Self-reported sleep disturbances are associated with osteoporosis: multivariable-adjusted and Mendelian randomization analyses in UK Biobank

Experimental studies suggest an association between sleep disturbances and osteoporosis risk, but epidemiological data remain inconclusive. This study investigated associations and possible causality between 4 sleep traits and osteoporosis risk, as well as BMD, in a large population-based cohort. We analyzed 402 533 UK Biobank participants with no history of osteoporosis at baseline (44.8% men, mean age 56.6 yr [SD 8.1], median follow-up 13.1 [IQR 12.8-14.4] yr). Multivariable-adjusted regression analyses assessed the associations between self-reported sleep traits at baseline and osteoporosis incidence, and BMD T-scores at the femoral neck, lumbar spine and radius. Two-sample Mendelian randomization (MR) was employed to provide evidence of potential causality. Self-reported short (<7 h) and long (>8 h) sleep durations, insomnia symptoms, daytime dozing, and evening chronotype were all associated with increased osteoporosis incidence. Conversely, no associations were observed between sleep traits and T-scores, except that an evening chronotype was associated with lower femoral neck T-score. Having a greater number of poor sleep behaviors was associated with increased osteoporosis risk and lower T-scores. MR did not support a causal relationship between sleep traits and osteoporosis risk or BMD. Since all sleep behaviors are associated with osteoporosis risk, assessing sleep patterns could be valuable to identify individuals-at-risk. However, the absence of causal evidence and limited associations with BMD suggest that sleep disturbances do not influence bone remodeling directly. Instead, the interaction between sleep and osteoporosis may involve unidentified mechanisms requiring further investigation.