The role of iron biomarkers in predicting type 2 Diabetes: An international, multi-cohort Study.

Abstract

We investigated whether adding iron biomarkers into existing type 2 diabetes risk models improves risk prediction. Data from three population-based cohorts were used; CoLaus|PsyCoLaus in Switzerland (5250 participants, 54.9 % females, mean age± standard deviation 51.8 ± 10.5 years, median follow-up of 14.5 years; PREVEND in the Netherlands (4784 participants, 51.8 % females, 52.2 ± 11.5 years, follow-up 7.3 years); and EPIPorto in Portugal (806 participants, 40 % females, 62 ± 13 years, follow-up 7.8 years). The effect of adding iron, ferritin, and transferrin in seven type 2 diabetes risk models was examined. 486 participants (9.3 %) in the CoLaus|PsyCoLaus, 170 (3.6 %) in PREVEND, and 22 (3.4 %) in EPIPorto developed diabetes. There was a substantial association between type 2 diabetes and all risk scores. In the CoLaus|PsyCoLaus and PREVEND, ferritin levels were positively and independently associated with the incidence of diabetes and considerably enhanced its prediction. Transferrin levels were positively and independently associated with the incidence of diabetes across all risk scores in all cohorts and improved its prediction in PREVEND, EpiPorto, and certain risk models in CoLaus|PsyCoLaus. There was found to be no association between iron levels and type 2 diabetes. Adding ferritin or transferrin slightly improved most diabetes risk prediction models.