Idiopathic rapid-eye-movement sleep disorder: Associations with antidepressants, psychiatric diagnoses, and other factors, in relation to age of onset

1. Introduction 

Rapid-eye-movement (REM) sleep behavior disorder (RBD) is a parasomnia characterized by excessive motor activity during REM sleep, often in the form of dream-enactment behavior, at times resulting in significant injury to either the patient or the bed partner [1], [2], [3]. The condition most often affects men in the sixth or seventh decade of life, and is frequently accompanied by neurodegenerative diseases, specifically synucleinopathies such as Parkinson’s disease (PD), dementia with Lewy Bodies (DLB), and multiple system atrophy (MSA) [2], [3], [4], [5], [6], [7], [8], [9], [10]. While the term idiopathic RBD is used to describe those without an accompanying neurological disease, this term may be misleading, as the condition can precede the frank onset of a neurodegenerative process by many years [2], [11], [12], [13].

Confusing the categorization of those with idiopathic RBD, there are many reports that associate the condition with various medications or substances. Reported inducing agents of RBD or REM sleep without atonia (RSWA), the presumed substrate for RBD, include the use or withdrawal of barbiturates, bisoprolol, caffeine, and alcohol [14], [15], [16], [17], [18], [19], [20], [21], [22]. Numerous antidepressants, including fluoxetine, venlafaxine, and clomipramine, have also been associated with RBD in those without accompanying neurodegenerative diseases [23], [24], [25], [26], [27]. In the sole investigation to systematically address potential associations of medications with the condition, Winkelman and James [28], utilizing a retrospective design, demonstrated a higher frequency of RSWA (RBD was not specifically assessed) in subjects taking serotonergic antidepressants compared to controls. Finally, there is controversy regarding the relationship between psychiatric illnesses themselves as etiological factors in RBD [29], with several studies showing a potential relationship between severe stress or post-traumatic stress disorder (PTSD) [30], [31], [32], [33], [34].

We sought to investigate the relationship between the age of onset of idiopathic RBD and many of the previously reported secondary associations. We hypothesized that patients with early-onset idiopathic RBD would be more likely to demonstrate the presence of antidepressant use as compared to age and gender matched patients without RBD. Conversely, we hypothesized that patients with late-onset idiopathic RBD would exhibit a similar frequency of secondary factors as compared to their respective control group.

2. Methods 

Medical records of consecutive patients diagnosed with RBD between 2002 and 2005 at the Mayo Sleep Disorders Center were reviewed, after approval by the Mayo Foundation Institutional Review Board. All diagnoses conformed to International Classification of Sleep Disorders (edition 2) (ICSD-2) criteria [1] and were established by the American Board of Sleep Medicine-certified physicians. Patients with concurrent neurologic disorders such as PD, DLB, MSA, mild cognitive impairment, and narcolepsy were excluded.

RBD patients were divided into two groups based on the age of onset which, in the case of the late-onset group, was determined solely from clinical histories. In the case of the early-onset group, premature commencement was further established by virtue of the fact that all patients underwent polysomnography (PSG) prior to the predetermined cutoff age for a late-onset diagnosis. Early-onset and late-onset designations were based on the mean age of RBD onset described in the two largest case series [2], [3], with cutoffs at ⩽50 and >50 years of age, respectively. Each patient was matched for age, gender and date of PSG to a control diagnosed with obstructive sleep apnea (OSA), but without RBD or a neurological disease. An OSA control group was chosen due to its high prevalence in the clinical setting and due to the fact that PSG is required for the diagnosis, thereby establishing that RSWA and RBD were not present.

The medical records were concomitantly examined for the presence of current or past psychiatric diagnoses and the use of medications or substances, including antidepressants, beta-blockers, caffeine, alcohol, and illicit drugs. The presence of a current psychiatric diagnosis was based on the sleep clinician’s statement as such in the medical record or based on its inclusion in the summary diagnoses section. Past psychiatric diagnoses were referenced only in the past medical history section of the record. Early-onset RBD cases and late-onset RBD cases were compared to their respective controls using chi-square or Fisher’s exact tests, as appropriate. Odds ratios with 95% confidence intervals were calculated for each contingency table. Early-onset and late-onset RBD cases were also compared using chi-square tests. A post hoc logistic regression analysis was used to assess the association of current and past psychiatric diagnosis and antidepressant use on one’s status as an RBD case or control, while adjusting for age. Analyses were performed using JMP statistical software (version 6, SAS Institute, Inc., Cary, NC). p-Values <0.05 were considered statistically significant.

3. Results 

Five hundred and twenty-nine consecutive records were reviewed. There were 20 patients with early-onset idiopathic RBD and 28 patients with late-onset idiopathic RBD. Mean ages at diagnosis were approximately 34 years for the early-onset and 70 years for the late-onset group, and the early-onset group had significantly more females (45% versus 11%, p=0.007). See Table 1 for further details.

Table 1. Demographic characteristics

		Early-onset idiopathic RBD (⩽50yo) N=20	Late-onset idiopathic RBD (>50yo) N=28	p-Value
	Gender	45% (9) female	11% (3) female	0.007
	Mean age at PSG	34.10±11.48 (range 13–49)	69.93±5.73 (range 59–81)	<0.001

At the time of PSG, the early-onset group had significantly more psychiatric diagnoses [85% vs. 46% and 85% vs. 36%, respectively, for past and present diagnoses (p<0.01 for both comparisons)] and antidepressant use (80% vs. 46%, p=0.02) in comparison to the late-onset group (see Table 2A for further details). Affective disorders (predominantly unipolar) comprised 63% and 68% of past and current psychiatric diagnoses (respectively) within the early-onset group and 67% and 92% of past and current psychiatric diagnoses (respectively) within the late-onset group. The remainder of diagnoses were categorized as anxiety disorders. Most of the patients on antidepressants were taking selective serotonin reuptake inhibitors (SSRIs) including citalopram, escitalopram, fluoxetine, sertraline, and paroxetine (see Table 2B). One of the early-onset RBD patients was taking a combination of amitriptyline and paroxetine. Buproprion was given in combination with other antidepressants to three of the early-onset RBD patients (bupropion and mirtazapine or venlafaxine), but not by itself. Bupropion was not prescribed in the late-onset RBD group.

Table 2A. Group characteristics

		Early-onset idiopathic RBD (⩽50 yo) N=20	Late-onset idiopathic RBD (>50 yo) N=28	p-Value
		Patients at time of PSG % (N)	Patients at time of PSG % (N)
	Past psychiatric diagnosis	85 (17)	46 (13)	<0.01
	Current psychiatric diagnosis	85 (17)	36 (10)	< 0.01
	Antidepressant use	80 (16)	46 (13)	0.02

Table 2B. Antidepressant characteristics

		Early-onset idiopathic RBD on antidepressants	Late-onset idiopathic RBD on antidepressants
		N=16a	N=13
		% (N)	% (N)
	SSRI (citalopram, escitalopram, fluoxetine, sertraline, paroxetine)	50 (8)	77 (10)
	Tricyclics (imipramine, amitriptyline, clomipramine, nortriptyline)	25 (4)	8 (1)
	Mixed mechanism (mirtazapine, venlafaxine, duloxetine)	31 (5)	15 (2)
	Buproprion (in combination with either mirtazapine or venlafaxine)	19 (3)	0

a One of the early-onset RBD patients was on a combination of amitriptyline and paroxetine increasing the total number to greater than 16.

Tables 3 and 4 compare idiopathic RBD patients to their respective non-RBD controls. For early-onset RBD patients, the association of variables with case–control status is shown in Table 3. Significant variables (all found more frequently in the RBD group) included past psychiatric diagnosis (p<0.01), current psychiatric diagnosis (p<0.01), and antidepressant use (p<0.01). The odds ratios and 95% confidence intervals associated with these findings were 17.0 [3.5–83.4], 17.0 [3.5–83.4], and 12.0 [2.7–53.3], respectively. There were no significant intergroup differences with respect to other assessed variables, which included use of beta-blockers, alcohol, caffeine, and illicit drugs.

Table 3. Early-onset idiopathic RBD vs. non-RBD controls

		Early-onset idiopathic RBD (⩽50 yo) N=20	Non-RBD controls N=20	Odds ratio (CI)	p-Value
		Patients at time of PSG % (N)	Patients at time of PSG % (N)
	Past psychiatric diagnosis	85 (17)	25 (5)	17.0 (3.5–83.4)	<0.01
	Current psychiatric diagnosis	85 (17)	25 (5)	17.0 (3.5–83.4)	<0.01
	Antidepressant use	80 (16)	25 (5)	12.0 (2.7–53.3)	<0.01
	Beta-blockers	10 (2)	0	NA	0.49
	Alcohol use	40 (8)	35 (7)	0.81 (0.22–2.9)	0.74
	Caffeine use	60 (12)	50 (10)	1.5 (0.43–5.3)	0.52
	Illicit drug use	0	5 (1)	NA	1.0

The results of the equivalent analyses in late-onset idiopathic RBD patients are shown in Table 4. Significant variables (found more commonly in the RBD group) included past psychiatric diagnosis (p<0.01) and current psychiatric diagnosis (p=0.03). The odds ratios and 95% confidence intervals for these variables were 7.2 [1.8–29.6] and 4.6 [1.1–19.3], respectively. There were otherwise no significant differences in antidepressant use or with respect to the other aforementioned variables.

Table 4. Late-onset idiopathic RBD vs. non-RBD controls

		Late-onset Idiopathic RBD (>50 yo) N=28	Non-RBD controls N=28	Odds ratio (CI)	p-Value
		Patients at time of PSG % (N)	Patients at time of PSG % (N)
	Past psychiatric diagnosis	46 (13)	11 (3)	7.2 (1.8–29. 6)	<0.01
	Current psychiatric diagnosis	36 (10)	11 (3)	4.6 (1.1–19.3)	0.03
	Antidepressants	46 (13)	25 (7)	2.6 (0.84–8.1)	0.09
	Beta-blockers	14 (4)	32 (9)	0.35 (0.09–1.3)	0.11
	Alcohol use	29 (8)	42 (12)	1.9 (0.62–5.7)	0.27
	Caffeine use	68 (19)	54 (15)	1.8 (0.62–5.4)	0.26
	Illicit drug use	3 (1)	0	NA	1.0

A post hoc logistic regression analysis adjusted for age at the time of PSG was performed comparing all the idiopathic RBD patients to non-RBD controls (Table 5). Psychiatric diagnoses increased the likelihood of RBD by approximately nine- to ten-fold and antidepressant use increased the likelihood by approximately five-fold.

Table 5. Logistic regression analysis of all idiopathic RBD vs. non-RBD controls

	Logistic regression, adjusted for age at PSG	Idiopathic RBD vs. non-RBD controls N=96	p-Value
		Odds ratio (CI)
	Past psychiatric diagnosis	10.2 (3.78–31.12)	<0.01
	Current psychiatric diagnosis	8.9 (3.18–29.21)	<0.01
	Antidepressant use	4.9 (2.02–12.38)	<0.01

4. Discussion 

Numerous reports have described idiopathic RBD in association with presumed inducing factors [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [30], [31], [32], [33], [34]. We hypothesized a higher frequency of these secondary factors in our early-onset cohort, which was supported by the demonstrated relationship with respect to antidepressants (the vast majority of which possessed serotonergic properties) and psychiatric diagnoses (primarily depressive disorders). The former finding expands upon the report of Winkelman and James and represents the first systematic assessment of antidepressant use in association with RBD, specifically within a younger population (mean age 34.10±11.48) [28]. As this is a retrospective study, however, direct causality with the demonstrated associations cannot be inferred, and there are other potential explanations for the findings.

If this association is indeed directly tied to medication use (with psychiatric diagnoses associated indirectly), one could conjecture that it is either due to an idiosyncratic reaction or that the antidepressant “unmasks” susceptibility to RBD that would have otherwise manifested later in life. Winkelman and James discuss a logical rationale for the chemical induction of RBD by antidepressants, due in particular to their serotonergic activity, and potential influence on REM motor tone either indirectly at brainstem levels or directly at spinal levels [28]. While these authors reported persistent effects of the medications on tonic (but not phasic) electromyographic REM activity, our analyses did not include these specific assessments. Future studies investigating the role of medications in this condition should include these measures whenever possible to determine whether differential patterns exist. Regardless of the possible role of antidepressant medications in idiopathic RBD, however, it appears that it is a relatively uncommon effect, particularly when one takes into account the number of individuals receiving these medications.

An alternative explanation is that antidepressants simply serve as a proxy for psychopathology, which may itself be yet another harbinger of an evolving neurodegenerative disease. In possible support of these suppositions, older RBD patients in this study also exhibited a significantly higher frequency of psychiatric diagnoses (but not antidepressant use) when compared to their respective control group, but not nearly to the same extent as their younger counterparts. In addition, the regression analysis most strongly supported an association with psychiatric diagnoses and idiopathic RBD. As a potentially unifying explanation, depression has been found to increase the risk of future development of PD by two-fold or more [35], and the majority of patients with relatively late-onset idiopathic RBD will later develop a neurodegenerative disorder, with a particular predilection for one of the alpha-synucleinopathies [2], [6], [11], [12]. In another study of PD patients, Remy and colleagues, utilizing positron emission tomography technology, demonstrated a unique pattern of neurodegeneration of non-dopaminergic neurotransmitter systems in depressed versus non-depressed controls and in the locus coeruleus (LC) [36]. Lesions of this region in animals are implicated in the development of RSWA, and LC pathology has also been demonstrated in select cases of RBD [37]. The applicability of these mechanisms in this particular study remains conjectural. While we assessed the presence of both current and past psychiatric diagnoses, we only documented present medication use. Therefore, the relative role of prior antidepressant exposure needs to be further explored.

A weak precedent may exist for a potential relationship between psychiatric illness and RBD, as several studies have supported an association between severe stress or PTSD and RBD [30], [31], [32], [33], [34]. However, this finding is controversial, as it relies solely on case reports or studies that describe REM motor abnormalities that have not been replicated [29] and do not meet specific criteria for RBD. We did not investigate an association between specific psychiatric entities and RBD, both due to the relatively small number of patients in the study and also due to uncertainty as to whether official criteria were used to make diagnoses. Future studies would benefit from utilization of standardized assessments, in addition to descriptions of associated features. A final possibility is that RBD itself predisposes to psychopathology, but our methodology did not afford the opportunity to assess a temporal relationship.

Recent literature shows that, even in the absence of a discretely associated neurological disorder, RBD patients frequently exhibit subclinical abnormalities on the testing of color vision, olfactory function, cognition, motor function, and autonomic function, with a pattern typically seen in patients with synucleinopathies [38], [39], [40], [41], [42], [43], [44]. Intriguing neuroimaging studies supplement these findings. Mazza et al., utilizing single photon emission computed tomography technology, demonstrated cerebral perfusion abnormalities in eight elderly subjects with idiopathic RBD compared to nine controls, with an anatomic metabolic profile consistent with those of PD patients [45]. A study investigating seventeen cognitively normal adults with dream-enactment behavior showed a reduced cerebral metabolic rate for glucose in regions preferentially affected in patients with DLB [46]. This had led some to suggest that the term “cryptogenic” be used in lieu of the term “idiopathic” [13], but available studies have not included younger cohorts such as the ones described in this study. It is therefore unclear whether they possess the same future risks of developing a neurodegenerative disease or whether they demonstrate the same subclinical signs discussed above. Moreover, it is uncertain whether the presence or absence of secondary factors alters prognostic implications.

Observed demographic differences between the early-onset and late-onset groups in our study are important as distinguishing factors when assessing an idiopathic RBD patient, but could potentially have other implications as well. While typically felt to be a male-predominant condition, we found a larger than expected percentage of female patients in the younger group compared to the older group. While this may be due in part to the increased prevalence of depression in women [47], an alternative possibility is that the younger-onset patients more frequently herald the future development of MSA, which has not been shown to demonstrate a particular gender preference [2].

In conclusion, further delineation of the implications of early-onset idiopathic RBD and the role of secondary factors is important for a variety of medical and ethical reasons. As it is unknown whether these patients possess the same risk for the development of neurodegenerative diseases as compared to their older counterparts, the question as to whether they should be told of these theoretical future risks (and whether any routine screening should be performed) presents an ethical dilemma. If such an association was indeed demonstrated, this younger population would be ideal candidates for trials of early neuroprotective interventions. On the other hand, if it were determined that RBD in younger populations was predominantly related to antidepressant medications, significant strides could be made in further understanding the pathophysiology underlying the condition. On a final note, neither explanation is mutually exclusive, as early-onset idiopathic RBD could be due to a combination of factors, and be representative of a variety of prognostic scenarios.

The retrospective case–control nature of this study prevents assumptions of direct causality. It does serve, however, to create a unique clinical profile for idiopathic RBD in younger patients, which should facilitate increased awareness and ultimately an understanding of its significance in this population. With long-term prospective studies (both with and without medications and other variables thought to be associated with RBD), the numerous uncertainties regarding this entity could be resolved.