Background & aims: Helicobacter pylori may disappear spontaneously with progressing precancerous changes and invalidate serologic studies of its association with gastric cancer. We reestimated the strength of the H. pylori-gastric cancer relationship, using both conventional immunoglobulin (Ig) G enzyme-linked immunosorbent assay (ELISA) and immunoblot (against cytotoxin-associated antigen A [CagA] antibodies that prevail longer after eradication) to detect past H. pylori exposure more relevant for time at cancer initiation.
Methods: In our population-based case-control study, the seroprevalence among 298 gastric adenocarcinoma cases was 72% (IgG ELISA) and 91% (immunoblot) vs. 55% and 56% among 244 controls frequency-matched for age and gender.
Results: Using IgG ELISA only, the adjusted OR for noncardia gastric cancer among H. pylori-positive subjects was 2.2 (95% confidence interval [CI], 1.4-3.6). When ELISA-/CagA+ subjects (odds ratio [OR], 68.0) were removed from the reference, the OR rose to 21.0 (95% CI, 8.3-53.4) and the previous effect modification by age disappeared. ELISA+/CagA- subjects had an OR of 5.0 (95% CI, 1.1-23.6). There were no associations with cardia cancer.
Conclusions: The weaker H. pylori-cancer relationships in studies based on IgG ELISA rather than CagA may be caused by misclassification of relevant exposure. A much stronger relationship emerges with more accurate exposure classification. In the general Swedish population, 71% of noncardia adenocarcinomas were attributable to H. pylori.