We aimed to examine the effect of a large population-based multifactorial screening and lifestyle intervention programme on 10-year incidence of diabetes. In a randomised trial of the general Danish population initiated in 1999-2001 59,616 men and women aged 30-60years were assigned to a five year screening and lifestyle counselling programme (n=11,629) or control group (n=47,987) and followed for ten years in nationwide registers. Intention to treat was applied and risk of diabetes was modeled by Cox regression and expressed as hazard ratios (HRs). We found that 1692 individuals had diabetes at baseline. Among 57,924 individuals without diabetes at baseline, 1267 emigrated, 2593 died and 3369 (Intervention group=684, Control group=2685) developed diabetes. We saw no significant difference in diabetes incidence between the groups after 10-year follow-up (Grey's test: p=0.22). In the first year of follow-up, incidence of diabetes was significantly higher in the intervention group than the control group (HR=1.68, 95%CI 1.29 to 2.29). We observed no difference in incidence of diabetes between the groups in the follow-up intervals from 1 to 6years or after 6-10years (HR=0.94, 0.83 to 1.06; HR=1.03, 0.91 to 1.17). Inviting the general population to participate in a repeated screening and lifestyle counselling programme over five years did not result in lower incidence of diabetes after 10years of follow-up. As expected, significantly more individuals were diagnosed with diabetes in the intervention group during the first year, but this was not followed by a decrease in the following years.

TRIALS REGISTRATION: Clinical trials NCT00289237.