RESUMO
BACKGROUND: Sleep disturbances are a prominent feature of bipolar disorder (BP). However, it remains unclear how sleep phenotypes may evolve among at-risk youth, and their relevance to BP onset. METHODS: Pittsburgh Bipolar Offspring Study (BIOS) offspring (ages 10-18) and their parents completed assessments approximately every two years pertaining to current psychopathology and offspring sleep habits. A latent transition analysis (LTA) identified latent sleep groups within offspring based on their ratings of six sleep domains using the School Sleep Habits Survey. Demographic and clinical characteristics were compared between sleep groups. Logistic regression tested links between sleep group and BP onset at the subsequent assessment. RESULTS: The LTA model identified latent groups of good, poor, and variable sleepers. We observed an overall trend of good sleep becoming variable, and then poor, as youth age. Offspring in the poor sleep group were more likely to have psychopathology. Adjusting for age and depression, poor sleepers had nearly twice the odds of developing BP relative to good (OR=1.99, CI=0.45-8.91) or variable (OR=2.03, CI=0.72-5.72) sleepers. LIMITATIONS: Limitations include the use of proximal sleep phenotypes to predict BP onset, and a self-report measure of sleep CONCLUSIONS: We found three non-overlapping sleep phenotype groups in a large sample of offspring of bipolar probands and offspring of demographically-matched community control parents. Clinicians should consider that youth will likely experience variable and/or poor sleep as they age, and that at-risk youth with poor sleep may be at increased risk of developing MDD and BP at their next assessment.