Complement C3 Associates With Incidence of Diabetes, but No Evidence of a Causal Relationship

Purpose: This study explored whether complement factor 3 (C3) in plasma is associated with incidence of diabetes in a population-based cohort. We also identified genetic variants related to C3 and explored whether C3 and diabetes share common genetic determinants. Methods: C3 was analyzed in plasma from 4368 nondiabetic subjects, 46 to 68 years old, from the Malmo Diet and Cancer Study. Incidence of diabetes was studied in relationship to C3 levels during 17.7 +/- 4.4 years of follow-up. Genotypes associated with C3 were identified in a genome-wide association study. Diabetes Genetics Replication and Meta-Analysis and the European Genetic Database were used for in silico look-up. Results: In all, 538 (12.3%) subjects developed diabetes during 18 years of follow-up. High C3 was significantly associated with incidence of diabetes after risk factor adjustments (hazard ratio comparing 4th vs 1st quartile, 1.54 (95% confidence interval, 1.13 to 2.09; P = 0.005). C3 was associated with polymorphisms at the complement factor H locus (P < 10(-8)). However, no relationship with diabetes was observed for this locus. Another eight loci were associated with C3 with P < 10(-5). One of them, the glucose kinase regulatory protein (GCKR) locus, has been previously associated with diabetes. The relationship between C3 levels and the GCKR locus was replicated in the European Genetic Database cohort. Conclusions: Plasma concentration of C3 is a risk marker for incidence of diabetes. The results suggest that this association could, in part, be explained by pleiotropic effects related to the GCKR gene.