OPENACCESS Citation: Caporali JFdM, Labanca L, Florentino KR, Background Souza BO, Utsch GoncËalves D (2018) Intrarater and interrater agreement and reliability of vestibular The vestibular evoked myogenic potential triggered by galvanic vestibular stimulation (gal- evoked myogenic potential triggered by galvanic vanic-VEMP) has been used to assess the function of the vestibulospinal motor tract and is vestibular stimulation (galvanic-VEMP) for HTLV-1 a candidate biomarker to predict and monitor the human T-cell lymphotropic virus type 1 associated myelopathy testing. PLoS ONE 13(9): (HTLV-1) associated myelopathy (HAM). This study determined the agreement and reliabil- e0204449. https://doi.org/10.1371/journal. pone.0204449 ity of this exam. Editor: Zheng Xing, University of Minnesota College of Veterinary Medicine, UNITED STATES Methods Received: May 10, 2018 Galvanic-VEMP was performed in 96 participants, of which 24 patients presented HAM, 27 Accepted: September 7, 2018 HTLV-1-asymptomatic carriers, and 45 HTLV-1-negative asymptomatic controls. Galvanic vestibular stimulation was achieved by passing a binaural and bipolar current at a 2 milliam- Published: September 27, 2018 peres (mA) intensity for 400 milliseconds (ms) between the mastoid processes. Galvanic- Copyright:© 2018 Caporali et al. This is an open VEMP electromyographic wave responses of short latency (SL) and medium latency (ML) access article distributed under the terms of the Creative Commons Attribution License, which were recorded from the gastrocnemius muscle. Intrarater (test-retest) and interrater (two permits unrestricted use, distribution, and independent examiners) agreement and reliability were assessed by standard error of mea- reproduction in any medium, provided the original surement (SEM), coefficient of repeatability (CR), intraclass correlation coefficient (ICC), author and source are credited. and Kappa coefficient. Data Availability Statement: All relevant data are within the paper and its Supporting Information Results files. In the total sample (n = 96), SL and ML medians were 56 ms (IQR 52±66) and 120 ms (IQR Funding: This work was supported by Pro-Reitoria de Pesquisa da Universidade Federal de Minas 107±130), respectively. The intrarater repeatability measures for SL and ML were, respec- Gerais (PRPQ/UFMG) - https://www.ufmg.br/prpq/; tively: SEM of 6 and 8 ms; CR of 16 and 22 ms; ICC of 0.80 (p<0.001) and 0.91 (p<0.001); Ä Á FundacËao de Amparo a Pesquisa de Minas Gerais and a Kappa coefficient of 0.53 (p<0.001) and 0.82 (p<0.001). The interrater reproducibility (FAPEMIG) - http://www.fapemig.br/en/; Conselho Nacional de Desenvolvimento CientõÂfico e measures for SL and ML were, respectively: SEM of 3 and 10 ms; CR of 8 and 27 ms; ICC PLOS ONE | https://doi.org/10.1371/journal.pone.0204449 September 27, 2018 1 / 13 Agreement and reliability of galvanic-VEMP in HAM Tecnolo Âgico (CNPq) - http://cnpq.br/; CoordenacËaÄo of 0.95 (p<0.001) and 0.86 (p<0.001); and a Kappa coefficient of 0.77 (p<0.001) and 0.88 de AperfeicËoamento de Pessoal de NõÂvel Superior (p<0.001). (CAPES/COFECUB) - http://www.capes.gov.br/. The funders had no role in study design, data Conclusion collection and analysis, decision to publish, or preparation of the manuscript. Galvanic-VEMP is a reliable and reproducible method to define the integrity of the vestibu- Competing interests: The authors have declared lospinal tract. Longitudinal studies will clarify its validity in the clinical context, aimed at that no competing interests exist. achieving an early diagnosis and the monitoring of HAM. Introduction The vestibular evoked myogenic potential triggered by galvanic vestibular stimulation (gal- vanic-VEMP) is an exam that assesses the function of the vestibulospinal motor tract  and has been used as an auxiliary tool in spinal cord diseases caused by tumor, trauma, and infec- tion [2±5]. In human T-cell lymphotropic virus type 1-associated myelopathy (HAM), gal- vanic-VEMP disclosed an electrophysiological altered response that ranged from a delayed latency among the asymptomatic carriers to a complete absence of response in those with established myelopathy . HAM is an insidious and irremissible meningomyelitis that affects 1±4% [6±10] of the 5±20 million people infected with HTLV-1 worldwide [11, 12]. This neurologic disease is more fre- quent in women than in men (2:1 to 3:1), and its symptoms onset is most often found in the fifth decade of life [13±17]. The first symptoms of HAM are weakness of the lower limbs, lum- bar pain, dizziness, and urinary and sexual impairments [17, 18±21]. Sensory changes may also be an early complaint . The progression is characterized by spastic paraparesis and lower limb hyperreflexia, Babinski sign, impaired vibratory sensitivity, positive Romberg test, and abnormal gait. After 10 years of symptoms, 20±50% of individuals with HAM become wheelchair-dependent [19, 23±25] HAM occurs due to an unbalanced inflammatory response to HTLV-1 infection [26±29]. The disease has a biphasic pathology pattern  in which the inflammatory phase is followed by an atrophic stage. The therapeutic strategies have been developed based mainly on inflam- mation control in the first phase, since irreversible neuron damage characterizes the later peri- ods of the disease. Thus, the earlier the diagnosis, the better the chance of obtaining a good response to treatment . In this scenario, along with the immunologic molecules, the neuro- physiology exams, such as galvanic-VEMP, are candidate biomarkers to predict HAM in its subclinical stage and monitor the disease activity during treatment . Vestibular evoked myogenic potential (VEMP) is an electrophysiological test in which a stimulus is offered to the vestibular system, triggering several interconnected motor responses comprising ocular and postural muscles. In VEMP triggered by galvanic vestibular stimulation (GVS), an electric stimulus is applied to the labyrinth organs through surface electrodes posi- tioned behind the ears. The stimulus generates a dipole between the labyrinths, which is inter- preted by the central nervous system (CNS) as a true head movement [1, 32]. Cathodal galvanic stimuli depolarize, whereas anodal stimuli hyperpolarize afferent vestibular fibers [33, 34]. The unanticipated vestibular stimulus elicits a protective reflex in all muscles engaged in posture control, leading the body to temporarily sway toward the anode. The motor reflexes that are evoked to maintain the postural equilibrium can be captured by surface electromyog- raphy (EMG) in the body muscles involved in one's posture. Galvanic-VEMP evaluates the brainstem function, as other VEMPs do , and further assesses the vestibulospinal motor tract. The chosen muscle to record the electrophysiological sign defines the tested level of the PLOS ONE | https://doi.org/10.1371/journal.pone.0204449 September 27, 2018 2 / 13 Agreement and reliability of galvanic-VEMP in HAM spine. The assessment of the spine is performed by recording the response in the sternocleido- mastoid muscle for the cervical level, the trunk (erectors spinae) muscles for the thoracic level, and the lower limb muscles, such as soleus or gastrocnemius, for the lumbar spinal level [1, 32]. Graphically, the galvanic-VEMP response taken from gastrocnemius muscle is character- ized by a biphasic wave, with a short-latency (SL) response around 60 ms, followed by a medium-latency (ML) response around 100 ms [1, 32, 36, 37]. A change in the waveform and the delay or the absence of any of the waves are considered altered results [2±5]. Galvanic-VEMP proved to be quite accurate in identifying spinal cord impairments based on the ROC curve in individuals with myeloradiculopathy caused by Schistosoma mansoni . However, to the best of our knowledge, the reliability and agreement of this exam have not been checked properly in prior studies, and this assessment is essential when the exam is used for diagnostic and monitoring purposes. This study determined the interrater and the intrara- ter agreement and reliability of galvanic-VEMP in individuals with HAM, asymptomatic HTLV-1 infection and controls. The concepts and importance of agreement and reliability The estimates of agreement (repeatability and reproducibility) and reliability are used to evalu- ate the measurement error of a quantity and its impact on the interpretation of measurements. Any measurement is susceptible to various types of errors that can cause the measured value to be different from the real value. Repeatability of the results (of a measurement) is the approxi- mation between the results of successive measurements of a quantity carried out under the same measurement conditions . These conditions are referred to as repeatability condi- tions, which include: the same measurement procedure; the same examiner (or rater); the same measuring instrument, used under the same conditions; the same place; and the repeti- tion should be performed within a short time. Reproducibility of the results (of a measure- ment) is the approximation between the results of the measurements of a quantity carried out with changes in the measurement conditions . Changes considered include the principle and method of measurement, the examiner, the instrument, the reference standard, the loca- tion, the conditions of use, and the time. Repeatability and reproducibility are grouped together in the concept of agreement, i.e., how far apart the repeated measures of the same quantity are. Reliability, on the other hand, correlates the magnitude of the measurement error of the repeated measurements with the inherent, error-free variability among individuals. Therefore, it depends on the variability of the population. If reliability is high, measurement errors are small relative to the actual differences among individuals in the population, and the method can differentiate well despite the measurement error . The measurement error of the repeated measurements may be due to intraindividual bio- logical variability, intrinsic variability to the measuring instrument, variability between one instrument and another, circumstances in which the measurement is performed, intrarater variability (the same examiner gives two different judgments at two different times) and inter- rater variability (one examiner gives a different judgment from the other examiner). By mea- suring and quantifying the measurement error (through repeatability, reproducibility, and reliability estimates), it is possible to judge whether this error is acceptable within the context in which the measurement is to be applied . Methods Ethical statement This study follows the ethical principles expressed in the Declaration of Helsinki . It was approved by the Research Ethics Committees of Universidade Federal de Minas Gerais PLOS ONE | https://doi.org/10.1371/journal.pone.0204449 September 27, 2018 3 / 13 Agreement and reliability of galvanic-VEMP in HAM Fig 1. Vestibular-evoked myogenic potential triggered by galvanic vestibular stimulation procedure. The standing position of the patient (barefoot on a hard flat surface with eyes closed, feet close together and body leaning forward in order to cause the gastrocnemius muscle contraction); the equipment used for stimulus generation (a); the electrode positions for GVS (b); the electrode position for electromyography on the gastrocnemius muscle (c); the equipment for signal processing (d); and the laptop (e) connected to (a) and (d). https://doi.org/10.1371/journal.pone.0204449.g001 (UFMG) and of the Hemominas Blood Transfusion Agency, in Brazil, under the protocol numbers, respectively, of 266/05 and 131. All participants gave their written informed consent. The individual in Fig 1 has given written informed consent (as outlined in PLoS consent form) for this photograph to be published. Study design and setting This is a repeatability and reproducibility study about the use of galvanic-VEMP to test HAM, which was conducted between 2014 and 2016 in the UFMG School of Medicine, Belo Hori- zonte, Brazil. Subjects and sample size The individuals were recruited from the open cohort of the Interdisciplinary HTLV Research Group (GIPH), formed in 1997, which has been following the individuals from 1997 to the present day [41±44]. The inclusion criteria for the infected individuals were positive serology in Enzyme-linked Immunosorbent Assay and Western Blot, as well as positive Protein Chain Reaction, for HTLV-1. The HTLV-1 infected individuals are divided into asymptomatic carri- ers (AC) and individuals with HAM, according to the revised diagnostic criteria by Castro- Costa et al. . The controls tested negative for HTLV-1. The exclusion criteria for all groups were: under 18 years of age, uncontrolled acute or chronic diseases, HIV coinfection, sus- pected or confirmed pregnancy, metallic prosthesis, being unable to stand in the upright posi- tion during the galvanic-VEMP procedures, neurologic diagnosis such as history of stroke, CNS tumor, CNS infection (other than HTLV-1 infection), vitamin B12 deficiency, spinal cord diseases (other than HAM), diabetic neuropathy, migraine, and, finally, vestibular dis- eases such as Benign Paroxysmal Positional Vertigo (BPPV), vestibular neuritis and Me Ânière's disease. All the subjects were submitted to a clinical interview and neurological examination before undergoing galvanic-VEMP procedures. PLOS ONE | https://doi.org/10.1371/journal.pone.0204449 September 27, 2018 4 / 13 Agreement and reliability of galvanic-VEMP in HAM Considering the study by Shoukri et al. (2004) , for a repeatability study to achieve reli- able results with two repeated measurements, a significance level of 5%, and a test power of 80%, a minimum sample of 86 participants is necessary. In the present study, the total sample included 96 participants. Since the interest variables (SL and ML) were collected from both legs of each individual, a randomization, performed by the statistical computer program, was performed to select which leg of each participant would be part of the analyses. Measurement process Technical aspects and protocol of the galvanic-VEMP. The galvanic-VEMP equipment used for stimulation and recording was the EvP4 / ATCPlus model (Contronic Ltda., Pelotas, Bra- zil) connected to a battery-powered portable computer. Self-adhesive surface electrodes, 3 centime- ters (cm) in diameter (model CF3200-Valutrode, Axelgaard, Fallbrook, CA, USA) were positioned on the participant's mastoid processes, anode on one side and cathode on the other, offering bipo- lar binaural stimulation. The stimulus was generated by a constant current stimulator, consisting of a single-phase, rectangular, direct current with an intensity of 2 mA for 400 ms [3±5]. Each examination consisted of 30 stimulations, 15 of which were performed with the anode in the right ear and 15 with anode in the left ear. Intervals between the stimuli were randomized between 4 and 6 seconds. The test was then immediately repeated once to evaluate repeatability. To perform the test, the subjects stood on a hard flat surface with their eyes closed, barefoot, with their body slightly bent forward, promoting contraction of the gastrocnemius muscle. Participants were instructed to turn their heads approximately 90Ê in the sagittal plane to the contralateral side of the lower limb from which the EMG signals would be drawn . The EMG activity was recorded by a pair of self-adhesive electrodes (model 2223BRQ, 3M, Saint Paul, MN, USA) placed on the medial head of the gastrocnemius muscle, and with their centers approximately 5 cm distant from one another. The reference electrode was placed on the back of the thigh, approximately 5cm above the recording electrode (Fig 1). Galvanic- VEMP was performed first on the left lower limb and then on the right lower limb. Performing the complete examination of a patient lasted about 20 minutes on average. The EMG signals were measured, rectified, filtered between 10 Hertz (Hz) and 1000 Hz, and scanned at a sampling frequency of 5 kHz, using one register channel. The data were col- lected during a period of 500 ms, beginning at 100 ms before the galvanic stimulus . The EMG responses to 15 consecutive stimuli with the same polarity setting were averaged, result- ing in a final tracing. The tracings could be observed online during the execution of the exam and were recorded for further analysis by the examiners, under blindness as to the group to which the participant belonged. Definition of the galvanic-VEMP variables. The measured variables were the latencies of each of the two components of the galvanic-VEMP wave. The rater analysis was based on pre- viously described criteria [36, 47], i.e.: SL is the wave starting at about 60 ms and ML is the fol- lowing wave, with opposite polarity, starting at about 100 ms. SL and ML reverse with the inversion of stimulus polarity. The responses were considered to be changed if they were delayed, absent, or with abnormal tracing. Delay was considered when response onset was later than 2 standard-deviations over the mean found in healthy controls , i.e., 63 ms for SL and 132 ms for ML. Fig 2 illustrates the normal, delayed, and abnormal tracing patterns. For a more detailed description of this test, go to dx.doi.org/10.17504/protocols.io.nxbdfin. Statistical analysis, agreement and reliability parameters This study analyzed the agreement and reliability of measurements of the EMG responses of galvanic-VEMP. For each of the two EMG responses (SL and ML), the estimates were PLOS ONE | https://doi.org/10.1371/journal.pone.0204449 September 27, 2018 5 / 13 Agreement and reliability of galvanic-VEMP in HAM Fig 2. Normal, delayed, and abnormal response patterns in vestibular-evoked myogenic potential triggered by galvanic vestibular stimulation (galvanic-VEMP). (A) Normal electromyographic (EMG) response recorded from the gastrocnemius muscle. The black line indicates the trace with the cathode on the right and the anode on the left, whereas the gray line indicates the opposite stimulation polarity. SL (~50 ms) and ML (~100 ms). (B) Delayed EMG responses. SL ~80 ms and ML ~150 ms. (C) Absent EMG response, no SL and no ML. https://doi.org/10.1371/journal.pone.0204449.g002 calculated based on two measurements done (a) in repeatability conditions, i.e., two immedi- ately repeated measurements in the same patient, analyzed by the same examiner±test-retest repeatability or intrarater repeatability (b) by two experienced independent examiners, blinded to the clinical condition of the participant±interrater reproducibility. The calculated agreement parameters included: standard error of measurement (SEM) = SD of the paired differences / 2, within-individual variation, limits of agreement, and the coefficient of repeatability (CR) = SD of paired differences x 1.96. SEM and CR represent the measurement error intrinsic to the measurement method and take into consideration the within-subject variation. The CR shows the expected variation of the results for 95% of the repeated measures, which is expressed in the same unit of measure. It is also known as the Smallest Real Difference (SRD). The reliability of the test was assessed by the intraclass correlation coefficient (ICC) and the Kappa coefficient. ICC indicates good reliability when equal to or higher than 0.70, [39, 48]. The Kappa coefficient was considered acceptable if greater than 0.6 . The Kappa coefficient was calculated after the categorization of the variables into normal, delayed, and absent, fol- lowing the criteria described in the previous section. The database was fed with double input using the EpiData 3.0 program (EpiData Data Entry, Data Management and basic Statistical Analysis System, EpiData Association, 2000± 2008, Odense, Denmark). The SPSS 15.0 program (SPSS, Inc., Chicago, IL, USA) was used to describe the variables and conduct statistical analysis. Continuous variables of interest were tested for normality with Shapiro-Wilk test and showed a non-normal distribution. The non- parametric Kruskal-Wallis test was used to compare continuous variables between groups. For categorized variables, a chi-square test (Pearson's or Fisher's) was used. The significance level was 5%. Results From a total of 100 individuals selected for the study, four were excluded: one reported a metal plaque implant in the skull, one had HIV infection, and we lost the galvanic-VEMP tracings in two patients due to interference in the software device. Of the 96 participants who completed the entire protocol, 45 were controls 27 were asymptomatic HTLV-1 carriers (AC), and 24 were individuals with HAM. The mean age was 55, 58, and 58 years in the control, AC, and PLOS ONE | https://doi.org/10.1371/journal.pone.0204449 September 27, 2018 6 / 13 Agreement and reliability of galvanic-VEMP in HAM HAM groups, respectively, with no statistical difference (p = 0.552). The proportion of male gender was 40, 41, and 29 percent in the control, AC, and HAM groups, respectively, with no statistical difference (p = 0.624). The comparison of the continuous and categorized (normal, delayed, or absent) galvanic-VEMP responses (SL and ML) are shown in Table 1. In the HAM group, the SL showed a tendency toward higher values (p = 0.089) and was more frequently delayed and absent (p = 0.067). The ML was delayed and more frequently absent in the HAM group when compared to the AC and control groups (p<0.001). Agreement and reliability of the galvanic-VEMP responses (SL and ML) The agreement and the reliability measures were acceptable in intrarater (test-retest) and inter- rater calculations for SL and ML in the total sample and in each group (Tables 2±5). There was no clinically relevant difference of these parameters between the groups. Discussion Galvanic-VEMP has been used to investigate the postural balance in normal individuals for more than four decades [50±55], and in recent years this exam has been considered to be an auxiliary tool for the diagnosis of myelopathies [2±5]. The accuracy of galvanic-VEMP has been described , but not the agreement and reliability, which are equally important to vali- date a diagnostic tool. The present study evaluated, for the first time, the agreement and the reliability of galvanic- VEMP between two repeated measurements (intrarater test-retest) and between measure- ments made by two examiners (interrater). Galvanic-VEMP is a test that measures the time, in milliseconds, of a postural reflex from its generation by electric stimulation of the vestibular nuclei until its muscular response, which is recorded by surface electromyography. Therefore, the response can be recorded only from the muscles involved in the balance control. Several factors can lead to a variability / measurement error of galvanic-VEMP latencies: 1) the circadian biological variations of individuals; 2) possible intrinsic instabilities of the devices; 3) the variability in the interpretation of the examiner when analyzing the electromy- ography curve; 4) the variability of interpretation of different examiners; 5) the variability of sensory perception, such as vision, hearing, and proprioception, which influences the EMG responses. Aimed at reducing external bias, the test is conducted in a silent environment, with a grounded electrical grid, and the patient must be able to maintain a correct posture during the exam, with eyes closed [36, 45, 56±58]. Table 1. Galvanic-VEMP variables (SL and ML): Comparison between groups. Variable HTLV-1 negative controls (n = 45) Asymptomatic carriers (n = 27) HAM (n = 24) p-value SL Median (IQR) 56 (53±64) 53 (52±63) 63 (56±75) 0.089 normal 29(64.4%) 14 (51.9%) 8 (33.3%) 0.067 delayed 10 (22.2%) 4 (14.8%) 8 (33.3%) absent 6 (13.4%) 9 (33.3%) 8 (33.3%) ML Median (IQR) 114 (105±126) 116 (101±130) 136 (124±144) <0.001 normal 42 (93.3%) 17 (63%) 8 (33.3%) <0.001 delayed 2 (4.4%) 4 (14.8%) 11 (45.8%) absent 1 (2.3%) 6 (22.2%) 5 (20.9%) Notes: SL delay: > 63ms; ML delay: > 132ms; IQR: interquartile range. Statistical tests: Kruskal-Wallis for continuous SL and ML values; Qui-square for categorized SL and ML. statistically different group. https://doi.org/10.1371/journal.pone.0204449.t001 PLOS ONE | https://doi.org/10.1371/journal.pone.0204449 September 27, 2018 7 / 13 Agreement and reliability of galvanic-VEMP in HAM Table 2. Intrarater (test-retest) and interrater agreement and reliability measures of galvanic-VEMP variables (SL and ML) in the total sample (n = 96). Variable SEM CR ICC P value Kappa P value Intrarater SL 6 16 0.803 <0.001 0.533 <0.001 ML 8 22 0.913 <0.001 0.829 <0.001 Interrater SL 3 8 0.953 <0.001 0.769 <0.001 ML 10 27 0.863 <0.001 0.884 <0.001 SEM: standard error of measurement. CR: coefficient of repeatability. ICC: intraclass correlation coefficient. https://doi.org/10.1371/journal.pone.0204449.t002 A practical way for clinicians to evaluate the error of measurement (both random and sys- tematic errors) is by observing the CR, which is expressed in the same unit as the measurement tool (in milliseconds, in our case). It is expected that the absolute difference between two mea- surements on a subject differs by no more than the repeatability coefficient in 95% of the occa- sions. For this reason, the CR is also referred to as the Smallest Real Difference (SRD) . In our results, CR was 16 ms for SL and 22 ms for ML in intrarater repeated measures, and 8 ms for SL and 27 ms for ML in interrater repeated measures, meaning that latency differences larger than these values are due to real differences and not measurement errors, considering a 95% probability. These estimates are important to be considered when the method is going to be used to detect the real difference within-subject in the disease progression or therapeutic response, which are, for instance, the proposed uses for galvanic-VEMP. The CR is calculated based on the standard error of measurement (SEM). SEM alone can be interpreted when there is an established concept of the differences that are clinically relevant. However, regarding the variables SL and ML, there is still no conclusion about how large the difference must be in order to be considered a significant change in the exam. In our study, SEM was 6 ms (intrara- ter) and 3 ms (interrater) for SL and 8 ms (intrarater) and 10 ms (interrater) for ML. As far as we know, the only available reference parameters are from two cross sectional studies. Cunha et al. found that SL was 67±8 ms in the group with HAM and 55±4 ms in the controls±a differ- ence of 12 ms between the means, while ML was 130±3 ms in HAM and 112±10 in controls±a difference of 18 ms between the means . In patients with schistosomal myeloradiculopathy the SL was 64 ms (60/74) and 59 ms (56/61) in the controls±a difference of 5 ms between the medians, while the ML was 138 ms (122/153) in patients and 109 ms (106/121) in controls± showing a larger difference of 29 ms . Longitudinal studies with larger samples are war- ranted to define the clinically relevant change in SL and ML when monitoring HAM and other myelopathies. For the risk prediction and the diagnosis, on the other hand, it is essential to determine if, despite the error, the method can distinguish the individuals, taking into consideration the variability between people. This aspect is linked to reliability and is assessed by the intra-class correlation coefficient (ICC) [46, 47]. A good ICC is considered to be 0.70, which means Table 3. Intrarater (test-retest) and interrater agreement and reliability measures of galvanic-VEMP variables (SL and ML) in HTLV-1 negative controls (n = 45). Variable SEM CR ICC P value Kappa P value Intrarater SL 5 14 0.688 <0.001 0.494 <0.001 ML 7 19 0.860 <0.001 0.567 <0.001 Interrater SL 2 5 0.963 <0.001 0.587 <0.001 ML 12 33 0.752 <0.001 0.395 <0.001 SEM: standard error of measurement. RC: repeatability coefficient. ICC: intraclass correlation coefficient. https://doi.org/10.1371/journal.pone.0204449.t003 PLOS ONE | https://doi.org/10.1371/journal.pone.0204449 September 27, 2018 8 / 13 Agreement and reliability of galvanic-VEMP in HAM Table 4. Intrarater (test-retest) and interrater agreement and reliability measures of galvanic-VEMP variables (SL and ML) in HTLV-1 asymptomatic carriers (n = 27). Variable SEM CR ICC P value Kappa P value Intrarater SL 6 16 0.694 0.014 0.567 <0.001 ML 8 23 0.923 <0.001 0.861 <0.001 Interrater SL 4 12 0.861 <0.001 0.878 <0.001 ML 4 12 0.946 <0.001 0.749 <0.001 SEM: standard error of measurement. CR: coefficient of repeatability. ICC: intraclass correlation coefficient. https://doi.org/10.1371/journal.pone.0204449.t004 that at least 70% of the variability in measurements is estimated to be due to real differences in the values, with the remaining 30% or less being due to errors in the measurement process [46, 47]. In our study, galvanic-VEMP proved to be reliable, with very good ICCs: 0.803 (SL) and 0.913 (ML) for intrarater measurements pairs and 0.953 (SL) and 0.863 (ML) for interrater pairs. The agreement and the reliability parameters described above are suitable for continuous variables. To include galvanic-VEMP in the battery to test the postural reflex, the responses must be categorized into normal, delayed, and absent (criteria described in the methods sec- tion). For the categorized results, we calculated the Kappa coefficient, which proved to be quite satisfactory for ML in intrarater and interrater analyses (greater than 0.80). For SL, the interra- ter Kappa was good (0.769), but for intrarater repeated measurements, it was not clinically acceptable (0.533). The Kappa was under the acceptance level especially in the control group. However, one limitation is that the normality cutoffs used in our study were based on the results of normal individuals from a study with a sample of 13 subjects , i.e. we considered the normality cutoff as being 2 standard-deviations over the mean found in this healthy small group. The ROC curve of the galvanic-VEMP showed good results (0.814 for SL, p = 0.001, and 0.861 for ML, p<0.001) in a study with schistosomal myeloradiculopaty , but the cutoff values of SL and ML were not described. Therefore, future studies on accuracy for definition of normality cutoffs should be conducted. Conclusion Galvanic-VEMP proved to have good accuracy , and the present results also show good repeatability, reproducibility, and reliability. For the time being, there is still no definition if a change in galvanic-VEMP in HTLV-1-asymptomatic carriers is a biomarker of HAM. A longi- tudinal study will fill this knowledge gap. We conclude that this test can be considered for the follow-up of HAM, since it proved to be a reliable low-cost, easy to perform, and safe tool to test the postural reflex. Table 5. Intrarater (test-retest) and interrater agreement and reliability measures of galvanic-VEMP variables (SL and ML) in individuals with HAM (n = 24). Variable SEM CR ICC P value Kappa P value Intrarater SL 7 19 0.850 0.001 0.438 0.002 ML 10 29 0.861 <0.001 0.869 <0.001 Interrater SL 3 8 0.978 <0.001 0.813 <0.001 ML 9 25 0.808 0.001 0.509 <0.001 SEM: standard error of measurement. CR: coefficient of repeatability. ICC: intraclass correlation coefficient. https://doi.org/10.1371/journal.pone.0204449.t005 PLOS ONE | https://doi.org/10.1371/journal.pone.0204449 September 27, 2018 9 / 13 Agreement and reliability of galvanic-VEMP in HAM Supporting information S1 Table. Abbreviations meaning. (DOCX) S1 Appendix. GRRAS checklist for reporting of studies of reliability and agreement. (PDF) S1 Dataset. (SAV) Acknowledgments We would like to thank the Interdisciplinary HTLV Research Group (GIPH) participants, the personnel of Fundac Ëão HEMOMINAS, Hospital das Clõnicas, and Universidade Federal de Minas Gerais for their collaboration in this study. Author Contributions Conceptualization: Julia Fonseca de Morais Caporali, Denise Utsch Gonc Ëalves. Data curation: Julia Fonseca de Morais Caporali. Formal analysis: Julia Fonseca de Morais Caporali. Funding acquisition: Julia Fonseca de Morais Caporali, Ludimila Labanca, Denise Utsch Gonc Ëalves. Investigation: Julia Fonseca de Morais Caporali, Ludimila Labanca, Kyonis Rodrigues Floren- tino, Barbara Oliveira Souza. Methodology: Julia Fonseca de Morais Caporali, Ludimila Labanca, Denise Utsch Gonc Ëalves. Project administration: Julia Fonseca de Morais Caporali, Ludimila Labanca, Denise Utsch Gonc Ëalves. Resources: Denise Utsch Gonc Ëalves. Supervision: Denise Utsch Gonc Ëalves. Validation: Julia Fonseca de Morais Caporali, Ludimila Labanca, Kyonis Rodrigues Floren- tino, Barbara Oliveira Souza. Visualization: Julia Fonseca de Morais Caporali, Denise Utsch Gonc Ëalves. Writing ± original draft: Julia Fonseca de Morais Caporali. Writing ± review & editing: Julia Fonseca de Morais Caporali, Ludimila Labanca, Denise Utsch Gonc Ëalves. References 1. Fitzpatrick RC, Day BL. Probing the human vestibular system with galvanic stimulation. J Appl Physiol. 2004; 96: 2301±2316. https://doi.org/10.1152/japplphysiol.00008.2004 PMID: 15133017 2. Iles JF, Ali AS, Savic G. Vestibular-evoked muscle responses in patients with spinal cord injury. Brain. 2004; 127: 1584±1592. https://doi.org/10.1093/brain/awh173 PMID: 15128616 3. Liechti M, Mu È ller R, Lam T, Curt A. Vestibulospinal responses in motor incomplete spinal cord injury. Clin Neurophysiol. 2008; 119: 2804±2812. https://doi.org/10.1016/j.clinph.2008.05.033 PMID: PLOS ONE | https://doi.org/10.1371/journal.pone.0204449 September 27, 2018 10 / 13 Agreement and reliability of galvanic-VEMP in HAM 4. Cunha LCM, Tavares MC, Criollo CJT, Labanca L, Paz CCSC, Martins HR, et al. Contribution of Gal- vanic Vestibular Stimulation for the Diagnosis of HTLV-1-Associated Myelopathy/Tropical Spastic Para- paresis. J Clin Neurol. 2013; 9: 252±258. https://doi.org/10.3988/jcn.2013.9.4.252 PMID: 24285967 5. Caporali JFM, Gonc Ë alves DU, Labanca L, Oliveira LD, Trindade GVM, Pereira TA, et al. Vestibular evoked myogenic potential (VEMP) triggered by galvanic vestibular stimulation (GVS): a promising tool to assess spinal cord function in Schistosomal Myeloradiculopathy. PLoS Negl Trop Dis. 2016 Apr 29; 10(4):e0004672. https://doi.org/10.1371/journal.pntd.0004672 PMID: 27128806 6. Kaplan JE, Osame M, Kubota H, Igata A, Nishitani H, Maeda Y, et al. The risk of development of HTLV-I-associated myelopathy/tropical spastic paraparesis among persons infected with HTLV-I. J. Acquir. Immune Defic. Syndr. 1990; 3: 1096±1101. PMID: 2213510 7. Murphy EL, Fridey J, Smith JW, Engstrom J, Sacher RA, Miller K, et al. HTLV-associated myelopathy in a cohort of HTLV-I and HTLV-II-infected blood donors. The REDS investigators. Neurology. 1997; 48: 315±320. PMID: 9040713 8. Orland JR, Engstrom J, Fridey J, Sacher RA, Smith JW, Nass C, et al. Prevalence and clinical features of HTLV neurologic disease in the HTLV Outcomes Study. Neurology. 2003; 61: 1588±1594. PMID: 9. Maloney EM, Cleghorn FR, Morgan OS, Rodgers-Johnson P, Cranston B, Jack N, et al. Incidence of HTLV-I-associated myelopathy/tropical spastic paraparesis (HAM/TSP) in Jamaica and Trinidad. J. Acquir. Immune Defic. Syndr. Hum. Retrovirol. 1998; 17: 167±170. PMID: 9473019 10. Romanelli LC, Caramelli P, Martins ML, Gonc Ë alves DU, Proietti FA, Ribas JG, et al. Incidence of human T cell lymphotropic virus type 1-associated myelopathy/ tropical spastic paraparesis in a long-term pro- spective cohort study of initially asymptomatic individuals in Brazil. AIDS Res. Hum. Retroviruses. 2013; 29: 1199±1202. https://doi.org/10.1089/AID.2013.0086 PMID: 23617363 11. de The G, Bomford R. An HTLV-I vaccine: why, how, for whom? AIDS Res. Hum. Retroviruses. 1993; 9: 381±386. PMID: 8318266 12. Gessain A, Cassar O. Epidemiological aspects and world distribution of HTLV-1 infection. Front. Micro- biol. 2012; 3: 388. https://doi.org/10.3389/fmicb.2012.00388 PMID: 23162541 13. Gessain A, Barin F, Vernant JC, Gout O, Maurs L, Calender A, et al. Antibodies to human T lymphotro- pic virus type I in patients with tropical spastic paraparesis. Lancet. 1985; 2: 407±410. PMID: 2863442 14. Osame M, Usuku K, Izumo S, Ijichi N, Amitani H, Igata A, et al. HTLV-I associated myelopathy: a new clinical entity. Lancet. 1986; 1:1031±1032. 15. Roma Â n GC, Osame M. Identity of HTLV-I-associated tropical spastic paraparesis and HTLV-I-associ- ated myelopathy. Lancet. 1988; 1:651. 16. Osame M. Review of WHO Kagoshima meeting and diagnostic guidelines for HAM/TSP. In: Blattner W, editor. Human retrovirology: HTLV. New York: Raven Press; 462 1990. pp 191±197. 17. Arau Â jo AQC, Andrade-Filho AS, Castro-Costa CM, Menna-Barreto M, Almeida SM. HTLV-1 Associated Myelopathy/Tropical Spastic Paraparesis in Brazil: a Nationwide Survey. J Acquir Immune Defic Syndr Hum Retrovirol. 1998 Dec 15; 19 (5):536±541. PMID: 9859969 18. Shibasaki H, Endo C, Kuroda Y, Kakigi R, Oda K, Komine S. Clinical picture of HTLV-I associated mye- lopathy. J Neurol Sci. 1988; 87: 15±24. PMID: 3193123 19. Martin F, Fedina A, Youshya S, Taylor GP. A 15-year prospective longitudinal study of disease progres- sion in patients with HTLV-1 associated myelopathy in the UK. J Neurol Neurosurg Psychiatry. 2010 Dec; 81(12):1336±1340. https://doi.org/10.1136/jnnp.2009.191239 PMID: 20660921 20. Labanca L, Starling AL, de Sousa-Pereira SR, Romanelli LC, de Freitas Carneiro-Proietti AB, Carvalho LN, et al. Electrophysiological analysis shows dizziness as the first symptom in human T cell lymphotro- pic virus type-associated myelopathy/tropical spastic paraparesis. AIDS Res Hum Retroviruses. 2015 Jun; 31(6):649±654. https://doi.org/10.1089/AID.2014.0153 PMID: 25760424 21. Oliveira P, Castro NM, Muniz AL, Tanajura D, Brandão JC, Porto AF, et al. Prevalence of erectile dys- function in HTLV-1-infected patients and its association with overactive bladder. Urology. 2010; 75:1100±1103. https://doi.org/10.1016/j.urology.2009.11.041 PMID: 20189229 22. Castillo JL, Cea JG, Verdugo RJ, Cartier L. Sensory dysfunction in HTLV-I-associated myelopathy/ tropical spastic paraparesis. A comprehensive neurophysiological study. Eur. Neurol. 1999; 42: 17±22. https://doi.org/10.1159/000008063 PMID: 10394043 23. Franzoi AC, Araujo AQ. Disability profile of patients with HTLV-I-associated myelopathy/tropical spastic paraparesis using the Functional Independence Measure (FIM). Spinal Cord. 2005; 43: 236±240. https://doi.org/10.1038/sj.sc.3101677 PMID: 15520834 24. Franzoi AC, Arau Â jo AQ. Disability and determinants of gait performance in tropical spastic paraparesis/ HTLV-I associated myelopathy (HAM/TSP). Spinal Cord. 2007 Jan; 45(1):64±68. https://doi.org/10. 1038/sj.sc.3101919 PMID: 16568145 PLOS ONE | https://doi.org/10.1371/journal.pone.0204449 September 27, 2018 11 / 13 Agreement and reliability of galvanic-VEMP in HAM 25. Olindo S, Cabre P, Le Â zin A, Merle H, Saint-Vil M, Signate A, et al. Natural history of human T-lymphotro- pic virus 1-associated myelopathy: a 14-year follow-up study. Arch Neurol. 2006 Nov; 63(11):1560± 1566. https://doi.org/10.1001/archneur.63.11.1560 PMID: 17101824 26. Narikawa K1, Fujihara K, Misu T, Feng J, Fujimori J, Nakashima I, et al. CSF-chemokines in HTLV-I-as- sociated myelopathy: CXCL10 up-regulation and therapeutic effect of interferon-α. J. Neuroimmunol. 2005; 159: 177±182. https://doi.org/10.1016/j.jneuroim.2004.10.011 PMID: 15652417 27. Sato T, Coler-Reilly A, Utsunomiya A, Araya N, Yagishita N, Ando H, et al. CSF CXCL10, CXCL9, and neopterin as candidate prognostic biomarkers for HTLV-1-associated myelopathy/tropical spastic para- paresis. PLoS Negl. Trop. Dis. 2013; 7: e2479. https://doi.org/10.1371/journal.pntd.0002479 PMID: 28. Guerreiro JB, Santos SB, Morgan DJ, Porto AF, Muniz AL, Ho JL, et al. Levels of serum chemokines discriminate clinical myelopathy associated with human T lymphotropic virus type 1 (HTLV-1)/tropical spastic paraparesis (HAM/TSP) disease from HTLV-1 carrier state. Clin. Exp. Immunol. 2006; 145: 296±301. https://doi.org/10.1111/j.1365-2249.2006.03150.x PMID: 16879249 29. Starling AL, Coelho-Dos-Reis JG, Peruhype-Magalhães V, Pascoal-Xavier MA, Gonc Ë alves DU, Be Â la SR, et al. Immunological signature of the different clinical stages of the HTLV-1 infection: establishing serum biomarkers for HTLV-1-associated disease morbidity. Biomarkers. 2015; 20(6±7): 502±512. https://doi.org/10.3109/1354750X.2015.1094141 PMID: 26474234 30. Iwasaki Y, Ohara Y, Kobayashi I, Akizuki S. Infiltration of helper/inducer T lymphocytes heralds central nervous system damage in human T-cel leukemia virus infection. Am. J. Pathol. 1992; 140: 1003± 1008. PMID: 1374584 31. Bangham CR, Araujo A, Yamano Y, Taylor GP. HTLV-1-associated myelopathy/tropical spastic para- paresis. Nat Rev Dis Primers. 2015 Jun 18; 1:15012. https://doi.org/10.1038/nrdp.2015.12 PMID: 32. Fitzpatrick RC, Burke D, Gandevia SC. Task-dependent reflex responses and movement illusions evoked by galvanic vestibular stimulation in standing humans. J Physiol. 1994; 15(478): 363±372. 33. Kim J, Curthoys I S. Responses of primary vestibular neurons to galvanic vestibular stimulation (GVS) in the anaesthetised guinea pig. Brain Res Bull. 2004; 64(3): 265±71. https://doi.org/10.1016/j. brainresbull.2004.07.008 PMID: 15464864 34. Curthoys IS, Macdougall H G. What galvanic vestibular stimulation actually activates. Front Neurol. 2012; 20 (117): 1±4. 35. de Natale ER, Ginatempo F, Paulus KS, Pes GM, Manca A, Tolu E, et al. Abnormalities of vestibular- evoked myogenic potentials in idiopathic Parkinson's disease are associated with clinical evidence of brainstem involvement. Neurol Sci. 2015; 36: 995±1001. https://doi.org/10.1007/s10072-014-2054-4 PMID: 25567081 36. Britton TC, Day BL, Brown P, Rothwell JC, Thompson PD Marsden CD. Postural electromyographic responses in the arm and leg following galvanic vestibular stimulation in man. Exp Brain Res. 1993; 94: 143±151. PMID: 8335069 37. Watson SR, Colebatch JG. Vestibular-evoked electromyographic responses in soleus: a comparison between click and galvanic stimulation. Exp Brain Res. 1998; 119: 504±510. PMID: 9588785 38. Joint Committee for Guides in Metrology (JCGM). International Vocabulary of Metrology±Basic and General Concepts and Associated Terms (VIM 3rd edition) JCGM 200; 2012. 39. Bartlett JW, Frost C. Reliability, repeatability and reproducibility: analysis of measurement errors in con- tinuous variables. Ultrasound Obstet Gynecol. 2008 Apr; 31(4): 466±475. https://doi.org/10.1002/uog. 5256 PMID: 18306169 40. World Medical Association. World Medical Association Declaration of Helsinki Ethical Principles for Medical Research Involving Human Subjects. JAMA. 2013 Nov 27; 310 (20): 2191±2194. https://doi. org/10.1001/jama.2013.281053 PMID: 24141714 41. Gonc Ë alves DU, Proietti FA, Ribas JG, Arau Â jo MG, Pinheiro SR, Guedes AC, et al. Epidemiology, treat- ment, and prevention of human T-cell leukemia virus type 1-associated diseases. Clin Microbiol Rev. 2010 Jul; 23(3): 577±589. https://doi.org/10.1128/CMR.00063-09 PMID: 20610824 42. Catalan-Soares B, Carneiro-Proietti AB, Proietti FA; Interdisciplinary HTLV Research Group. Heteroge- neous geographic distribution of human T-cell lymphotropic viruses I and II (HTLV-I/II): serological screening prevalence rates in blood donors from large urban areas in Brazil. Cad Saude Publica. 2005 May-Jun; 21(3): 926±931. PMID: 15868051 43. Carneiro-Proietti AB, Ribas JG, Catalan-Soares BC, Martins ML, Brito-Melo GE, Martins-Filho O, et al. Infecc Ëão e doenc Ë a pelos võ Ârus linfotro Â picos humanos de ce Â lulas T (HTLV-I/II) no Brasil. Rev Soc Bras Med Trop. 2002 Sep-Oct; 35(5): 499±508. PMID: 12621671 PLOS ONE | https://doi.org/10.1371/journal.pone.0204449 September 27, 2018 12 / 13 Agreement and reliability of galvanic-VEMP in HAM 44. Reiss DB, Freitas GS, Bastos RHC, de Souza MA, Horiguchi CLF, Martins ML, et al. Neurological out- comes analysis of HTLV-1 seropositive patients of the Interdisciplinary Research HTLV Group (GIPH) cohort, Brazil. Retrovirology. 2014; 11(Suppl 1): 51. 45. De Castro-costa CM, Arau Â jo AQC, Barreto MM, Takayanagui OM, Sohler MP, da Silva EL, et al. Pro- posal for diagnostic criteria of tropical spastic paraparesis/HTLV-I associated myelopathy (HAM/TSP). AIDS Res Hum Retroviruses. 2006; 22: 931±935. https://doi.org/10.1089/aid.2006.22.931 PMID: 46. Shoukri MM, Asyali MH, Donner A. Sample size requirements for the design of reliability study: review and new results. Stat. Methods Med. Res. 2004; 13: 251±271. 47. Muise SB, Lam CK, Bent LR. Reduced input from foot sole skin through cooling differentially modulates the short latency and medium latency vestibular reflex responses to galvanic vestibular stimulation. Exp Brain Res. 2012; 218: 63±71. https://doi.org/10.1007/s00221-012-3002-2 PMID: 22278107 48. de Vet HC, Terwee CB, Knol DL, Bouter LM. When to use agreement versus reliability measures. J Clin Epidemiol 2006; 59:1033±1039. https://doi.org/10.1016/j.jclinepi.2005.10.015 PMID: 16980142 49. McHugh ML. Interrater reliability: the kappa statistic. Biochem Med. 2012 Oct; 22(3): 276±282. 50. Melyill JG, Watt DGD. Muscular control of landing from unexpected falls in man. J. Physiol. 1971; 219: 729±737. PMID: 5157599 51. Nashner LA. Adapting reflexes controlling the human posture. Exp Brain Res. 1976; 26: 59±72. PMID: 52. Nashner LM, Wvolfson P. Influence of head position and proprioceptive cues on short latency postural reflexes evoked by galvanic stimulation of the human labyrinth. Brain Res. 1974; 67: 255±268. PMID: 53. Allum JHJ, Pfaltz CR. Visual and vestibular contributions to pitch sway stabilization in the ankle muscles of normals and patients with bilateral vestibular deficits. Exp Brain Res. 1985; 58: 82±94. PMID: 54. Bussel B, Katz R, Pierrot-Deseilligny E, Bergego C, Hayat A. Vestibular and proprioceptive influences on the postural reactions to a sudden body displacement in man. In: Desmedt JE, editor. Progress in Clinical Neurophysiology, vol. 8, Spinal and Supraspinal Mechanisms of Voluntary Motor Control and Locomotion. Basel: Karger; 1980. pp. 310±322. 55. Horstmiann GA, Dietz V. The contribution of vestibular input to the stabilization of human posture: a new experimental approach. Neurosci. Lett. 1988; 95: 179±184. 56. Grillner S, Hongo T, Lund S. Convergent effects on alpha motoneurones from the vestibulospinal tract and a pathway descending in the medial longitudinal fasciculus. Exp Brain Res. 1971; 12: 457±479. PMID: 5093725 57. Day BL, Cole J. Vestibular-evoked postural responses in the absence of somatosensory information. Brain. 2002; 125: 2081±2088. PMID: 12183353 58. Cathers I, Day BL, Fitzpatrick RC. Otolith and canal reflexes in human standing. J Physiol. 2005; 563: 229±34. https://doi.org/10.1113/jphysiol.2004.079525 PMID: 15618274 59. Vaz S, Falkmer T, Passmore AE, Parsons R, Andreou P. The Case for Using the Repeatability Coeffi- cient When Calculation Test-Retest Reliability. PLoS One. 2013 Sep 9; 8(9):e73990. https://doi.org/10. 1371/journal.pone.0073990 PMID: 24040139 PLOS ONE | https://doi.org/10.1371/journal.pone.0204449 September 27, 2018 13 / 13
PLoS ONE – Public Library of Science (PLoS) Journal
Published: Sep 27, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera