To examine a long-term effect of community-based education program for hypercholesterolemia and an effect modification by apolipoprotein E polymorphism, we conducted a 1-year randomized clinical trial with 8 year-follow-up. One hundred four persons aged 40-64 years who had serum total cholesterol levels between 6.21 and 7.73 mmol/l (240 and 299 mg/dl) in 1988-89 cardiovascular risk surveys were enrolled in the trial. The intervention group (n=51, 82% for women) attended eight education classes in 1 year, while the control group (n=53, 85% for women) attended only two classes. Both groups were invited to the subsequent annual surveys. The mean serum cholesterol was 0.24-0.26 mmol/l less in the intervention than in the control group at both 6 month and 1 year (P=0.03, each) while the proportion of subjects using hypolipidemic agents was 0 and 6% in both groups, respectively. During 8-year follow-up, the probability of using hypolipidemic agents and/or total cholesterol > or =7.76 mmol/l was 51% in the education group and 69% in the control group; the risk ratio in the intervention vs control groups was 0.62 (95% CI: 0.36-1.06). When stratified by the apolipoprotein E polymorphism examined for 78% of the subjects, the risk ratio was 0.61 (0.31-1.18) among subjects without e4 allele (n=59) and 0.55 (0.14-2.14) among those with e4 allele (n=22). The intervention group had reduced intake of egg, fish egg, butter, mayonnaise and fatty meat compared to the control group at 6-month, 1- and 8-year follow-up. In conclusion, our community-based program was effective in reducing serum total cholesterol levels non-pharmacologically during the first year, and also reduced the likelihood of progressive worsening of hypercholesterolemia during the subsequent 8 years, regardless of the apolipoprotein E polymorphism.