Risk Factors and Incidence of Gastric Cancer After Detection of Helicobacter pylori Infection

Based on an analysis of 371,813 H. pylori–exposed US veterans, as defined by either a prescription for eradication therapy, which comprised the vast majority (71%), administrative code (18%), or non-serologic testing confirming active infection (10.4%), Kumar and colleagues demonstrated that the cumulative risk of incident intestinal-type, non-cardia gastric adenocarcinoma (NCGA) was 0.37%, 0.50%, and 0.65% at 5-, 10-, and 20-years after H. pylori diagnosis, and slightly lower when cancers diagnosed within 6 or 12 months of the H. pylori diagnosis were excluded.

Among the small subset prescribed H. pylori eradication therapy and who had subsequent eradication confirmation testing (2.2% of the full cohort), there was a significantly lower risk of incident NCGA. In addition to increased NCGA risk associated with age, male sex, and smoking, there were also clear racial/ethnic differences in this cohort that warrant emphasis, with blacks, Asians, and Hispanics having a 2-, 2.5-, and 1.6-fold higher risk of incident NCGA compared with non-Hispanic whites.

There are some limitations to this large cohort study, which should be noted to ensure appropriate interpretation in context—the most relevant being potential biases, potential incomplete measurement for exposure and outcome, inability to control for important confounders, such as duration of infection, and generalizability to non-veteran populations. Indeed, because H. pylori testing is prompted by clinical symptoms or associated pathology (eg, peptic ulcer disease, unexplained iron-deficiency anemia), there is inherent risk of certain biases, including ascertainment bias given that prospective epidemiological studies conducted in both veteran and civilian populations confirm that a large percentage of individuals colonized with H. pylori are asymptomatic and thus testing would not be prompted.

There is also a real possibility that the true incidence of NCGA may be underestimated in light of the following: 1) possible competing risks for censorship prior to NCGA outcome, particularly because NCGA is asymptomatic until advanced stages; 2) the possibility that veterans were diagnosed outside of the VA system; and 3) the exclusion of NCGA cases that were not distinctly coded as intestinal-type on histology. How these and other limitations might affect the risk estimates for cumulative incidence of intestinal-type NCGA and associated determinants is unpredictable. Notwithstanding, the findings of this study importantly emphasize the unmet need to more clearly define high-risk populations in the US, both veteran and civilian, who might benefit from targeted gastric cancer prevention and early detection efforts.