Helicobacter pylori-associated gastritis has been put forward as a distinct risk factor for gastric cancer. Furthermore, among H. pylori-positive individuals a correlation between a high serum level of H. pylori antibodies and the risk of gastric cancer has been found in two different studies. Other studies have challenged this hypothesis. We therefore studied the presence and level of H. pylori serum antibodies, using an enzyme-linked immunosorbent assay technique in 116 gastric cancer patients (65 men; mean age, 67 years; range, 23-92 years) and 116 controls matched for age and sex. Patients and controls were selected on referral for gastroscopy. The prevalence of infection in gastric cancer patients was 77% (89 of 116) and in controls 79% (92 of 116). This difference is not statistically significant, nor is the prevalence of infection in cases and controls of different age cohorts significantly different. High levels of serum antibodies were found in 46% (53 of 116) of gastric cancer patients and 40% (46 of 116) of controls. Comparison of the prevalence of high serum levels of antibodies for the total population and for the different age cohorts did not show significant differences either. We conclude that the comparison of actual H. pylori infection in a cross-sectional study of gastric cancer patients and controls does not enable relative risk calculation in the study of the role of H. pylori infection in gastric carcinogenesis. Prospective studies showing diminishment of the risk for gastric cancer after eradication of H. pylori are required.