An immunoblot method has been evaluated to diagnose Helicobacter pylori infection serologically by comparing 69 serum specimens from patients with a positive Gram stain and/or culture result and a positive urease test on biopsy material, as well as 51 serum specimens from patients with at least 4 negative urease tests, and negative microscopy and culture results. Sensitivity and specificity was found to be 100%. Recognition of the cross-reacting flagellin (66 kDa), flagellar sheath protein (51 kDa), and a 14 kDa protein are not a criterion for a current H. pylori infection. On the other hand, any combination of at least two of the 180, 120, 90, 75, 67, 29.5 and 19 kDa bands were diagnostic of infection. Three H. pylori strains, which were compared with both gel electrophoretic analyses and immunoblot reactivity, exhibited in part strong qualitative and quantitative differences that particularly affect the 120 kDa pathogenic factor, the large urease subunit and other proteins especially in the molecular mass range from 50 to 67 kDa. IgG immunoblot patterns showed that the choice of H. pylori strain, as well as a reproducible and standardizable antigen preparation, is of great importance for the reliability of serodiagnostic tests. The immunoblot method was found to be a valuable tool for the semi-quantitative confirmation of results achieved with other serological methods as well as optimization and quality control of the antigens used for serodiagnostic purposes.