Our study investigated the relationship between gastric cancer development and activity of Helicobacter pylori-associated chronic gastritis or the resulting chronic atrophic gastritis (CAG). A cohort of 4,655 healthy asymptomatic subjects, in whom serum pepsinogen (PG) and H. pylori antibody titer had been measured to assess the activity and stage of H. pylori-associated chronic gastritis, was followed for up to 16 years, and cancer development was investigated. In subjects with a serologically diagnosed healthy stomach (H. pylori-negative/CAG-negative), cancer incidence rate was low, at 16/100,000 person-years. With the establishment of H. pylori infection and progression of chronic gastritis, significant stepwise cancer risk elevations were seen from CAG-free subjects (H. pylori-positive/CAG-negative) [hazard ratio (HR) = 8.9, 95% confidence interval (CI) = 2.7-54.7] to subjects with CAG (H. pylori-positive/CAG-positive) (HR = 17.7, 95% CI = 5.4-108.6) and finally to subjects with metaplastic gastritis (H. pylori-negative/CAG-positive) (HR = 69.7, 95% CI = 13.6-502.9). In H. pylori-infected CAG-free subjects, significantly elevated cancer risk was observed in the subgroup with active inflammation-based high PG II level or potent immune response-based high H. pylori antibody titer; the former was associated with a particularly high risk of diffuse-type cancer, and both subgroups showed high cancer incidence rates of around 250/100,000 person-years, comparable to that in subjects with CAG. No such risk elevation was observed in H. pylori-infected subjects with CAG. These results clearly indicate that gastric cancer develops mainly from the gastritis-atrophy-metaplasia-cancer sequence and partly from active inflammation-based direct carcinogenesis, and that serum levels of PG and H. pylori antibody titer provide indices of cancer development in H. pylori-infected subjects.