Nonsteroidal anti-inflammatory drugs and risk of esophageal and gastric cancer.
Cancer Epidemiology, Biomarkers and Prevention. 2005;14(2):444-450
ABSTRACT
Methods: We tested these associations in a prospective, nested case-control study based on the General Practitioners Research Database, including over 2 million persons in the United Kingdom between 1994 and 2001. In multivariate analyses we calculated odds ratios (ORs) with 95% confidence intervals (CI). Data were stratified by history of upper gastrointestinal (UGI) disorders and recalculated using two years lag time on data, i.e. excluding all information two years before index date.
Results: Among 4,340,207 person-years of follow-up, we identified 909 patients with esophageal cancer and 1,023 patients with gastric cancer. We randomly selected 10,000 control subjects. Overall analysis suggested that long-term users of non-aspirin NSAIDs were at reduced risks of esophageal (OR 0.82, CI 0.57-1.18) and gastric cancer (OR 0.65, CI 0.44-0.94), while long-term aspirin users might be at decreased risk of esophageal cancer (OR 0.76, CI 0.53-1.08), but not of gastric cancer (OR 1.09, CI 0.82-1.45). All estimates of reduced risk were weakened in the two years lag time analysis except the association between non-aspirin NSAIDs long-term users and gastric cancer. Potentially protective effects were suggestive of being more marked among subjects with a history of UGI disorders.
Conclusions: Non-aspirin NSAIDs long-term use was associated with a reduced risk of gastric cancer, while no other studied associations could be firmly established. Our results suggest that UGI disorders could distort the associations, even though we could not show this with statistical significance. If such bias was to be true, the previously reported inverse associations might, at least partly, be explained by lack of appropriate adjustment for such disorders.
Introduction: Nonsteroidal anti-inflammatory drugs (NSAIDs) might reduce risks of esophageal and gastric cancer, but previous studies are limited and provide somewhat conflicting results.