Personal review: alarmism or legitimate concerns about long-term suppression of gastric acid secretion?


Dent Department of Gastrointestinal Medicine, Royal Adelaide Hospital, Adelaide SA 5000, Australia. E-mail:


This article responds to controversial issues about the long-term use of acid suppression raised in a recent article in this journal by Waldum & Brenna. Although rebound acid secretion occurs following proton pump inhibitor therapy, the clinical significance of this is unclear, but the proposal that this is a major driver of acid-related diseases is considered implausible. The polypoid deformity of the gastric corpus that can occur with long-term proton pump inhibitor therapy is not neoplastic, and therefore has no bearing on other issues raised about proton pump inhibitor therapy and gastric malignancy. Current data in humans suggest that the magnitude of serum gastrin elevation from proton pump inhibitor treatment of up to 10 years, and any theoretical risks from this, have been overstated by Waldum & Brenna. Pernicious anaemia is a model of very doubtful validity for the risks of proton pump inhibitor therapy on several grounds. The proposal that diffuse gastric carcinoma arises from acid suppression-induced stimulation of enterochromaffin-like cells is challenged vigorously, because this is based on an implausible and substantially criticized interpretation of histopathology. It is agreed that it is appropriate to be cautious about the safety of long-term acid suppression, because no data are available for lifelong treatment in humans. Such caution should be tempered by a critical assessment of the benefits of this treatment in relation to any possible risks. The substantial data that now exist from long-term treatment of humans with proton pump inhibitors has not thus far revealed any definite risks. The risk of death from anti-reflux surgery, although small, would seem to far exceed any possible risks associated with long-term proton pump inhibitor use. Available data suggest that denial of the benefits of effective acid suppressant therapy to patients with clear-cut troublesome acid related disorders is an overreaction to concerns about the biological effects of inhibiting acid secretion with proton pump inhibitors.


The long-term safety of acid suppression has been controversial ever since the development of the histamine2-receptor antagonists (H2RAs). In an article that appeared recently in this journal,1 Waldum & Brenna question the wisdom of long-term use of acid suppression on the basis of several concerns, directed particularly at proton pump inhibitors because of their greater effects on acid secretion and intra-gastric pH. In our opinion, their view of the risks of acid suppression is unreasonably gloomy, fails to put the benefits of this treatment in correct context and, most importantly, contains what we consider to be serious misinterpretations of pathological findings on the origins of diffuse gastric carcinoma.

There are three main questions that should be asked about each argument that Waldum & Brenna or others put forward against the use of acid suppression, these being: (1) are the observations valid and correctly interpreted, (2) what do the findings, if correct, tell us about the risks of use of proton pump inhibitors in humans, and (3) is the balance of risks from the use of proton pump inhibitors such that we should not be using these agents to treat troublesome acid peptic disorders? Our comments below attempt to make this assessment on each of the issues raised by Waldum & Brenna.


Yes, there may be a slightly increased risk for enteric infection during proton pump inhibitor therapy as stated by Waldum & Brenna. The benefits of appropriately used proton pump inhibitor therapy would seem to outweigh any risks or inconvenience associated with this effect.


This does occur, but its clinical significance is not well understood. We would submit that Waldum & Brenna are drawing a very long bow to suggest that rebound hyper secretion is a major driver of morbidity from acid peptic disorders, and consequently the expansion of use of proton pump inhibitors. The treatment of gastro-oesophageal reflux disease (GERD) accounts for a major part of proton pump inhibitor prescriptions. It is true that the chronicity of reflux disease has only been generally recognized during the period that proton pump inhibitors have been available, but the literature is replete with evidence that reflux disease is a chronic problem regardless of whether proton pump inhibitors have been used or not. Similarly, chronic peptic ulcer is by nature a chronic disorder, whether occurring as a result of Helicobacter pylori infection or in association with aspirin or nonaspirin nonsteroidal anti-inflammatory drug use. If removal of risk factors for chronic peptic ulceration such as H. pylori infection or ulcerogenic drugs is not feasible, then proton pump inhibitor therapy is extremely successful in reducing the substantial risks that these patients face from having peptic ulcer recurrence.

Whilst we would accept that there is a need for more research into the clinical significance of acid rebound, we do not think that this is actually driving the severity of acid peptic disorders or the sales of proton pump inhibitors. Rather, these agents have been so successful because they work very well and better than anything else. Compared to the gains derived from effective treatment of acid peptic disorders, we consider acid rebound to be a minor issue.


Waldum & Brenna are correct in noting that case reports suggest that a polypoid deformity of the gastric corpus mucosa develops occasionally in people who take proton pump inhibitors continuously. Such reports don’t tell us what the risk for development of this change is, because there is no denominator. Structured surveillance of patients treated with continuous proton pump inhibitor therapy for periods of up to 11 years suggest that this change develops rarely.2 More importantly, as Waldum & Brenna point out, this is not a neoplastic polypoid change, but rather a cystic response of the gastric mucosa to the alteration of its physiology by proton pump inhibitors. There is no indication whatsoever that this pattern of mucosal change carries any risk to the patient. It is therefore somewhat misleading for these changes to be referred to within a discussion of possible mechanisms of tumorigenesis associated with acid suppression. The polypoid gastric mucosal deformity appears only to be a minor, even trivial matter that might on occasion raise concerns when seen endoscopically.


Those familiar with the development of the proton pump inhibitors will remember the flurry of research that followed the finding that omeprazole induced enterochromaffin-like (ECL) cell carcinoids in aged female rats, which is referred to by Waldum & Brenna.3 The setting was high dose administration, leading to virtual achlorhydria for most of the rat life span. This effect was not seen in mice or dogs. Subsequent work showed that the mediator was the very high levels of plasma gastrin produced in these animals by acid suppression, in combination with the susceptibility of females of this particular rat strain to this effect. This conclusion was strongly supported by the finding that ECL cell carcinoids developed in rats rendered markedly hypergastrinaemic by other means, such as with very high dose ranitidine4 or subtotal resection of the oxyntic mucosa.3

ECL cell carcinoids do sometimes occur in patients with the Zollinger–Ellison syndrome (especially those with the multiple endocrine neoplasia syndrome [MEN]),5 or longstanding pernicious anaemia as pointed out by Waldum & Brenna.6 Both of these situations are characterized by very marked hypergastrinaemia for a long period. Even then, the great majority run a benign course. In the series of Kokkola et al.6 referred to by Waldum & Brenna, none of the gastric carcinoids had metastasized during the period of observation. The rarity of malignant ECL cell tumours of whatever origin is shown by the small number accumulated in a referral centre with an international reputation for its interest in such tumours7—giving an incidence that is conservatively less than one per 10 million people per annum in their referring population.

There is no doubt though, that some stimulation of ECL cells does occur as a result of acid suppression. Sanduleanu et al.8 found that serum levels of chromogranin A (an endocrine cell product) were increased in patients treated with omeprazole.

How well-based then are Waldum & Brenna’s concerns about the postulated sequence of high luminal pH and hypergastrinaemia, leading to ECL tumours in patients taking acid secretion inhibitors? Waldum & Brenna state that ‘…to raise pH above 4…will totally abolish inhibition of gastrin release by acid’. This is simply incorrect. For example, in our own 12-month study of omeprazole for treatment of GERD,9 the mean plasma gastrin concentration, although increased, remained in the normal range. There were no significant changes in ECL cell numerical densities, and no dysplastic ECL cell changes. Similar results have been reported by Lundell et al.10 Freston et al.11 recently reported the US data on lansoprazole treatment in more than 3000 patients treated for up to 1 year. Furthermore, in this study, median plasma gastrin concentrations were within the normal range, with less than 1% of patients having two or more values greater than 500 pg/mL. Again, there were no significant changes in gastric endocrine cell histology. The greatest experience is with omeprazole given for up to 11 years.2 In this study, the findings on serum gastrin and gastric histology in a cohort of 230 patients with a mean follow-up time of 6.5 years showed only minor changes. In our opinion, these data on long-term use of proton pump inhibitors do not support Waldum & Brenna’s concerns.


It is well established that gastric carcinoids and gastric adenocarcinoma occur infrequently in patients who have pernicious anaemia, but still apparently at a higher rate than those without pernicious anaemia. Although the risk to pernicious anaemia patients from the development of gastric carcinoid tumours is small, this is not the major issue. Instead, we advise caution about generalizing from pernicious anaemia patients to those treated long-term with proton pump inhibitors on two main grounds. First, the impact of pernicious anaemia on luminal pH and serum gastrin is far in excess of that produced by proton pump inhibitor therapy. Second, the mucosa of these pernicious anaemia patients is already highly abnormal as a result of the disease processes that produced pernicious anaemia. This mucosal disease is more likely to be the cause of the modestly increased risk for development of gastric carcinoma than the gross changes in serum gastrin or luminal pH.


We are particularly concerned that Waldum & Brenna state that ‘…the first report of gastric carcinoma ascribed (our emphasis) to the use of inhibitors of gastric acid secretion, has been published…’. The authors of the case report referred to12 did not ascribe the patient’s carcinoma to the 6 months of H2-receptor antagonist and proton pump inhibitor therapy, they merely reported an association. Given what we know about the frequency of use of acid suppressant drugs and its minor impacts on gastric histology, this is probably a coincidence.

Waldum & Brenna have broadened the debate substantially by claiming that ‘…an important proportion of gastric carcinomas…seems to be derived from the ECL cell’. They support this by citing two papers from their group.13, 14 In the first13 some general endocrine markers seen in mature ECL cells, as well as several other cell types, were detectable in a minority of cells from seven of 19 gastric cancers of the diffuse type. Demonstration of the presence of diverse gene products is common in tumour cells of many origins15, 16 and certainly should not be taken as evidence that these gastric cancers arose from ECL cells. Waldum & Brenna seem to have lost sight of their own finding that all 19 tumours that they reported had evidence of mucous cell differentiation,13 a finding that also argues strongly against their origin from ECL cells. In Waldum et al.’s more recent paper14 three electron micrographs have been interpreted as showing neuroendocrine granules in tumour cells. Without electron microscopic immunochemistry this conclusion is not sustainable—the granules might even have been composed of the non-glycoprotein core material of mucous granules.17 The interpretation of the histopathology in these patients by Waldum et al. has been challenged vigorously by pathologists expert in this field.18 We think it much more likely that diffuse gastric cancers arise from gastric stem cells or near stem cells.


It is not possible to say that pharmacological suppression of acid secretion will never lead to harm under any circumstance, but it is important to keep a balanced perspective of the risk/benefit equation. Waldum & Brenna are correct in reminding us that we should be thinking in terms of any risks over decades. As yet, no patient has taken a proton pump inhibitor for more than 20 years, and very few have reached this milestone with H2-receptor antagonist therapy. The best we can do is to try to estimate any risks, and weigh them against the known benefits—especially in troublesome reflux disease. Based on the data summarized above, our judgement is that any risk will be very small.

If proton pump inhibitors are considered to carry unacceptable risk, then what are the alternatives for treatment of reflux disease? Withholding of effective therapy is certainly unreasonable, because reflux disease has a significant impact on quality of life.19 Furthermore, inadequate treatment of oesophagitis carries with it the possibility of development of Barrett’s oesophagus, with its attendant risk for oesophageal adenocarcinoma. It is impossible to estimate how many cases of oesophageal adenocarcinoma may be prevented by proton pump inhibitor therapy, but it is difficult to believe that this benefit would be outweighed by any putative risk from development of ECL cell neoplasia as a result of acid suppressant therapy. Failure to treat oesophagitis adequately is also associated with development of peptic stricture, dilatation of which carries a small but measurable risk of death. Anti-reflux surgery is a commonly used and appropriate management option in some patients, but it is certainly not without risk. For instance, in two large recent series20, 21 the operative mortality was around 0.2%, a risk that must be greatly in excess of any possible risk from use of proton pump inhibitors.

Blind faith in the safety of newer treatments is inappropriate, and we can thank Waldum & Brenna for reminding us of some putative risks that have at least some basis in cell biology. However, to express concerns about putative risks, based at least in part on misinterpretation of pathology, is unbalanced, alarmist and unhelpful to our patients. We remain happy to treat our own reflux disease, and that of our patients with acid suppression.


Both authors are consultants to AstraZeneca in the field of the treatment of acid peptic disorders, and have been recipients of research grants from other manufacturers of acid suppressant drugs.