LAUSR

We recently discovered that we used a suboptimal method for a part of our statistical analyses in our paper published recently in Gut. This has affected the duration analyses (time since eradication of Helicobacter pylori) and associated risk of gastric cancer. Previously, we only assessed the risk of gastric cancer in those eradicated during the selected followup periods, and disregarded the other individuals who ideally should also have contributed persontime. Therefore, persontime was underestimated in the first followup period, leading to a (biased) overestimation of the number of cases relative to the persontime, and therefore showed too highrisk estimates of gastric cancer. In the latter time intervals, too much followup time was included, that is, years that the individual could not develop cancer because of our inclusion criteria for these specific analyses. We have now rerun the analysis using the same cohort. As seen in tables 1 and 2, the results did not change dramatically compared with the results presented before, but the estimates are less extreme compared with the original method. Previously, we also reported an unexpected lower risk of gastric cancer in individuals who were eradicated at least 5 years earlier compared with the background population, which was not confirmed in the corrected analyses. Thus, as presented in table 1, the risk of gastric cancer does still decrease over time since eradication, but does not become lower than the overall risk of gastric cancer in Sweden, standardised for age, sex and calendar period.