Medical interventions and gastric cancer risk : an observational approach

Abstract

Helicobacter pylori infection is undoubtedly the most important risk factor for gastric cancer. A causal relationship with gastric cancer, however, has not been clearly established and the carcinogenic mechanism is still ambiguous. There is a growing consensus for introduction of large-scale eradication programs, but it is yet unknown if elimination of H. pylori would reduce the incidence of gastric cancer. To address this question, randomized intervention trials using gastric cancer as end-point have been initiated, but conclusive results will not be available in the near future since these trials require very large samples and lengthy follow-up. Using observational data on medical interventions, the aim of this thesis was to investigate characteristics of the relationship between H. pylori infection and gastric cancer, and to assess the risk of gastric cancer following antibiotic and anti-inflammatory treatments.

In a previous cohort study of 39,000 hip replacement patients, identified in the Swedish Inpatient Register, a declining risk of gastric cancer was observed with increasing time after hip surgery. Here, we tested the hypothesis that the risk reduction was mediated through incidental eradication of H. pylori, caused by the prophylactic antibiotic treatment given at time of surgery. In a case-control study, 'nested' within the original cohort (174 cases, 462 controls), we found that both long-term antibiotic treatment before surgery (OR, 0.3 95% Cl, 0. 1-0.7) and prophylactic antibiotic treatment (OR, 0.7 95% Cl, 0.5- 1. 1) conferred a reduction in gastric cancer risk.

To further test the hypothesis that exposure to heavy antibiotic treatment reduces the risk of gastric cancer, we assessed cancer risk among 336,017 patients hospitalized for one of ten selected infectious diseases, identified in the Swedish Inpatient Register. The gastric cancer risk in the infectious disease cohort was similar to that expected in the background population, and the data did thus not further support our hypothesis.

As an alternative explanation, the apparent reduction in gastric cancer risk may have resulted from intense exposure to aspirin and other non-steroid anti-inflammatory drugs used by the hip replacement patients, who typically have intractable pain prior to their operations. Analysis of data from a population-based case-control study (567 cases, 1165 controls) showed a 30% (OR, 0.7 95% Cl, 0.6-1.0) reduced gastric cancer risk among users of aspirin compared to non-users, as well as an inverse trend with increased number of tablets used (p=0.02).

Although largely unclear, the carcinogenic action of H. pylori is believed to be mediated through induction and maintenance of chronic inflammation. We aimed to test whether a noninfectious chronic inflammation, i.e. reactive gastritis induced by duodeno-gastric bile reflux following cholecystectomy, also increases the risk of gastric cancer. To address this issue, we assessed the gastric cancer risk within a cohort of 251,672 cholecystecomized patients that were followed for up to 27 years after surgery. A modest excess risk of gastric cancer was observed, but the association was inconsistent over gender strata and disappeared over time.

Although not discarding the prospect of gastric cancer prevention by eradication of H. pylori, neither do these results unambiguously support the notion that antibiotic treatment reduces the risk of gastric cancer. Highlighting the role of inflammation - possibly H. pylori specific - a protective effect of nonsteroid anti-inflammatory drugs (aspirin) on gastric carcinogenesis was suggested.