Sex differences in the relationship between serum total bilirubin and risk of incident metabolic syndrome in community-dwelling adults: Propensity score analysis using longitudinal cohort data over 16 years

Abstract

Abstract Background Research on identifiable risks for metabolic syndrome (MetS) is ongoing, and growing evidence suggests that bilirubin is a potent antioxidant and cytoprotective agent against MetS. However, there have been conflicting results on the association between bilirubin and MetS. Our study aimed to validate the association by separately stratifying data for men and women in a longitudinal prospective study. Methods Data were derived from the Korean Genome Epidemiology Study provided by the Korea Centers for Disease Control and Prevention. Data from 5,185 adults aged 40–69 years (3,089 men and 2,096 women) without MetS were analyzed. The participants were divided according to sex-specific quartiles of serum total bilirubin levels and followed up biennially for 16 years (until 2018). The log-rank test was used for obtaining the Kaplan-Meier curves of cumulative incidence of MetS according to sex-specific serum total bilirubin quartiles, and the hazard ratios (HRs) with 95% confidence intervals (CIs) for incident metabolic syndrome were analyzed with a multiple Cox proportional hazard regression analysis model, after propensity score matching for removing differences at baseline. Results With increasing serum total bilirubin quartiles, the incidence rate per 1000 person-years proportionally decreased in both men and women. After propensity score matching and adjusting for confounding variables, the HRs (95% CIs) for MetS of the highest quartile in reference to the lowest quartile were 1.00 (0.80–1.24) for men and 0.80 (0.65–0.99) for women. Higher quartiles of serum total bilirubin showed significantly lower cumulative incidence of MetS in women (log-rank test p = 0.009), but not in men (log-rank test p = 0.285). Conclusion Serum total bilirubin levels were significantly inversely associated with MetS in women, but there was no significant association observed in men. Sex differences in the effects of serum total bilirubin should be noted when predicting incident MetS by sex in clinical settings.