Background: Obstructive sleep apnea (OSA) is one of the leading non-psychiatric comorbidities in bipolar disorders (BD). We sought to explore associations between risk of OSA in BD, clinical characteristics alongside with both subjective sleep complaints and objective sleep abnormalities.
Methods: Euthymic patients with BD (n = 144) were assessed over a three-week period, by actigraphy, clinical interviews and questionnaires.
Results: Of the study sample, 32 (22%) individuals were at high risk of OSA (HR-OSA) and 112 (78%) had a low risk (LR-OSA), as assessed with the Berlin questionnaire. HR-OSA, compared to LR-OSA, were older (p = 0.031), had higher BMI (p < 0.0005), larger neck circumference (p = 0.002), and more residual depressive symptoms (p < 0.0005). HR-OSA was also associated with greater sleepiness (p = 0.003), poorer sleep quality (p = 0.003), insomnia complaints (p = 0.027), "languid" chronotype (p = 0.002), and higher actigraphy-derived fragmentation index (p = 0.015). Backward stepwise linear regression retained BMI and depressive symptoms (correct classification of 83% of participants). Classification increased up to 85.4% when adding sleepiness and languid-vigorous scales and up to 87.8% when adding fragmentation index. Combining ROC curve analysis and Youden Index provided best cut-offs (HR-OSA if cut-off greater than or equal to) of 29.84 for BMI (Sensibility(Se) = 0.47, Specificity(Spe) = 0.96) and 1.5 for MADRS total score (Se = 0.84, Spe = 0.58).
Limitations: No confirmation of OSA diagnosis with polysomnography.
Conclusions: Higher BMI and residual depressive symptoms are the two best independent predictors of OSA in BD. Such information contributes to improving the screening and management of OSA in BD.
Trial number: NCT02627404.
Keywords: Bipolar disorder; Circadian rhythms; Depression; Obesity; Sleep.
Copyright © 2018 Elsevier B.V. All rights reserved.