Helicobacter pylori infection and long term proton pump inhibitor therapy.

S hould Helicobacter pylori infection be eradicated in patients requiring maintenance proton pump inhibitor therapy for gastro-oesophageal reflux disease? This question has stimulated heated debate and contention over the past few years. The issue first came to prominence in 1996 when Kuipers et al published their study purporting to demonstrate that omeprazole accelerated the development of corpus atrophic gastritis in H pylori infected subjects. This rung alarm bells due to the fact that atrophic gastritis is a well recognised risk factor for gastric cancer in H pylori infected subjects. There was therefore concern that proton pump inhibitor therapy was modifying the inflammatory response to H pylori infection in such a way as to increase the risk of gastric cancer. For this reason, some experts have recommended that H pylori infection should be eradicated prior to long term proton pump inhibitor therapy. The original paper by Kuipers et al was widely criticised due to weaknesses in its design, and its claim that proton pump inhibitor therapy accelerated atrophy in H pylori infected subjects was not supported by the FDA Gastrointestinal Drugs Advisory Committee. In 1999, Lundell et al published a study claiming that proton pump inhibitor therapy did not accelerate the development of corpus atrophy in H pylori infected subjects. However, their conclusion was challenged because there was evidence of accelerated development of moderate and severe atrophy in the H pylori infected group on proton pump inhibitor therapy and the size of this effect was similar to that reported by Kuipers and colleagues. The paper by Kuipers and colleagues in the current issue of Gut confuses the issue further as they did not observe any progression of atrophy in their H pylori infected subjects [see page 12]. Indeed, there was not even a trend in favour of progression of atrophy that might have become significant in a larger study. Several other recent studies have also found no evidence of acceleration of corpus atrophy in H pylori infected subjects on proton pump inhibitor therapy. Consequently, there is little evidence in support of the original concern that proton pump inhibitor therapy accelerates corpus atrophy in H pylori infected subjects. One consistent finding however of all of these studies is that proton pump inhibitor therapy does change the pattern of H pylori induced gastritis, causing it to move from the antrum into the more proximal corpus mucosa of the stomach. In this way, it induces what is referred to as a corpus or body predominant gastritis. This may have significance with respect to the subsequent risk of H pylori associated gastric cancer. Uemura et al recently examined the association between the pattern of H pylori induced gastritis and the subsequent development of gastric cancer in patients not receiving proton pump inhibitor therapy. Their study indicated that the strongest risk factor for cancer was the presence of corpus predominant gastritis and that this was a greater risk factor than either atrophy or intestinal metaplasia. Irrespective of whether proton pump inhibitor therapy accelerates atrophy, there is cause for concern that it does induce the pattern of gastritis most associated with increased risk of gastric cancer. However, the association of two factors does not confirm a cause and effect relationship. It is not known whether corpus gastritis by itself increases the risk of cancer or whether it is just an epiphenomenon induced by some underlying factor which represents the link with cancer. Consequently, we do not know whether inducing a corpus predominant gastritis by proton pump inhibitor therapy will, in itself, increase the subsequent risk of gastric cancer. The paper by Kuipers and colleagues in this issue of Gut is useful in that it demonstrates that it is readily feasible to eradicate H pylori infection in patients on long term proton pump inhibitor therapy and by so doing achieve resolution of the corpus predominant gastritis. The paper also claims that treating H pylori infection results in some resolution of corpus atrophy. However, there are concerns about assessing the severity of atrophy following eradication of H pylori infection. Resolution of inflammation makes quantification of atrophy difficult and, of course, also makes it impossible for the observer to be blinded to the patient’s H pylori status. However, as discussed above, resolution of the inflammation might be more important than any possible resolution of atrophy with respect to the risk of gastric cancer. The question regarding the appropriateness of eradicating H pylori infection in patients with reflux disease must also address the effect this may have on the reflux disease itself and its response to treatment. There are now reliable data indicating that H pylori infected subjects have a lower prevalence of reflux disease and some data indicating that patients with the more virulent CagA positive strain of H pylori infection have a lower incidence of reflux oesophagitis and its complications than those with CagA negative strains. 18 These observations have stimulated interest in the possibility that H pylori infection may afford some protection from reflux disease and that the increasing prevalence of this disease and its complications in the Western world may be partly explained by the fall in prevalence of H pylori infection. Again, however, we must recognise that associations do not confirm causality as they may be due to confounding factors. However, the observation that CagA positive strains are associated with less reflux oesophagitis than CagA negative strains does raise the real possibility of a protecting effect as this comparison within H pylori infected subjects removes confounding factors related to susceptibility to the infection. If H pylori infection does provide protection against reflux disease then eradicating the infection should induce or aggravate the condition. The data on this question are also conflicting. Labenz et al reported that eradicating H pylori increased the incidence of oesophagitis in ulcer patients. However, Moayyedi et al did not find any increase in reflux symptoms following eradication of H pylori in patients with symptomatic heartburn. Swhwizer et al reported improvement in reflux symptoms following H pylori treatment but their study was criticised for its small size and inadequate matching of randomised groups. Two studies have observed development of reflux disease COMMENTARIES 5