Data_Sheet_1_Linear and non-linear Mendelian randomization analyses of sex-specific associations between sleep duration and hyperuricemia.docx

BackgroundObservational studies have suggested a potential non-linear association between sleep duration and hyperuricemia. However, the causal nature and sex-specific differences are poorly understood. We aimed to determine the shape of sex-specific causal associations between sleep duration and hyperuricemia in the UK Biobank.
MethodsLogistic regression was used to investigate the observational association between self-reported sleep duration and hyperuricemia among 387,980 white British participants (mean age: 56.9 years and 46.0% males). Linear and non-linear Mendelian Randomization (MR) analyses were performed to assess the causal association between continuous sleep duration and hyperuricemia. The causal effects of genetically predicted short (<7 h) and long (>8 h) sleep durations on hyperuricemia were further estimated, respectively.
ResultsTraditional observational analysis suggested U- and J-shaped associations between sleep duration and hyperuricemia in females and males, respectively. Linear MR did not support the causal effect of sleep duration on hyperuricemia. Non-linear MR demonstrated an approximately U-shaped causal association between continuous sleep duration and hyperuricemia in overall participants and females, but not in males. Genetically predicted short sleep duration was significantly associated with hyperuricemia in females (OR [95% CI]: 1.21 [1.08–1.36]; P = 0.001), but not in males (1.08 [0.98–1.18]; P = 0.137). By contrast, genetically predicted long sleep duration was not significantly associated with the risk of hyperuricemia in either females or males.
ConclusionGenetically predicted short sleep duration is a potential causal risk factor for hyperuricemia for females but has little effect on males. Long sleep duration does not appear to be causally associated with hyperuricemia.