Sleep-Disordered Breathing in REM Sleep Behavior Disorder with or without Parkinson’s Disease

Abstract

REM sleep behavior disorder (RBD) might render patients with Parkinson’s disease prone to sleep-disordered breathing. This retrospective polysomonographic study assessed the prevalence of sleep-disordered breathing in 108 consecutive patients with either both Parkinson’s disease and RBD (n = 37), Parkinson’s disease without RBD (n = 21), or isolated RBD (n = 50). Across all patients, 25% had at least moderate sleep-related breathing disorder, without significant differences between groups. Following multivariable analysis, RBD influenced sleep-related breathing parameters modestly but not significantly, whereas body mass index had a prominent impact. Further studies with larger patient cohorts are needed, and confounders like body mass index must adequately be controlled for.

Keywords

Body mass index Parkinson’s disease REM sleep behavior disorder sleep apnea syndromes

INTRODUCTION

Rapid eye movement (REM) sleep behavior disorder (RBD) is a parasomnia characterized by loss of physiological muscle atonia during REM sleep, resulting in motor activity like punching, kicking, and rolling over. Caused by neurodegeneration in the subcoeruleus nucleus within the brainstem, RBD is a typical feature or prodrome of synucleopathies like idiopathic Parkinson’s disease (PD) with a prevalence of 46–58% in PD patients [1 –3]. Some authors found an increased prevalence of mainly obstructive sleep apnea in PD patients with vs. without RBD [1, 2] while others did not [3 –5]. Potential pathophysiological mechanisms are unclear. The restoration of muscle tone during REM sleep per se is implausible to cause sleep-disordered breathing [1].

The study presented here aimed at re-evaluating the association between RBD and sleep-disordered breathing. Sleep-disordered breathing was hypothesized to be more prevalent among patients with PD and RBD (PD + RBD) and isolated RBD (iRBD) than in patients with PD but without RBD (PD–RBD). A multivariable analysis was initiated to assess if the effect of RBD on sleep-disordered breathing is independent from sex, age, and body mass index (BMI) which are well-known risk factors for sleep-related breathing disorders in general [6, 7].

METHODS

This retrospective study was performed at the Department of Neurology, Charité –Universitätsmedizin Berlin, and the Department of Epileptology, Ev. Krankenhaus KEH, Berlin, Germany. It was carried out in accordance with the declaration of Helsinki and approved by the ethical committee at Charité –Universitätsmedizin Berlin. Patients presenting to the outpatient clinic for neurological sleep medicine either because of suspected RBD or because of sleep complaints (mostly insomnia) in PD were referred to diagnostic video polysomnography for 3–4 nights. Deep brain stimulators were switched off overnight if producing relevant artefacts. Sleep stages and associated events were scored according to the manual of the American Academy of Sleep Medicine [8]. REM sleep without atonia was scored visually by a certified sleep specialist. RBD was diagnosed in case of REM sleep without atonia plus either video recorded RBD or a history of witnessed actions during sleep that were typical for RBD. Patients with diagnosed PD and/or RBD were included into the study, excluding only those subjects with ongoing GABAergic treatment (e.g., benzodiazepines), nocturnal continuous positive airway pressure (CPAP) ventilation, or traumatic etiology of RBD. Patients were divided into three diagnostic groups: PD + RBD, PD–RBD, and iRBD. Age, sex, and BMI were assessed for all patients. Daily levodopa equivalent dose (LED) was calculated as a surrogate marker for the progression of PD [9]. In PD patients with deep brain stimulation, LED was virtually increased by 450 mg [10]. To assess sleep-disordered breathing, only diagnostic nights without GABAergic medication and with at least 5 minutes of REM sleep were evaluated. Per patient, results of all available nights were averaged, weighted by the respective total sleep times per night, in order to account for intra-individual inter-night variability. As key measures of sleep-disordered breathing, oxygen desaturation index (ODI) at 3% -threshold and apnea-hypopnea index (AHI) were the main outcome parameters.

Statistics were performed with SPSS Statistics 23.0 (IBM, NY, USA). Data are given as percentages or as median and interquartile range (IQR). Categorical data were compared between the three diagnostic groups using Pearson’s two-tailed χ² test (omnibus and post-hoc, Bonferroni-corrected) as were continuous data with Kruskal-Wallis H test (omnibus) and two-tailed Mann-Whitney U test (post-hoc, Bonferroni-corrected). Effect sizes were estimated using Cohen’s d [11]. Correlations were evaluated using Spearman’s method. Statistical significance was set at p < 0.05 and high statistical significance at p < 0.01. In order to minimize the risk of type II errors in this exploratory study, the testing of multiple variables was not corrected for. Multiple linear regression was performed to identify variables independently associated with ODI and AHI, respectively. Sex, age, BMI, PD, RBD, and LED were included as independent variables. Since ODI and AHI values were extremely right skewed, they were transformed using Johnson transformation prior to linear regression: ODI_transformed = –1.71 + 0.95 * Ln(ODI + 0.61); AHI_transformed = 1.81 + 0.71 * Ln((AHI –0.22)/(95.75 + 0.22 –AHI)) [12].

RESULTS

One hundred twenty-eight eligible patients were screened between July 2006 and July 2019. Twenty subjects were excluded from further analysis due to GABAergic treatment (n = 9), nocturnal CPAP (n = 5), missing information on BMI (n = 5), or traumatic etiology of RBD (n = 1). Among the 108 subjects remaining, 87 were diagnosed with RBD. Of those, 82 had video confirmed RBD while another 5 had REM sleep without muscle atonia and a history of witnessed actions during sleep typical for RBD. Overall, 37 subjects had PD + RBD, 21 had PD–RBD, and 50 had iRBD. Four diagnostic nights could be evaluated in 9 patients, three nights in 52 patients, two nights in 37 patients, and one night in 10 patients. As an indicator of internal data validity, average ODI and average AHI were highly correlated (Spearman’s ρ= 0.95; p < 0.001; Supplementary Figure 1), as were first-night ODI and average ODI (ρ= 0.91; p < 0.001) and first-night AHI and average AHI (ρ= 0.94; p < 0.001).

Demographical, clinical and polysomnographic data on the respective patient groups are given in Table 1. Overall, 27 patients (25%) had an AHI of at least 15/h, indicating moderate to severe sleep-related breathing disorder (Fig. 1). Ninety-one patients (84%) had obstructive apnea predominance (PD + RBD, n = 33; PD–RBD, n = 16; iRBD, n = 42). There was no significant AHI difference between REM and non-REM sleep stages in any of the groups (two-tailed Wilcoxon rank sum test; Table 1). ODI and AHI were highest in patients with PD + RBD, albeit not significant, but equally low among patients with PD–RBD and iRBD. AHI differences during REM sleep alone were not significant either.

Table 1

Demographic, clinical, and polysomnographic data

	PD + RBD	PD–RBD	iRBD
	(n = 37)	(n = 21)	(n = 50)
Females	n = 10 (27%)	n = 11 (52%)	n = 11 (22%)
Age	68 years [61–71]	65 years [52–72]	69 years [63–73]
BMI	26.4 kg/m² [23.9–28.7]	25.4 kg/m² [22.8–27.5]	26.6 kg/m² [24.6–29.5]
Deep brain stimulation	n = 7 (19%)	n = 4 (19%)	n = 0 (0%)^*
LED^a	726 mg/d [450–1,402]	664 mg/d [385–1,072]	0 mg/d [0–0]^*
Total sleep time^b	334 min [271–394]	324 min [282–385]	399 min [376–434]
Sleep efficiency^b	81% [68–84]	78% [64–84]	82% [77–87]
Relative REM time^b	15% [10–21]	15% [11–17]	20% [14–25]
ODI^b	6.7 /h [3.5–17.8]	4.6 /h [1.7–7.5]	4.4 /h [1.9–12.0]
AHI^b	11.0 /h [3.8–24.2]	4.4 /h [2.2–9.5]	5.5 /h [2.3–13.5]
AHI during REM^b	9.4 /h [2.9–18.4]	3.9 /h [2.2–7.8]	5.4 /h [1.9–12.8]
AHI during non-REM^b	10.6 /h [3.1–24.9]	3.7 /h [1.9–10.1]	5.7 /h [2.2–13.1]
AHI≥15/h^b	n = 13 (35%)	n = 3 (14%)	n = 11 (22%)
Obstructive apneas^b	2.5 /h [0.9–6.0]	0.8 /h [0.2–3.8]	1.5 /h [0.5–5.0]
Central abnoeas^b	0.2 /h [0.0–1.2]	0.1 /h [0.0–0.4]	0.1 /h [0.0–0.4]
Mixed apneas^b	0.3 /h [0.0–1.9]^#	0.1 /h [0.0–0.2]	0.0 /h [0.0–0.2] ^#
Hypopneas^b	5.2 /h [2.0–9.2]	2.0 /h [1.2–4.1]	3.0 /h [1.2–5.7]

Data are given as n (%) or median [interquartile range]. PD, Parkinson’s disease; RBD, REM sleep behavior disorder; iRBD, isolated RBD; BMI, body mass index; LED, levodopa equivalent dose; ODI, oxygen desaturation index; AHI, apnea-hypopnea-index. ^aIn patients with deep brain stimulation, LED was virtually increased by 450 mg. Two subjects with iRBD took low-dose levodopa for restless legs symptoms. ^bPer patient, polysomnographic data were averaged from all available diagnostic nights (range 1–4), weighted by the respective total sleep times per night. *p < 0.01, PD + RBD vs. iRBD and PD–RBD vs. iRBD, respectively. ^#p < 0.05, PD + RBD vs. iRBD.

All patients taken together, in bivariate analyses, BMI was the only variable significantly associated with ODI (ρ= 0.34; p < 0.001) and AHI (ρ= 0.29; p = 0.002) while all other variables had weaker and statistically not significant effects (Table 2). Following multivariable analyses, BMI again was the only variable that had a substantial and significant effect on ODI (standardized β= 0.38; p < 0.001) and AHI (standardized β= 0.33; p < 0.001). The effect sizes of LED, age, and RBD were smaller (standardized β 0.14–0.24, p > 0.05) while PD and sex had virtually no effect (Table 2). The multivariable model for ODI fit the data somewhat more exactly (corrected R² = 0.18; ANOVA: p < 0.001) than that for AHI (corrected R² = 0.15; ANOVA: p = 0.001).

Table 2

Bivariate and multivariate analyses of variables associated with ODI and AHI

Dependent variables	Independent variables	Bivariate analyses^a		Multivariable analyses^b
		Cohen’s d/Spearman’s ρ	p	β (95% –CI)	Standardized β	P
ODI	Sex [female]	d = 0.13	0.49	–0.01 (–0.40 –0.39)	–0.00	0.97
	Age [years]	ρ= 0.14	0.16	0.02 (–0.00 –0.03)	0.17	0.06
	BMI [kg/m²]	ρ= 0.34	<0.001	0.10 (0.05 –0.15)	0.38	<0.001
	RBD	d = 0.26	0.19	0.35 (–0.16 –0.86)	0.14	0.18
	PD	d = 0.17	0.38	0.05 (–0.48 –0.58)	0.03	0.85
	LED [100 mg/d]	ρ= 0.17	0.09	0.04 (–0.00 –0.08)	0.24	0.06
AHI	Sex [female]	d = 0.18	0.34	–0.04 (–0.46 –0.39)	–0.02	0.87
	Age [years]	ρ= 0.15	0.12	0.02 (–0.00 –0.04)	0.17	0.06
	BMI [kg/m²]	ρ= 0.29	0.002	0.09 (0.04 –0.14)	0.33	<0.001
	RBD	d = 0.32	0.10	0.42 (–0.13 –0.97)	0.16	0.13
	PD	d = 0.17	0.38	0.09 (–0.48 –0.66)	0.04	0.76
	LED [100 mg/d]	ρ= 0.17	0.08	0.04 (–0.00 –0.08)	0.23	0.08

CI, confidence interval; ODI, oxygen desaturation index; AHI, apnea-hypopnea-index; PD, Parkinson’s disease; RBD, REM sleep behavior disorder; BMI, body mass index; LED, levodopa equivalent dose. ^aDichotomous variables were compared to ODI and AHI, respectively, using Mann-Whitney U test with Cohen’s d as a measure of effect size. Continuous variables were correlated to ODI and AHI using Spearman’s method. ^bODI and AHI were transformed using Johnson transformation prior to multiple linear regression. Model for ODI: corrected R² = 0.18; ANOVA: p < 0.01. Model for AHI: corrected R² = 0.15; ANOVA: p < 0.01.

DISCUSSION

In this study, both during REM sleep and non-REM sleep, ODI and AHI tended to be highest among patients with PD + RBD as compared to those with iRBD (positive control) and PD–RBD (negative control). However, differences between the groups were statistically not significant. Following multivariable analysis, a statistical effect of RBD on measures of sleep-disordered breathing was present, but it was weak and statistically not significant. Among all variables included, the one with the strongest statistical effect on ODI and AHI was BMI. BMI is a well-known risk factor for sleep-related breathing disorder in the general population [6, 7] although this has been questioned to be also true in PD [13]. Age and LED also contributed but weakly to ODI and AHI, which is consistent with the literature since sleep-disordered breathing has been shown to worsen with advancing age [6]. Dopaminergic medication per se does not aggravate sleep-disordered breathing [14] but the extent of neurodegeneration apparently does [15], and LED is a surrogate marker for the latter.

This polysomnographic study is the first to assess sleep-related breathing disorder in the three conditions PD + RBD, PD–RBD, and iRBD within one patient cohort. Patients with RBD were predominantly male, consistent with other studies [2 , 16]. Across all patients, apneas were rather obstructive than central, again consistent with the literature [1, 17]. The 25% prevalence of at least moderate sleep-related breathing disorder (AHI≥15/h) among patients with PD was somewhat higher than 15–22% as described in the literature [18], probably due to selection bias: the study cohort exclusively consisted of patients with sleep complaints. In the general population, the prevalence of sleep-disordered in subjects aged 60 to 70 is about 20% [7].

Zhang et al. reported sleep-related breathing disorder in 51.4% of 35 patients with PD + RBD compared to 9.1% in 11 patients with PD–RBD (p < 0.05, uncorrected). However, as in the present study, their PD + RBD group had higher BMI (mean 23.9 kg/m² vs. 22.5; p > 0.05) and higher age (mean 66.7 years vs. 62.4; p > 0.05) than the PD–RBD group. These probable confounders were not accounted for, nor were multivariable analyses performed [2]. Other authors found no significant differences in sleep-related breathing either between PD + RBD and iRBD [4, 16] or between PD + RBD and PD–RBD [3, 5]. In these studies, BMI values were homogeneous across the groups compared [3 –5] or not denoted [16].

Based on the results depicted above, the following is left to conclude. Statistically, RBD indeed seems to render patients prone to sleep-disordered breathing. However, the effect is too weak to be proven in monocentric studies. Power estimation reveals that about 110 patients per group are required. Other than previously reported [13], sleep-related breathing disorder in PD and/or RBD is highly dependent on BMI. This must adequately be controlled for in further studies. The pathophysiology of sleep-disordered breathing in RBD remains to be elucidated.

CONFLICTS OF INTEREST

The authors have no conflict of interest to report.

Footnotes

ACKNOWLEDGMENTS

The authors wish to thank Pascal Grosse, MD, for initiating this study and for diagnosing the patients, and Ulrike Grittner, PhD, for supervising the statistical analyses. BJV received funding from Deutsche Forschungsgemeinschaft (DFG 422589384) and from von Bodelschwingh Foundation.

The supplementary material is available in the electronic version of this article: .

References

Cochen

, V, Abouda

, Leu

, Oudiette

, Roze

, Vidailhet

, Similowski

, Arnulf

(2010) Is obstructive sleep apnea a problem in Parkinson’s disease?Sleep Med11, 247–252.

Zhang

, Liu

, Kang

, Yang

, Wang

, Ding

, Chen

, Liu

(2016) Association of rapid eye movement sleep behavior disorder with sleep-disordered breathing in Parkinson’s disease. Sleep Med20, 110–115.

Gong

, Xiong

, Mao

, Shen

, Hu

, Huang

, Han

, Chen

, Liu

(2014) Clinical manifestations of Parkinson disease and the onset of rapid eye movement sleep behavior disorder. Sleep Med15, 647–653.

Koo

, Lee

, Nam

(2018) Difference in severity of sleep apnea in patients with rapid eye movement sleep behavior disorder with or without parkinsonism. Sleep Med49, 99–104.

Huang

, Zhang

, Shen

, Zhang

, Cao

, Mao

, Yang

, Chen

, Liu

, Li

(2018) Effect of rapid eye movement sleep behavior disorder on obstructive sleep apnea severity and cognition of Parkinson’s disease patients. Chin Med J (Engl)131, 899–906.

Tufik

, Santos-Silva

, Taddei

, Bittencourt

(2010) Obstructive sleep apnea syndrome in the Sao Paulo Epidemiologic Sleep Study. Sleep Med11, 441–446.

Young

, Shahar

, Nieto

, Redline

, Newman

, Gottlieb

, Walsleben

, Finn

, Enright

, Samet

(2002) Predictors of sleep-disordered breathing in community-dwelling adults: the Sleep Heart Health Study. Arch Intern Med162, 893–900.

Berry

, Brooks

, Gamaldo

, Harding

, Marcus

, Vaughn

(2018) The AASM manual for the scoring of sleep and associated events: Rules, terminology and technical specifications, American Academy of Sleep Medicine, Darien, IL.

Tomlinson

, Stowe

, Patel

, Rick

, Gray

, Clarke

(2010) Systematic review of levodopa dose equivalency reporting in Parkinson’s disease. Mov Disord25, 2649–2653.

10.

Perestelo-Perez

, Rivero-Santana

, Perez-Ramos

, Serrano-Perez

, Panetta

, Hilarion

(2014) Deep brain stimulation in Parkinson’s disease: meta-analysis of randomized controlled trials. J Neurol261, 2051–2060.

11.

Lenhard

, Psychometrica. Computation of Effect Sizes, https://www.psychometrica.de/effect_size.html, Accessed October 8, 2019.

12.

Hemmerich

, Johnson Transformation berechnen: StatistikGuru, https://statistikguru.de/rechner/johnson-transformation-berechnen.html, Accessed July 18, 2019.

13.

Trotti

, Bliwise

(2010) No increased risk of obstructive sleep apnea in Parkinson’s disease. Mov Disord25, 2246–2249.

14.

Gros

, Mery

, Lafontaine

, Robinson

, Benedetti

, Kimoff

, Kaminska

(2016) Obstructive sleep apnea in Parkinson’s disease patients: effect of Sinemet CR taken at bedtime. Sleep Breath20, 205–212.

15.

Kaminska

, Lafontaine

, Kimoff

(2015) The interaction between obstructive sleep apnea and Parkinson’s disease: possible mechanisms and implications for cognitive function. Parkinsons Dis2015, 849472.

16.

Iranzo

, Santamaria

, Tolosa

(2009) The clinical and pathophysiological relevance of REM sleep behavior disorder in neurodegenerative diseases. Sleep Med Rev13, 385–401.

17.

Crosta

, Desideri

, Marini

(2017) Obstructive sleep apnea syndrome in Parkinson’s disease and other parkinsonisms. Funct Neurol32, 137–141.

18.

Nomura

, Inoue

, Kobayashi

, Namba

, Nakashima

(2013) Characteristics of obstructive sleep apnea in patients with Parkinson’s disease. J Neurol Sci327, 22–24.