The hypothesis that apolipoprotein E (apo E)-isoform-related differences in plasma and LDL cholesterol concentrations are due to differential responses to dietary lipids was explored in 110 North Karelian subjects who had previously participated in dietary intervention studies. This was accomplished by collecting fresh blood samples for apo E phenotyping and by re-analysis of the original plasma lipid data according to apo E phenotypes. During high fat, high cholesterol baseline (p = 0.003) and switchback diets (p = 0.002), plasma cholesterol correlated inversely with the sum of subscript numbers (e.g., apo E3/4 = 7). Thus, subjects with the apo E4/4 phenotype had the highest (7.63 +/- 1.32 mmol/l), and subjects with the apo E3/2 phenotype had the lowest baseline levels of plasma cholesterol (5.85 +/- 1.48 mmol/l). This association became weaker during a low fat, low cholesterol diet intervention (p = 0.069). Greater reductions in plasma cholesterol occurred in subjects homozygous for the apo epsilon 4 allele (-1.84 mmol/l) as compared to subjects with other genotypes (-1.13 mmol/l) (p = 0.0097). Moreover, these subjects responded to the switchback diet by greater increases in plasma cholesterol (1.52 mmol/l) than others (0.92 mmol/l, p = 0.0141). The results suggest that the effect of apo epsilon genotype on plasma cholesterol is modulated by dietary fat and cholesterol intake.