The educational gradient in premature cardiovascular mortality: Examining mediation by risk factors in cohorts born in the 1930s, 1940s and 1950s.

Aims Educational inequality in cardiovascular disease and in modifiable risk factors changes over time and between birth cohorts. We aimed to assess how cardiovascular disease risk factors mediate educational differences in premature cardiovascular disease mortality and how this varies over birth cohorts and sex. Methods We followed 360,008 40-45-year-olds born in the 1930s, 1940s or 1950s from Norwegian health examination surveys (1974-1997) for premature cardiovascular disease mortality. Cox proportional hazard and Aalen's additive survival analyses provided hazard ratios and rate differences of excess deaths in participants with basic versus tertiary education. Results Relative educational differences in premature cardiovascular disease mortality were stable, whereas absolute differences narrowed from the 1930s to the 1950s cohorts; rate differences per 100 000 person years declined from 170 (95% confidence interval 117, 224) to 49 (36, 61) in men and from 60 (34, 85) to 23 (16, 29) in women. Cardiovascular disease risk factors attenuated rate differences by 69% in both cohorts in men, and in women by 102% in 1930s and 61% in 1950s cohorts. Smoking had the single strongest influence on the educational differences for men in all three cohorts, and for women in the two most recent cohorts. Conclusion Smoking appeared to be the driving force behind educational differences in premature cardiovascular disease mortality in the 1930s to 1950s birth cohorts for men and in the two recent birth cohorts for women. This suggests that strategies for smoking prevention and cessation might have the strongest impact for reducing educational inequality in premature cardiovascular disease mortality.