Discontinuation and poor adherence to therapy are major problems during long-term treatment, particularly with cholesterol lowering drugs. Several studies have indicated that the cholesterol lowering effect of statins differs according to apolipoprotein (apo)E genotypes. Low-density lipoprotein-cholesterol lowering capacity appears to be smaller in subjects with the epsilon(4) allele. To assess whether the use of statins in daily practice differs according to apoE genotypes, we used data from the Rotterdam Study, a population-based prospective cohort study in the Netherlands, which started in 1990 and included 7983 subjects aged 55 years or more. During follow-up, there were 798 subjects who started to use statins. We used a Cox proportional hazard model to determine the rate of discontinuation in the first 3 years of statin use. Subjects on statin therapy with epsilon(2)epsilon(2) and epsilon(4)epsilon(4) genotypes showed a trend towards higher dosages than subjects with the other genotypes. Relative to subjects with the epsilon(2)epsilon(3) genotype, those with the epsilon(4)epsilon(4) genotype had a risk of 2.28 [95% confidence interval (CI) 1.02-5.12] to discontinue statins within 3 years. In women, this relative risk was 1.70 (CI 0.53-5.42) versus 3.18 (CI 1.01-10.03) in men. The apoE genotype is associated with discontinuation of statins. This suggests that subjects who are genetically prone to develop hypercholesterolemia show the highest risk of discontinuation of treatment.