Genetic diversity within the cag pathogenicity island (PAI) of Helicobacter pylori may have a modifying effect on the pathogenic potential of the infecting strain. The genetic structure of the cag PAI was examined in Japanese isolates. The composition and nucleotide sequences of the cag PAI were quite similar among strains; however, diversity between 2 cag genes (virB10 and cagA) was observed. The variety in the number of repetition of the 5-amino acid sequence R1 (EPIYA) in the 3' region of the cagA gene was identified. The frequencies of the genotypes that contained >4 R1 sequences were significantly higher in atrophic gastritis-causing strains than in duodenal ulcer-causing strains. One-third of strains with >4 R1 sequences were gastric cancer-causing strains. Although the cag PAI is conserved in H. pylori isolates in Japan, H. pylori infection with the cagA genotype with >4 R1 sequences may correlate with the pathogenesis of atrophic gastritis and gastric cancer.