RESUMO
BACKGROUND: The association between testosterone concentrations and sleep duration is poorly understood. OBJECTIVE: To evaluate the association between sleep duration and quality with serum testosterone concentrations and its variation by sex and age. METHODS: Data were analyzed for 8748 men and women (≥20 years old) who participated in the cycles of the National Health and Nutrition Examination Survey 2011-2016, a cross-sectional study. Total testosterone (ng/dL) was measured and categorized (low, moderate, and high) based on established cut-offs for men and its tertile distribution among women. Sleep duration was classified as ≤6, 7-8, and ≥9 h. Sleep quality was classified as poor or good based on the frequency of trouble falling or staying asleep or sleeping too much. Weighted multivariable adjusted and multinomial logistic regression models were conducted to assess these associations. RESULTS: The association between sleep duration and testosterone concentrations, varied according to sex and age. Sleep deprivation (≤6 h) was associated with high testosterone (odds ratio = 3.62; 95% confidence interval: 1.37, 9.53) among young men (20-40 years old); meanwhile, middle-aged men (41-64 years old) who reported more sleep duration had low testosterone (odds ratio = 2.03; 95% confidence interval: 1.10, 3.73). A J-shaped association between sleep duration and low testosterone (odds ratio≤6 h = 1.57; 95% confidence interval: 1.10, 2.27; odds ratio≥9 h = 2.06; 95% confidence interval: 1.18, 3.59) was observed in women aged 41-64 years. We did not find any association with sleep quality. CONCLUSION: The association of sleep duration with serum testosterone concentrations varies with sex and age group. Prospective studies are warranted to confirm these sex and age group differences.