Background
Rapid eye movement (REM) sleep behavior disorder (RBD) is a parasomnia that occurs frequently in patients with Parkinson’s disease (PD), with an estimated prevalence of 30-60% [
1‐
3]. RBD is characterized by the enactment of dreams as a result of the loss of physiological atonia during REM sleep. Behaviors displayed include hitting, kicking, shouting but also laughing. The associated dreams are often violent or frightening in origin, and sometimes patients harm themselves or their bed partners with their movements [
1]. Remarkably, the RBD-associated movements of PD patients are usually much faster, stronger and smoother than during the day, suggesting that the movements “bypass” the affected extrapyramidal systems [
4].
To make a diagnosis of RBD, the current 2nd edition of the International Classification of Sleep Disorders (ICSD-II) requires the combination of clinical features (either by history or on nocturnal video recordings) and the presence of REM sleep without atonia as an electromyographic (EMG) finding during sleep recordings (Table
1) [
5]. The gold standard for the diagnosis of RBD therefore entails a clinical interview, preferably by a sleep medicine specialist, together with at least one night of polysomnography with audiovisual recordings (video-PSG). However, referring every PD patient with complaints of nocturnal restlessness to a sleep medicine center is time-consuming, expensive and may not always be feasible in clinical practice. As a result, many movement disorder specialists base their diagnosis of RBD solely on the description of the typical behaviors by the bed partner of the patient. Studies show that this practice results in frequent misdiagnoses [
6] and, consequently, overtreatment. Mimicking disorders such as obstructive sleep apnea, confusional arousals and nocturnal hallucinations should be excluded [
7], especially because some of these may worsen with clonazepam, the first-line treatment of RBD. Therefore, there is a need for less expensive, easy to use methods to diagnose RBD.
Table 1
ICSD-II criteria for REM sleep behavior disorder
A | Presence of REM sleep without atonia; the EMG finding of excessive amounts of sustained or intermittend elevation of sub-mental EMG tone or excessive phasic submental or (upper or lower) limb EMG twitching. |
B | At least one of the following is present: |
| I. Sleep related injurious, or disruptive behaviors by history |
| II. Abnormal REM sleep behaviors documented during PSG monitoring |
C | Absence of EEG epileptiform activity during REM sleep unless RBD can be clearly distinguished from any concurrent REM sleep-related seizure disorder. |
D | The sleep disturbance is not better explained by any other sleep disorder, medical or neurological disorders, mental disorders, medication use, or substance abuse |
The REM sleep behavior disorder screening questionnaire (RBDSQ) was developed as an easy to use screening method [
8,
9]. The scale was created in German and English and more recently a Japanese version has been validated [
10]. Although the questionnaire shows good internal consistency and a high sensitivity (96%) compared to the clinical interview, it has a low specificity (56%) [
8]. The REM sleep behavior disorder questionnaire Hong Kong (RBDQ-HK) has been developed, tested and validated in Chinese patients based on the ICSD-II criteria [
11]. It was validated in a group of PSG-confirmed RBD patients and controls. The overall RBDQ-HK score was significantly higher in the RBD group. ROC analysis showed that a cut off score of 18/19 had moderate sensitivity and specificity [
11]. More recently Frauscher et al. published a validation study of the Innsbruck REM sleep behavior disorders inventory [
12]. The scale had a sensitivity of 91.4% and a specificity of 85.7% for both idiopathic and PD related RBD (AUC, 0.886). Interestingly, the scores of patients sleeping alone were comparable with patients with a bed partner.
Actigraphy has been suggested as another possible diagnostic tool for RBD. Actigraphy may be a useful instrument to obtain general measures such as total sleep time, sleep efficiency and wake after sleep onset [
13,
14]. Compared to questionnaires, actigraphy should give a more objective representation of actual motor activity during the night. In addition, actigraphy is much less expensive and cumbersome compared to video-PSG and could be used in a home setting for several days, which may compensate for night-to-night fluctuations in the presence or severity of RBD symptoms. As such the use of actigraphy in the diagnosis of RBD seems attractive and the first results on its use are indeed promising. Naismith et al. found that PD patients with RBD had a higher number of bouts scored as “wake” by actigraphy, compared to patients without RBD, based on questionnaires [
15]. In the current study, we sought to confirm these findings in a larger group of well-defined PD patients. We compared actigraphy outcomes in PD patients with and without RBD, based on the gold standard of a clinical interview in combination with video-PSG. Furthermore we searched for an optimal cut-off point to actually implement the use of actigraphy in clinical practice.
Discussion
Solely using the clinical interview to assess the possible presence of RBD in PD patients, often results in misdiagnoses. However, even judicious use of video-PSG is costly and not always feasible. Therefore, there is a clear need for new screening tools for RBD. Our results show that using actigraphy, the number of bouts classified as “wake” is significantly higher in PD patient with RBD compared to PD patients without. Accordingly, we show that actigraphy has a very high specificity and a good positive predictive value for diagnosing RBD in PD patients.
Wake bouts as scored by actigraphy were previously suggested as a possible useful marker in the diagnostic workup of RBD in PD: our findings are in agreement with Naismith et al., who studied 22 patients with 14 consecutive nights of actigraphy [
15]. However, the actual number of wake bouts was almost twice as high in our patient group, compared to theirs. Since sensitivity settings of the actigraphs were the same, this difference may have been caused by different epoch length settings, which was 0.25 min in our study and 0.50 in the study of Naismith et al. [
15]. In addition, we used video-PSG in combination with a clinical interview by a sleep medicine specialist as the gold standard for the diagnosis of RBD, instead of questionnaires.
Previous studies have suggested that actigraphy is an useful method to measure sleep quality in PD patients. Correlations were found between actigraphy and total sleep time, wake after sleep onset and subjective complaints about nocturnal sleep [
13,
14]. Our results however showed a difference between total sleep time and sleep efficiency measured with actigraphy and PSG. Although the actigraph was measured on the least affect side, we cannot exclude that the presence of tremor, on-off fluctuations and/or dyskinesias may have influenced the results. More research is needed to study the influence of PD motors symptoms on actigraphic results during the night.
There was no increase in either total or mean activity levels during sleep, which could have been expected in patients with REM related movements. However, as RBD associated movements lead to activity well above the threshold that is represented as “sleep” by actigraphy, they are almost always scored as “wake bouts” rather than increased activity during sleep. Variables other than the presence of RBD may have influenced the number of bouts classified as wake. The periodic limb movement index was, although not significantly, higher in the PD + RBD group compared to the PD-RBD group. The lack of significance could be caused by a large difference in variance. Periodic limb movements can cause sleep disturbances and therefore increase the number of wake bouts. Our groups were not matched with respect to age, disease duration, disease stage and medication use, and these factors may also influence sleep. However, regression analyses correcting for these clinical characteristics and PSG-determined actual wake time during the night, still showed a significant differences in the number of wake bouts between groups. These findings suggest that the increased number of wake bouts is primarily the result of the presence of RBD.
Results showed that using an epoch length of 0.25 min and a cut-off of 95 wake bouts per night, actigraphy is a highly specific tool for RBD in PD patients, albeit with a low sensitivity. As the prevalence of RBD in PD ranges between 30% and 60% [
1‐
3], actigraphy has a positive predictive value between 70 and 90% which is reasonable. Based on a semi-structured clinical interview alone, we found seven patients incorrectly suspected of having RBD. Of these, only one patient scored above the threshold of 95 wake bouts per night. Therefore, these results show an additional value of using actigraphy next to a clinical interview in the diagnostic trajectory of RBD. Seven patients had no clinical history of RBD-like behavior but still fulfilled the diagnostic PSG criteria of RBD, and actigraphy did not differentiate these patients from the group without RBD. Actigraphy therefore mainly has a role in combination with at least a clinical suspicion of RBD, rather than a screening instrument in PD patients without complaints of RBD. Actigraphy should not be used in the diagnosis of idiopathic RBD: although clear studies about the prevalence of RBD in the general elderly population are lacking, rates are estimated between 0.38% and 0.50%, leading to a positive predictive value below 5% [
19,
20].
Contrary to the high specificity and low sensitivity of actigraphy, previous research showed that RBD questionnaires have a high sensitivity and a somewhat low specificity [
8,
10‐
12]. Combining actigraphy and RBD questionnaires could therefore lead to a more accurate diagnosis of RBD. The combination of these two tools could reduce the need for video-PSG even more. Future research should focus on the clinical value of using a combination of both methods.
Our study used the ICSD-II criteria for the diagnosis of RBD. These criteria are not unambiguous unfortunately. They include presence of atonia during REM sleep, which represents a pathological increase of either phasic EMG activity, tonic EMG activity or both. Cut-off points to diagnose pathological increased phasic and tonic EMG activity are not mentioned in the criteria however, and no agreement has been reached on this point among international research groups. Here, we therefore adopted the criteria developed by the SinBar group, although several other visual and computerized scoring methods have been mentioned in literature [
2,
18,
21‐
26].
Conclusions
PD patients with RBD showed a significantly higher number of bouts scored as “wake” using actigraphy, compared to patients without RBD. In clinical practice, actigraphy has a high specificity, but low sensitivity in the diagnosis of RBD. According to our results and previous studies on the use of RBD questionnaires, the combination of both tools could be a promising method to diagnose RBD in PD patients, leading to a decrease in the need for the costly and time-consuming video-PSG.
Competing interests
BRB was supported by a VIDI research grant from the Netherlands Organization for Scientific Research (grant no. 016.076.352). SO was supported by a VIDI research grant from the Netherlands Organization for Scientific Research (grant no. 016.116.371). The other authors have no conflict of interest.
Authors’ contributions
ML: conception and design, acquisition of data, analysis and interpretation of data, drafting of the manuscript, final approval for publishing. JA: conception and design, interpretation of data, critical review of manuscript, final approval for publishing. BB: conception and design, critical review of manuscript, final approval for publishing. SO: conception and design, acquisition of data, interpretation of data, drafting of the manuscript, final approval for publishing. All authors read and approved the final manuscript.