Objective It has been questioned if the excess cardiovascular disease (CVD) mortality by lower educational level can be fully explained by conventional modifiable CVD risk factors. Our objective was to examine whether repeated measures over time of risk factors (smoking, physical inactivity, blood pressure, total cholesterol and body mass index) explain more of the socioeconomic gradient in CVD mortality than if they are measured only once.

Methods A cohort of 34 884 men and women attended all three screenings (1974–1978, 1977–1983 and 1985–1988) in the Norwegian Counties Study and were followed for CVD mortality through 2009 by linkage to the Norwegian Cause of Death Registry.

Results Age-adjusted and sex-adjusted HR of CVD mortality was 2.32 (95% CI 1.93 to 2.80) for basic relative to tertiary educated individuals. The HR was attenuated by 48% (HR 1.54 (1.28 to 1.87)) when adjusted for CVD risk factors measured at baseline and by 56% (HR 1.45 (1.20 to 1.75)) when two repeated measurements ascertained 5 years apart were added to the model. Similarly, absolute risk difference in CVD mortality by education was attenuated by 62% when adjusted for baseline and by 72% when adjusted for repeated measurements of risk factors.

Conclusions In this cohort, repeated measurements of risk factors seemed to explain more of the educational gradient in CVD mortality. This suggests that a substantial part of the excess CVD mortality among those with lower education might be explained by conventional risk factors.