Aspirin Use and Gastric Adenocarcinoma-Letter.

Abstract

One retrospective cohort study conducted by Kwon and colleagues disclosed that regular use of aspirin (≧two times of use every 1 week) correlated with a decreased risk of gastric cancer as compared with nonregular use of aspirin in women (adjusted HR = 0.52 and 95% confidence interval = 0.37–0.73; ref. 1). Some points not shown in the original article are discussed.First, based on Table 2 in Kwon and colleagues study, we estimated that the incidence rates of gastric cancer were about 4 every 100,000 person-years (or 51/1,258,158 person-years) in women with regular use of aspirin and 6 every 100,000 person-years (or 125/2,097,414 person-years) in women with nonregular use of aspirin. These results indicated the relative risk of gastric cancer was 0.67 for regular use versus nonregular use of aspirin, but the absolute risk reduction was about 0.00002. If we only see the relative risk, the protective effect of regular use of aspirin seems to be great. However, the relative risk might mislead. If we focus on the absolute risk reduction, regular use of aspirin would lead to two fewer cases of gastric cancer every 100,000 person-years in women. According to the traditional formula, the number needed for regular use of aspirin to prevent one case of gastric cancer was 50,000 in women. Only one woman would get a protective effect against gastric cancer among 50,000 women who received regular use of aspirin. From a view of public health, there is no advantage for regular use of aspirin on preventing gastric cancer because upper gastrointestinal bleeding cannot be negligible. We suggest that the absolute risk reduction would provide the real-world concept when the relative risk shows a good number. Second, the definition of the index date was not mentioned in Kwon and colleagues study. Thus, the readers did not know whether the immortal time bias was excluded from the study. In the absence of this information, the readers would believe that the reduced risk of gastric cancer is distorted by the immortal time bias (2). Third, Kwon and colleagues commented that the mechanism underlying the relation between aspirin use and gastric cancer might be mediated by the effect of aspirin on Helicobacter pylori (1). Aspirin has an anti-inflammatory effect but does not have an antibiotic effect. The first-line eradication treatment of Helicobacter pylori infection includes one proton pump inhibitor plus two types of antibiotics, but aspirin not included. It is difficult to interpret that only aspirin use can reduce the risk of Helicobacter pylori infection. Four, the USPSTF in 2016 has ever reported that only colorectal cancer may be prevented by use of low-dose aspirin in the specific populations (3), but these statements have been replaced now (4). On the basis of the above discussion, the current evidence is insufficient to evaluate the balance of benefit and harm of regular use of aspirin on prevention of gastric cancer. We agree with Kwon and colleagues’ conclusion that further studies are needed to confirm their results.See the Response, p. 555The authors disclose no conflicts of interest.The authors received no funding from an external source.