Continuous positive airway pressure in patients with rapid eye movement ( REM )-specific obstructive sleep apnea , a retrospective match-controlled chart review

Rapid eye movement (REM) obstructive sleep apnea (OSA) Background represents 13 to 35% of all OSA cases and is more common in women. Continuous positive airway pressure (CPAP) is the gold standard for treatment of all forms of OSA but we do not know if patients with REM OSA have different pressure requirements than those with non-stage dependent OSA. : This was a retrospective case control study. We first identified Methods individuals with REM OSA and then tried to identify apnea hypopnea index (AHI), gender, and body mass index (BMI) matching controls that had non-stage specific OSA. Individuals were considered to have REM OSA if the REM AHI was greater than 5 events/hour, and the ratio of REM AHI / non-rapid eye movement (NREM) AHI was greater than 2. Demographic variables and the recommended CPAP pressure were analyzed using paired Student’s T-Tests. : Our study included a total of 16 individuals with REM OSA and Results equal number of AHI, gender, and BMI matching controls. Both groups had similar demographic and polysomnographic characteristics. Individuals with REM OSA required similar CPAP pressures as controls (7.5 cm H O vs 7.4 cm H O p=0.78). : Individuals with REM require similar CPAP pressures as their Conclusion AHI, gender, and BMI matching controls.


Introduction
The prevalence of obstructive sleep apnea (OSA) in adults has been estimated to be between 1% to 2% in women and 3% to 4% in men 1 .OSA is associated with increased cardiovascular mortality and it is considered an independent risk factor for all cause mortality [2][3][4] .The severity of the disease is usually expressed using the apnea hypopnea index (AHI) 5 .
In a subset of those with OSA, respiratory events happen predominantly or exclusively during rapid eye movement (REM) sleep; this condition is referred to as "REM-related OSA (REM OSA)".Its prevalence among patients with OSA is 13 to 35% depending on the criteria used to define AHIs [6][7][8] .REM OSA has been a topic of debate in the sleep medicine community since we do not know if this form of sleep apnea is part of the spectrum of "general" sleep apnea or a completely different entity with unique risk factors and different treatment needs.
REM OSA does not have a male predilection, in fact it may be even more common in women.Unlike OSA the likelihood of REM OSA decreases with age 7,8 .REM OSA is not associated with daytime sleepiness or diminished quality of life (QOL) 9 .
It is, however, linked to increased incidence of hypertension 10 and type 2 diabetes 11 .
CPAP is the gold standard for treatment of both REM and NREM OSA, and both forms of sleep apnea seem to respond in similar ways to CPAP therapy. 12.We still don't know if CPAP pressures required to control REM OSA are significantly different than those needed for OSA in general.
We designed this study with the initial hypothesis that individuals with REM OSA would require a lower CPAP pressure than controls.This hypothesis was based on day-to-day observations at our sleep center.

Methods
The Institutional Review Board (IRB) at HealthEast care system approved this study (IRB #HE1511002).The search included all polysomnography (PSG) tests that were completed from January 1 st 2014 to December 31 st 2014.We searched all patients older than 18 years old who had a baseline evaluation as well as a CPAP titration.REM OSA subjects and controls were required to have NREM and REM sleep during the baseline test and a minimum of 10 minutes of stage REM sleep during the titration.

Identification of individuals and controls
We first identified individuals that met criteria for REM OSA using the sleep center's database.Data was extracted by reviewing the database manually, record by record, and reviewing the polysomnography report; all the variables that were needed for this study were already part of the reports.The dataset that is provided, was used to collect the information on an MS Excel 2013 spreadsheet (Dataset 1 13 ).No information regarding CPAP pressure recommendations was accessed during this phase to avoid selection bias.The database was searched a second time in order to identify gender, BMI, and AHI matching controls.Indi-viduals were considered a matching control if the BMI and the AHI were within 5 kg/m 2 and 5 events/hour respectively.

Sample Size and controls
Based on preliminary sample-size calculations, we aimed to identify 50 individuals with REM OSA and 50 controls in order to detect a pressure difference of 1 cwp with a significance level of 0.05 Power 0.8.This was based on a separate estimate of the mean CPAP pressure recommended at the lab (9 cwp; SD=2.45;D=0.41).We only found 16 matching controls.

Polysomnography (PSG) values and definitions
PSG tests were scored by a certified PSG technologist the night of the test using the American Academy of Sleep Medicine Scoring Manual v 2.0 5 .

REM OSA definition
Individuals were considered to have REM OSA if the REM AHI was greater than 5 events/hour, and the ratio of REM AHI / non-rapid eye movement (NREM) AHI was greater than 2. REM sleep duration had to be greater than 15 minutes.

OSA definition for controls
Individuals were considered appropriate controls if they had an AHI greater than 5 events/hour and within 5 events/hour of the REM OSA match.The NREM AHI had to be greater than 5 events/hour and the ratio of REM AHI to NREM AHI had to be 1 or less.REM sleep duration had to be greater than 15 minutes.

CPAP pressure
The recommended CPAP pressure was the lowest pressure that eliminated respiratory events during the test, including supine REM sleep.

Other variables
We collected demographic, clinical, and polysomnographical variables such as age, gender, Body mass index (BMI), presence of diabetes, presence of hypertension, recording time, total sleep time, Epworth score (to assess daytime sleepiness), STOP BANG Score (to assess risk for OSA), and percentage of each sleep stage.

Statistical analysis
Descriptive characteristics for demographic information were summarized as mean with standard deviation for continuous variables and as frequency for categorical variables.
Continuous and categorical variables were compared using paired Student's T-Tests and Chi-square tests, respectively.All hypothesis tests were 2-sided, with a significance level of 0.05.All analyses were performed using R® for Mac OS X GUI version 1.67.

Results
We were able to identify a total of 16 individuals with REM OSA who had gender, AHI, and BMI matching controls for a total of 32 subjects in the study.Table 1 summarizes the demographic and clinical characteristics of individuals with REM OSA and controls.
Both groups were very similar in terms of age, daytime sleepiness, diabetes, hypertension, and sleep architecture parameters, .Subjects with REM-OSA had a slightly lower score in the STOP-BANG questionnaire (3.5 vs. 4.0 p= 0.03).
Figure 1 shows the recommended CPAP pressure for the REM OSA group and controls.There was no significant difference between the groups.

Discussion
To our knowledge our study is the first comparing nPAP treatment requirements for individuals with REM OSA.We found that Age, gender, and AHI have been identified as predictors of CPAP pressure needs but none of those studies have looked at the stage in which the respiratory events happen. 14,15.Since REM OSA has a different gender and BMI association than the general OSA population, we were able to isolate the effect of the REM component by looking at gender, AHI, and BMI matched controls 16,17 .
Our study had the limitation of a small sample size because it was difficult to identify an adequate number of matching controls.With our small sample size, the study was underpowered to detect smaller differences in CPAP pressure needs.We still think that publishing our results could be important because there are no other studies, regardless of sample size, that have looked into this topic.
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias You can publish traditional articles, null/negative results, case reports, data notes and more The peer review process is transparent and collaborative Your article is indexed in PubMed after passing peer review Dedicated customer support at every stage For pre-submission enquiries, contact research@f1000.com /doi.org/10.12688/f1000research.16688.1 26 Nov 2018, :1848 ( Latest published: 7 ) https://doi.org/10.12688/f1000research.16688.1 v1 2 2