Does sleep affect weight gain? Assessing subjective sleep and polysomnography measures in a population-based cohort study (CoLaus/HypnoLaus)

Study Objectives: Although several studies have linked short and long sleep duration to weight gain, mixed results exist. Contrarily, few studies associated objectively measured sleep characteristics with weight gain. We investigated the association between several sleep characteristics measured by questionnaire and polysomnography with prospective weight gain in a population-based, middle-aged cohort. Methods: Three samples were analyzed: sample 1 (n = 2551, 47.3% men, 56.9 +/- 10.3 years) had data for subjective sleep characteristics, sample 2 (n = 1422, 49.4% men, 57.6 +/- 10.4 years) had objective sleep assessment (polysomnography), and sample 3 consisting of 1259 subjects included in both samples. Multivariable logistic regressions were performed to assess the relationship between sleep characteristics and >= 5 kg weight gain during a median follow-up of 5.3 years. Results: In both study samples, 12% of the subjects gained >= 5 kg during follow-up. Multivariable analyses showed poor subjective sleep quality (as assessed by Pittsburgh Sleep Quality Index: odds ratio [95% confidence interval] = 1.54 [1.19 to 1.99]), percentage of sleep spent in stage 2 (1.32 [1.10 to 1.58]), and less than 90% oxygen saturation (SpO(2) < 90) (1.23 [1.07 to 1.41]); moderate/severe Oxygen Desaturation Index (1.70 [1.01 to 2.85]) and autonomic arousal duration (1.22 [1.02 to 1.45]) were related to >= 5 kg weight gain. Only poor subjective sleep quality was robustly associated with weight gain in all sensitivity analyses, except in female subsamples. Conclusions: Poor subjective sleep quality, and to some extent moderate to severe oxygen desaturation, but no other sleep characteristics, were robustly associated with weight gain. Future studies should confirm the relationship between sleep quality and weight gain, assess sex differences, and investigate underlying mechanisms.