Hypersomnia: an overlooked, but not overestimated, sleep disturbance in bipolar disorder

ABSTRACT FROM: Steinan MK, Scott J, Lagerberg TV, et al. Sleep problems in bipolar disorders: more than just insomnia. Acta Psychiatr Scand 2016;133:368–77.

What is already known on this topic

Sleep disturbances are common across bipolar spectrum disorders. Reduced need for sleep, insomnia and hypersomnia (excessive sleep or sleepiness) are diagnostic features of illness episodes; these sleep disturbances persist into the interepisode period and are associated with functional impairment and risk for relapse.¹ Few studies have investigated subtypes of sleep disturbances within bipolar spectrum disorders, and these have been limited to specific illness states (eg, depression)² or to specific bipolar subtypes (eg, bipolar I disorder).³

Methods of the study

This cross-sectional study recruited 563 inpatient and outpatient adults with bipolar disorder types I and II (BD I and BD II) across Norway. Two semistructured clinician-administered interviews, the Inventory of Depressive Symptomatology (IDS-C) and Young Mania Rating Scale (YMRS), were used to assess mood state. Sleep disturbance profiles were established using items from the IDS-C (items 1–3 for insomnia, item 4 for hypersomnia). Participants were separated into three groups (insomnia, hypersomnia or no sleep problems), and univariate and multivariate associations between sleep disturbance and mood/demographic variables were explored.

What this paper adds

• ▸Steinan et al extend previous work by examining subtypes of sleep disturbance across a range of mood states and subtypes of bipolar disorder. They recruited a sizeable sample sufficient for exploring clinical and demographic features that might predict sleep disturbance subtypes.
• ▸While insomnia was common overall (43%), a significant number of participants also endorsed hypersomnia (29%); both sleep disturbance subtypes outnumbered those with no sleep problems (28%). Examining sleep disturbances by subtype and mood, 35% of individuals with BD I endorsed hypersomnia while euthymic, consistent with previous reports of long interepisode sleep durations.⁴ Participants with BD II were twice as likely to report insomnia as to report hypersomnia or no sleep problems.
• ▸Comparing the sleep subgroups, participants with hypersomnia were significantly younger than those with insomnia or no sleep problems (35 vs 40 and 39, respectively) and were more likely to be prescribed antidepressants (42% vs 35% and 23%, respectively), whereas those with insomnia were more often prescribed benzodiazepines and hypnotics (approximately one-third of the insomnia subgroup vs one-quarter of the no sleep problems subgroup and one-fifth of the hypersomnia subgroup). In a multiple regression model, when compared with the no sleep problems group, the hypersomnia and insomnia groups were more likely to be in a depressed (ORs of 6.26 and 9.46, respectively) or mixed (ORs of 7.21 and 21.38) state.

Limitations

• ▸In forcing individuals into one of three sleep disturbance groups, the authors may have created false categories. Insomnia and hypersomnia in BD are known to overlap at rates as high as 31%,² and previous investigations of sleep disturbance subtypes have acknowledged and allowed for their co-occurrence.³ Patients who report hypersomnia and insomnia symptoms tend to suffer from more severe BD.² Without a fourth group investigating the co-occurrence of these symptoms, the representativeness of remaining groups is unclear.
• ▸Sleep disturbance and depression severity were both determined using the IDS-C. While the authors note cut-off scores were raised to ensure mood states were not artificially elevated by endorsement of sleep disruption, removing the sleep items from the mood scores and examining symptomatology continuously would have offered stronger evidence of the relationship between sleep and mood disturbance.
• ▸The study defined insomnia and hypersomnia using single-item sleep ratings derived from mood scales, instead of using more widely used and comprehensive diagnostic sleep questionnaires. Further, hypersomnia classification in this study was based solely on total sleep time, not excessive sleepiness, though the latter is also a diagnostic feature of hypersomnia.

What is next in research and why

Longitudinal approaches, ideally spanning 6 months or more, are needed to better examine dynamic relationships between BD subtypes, mood states and sleep changes. Furthermore, research that examines the co-occurrence of specific sleep disturbances will help clarify their relationship to prospective mood changes. Finally, validated and comprehensive sleep assessments including objective measures of sleep are needed, particularly for investigation of hypersomnia.

Do these results change your practices and why?

Yes. This research echoes previous findings that sleep disturbances in BD are quite common, and co-occur with illness states. It further underscores the importance of hypersomnia in BD, which appears particularly relevant to BD I. Given that treatment of sleep disturbances may influence illness course in BD, providers should be mindful to assess and treat sleep disturbances in BD.^{5 6} Non-pharmacological approaches to hypersomnia treatment are sorely needed.