Background: Various indicators of childhood socioeconomic position have been related to cardiovascular disease (CVD) risk in adulthood. We investigated the impact of shared family factors on the educational gradient in midlife CVD risk factors by assessing within sibling similarities in the gradient using a discordant sibling design.

Methods: Norwegian health survey data (1980–2003) was linked to educational and generational data. Participants with a full sibling in the health surveys (228,346 individuals in 98,046 sibships) were included. Associations between attained educational level (7–9 years, 10–11 years, 12 years, 13–16 years, or >16 years) and CVD risk factor levels in the study population was compared with the corresponding associations within siblings.

Results: Educational gradients in risk factors were attenuated when factors shared by siblings was taken into account: A one category lower educational level was associated with 0.7 (95% confidence interval 0.6 to 0.8) mm Hg higher systolic blood pressure (27% attenuation), 0.4 (0.4 to 0.5) mmHg higher diastolic blood pressure (30%), 1.0 (1.0 to 1.1) more beats per minute higher heart rate (21%), 0.07 (0.06 to 0.07) mmol/l higher serum total cholesterol (32%), 0.2 (0.2 to 0.2) higher smoking level (5 categories) (30%), 0.15 (0.13 to 0.17) kg/m2 higher BMI (43%), and 0.2 (0.2 to 0.2) cm lower height (52%). Attenuation increased with shorter age-difference between siblings.

Conclusion: About one third of the educational gradients in modifiable CVD risk factors may be explained by factors that siblings share. This implies that childhood environment is important for the prevention of CVD.