Atrophic gastritis and helicobacter pylori infection in patients with reflux esophagitis treated with omeprazole or fundoplication

Authors: Kuipers EJ, Lundell L, Klinkenberg-Knoll EC, Havu N, Festen HPM et al.
Source: New England Journal of Medicine. 334:1018-22. April 18, 1996.
Institutions: Multiple university hospitals, five in the Netherlands one in Sweden.
Financial support: Astra, Sweden.



Helicobacter pylori infection causes chronic gastritis, which can lead to atrophic gastritis, itself a precursor of gastric cancer. Raising the gastric pH can alter or worsen chronic gastritis in patients infected with H. pylori, and can also raise gastrin levels, both of which favor atrophic gastritis. This study was designed to look at the interaction between H. pylori infection, omeprazole therapy and the development of atrophic gastritis.


The study compared gastric biopsies in two cohorts of patients with reflux esophagitis, one treated with omeprazole and one treated surgically with fundoplication.

Results Authors' Discussion

The authors conclude that patients receiving prolonged acid-suppressive therapy with omeprazole who are also infected with H. pylori are at risk for developing atrophic gastritis. They imply that omeprazole-treated patients who are H. pylori negative, however, are not at increased risk for atrophic gastritis. They note that there are some differences between the two cohorts (the omeprazole group was 9 years older, on average). Nevertheless, they believe that these differences are not sufficient to explain the results of their study.


This is not a randomized trial but a comparison between two cohorts of patients. As such, any comparisons between the groups are of dubious validity, especially since the two groups are from different countries and little baseline data is given about them.

The data from the fundoplication cohort are incomplete (53 out of 137 were not analyzed because of an incomplete series of biopsy specimens). Because of the small numbers, we cannot conclude that H. pylori infection is not associated with atrophic gastritis in the absence of acid-suppression.

The data from the omeprazole group are more complete, and the results do suggest that, in this group, patients infected with H. pylori are at significantly higher risk for developing atrophic gastritis than those who are not infected.

Because of the problems with comparing the two cohorts, I do not believe the data show that patients receiving omeprazole therapy who are not infected are not at increased risk for atrophic gastritis. Furthermore, it is not possible to extrapolate these results to therapeutic efficacy -- we don't know if antibiotic treatment will decrease the incidence of atrophic gastritis in these patients and thus make omeprazole therapy "safer".

This study is another one that adds to the growing number of reports indicating a significant pathogenetic role for H. pylori. Given its high prevalence in the upper age groups, treatment in order to prevent atrophic gastritis and gastric cancer would be an enormously expensive proposition. Targeted treatment of high risk groups (such as the acid-suppressed group examined here) needs to be investigated by randomized trials.


Reader comments

June 24, 1996

The authors of this paper reply:

We appreciate the attention and comments of Dr. Jacobson made as contribution to this journalclub on our recent paper in the NEJM. We would like to use the opportunity to address the issues raised by Dr. Jacobson in his comments.

Dr. Jacobson states that we can not conclude from our data that H. pylori infection is not associated with atrophic gastritis in the absence of acid-suppression. That is a valid statement, which however, with all due respect, is besides the point of our paper. The introduction of our paper actually starts with giving an overview of the strong association between H. pylori and atrophic gastritis, ending with the sentence that 'the development of atrophic gastritis induced by persistent H. pylori infection is an essential step in the cascade of events leading to gastric cancer' (NEJM 1996; 334: 1018*).

H. pylori causes chronic gastritis in virtually all infected individuals. In many of them, this persistent inflammation leads to ultimate development of multifocal atrophic gastritis. However, this is generally a very slow process. As pointed out in our discussion, various cohort follow-up studies described an annual increase in the prevalence of atrophy of approximately 1 - 3% annually. These studies were done both in developed countries (Finland, the Netherlands) and developing countries (Estonia, Colombia), but nevertheless had very concordant results. The first study that differentiated between H. pylori positive and negative subjects in its cohort, was our study from Amsterdam, showing an annual 0.3% increase of atrophy prevalence in uninfected compared to a significantly higher 1.8% in infected subjects (Lancet 1995; 345: 1525-8*). The point thus is that H. pylori is a very important factor in the process of development of atrophic gastritis, but that this process is slow. After 20 years of follow-up, approximately 1 out of 3 infected adults will have signs of atrophy.

The chance to develop atrophy is however dependent upon the severity of gastritis, which is determined by characteristics of the bacterial strain and of the infected host. We have recently shown that within the population of H. pylori infected subjects, those that are infected with a cagA positive strain have more active gastritis and a 2-fold higher chance to develop atrophy than those that are infected with a cagA negative strain (J Natl Cancer Inst 1995; 87: 1777-80*). This sheds new light on the observation that acid suppression also increases the severity of gastritis, in particular in the gastric body. This effect has consistently been observed by various research groups, including Dr. Logan from the UK (Gut 1995; 36: 12-6), Dr. Solcia from Italy (Scand J Gastroenterol 1994; 201: 28-34) and ourselves (Am J Gastroenterol 1995; 90: 1401-6*). In the past, the same effect had also been observed in DU patients treated with a vagotomy. This led to our hypothesis that acid suppressive maintenance therapy may increase the risk for atrophic gastritis (see Am J Gastroenterol 1995; 90: 1401-6, or Aliment Pharmacol Ther 1995; 9: 331-40*). For that purpose, we evaluated the development of atrophy in our cohort of GERD patients treated with omeprazole. As pointed out in our NEJM paper, the H. pylori negative GERD patients had an annual rate of atrophy development (0.8%) that was very low and not significantly differed from our Amsterdam volunteer population of the same age not treated with omeprazole (0.3% ; Lancet 1995; see above). However, the H. pylori positive omeprazole treated GERD patients had indeed a high rate of atrophy development (6.1%), which significantly differed from the 1.8% found in our Amsterdam volunteer population of the same age not treated with omeprazole (Lancet 1995; see above). At this rate, it did not take 20 years, but only 5 years for atrophy to develop in 1 out of every 3 infected subjects! We presented these data on the 1995 AGA meeting in San Diego and submitted them to the NEJM, including the comparison with the Lancet - Amsterdam cohort of subjects without specific disease or treatment. The main comment we received, was that, although both cohorts were from the same city and had comparable mean ages, we could not exclude the possibility that GERD by itself increased the chance for atrophy development. For that reason, we then made another comparison with the Swedish fundoplication cohort, again showing that H. pylori negatives did not differ whether treated with omeprazole, fundoplication, or nothing at all (Amsterdam - Lancet cohort), whereas in H. pylori positives the acid suppressive therapy significantly increased the rate of atrophy. This hypothesis receives further strong support from a comparison with all data from other histological cohort follow-ups. I refer for this comparison to the graphical display of these data as presented in the Am J Gastroenterol 1995; 90: 1401-6*. The new data published in the NEJM fit very well the data presented in this graph.

Thus, the point of our paper is that H. pylori causes gastritis, which can lead to atrophy. The chance to develop atrophy is dependent upon the severity of gastritis. More than ten different short-term studies have now consistently shown that profound acid suppression increases gastritis activity and our NEJM study is the first long-term study on this topic. The data from this study strongly suggest that the permanently increased gastritis activity has important long-term implications. These data are supported by the other limited available cohort follow-up data as extensively discussed in the discussion of our paper.

This brings me to the specific points raised by Dr. Jacobson. First of all, this is indeed not a randomized study, although I do not see why this automatically without any specific arguments makes the comparison of dubious validity. As I already pointed out, the data do fit with those of other available cohort studies and they also do fit a plausible explanation for the effect of an increased development of atrophy. Therefore, the comparison may actually be very valid. The two populations are indeed from two countries (Sweden and the Netherlands), but it would serve the American reader to know that the distance between the two cities is about the same as the distance between San Francisco and San Diego within one state of the USA. Furthermore, both countries contain a largely caucasian population and are very similar with respect to socio-economic status, dietary, drinking and smoking habits, medication use, life expectancy and incidence of atrophic gastritis and gastric cancer. Baseline data of both cohorts are not incomplete, they are given in both the NEJM paper and in the two other papers that we refer to. Dr. Jacobson raises the need for randomized controlled studies. Although we agree with this need, the chance that they will ever be performed on this topic is close to nil, as no ethical committee will allow the randomization of GERD patients to medication or surgery.

A total of 137 Swedish patients was treated with a fundoplication cohort, yet only in one hospital were the biopsy specimens taken according to the same protocol as in our omeprazole treated patients. Therefore, we had to use only 72 of the 137 fundoplication patients. These did not significantly differ from the excluded patients.

The discussion then comes back to the comparison between the two cohorts. Dr. Jacobson states that he believes that H. pylori uninfected patients receiving omeprazole may still be at increased risk to develop atrophy. What can we say? We observed an approximately 10-fold higher chance for omeprazole treated GERD patients to develop atrophy in the presence of H. pylori compared to those who were not infected, and an even 35-fold higher chance to develop argyrophil cell hyperplasia. Furthermore, the rate of atrophy development was as low in the non-infected omeprazole treated GERD patients as it was in a non-treated non-GERD volunteer population (see Lancet publication). This is strong evidence against the belief of Dr. Jacobson and in favour of the safety of omeprazole maintenance treatment. Finally, we fully agree with Dr. Jacobson that the results can not be extrapolated to therapeutic efficacy. This is actually the last sentence of our paper, saying that 'it remains to be seen whether eradication therapy can prevent atrophy and argyrophil-cell hyperplasia'. This needs to be determined in future studies. Therefore, we suggest in our paper that at this time we should consider H. pylori eradication in H. pylori infected GERD patients requiring profound acid suppressive maintenance therapy.

Ernst J. Kuipers, M.D., Ph.D.
Dept. of Gastroenterology, Free University Hospital, Amsterdam, The Netherlands
Vanderbilt University School of Medicine, Nashville, TN

* Please submit any reprint requests to: Ms. Sylvia Band, e-mail address:

From: Michael Jacobson

Date: Mon, 01 Jul 1996
From: Howard Homler MD <>

Prilosec seemed too good to be true! Once the fear of inducing gastric carcinoid tumors quieted down, we really thought we had a wonder drug. The study reviewed here is a timely warning to avoid the cavalier use of proton pump inhibitors in our GERD patients. I would suspect that the incidence of chronic atrophic gastritis in H pylori infected individuals treated with H-2 blockers would be intermediate between the control group and the group treated with Prilosec...seem reasonable?

Dr. Kuipers (author of the paper) responds:

 Date: Tue, 02 Jul 1996
From: kuipere@ctrvax.Vanderbilt.Edu (Ernst Kuipers)

I fully agree that the hypothesis brought forward makes much sense. The data to support or refute it are however still very scarce. I am aware of two studies that claimed some increase of H. pylori associated body gastritis during H2-blocker and also during antacid therapy (Stolte et al. Ir J Med Science 1992; proceedings of the IVth (?) International meeting on H. pylori, respectively Lanza et al. Am J Gastroenterol 1994). However, there have been no cohort follow-up data published yet to show whether or not this effect would make any difference with respect to the development of atrophy. There has been one cross-sectional study (Penston et al. Aliment Pharmacol Therap 1990) that focused on long-term effectivity of 150-300 mg ranitidin maintenance treatment in DU patients and did not find atrophic gastritis at a single observation after five years of therapy. However, there is obviously a clear need for additional data, to address this issue.

Ernst J. Kuipers, M.D., Ph.D.
Dept. of Gastroenterology, Free University Hospital, Amsterdam, The Netherlands
Vanderbilt University School of Medicine, Nashville, TN

Date: Sat, 10 Aug 1996
From: (guest login)

What about lansoprazole (prevacid)? In vitro, it is more active against H. pylori than omeprazole, but costs a bit more. For formulary consideration, I would keep omeprazole.


Dr. Kuipers' response:

 Date: Tue, 13 Aug 1996
From: kuipere@ctrvax.Vanderbilt.Edu (Ernst Kuipers)

 With respect to lansoprazole, my reply is twofold:

1/ The effects of proton pump inhibitors on H.pylori are not exerted by the inactive prodrug that we prescribe, but by the activated sulphenamide form of the drug. This means that the PPI's first have to bind a proton before they have any effect on Hp. Such binding of a proton in the lab is obviously performed by acidifying the solution. In vivo however, this occurs predominantly within the vesicles of the parietal cell, where it is then immediately and irreversibly bound to the proton-pump. It is therefore on theoretical grounds very questionable whether PPI's have any significant anti-Hp effect in vivo. There are to my knowledge no data that prove otherwise. Patients can be treated with high dose PPI's for years and will not become Hp negative (in our recent NEJM study, we did not observe any patient to become Hp negative during 3-8 years omeprazole therapy for GERD). The important contribution of PPI's in Hp eradication therapies is particularly due to the acid suppressive effect. This is true for omeprazole, but also for lansoprazole and pantoprazole. Even though lansoprazole may have a greater in vitro antibacterial effect on Hp, the mechanism is still the same and we have to assume that the clinical importance of this antibacterial effect is minimal .

2/ Hp positive patients treated with lansoprazole do get exactly the same increase in corpus gastritis as observed during omeprazole therapy. With maintenance therapy, this does place them at risk for the development of atrophic gastritis and argyrophil cell hyperplasia just as with omeprazole therapy (see for instance Eissele et al. Gastroenterology 1996; 110 (4): A101 (abstract). As mentioned in our NEJM paper, the effect is associated with the level of acid suppression and not with the method by which the acid suppression is being achieved (e.g. it also occurs after vagotomy). Therefore, the same advice is to treat these patients with Hp eradication at the start of lansoprazole maintenance therapy.

With kind regards,


Ernst Kuipers

Dr. Ernst Kuipers
Vanderbilt University Medical Center
Department of Infectious Diseases

September 5, 1996

Letters to the Editor about this article from the NEJM site.

Journal Club on the Web main screen

Submit a comment about this article

Site Meter