Abstract: Abstract: Objective To evaluate whether serum complement C3 concentration was associated with the prevalence and incidence of prediabetes in an adult population. Methods A cross-sectional (n=10 539) and prospective cohort (n= 3 064, followed up for -6 years, mean: 2.8 y) study was performed on subjects recruited from the Health Management Center of Tianjin Medical University General Hospital in Tianjin. Measurements of serum C3 concentration, blood fasting glucose and other potential confounding factors were assessed at baseline and per year during the follow-up period. Prediabetes was defined according to the criteria of American Diabetes Association. Adjusted Logistic and Cox proportional hazards regression models were used to assess the associations between C3 quintiles and prediabetes. Results The prevalence and incidence of prediabetes were 19.9% and 99.5 per 1 000 person- year, respectively. In cross- sectional analysis, after adjusted for potential confounders, the odds ratios (95% confidence interval) of prediabetes for increasing quintiles of C3 were 1.00 (reference), 1.18 (0.98-1.42), 1.11 (0.92-1.34), 1.38 (1.15-1.65) and 1.63 (1.36-1.95) (P for trend <0.000 1). In cohort analysis, in the final multivariate models, the hazard ratios (95% confidence interval) for prediabetes across C3 quintiles were 1.00 (reference), 1.20 (0.94-1.54), 1.48 (1.16-1.88), 1.38 (1.09-1.76) and 1.53 (1.21-1.95) (P for trend < 0.001), respectively. Conclusion The study suggests that the elevated C3 level is significantly associated with the prevalence and incidence of prediabetes, which means that C3 can be used as a biomarker in early prevention of prediabetes and diabetes.

Key words: prediabetic state, complement C3, inflammation, prevalence, cross-sectional studies, cohort studies