Aims: To explore the confounding effects of early family factors shared by siblings and cardiovascular risk factors in midlife on the educational differences in mortality from cardiovascular disease (CVD). Methods: Data from national and regional health surveys in Norway (1974–2003) were linked with data from the Norwegian Family Based Life Course Study, the National Educational Registry and the Cause of Death Registry. The study population consisted of participants with at least one full sibling among the health survey participants (n=271,310). Data were available on CVD risk factors, including weight, height, blood pressure, total cholesterol and smoking. Results: The hazards ratio (HR) of CVD mortality was 3.44 (95% confidence interval (CI) 2.98–3.96) in the lowest educational group relative to the highest. The HRs were little altered in the within-sibship analyses. Adjusted for risk factors, the HR for CVD mortality in the cohort analyses was 2.05 (CI 1.77–2.37) in the lowest educational group relative to the highest. The respective HR in the within-sibship analyses was 2.46 (CI 1.48–2.24). Conclusions: Using a sibling design, we did not find that the association between education and CVD mortality was confounded by early life factors shared by siblings, but it was explained to a large extent by CVD risk factors. These results suggest that reducing levels of CVD risk factors could have the greatest effect on mortality in less well-educated people.