Metabolic syndrome and incident diabetes: current state of the evidence.

Abstract

Our objective was to perform a quantitative review of prospective studies examining the association between the metabolic syndrome and incident diabetes. Using the title terms "diabetes" and "metabolic syndrome" in PubMed, we searched for articles published since 1998. Based on the results from 16 cohorts, we performed a meta-analysis of estimates of relative risk (RR) and incident diabetes. The random-effects summary RRs were 5.17 (95% CI 3.99-6.69) for the 1999 World Health Organization definition (ten cohorts); 4.45 (2.41-8.22) for the 1999 European Group for the Study of Insulin Resistance definition (four cohorts); 3.53 (2.84-4.39) for the 2001 National Cholesterol Education Program definition (thirteen cohorts); 5.12 (3.26-8.05) for the 2005 American Heart Association/National Heart, Lung, and Blood Institute definition (five cohorts); and 4.42 (3.30-5.92) for the 2005 International Diabetes Federation definition (nine cohorts). The fixed-effects summary RR for the 2004 National Heart, Lung, and Blood Institute/American Heart Association definition was 5.16 (4.43-6.00) (six cohorts). Higher number of abnormal components was strongly related to incident diabetes. Compared with participants without an abnormality, estimates of RR for those with four or more abnormal components ranged from 10.88 to 24.4. Limited evidence suggests fasting glucose alone may be as good as metabolic syndrome for diabetes prediction. The metabolic syndrome, however defined, has a stronger association with incident diabetes than that previously demonstrated for coronary heart disease. Its clinical value for diabetes prediction remains uncertain.