Several protein and DNA-based methods have been previously described for the identification of apolipoprotein E isoforms or genotypes. However, all of them generate frequently false-positive results. The purpose of this study was to set up a new, simple, and effective method for the analysis of the apoE polymorphism. A total of 1,253 subjects previously examined for the apolipoprotein E polymorphism by restriction fragment length polymorphism were reanalyzed by our new method based on Taq DNA polymerase's inability to correctly initiate the replication in the presence of a mismatch at the 3' end of the primer. We conceived a combination of 4 specific primers in 3 different pairs sharing the same stringent polymerase chain reaction conditions to directly detect the presence/absence of polymerase chain reaction products, and thus reveal the 6 apolipoprotein E genotypes. We confirm our previous results in 1,171 subjects, whereas in 82 subjects out of 1,253 (about 6%), the results have been reinterpreted. The final analysis revealed a total of 12 homozygotic subjects for the e2 allele (1.0%), 874 homozygotes for the e3 allele (69.8 %), and 8 homozygotes for the e4 allele (0.6 %). The frequence of heterozygotes was 8.7% for the e2/e3 genotype (n=109), 1.4% for the e2/e4 genotype (n=17), and 0.6% for the e3/e4 genotype (n=8). Relative allele frequencies were e2=0.060, e3=0.834, and e4=0.106. We describe a new, simple, unequivocal, and nonexpensive method for the identification of the 6 apoE genotypes.