Abstract Background The increased risk of cardiovascular disease (CVD) in diabetic compared to non-diabetic subjects seems to decrease with age. Whether this age-related reduction applies to CVD risk factors, and whether it is limited to established diabetes mellitus (DM) or also applies to pre-diabetic conditions are not well known. Methods Using a cross-sectional design we compared the strength of the correlation between glucometabolic disturbances (by grouping), CVD risk factor burden and self-rated health, in two age groups: middle-aged (57-69 years) and older (70-86 years) subjects, (63% men), participating in the Malmö Preventive Project Re-examination Study (n = 18,238). Simple (unadjusted) logistic regression analysis was applied to estimate between-group differences and trends. Interaction analysis was applied to estimate differences between age groups. Results CVD risk factor burden and the proportion of subjects reporting poor self-rated health increased with increasing glucometabolic disturbance for men and women in both age groups (p-trend < 0.0001 for all). The slope of the trend curve with increasing CVD risk factor burden was significantly steeper for older women than for older men (p-interaction = 0.002). The slope of the trend curve for poor self-rated health was significantly steeper for middle-aged than for older men (p-interaction = 0.005), while no difference was observed between the age groups among women (p-interaction = 0.97). Conclusions We found no reduction in risk factor accumulation with increasing glucometabolic disturbance between middle-aged and older subjects. Our results indicate life-long CVD risk factor clustering with increased glucometabolic disturbance, and suggest that previously observed age-related reduction in excess CVD risk for subjects with DM might be due to a survival bias. However, our observations indicate more pronounced risk factor clustering and worse self-rated health with increased glucometabolic disturbance in older women than in older men.