BackgroundAmong people with type 2 diabetes the relationship between central obesity and cardiovascular mortality has not been definitely assessed. Moreover, NT-proBNP is negatively associated with central obesity, but no study has examined their combined effect on survival. We have examined these issues in a well-characterized population-based cohort.Methods and FindingsSurvival data of 2272 diabetic people recruited in 2000 who had no other chronic disease have been updated to 31 December 2006. NT-proBNP was measured in a subgroup of 1690 patients. Cox proportional hazards modeling was employed to estimate the independent associations between cardiovascular and all-cause mortality and waist circumference. Mean age was 67.9 years, 49.3% were men. Both age and NT-proBNP were negatively correlated with waist circumference (r = −0.11, p<0.001 and r = −0.07, p = 0.002). Out of 2272 subjects, 520 deaths (221 for CV mortality) occurred during a median follow-up of 5.4 years. Central obesity was not associated with CV mortality (hazard ratio, HR, adjusted for age, sex, diabetes duration, 1.14, 95% CI 0.86–1.52). NTproBNP was a negative confounder and age a strong modifier of this relationship (p for interaction<0.001): age<70 years, fully adjusted model HR = 3.52 (1.17–10.57) and age ≥70 years, HR = 0.80 (0.46–1.40). Respective HRs for all-cause mortality were 1.86 (1.03–3.32) and 0.73 (0.51–1.04).ConclusionsIn diabetic people aged 70 years and lower, central obesity was independently associated with increased cardiovascular mortality, independently of the negative effect of NT-proBNP. In contrast, no effect on 6-years survival was evident in diabetic people who have yet survived up to 70 years.