Statistics from Altmetric.com
We read with great interest the paper by Cheung et al recently published in your journal.1 The extensive worldwide use of proton pump inhibitors (PPIs) makes the debate about a related risk of gastric cancer a very popular and current topic.2 Potentially, the PPI-induced reduction of acid content in the stomach may contribute to gastric cancer pathogenesis, possibly by increasing gastrin secretion, with a resultant constant trophic stimulus on the gastric mucosa, similar to Helicobacter pylori-related chronic atrophic gastritis.3 These conditions may both feasibly share a similar alteration of the gastric microbiota.4
However, evidence on this topic is not definitive: two meta-analyses of randomised controlled trials found no correlation between gastric cancer and long-term PPI use,5 6 whereas a meta-analysis of observational studies found a 40% increased risk associated with PPI use,7 similar to that of a recent nationwide population-based cohort study in Sweden (over threefold increase).8 The main reason for these contrasting results is generally considered the patient selection bias in observational studies, as patients taking long-term PPIs are usually older and have more comorbidities than control groups and could present independent risk factors for gastric cancer, not necessarily related to PPI use. Cheung et al performed a population-based study and elegantly attempted to reduce these confounding factors, excluding, among others, subjects with a diagnosis of gastric cancer after less than 6 months of PPI use or within 1 year of H. pylori eradication. They concluded that long-term use of PPIs was still associated with an increased gastric cancer risk despite H. pylori eradication therapy. However, this study still contains many confounding factors. First, the target and control populations had many baseline differences. Although the statistical significance of these differences is not stated, PPI users were about 10 years older than the control population (64.1 vs 54.3 years) at the time of eradication therapy. These data suggest a longer course of H. pylori infection and a likely deeper resultant alteration in both the gut microbiota and the gastric mucosal immunology. Furthermore, no data is reported regarding the prevalence of atrophic gastritis, intestinal metaplasia and the gastric microbiota at baseline, all conditions which may be only partially reversible following eradication in patients with long-standing H. pylori infection. Consequently, these patients may represent a particularly high-risk group when exposed to potent PPI-induced acid suppression.
Second, PPI users were more frequently affected by cardiovascular and metabolic diseases, which perhaps may suggest a higher exposure to other common risk factors, for example, the use of salty and processed meat or a higher body weight, factors which are usually not reported in electronic databases.
Finally, H. pylori eradication was not evaluated sufficiently. In fact, the authors defined the eradication success rate as a second course of eradication therapy prescription and not by urea breath testing. Such an approximation may have allowed the inclusion of a small, but still significant, group of subjects with ongoing H. pylori infection within the PPI users. Thus, although the resultant 2.4-fold increase in gastric cancer risk may be clinically relevant in similar high-risk populations, it may have a more limited impact in terms of absolute risk when applied to low-risk populations (ie, younger patients with a shorter course of H. pylori infection). In conclusion, the results from this high-quality paper recommend the consideration of the potential harm of long-term PPI use, leading to a sensible cessation of PPIs when possible. However, in patients in whom long-term PPIs are indicated, larger prospective observational studies are required to evaluate the real PPI-associated risks and to identify potential concomitant risk factors that could be treated to reduce the risk of gastric cancer.
Contributors LL wrote the commentary. FS and AG provided intellectual contribution and correction of the manuscript.
Competing interests None declared.
Provenance and peer review Not commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.