Background: Our purpose was to estimate the incidence of type 2 diabetes among a high risk population with or without impaired glucose tolerance (IGT), analysing the progression to diabetes.
Patients and method: Multicenter cohort study on high risk individuals without diabetes (WHO-85 criteria) in primary health care. Subjects underwent an oral glucose tolerance test measuring fasting plasma glucose (FPG) and plasma glucose at 2 hours (2hPG). Mean follow-up was 37.2 months (4.3-69.7). Phenotypic features, diagnostic variation, diabetes incidence and predictive factors (multivariate analysis and Cox proportional hazards model) were assessed.
Results: We included 243 individuals (148 females), aged 59.7 (10) years, with FPG < 7.8 mM and 2hPG < 11.1 mM. 137 IGT subjects (56.4%) and 106 (43.6%) normal glucose tolerance (NGT) subjects with a similar risk factor impact were evaluated. After the study was closed, 63 (25.9%) subjects developed diabetes: 43 (31.4%) with baseline IGT and 20 (18.9%) with NGT. Overall diabetes incidence increased over time but not proportionally. Mean annual incidence was 9.2% and it dropped to 4.6% when FPG was used as the unique diagnosis criterion (ADA-97). Male under 65 years with both overweight and IGT reported the highest incidence. HbA1c, FPG and 2hPG were independent predictors. Increased HDL cholesterol showed a protective effect on diabetes incidence. The IGT diagnosis interval was a much better predictor [OR = 2.06 (1.56-2.72)] of diabetes development than the impaired FPG diagnosis interval [OR = 1.37 (0.93-2.04)].
Conclusions: FPG predicted but undervalued diabetes incidence in high risk population. The IGT (2hPG) diagnosis interval predicted diabetes development better than the impaired fasting plasma glucose diagnosis interval. Increased diabetes incidence in high risk Spanish population, particularly with regard to IGT, means that primary preventive resources should be increased.