ADVERTISEMENT

Upcoming Website Maintenance Notification

On Sunday October 20th, 2019 from 5:00 am cst to 8:00 am cst there will be approximately 3 hours of scheduled downtime while we make upgrades to our servers. During this time-window the JCSM website will be unavailable. Thank you for your patience.

gears

Issue Navigator

Volume 11 No. 08
Earn CME
Accepted Papers





Scientific Investigations

Altered Sleep Stage Transitions of REM Sleep: A Novel and Stable Biomarker of Narcolepsy

Yaping Liu, MD1; Jihui Zhang, MD, PhD1; Venny Lam, RPSGT1; Crover Kwok Wah Ho, RPSGT1; Junying Zhou, MD, PhD1,2; Shirley Xin Li, PhD1; Siu Ping Lam, MRCPsych, FHKAM (Psych)1; Mandy Wai Man Yu, MPH, RPSGT1; Xiangdong Tang, MD, PhD2; Yun-Kwok Wing, FRCPsych, FHKAM (Psych)1
1Department of Psychiatry, Faculty of Medicine, The Chinese University of Hong Kong, Shatin, Hong Kong SAR, China; 2Sleep Medicine Center, West China Hospital, Sichuan University, Chengdu, Sichuan, China

ABSTRACT

Objectives:

To determine the diagnostic values, longitudinal stability, and HLA association of the sleep stage transitions in narcolepsy.

Methods:

To compare the baseline differences in the sleep stage transition to REM sleep among 35 patients with type 1 narcolepsy, 39 patients with type 2 narcolepsy, 26 unaffected relatives, and 159 non-narcoleptic sleep patient controls, followed by a reassessment at a mean duration of 37.4 months.

Results:

The highest prevalence of altered transition from stage non-N2/N3 to stage R in multiple sleep latency test (MSLT) and nocturnal polysomnography (NPSG) was found in patients with type 1 narcolepsy (92.0% and 57.1%), followed by patients with type 2 narcolepsy (69.4% and 12.8%), unaffected relatives (46.2% and 0%), and controls (39.3% and 1.3%). Individual sleep variables had varied sensitivity and specificity in diagnosing narcolepsy. By incorporating a combination of sleep variables, the decision tree analysis improved the sensitivity to 94.3% and 82.1% and enhanced specificity to 82.4% and 83% for the diagnosis of type 1 and type 2 narcolepsy, respectively. There was a significant association of DBQ1*0602 with the altered sleep stage transition (OR = 16.0, 95% CI: 1.7–149.8, p = 0.015). The persistence of the altered sleep stage transition in both MSLT and NPSG was high for both type 1 (90.5% and 64.7%) and type 2 narcolepsy (92.3% and 100%), respectively.

Conclusions:

Altered sleep stage transition is a significant and stable marker of narcolepsy, which suggests a vulnerable wake-sleep dysregulation trait in narcolepsy. Altered sleep stage transition has a significant diagnostic value in the differential diagnosis of hypersomnias, especially when combined with other diagnostic sleep variables in decision tree analysis.

Citation:

Liu Y, Zhang J, Lam V, Ho CK, Zhou J, Li SX, Lam SP, Yu MW, Tang X, Wing YK. Altered sleep stage transitions of REM sleep: a novel and stable biomarker of narcolepsy. J Clin Sleep Med 2015;11(8):885–894.


Narcolepsy is a lifelong debilitating sleep disorder characterized by the instability of sleep-wake regulation and a pentad of manifestations, including excessive daytime sleepiness (EDS), cataplexy, hypnagogic hallucination, sleep paralysis, and nocturnal sleep disruption.1,2 The cerebrospinal fluid (CSF) hypocretin-1 level is currently considered as one of the best criteria for the diagnosis of narcolepsy. However, CSF hypocretin remains a less ideal option in clinical practice, as lumbar puncture is an invasive procedure that entails a degree of risk and potential side effects (e.g., headache), and the majority of patients in some Asian countries are reluctant to undergo this diagnostic testing.3 Thus, multiple sleep latency test (MSLT) remains essential for making a diagnosis of narcolepsy. Mean sleep latency (MSL) ≤ 8 minutes and the presence of two or more sleep onset REM periods (SOREMPs) on daytime MSLT and nocturnal polysomnography (NPSG) are considered diagnostic for narcolepsy according to the International Classification of Sleep Disorders, third edition (ICSD-3).1 The sensitivity and specificity of MSLT in diagnosing narcolepsy, however, are not as satisfactory as originally expected.4 Patients with other sleep disorders present with EDS, such as obstructive sleep apnea syndrome (OSAS),5,6 behaviorally induced insufficient sleep syndrome (BIISS),7 and sleep interruption,8 can also show similar diagnostic features of narcolepsy on MSLT. It was found that 3.9% to 17% of the general population had multiple SOREMPs,9,10 and 3.6% could meet the MSLT diagnostic criteria of narcolepsy.11 In other words, the MSLT criteria of ICSD-3 would likely give rise to a high false positive rate in diagnosing narcolepsy.

BRIEF SUMMARY

Current Knowledge/Study Rationale: Recent data suggested that altered sleep stage transition of REM sleep might be able to differentiate narcolepsy from other hypersomnias, but its longitudinal stability and association with HLA typing are unclear.

Study Impact: Altered sleep stage transition from stage non-N2/N3 to stage R has good sensitivity and specificity in diagnosing narcolepsy especially with the help of additional sleep variables in the decision tree analysis, which suggests its clinical utility in the differential diagnosis of hypersomnias. The longitudinal stability and association with HLA DBQ1*0602 further demonstrates that altered sleep stage transition is a trait marker of the instability of NREM-REM sleep regulation in narcolepsy.

Some recent studies employed a novel concept of sleep stage sequence analysis to investigate the sleep stage transition prior to the onset of REM sleep (stage R) so as to facilitate the differentiation of narcolepsy from other hypersomnias. Marti et al. found that the frequency of the transition from NREM sleep stage 1 (stage N1) to stage R on MSLT was 71% in patients diagnosed with type 1 narcolepsy (narcolepsy with cataplexy) but only 15% in the patients with BIISS (p < 0.001).7 The sensitivity of this sleep stage sequence analysis was 78.1% and 54.9% for diagnosing type 1 and 2 narcolepsy, respectively, from other hypersomnias, while the specificity was nearly 100%.12 During the NPSG assessment, 15 of 24 (62.5%) patients with type 1 narcolepsy and 8 of 38 (21%) patients with type 2 narcolepsy were found to enter their first REM sleep period (first stage R) from wakefulness (stage W) or stage N1 (stage non-N2/N3), but none was observed in the other groups.13 These studies demonstrated that the altered sleep stage transition of REM sleep might be a key characteristic of NREM-REM sleep dysregulation in narcolepsy. However, previous studies have often been limited by a cross-sectional design with a modest sample size. In addition, before we consider the altered sleep stage transitions of REM sleep as a possible diagnostic marker of narcolepsy, there is a need to investigate whether it is a stable trait associated with familial transmission and human leukocyte antigen (HLA) typing.

In this study, we hypothesized that: (1) the sleep stage transitions of REM sleep are different between narcoleptic and non-narcoleptic patients; (2) altered sleep stage transition of REM sleep is a stable characteristic specific to narcolepsy over time and (3) is associated with specific HLA typing; (4) there is a familial aggregation of altered sleep stage transition of REM sleep in narcoleptic patients; (5) the altered sleep stage transition of REM sleep has a potential diagnostic value by using the decision tree analysis.

METHODS

Study Design and Subject Recruitment

This study consisted of 2 parts, including a retrospective review of our case records to examine the diagnostic value of the sleep stage sequence analysis, using a case-control approach followed by a review of the follow-up data to evaluate the stability of the diagnostic markers. The narcoleptic and control subjects were the consecutive patients who underwent NPSG and MSLT in our sleep center from January 2001 to September 2013. The diagnosis of narcolepsy (type 1 and type 2) was based on ICSD-3 criteria: (1) a complaint of EDS for ≥ 3 months, and/or (2) a definite history of cataplexy, and (3) MSL ≤ 8 min on MSLT with ≥ 2 SOREMPs on NPSG and MSLT, and (4) hypersomnia that is not better explained by other sleep, mental, or medical disorders. As most local Chinese patients were reluctant to undergo lumbar punctures, there were limited data available on CSF hypocretin-1.3 In our previous family study, we identified a group of relatives with narcoleptic spectrum (unaffected) who had one or more SOREMPs on MSLT but did not fulfil the diagnostic criteria of narcolepsy.3,14,15 Another group of relatives who were diagnosed with narcolepsy (affected) were included into the 2 narcoleptic groups. Non-narcoleptic sleep patient controls were selected from our existing database based on the following inclusion criteria: (1) completing both NPSG and MSLT; (2) showing ≥ 1 SOREMP on MSLT; and (3) not having a diagnosis of narcolepsy. The diagnoses of controls included: OSAS,16 circadian rhythm disorder, BIISS, mood disorder, REM sleep behavior disorder (RBD), other NREM parasomnias, idiopathic hypersomnia, and insomnia. The inclusion criterion of at least one nap with SOREMP on MSLT for the unaffected relatives and sleep patient controls allowed the comparison of scoring, coding, and analyzing the sleep stage sequence prior to the onset of REM sleep. Those cases who had clinical follow-up assessments of NPSG and MSLT were included in the second part of the study. All the subjects had signed relevant informed consents of the respective studies,3,14,15 which were approved by the institutional ethics committee.

Questionnaire Assessment and HLA Typing

All the subjects completed Beck Depression Inventory (BDI),17 Ullanlinna Narcolepsy Scale (UNS),18 Epworth Sleepiness Scale (ESS),19 and a one-week sleep log. The usage of medications and continuous positive airway pressure (CPAP) were also recorded.

HLA typing was conducted in a subset of the subjects based on sequencing-based typing (SBT) of HLA-DRB and DQB according to the protocols established by the International Histocompatibility Working Group.20 Sequencing data were analyzed using the SBT engine software (Genome Diagnostics, The Netherlands).

Sleep Stage Sequence Scoring

Sleep Stage Scoring

NPSG and MSLT were performed using CNS SL-1000p polygraph (CNS, Chanhassen, MN), Nicolet Ultrasom NT (Nicolet Biomedical Inc., Madison, USA) and Compumedics Siesta 802 PSG system (Compumedics, Abbotsford, Australia). The NPSG included the measurement of central (C3-A2, C4-A1) and occipital (O2-A1) electroencephalogram (EEG), bilateral electrooculogram (EOG), electromyogram (EMG) of mentalis and bilateral anterior tibialis muscles, electrocardiogram (ECG), respiratory airflow and efforts, arterial oxygen saturation (SpO2), and body position. An MSLT with 5 scheduled naps was performed to determine the sleep latency and SOREMPs on the following day after NPSG for all subjects.21 All raw data of NPSG and MSLT were staged and scored visually by 2 board certified sleep technicians who were blind to the clinical diagnoses of the subjects. For those inconsistent scoring results, a consensus was reached after the raw data being reviewed in a consensus meeting. There was a high inter-rater reliability for the scoring of sleep stage transitions (κ: 0.83), which was comparable to inter-rater (κ: 0.82–0.94) and intra-rater (κ: 0.62–0.99) reliability of the MSLT scoring in our center.22

All the sleep stages were scored in 30-s epochs according to American Academy of Sleep Medicine criteria (2012).23 As the sleep stage transition from stage W/N1 to stage R was not well described and defined by the AASM criteria, we scored the epoch immediately after stage W as definite stage R only if all of the following PSG features were present in more than half of the epoch: (1) change in the background EEG rhythm to low amplitude mixed frequency waves, with the absence of alpha rhythm, vertex sharp wave, K-complex, or sleep spindles; (2) reduction in the chin muscle tone that reached the lowest level of the entire recording; and (3) presence of REMs. If there were ambiguous epochs (for example, all the following features presenting in one 30-s epoch: chin muscle tone decreased to a level that was at the lowest of the entire recording, with low amplitude mixed-frequency activity EEG but without sleep spindles and K complexes, without slow eye movement, and without REMs) between stage W and definite stage R, these epoch(s) were scored as stage N1 based on the recent suggestions of the new AASM scoring criteria.24

Sleep Stage Transition Coding

As the scoring of sleep stages could be easily influenced by the arousals, such as the apnea-related arousals in the OSAS patients, the epoch immediately preceding stage R might not adequately reflect the NREM-REM sleep dysregulation. For this reason, the sleep stage transition coding in the current study focused on the entire sleep stages prior to the stage R using a hierarchical coding method. It is expected that a normal wake-to-sleep transition would show a sequential pattern, going from stage W into stage N1, N2, and sometimes stage N3, prior to stage R. Thus, the sleep stage sequence between stage W and stage R was scored on the basis of a hierarchical order. If stage N3 occurred prior to stage R, regardless of the presence of other sleep stages, the sleep stage sequence prior to stage R would be coded as “sleep stage transition from stage N3 to stage R.” Similarly, the sleep stage sequence would be coded as “sleep stage transition from stage N2 to stage R” or “stage N1 to stage R” when stage N2 (without stage N3) presented prior to stage R, or when N1 and stage W (without stage N3 and N2) occurred before stage R, respectively. The sleep stage transition from stage W to stage R would be coded only when stage W (without stage N3, N2, and N1) occurred immediately prior to stage R.

As we believed that REM sleep transition from stage N2 or N3 is physiological while REM sleep transition from stage W or stage N1 is pathological, we classified the pattern of the sleep stage transitions into 2 dichotomous categories for the analysis: (1) altered transition from stage W/N1 to stage R (described as stage non-N2/N3 to stage R), and (2) normal transition from stage N2/N3 to stage R.

Statistical Analysis

Descriptive data are presented in means ± standard deviation or frequencies (%). The inter-rater reliability of the 2 scorers was estimated by the κ value. Comparisons of age, questionnaire scores, and other sleep parameters between groups were conducted using one-way analysis of variance (ANOVA) with Bonferroni post hoc test or Kruskal-Wallis test. Other tests included χ2 test, Fisher exact test, or Mann-Whitney U test. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and odds ratio (OR) were also calculated. In order to explore the most significant classification value among 5 sleep variables (“MSL ≤ 8 min,” “multiple SOREMPs,” “any nap with transition from stage non-N2/N3 to stage R on MSLT,” “transition from stage non-N2/N3 to first stage R on NPSG,” and “nocturnal SOREMP”) and to assess their combined sensitivity and specificity, a decision tree analysis was performed to differentiate cases with narcolepsy type 1 and type 2 from sleep patient controls. In view of the potential correlations among naps within the same subject, generalized estimating equations (GEE) model was used to explore the association between HLA typing and sleep stage transition from stage non-N2/N3 to stage R. A p value < 0.05 was considered significant except for the post hoc analyses, in which a p value < 0.008 was considered significant following the Bonferroni adjustment due to the concern of multiple comparisons. All the statistical tests were performed using SPSS 20.0 for windows (SPSS Inc., Chicago, IL).

RESULTS

Recruitment, Demographic, Clinical and Polysomnographic Features of the Subjects

In the final analysis we included 74 narcoleptic subjects (type 1 narcolepsy = 35, type 2 narcolepsy = 39) and 26 unaffected relatives with narcoleptic spectrum (who had ≥ 1 SOREMP on MSLT) were recruited from previous clinical and family studies.3,14,15 Another 7 relatives who were diagnosed with narcolepsy (affected) were included into the narcoleptic groups. A total of 159 non-narcoleptic controls who had ≥ 1 SOREMP on MSLT were also included and their diagnoses included: OSAS (75.5%),16 circadian rhythm disorder (22.0%), periodic limb movement disorder (20.8%), BIISS (15.7%), mood disorder (15.1%), REM sleep behavior disorder (RBD) and other parasomnias (8.8%), idiopathic hypersomnia (5.7%), and insomnia (4.4%).

At a mean interval of 37.4 months (SD: 28.1, range: 5–102 months), 24 patients with type 1 narcolepsy, 19 patients with type 2 narcolepsy and 57 sleep patient controls were reassessed by NPSG. Among these subjects, 23 patients with type 1 narcolepsy, 16 patients with type 2 narcolepsy, and 28 sleep patient controls also completed the MSLT reassessment. Thirteen of 39 (33.3%) narcoleptic patients were taking medications at follow-up, including stimulants (17.9%), antidepressants (5.1%), hypnotics (2.6%), and other drugs (12.8%). The detail of the study flowchart is summarized in Figure 1.

Flowchart of subjects recruitment.

NPSG, nocturnal polysomnography; MSLT, multiple sleep latency test; SOREMP, sleep onset REM period.

jcsm.11.8.885a.jpg

jcsm.11.8.885a.jpg
Figure 1

Flowchart of subjects recruitment. NPSG, nocturnal polysomnography; MSLT, multiple sleep latency test; SOREMP, sleep onset REM period.

(more ...)

The demographic, clinical, and polysomnographic characteristics of the study subjects are shown in Table 1. The highest ESS and UNS scores were observed in the patients with type 1 narcolepsy, followed by the patients with type 2 narcolepsy, sleep patient controls, and relatives (p < 0.001). In terms of the NPSG features, the non-narcoleptic sleep patient controls had a higher level of arousal and apnea-hypopnea index (AHI) than the other 3 groups (p < 0.001). As expected, the highest numbers of SOREMPs and the shortest MSL and REM sleep latency were found in patients with type 1 narcolepsy, followed by patients with type 2 narcolepsy (p < 0.001).

Demographic, clinical, and polysomnographic characteristics of subjects.

jcsm.11.8.885.t01.jpg

table icon
Table 1

Demographic, clinical, and polysomnographic characteristics of subjects.

(more ...)

Prevalence of Sleep Stage Transitions of REM Sleep in Different Groups

For the baseline MSLT results, a total of 1,295 naps were scored and 535 naps (41.3%) were found to have SOREMPs. The highest percentage of the naps containing SOREMPs was found in the patients with type 1 narcolepsy (78.9%), followed by the patients with type 2 narcolepsy (63.6%), relatives (30.0%), and sleep patient controls (29.4%). There was a significant difference in the prevalence of altered sleep stage transition from stage non-N2/N3 to stage R among the 4 groups (Figure 2A). Patients with type 1 narcolepsy had significantly more altered transitions from stage non-N2/N3 to stage R (92.0%; stage W to stage R: 39.1%) when compared with patients with narcolepsy type 2 (69.4%; stage W to stage R: 8.1%, p < 0.001), relatives (46.2%; stage W to stage R: 2.6%, p < 0.001), and sleep patient controls (39.3%; stage W to stage R: 0%, p < 0.001). In addition, the percentage of altered sleep stage transition from stage non-N2/N3 to stage R in the patients with type 2 narcolepsy was higher than that of relatives and controls (p = 0.009 and p < 0.001, respectively).

Transition of sleep stage prior to onset of REM sleep on MSLT and NPSG.

(A) MSLT. (B) NPSG. Non-N2/N3, stage W or stage N1; MSLT, multiple sleep latency test; NPSG, nocturnal polysomnography.

jcsm.11.8.885b.jpg

jcsm.11.8.885b.jpg
Figure 2

Transition of sleep stage prior to onset of REM sleep on MSLT and NPSG. (A) MSLT. (B) NPSG. Non-N2/N3, stage W or stage N1; MSLT, multiple sleep latency test; NPSG, nocturnal polysomnography.

(more ...)

A total of 259 NPSG at baseline were analyzed in which there were 35 SOREMPs (Figure 2B). The prevalence of altered transition from stage non-N2/N3 to first stage R on NPSG was 57.1%, 12.8%, and 1.3% in the patients with type 1 narcolepsy, patients with type 2 narcolepsy, and sleep patient controls, respectively. None of the relatives was found to have altered sleep stage transition. Nonetheless, there were no significant differences in the altered sleep stage transition between relatives and sleep patient controls on MSLT and NPSG (p = 0.42 and p = 0.57).

Table 2 shows the diagnostic values of individual parameters (multiple SOREMPs, MSL ≤ 8 minutes, sleep stage transitions, nocturnal SOREMPs) and decision tree analysis in differentiating patients with narcolepsy type 1 and type 2 and relatives from sleep patient controls. When “multiple SOREMPs” and “MSL ≤ 8 min on MSLT” were individually used as the criterion for diagnosing narcolepsy, they had 100% sensitivity but low to moderate specificity (36.5% for ≤ 8 min on MSLT, 69.8% for multiple SOREMPs) in differentiating type 1 and type 2 narcolepsy from sleep patient controls. The variable “any nap with transition from stage non-N2/N3 to stage R on MSLT” had high sensitivity (94.3% and 82.1%) but only moderate specificity (61.6%). The specificity would rise to 98.7% but the sensitivity declined (57.1% and 12.8%) when “transition from stage non-N2/N3 to first stage R on NPSG” was used as the criterion to differentiate type 1 and type 2 narcolepsy from sleep patient controls. “Nocturnal SOREMPs” had comparable diagnostic capacity to that of “transition from stage non-N2/N3 to first stage R on NPSG.” Finally, a decision tree analysis was performed by incorporating a series of sleep variables to classify the cases and controls. The sensitivity was improved to 94.3% and 82.1% and the specificity was increased to 82.4% and 86.8% in differentiating type 1 and type 2 narcolepsy from sleep patient controls, respectively. Interestingly, the variable “MSL ≤ 8 min” did not have a higher differentiating capacity than the other 3 variables in the decision tree analysis for classifying type 1 narcolepsy (Figure 3A). Moreover, “nocturnal SOREMP” did not have a higher differentiating capacity than the other 3 variables in the decision tree analysis for classifying both types of narcolepsy (Figure 3A, 3B).

The diagnostic values of different NPSG/MSLT characteristics.

jcsm.11.8.885.t02.jpg

table icon
Table 2

The diagnostic values of different NPSG/MSLT characteristics.

(more ...)

Decision tree analysis using “MSL ≤ 8 min,” Multiple SOREMPs,” “Any nap with transition from stage non-N2/N3 to stage R on MSLT,” “Transition from stage non-N2/N3 to first stage R on NPSG,” and “Nocturnal SOREMP” as independent variables for differentiating narcolepsy type 1 patients and narcolepsy type 2 patients from controls.

(A) Decision tree shows the differentiation of narcolepsy type 1 from controls (sensitivity: 94.3%; specificity: 82.4%). (B) Decision tree shows the differentiation of narcolepsy type 2 from controls (sensitivity: 82.1%; specificity: 86.8%). N+C, narcolepsy type 1; N-C, narcolepsy type 2; non-N2/N3, stage W or stage N1; NPSG, nocturnal polysomnography; MSLT, multiple sleep latency test; SOREMP, sleep onset REM period; MSL, mean sleep latency.

jcsm.11.8.885c.jpg

jcsm.11.8.885c.jpg
Figure 3

Decision tree analysis using “MSL ≤ 8 min,” Multiple SOREMPs,” “Any nap with transition from stage non-N2/N3 to stage R on MSLT,” “Transition from stage non-N2/N3 to first stage R on NPSG,” and “Nocturnal SOREMP” as independent variables for differentiating narcolepsy type 1 patients and narcolepsy type 2 patients from...

(more ...)

Longitudinal Stability of Sleep Stage Transitions of REM Sleep

With a mean follow-up period of more than 3 years, 24 patients with type 1 narcolepsy (68.6%), 19 patients with type 2 narcolepsy (48.7%), and 57 sleep patient controls (35.8%) completed a PSG reassessment. Among the subjects with the follow-up PSG data, 23 patients with type 1 narcolepsy, 16 patients with type 2 narcolepsy, and 28 sleep patient controls also completed the MSLT reassessment. The longitudinal stability of sleep stage transitions on MSLT for each group is presented in Figure 4. Subjects with the altered sleep stage transition from stage non-N2/N3 to stage R on MSLT and NPSG at baseline were more likely to show this altered sleep stage transition at follow-up. Twenty out of 21 patients with type 1 narcolepsy (90.5%) had “any nap with transition from stage non-N2/N3 to stage R on MSLT” at both baseline and follow-up (Figure 4A). Similarly, the stability of this altered sleep stage transition in patients with type 2 narcolepsy (92.3%) was comparable to that of patients with type 1 narcolepsy (Figure 4B). However, the stability of this sleep stage transition on MSLT was not seen in sleep patient controls (Figure 4C). On the other hand, the normal sleep stage transition from stage N2/N3 to stage R was stable in the majority of sleep patient controls (80%). For NPSG, the pattern of the transition from stage non-N2/N3 to stage R persisted at follow-up in 64.7% of patients with type 1 narcolepsy (Figure 4D). One patient with type 2 narcolepsy had the transition from stage non-N2/N3 to stage R at baseline, which also persisted at follow-up (100%) (Figure 4E). Of note, the normal transition from stage N2/N3 to stage R on NPSG observed at baseline persisted at follow-up in all the sleep patient controls (100%) (Figure 4F).

Persistence of sleep stage transition from stage non-N2/N3 to stage R between baseline and follow-up.

(A–C) Persistence of sleep stage transition from stage non-N2/N3 to stage R between baseline and follow-up on MSLT. (D–F) Persistence of sleep stage transition from stage non-N2/N3 to first stage R between baseline and follow-up on NPSG. MSLT, multiple sleep latency test; NPSG, nocturnal polysomnography.

jcsm.11.8.885d.jpg

jcsm.11.8.885d.jpg
Figure 4

Persistence of sleep stage transition from stage non-N2/N3 to stage R between baseline and follow-up. (A–C) Persistence of sleep stage transition from stage non-N2/N3 to stage R between baseline and follow-up on MSLT. (D–F) Persistence of sleep stage transition from stage non-N2/N3 to first stage R between baseline and follow-up on...

(more ...)

Association between Sleep Stage Transition and HLA Typing

A total of 55 subjects (narcolepsy type 1 = 22, narcolepsy type 2 = 13, relatives = 20) had DBQ1*0602 typing and 32 subjects (narcolepsy type 1 = 21, narcolepsy type 2 = 9, relatives = 2) had DRB1*1501 typing. Using the “transition from stage non-N2/N3 to stage R” as the dependent variable and “HLA typing” as the predicting factor, after adjusting for the diagnostic group, age, sex, AHI, periodic limb movement index, and total sleep time, GEE model revealed that the “transition from stage non-N2/N3 to stage R on MSLT” was significantly associated with DBQ1*0602 typing (OR = 16.0, 95% CI: 1.7–149.8, p = 0.015) and was associated with DRB1*1501 typing with a borderline significance (OR = 11.9, 95% CI: 0.8–174.9, p = 0.07).

DISCUSSION

We found that the altered “sleep stage transition from stage non-N2/N3 to stage R” in both MSLT and NPSG was a sensitive and specific marker of narcolepsy, especially when used in combination of other sleep variables in the decision tree analysis. We also showed that this altered sleep stage transition of REM sleep is a stable marker of narcolepsy over time. Furthermore, GEE model revealed that this altered sleep stage transition was significantly linked to HLA typing especially DBQ1*0602. However, this pattern of the altered sleep stage transition did not seem to aggregate in narcoleptic families.

Although almost one-third of the narcoleptic patients in this study were taking the stimulants and/or psychotropics at follow-up, the pattern of the altered sleep stage transition from stage non-N2/N3 to stage R persisted at the 3-year follow-up. In other words, the altered sleep stage transition of REM sleep might be a stable biomarker of narcolepsy. Not only could this particular pattern of the sleep stage transition be considered as a diagnostic marker, but also it might suggest a trait of the instability of wake-sleep regulation in narcolepsy, which is related to the deficiency of CSF hypocretin-1.2528 Although the exact function of hypocretin-1 is still elusive,29 a flip-flop mechanism of wake-sleep transition, at which hypocretin-1 plays a stabilizing role in preventing the sudden and inappropriate transitions from wake to sleep or vice versa, has been proposed.30 In this regard, the findings of our study lend a further support to the disrupted wake-to-sleep transition in narcolepsy as reflected by the altered sleep stage transition of NREM-REM sleep. Nonetheless, only five subjects in our study had the valid CSF hypocretin-1 results, which limited any further analysis.

In addition, the GEE model revealed a close association of the altered sleep stage transition with HLA DBQ1*0602 and possibly with DRB1*1501 typing. In light of the association of DBQ1*0602 with the shortened REM sleep latency,31,32 the association between DBQ1*0602 and narcolepsy might manifest as the disruption of the normal wake-to-sleep transition, probably through the immune-mediated destruction of hypocretin neurons.33,34

Our previous study reported a significant familial aggregation of narcolepsy symptoms and abnormal MSLT findings.15,35 In this study, the mean sleep latency and the mean REM sleep latency on MSLT was 7.3 min and 5.6 min, respectively for the unaffected relatives, which were comparable to that of the sleep patient controls. There was also a similar prevalence of the transition from stage non-N2/N3 to stage R in the unaffected relatives and sleep patient controls. These findings suggested that the altered sleep stage transition from stage non-N2/N3 to stage R is not a phenomenon of familial transmission but a potential disease marker.

Our present study demonstrated a higher prevalence of the altered sleep stage transition from stage non-N2/N3 to stage R in the narcoleptic patients, especially the direct transition from stage W to stage R, which differed from the findings in the previous studies.7,12,13 The reason for this discrepancy was unclear but there might be several possible explanations, including different scoring and coding methods of the sleep stage sequences especially related to the ambiguous sleep stages that were closely related to narcolepsy, probable inter-center and inter-rater variability, and ethnic differences. The sleep stage sequence analysis used in the present study may potentially have a practical implication for future clinical practice of making a differential diagnosis of hypersomnia. As the scoring of the presence of sleep onset and SOREMP is already part of the routine standard requirement for MSLT, only some additional effort is required to score the individual sleep stages prior to onset of REM sleep (e.g., stage N1, N2 or N3). In other words, the sleep stage sequence scoring is feasible and convenient and could be incorporated as a standard measurement. Nonetheless, further multicenter sleep stage scoring study is warranted to evaluate the inter-center/rater reliability.

Our study supported the observation that the shortened MSL and SOREMPs could occur among various sleep disorders, such as OSAS and BIISS.57 Thus, the observation of the altered sleep stage transition from stage non-N2/N3 to stage R may provide an additional diagnostic value to differentiate narcolepsy from the other types of hypersomnia. A recent study by Andlauer et al. investigated the diagnostic value of the nocturnal SOREMP in narcolepsy.35 They found that the SOREMP in the NPSG was highly specific (99.2%–99.6%) but had only low to moderate sensitivity (35.7%–50.6%). In the present study, we found that using decision tree analysis with the incorporation of a series of sleep variables would result in a comparable specificity but a higher sensitivity than simply using any single sleep variable as the sole diagnostic criterion. Interestingly, the altered “sleep stage transition from stage non-N2/N3 to first stage R” had a significantly higher differentiating capacity than “nocturnal SOREMP.” Further studies will be needed to determine the relative diagnostic advantages and cost-effectiveness of using different biomarkers of narcolepsy.

Limitations

This study is of a retrospective design by examining the data gathered from previous cohorts. This design is inevitably less ideal as compared to a carefully planned prospective study. This study had a relatively modest sample size of narcoleptic patients and relatives, especially for the data from the follow-up assessment and HLA typing. As most local Chinese subjects were reluctant to undergo lumbar punctures,3,36 the lack of CSF hypocretin-1 measurement precluded us to explore the association between the pattern of the altered sleep stage transition and the CSF hypocretin-1 level. In addition, medications may have theoretically affected the sleep stage transitions of the follow-up MSLTs. However, the pattern of the altered sleep stage transition persisted on MSLTs even among the narcoleptic patients taking medications.

CONCLUSION

Altered sleep stage transition from stage non-N2/N3 to stage R is a specific and stable marker of narcolepsy, especially those with cataplexy. There was a strong association between altered sleep stage transition and HLA DBQ1*0602. The altered sleep stage transition on both MSLT and NPSG, especially when combined with other diagnostic PSG variables in the decision tree analysis, has a good sensitivity and specificity in diagnosing narcolepsy, which suggested its clinical utility in the differential diagnosis of hypersomnia. The longitudinal stability and association with HLA DBQ1*0602 suggested the altered sleep stage transition as a trait of the instability of wake-sleep regulation in narcolepsy.

DISCLOSURE STATEMENT

This was not an industry supported study. This work was partly supported by the National Natural Science Foundation of China (81328010, Yun-Kwok Wing). Dr. Yun-Kwok Wing has received sponsorship from Lundbeck Export A/S, Servier Hong Kong Ltd and Celki Medical Company and was a part-time paid consultant for Renascence Therapeutics Limited. The other authors have indicated no conflicts of interest. Dr. Junying Zhou is partly supported by the Faculty Postdoctoral Fellowship Scheme of The Chinese University of Hong Kong. The funding agencies had no role in study design, methods, subject recruitment, data collection, analysis of data, or preparation of the paper.

ABBREVIATIONS

95% CI

95% confidence interval

AHI

apnea-hypopnea index

BDI

Beck Depression Inventory

BIISS

behaviorally induced insufficient sleep syndrome

CSF

cerebrospinal fluid

CPAP

continuous positive airway pressure

ECG

electrocardiogram

EDS

excessive daytime sleepiness

EEG

electroencephalogram

EMG

electromyogram

EOG

electrooculogram

ESS

Epworth Sleepiness Scale

GEE

generalized estimating equations

HLA

human leukocyte antigen

ICSD

International Classification of Sleep Disorders

MSL

mean sleep latency

MSLT

multiple sleep latency test

NA

not available

NPSG

nocturnal polysomnography

NPV

negative predictive value

OR

odds ratio

PLMI

periodic limb movement index

PPV

positive predictive value

REM SL

REM sleep latency

SD

standard deviation

SOREMP

sleep onset REM period

SpO2

arterial oxygen saturation

TIB

time in bed

TST

total sleep time

UNS

Ullannina Narcolepsy Scale

WASO

wake after sleep onset

ACKNOWLEDGMENTS

Author contributions: All authors contributed to the collection of the data, analysis and preparation of the manuscript. Dr. Yun-Kwok Wing, the corresponding author, had full access to all the data in the study and take responsibility for the integrity and accuracy of the data analysis.

REFERENCES

1 

American Academy of Sleep Medicine. International classification of sleep disorders. 3rd ed. Darien, IL: American Academy of Sleep Medicine, 2014.

2 

Scammell TE, author. The neurobiology, diagnosis, and treatment of narcolepsy. Ann Neurol. 2003;53:154–66. [PubMed]

3 

Wing YK, Chen L, Fong SY, et al., authors. Narcolepsy in Southern Chinese patients: clinical characteristics, HLA typing and seasonality of birth. J Neurol Neurosurg Psychiatry. 2008;79:1262–7. [PubMed]

4 

Johns MW, author. Sensitivity and specificity of the multiple sleep latency test (MSLT), the maintenance of wakefulness test and the epworth sleepiness scale: failure of the MSLT as a gold standard. J Sleep Res. 2000;9:5–11. [PubMed]

5 

Chervin RD, Aldrich MS, authors. Sleep onset REM periods during multiple sleep latency tests in patients evaluated for sleep apnea. Am J Respir Crit Care Med. 2000;161:426–31. [PubMed]

6 

Aldrich MS, Chervin RD, Malow BA, authors. Value of the multiple sleep latency test (MSLT) for the diagnosis of narcolepsy. Sleep. 1997;20:620–9. [PubMed]

7 

Marti I, Valko PO, Khatami R, Bassetti CL, Baumann CR, authors. Multiple sleep latency measures in narcolepsy and behaviourally induced insufficient sleep syndrome. Sleep Med. 2009;10:1146–50. [PubMed]

8 

Takeuchi T, Fukuda K, Murphy TI, authors. Elicitation of sleep-onset REM periods in normal individuals using the sleep interruption technique (SIT). Sleep Med. 2002;3:479–88. [PubMed]

9 

Bishop C, Rosenthal L, Helmus T, Roehrs T, Roth T, authors. The frequency of multiple sleep onset REM periods among subjects with no excessive daytime sleepiness. Sleep. 1996;19:727–30. [PubMed]

10 

Singh M, Drake CL, Roth T, authors. The prevalence of multiple sleep-onset REM periods in a population-based sample. Sleep. 2006;29:890–5. [PubMed]

11 

Mignot E, Lin L, Finn L, et al., authors. Correlates of sleep-onset REM periods during the Multiple Sleep Latency Test in community adults. Brain. 2006;129:1609–23. [PubMed]

12 

Drakatos P, Suri A, Higgins SE, et al., authors. Sleep stage sequence analysis of sleep onset REM periods in the hypersomnias. J Neurol Neurosurg Psychiatry. 2013;84:223–7. [PubMed]

13 

Drakatos P, Kosky CA, Higgins SE, Muza RT, Williams AJ, Leschziner GD, authors. First rapid eye movement sleep periods and sleep-onset rapid eye movement periods in sleep-stage sequencing of hypersomnias. Sleep Med. 2013;14:897–901. [PubMed]

14 

Chen L, Fong SY, Lam CW, et al., authors. The familial risk and HLA susceptibility among narcolepsy patients in Hong Kong Chinese. Sleep. 2007;30:851–58. [PubMed Central][PubMed]

15 

Wing YK, Chen L, Lam SP, et al., authors. Familial aggregation of narcolepsy. Sleep Med. 2011;12:947–51. [PubMed]

16 

Fong SYY, Ho CKW, Li AM, Wing YK, authors. Excessive daytime sleepiness in obstructive sleep apnea patients after home CPAP treatment: a long-term outcome study. Sleep Biol Rhythms. 2009;7:193–200.

17 

Shek DT, author. Reliability and factorial structure of the Chinese version of the Beck Depression Inventory. J Clin Psychol. 1990;46:35–43. [PubMed]

18 

Wing YK, Li RH, Ho CK, Fong SY, Chow LY, Leung T, authors. A validity study of Ullanlinna Narcolepsy Scale in Hong Kong Chinese. J Psychosom Res. 2000;49:355–61. [PubMed]

19 

Fong SY, Ho CK, Wing YK, authors. Comparing MSLT and ESS in the measurement of excessive daytime sleepiness in obstructive sleep apnoea syndrome. J Psychosom Res. 2005;58:55–60. [PubMed]

20 

Kimura A, Sasazuki T, authors. Eleventh International Histocompatibility Workshop reference protocol for the HLA DNA-typing technique. HLA. 1991;1:397–419.

21 

Carskadon MA, Dement WC, Mitler MM, Roth T, Westbrook PR, Keenan S, authors. Guidelines for the multiple sleep latency test (MSLT): a standard measure of sleepiness. Sleep. 1986;9:519–24. [PubMed]

22 

Chen L, Ho CK, Lam VK, et al., authors. Interrater and intrarater reliability in multiple sleep latency test. J Clin Neurophysiol. 2008;25:218–21. [PubMed]

23 

Iber C, Ancoli-Israel S, Chesson A, Quan SF, authors; for the American Academy of Sleep Medicine. The AASM manual for the scoring of sleep and associated events: rules, terminology and technical specifications. 1st ed. Westchester, IL: American Academy of Sleep Medicine, 2007.

24 

American Academy of Sleep Medicine Scoring Manual Steering Committee. Scoring Manual FAQ. [cited 2014 May 25]; Available from: http://www.aasmnet.org/scoringmanualfaq.aspx#63.

25 

Mignot E, Young T, Lin L, Finn L, authors. Nocturnal sleep and daytime sleepiness in normal subjects with HLA-DQB1*0602. Sleep. 1999;22:347–52. [PubMed]

26 

Nishino S, Ripley B, Overeem S, Lammers GJ, Mignot E, authors. Hypocretin (orexin) deficiency in human narcolepsy. Lancet. 2000;355:39–40. [PubMed]

27 

Peyron C, Faraco J, Rogers W, et al., authors. A mutation in a case of early onset narcolepsy and a generalized absence of hypocretin peptides in human narcoleptic brains. Nat Med. 2000;6:991–7. [PubMed]

28 

Sorensen GL, Knudsen S, Jennum P, authors. Sleep transitions in hypocretin-deficient narcolepsy. Sleep. 2013;36:1173–7. [PubMed Central][PubMed]

29 

Baumann CR, Bassetti CL, authors. Hypocretins (orexins) and sleep-wake disorders. Lancet Neurol. 2005;4:673–82. [PubMed]

30 

Saper CB, Chou TC, Scammell TE, authors. The sleep switch: hypothalamic control of sleep and wakefulness. Trends Neurosci. 2001;24:726–31. [PubMed]

31 

Mignot E, Hayduk R, Black J, Grumet FC, Guilleminault C, authors. HLA DQB1*0602 is associated with cataplexy in 509 narcoleptic patients. Sleep. 1997;20:1012–20. [PubMed]

32 

Mignot E, Young T, Lin L, Finn L, authors. Nocturnal sleep and daytime sleepiness in normal subjects with HLA-DQB1*0602. Sleep. 1999;22:347–52. [PubMed]

33 

Partinen M, Kornum BR, Plazzi G, Jennum P, Julkunen I, Vaarala O, authors. Narcolepsy as an autoimmune disease: the role of H1N1 infection and vaccination. Lancet Neurol. 2014;13:600–13. [PubMed]

34 

Han F, Lin L, Warby SC, et al., authors. Narcolepsy onset is seasonal and increased following the 2009 H1N1 pandemic in China. Ann Neurol. 2011;70:410–17. [PubMed]

35 

Andlauer O, Moore H, Jouhier L, et al., authors. Nocturnal rapid eye movement sleep latency for identifying patients with narcolepsy/hypocretin deficiency. JAMA Neurol. 2013;70:891–902. [PubMed Central][PubMed]

36 

Wing YK, Li RH, Lam CW, Ho CK, Fong SY, Leung T, authors. The prevalence of narcolepsy among Chinese in Hong Kong. Ann Neurol. 2002;51:578–84. [PubMed]