Incidence and predictors of drug-treated diabetes in elderly French subjects. The PAQUID Epidemiological Survey.

Aims: To estimate the incidence and predictors of drug-treated diabetes in elderly subjects.

Methods: The PAQUID epidemiological survey, a population-based study, has followed up 3,777 subjects older than 65 years since 1988. At each visit (baseline, 1, 3, 5 and 8 years), treatment regimen was used to identify new drug-treated diabetic subjects. Potential predictors of drug-treated diabetes were collected during the baseline visit (body mass index (BMI), educational level, cigarette smoking and wine consumption, physical activity, depressive symptomatology, subjective health, treatment, and hypertension) and analysed by using a multivariate backward stepwise regression Cox model with delayed entry.

Results: The prevalence rate of drug-treated diabetes was 7.5% at baseline and 7.1% after 8 years' follow-up. The incidence rate of drug-treated diabetes was 3.8/1,000 person-years, 5.9/1,000 person-years in men and 2.4/1,000 person-years in women, with no significant variation according to age group. Male sex (relative risk (RR) 2.4, 95% confidence interval (CI) 1.4-4.0, P < 0.001, attributable risk (AR) 0.36), elevated BMI (for one point increase, RR 1.1, 95% CI 1.1-1.1, P < 0.001, > or = 25 vs. < 25, RR 2.1, 95% CI 1.2-3.5, AR 0.33), thiazide diuretics used alone (RR 5.9, 95% CI 1.8-19.6, P = 0.02), and poorer subjective health ('the same' vs. 'better' RR 1.8, 95% CI 1.0-3.1, P = 0.04; 'worse' vs. 'better' RR 2.3, 95% CI 0.9-5.7, P = 0.06) were independent predictors of drug-treated diabetes in this population.

Conclusions: In older French individuals, men seem to be particularly exposed to drug-treated diabetes although being overweight was found to be a strong predictor as in younger populations.