Infection with Helicobacter pylori has been associated with an increased risk of gastric cancer in low-risk populations. However, our previous results (P. M. Webb et al., Int. J. Cancer, 67: 603-607, 1996) from an ongoing prospective study in Shanghai, China, a relatively high-risk population, failed to show an association between H. pylori infection and the subsequent risk of gastric cancer. That previous study had a relatively short time period of follow-up and the enzyme-linked immunosorbent assay (ELISA) used was based on strains found in Southern England and without validation among the Chinese. Either one of these two factors could have had an impact on the validity of those earlier observations. An ELISA developed and validated among Shanghai residents was used in the present study to reexamine specific antibodies to H. pylori in 188 gastric cancer patients and 548 control subjects. All of the cases of gastric cancer were identified during the first 12 years of follow-up of a cohort of 18,244 men, ages 45-64 years in Shanghai, from whom blood samples were collected at enrollment during 1986-1989. For each cancer case, three cancer-free control subjects were randomly selected from the cohort and matched to the index cases by age (within 2 years), month and year of sample collection, and neighborhood of residence. The Shanghai-based ELISA detected a higher prevalence of serum antibodies to H. pylori than the English-based assay in both gastric cancer cases (86 versus 53%) and control subjects (85 versus 56%). Virtually all of the subjects (98%) who were H. pylori-seropositive by the English-based assay tested positive by the Shanghai-based assay. On the other hand, 73% of gastric cancer cases and 68% of control subjects who were seronegative according to the English-based assay tested positive by the Shanghai-based assay. Using this alternative assay, combined with increased follow-up, our latest data contradict our earlier findings and show a statistically significant association between H. pylori seropositivity and gastric cancer risk (odds ratio, 1.84; 95% confidence interval, 1.08-3.11). We noted an increasing rate of seropositivity among cases as the time interval between cohort enrollment and cancer diagnosis increased. Among subjects followed for 5 or more years after enrollment, the odds ratio for gastric cancer related to H. pylori seropositivity was 3.74 (95% confidence interval, 1.51-9.30).