AIMS: To determine the relation among the cytotoxin associated gene (cagA) and vacuolating cytotoxin gene (vacA) status of Helicobacter pylori isolates, the associated clinical diseases, and the severity and pattern of chronic gastritis. METHODS: Helicobacter pylori was cultured from gastric biopsies obtained from dyspeptic patients. DNA was extracted from the isolates and the cagA and vacA status determined by the polymerase chain reaction (PCR). The prevalence of the different cagA and vacA genotypes in three clinical groups, duodenal ulcer, gastric ulcer, and non-ulcer dyspepsia was compared. The histological features in sections from two antral and two corpus biopsies were graded by one blinded observer. The grades were compared with age and sex matched groups with different cagA and vacA genotypes, and with duodenal ulcers, or non-ulcer dyspepsia. RESULTS: Isolates from 161 patients were included. One hundred and nine (68%) harboured a cagA+ strain and 143 (89%) harboured a vacA s1 strain. The prevalence of cagA+ strains in duodenal ulcer patients (94%) was highly significantly greater than in those with non-ulcer dyspepsia (56%). However, of the patients infected with a cagA+ strain, almost equal numbers had non-ulcer dyspepsia or peptic ulceration. Chronic inflammation, polymorph activity, surface epithelial degeneration, atrophy, and intestinal metaplasia were all significantly more severe in the cagA+ than in the cagA- group, whereas only corpus epithelial degeneration was significantly more severe in the vacA s1 group compared with the vacA s2 group. Patients infected with cagA+ strains were almost four times more likely to have antral intestinal metaplasia than cagA- patients. An antral predominant gastritis was present in duodenal ulcer patients compared with matched non-ulcer dyspepsia patients, but this was not attributable to cagA or vacA status. CONCLUSIONS: Helicobacter pylori strains showing cagA positively and the vacA s1 genotype are associated with more severe gastritis but these virulence factors do not appear to determine the overall pattern. The pattern is closely linked to clinical disease. Therefore, it is likely that the nature of the disease complicating chronic infection is determined by host and environmental factors, while bacterial factors determine the magnitude of the risk of developing such disease.