Impaired fasting glucose, ancestry and waist-to-height ratio: main predictors of incident diagnosed diabetes in the Canary Islands.

AIMS To estimate the incidence rate and risk factors for diabetes in the Canary Islands. METHODS A total of 5521 adults without diabetes were followed for a median of 3.5 years. Incident cases of diabetes were self-declared and validated in medical records. The following factors were assessed by Cox regression to estimate the hazard ratios for diabetes: impaired fasting glucose (5.6 mmol/l ≤ fasting glucose ≤ 6.9 mmol/l), BMI, waist-to-height ratio (≥ 0.55), insulin resistance (defined as triglycerides/HDL cholesterol ≥ 3), familial antecedents of diabetes, Canarian ancestry, smoking, alcohol intake, sedentary lifestyle, Mediterranean diet, social class and the metabolic syndrome. RESULTS The incidence rate was 7.5/10(3) person-years (95% CI 6.4-8.8). The greatest risks were obtained for impaired fasting glucose (hazard ratio 2.6; 95% CI 1.8-3.8), Canarian ancestry (hazard ratio 1.9; 95% CI 1.0-3.4), waist-to-height ratio (hazard ratio 1.7; 95% CI 1.1-2.5), insulin resistance (hazard ratio 1.5; 95% CI 1.0-2.2) and paternal history of diabetes (hazard ratio 1.5; 95% CI 1.0-2.3). The metabolic syndrome (hazard ratio 1.9; 95% CI 1.3-2.8) and BMI ≥ 30 kg/m(2) (hazard ratio 1.7; 95% CI 1.0-2.7) were significant only when their effects were not adjusted for impaired fasting glucose and waist-to-height ratio, respectively. CONCLUSIONS The incidence of diabetes in the Canary Islands is 1.5-fold higher than that in continental Spain and 1.7-fold higher than in the UK. The main predictors of diabetes were impaired fasting glucose, Canarian ancestry, waist-to-height ratio and insulin resistance. The metabolic syndrome predicted diabetes only when its effect was not adjusted for impaired fasting glucose. In individuals with Canarian ancestry, genetic susceptibility studies may be advisable. In order to propose preventive strategies, impaired fasting glucose, waist-to-height ratio and triglyceride/HDL cholesterol should be used to identify subjects with an increased risk of developing diabetes.