Metabolic syndrome and risk of incident diabetes: findings from the European Prospective Investigation into Cancer and Nutrition-Potsdam Study

Abstract

Background: Several aspects concerning the relationship between the metabolic syndrome and incident diabetes are incompletely understood including the magnitude of the risk estimate, potential gender differences in the associations between the metabolic syndrome and incident diabetes, the associations between the components of the metabolic syndrome and incident diabetes, and whether the metabolic syndrome provides additional prediction beyond its components. To shed light on these issues, we examined the prospective association between the metabolic syndrome defined by the National Cholesterol Education Program (NCEP) and International Diabetes Federation (IDF) and diabetes.

Methods: We used data for 2796 men and women aged 35-65 years from the European Prospective Investigation into Cancer and Nutrition-Potsdam Study followed for an average of 6.9 years. This analysis employed a case-cohort design that included 697 participants who developed diabetes and 2099 participants who did not. Incident diabetes was identified on the basis of self-reports and verified by contacting the patient's attending physician.

Results: The adjusted hazard ratio for the NCEP definition was 4.62 (95% confidence interval [CI]: 3.90-5.48) and that for the IDF definition was 4.59 (95% CI: 3.84-5.50). The adjusted hazard ratios for the NCEP but not IDF definition were higher for women than men. When participants who had no cardiometabolic abnormalities were used as the reference group for the NCEP definition, the adjusted hazard ratio for having 3 or more abnormalities increased to 22.50 (95% CI: 11.21-45.19). Of the five components, abdominal obesity and hyperglycemia were most strongly associated with incident diabetes.

Conclusion: In this study population, both definitions of the metabolic syndrome provided similar estimates of relative risk for incident diabetes. The increase in risk for participants with the metabolic syndrome according to the NCEP definition was very large when contrasted with the risk among those who had no cardiometabolic abnormalities.

Figure 1

Hazard ratios and 95% confidence intervals for incident diabetes among 2796 participants aged 35–65 years, by number of cardiometabolic abnormalities, EPIC-Potsdam 1994–1998 to 2005. Hazard ratios are adjusted for age, sex, educational status, smoking status, alcohol use, occupational activity, physical activity, concentrations of total cholesterol and C-reactive protein, and other components of the metabolic syndrome.