BACKGROUND: This study aimed to investigate the relationship between serum total bilirubin (TBIL) and diabetes mellitus (DM) in a large health checkup population in China. METHODS: A cross-sectional study was conducted in 2011-2016 at the Physical Examination Center of the First Affiliated Hospital of Wannan Medical College, involving 302,515 subjects. Logistic regression and restricted cubic splines were used to determine the associations between TBIL and DM. The relationship between TBIL and fasting plasma glucose (FPG) was explored through the application of generalized linear models. RESULTS: Quartiles of TBIL were negatively associated with DM after multivariable adjustment, compared with Q1, the ORs of Q2, Q3 and Q4 for DM were 0.971 (95% CI: 0.923-1.022), 0.922 (95% CI: 0.876-0.971), and 0.880 (95% CI: 0.836-0.927), respectively (p < 0.001). The restricted cubic spline model showed a non-linear relationship between TBIL and DM (p < 0.001), with the ORs for DM increasing progressively, reaching a threshold at the TBIL level of 14.54 µmol/L. The relationship between TBIL and FPG was also non-linear, with a threshold at TBIL of 20.10 µmol/L. FPG decreased with TBIL when TBIL < 20.10 µmol/L and increased when TBIL ≥ 20.10 µmol/L (p < 0.001). After stratifying by gender and age, the results revealed similar patterns. CONCLUSION: Our study identified that moderate to low levels of TBIL are significantly negatively correlated with DM. In contrast, high levels of TBIL are positively correlated with DM, with a threshold of 14.54 µmol/L. A non-linear relationship was also observed between TBIL and FPG, with a threshold of 20.10 µmol/L.
