Population-based incidence rates and risk factors for type 2 diabetes in white individuals: the Bruneck study

Abstract

Incidence rates and risk factors for type 2 diabetes in low-risk populations are not well documented. We investigated these in white individuals who were aged 40-79 years and from the population of Bruneck, Italy. Of an age- and sex-stratified random sample of 1,000 individuals who were identified in 1990, 919 underwent an oral glucose tolerance test (OGTT) and an assessment of physiological risk factors for diabetes, including insulin resistance (homeostasis model assessment, HOMA-IR), and postchallenge insulin response (Sluiter's Index). Diabetes at baseline by fasting or 2-h OGTT plasma glucose (World Health Organization criteria, n = 82) was excluded, leaving 837 individuals who were followed for 10 years. Incident cases of diabetes were ascertained by confirmed diabetes treatment or a fasting glucose >or=7.0 mmol/l. At follow-up, 64 individuals had developed diabetes, corresponding to a population-standardized incidence rate of 7.6 per 1,000 person-years. Sex- and age-adjusted incidence rates were elevated 11-fold in individuals with impaired fasting glucose at baseline, 4-fold in those with impaired glucose tolerance, 3-fold in overweight individuals, 10-fold in obese individuals, and approximately 2-fold in individuals with dyslipidemia or hypertension. Incidence rates increased with increasing HOMA-IR and decreasing Sluiter's Index. As compared with normal insulin sensitivity and normal insulin response, individuals with low insulin sensitivity and low insulin response had a sevenfold higher risk of diabetes. Baseline impaired fasting glucose, BMI, HOMA-IR, and Sluiter's Index were the only independent predictors of incident diabetes in multivariate analyses. We conclude that approximately 1% of European white individuals aged 40-79 years develop type 2 diabetes annually and that "subdiabetic" hyperglycemia, obesity, insulin resistance, and impaired insulin response to glucose are independent predictors of diabetes.