (CAP) in human sleep . Sleep Med. 2002 ; 3 ( 2 ): 187 – 199 . Google Scholar Crossref Search ADS PubMed WorldCat 23. Vetrugno R , et al. Impaired cortical and autonomic arousal during sleep in multiple system atrophy . Clin Neurophysiol. 2007 ; 118 ( 11 ): 2512 – 2518 . Google Scholar Crossref Search ADS PubMed WorldCat 24. Maestri M , et al. Non-rapid eye movement sleep instability in mild cognitive impairment: a pilot study . Sleep Med. 2015 ; 16 ( 9 ): 1139 – 1145 . Google Scholar Crossref Search ADS PubMed WorldCat 25. Boeve BF , et al. Pathophysiology of REM sleep behaviour disorder and relevance to neurodegenerative disease . Brain. 2007 ; 130 ( Pt 11 ): 2770 – 2788 . Google Scholar Crossref Search ADS PubMed WorldCat 26. Schenck CH , et al. 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 27. Ferri R , et al. 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 28. Hughes AJ , et al. Accuracy of clinical diagnosis of idiopathic Parkinson’s disease: a clinico-pathological study of 100 cases . J Neurol Neurosurg Psychiatry. 1992 ; 55 ( 3 ): 181 – 184 . Google Scholar Crossref Search ADS PubMed WorldCat 29. McKeith IG , et al. ; Consortium on DLB . Diagnosis and management of dementia with Lewy bodies: third report of the DLB Consortium . Neurology. 2005 ; 65 ( 12 ): 1863 – 1872 . Google Scholar Crossref Search ADS PubMed WorldCat 30. Gilman S , et al. Second consensus statement on the diagnosis of multiple system atrophy . Neurology. 2008 ; 71 ( 9 ): 670 – 676 . Google Scholar Crossref Search ADS PubMed WorldCat 31. Petersen RC . Mild cognitive impairment as a diagnostic entity . J Intern Med. 2004 ; 256 ( 3 ): 183 – 194 . Google Scholar Crossref Search ADS PubMed WorldCat 32. Iber C , et al. The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications . 1st ed. Westchester, IL : American Academy of Sleep Medicine ; 2007 . Google Preview WorldCat COPAC 33. Zucconi M , 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 34. Parrino L , et al. The resilient brain and the guardians of sleep: new perspectives on old assumptions . Sleep Med Rev. 2018 ; 39 : 98 – 107 . Google Scholar Crossref Search ADS PubMed WorldCat 35. Hayashi Y , et al. Cells of a common developmental origin regulate REM/non-REM sleep and wakefulness in mice . Science. 2015 ; 350 ( 6263 ): 957 – 961 . Google Scholar Crossref Search ADS PubMed WorldCat 36. Niwa Y , et al. Muscarinic Acetylcholine Receptors Chrm1 and Chrm3 Are Essential for REM Sleep . Cell Rep. 2018 ; 24 ( 9 ): 2231 – 2247.e7 . Google Scholar Crossref Search ADS PubMed WorldCat 37. Saper CB , et al. Sleep state switching . Neuron. 2010 ; 68 ( 6 ): 1023 – 1042 . Google Scholar Crossref Search ADS PubMed WorldCat 38. Schwartz JR , et al. Neurophysiology of sleep and wakefulness: basic science and clinical implications . Curr Neuropharmacol. 2008 ; 6 ( 4 ): 367 – 378 . Google Scholar Crossref Search ADS PubMed WorldCat 39. Peever J , et al. Breakdown in REM sleep circuitry underlies REM sleep behavior disorder . Trends Neurosci. 2014 ; 37 ( 5 ): 279 – 288 . Google Scholar Crossref Search ADS PubMed WorldCat 40. Saper CB , et al. The sleep switch: hypothalamic control of sleep and wakefulness . Trends Neurosci. 2001 ; 24 ( 12 ): 726 – 731 . Google Scholar Crossref Search ADS PubMed WorldCat 41. Lu J , et al. A putative flip-flop switch for control of REM sleep . Nature. 2006 ; 441 ( 7093 ): 589 – 594 . Google Scholar Crossref Search ADS PubMed WorldCat 42. Parrino L , et al. The cyclic alternating pattern and the brain body coupling during sleep . Epileptologie . 2016 ; 33 : 150 – 60 . WorldCat 43. Smerieri A , et al. Cyclic alternating pattern sequences and non-cyclic alternating pattern periods in human sleep . Clin Neurophysiol. 2007 ; 118 ( 10 ): 2305 – 2313 . Google Scholar Crossref Search ADS PubMed WorldCat 44. Fantini ML , et al. Longitudinal study of cognitive function in idiopathic REM sleep behavior disorder . Sleep. 2011 ; 34 ( 5 ): 619 – 625 . Google Scholar PubMed WorldCat 45. Iranzo A , et al. Brainstem proton magnetic resonance spectroscopy in idopathic REM sleep behavior disorder . Sleep. 2002 ; 25 ( 8 ): 867 – 870 . Google Scholar Crossref Search ADS PubMed WorldCat 46. Terzano MG , et al. Cyclic alternating pattern (CAP) alterations in narcolepsy . Sleep Med. 2006 ; 7 ( 8 ): 619 – 626 . Google Scholar Crossref Search ADS PubMed WorldCat 47. Vanini G , et al. GABA-to-ACh ratio in basal forebrain and cerebral cortex varies significantly during sleep . Sleep. 2012 ; 35 ( 10 ): 1325 – 1334 . Google Scholar Crossref Search ADS PubMed WorldCat 48. Nakamura S , et al. Large neurons in the tuberomammillary nucleus in patients with Parkinson’s disease and multiple system atrophy . Neurology. 1996 ; 46 ( 6 ): 1693 – 1696 . Google Scholar Crossref Search ADS PubMed WorldCat 49. Benarroch EE . Brainstem in multiple system atrophy: clinicopathological correlations . Cell Mol Neurobiol. 2003 ; 23 ( 4-5 ): 519 – 526 . Google Scholar Crossref Search ADS PubMed WorldCat 50. Terzano MG , et al. CAP components and EEG synchronization in the first 3 sleep cycles . Clin Neurophysiol. 2000 ; 111 ( 2 ): 283 – 290 . Google Scholar Crossref Search ADS PubMed WorldCat 51. Braak H , et al. Staging of brain pathology related to sporadic Parkinson’s disease . Neurobiol Aging. 2003 ; 24 ( 2 ): 197 – 211 . Google Scholar Crossref Search ADS PubMed WorldCat 52. Braak H , et al. Stages in the development of Parkinson’s disease-related pathology . Cell Tissue Res. 2004 ; 318 ( 1 ): 121 – 134 . Google Scholar Crossref Search ADS PubMed WorldCat 53. Peter-Derex L , et al. Heterogeneity of arousals in human sleep: a stereo-electroencephalographic study . Neuroimage. 2015 ; 123 : 229 – 244 . Google Scholar Crossref Search ADS PubMed WorldCat 54. Halász P . The K-complex as a special reactive sleep slow wave - A theoretical update . Sleep Med Rev. 2016 ; 29 : 34 – 40 . Google Scholar Crossref Search ADS PubMed WorldCat 55. Postuma RB , et al. Potential early markers of Parkinson disease in idiopathic REM sleep behavior disorder . Neurology. 2006 ; 66 ( 6 ): 845 – 851 . Google Scholar Crossref Search ADS PubMed WorldCat 56. Christensen JAE , 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 57. Rodrigues Brazète J , et al. Electroencephalogram slowing predicts neurodegeneration in rapid eye movement sleep behavior disorder . Neurobiol Aging. 2016 ; 37 : 74 – 81 . Google Scholar Crossref Search ADS PubMed WorldCat 58. Iranzo A , 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 59. Massicotte-Marquez J , 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 60. Guo T , et al. Alterations of brain structural network in Parkinson’s disease with and without rapid eye movement sleep behavior disorder . Front Neurol. 2018 ; 9 : 334 . Google Scholar Crossref Search ADS PubMed WorldCat 61. Ferri R , et al. The functional connectivity of different EEG bands moves towards small-world network organization during sleep . Clin Neurophysiol. 2008 ; 119 ( 9 ): 2026 – 2036 . Google Scholar Crossref Search ADS PubMed WorldCat © Sleep Research Society 2019. Published by Oxford University Press on behalf of the Sleep Research Society. All rights reserved. For permissions, please e-mail [email protected]. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Ong, Jason, C

doi: 10.1093/sleep/zsz172pmid: 31595967

Sleep, exercise, and diet are well-known lifestyle factors associated with health and well-being. In their letter to the editor, Lamp and colleagues [1] propose that waking rest should be a new lifestyle factor considered alongside these traditional lifestyle factors. They define waking rest as a period of quiet, reflective thought that is void of effortful, focused thought, and distracting stimuli (e.g., watching television or using social media). It can include reflection on past experiences or thoughts about the future. The waking rest period can last from 5 to 20 min and can occur at multiple times throughout the day or night. Lamp and colleagues further argue that waking rest can serve to facilitate mental rejuvenation and emotion regulation, thus maintaining proper mental health. This proposal challenges established paradigms of sleep and health. Typically, sleep and wake are considered dichotomous states with specific biological and behavioral indicators associated with each state. One of the main benefits of sleep is rest and recovery, which is important for optimizing mental and physical health during wakefulness. However, there are some known situations or conditions that challenge this dichotomy and could fit the definition of waking rest. For example, paradoxical insomnia (previously known as sleep-state misperception) is a subtype of insomnia where the individual complains of little or no sleep with racing thoughts or difficulty shutting off the mind in bed. However, objective measures of sleep using actigraphy or polysomnography reveal normal or near-normal amounts of sleep. Interestingly, the degree of daytime impairment is usually minimal and inconsistent with the self-reported amount of sleep. Could this be an example of waking rest? Lucid dreaming is another condition where there is an overlap between intentional, conscious thought and the sleep state. During lucid dreaming, there is conscious awareness of dreaming while the individual is still asleep and individuals describe an ability to control their dreams. Voss and colleagues [2] conducted a laboratory experiment with up to five nights of electroencephalogram (EEG) recording and found evidence for specific EEG patterns in the frontal regions that were distinct from wakefulness and rapid eye movement sleep. Could this be another example of waking rest? Lamp and colleagues also discuss the rise of mindfulness-based interventions as support for the need to consider the concept of waking rest. Mindfulness meditation is similar to the concept of waking rest in that the period of meditation involves openness and nonjudging of thoughts as they arise without attachment to outcomes. In a recent report from our lab, we found that participants who completed an 8 week mindfulness intervention showed increases in beta EEG from baseline to 6 month follow-up [3]. During the same period, Insomnia Severity Index scores decreased significantly but objectively measured total sleep time did not increase significantly [4]. An argument could be made that these individuals were in a state of waking rest, which might have reduced the perceived severity of their insomnia symptoms. Although these examples can be interpreted as waking rest, there has been very little research directly investigating this concept that would distinguish it from other states. Therefore, Lamp and colleagues propose a call to action for more research to explore the concept of waking rest and evaluate the benefits on well-being. If we are to embark on this journey, several important issues will need to be investigated. An important first step would be to establish the operational definition of waking rest. Is there an EEG signature that can clearly distinguish waking rest from sleep and active wakefulness? Can this be validated with a test or some gold standard? Are there behavioral criteria that also need to be considered and how would those be measured? How would this state be differentiated from other activities such as meditation or daydreaming? If empirically supported parameters of waking rest can be defined, then research could examine clinical implications. For example, Lamp and colleagues suggest that waking rest could serve as an alternative for stimulus control. However, this would need to be tested with specific ways to measure adherence to each strategy. Lamp and colleagues also suggest that waking rest could have benefits on mental health broadly, but studies would need to examine more specific aspects of mental health (e.g., symptoms of depression or anxiety and memory consolidation) and specific neurocognitive mechanisms underlying this relationship. Finally, it would need to be determined if waking rest is a necessary component for health and well-being or if it should be used as an intervention to improve well-being, similar to mindfulness meditation. It has been posited that sleep health is a multidimensional concept that emphasizes the positive role of sleep in overall health [5]. Using a similar framework, a period of waking rest might serve as a dimension of overall health that is needed for emotion regulation. To this end, further research is needed to see if waking rest could serve as a new lifestyle factor or if it is just another name for a form of mental relaxation such as meditation. Funding None. Conflict of interest statement. None declared. References 1. Lamp A , et al. Exercise, nutrition, sleep, and waking rest? . Sleep. 2019 ; 42 ( 10 ): 1 – 2 . doi: https://doi.org/10.1093/sleep/zsz138 . WorldCat 2. Voss U , et al. Lucid dreaming: a state of consciousness with features of both waking and non-lucid dreaming . Sleep. 2009 ; 32 ( 9 ): 1191 – 1200 . doi: https://doi.org/10.1093/sleep/32.9.1191 . Google Scholar Crossref Search ADS PubMed WorldCat 3. Goldstein MR , et al. . Increased high-frequency NREM EEG power associated with mindfulness-based interventions for chronic insomnia: preliminary findings from spectral analysis . J Psychosom Res. 2019 ; 120 : 12 – 19 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Ong JC , et al. . A randomized controlled trial of mindfulness meditation for chronic insomnia . Sleep. 2014 ; 37 ( 9 ): 1553 – 1563 . doi: https://doi.org/10.5665/sleep.4010 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Buysse DJ . Sleep health: can we define it? Does it matter? Sleep. 2014 ; 37 ( 1 ): 9 – 17 . doi: https://doi.org/10.5665/sleep.3298 . Google Scholar Crossref Search ADS PubMed WorldCat © Sleep Research Society 2019. Published by Oxford University Press on behalf of the Sleep Research Society. All rights reserved. For permissions, please e-mail [email protected]. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Komarzynski,, Sandra;Huang,, Qi;Lévi, Francis, A;Palesh, Oxana, G;Ulusakarya,, Ayhan;Bouchahda,, Mohamed;Haydar,, Mazen;Wreglesworth, Nicholas, I;Morère,, Jean-François;Adam,, René;Innominato, Pasquale, F

doi:

Wynchank,, Dora;Bijlenga,, Denise;Penninx, Brenda, W;Lamers,, Femke;Beekman, Aartjan, T;Kooij, J J, Sandra;Verhoeven, Josine, E

doi: 10.1093/sleep/zsz139pmid: 31270544

Abstract Study Objectives We evaluated the relationship between leukocyte telomere length (LTL) and sleep duration, insomnia symptoms, and circadian rhythm, to test whether sleep and chronobiological dysregulations are associated with cellular aging. Methods Data from the Netherlands Study of Depression and Anxiety (N = 2,936) were used at two waves 6 years apart, to measure LTL. Telomeres shorten during the life span and are important biomarkers for cellular aging. LTL was assessed by qualitative polymerase chain reaction and converted into base pair number. Sleep parameters were: sleep duration and insomnia symptoms from the Insomnia Rating Scale. Circadian rhythm variables were: indication of Delayed Sleep Phase Syndrome (DSPS), mid-sleep corrected for sleep debt on free days (MSFsc), sleep-onset time, and self-reported chronotype, from the Munich Chronotype Questionnaire. Generalized estimating equations analyzed the associations between LTL, sleep, and chronobiological factors, adjusted for baseline age, sex, North European ancestry, and additionally for current smoking, depression severity, obesity, and childhood trauma. Results Indicators of delayed circadian rhythm showed a strong and consistent effect on LTL, after adjustment for sociodemographic and health indicators. Late MSFsc (B = −49.9, p = .004), late sleep-onset time (B = −32.4, p = .001), indication of DSPS (B = −73.8, p = .036), and moderately late chronotype in adulthood (B = −71.6, p = .003) were associated with significantly shorter LTL across both waves; whereas sleep duration and insomnia symptoms were not. Extremely early chronotype showed significantly less LTL shortening than intermediate chronotype (B = 161.40, p = .037). No predictors showed accelerated LTL attrition over 6 years. Conclusions Individuals with delayed circadian rhythm have significantly shorter LTL, but not faster LTL attrition rates. aging, leukocyte telomere length, delayed sleep phase, circadian rhythms, insomnia Statement of Significance Late sleep-onset and insomnia symptoms are common. We studied the impact of sleep duration, insomnia symptoms, and circadian rhythm dysregulation on cellular aging by measuring leukocyte telomere length (LTL) in 2,936 adults, twice over 6 years. For the first time, we show an association between delayed circadian rhythm and shorter LTL. Persons with a late self-reported chronotype and indicators of delayed circadian rhythm had significantly shorter LTL (but not faster LTL attrition), after adjustment for sociodemographic and health indicators. Those with delayed circadian rhythm may thus be at risk for significant morbidity. Insomnia symptoms and sleep duration were not associated with shorter LTL. Extended longitudinal studies should focus on the lifetime risks associated with circadian rhythm disturbance and cellular aging. Introduction Telomeres are stretches of DNA, situated at the ends of chromosomes, which cap and protect chromosomes during successive cellular divisions [1], preventing end-to-end fusion, genomic instability, and base pair loss of chromosomal DNA [2]. As humans age, telomeres shorten until they become too short for further cellular division, resulting in cellular senescence [3]. Critically short telomeres may also lead to carcinogenic transformation [4]. Leukocyte telomere length (LTL), expressed as number of base pairs, has been suggested as a potential biomarker for cellular aging [3]. Telomere length and rate of shortening with aging (attrition rate) vary enormously among individuals, and even between chromosomes [5, 6]. Genetic factors contribute to approximately 64%–70% of the variability in LTL [7] and other factors believed to play a role are lifestyle and illness [3]. Cross-sectional studies have shown that shorter LTL is associated with older chronological age [5, 6, 8], male sex [9], non-North European ancestry [10], lifestyle factors such as heavy alcohol use [11], less physical activity [10, 12], psychiatric factors [13] including affective disorders, and somatic illnesses [3, 14, 15]. In terms of the telomere attrition rate, there appears to be a relationship between LTL attrition and chronological age: in the first year of life, telomere loss is very rapid [16]; it stabilizes from childhood to early adulthood; and begins to decrease in older adulthood [5, 8]. Baseline LTL appears to be an important determinant [10], as are male sex [17], greater age [8], and lifestyle factors such as cigarette smoking [12, 17] and less exercise [12]. Other determinants of accelerated LTL attrition include genetic heritability for cellular decline [7] and the presence of childhood trauma [10]. Adequate sleep is an essential requirement for health [18], yet exactly what constitutes adequate sleep duration varies within and between individuals [19]. For adults, optimal sleep duration is 7 or 8 h, as both shorter and longer sleep durations have been associated with significant health risks. A recent systematic review and dose-response meta-analysis showed significant U-shaped relationships between sleep duration and all-cause mortality as well as cardiovascular events [20]. Sleep of 7–8 h was associated with the lowest cardiometabolic risk score in another study [21]. Several recent studies have shown significant U-shaped relationships between sleep duration and metabolic syndrome and its risk factors [22, 23], all-cause and cause-specific mortality in women [24], and moderate diabetic retinopathy in older adults (mean age 64 years) [25]. Several quality of life measures (including physical, mental, emotional, and social functioning) also showed U-shaped relationships with sleep duration in both adults with chronic disease (N = 277,757) and those without (N = 172,052), after adjustment for sociodemographics and health-risk behaviors [26]. Sleep duration ≥9 h was partially associated with higher stroke prevalence [23], obesity [27], and with systemic inflammation in women [28]. Short sleep duration (<7 h) was associated with obesity independent of diet and physical activity [29] and systemic inflammation in men [28]. The research on LTL and sleep factors such as sleep duration, insomnia, and obstructive sleep apnea is scant and results are mixed. Short sleep duration has been associated with shorter LTL in older but not in middle-aged adults [30] or middle-aged women [31], in women under 50 years [32], and in men [33]. Insomnia has also been associated with shorter LTL in people older than 70 years [34]. But long sleep duration (>9 h) has also been associated with shorter LTL [10], hence the association between LTL and sleep duration appears to be nonlinear, with both shorter and longer sleep duration associated with shorter LTL. Obstructive sleep apnea syndrome and snoring are associated with shorter LTL [35, 36]. However, these sleep factors have not been examined in terms of LTL attrition over time. Linked to sleep duration are the disorders of the circadian rhythm. The association between circadian dysregulation and LTL has not been studied in humans. In animal models however, circadian desynchronization has already been shown to trigger premature cellular aging [37] and has been suggested as a mechanism underlying telomere activity control, linking the chronobiological systems to the aging process [38]. The circadian rhythm is set by the master biological clock, situated in the suprachiasmatic nuclei in the brain. One’s chronotype refers to the behavioral manifestation of underlying circadian rhythms, as evidenced by regular rising and sleep-onset times on free days (i.e. without daytime obligations) [39]. People can have early, intermediate, or late chronotype [40]. People with a late chronotype may have the persistent Delayed Sleep Phase Syndrome (DSPS), in which there is a chronic (>3 months) pattern of delayed sleep-onset, difficulty getting to sleep at a desired earlier time, difficulty awaking in the morning, resulting in daytime sleepiness and impaired social and occupational functioning [41]. Where morning obligations exist, it may also result in chronic short sleep and sleep debt. Disturbed circadian rhythm may result in sleep and energy metabolism disturbance, cardiovascular disease [42], and increased carcinogenesis [43]. On a cellular level, disturbed circadian rhythm affects the oscillatory behavior of many blood metabolites, which fluctuate in accordance with the circadian rhythm [44]. We hypothesize that late sleep (indicating disturbed circadian rhythm) will be a significant predictor of both short LTL and accelerated LTL attrition. We investigated whether sleep duration, insomnia, and chronobiological variables were related to LTL and LTL attrition over a 6-year time period, in a large cohort of adults participating in an epidemiological study on anxiety and depression. Methods Participants Subjects participated in the Netherlands Study of Depression and Anxiety (NESDA), an ongoing longitudinal, naturalistic cohort study of 2,981 participants at wave 1, aged 18–65 years. Participants were recruited from different health care settings (community, primary, and specialized mental health care) and included a group without lifetime affective disorders (“controls”) and others with current and remitted affective disorders. A full description of NESDA has been reported elsewhere [45]. Ethical Review Boards of all participating centers approved the NESDA study protocol and it was carried out in accordance with the latest Declaration of Helsinki. All participants gave written informed consent at enrolment after the study procedures had been fully explained. Measures in this study were taken from the baseline (wave 1), 2-year (wave 3), 4-year (wave 4), and 6-year (wave 5) follow-up assessments of NESDA, where the response rates were 87.1% (N = 2,596), 81% (N = 2,414), and 75.7% (N = 2,256), respectively. Our initial sample was 98.5% (N = 2,936) of the wave 1 group, which included those who had LTL measured and their sleep assessed. At wave 5, 1,883 participants had LTL measured (83.4% of the sample). Those who had LTL measurement at waves 1 and 5 were significantly: older, less likely to be current smokers, had more years of education, less severe anxiety and depression symptoms, shorter baseline LTL, less short (<7 h) and long (>9 h) sleep duration, and fewer insomnia symptoms than those without wave 5 LTL measurement. Sleep measures Sleep duration was measured at waves 1, 3, 4, and 5 as part of a self-report questionnaire where participants were asked to estimate the average number of hours of sleep during the past 4 weeks. Answer options were: “10 or more hours,” “9 h,” “8 h,” “7 h,” “6 h,” “5 or less hours.” In descriptive analyses, the single variable sleep duration was subcategorized as ≤5 h, 6 h, 7 or 8 h, 9 h or ≥10 h per night. We defined chronic short sleep as a dichotomous variable with sleep duration ≤6 h across waves 1, 3, 4, and 5 compared to sleep of 7 or 8 h across these waves, allowing for a maximum of one missing value on these four waves. Those with sleep duration of 9 or 10 h were excluded from this variable. Insomnia symptoms were measured at waves 1, 3, 4, and 5 with the Women’s Health Initiative Insomnia Rating Scale (WHIIRS) [46] which consists of five questions addressing sleep in the last 4 weeks (trouble falling asleep, waking up during the night, early morning awakening, difficulty getting back to sleep after waking up, and sleep quality). Answers are on a 4-point scale, with a sum score ranging from 0 to 20. The WHIIRS has good test-retest reliability and has high convergent correlation with objective actigraphy sleep measures [47]. In our sample, the WHIIRS showed good internal validity (Cronbach’s α = 0.83). In all analyses, WHIIRS scores were dichotomized at the cutoff point of 9 or higher, which indicates clinically significant insomnia symptoms [46]. We also assessed the WHIIRS as a continuous variable and examined the five questions individually with regard to LTL as outcome. We defined chronic insomnia as a WHIIRS score of ≥9 across waves 1, 3, 4, and 5 compared to no insomnia across these waves and allowed for a maximum of one missing value on these four waves. We considered these chronic variables course variables. Chronotype Chronotype was assessed with the Munich Chronotype Questionnaire (MCTQ) [39] at wave 3, which we deemed acceptable as chronotype is considered to be relatively stable over the life span [48]. The MCTQ is a self-report measure with questions on sleep timing on nights before work and free days, both in adulthood and “as a child.” Subjects rated their chronotype on a 7-point Likert scale ranging from “extremely early” to “extremely late.” Parameters of circadian rhythm used were the continuous measures sleep-onset time on nights before free days, and the time of mid-sleep on free days corrected for sleep debt during the work week (MSFsc) [39]. For self-reported chronotype in adulthood, we created a 5-category variable, where slightly early, normal, and slightly late were grouped together, forming the middle, intermediate chronotype category. The other categories were extremely early, moderately early, moderately late, and extremely late. We defined “indication for DSPS” as the combination of an inability to fall asleep before 12.30 am on work days, and a sleep-onset latency of 30 min or more on work days or a self-rating of being an extremely late chronotype in childhood or adulthood, as reported on the MCTQ. We considered chronobiological predictors constant variables. To explore the at-risk group of chronotype with moderate to extremely late sleep onset, we analyzed LTL in three groups: the highest risk group with moderate to extremely late chronotype with short sleep ≤6 h, moderate or extremely late chronotype with sleep duration 7–8 h, and early/intermediate chronotype with sleep duration 7–8 h. Leukocyte telomere length LTL was measured at waves 1 and 5. An extensive description of LTL assessment in our study has been reported before [49]. In summary, DNA was prepared from fasting blood samples drawn in the morning and stored in a −20°C freezer. Subsequently, quantitative polymerase chain reaction (qPCR) was used to determine baseline and 6-year LTL at the laboratories of Telomere Diagnostics (Menlo Park, CA) and the University of California, San Francisco, in 2012 and 2014, respectively. The qPCR was adapted from the published original method by Cawthon [50]. Telomere sequence copy number (T) in each patient’s sample was compared to a single-copy gene copy number (S), relative to a reference sample, where the resulting T/S ratio is proportional to mean LTL [51]. As previously described in our study [49], T/S ratios were converted into number of base pairs (bp) with the following formula: bp = 3274 + 2413 × ((T/S − 0.0545)/1.16). To address possible systematic differences caused by different reference samples, the wave 5 T/S ratios were adjusted relative to the wave 1 samples by rerunning and comparing samples from wave 1 sample plates 2 years later (N = 226, up to eight samples from each of the wave 1 plates), together with wave 5 samples. On average, the T/S ratios of the wave 5 follow-up runs were at 76% of the T/S ratios of wave 1; consequently, the follow-up T/S ratios were divided by 0.76. DNA samples were de-identified, and the laboratories that performed the assays were blind to all other measurements, and thus samples for case patients and control subjects were randomly distributed over the plates [52]. This procedure was also used in a previous study on NESDA data [52]. Of the entire NESDA cohort, 2,936 subjects had LTL measurement at wave 1 and 1,883 at wave 5. Of these, 1,860 had complete LTL data at both time points. Covariates Sex, age, years of education, and ancestry were determined at the wave 1 interview. Smoking status was categorized as current or not current smoker. Alcohol use was defined as number of alcohol consumptions/week. The majority of published studies have found statistically significant inverse associations between body mass index (BMI) and LTL [53]. Observational and experimental sleep studies have related poor sleep quality and short sleep duration to obesity and shortened LTL [54, 55]. We based our choice of BMI as a proxy for lifestyle following other studies in the NESDA cohort [52, 56]. Measured BMI (calculated by mass divided by height [2]) was categorized as underweight (<18.5), normal (18.5–24.9), overweight (25.0–29.9), and obese (≥30.0). The number of self-reported chronic diseases diagnosed and under treatment was determined with a 21-item, face-to-face interview. Physical activity was assessed with the International Physical Activity Questionnaire, expressed in Metabolic Equivalent Total (MET) minutes per week. We assessed depressive and anxiety disorders using the DSM-IV Composite International Diagnostic Interview (CIDI-2.1), administered by trained clinical research staff [57]. Affective disorders fell into two categories: no lifetime diagnosis; remitted (present during the lifetime history but not in the last year) or current (present in the last 6 months). Depressive and anxiety symptom severity were investigated using the Inventory of Depressive Symptomatology (IDS-SR) [58] and the Beck Anxiety Inventory (BAI) scales [59], respectively. We excluded the four sleep-related items of the IDS-SR because of overlap with our predictor variables, resulting in a range of 0–72. The Cronbach’s α for this adjusted scale was 0.83 [60]. Results from this cohort showed that associations were similar for depressive and anxiety symptom severity. Due to multicollinearity we did not additionally correct for anxiety severity. We used the Childhood Trauma Interview (CTI), in which participants were asked whether they were emotionally neglected; psychologically, physically, or sexually abused before the age of 16, as previously described [61]. The CTI reports the sum of the categories scored from 0 to 2 (0: never happened; 1: sometimes; 2: happened regularly); resulting in an index score between 0 and 8, which was used as a continuous variable. Statistical analyses We reported general characteristics using means and standard deviations for continuous data, and frequencies and percentages for categorical data. Normality assumptions for continuous variables were checked. Correlations between predictor variables were tested. The relationship between sociodemographic factors, lifestyle, depression, childhood trauma, and LTL was tested using generalized estimating equations (GEE) with LTL as outcome variable. We used GEE analyses with an exchangeable correlation structure as these take within-person correlations into account when examining multiple observations per participant [62]. We included all participants who had an LTL and a sleep/chronobiological assessment at least once, because GEE analyses tolerate missing observations. We performed separate GEE analyses. The first examined the associations between predictors sleep/chronobiological indicators and outcome LTL which was normally distributed, across two time points, 6 years apart. Wave, that is the within-subject variable defining the order of measurements, was categorized as 1 (wave 1) and 2 (wave 5). In the second GEE analyses, by adding time interactions, we investigated the longitudinal associations between sleep/chronobiological predictors and LTL attrition over 6 years. For non-time-varying predictors (course or constant), the time interactions describe differences between LTL attrition between the categories of the variable. Using a fully adjusted, multivariable GEE model, we examined chronic short sleep and sleep-onset time (free days) in one analysis simultaneously. We also conducted GEE analyses examining the associations between the three combinations of chronotype/sleep duration and outcome LTL, across two time points, 6 years apart, with time interactions. In all GEE analyses, we adjusted for age at baseline, sex, North European ancestry, current smoking, depression severity (both time-dependent), obesity, and childhood trauma index (fully adjusted), as these have previously been associated with longitudinal change in LTL in this cohort [10] and with sleep apnea and shorter LTL in other cohorts [35]. To account for multiple testing, in all GEE analyses, the Benjamini–Hochberg false discovery rate was calculated for significant findings [63]. Data were analyzed using SPSS for Windows (version 23.0; IBM Company, Chicago, IL). Statistical significance was inferred at p-value = .05. Results Descriptive analyses Table 1 shows the sample characteristics of the 2,936 subjects who had complete LTL measures at wave 1, those who had LTL measures at wave 5 (N = 1,883), and those who had chronobiological characteristics assessed at wave 3 (N = 2,561). LTL was 5,467 bp at wave 1 (SD = 617) and 5,386 bp at wave 5 (SD = 433), indicating a mean LTL attrition of 60 bp over the 6 years of the study (SD = 573). Table 1. Sample characteristics at wave 1, wave 3, and wave 5 Wave 1 (N = 2,936) Wave 5 (N = 1,883) Group comparisons* ,† Value (SE) P-value Demographics  Age, mean (SD) 41.8 (13.1) 48.6 (12.9)  Years of education, mean (SD) 12.2 (3.3) 12.9 (3.3)  Sex, female % (n) 66.4 (1,950) 65.4 (1,232)  North European ancestry, % (n) 94.8 (2,783) 96.0 (1,807) Lifestyle and health  Current smoker, % (n) 38.7 (1,136) 28.1 (529) χ 2 = 391.19 <.001  No. of alcoholic drinks per week, mean (SD) 7.0 (10.0) 6.2 (8.6) B = −1.01 (0.17) <.001  Body mass index, mean (SD) 25.6 (5.0) 26.3 (5.0) B = 0.63 (0.15) <.001   Underweight, % (n) 2.3 (50) 6.8 (128)   Normal, % (n) 51.8 (1,144) 42.2 (790)   Overweight, % (n) 30.1 (664) 32.0 (598)   Obese, % (n) 15.9 (350) 18.9 (354)  No. of chronic diseases under treatment, mean (SD) 0.61 (0.88) 0.62 (0.86) B = −0.1 (0.02) .522  Physical activity (in 1,000 MET-minutes per week), mean (SD)‡ 3.7 (3.1) 4.0 (3.4) B = 253.9 (77.86) .001 Psychiatric characteristics  IDS§ mean (SD) 17.6 (14.3) 9.9 (17.4) B = −8.92 (0.43) <.001  BAI|| mean (SD) 11.9 (10.6) 7.9 (8.6) B = −4.08 (0.20) <.001  Any lifetime depressive or anxiety disorder, % (n) 78 (2,292) 79 (1,495) χ 2 = 98.0 <.001  Childhood trauma index, mean (SD) 0.87 (1.12) Not measured Leukocyte telomere length (LTL), mean (SD) 5,467.8 (617.0) 5,386.8 (433.2) B = −80.76 (12.04) <.001 Sleep  Sleep duration, % (n) B = −0.06 (0.03) .074   ≤5 h 7.5 (154) 6.2 (112)   6 h 18.5 (381) 18.8 (340)   7 or 8 h 58.2 (1,197) 61.1 (1,106)   9 h 11.9 (244) 11.4 (206)   ≥10 h 3.8 (79) 2.6 (47)  Insomnia symptoms, % (n) 38.4 (1,128) 35.2 (662) χ 2 = 16.45 <.001  Insomnia continuous, mean (SD) 8.14 (5.11) 7.30 (4.77) B = −0.32 (0.12) .007 Course variables  Chronic short sleep (≤6h), % (n) 8.1 (238)  Chronic insomnia, % (n) 12.0 (352) Constant variables—chronobiology Wave 3 (N = 2,561)  Sleep-onset time free days, mean (SD) 00:10 am (1 h 07 min)  Mid-sleep on free days (MSFsc), mean (SD)¶ 03:47 am (37 min)  Indication of Delayed Sleep Phase Syndrome, % (n) 7.9 (148)  Chronotype adult (self-reported)   Extremely early, % (n) 2.3 (50)   Moderately early, % (n) 18.0 (387)   Intermediate, % (n) 48.6 (1,043)   Moderately late, % (n) 26.2 (562)   Extremely late, % (n) 4.9 (106)  Chronotype child (self-reported)   Extremely early, % (n) 3.5 (74)   Moderately early, % (n) 22.2 (473)   Intermediate, % (n) 57.5 (1,222)   Moderately late, % (n) 14.1 (300)   Extremely late, % (n) 2.7 (57) Wave 1 (N = 2,936) Wave 5 (N = 1,883) Group comparisons* ,† Value (SE) P-value Demographics  Age, mean (SD) 41.8 (13.1) 48.6 (12.9)  Years of education, mean (SD) 12.2 (3.3) 12.9 (3.3)  Sex, female % (n) 66.4 (1,950) 65.4 (1,232)  North European ancestry, % (n) 94.8 (2,783) 96.0 (1,807) Lifestyle and health  Current smoker, % (n) 38.7 (1,136) 28.1 (529) χ 2 = 391.19 <.001  No. of alcoholic drinks per week, mean (SD) 7.0 (10.0) 6.2 (8.6) B = −1.01 (0.17) <.001  Body mass index, mean (SD) 25.6 (5.0) 26.3 (5.0) B = 0.63 (0.15) <.001   Underweight, % (n) 2.3 (50) 6.8 (128)   Normal, % (n) 51.8 (1,144) 42.2 (790)   Overweight, % (n) 30.1 (664) 32.0 (598)   Obese, % (n) 15.9 (350) 18.9 (354)  No. of chronic diseases under treatment, mean (SD) 0.61 (0.88) 0.62 (0.86) B = −0.1 (0.02) .522  Physical activity (in 1,000 MET-minutes per week), mean (SD)‡ 3.7 (3.1) 4.0 (3.4) B = 253.9 (77.86) .001 Psychiatric characteristics  IDS§ mean (SD) 17.6 (14.3) 9.9 (17.4) B = −8.92 (0.43) <.001  BAI|| mean (SD) 11.9 (10.6) 7.9 (8.6) B = −4.08 (0.20) <.001  Any lifetime depressive or anxiety disorder, % (n) 78 (2,292) 79 (1,495) χ 2 = 98.0 <.001  Childhood trauma index, mean (SD) 0.87 (1.12) Not measured Leukocyte telomere length (LTL), mean (SD) 5,467.8 (617.0) 5,386.8 (433.2) B = −80.76 (12.04) <.001 Sleep  Sleep duration, % (n) B = −0.06 (0.03) .074   ≤5 h 7.5 (154) 6.2 (112)   6 h 18.5 (381) 18.8 (340)   7 or 8 h 58.2 (1,197) 61.1 (1,106)   9 h 11.9 (244) 11.4 (206)   ≥10 h 3.8 (79) 2.6 (47)  Insomnia symptoms, % (n) 38.4 (1,128) 35.2 (662) χ 2 = 16.45 <.001  Insomnia continuous, mean (SD) 8.14 (5.11) 7.30 (4.77) B = −0.32 (0.12) .007 Course variables  Chronic short sleep (≤6h), % (n) 8.1 (238)  Chronic insomnia, % (n) 12.0 (352) Constant variables—chronobiology Wave 3 (N = 2,561)  Sleep-onset time free days, mean (SD) 00:10 am (1 h 07 min)  Mid-sleep on free days (MSFsc), mean (SD)¶ 03:47 am (37 min)  Indication of Delayed Sleep Phase Syndrome, % (n) 7.9 (148)  Chronotype adult (self-reported)   Extremely early, % (n) 2.3 (50)   Moderately early, % (n) 18.0 (387)   Intermediate, % (n) 48.6 (1,043)   Moderately late, % (n) 26.2 (562)   Extremely late, % (n) 4.9 (106)  Chronotype child (self-reported)   Extremely early, % (n) 3.5 (74)   Moderately early, % (n) 22.2 (473)   Intermediate, % (n) 57.5 (1,222)   Moderately late, % (n) 14.1 (300)   Extremely late, % (n) 2.7 (57) *Group comparisons used McNemar’s test (χ 2) for dichotomous variables in repeated measures. †Group comparisons used GEE model for continuous variables in repeated measures. ‡Metabolic Equivalent Total minutes (MET-minutes). §Inventory of Depressive Symptomatology (IDS). ||Beck Anxiety Inventory (BAI). ¶MSFsc: mid-sleep on free days corrected for sleep debt on working days. Open in new tab Table 1. Sample characteristics at wave 1, wave 3, and wave 5 Wave 1 (N = 2,936) Wave 5 (N = 1,883) Group comparisons* ,† Value (SE) P-value Demographics  Age, mean (SD) 41.8 (13.1) 48.6 (12.9)  Years of education, mean (SD) 12.2 (3.3) 12.9 (3.3)  Sex, female % (n) 66.4 (1,950) 65.4 (1,232)  North European ancestry, % (n) 94.8 (2,783) 96.0 (1,807) Lifestyle and health  Current smoker, % (n) 38.7 (1,136) 28.1 (529) χ 2 = 391.19 <.001  No. of alcoholic drinks per week, mean (SD) 7.0 (10.0) 6.2 (8.6) B = −1.01 (0.17) <.001  Body mass index, mean (SD) 25.6 (5.0) 26.3 (5.0) B = 0.63 (0.15) <.001   Underweight, % (n) 2.3 (50) 6.8 (128)   Normal, % (n) 51.8 (1,144) 42.2 (790)   Overweight, % (n) 30.1 (664) 32.0 (598)   Obese, % (n) 15.9 (350) 18.9 (354)  No. of chronic diseases under treatment, mean (SD) 0.61 (0.88) 0.62 (0.86) B = −0.1 (0.02) .522  Physical activity (in 1,000 MET-minutes per week), mean (SD)‡ 3.7 (3.1) 4.0 (3.4) B = 253.9 (77.86) .001 Psychiatric characteristics  IDS§ mean (SD) 17.6 (14.3) 9.9 (17.4) B = −8.92 (0.43) <.001  BAI|| mean (SD) 11.9 (10.6) 7.9 (8.6) B = −4.08 (0.20) <.001  Any lifetime depressive or anxiety disorder, % (n) 78 (2,292) 79 (1,495) χ 2 = 98.0 <.001  Childhood trauma index, mean (SD) 0.87 (1.12) Not measured Leukocyte telomere length (LTL), mean (SD) 5,467.8 (617.0) 5,386.8 (433.2) B = −80.76 (12.04) <.001 Sleep  Sleep duration, % (n) B = −0.06 (0.03) .074   ≤5 h 7.5 (154) 6.2 (112)   6 h 18.5 (381) 18.8 (340)   7 or 8 h 58.2 (1,197) 61.1 (1,106)   9 h 11.9 (244) 11.4 (206)   ≥10 h 3.8 (79) 2.6 (47)  Insomnia symptoms, % (n) 38.4 (1,128) 35.2 (662) χ 2 = 16.45 <.001  Insomnia continuous, mean (SD) 8.14 (5.11) 7.30 (4.77) B = −0.32 (0.12) .007 Course variables  Chronic short sleep (≤6h), % (n) 8.1 (238)  Chronic insomnia, % (n) 12.0 (352) Constant variables—chronobiology Wave 3 (N = 2,561)  Sleep-onset time free days, mean (SD) 00:10 am (1 h 07 min)  Mid-sleep on free days (MSFsc), mean (SD)¶ 03:47 am (37 min)  Indication of Delayed Sleep Phase Syndrome, % (n) 7.9 (148)  Chronotype adult (self-reported)   Extremely early, % (n) 2.3 (50)   Moderately early, % (n) 18.0 (387)   Intermediate, % (n) 48.6 (1,043)   Moderately late, % (n) 26.2 (562)   Extremely late, % (n) 4.9 (106)  Chronotype child (self-reported)   Extremely early, % (n) 3.5 (74)   Moderately early, % (n) 22.2 (473)   Intermediate, % (n) 57.5 (1,222)   Moderately late, % (n) 14.1 (300)   Extremely late, % (n) 2.7 (57) Wave 1 (N = 2,936) Wave 5 (N = 1,883) Group comparisons* ,† Value (SE) P-value Demographics  Age, mean (SD) 41.8 (13.1) 48.6 (12.9)  Years of education, mean (SD) 12.2 (3.3) 12.9 (3.3)  Sex, female % (n) 66.4 (1,950) 65.4 (1,232)  North European ancestry, % (n) 94.8 (2,783) 96.0 (1,807) Lifestyle and health  Current smoker, % (n) 38.7 (1,136) 28.1 (529) χ 2 = 391.19 <.001  No. of alcoholic drinks per week, mean (SD) 7.0 (10.0) 6.2 (8.6) B = −1.01 (0.17) <.001  Body mass index, mean (SD) 25.6 (5.0) 26.3 (5.0) B = 0.63 (0.15) <.001   Underweight, % (n) 2.3 (50) 6.8 (128)   Normal, % (n) 51.8 (1,144) 42.2 (790)   Overweight, % (n) 30.1 (664) 32.0 (598)   Obese, % (n) 15.9 (350) 18.9 (354)  No. of chronic diseases under treatment, mean (SD) 0.61 (0.88) 0.62 (0.86) B = −0.1 (0.02) .522  Physical activity (in 1,000 MET-minutes per week), mean (SD)‡ 3.7 (3.1) 4.0 (3.4) B = 253.9 (77.86) .001 Psychiatric characteristics  IDS§ mean (SD) 17.6 (14.3) 9.9 (17.4) B = −8.92 (0.43) <.001  BAI|| mean (SD) 11.9 (10.6) 7.9 (8.6) B = −4.08 (0.20) <.001  Any lifetime depressive or anxiety disorder, % (n) 78 (2,292) 79 (1,495) χ 2 = 98.0 <.001  Childhood trauma index, mean (SD) 0.87 (1.12) Not measured Leukocyte telomere length (LTL), mean (SD) 5,467.8 (617.0) 5,386.8 (433.2) B = −80.76 (12.04) <.001 Sleep  Sleep duration, % (n) B = −0.06 (0.03) .074   ≤5 h 7.5 (154) 6.2 (112)   6 h 18.5 (381) 18.8 (340)   7 or 8 h 58.2 (1,197) 61.1 (1,106)   9 h 11.9 (244) 11.4 (206)   ≥10 h 3.8 (79) 2.6 (47)  Insomnia symptoms, % (n) 38.4 (1,128) 35.2 (662) χ 2 = 16.45 <.001  Insomnia continuous, mean (SD) 8.14 (5.11) 7.30 (4.77) B = −0.32 (0.12) .007 Course variables  Chronic short sleep (≤6h), % (n) 8.1 (238)  Chronic insomnia, % (n) 12.0 (352) Constant variables—chronobiology Wave 3 (N = 2,561)  Sleep-onset time free days, mean (SD) 00:10 am (1 h 07 min)  Mid-sleep on free days (MSFsc), mean (SD)¶ 03:47 am (37 min)  Indication of Delayed Sleep Phase Syndrome, % (n) 7.9 (148)  Chronotype adult (self-reported)   Extremely early, % (n) 2.3 (50)   Moderately early, % (n) 18.0 (387)   Intermediate, % (n) 48.6 (1,043)   Moderately late, % (n) 26.2 (562)   Extremely late, % (n) 4.9 (106)  Chronotype child (self-reported)   Extremely early, % (n) 3.5 (74)   Moderately early, % (n) 22.2 (473)   Intermediate, % (n) 57.5 (1,222)   Moderately late, % (n) 14.1 (300)   Extremely late, % (n) 2.7 (57) *Group comparisons used McNemar’s test (χ 2) for dichotomous variables in repeated measures. †Group comparisons used GEE model for continuous variables in repeated measures. ‡Metabolic Equivalent Total minutes (MET-minutes). §Inventory of Depressive Symptomatology (IDS). ||Beck Anxiety Inventory (BAI). ¶MSFsc: mid-sleep on free days corrected for sleep debt on working days. Open in new tab Table 2 shows correlations between the different sleep indicators. Late sleep-onset was significantly positively correlated with later MSFsc (r = 0.46; p < .001). Longer sleep duration was significantly and negatively correlated with later sleep-onset time on free days (r = −0.37; p < .001), meaning that longer sleepers tended to fall asleep earlier. Long sleep was also positively correlated with later MSFsc (r = 0.43, p < .001), indicating that mid-sleep is very much affected by sleep duration. Therefore, sleep-onset time may be a more important indicator of chronotype. There was also a moderate association between self-reported chronotype in adulthood and childhood (not shown, Spearman’s correlation coefficient = 0.47, p < .001), which corresponds to the description of chronotype as a lifelong trait [64]. Table 2. Pearson’s intercorrelations between different sleep indicators and chronotype Sleep duration (free days) Insomnia Rating Scale score (wave 1) Sleep-onset time (free days) Mid-sleep on free days (MSFsc) Indication of Delayed Sleep Phase Syndrome Chronotype adult (self-reported) Sleep duration 1 Insomnia Rating Scale score (wave 1) −0.25* 1 Sleep-onset time (free days, hours) −0.37* 0.03 1 Mid-sleep on free days (MSFsc)† 0.43* −0.17* 0.46* 1 Indication of Delayed Sleep Phase Syndrome −0.18* 0.10* 0.37* 0.00 1 Chronotype adult (self-reported, from early to late) 0.15* −0.10* 0.46* 0.46* 0.26* 1 Sleep duration (free days) Insomnia Rating Scale score (wave 1) Sleep-onset time (free days) Mid-sleep on free days (MSFsc) Indication of Delayed Sleep Phase Syndrome Chronotype adult (self-reported) Sleep duration 1 Insomnia Rating Scale score (wave 1) −0.25* 1 Sleep-onset time (free days, hours) −0.37* 0.03 1 Mid-sleep on free days (MSFsc)† 0.43* −0.17* 0.46* 1 Indication of Delayed Sleep Phase Syndrome −0.18* 0.10* 0.37* 0.00 1 Chronotype adult (self-reported, from early to late) 0.15* −0.10* 0.46* 0.46* 0.26* 1 †MSFsc: mid-sleep on free days corrected for sleep debt on working days. *p < .001. Open in new tab Table 2. Pearson’s intercorrelations between different sleep indicators and chronotype Sleep duration (free days) Insomnia Rating Scale score (wave 1) Sleep-onset time (free days) Mid-sleep on free days (MSFsc) Indication of Delayed Sleep Phase Syndrome Chronotype adult (self-reported) Sleep duration 1 Insomnia Rating Scale score (wave 1) −0.25* 1 Sleep-onset time (free days, hours) −0.37* 0.03 1 Mid-sleep on free days (MSFsc)† 0.43* −0.17* 0.46* 1 Indication of Delayed Sleep Phase Syndrome −0.18* 0.10* 0.37* 0.00 1 Chronotype adult (self-reported, from early to late) 0.15* −0.10* 0.46* 0.46* 0.26* 1 Sleep duration (free days) Insomnia Rating Scale score (wave 1) Sleep-onset time (free days) Mid-sleep on free days (MSFsc) Indication of Delayed Sleep Phase Syndrome Chronotype adult (self-reported) Sleep duration 1 Insomnia Rating Scale score (wave 1) −0.25* 1 Sleep-onset time (free days, hours) −0.37* 0.03 1 Mid-sleep on free days (MSFsc)† 0.43* −0.17* 0.46* 1 Indication of Delayed Sleep Phase Syndrome −0.18* 0.10* 0.37* 0.00 1 Chronotype adult (self-reported, from early to late) 0.15* −0.10* 0.46* 0.46* 0.26* 1 †MSFsc: mid-sleep on free days corrected for sleep debt on working days. *p < .001. Open in new tab Associations with LTL over 6 years In Table 3, the GEE analyses show that LTL was significantly shorter in those with delayed circadian rhythm, in both the partially and fully adjusted models, taking baseline LTL into account. Specifically, those with later MSFsc, later sleep-onset time or an indication of DSPS had significantly shorter LTL compared to those without these disturbances, in both models (fully adjusted model: MSFsc B = −49.9, p = .004; sleep-onset time B = −32.4, p = .001; indication of DSPS B = −73.8, p = .036). Moderately late self-reported chronotype in adulthood was significantly associated with shorter LTL, compared to the reference of intermediate chronotype (B = −71.6, p = .003). In contrast, self-reported extremely early chronotype (in adulthood) was associated with significantly less shortening (fully adjusted model: B = 161.4, p = .037) than intermediate chronotype. Accounting for multiple testing using the Benjamini–Hochberg false discovery rate did not affect any of the significant findings. In contrast, GEE analyses with sleep duration, chronic short sleep, insomnia symptoms (cutoff score of 9, continuous, chronic insomnia), and self-reported chronotype in childhood did not show a significant relationship with LTL in either model. We also examined the four questions of the WHIIRS individually at both waves. For the question concerning problems falling asleep, we found significantly shorter LTL where problems occurred five or more times per week, but only in the partially adjusted model (not shown). Table 3. Associations between predictors sleep and chronobiological parameters, and outcome leukocyte telomere length (LTL) over two time points (N = 2,936)* Mean LTL, partially adjusted Mean LTL, fully adjusted B (SE) F (df) P-value ηp2 B (SE) F (df) P-value ηp2 Time-varying variables  Sleep duration   ≤5 h −25.50 (35.23) 0.23 (4,365) .469 .000 −7.06 (35.66) 0.14 (4,3822) .843 .000   6 h −12.43 (22.46) .580 7.09 (21.63) .743   7 or 8 h Reference — .756 Reference — .685   9 h −8.36 (.26.89) .763 10.22 (25.21) .447   ≥10 h 16.28(54.01) 38.76 (50.96)  Insomnia symptoms −4.11 (18.36) 0.00 (1,3837) .823 .000 11.91 (18.91) 0.86 (1,3795) .529 .000  Insomnia continuous −1.04 (1.88) 0.12 (1,3837) .580 .000 1.36 (2.08) 0.49 (1,3795) .512 .000 Course variables  Chronic short sleep (≤6 h) −73.38 (35.91) 3.37 (1,1976) .041 .002 −66.26 (37.17) 2.57 (1,1951) .075 .001  Chronic insomnia −9.21 (27.20) 0.17 (1,3686) .735 .000 −0.50 (28.46) 0.01 (1,3643) .986 .000 Constant variables—chronobiology  Mid-sleep on free days (MSFsc, hours)† −48.60 (17.01) 9.32 (1,3093) .004+ .003 −49.89 (17.36) 10.12 (1,3055) .004+ .003  Sleep-onset time (free days, hours) −34.04 (8.99) 17.12 (1,3181) <.001+ .005 −32.36 (9.39) 13.64 (1,3143) .001+ .004  Indication of Delayed Sleep Phase Syndrome −84.84(33.96) 5.81 (1,2996) .012+ .002 −73.78 (35.17) 3.51 (1,2961) .036+ .001  Chronotype adult (self-reported) 7.08 (4,3441) .008 6.35 (1,3400) .007   Extremely early 141.05 (76.50) .062 — 161.40 (77.20) — .037+ —   Moderately early 30.91 (28.67) .281 — 28.64 (29.04) — .324 —   Intermediate Reference — Reference —   Moderately late −80.43 (23.63) .001+ — −71.63 (24.08) — .003+ —   Extremely late −42.98 (44.91) .339 — −29.09 (45.58) — .407 —  Chronotype child (self-reported) 2.03 (4,3410) .002 1.96 (4,3371) .002   Extremely early 54.45 (61.11) .373 — 73.77 (63.11) — .242 —   Moderately early 17.36 (25.03) .488 15.24 (25.76) — .554 —   Intermediate Reference — Reference — —   Moderately late −45.50 (30.33) .134 −40.54 (30.57) .185   Extremely late −88.61 (62.26) .155 −74.91 (62.46) — .230 — Mean LTL, partially adjusted Mean LTL, fully adjusted B (SE) F (df) P-value ηp2 B (SE) F (df) P-value ηp2 Time-varying variables  Sleep duration   ≤5 h −25.50 (35.23) 0.23 (4,365) .469 .000 −7.06 (35.66) 0.14 (4,3822) .843 .000   6 h −12.43 (22.46) .580 7.09 (21.63) .743   7 or 8 h Reference — .756 Reference — .685   9 h −8.36 (.26.89) .763 10.22 (25.21) .447   ≥10 h 16.28(54.01) 38.76 (50.96)  Insomnia symptoms −4.11 (18.36) 0.00 (1,3837) .823 .000 11.91 (18.91) 0.86 (1,3795) .529 .000  Insomnia continuous −1.04 (1.88) 0.12 (1,3837) .580 .000 1.36 (2.08) 0.49 (1,3795) .512 .000 Course variables  Chronic short sleep (≤6 h) −73.38 (35.91) 3.37 (1,1976) .041 .002 −66.26 (37.17) 2.57 (1,1951) .075 .001  Chronic insomnia −9.21 (27.20) 0.17 (1,3686) .735 .000 −0.50 (28.46) 0.01 (1,3643) .986 .000 Constant variables—chronobiology  Mid-sleep on free days (MSFsc, hours)† −48.60 (17.01) 9.32 (1,3093) .004+ .003 −49.89 (17.36) 10.12 (1,3055) .004+ .003  Sleep-onset time (free days, hours) −34.04 (8.99) 17.12 (1,3181) <.001+ .005 −32.36 (9.39) 13.64 (1,3143) .001+ .004  Indication of Delayed Sleep Phase Syndrome −84.84(33.96) 5.81 (1,2996) .012+ .002 −73.78 (35.17) 3.51 (1,2961) .036+ .001  Chronotype adult (self-reported) 7.08 (4,3441) .008 6.35 (1,3400) .007   Extremely early 141.05 (76.50) .062 — 161.40 (77.20) — .037+ —   Moderately early 30.91 (28.67) .281 — 28.64 (29.04) — .324 —   Intermediate Reference — Reference —   Moderately late −80.43 (23.63) .001+ — −71.63 (24.08) — .003+ —   Extremely late −42.98 (44.91) .339 — −29.09 (45.58) — .407 —  Chronotype child (self-reported) 2.03 (4,3410) .002 1.96 (4,3371) .002   Extremely early 54.45 (61.11) .373 — 73.77 (63.11) — .242 —   Moderately early 17.36 (25.03) .488 15.24 (25.76) — .554 —   Intermediate Reference — Reference — —   Moderately late −45.50 (30.33) .134 −40.54 (30.57) .185   Extremely late −88.61 (62.26) .155 −74.91 (62.46) — .230 — Partially adjusted: corrected for age at wave 1, female sex, North European ancestry, wave; fully adjusted: additionally corrected for current smoker, depression severity, obesity, childhood trauma. *Generalized estimating equation (GEE) analyses where effect sizes are partial eta squared (ηp2), N of observations range = 2,987–3,866. †MSFsc: mid-sleep on free days corrected for sleep debt on working days. +Remained significant when applying the Benjamini–Hochberg false discovery procedure. Open in new tab Table 3. Associations between predictors sleep and chronobiological parameters, and outcome leukocyte telomere length (LTL) over two time points (N = 2,936)* Mean LTL, partially adjusted Mean LTL, fully adjusted B (SE) F (df) P-value ηp2 B (SE) F (df) P-value ηp2 Time-varying variables  Sleep duration   ≤5 h −25.50 (35.23) 0.23 (4,365) .469 .000 −7.06 (35.66) 0.14 (4,3822) .843 .000   6 h −12.43 (22.46) .580 7.09 (21.63) .743   7 or 8 h Reference — .756 Reference — .685   9 h −8.36 (.26.89) .763 10.22 (25.21) .447   ≥10 h 16.28(54.01) 38.76 (50.96)  Insomnia symptoms −4.11 (18.36) 0.00 (1,3837) .823 .000 11.91 (18.91) 0.86 (1,3795) .529 .000  Insomnia continuous −1.04 (1.88) 0.12 (1,3837) .580 .000 1.36 (2.08) 0.49 (1,3795) .512 .000 Course variables  Chronic short sleep (≤6 h) −73.38 (35.91) 3.37 (1,1976) .041 .002 −66.26 (37.17) 2.57 (1,1951) .075 .001  Chronic insomnia −9.21 (27.20) 0.17 (1,3686) .735 .000 −0.50 (28.46) 0.01 (1,3643) .986 .000 Constant variables—chronobiology  Mid-sleep on free days (MSFsc, hours)† −48.60 (17.01) 9.32 (1,3093) .004+ .003 −49.89 (17.36) 10.12 (1,3055) .004+ .003  Sleep-onset time (free days, hours) −34.04 (8.99) 17.12 (1,3181) <.001+ .005 −32.36 (9.39) 13.64 (1,3143) .001+ .004  Indication of Delayed Sleep Phase Syndrome −84.84(33.96) 5.81 (1,2996) .012+ .002 −73.78 (35.17) 3.51 (1,2961) .036+ .001  Chronotype adult (self-reported) 7.08 (4,3441) .008 6.35 (1,3400) .007   Extremely early 141.05 (76.50) .062 — 161.40 (77.20) — .037+ —   Moderately early 30.91 (28.67) .281 — 28.64 (29.04) — .324 —   Intermediate Reference — Reference —   Moderately late −80.43 (23.63) .001+ — −71.63 (24.08) — .003+ —   Extremely late −42.98 (44.91) .339 — −29.09 (45.58) — .407 —  Chronotype child (self-reported) 2.03 (4,3410) .002 1.96 (4,3371) .002   Extremely early 54.45 (61.11) .373 — 73.77 (63.11) — .242 —   Moderately early 17.36 (25.03) .488 15.24 (25.76) — .554 —   Intermediate Reference — Reference — —   Moderately late −45.50 (30.33) .134 −40.54 (30.57) .185   Extremely late −88.61 (62.26) .155 −74.91 (62.46) — .230 — Mean LTL, partially adjusted Mean LTL, fully adjusted B (SE) F (df) P-value ηp2 B (SE) F (df) P-value ηp2 Time-varying variables  Sleep duration   ≤5 h −25.50 (35.23) 0.23 (4,365) .469 .000 −7.06 (35.66) 0.14 (4,3822) .843 .000   6 h −12.43 (22.46) .580 7.09 (21.63) .743   7 or 8 h Reference — .756 Reference — .685   9 h −8.36 (.26.89) .763 10.22 (25.21) .447   ≥10 h 16.28(54.01) 38.76 (50.96)  Insomnia symptoms −4.11 (18.36) 0.00 (1,3837) .823 .000 11.91 (18.91) 0.86 (1,3795) .529 .000  Insomnia continuous −1.04 (1.88) 0.12 (1,3837) .580 .000 1.36 (2.08) 0.49 (1,3795) .512 .000 Course variables  Chronic short sleep (≤6 h) −73.38 (35.91) 3.37 (1,1976) .041 .002 −66.26 (37.17) 2.57 (1,1951) .075 .001  Chronic insomnia −9.21 (27.20) 0.17 (1,3686) .735 .000 −0.50 (28.46) 0.01 (1,3643) .986 .000 Constant variables—chronobiology  Mid-sleep on free days (MSFsc, hours)† −48.60 (17.01) 9.32 (1,3093) .004+ .003 −49.89 (17.36) 10.12 (1,3055) .004+ .003  Sleep-onset time (free days, hours) −34.04 (8.99) 17.12 (1,3181) <.001+ .005 −32.36 (9.39) 13.64 (1,3143) .001+ .004  Indication of Delayed Sleep Phase Syndrome −84.84(33.96) 5.81 (1,2996) .012+ .002 −73.78 (35.17) 3.51 (1,2961) .036+ .001  Chronotype adult (self-reported) 7.08 (4,3441) .008 6.35 (1,3400) .007   Extremely early 141.05 (76.50) .062 — 161.40 (77.20) — .037+ —   Moderately early 30.91 (28.67) .281 — 28.64 (29.04) — .324 —   Intermediate Reference — Reference —   Moderately late −80.43 (23.63) .001+ — −71.63 (24.08) — .003+ —   Extremely late −42.98 (44.91) .339 — −29.09 (45.58) — .407 —  Chronotype child (self-reported) 2.03 (4,3410) .002 1.96 (4,3371) .002   Extremely early 54.45 (61.11) .373 — 73.77 (63.11) — .242 —   Moderately early 17.36 (25.03) .488 15.24 (25.76) — .554 —   Intermediate Reference — Reference — —   Moderately late −45.50 (30.33) .134 −40.54 (30.57) .185   Extremely late −88.61 (62.26) .155 −74.91 (62.46) — .230 — Partially adjusted: corrected for age at wave 1, female sex, North European ancestry, wave; fully adjusted: additionally corrected for current smoker, depression severity, obesity, childhood trauma. *Generalized estimating equation (GEE) analyses where effect sizes are partial eta squared (ηp2), N of observations range = 2,987–3,866. †MSFsc: mid-sleep on free days corrected for sleep debt on working days. +Remained significant when applying the Benjamini–Hochberg false discovery procedure. Open in new tab Table 4 shows sleep-onset time (free days), a robust marker of altered circadian rhythm and chronic short sleep in one fully adjusted, multivariable GEE analysis. Again, we found significantly shorter LTL for age (B = −11.92, p < .001), male sex (B = −58.13, p = .044), chronic short sleep (B = −87.18, p = .033), and sleep-onset time (B = −36.90, p = .002). Table 4. Multivariate associations between predictors sleep and chronobiological parameters, and outcome leukocyte telomere length (LTL) over two time points in one model (N = 2,936)* Mean LTL, fully adjusted B (SE) P-value Age −11.92 (1.12) <.001+ Sex (male = reference) 58.13 (28.92) .044 Wave −59.99 (26.27) .022+ Chronic short sleep −87.18 (40.92) .033+ Sleep-onset time (free days, hours) −36.90 (11.90) .002+ Mean LTL, fully adjusted B (SE) P-value Age −11.92 (1.12) <.001+ Sex (male = reference) 58.13 (28.92) .044 Wave −59.99 (26.27) .022+ Chronic short sleep −87.18 (40.92) .033+ Sleep-onset time (free days, hours) −36.90 (11.90) .002+ Fully adjusted: corrected for North European ancestry, current smoker, depression severity, obesity, childhood trauma. *Generalized estimating equation (GEE) with N of observations = 1,598. +Remained significant when applying the Benjamini–Hochberg false discovery procedure. Open in new tab Table 4. Multivariate associations between predictors sleep and chronobiological parameters, and outcome leukocyte telomere length (LTL) over two time points in one model (N = 2,936)* Mean LTL, fully adjusted B (SE) P-value Age −11.92 (1.12) <.001+ Sex (male = reference) 58.13 (28.92) .044 Wave −59.99 (26.27) .022+ Chronic short sleep −87.18 (40.92) .033+ Sleep-onset time (free days, hours) −36.90 (11.90) .002+ Mean LTL, fully adjusted B (SE) P-value Age −11.92 (1.12) <.001+ Sex (male = reference) 58.13 (28.92) .044 Wave −59.99 (26.27) .022+ Chronic short sleep −87.18 (40.92) .033+ Sleep-onset time (free days, hours) −36.90 (11.90) .002+ Fully adjusted: corrected for North European ancestry, current smoker, depression severity, obesity, childhood trauma. *Generalized estimating equation (GEE) with N of observations = 1,598. +Remained significant when applying the Benjamini–Hochberg false discovery procedure. Open in new tab In terms of the covariates age, sex, North European ancestry, in all GEE analyses, in both models, older age was consistently significant (p < .001). In analyses in the fully adjusted model, significance of p < .05 was attained for female sex where self-reported chronotype in adulthood was the predictor. Similarly, in all analyses, North European ancestry showed significance of p < .05, except where self-reported chronotype in adulthood, chronic short sleep, or chronic insomnia were the predictors. Current smoking attained significance of p < .05 in analyses where sleep duration, insomnia symptoms, chronic insomnia, and DSPS were the predictors (former three not shown). Childhood trauma, obesity, and depression severity were not significant in any analyses. Results of the GEE analyses exploring the groups with varying sleep duration and chronotype over 6 years showed significantly shorter LTL in both the partially and fully adjusted models for the highest risk group (moderate to extremely late chronotype with short sleep ≤6 h, fully adjusted model, B = −91.2, p = .028) and for moderate or extremely late chronotype with sleep duration 7–8 h (fully adjusted model, B = −79.1, p = .028), compared to those with early/intermediate chronotype and with sleep duration 7–8 h (not shown). Figure 1 plots the estimated means from the GEE models of LTL at both time points, for each of the three groups. Those with late chronotype and sleep duration of ≤6 h and 7–8 h showed statistically significant LTL shortening compared to the reference group in the fully adjusted model. Figure 1. Open in new tabDownload slide Mean leukocyte telomere length (LTL) at wave 1 and wave 5 in different groups of chronotype and sleep duration in adulthood. *p < .05, in Generalized Estimating Equations models with time interactions including age at wave 1, female sex, North European ancestry, current smoker, depression severity, childhood trauma, obesity, wave. Figure 1. Open in new tabDownload slide Mean leukocyte telomere length (LTL) at wave 1 and wave 5 in different groups of chronotype and sleep duration in adulthood. *p < .05, in Generalized Estimating Equations models with time interactions including age at wave 1, female sex, North European ancestry, current smoker, depression severity, childhood trauma, obesity, wave. LTL attrition over 6 years In order to examine longitudinal associations between sleep duration/chronobiological predictors and outcome LTL attrition, we conducted further GEE analyses, adding time interactions (not shown). No results from these analyses showed statistically significant time interactions. While LTL did shorten on average over the 6 years of the study, and time was significant throughout the GEE analyses, we found no substantial difference in slopes of LTL over time, as indicated by the nonsignificant interaction terms. For the time-varying predictors (insomnia measured continuously and sleep duration), the time interactions showed no significance (not shown). The overall picture emerging from these analyses is that chronobiological dysregulations are consistently associated with shorter LTL throughout the entire follow-up, and this relationship remains consistent over time. Figure 2 illustrates these results in four panels, in fully adjusted models. We plotted the estimated means from the GEE models of LTL at both time points according to different chronobiological predictors. For illustrative purposes, continuous predictors MSFsc and sleep-onset time were categorized into quintiles ranging from early to late. For MSFsc, there were no significant differences for any quintiles compared to the intermediate reference group. However, sleep-onset time in quintiles showed significant differences between the latest two quintiles and the intermediate reference quintile. An indication of DSPS at both time points showed significantly shorter mean LTL, in the fully adjusted model. From these results, it appears that the presence of DSPS accelerates cellular aging by 6 years. While moderately late chronotype showed significant significantly shorter means for LTL, extremely early chronotype shows significantly less shortening for LTL (both compared to the intermediate chronotype, in adulthood). In the panel showing significant differences between mean LTL, the slopes are parallel, illustrating the finding that there was no difference in LTL attrition rate over time. Figure 2. Open in new tabDownload slide Mean leukocyte telomere length (LTL) at wave 1 and wave 5 in different indicators of circadian rhythm in adulthood. **p < .01; *p < .05, in Generalized Estimating Equations models with time interactions (Table 4) including age at wave 1, female sex, North European ancestry, current smoker, depression severity, childhood trauma, obesity, wave. aFive different self-reported chronotypes in adulthood where intermediate chronotype is reference. bQuintiles of sleep-onset time (free days) from early to late: Q1: 7:35–11:10 pm; Q2: 11:13–11:50 pm; Q3 (reference): 11:53 pm–12:15 am; Q4: 12:18–1:04 am; Q5: 1:05–4:18 am. cSubjects with and without (reference) indication of Delayed Sleep Phase Syndrome (DSPS). dQuintiles of mid-sleep on free days (MSFsc) from early to late: Q1: 1–3:20 am; Q2: 3:20–3:38 am; Q3 (reference): 3:38–3:53 am; Q4: 3:53–4:14 am; Q5: 4:14–7:59 am. Figure 2. Open in new tabDownload slide Mean leukocyte telomere length (LTL) at wave 1 and wave 5 in different indicators of circadian rhythm in adulthood. **p < .01; *p < .05, in Generalized Estimating Equations models with time interactions (Table 4) including age at wave 1, female sex, North European ancestry, current smoker, depression severity, childhood trauma, obesity, wave. aFive different self-reported chronotypes in adulthood where intermediate chronotype is reference. bQuintiles of sleep-onset time (free days) from early to late: Q1: 7:35–11:10 pm; Q2: 11:13–11:50 pm; Q3 (reference): 11:53 pm–12:15 am; Q4: 12:18–1:04 am; Q5: 1:05–4:18 am. cSubjects with and without (reference) indication of Delayed Sleep Phase Syndrome (DSPS). dQuintiles of mid-sleep on free days (MSFsc) from early to late: Q1: 1–3:20 am; Q2: 3:20–3:38 am; Q3 (reference): 3:38–3:53 am; Q4: 3:53–4:14 am; Q5: 4:14–7:59 am. Discussion While circadian dysregulation has been associated with many deleterious health consequences, to our knowledge there are no published data on circadian desynchronization and cellular aging. In this large-scale study, we explored the relationship between LTL and measures of sleep length, insomnia symptoms, and markers of circadian dysregulation. We also examined whether these factors were associated with accelerated LTL attrition over 6 years, taking into account baseline LTL [10]. Our results uniformly show that delayed circadian rhythm is associated with shorter LTL, rather than short sleep duration or insomnia. However, chronic short sleep (≤6 h) showed an association with significantly shorter LTL, even in a multivariable model that also included sleep-onset time on free days. Indicators of delayed circadian rhythm included late MSFsc, later sleep-onset time (on free days), indication of DSPS, and moderately late self-reported chronotype in adulthood. These were all significantly associated with shorter LTL across wave 1 and 6-year time points, after correction for wide-ranging sociodemographic and lifestyle factors. Combining chronotype and sleep duration into groups revealed significantly shorter LTL in the groups with both late chronotype and normal to short sleep. These appear to be an at-risk group and deserve further investigation. When we examined LTL attrition rate over time, there was no evidence of accelerated LTL attrition rate for any predictor over 6 years. We report that the average telomere shortening per year is 8 bp. This is less than that reported in other studies [8]. However, in the NESDA cohort, there is a wide range of telomere shortening, and it is the individual difference of change over 6 years, not the average shortening rate, that we correlated with the sleep and circadian variables. We used the WHIIRS to estimate insomnia symptoms. Using a cutoff score of 9, we found a prevalence of 38% and 35% per waves 1 and 5, respectively. In another study from the NESDA cohort, the highest WHIIRS scores were associated with increasing depressive/anxiety psychopathology [65]. However, in a study of women (N = 93,532), insomnia prevalence at baseline was 24.5%, which is lower than our finding. This may be explained by the selection of our sample to study depressive and anxiety disorders, where more insomnia comorbidity may have been present [66]. Although delayed sleep-onset is an indicator of circadian rhythm delay, it is not the only reason for delayed sleep. Indeed, when we examined the five questions of the WHIIRS individually with regard to LTL as outcome, self-reported problems falling asleep which are part of the insomnia symptom complex, were found to be significantly associated with shorter LTL. Our significant findings of shorter LTL in those with delayed circadian rhythm may have important consequences for health. Circadian desynchrony has been related to several psychiatric disorders (Attention Deficit Hyperactivity Disorder, bipolar and other mood disorders) [67–70]; age [71], environmental factors such as night shift work [72], and metabolic changes (obesity, cardiovascular risk) [18, 42]. Recent research has highlighted that nurses with long periods of consecutive night shifts have telomere shortening associated with an increased breast cancer risk [73]. While it is recognized that the large interindividual variation in LTL is explained by genetic factors in over two-thirds [6, 7] research has shown that age [8], lifestyle, night shift work [73], and illness also play a role [3]. Our results further show that other possibly modifiable factors, such as delayed sleep-onset time, appear to play a significant role too. Identifying those at high risk for delayed circadian rhythm may be a first step in addressing these deleterious consequences. The mechanism linking the circadian and metabolic systems is an intriguing area of study, and is only partially understood (reviewed in Ray [18]). The effects of the circadian rhythm dysregulation on telomeres may relate to oxidative stress. Telomere length is associated with inflammation [7C4] and possibly shorten as a result of increased oxidative stress [75]. Senescent cells with shortened telomeres also show increased secretion of pro-inflammatory cytokines and extracellular matrix-degrading enzymes [1, 76]. This may drive accelerated disease progression. Hence telomeres may be active and dynamic structures, increasingly considered to be a reflection of a (long-term) state [77]. Dysregulated sleep is also associated with chronic inflammation [78], which can be seen on a cellular level [79, 80]. For example, sleep deprivation in humans adversely affects the oscillatory behavior of many blood metabolites, which fluctuate in accordance with the circadian rhythm [44]. Healthy humans deprived of sleep show significant elevations in inflammatory activity compared to an undisturbed sleep condition [81, 82], even after one night of sleep deprivation [83]. A cumulative sleep debt frequently ensues from delayed sleep-onset in those with a late chronotype. Mechanisms underlying the shorter LTL seen in delayed circadian rhythm may include increased inflammation and circadian alteration of cellular metabolites, leading to cellular damage and premature cellular aging. Ultimately, this may increase risk for carcinogenic transformation and somatic disease [4]. In contrast, shorter LTL itself may predispose to delayed circadian rhythm: the direction of the relationship is not certain. We also showed that persons with self-reported, extremely early chronotype in adulthood, had significantly less LTL shortening than those with intermediate chronotype. While these results are to be interpreted with caution, they may imply a protective role of having an early chronotype. It should be noted that this finding is preliminary. Therefore, we consider our results promising but replication in follow-up studies is required [84, 85]. While we demonstrated a significant association between indicators of delayed circadian rhythm and shorter LTL, we found no association between LTL, sleep duration, and insomnia symptoms. As regards associations between sleep duration, insomnia, and LTL, the literature is scant and study populations are diverse in terms of sex, age, obesity, and disease status. Our results concur with the findings of some studies [30, 31, 56, 86] but contradict others [32, 33, 87]. Sleep duration was unrelated to LTL in a sample of healthy women aged between 50 and 65, although sleep quality was inversely related to LTL [31] In another study among healthy women, average sleep duration was not associated with LTL after controlling for BMI, activity, stress, and smoking [86]. Among middle-aged adults, sleep duration was unrelated to LTL, but adequate sleep duration on LTL was beneficial for those 60 years of age or older [30]. Another important study investigated LTL in multiple immune cell subsets in obese adults. Here, it was poorer subjective sleep quality but not sleep duration that was associated with shorter LTL [56]. Unlike our results, short sleep duration was associated with shorter LTL in three studies: in women under 50 years (but not in older women) [32], in men sleeping 5 h or less [33], and in HIV positive adults with sleep duration of at least 7 h, measured with wrist actigraphy [87]. Although some of these findings suggest a relationship between LTL and sleep duration in the same direction (shorter LTL = shorter sleep duration), such an association might be dependent on age, sex, and presence of affective disorder, or may even disappear when a circadian sleep parameter is taken into account. There is less published on insomnia and LTL, barring one negative study in breast cancer survivors with and without insomnia [88]. Another study of older adults found an association between insomnia and LTL in those aged 70–88 years, but not in those aged 60–69 years [34]. As regards mortality associated with sleep dysregulation, both persistently short (≤5 h) and long sleep (≥9 h) were recently associated with an increased risk of all-cause mortality [89]. Night or evening shift work were also associated with increased mortality, compared to day shift [90]. In a population-based cohort, persistent insomnia was associated with an increased risk for all-cause, cardiopulmonary mortality and increased inflammation [91]. The consequences of delayed sleep and circadian dysregulation of the sleep–wake cycle are severe. Our hypothesis that late sleep (indicating disturbed circadian rhythm) would be a significant predictor of short LTL was shown to be true. However, disturbed circadian rhythm did not predict accelerated LTL attrition over the 6 years of the study. On average, mean LTL did shorten over the 6 years of the study, but this was not a substantial change, a finding extensively discussed by Révész et al. in a previous longitudinal NESDA study [10]. Our mean telomere attrition rate of 8 bp per year was lower than that described in a systematic review (32–46 bp per year in longitudinal studies, and 20–30 bp in the larger cross-sectional studies) [8]. In short, this may be due to our relatively young and healthy study sample compared to the longitudinal studies reviewed where samples were smaller, older, and had higher morbidity and mortality rates during follow-up [8]. Our sample also showed a large variation in TL change, with a considerable number of subjects who had stable TL or even lengthened their LTL, similar to those reported in earlier research, as previously outlined by Révész et al. [10]. In this study, we correlated the sleep and circadian variables with the individual difference of LTL change over 6 years, and not the average LTL shortening rate. Yet still we found that only circadian rhythm variables were associated with LTL change. Some questions remain to be answered concerning LTL attrition rate in delayed circadian rhythm. It is possible that we did not find faster LTL attrition because the 6 years of this study were too short a time period to measure any significant difference. Secondly, the timing of LTL attrition may not have been captured in our study. It may have occurred earlier in the life span. It is known that in the first year of life, telomere loss is very rapid [16], perhaps more so in those with circadian dysregulation, a lifelong trait [64]. Thirdly, genetic factors are known to determine cellular decline [7] and baseline LTL (which we corrected for) is an important determinant for LTL attrition [10]. Subjects with delayed circadian rhythm may be born with shorter LTL with less subsequent attrition. LTL has been postulated to be determined well before adulthood, and possibly in utero [92]. The large sample size, wide age range, variety of sleep variables, and longitudinal design are strengths of this study. However, some limitations should be noted. First, we did not include an objective marker for the circadian rhythm, such as melatonin curves or actigraphy, and relied on self-report for chronotype. However, a recent study has suggested that as proxies for determining chronotype, melatonin curves [93] correlate well with MCTQ measurements. Objectively measured sleep (with polysomnography or actigraphy) is also associated with self-reported sleep duration [94]. Second, we did not measure obstructive sleep apnea, which has been linked to cardiovascular disease, metabolic syndrome, and shorter LTL in adults [35, 36]. We did control for obesity, which is associated with both obstructive sleep apnea and shorter LTL [35, 95]. Recent work has shown that the circadian system is implicated in both the severity and duration of the apneic events. At different circadian phases, apnea and hypopnea durations and the number of apneic events differed rhythmically. These results suggest an intriguing link between the circadian system and sleep apnea, which may impact on LTL and should be further studied [96]. Third, we measured TL over 6 years, while ideally studies should track telomere length over a lifetime, hence limiting our ability to draw long-term conclusions. Furthermore, LTL from wave 1 and wave 5 was measured 2 years apart, which could have caused noise between the time points. It was not possible or feasible to rerun the two samples together, but this limitation may be offset by certain disadvantages, such as the unknown impact of storage conditions on the samples. A subset of samples from both time points were rerun together (N = 226, up to eight samples from each of the baseline plates) to estimate and correct for the difference in telomere length estimates derived from two distinct assay runs. This way, we were able to adjust for possible systematic differences. However, these results in this field may be partly influenced by measurement error or other methodological shortcomings, as discussed in the literature [85]. Fourth, TL was measured in leukocytes as opposed to multiple immune cell subsets. It has been shown in vivo that TL attrition occurs at different rates in different cell types, such as T cells, B cells, and monocytes, which are also differentially distributed [97]. This may have biased our findings, making it impossible to distinguish what proportion of the lengtheners in NESDA showed actual lengthening and what proportion was a result of differences in cell composition, or alternatively, measurement error [85]. However, it has been shown that TL in leukocytes, muscle, skin, and fat tissue display similar rates of age-dependent attrition [98], which suggests that results from LTL studies may be applicable to other cell types. Finally, another limitation is that we did not measure telomerase activity (the enzyme that lengthens LTL), and hence cannot comment on the telomere maintenance system, giving further insight in the dynamics of LTL [99]. In conclusion, delayed circadian rhythm rather than insomnia or altered sleep duration seems to be deleterious for cellular aging and therefore perhaps for general health and disease status. While a genetically driven late chronotype may not be entirely modifiable, delayed sleep phase advancement can be achieved with sleep hygiene interventions [100]. Other therapies include timed evening melatonin administration, morning bright light therapy, and chronotherapy [101]. An initial open-label report suggests that delayed sleep phase advances with the use of blue light-blocking glasses in the evening [102]. Such interventions may protect against or even reverse accelerated cellular aging and the morbidity with which it is associated. However, the protective effect of such measures on cellular aging should be studied in a large prospective study. Further longitudinal studies of cellular aging in sleep and circadian dysregulation may clarify if the process of telomeric attrition can be slowed or reversed. Acknowledgments We thank the support staff of PsyQ and VUMc for their encouragement and interest in this study. Funding This study is supported by an NWO-VICI grant (number 91811602) to Prof. Penninx and Dr. Verhoeven for telomere length assaying. The infrastructure for the Netherlands Study of Depression and Anxiety (www.nesda.nl) is funded through the Geestkracht program of the Netherlands Organization for Health Research and Development (Zon-Mw, grant number 10-000-1002) and financial contributions by participating universities and mental health care organizations (VU University Medical Center, GGZ inGeest, Leiden University Medical Center, GGZ Rivierduinen, University Medical Center Groningen, University of Groningen, Lentis, GGZ Friesland, GGZ Drenthe, Rob Giel Onderzoekscentrum). These organizations had no further role in study design, collection, analysis and interpretation of data, writing of the report, and in the decision to submit the paper for publication. Conflict of interest statement Dr. Wynchank has served on the advisory boards of Janssen BV, Novartis, and Eli Lilly (2009–2014) for activities outside of the scope of this paper. Dr. Lamers has received funding from the European Union Seventh Framework Programme (FP7/2007-2013) under grant agreement n° PCIG12-GA-2012-334065 for activities outside of the scope of this paper. Prof. Penninx has received research grants from Johnson & Johnson, Johnson, Boehringer Ingelheim, NWO, BBRMI-NL, NIMH, and the EU-FP7 program (2014–2021) for research in the Netherlands Study of Depression and Anxiety (NESDA), activities outside the scope of this paper. Prof . Beekman has received funds through the speakers’ bureau of Lundbeck and Eli Lilly. Drs. Kooij, Bijlenga, and Verhoeven declare no financial or other relationship relevant to the subject of this article. References 1. Blackburn EH , et al. Human telomere biology: a contributory and interactive factor in aging, disease risks, and protection . Science. 2015 ; 350 ( 6265 ): 1193 – 1198 . Google Scholar Crossref Search ADS PubMed WorldCat 2. von Zglinicki T . Role of oxidative stress in telomere length regulation and replicative senescence . Ann N Y Acad Sci. 2000 ; 908 : 99 – 110 . Google Scholar Crossref Search ADS PubMed WorldCat 3. Aubert G , et al. Telomeres and aging . Physiol Rev. 2008 ; 88 ( 2 ): 557 – 579 . Google Scholar Crossref Search ADS PubMed WorldCat 4. Deng Y , et al. Telomere dysfunction and tumour suppression: the senescence connection . Nat Rev Cancer. 2008 ; 8 ( 6 ): 450 – 458 . Google Scholar Crossref Search ADS PubMed WorldCat 5. Oeseburg H , et al. Telomere biology in healthy aging and disease . Pflugers Arch. 2010 ; 459 ( 2 ): 259 – 268 . Google Scholar Crossref Search ADS PubMed WorldCat 6. Broer L , et al. Meta-analysis of telomere length reveals high heritability, stronger maternal inheritance and a paternal age effect . Eur J Hum Genet. 2013 ; 21 ( 10 ): 1163 – 1168 . Google Scholar Crossref Search ADS PubMed WorldCat 7. Hjelmborg JB , et al. The heritability of leucocyte telomere length dynamics . J Med Genet. 2015 ; 52 ( 5 ): 297 – 302 . Google Scholar Crossref Search ADS PubMed WorldCat 8. Müezzinler A , et al. A systematic review of leukocyte telomere length and age in adults . Ageing Res Rev. 2013 ; 12 ( 2 ): 509 – 519 . Google Scholar Crossref Search ADS PubMed WorldCat 9. Gardner M , et al. ; Halcyon Study Team . Gender and telomere length: systematic review and meta-analysis . Exp Gerontol. 2014 ; 51 : 15 – 27 . Google Scholar Crossref Search ADS PubMed WorldCat 10. Révész D , et al. Baseline biopsychosocial determinants of telomere length and 6-year attrition rate . Psychoneuroendocrinology. 2016 ; 67 : 153 – 162 . Google Scholar Crossref Search ADS PubMed WorldCat 11. Wang H , et al. Associations of alcohol consumption and alcohol flush reaction with leukocyte telomere length in Korean adults . Nutr Res Pract. 2017 ; 11 ( 4 ): 334 – 339 . Google Scholar Crossref Search ADS PubMed WorldCat 12. Puterman E , et al. Determinants of telomere attrition over 1 year in healthy older women: stress and health behaviors matter . Mol Psychiatry. 2015 ; 20 ( 4 ): 529 – 535 . Google Scholar Crossref Search ADS PubMed WorldCat 13. Darrow SM , et al. The association between psychiatric disorders and telomere length: a meta-analysis involving 14,827 persons . Psychosom Med. 2016 ; 78 ( 7 ): 776 – 787 . Google Scholar Crossref Search ADS PubMed WorldCat 14. Haycock PC , et al. Leucocyte telomere length and risk of cardiovascular disease: systematic review and meta-analysis . BMJ. 2014 ; 349 : g4227 . Google Scholar Crossref Search ADS PubMed WorldCat 15. Willeit P , et al. Leucocyte telomere length and risk of type 2 diabetes mellitus: new prospective cohort study and literature-based meta-analysis . PLoS One. 2014 ; 9 ( 11 ): e112483 . Google Scholar Crossref Search ADS PubMed WorldCat 16. Rufer N , et al. Telomere fluorescence measurements in granulocytes and T lymphocyte subsets point to a high turnover of hematopoietic stem cells and memory T cells in early childhood . J Exp Med. 1999 ; 190 ( 2 ): 157 – 167 . Google Scholar Crossref Search ADS PubMed WorldCat 17. Huzen J , et al. Telomere length loss due to smoking and metabolic traits . J Intern Med. 2014 ; 275 ( 2 ): 155 – 163 . Google Scholar Crossref Search ADS PubMed WorldCat 18. Ray S , et al. Cross-talk between circadian clocks, sleep-wake cycles, and metabolic networks: dispelling the darkness . Bioessays. 2016 ; 38 ( 4 ): 394 – 405 . Google Scholar Crossref Search ADS PubMed WorldCat 19. Consensus Conference Panel , Watson NF , et 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 20. Yin J , et al. Relationship of sleep duration with all-cause mortality and cardiovascular events: a systematic review and dose-response meta-analysis of prospective Cohort studies . J Am Heart Assoc. 2017 ; 6 ( 9 ). pii: e005947. doi: https://doi.org/10.1161/JAHA.117.005947. WorldCat 21. Kanagasabai T , et al. Sleep duration and the associated cardiometabolic risk scores in adults . Sleep Health. 2017 ; 3 ( 3 ): 195 – 203 . Google Scholar Crossref Search ADS PubMed WorldCat 22. Kim CE , et al. Association between sleep duration and metabolic syndrome: a cross-sectional study . BMC Public Health. 2018 ; 18 ( 1 ): 720 . Google Scholar Crossref Search ADS PubMed WorldCat 23. Kim MY , et al. Association between sleep duration and stroke prevalence in Korean adults: a cross-sectional study . BMJ Open. 2018 ; 8 ( 6 ): e021491 . Google Scholar Crossref Search ADS PubMed WorldCat 24. Kabat GC , et al. The association of sleep duration and quality with all-cause and cause-specific mortality in the Women’s Health Initiative . Sleep Med. 2018 ; 50 : 48 – 54 . Google Scholar Crossref Search ADS PubMed WorldCat 25. Tan NYQ , et al. Associations between sleep duration, sleep quality and diabetic retinopathy . PLoS One. 2018 ; 13 ( 5 ): e0196399 . Google Scholar Crossref Search ADS PubMed WorldCat 26. Liu Y , et al. Relationship between sleep duration and self-reported health-related quality of life among US adults with or without major chronic diseases, 2014 . Sleep Health. 2018 ; 4 ( 3 ): 265 – 272 . Google Scholar Crossref Search ADS PubMed WorldCat 27. Liu W , et al. Long sleep duration predicts a higher risk of obesity in adults: a meta-analysis of prospective cohort studies . J Public Health (Oxf) . 2018 . doi: https://doi.org/10.1093/pubmed/fdy135. [Epub ahead of print]. PubMed PMID: 30107483. WorldCat 28. Bakour C , et al. Sleep duration trajectories and systemic inflammation in young adults: results from the National Longitudinal Study of Adolescent to Adult Health (Add Health) . Sleep. 2017 ; 40 ( 11 ). doi: https://doi.org/10.1093/sleep/zsx156 WorldCat 29. Maugeri A , et al. Sleep duration and excessive daytime sleepiness are associated with obesity independent of diet and physical activity . Nutrients. 2018 ; 10 ( 9 ). pii: E1219. doi: https://doi.org/10.3390/nu10091219. WorldCat 30. Cribbet MR , et al. Cellular aging and restorative processes: subjective sleep quality and duration moderate the association between age and telomere length in a sample of middle-aged and older adults . Sleep. 2014 ; 37 ( 1 ): 65 – 70 . Google Scholar Crossref Search ADS PubMed WorldCat 31. Prather AA , et al. Shorter leukocyte telomere length in midlife women with poor sleep quality . J Aging Res. 2011 ; 2011 : 721390 . Google Scholar Crossref Search ADS PubMed WorldCat 32. Liang G , et al. Associations between rotating night shifts, sleep duration, and telomere length in women . PLoS One. 2011 ; 6 ( 8 ): e23462 . Google Scholar Crossref Search ADS PubMed WorldCat 33. Jackowska M , et al. Short sleep duration is associated with shorter telomere length in healthy men: findings from the Whitehall II cohort study . PLoS One. 2012 ; 7 ( 10 ): e47292 . Google Scholar Crossref Search ADS PubMed WorldCat 34. Carroll JE , et al. Insomnia and telomere length in older adults . Sleep. 2016 ; 39 ( 3 ): 559 – 564 . Google Scholar Crossref Search ADS PubMed WorldCat 35. Barceló A , et al. Telomere shortening in sleep apnea syndrome . Respir Med. 2010 ; 104 ( 8 ): 1225 – 1229 . Google Scholar Crossref Search ADS PubMed WorldCat 36. Shin C , et al. Association between snoring and leukocyte telomere length . Sleep. 2016 ; 39 ( 4 ): 767 – 772 . Google Scholar Crossref Search ADS PubMed WorldCat 37. Grosbellet E , et al. Circadian desynchronization triggers premature cellular aging in a diurnal rodent . FASEB J. 2015 ; 29 ( 12 ): 4794 – 4803 . Google Scholar Crossref Search ADS PubMed WorldCat 38. Chen WD , et al. The circadian rhythm controls telomeres and telomerase activity . Biochem Biophys Res Commun. 2014 ; 451 ( 3 ): 408 – 414 . Google Scholar Crossref Search ADS PubMed WorldCat 39. Roenneberg T , et al. Life between clocks: daily temporal patterns of human chronotypes . J Biol Rhythms. 2003 ; 18 ( 1 ): 80 – 90 . Google Scholar Crossref Search ADS PubMed WorldCat 40. Adan A , et al. Circadian typology: a comprehensive review . Chronobiol Int. 2012 ; 29 ( 9 ): 1153 – 1175 . Google Scholar Crossref Search ADS PubMed WorldCat 41. American Psychiatric Association . Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5) . 5th ed. Washington DC : American Psychiatric Association ; 2013 . WorldCat COPAC 42. McGinnis GR , et al. Circadian regulation of metabolic homeostasis: causes and consequences . Nat Sci Sleep. 2016 ; 8 : 163 – 180 . Google Scholar PubMed WorldCat 43. Feillet C , et al. Coupling between the circadian clock and cell cycle oscillators: implication for healthy cells and malignant growth . Front Neurol. 2015 ; 6 : 96 . Google Scholar Crossref Search ADS PubMed WorldCat 44. Davies SK , et al. Effect of sleep deprivation on the human metabolome . Proc Natl Acad Sci U S A. 2014 ; 111 ( 29 ): 10761 – 10766 . Google Scholar Crossref Search ADS PubMed WorldCat 45. Penninx BW , et al. ; NESDA Research Consortium . The Netherlands Study of Depression and Anxiety (NESDA): rationale, objectives and methods . Int J Methods Psychiatr Res. 2008 ; 17 ( 3 ): 121 – 140 . Google Scholar Crossref Search ADS PubMed WorldCat 46. Levine DW , et al. Factor structure and measurement invariance of the Women’s Health Initiative Insomnia Rating Scale . Psychol Assess. 2003 ; 15 ( 2 ): 123 – 136 . Google Scholar Crossref Search ADS PubMed WorldCat 47. Levine DW , et al. Reliability and validity of the Women’s Health Initiative Insomnia Rating Scale . Psychol Assess. 2003 ; 15 ( 2 ): 137 – 148 . Google Scholar Crossref Search ADS PubMed WorldCat 48. Skeldon AC , et al. Modelling changes in sleep timing and duration across the lifespan: changes in circadian rhythmicity or sleep homeostasis? Sleep Med Rev. 2016 ; 28 : 96 – 107 . Google Scholar Crossref Search ADS PubMed WorldCat 49. Révész D , et al. Telomere length as a marker of cellular aging is associated with prevalence and progression of metabolic syndrome . J Clin Endocrinol Metab. 2014 ; 99 ( 12 ): 4607 – 4615 . Google Scholar Crossref Search ADS PubMed WorldCat 50. Cawthon RM . Telomere measurement by quantitative PCR . Nucleic Acids Res. 2002 ; 30 ( 10 ): e47 . Google Scholar Crossref Search ADS PubMed WorldCat 51. Wolkowitz OM , et al. Resting leukocyte telomerase activity is elevated in major depression and predicts treatment response . Mol Psychiatry. 2012 ; 17 ( 2 ): 164 – 172 . Google Scholar Crossref Search ADS PubMed WorldCat 52. Verhoeven JE , et al. Depressive and anxiety disorders showing robust, but non-dynamic, 6-year longitudinal association with short leukocyte telomere length . Am J Psychiatry. 2016 ; 173 ( 6 ): 617 – 624 . Google Scholar Crossref Search ADS PubMed WorldCat 53. Müezzinler A , et al. Body mass index and leukocyte telomere length in adults: a systematic review and meta-analysis . Obes Rev. 2014 ; 15 ( 3 ): 192 – 201 . Google Scholar Crossref Search ADS PubMed WorldCat 54. Cappuccio FP , et al. Meta-analysis of short sleep duration and obesity in children and adults . Sleep. 2008 ; 31 ( 5 ): 619 – 626 . Google Scholar Crossref Search ADS PubMed WorldCat 55. Markwald RR , et al. Impact of insufficient sleep on total daily energy expenditure, food intake, and weight gain . Proc Natl Acad Sci U S A. 2013 ; 110 ( 14 ): 5695 – 5700 . Google Scholar Crossref Search ADS PubMed WorldCat 56. Prather AA , et al. Tired telomeres: poor global sleep quality, perceived stress, and telomere length in immune cell subsets in obese men and women . Brain Behav Immun. 2015 ; 47 : 155 – 162 . Google Scholar Crossref Search ADS PubMed WorldCat 57. World Health Organization . Composite International Diagnostic Interview, Core Version 2.1: Interviewer’s Manual . Sydney, Australia : World Health Organization ; 1997 . WorldCat COPAC 58. Rush AJ , et al. The Inventory of Depressive Symptomatology (IDS): psychometric properties . Psychol Med. 1996 ; 26 ( 3 ): 477 – 486 . Google Scholar Crossref Search ADS PubMed WorldCat 59. Beck AT , et al. An inventory for measuring clinical anxiety: psychometric properties . J Consult Clin Psychol. 1988 ; 56 ( 6 ): 893 – 897 . Google Scholar Crossref Search ADS PubMed WorldCat 60. Wardenaar KJ , et al. The structure and dimensionality of the Inventory of Depressive Symptomatology Self Report (IDS-SR) in patients with depressive disorders and healthy controls . J Affect Disord. 2010 ; 125 ( 1–3 ): 146 – 154 . Google Scholar Crossref Search ADS PubMed WorldCat 61. Verhoeven JE , et al. The association of early and recent psychosocial life stress with leukocyte telomere length . Psychosom Med. 2015 ; 77 ( 8 ): 882 – 891 . Google Scholar Crossref Search ADS PubMed WorldCat 62. Twisk JW . Longitudinal data analysis. A comparison between generalized estimating equations and random coefficient analysis . Eur J Epidemiol. 2004 ; 19 ( 8 ): 769 – 776 . Google Scholar Crossref Search ADS PubMed WorldCat 63. Benjamini Y , et al. Controlling the false discovery rate: a practical and powerful approach to multiple testing . J R Stat Soc. 1995 ; 57 ( 1 ): 289 – 300 . WorldCat 64. Jones CR , et al. Genetic basis of human circadian rhythm disorders . Exp Neurol. 2013 ; 243 : 28 – 33 . Google Scholar Crossref Search ADS PubMed WorldCat 65. van Mill JG , et al. Insomnia and sleep duration in a large cohort of patients with major depressive disorder and anxiety disorders . J Clin Psychiatry. 2010 ; 71 ( 3 ): 239 – 246 . Google Scholar Crossref Search ADS PubMed WorldCat 66. Zaslavsky O , et al. Longitudinal changes in insomnia status and incidence of physical, emotional, or mixed impairment in postmenopausal women participating in the Women’s Health Initiative (WHI) study . Sleep Med. 2015 ; 16 ( 3 ): 364 – 371 . Google Scholar Crossref Search ADS PubMed WorldCat 67. Kooij JJ , et al. The circadian rhythm in adult attention-deficit/hyperactivity disorder: current state of affairs . Expert Rev Neurother. 2013 ; 13 ( 10 ): 1107 – 1116 . Google Scholar Crossref Search ADS PubMed WorldCat 68. Abreu T , et al. The bipolarity of light and dark: a review on bipolar disorder and circadian cycles . J Affect Disord. 2015 ; 185 : 219 – 229 . Google Scholar Crossref Search ADS PubMed WorldCat 69. Vadnie CA , et al. Circadian rhythm disturbances in mood disorders: insights into the role of the suprachiasmatic nucleus . Neural Plast. 2017 ; 2017 : 1504507 . Google Scholar Crossref Search ADS PubMed WorldCat 70. Wynchank DS , et al. ADHD, circadian rhythms and seasonality . J Psychiatr Res. 2016 ; 81 : 87 – 94 . Google Scholar Crossref Search ADS PubMed WorldCat 71. Fonseca Costa SS , et al. Impact of the circadian clock on the aging process . Front Neurol. 2015 ; 6 : 43 . Google Scholar Crossref Search ADS PubMed WorldCat 72. Coogan AN , et al. Chronotherapeutics and psychiatry: setting the clock to relieve the symptoms . World J Biol Psychiatry. 2011 ; 12 ( Suppl. 1 ): 40 – 43 . Google Scholar Crossref Search ADS PubMed WorldCat 73. Samulin Erdem J , et al. Mechanisms of breast cancer risk in shift workers: association of telomere shortening with the duration and intensity of night work . Cancer Med. 2017 ; 6 ( 8 ): 1988 – 1997 . Google Scholar Crossref Search ADS PubMed WorldCat 74. Yeh JK , et al. Telomeres and telomerase in cardiovascular diseases . Genes (Basel). 2016 ; 7 ( 9 ). pii: E58. doi: https://doi.org/10.3390/genes7090058. WorldCat 75. Epel ES , et al. Accelerated telomere shortening in response to life stress . Proc Natl Acad Sci U S A. 2004 ; 101 ( 49 ): 17312 – 17315 . Google Scholar Crossref Search ADS PubMed WorldCat 76. Effros RB . The role of CD8 T cell replicative senescence in human aging . Discov Med. 2005 ; 5 ( 27 ): 293 – 297 . Google Scholar PubMed WorldCat 77. Verhoeven JE , et al. Cellular aging in depression: permanent imprint or reversible process?: an overview of the current evidence, mechanistic pathways, and targets for interventions . Bioessays. 2014 ; 36 ( 10 ): 968 – 978 . Google Scholar Crossref Search ADS PubMed WorldCat 78. Irwin MR , et al. Sleep health: reciprocal regulation of sleep and innate immunity . Neuropsychopharmacology. 2017 ; 42 ( 1 ): 129 – 155 . Google Scholar Crossref Search ADS PubMed WorldCat 79. von Zglinicki T . Oxidative stress shortens telomeres . Trends Biochem Sci. 2002 ; 27 ( 7 ): 339 – 344 . Google Scholar Crossref Search ADS PubMed WorldCat 80. Révész D , et al. Dysregulated physiological stress systems and accelerated cellular aging . Neurobiol Aging. 2014 ; 35 ( 6 ): 1422 – 1430 . Google Scholar Crossref Search ADS PubMed WorldCat 81. Irwin MR , et al. Sleep deprivation and activation of morning levels of cellular and genomic markers of inflammation . Arch Intern Med. 2006 ; 166 ( 16 ): 1756 – 1762 . Google Scholar Crossref Search ADS PubMed WorldCat 82. Vgontzas AN , et al. Adverse effects of modest sleep restriction on sleepiness, performance, and inflammatory cytokines . J Clin Endocrinol Metab. 2004 ; 89 ( 5 ): 2119 – 2126 . Google Scholar Crossref Search ADS PubMed WorldCat 83. Meier-Ewert HK , et al. Effect of sleep loss on C-reactive protein, an inflammatory marker of cardiovascular risk . J Am Coll Cardiol. 2004 ; 43 ( 4 ): 678 – 683 . Google Scholar Crossref Search ADS PubMed WorldCat 84. Susser E , et al. Non-dynamic association of depressive and anxiety disorders with leukocyte telomere length? Am J Psychiatry. 2016 ; 173 ( 11 ): 1147 . Google Scholar Crossref Search ADS PubMed WorldCat 85. Verhoeven JE , et al. Unresolved issues in longitudinal telomere length research: response to Susser et al . Am J Psychiatry. 2016 ; 173 ( 11 ): 1147 – 1149 . Google Scholar Crossref Search ADS PubMed WorldCat 86. Parks CG , et al. Telomere length, current perceived stress, and urinary stress hormones in women . Cancer Epidemiol Biomarkers Prev. 2009 ; 18 ( 2 ): 551 – 560 . Google Scholar Crossref Search ADS PubMed WorldCat 87. Lee KA , et al. Telomere length is associated with sleep duration but not sleep quality in adults with human immunodeficiency virus . Sleep. 2014 ; 37 ( 1 ): 157 – 166 . Google Scholar Crossref Search ADS PubMed WorldCat 88. Garland SN , et al. A nested case-controlled comparison of telomere length and psychological functioning in breast cancer survivors with and without insomnia symptoms . Rejuvenation Res. 2014 ; 17 ( 5 ): 453 – 457 . Google Scholar Crossref Search ADS PubMed WorldCat 89. Soh AZ , Chee MWL , Yuan JM , Koh WP . Sleep lengthening in late adulthood signals increased risk of mortality . Sleep. 2018 ; 41 ( 3 ). doi: https://doi.org/10.1093/sleep/zsy005 WorldCat 90. Jørgensen JT , et al. Shift work and overall and cause-specific mortality in the Danish nurse cohort . Scand J Work Environ Health. 2017 ; 43 ( 2 ): 117 – 126 . Google Scholar Crossref Search ADS PubMed WorldCat 91. Parthasarathy S , et al. Persistent insomnia is associated with mortality risk . Am J Med. 2015 ; 128 ( 3 ): 268 – 275.e2 . Google Scholar Crossref Search ADS PubMed WorldCat 92. Factor-Litvak P , et al. The importance of early life studies of telomere attrition . Paediatr Perinat Epidemiol. 2015 ; 29 ( 2 ): 144 – 145 . Google Scholar Crossref Search ADS PubMed WorldCat 93. Kantermann T , et al. Comparing the Morningness-Eveningness Questionnaire and Munich ChronoType Questionnaire to the Dim Light Melatonin Onset . J Biol Rhythms. 2015 ; 30 ( 5 ): 449 – 453 . Google Scholar Crossref Search ADS PubMed WorldCat 94. Lauderdale DS , et al. Self-reported and measured sleep duration: how similar are they? Epidemiology. 2008 ; 19 ( 6 ): 838 – 845 . Google Scholar Crossref Search ADS PubMed WorldCat 95. Riestra P , et al. Obstructive sleep apnea risk and leukocyte telomere length in African Americans from the MH-GRID study . Sleep Breath. 2017 ; 21 ( 3 ): 751 – 757 . Google Scholar Crossref Search ADS PubMed WorldCat 96. Butler MP , et al. The Circadian system contributes to apnea lengthening across the night in obstructive sleep apnea . Sleep. 2015 ; 38 ( 11 ): 1793 – 1801 . Google Scholar Crossref Search ADS PubMed WorldCat 97. Lin Y , et al. Age-associated telomere attrition of lymphocytes in vivo is co-ordinated with changes in telomerase activity, composition of lymphocyte subsets and health conditions . Clin Sci (Lond). 2015 ; 128 ( 6 ): 367 – 377 . Google Scholar Crossref Search ADS PubMed WorldCat 98. Daniali L , et al. Telomeres shorten at equivalent rates in somatic tissues of adults . Nat Commun. 2013 ; 4 : 1597 . Google Scholar Crossref Search ADS PubMed WorldCat 99. Blackburn EH . Switching and signaling at the telomere . Cell. 2001 ; 106 ( 6 ): 661 – 673 . Google Scholar Crossref Search ADS PubMed WorldCat 100. Auger RR , et al. Clinical Practice Guideline for the Treatment of Intrinsic Circadian Rhythm Sleep-Wake Disorders: Advanced Sleep-Wake Phase Disorder (ASWPD), Delayed Sleep-Wake Phase Disorder (DSWPD), Non-24-Hour Sleep-Wake Rhythm Disorder (N24SWD), and Irregular Sleep-Wake Rhythm Disorder (ISWRD). An update for 2015: an American Academy of Sleep Medicine Clinical Practice Guideline . J Clin Sleep Med. 2015 ; 11 ( 10 ): 1199 – 1236 . Google Scholar Crossref Search ADS PubMed WorldCat 101. Kim MJ , et al. Circadian rhythm sleep disorders . J Clin Outcomes Manag. 2013 ; 20 ( 11 ): 513 – 528 . Google Scholar PubMed WorldCat 102. Esaki Y , et al. Wearing blue light-blocking glasses in the evening advances circadian rhythms in the patients with delayed sleep phase disorder: An open-label trial . Chronobiol Int. 2016 ; 33 ( 8 ): 1037 – 1044 . Google Scholar Crossref Search ADS PubMed WorldCat © Sleep Research Society 2019. Published by Oxford University Press [on behalf of the Sleep Research Society]. All rights reserved. For permissions, please e-mail: [email protected] This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)