Personal review: is profound acid inhibition safe?
Waldum Dr Department of Medicine, Trondheim University Hospital, N-7006 Trondheim, Norway. E-mail: email@example.com
Inhibitors of gastric acid secretion, particular proton pump inhibitors, are effective drugs in the treatment and prophylaxis of acid-related diseases. Proton pump inhibitors are therefore prescribed widely, often for minor complaints.
Gastric acidity kills swallowed microorganisms, and acid secretion must be of biological importance because it is maintained in phylogenesis. Acid secretion is controlled by feedback mechanisms, mainly via gastrin. A decrease in acidity always causes an increase in plasma gastrin. The trophic effect of gastrin leads to hyperplasia and neoplasia of the enterochromaffin-like (ECL) cell. ECL cell derived tumours in man were previously regarded as rare, and also as rather benign. It is now clear that the ECL cell gives rise to a significant proportion of gastric carcinomas. Moreover, ECL cell carcinoids secondary to hypergastrinaemia may develop into highly malignant tumours.
Treatment with a proton pump inhibitor is followed by rebound acid hypersecretion and decreased efficiency of H2-blockers, thus such treatment may induce a type of physical dependence. It is therefore reasonable to be cautious and not to treat younger (< 50 years) patients for long periods of time with profound inhibitors of gastric acid secretion. Chromogranin A in the blood is a sensitive marker of the ECL cell mass, and it could be used to survey patients on long-term proton pump inhibitors.
With the development of histamine-2 (H2) blockers1 and later the proton pump inhibitors,2 efficient drugs in the treatment of acid-related diseases such as peptic ulcer and gastro-oesophageal reflux disease (GERD) became available. The proton pump inhibitors are far more efficient in inhibiting gastric acid secretion, and have been found to be superior to H2-blockers in healing peptic ulcer,3 and particularly in treating reflux oesophagitis.4 Proton pump inhibitors are therefore used increasingly, and are also now used for minor complaints like heartburn.5
Although proton pump inhibitors have been reported to cause rather few side-effects related to the compounds themselves, it should be recalled that the development of the first proton pump inhibitor, omeprazole, was temporarily stopped due to enterochromaffin-like (ECL) cell carcinoids in the oxyntic mucosa of rats secondary to hypergastrinaemia following life-long treatment with omeprazole.6 Thus, the drug induced tumours in its target organ.6 Nevertheless, omeprazole was allowed to proceed into clinical use. It is generally accepted that the proton pump inhibitors in clinical use have few side-effects by themselves.7 Their toxicity is related to their biological effect, i.e. reduction of gastric acid secretion, resulting in hypo/anacidity and secondary hypergastrinaemia. Moreover, the concentration–response curve for their inhibition of acid secretion is very steep,8 making it difficult to induce a moderate reduction in gastric acidity. Furthermore, due to individual differences in sensitivity towards these agents,9 dose adjustments may only affect the proportion of patients over-treated to those under-treated. Not only the proton pump inhibitors,6 but also other inhibitors of gastric acid secretion, like the so-called insurmountable H2-blockers,10 tend to induce ECL cell carcinoids in the rat. Moreover, even ranitidine at high doses induces ECL cell carcinoids in the rat,11 further demonstrating that the tumorigenic effect of inhibitors of acid secretion is related to degree of acid inhibition and not to any specific group of drugs. However, proton pump inhibitors are used in doses giving much more profound acid inhibition and thus more severe hypergastrinaemia than H2-blockers.12 In contrast to proton pump inhibitors, there is a tendency to tolerance development with the long-term use of H2-blockers,13 which reduces the efficacy of these drugs and thus their hypergastrinaemic effect. Therefore, in ordinary clinical doses the risks of long-term treatment with H2-blockers are low compared to the risk when using proton pump inhibitors. Partial corpectomy without any use of drugs also induces ECL cell carcinoids,14 further establishing that hypo/anacidity and hypergastrinaemia cause these tumours.
In this short review we will discuss the risks of hypo/anacidity accompanied by hypergastrinaemia, and then discuss individually the risks of gastric hypoacidity and hypergastrinaemia.
HYPO/ANACIDITY AND HYPERGASTRINAEMIA
Acid (H+) inhibits gastrin release up to pH 4.0.15 Therefore, patients with atrophic gastritis affecting the oxyntic mucosa resulting in the loss of oxyntic glands and reduced gastric acidity will have hypergastrinaemia.16, 17 It has long been known, even when omeprazole was accepted for clinical use, that patients with pernicious anaemia have an increased risk of developing both carcinoids originating from the ECL cell18 and adenocarcinomas in the oxyntic mucosa.19 However, the increased risk of adenocarcinoma in patients with pernicious anaemia is fairly low (a factor of about 3),20 and because patients with pernicious anaemia have had a profound hypergastrinaemia for many years, the slight increase in gastric adenocarcinomas in these patients has not been regarded as relevant for patients treated with proton pump inhibitors. Remarkably, nobody seems to have taken into account that atrophic gastritis and thus pernicious anaemia is a disease diagnosed at a rather old age,21 and therefore occurs in patients with reduced lifespan after diagnosis.
Because carcinogenesis may be a very slow process, as exemplified by the lag between gastric resection and the development of gastric stump cancer,22 knowledge of the cancer frequency in patients with pernicious anaemia20 cannot be taken as proof of the safety of iatrogenic hypo/anacidity in young individuals. Furthermore, compared to a median duration of 39 years and minimum lag time of 21 years between gastric resection and diagnosis of stump cancer,22 the 10–15 years’ experience with proton pump inhibitors23 is insufficient to show that these agents will not cause malignancies. It should also be mentioned that the prevalence of ECL cell carcinoids in patients with pernicious anaemia has previously been underestimated.24
DIRECT CONSEQUENCES OF HYPO/ANACIDITY
It is generally accepted that the main function of gastric acid is to kill swallowed microorganisms. The combination of the strong acid with the proteolytic enzyme pepsin and the gastric lipase makes the gastric juice unique in its ability to destroy microorganisms. The importance of this function is indirectly shown by the preservation of acid secretion in the upper gastrointestinal tract in vertebrates,25 and by the regulatory mechanisms found in each individual which try to keep the pH below a certain level via regulation of gastrin release. In man, pH is maintained below 4.0.26 Above this pH, there is no inhibition of gastrin release by H+.15 Interestingly, the bactericidal effect of gastric juice is abruptly lost or reduced at this pH,27 further indicating the role of the gastric juice in the defence against microorganisms. The gastric juice also destroys most viruses, although a few, the so-called enteropathogenic viruses, have a capsid protecting them from destruction by the gastric juice. The clinical consequences of removal of gastric acidity is not fully known, but because microorganisms are living organisms, even short-term iatrogenic reduction of gastric acidity could have serious long-term consequences by allowing microorganisms to enter the gut where they could gain access into the body. It should be recalled that the alimentary tract above the stomach is lined by stratified squamous epithelium, which is relatively resistant to penetration by microorganisms. Actually, patients with hypo/anacidity are known to be more susceptible to infection with enteropathogenic bacteria,28 and more prone to developing bacterial contamination of the upper small bowel.29 Moreover, N-nitroso compounds have been incriminated in gastric carcinogenesis,30 whereas vitamin C is thought to protect against the harmful effects of N-nitroso compounds by preventing nitrosation.31 Interestingly, treatment with omeprazole was recently reported to increase the level of nitrate and reduce the level of vitamin C in gastric juice,32 thus unfavourably changing the intragastric environment. Furthermore, the negative consequences of drug-induced gastric hypo/anacidity for the development of Helicobacter pylori gastritis are now increasingly recognized.33
The main actions of gastrin are to stimulate acid secretion and the growth of the oxyntic mucosa. The acid stimulatory effect of gastrin is exerted by stimulation of histamine release from the ECL cell.34, 35 The ECL cell has a gastrin (CCK-B) receptor.36 On the other hand, it is questionable whether there is a gastrin receptor on the parietal cell,35 although the gastrin receptor was cloned from dispersed oxyntic mucosal cells highly purified in parietal cells.37 The trophic effect of gastrin on the oxyntic mucosa was thought to be due to a direct stimulation of the progenitor cell,38 but this effect could be indirect, mediated by histamine39 or other mediators from the ECL cell.40 Unequivocally, there is a gastrin (CCK-B) receptor on the ECL cell,34–36 regulating both its function34 and its proliferation.41–43 The ECL cell must be regarded as a main target cell of gastrin in the oxyntic mucosa. The ECL cell is self-replicative,41, 43 and the concentration–response curve for the functional and trophic effects appears to be similar both in the rat34, 42 and man.44, 45 Moreover, it should be realized that the maximal trophic effect of gastrin on the ECL cell is reached at a concentration of about 500 p M.42, 45 Every condition with long-term hypergastrinaemia, whether secondary to hypo/anacidity6, 10, 11, 14, 24, 45, 46 or not,47, 48 tends to cause ECL cell tumours. It has been claimed that gastritis itself could play a role in the development of ECL cell carcinoids,49 but this hypothesis lacks experimental evidence. On the other hand, hypergastrinaemia alone has indeed been shown to be sufficient to induce ECL cell carcinoids, as demonstrated in transgenic mice which were made moderately hypergastrinaemic by making the gastrin gene be controlled by the insulin promotor.50 The much higher frequency of ECL cell carcinoids in patients with Zollinger–Ellison syndrome as a part of the multiple endocrine neoplasia type-I (MEN-I) syndrome compared with those with spontaneous gastrinomas,47, 51 suggests that genetic susceptibility also plays a role in the development of ECL cell carcinoid tumours. However, ECL cell carcinoids also occur in patients with sporadic gastrinomas.52 It should also be recalled that the biological behaviour of gastrinomas as part of the MEN-I syndrome probably differs from those occurring sporadically.53 Thus, they tend to be more benign than those occurring sporadically.53 This biological difference may result from a longer period with hypergastrinaemia in patients with gastrinoma as part of MEN-I syndrome compared with the spontaneous ones, and the difference in the frequency of ECL cell carcinoids between these two groups of patients could just reflect the duration of hypergastrinaemia.54
It is now generally accepted that the ECL cell gives rise not only to the relatively benign ECL cell carcinoids, but also to poorly differentiated neuroendocrine carcinomas.55 A significant proportion of gastric carcinomas, particularly those of the diffuse type, appear to be derived from the ECL cell.56, 57 The ECL cell plays an important role in spontaneous gastric carcinogenesis not only in man, but also in the African rodent mastomys,58 and in a strain of Japanese cotton rats.46 Therefore, the argument for accepting omeprazole for clinical use, namely that the ECL cell does not play any role in the aetiopathogenesis of human gastric carcinomas, does not seem to be tenable. In fact, the first report of gastric carcinoids ascribed to the use of inhibitors of gastric acid secretion has been published,59 and there have been some reports on the occurrence of benign non-neoplastic polyps in the stomach of patients treated with long-term omeprazole.60, 61 Such polyps are claimed to be harmless,62 but it should be recalled that hyperplastic polyposis has been described together with diffuse gastric carcinoma.63 It has been reported that neuroendocrine-differentiated tumour cells in gastric adenocarcinomas do not replicate and thus represent terminal differentiation.64 However, when applying the tyramide amplification method together with CgA monoclonal antibodies, we could show that not only the majority of the tumour cells in up to 50% of the gastric carcinomas were CgA immunoreactive,65 but also that CgA immunoreactive tumour cells do proliferate.65, 66
ECL cell tumours have been divided into three types of carcinoids, based upon their pathogenesis, and a fourth group of ECL cell neuroendocrine carcinoma.67 It has been thought that the more benign ECL cell carcinoids developing secondary to hypergastrinaemia do not develop into the more malignant ECL cell neuroendocrine carcinoma.67 However, reports describing such a transition exist,68, 69 which agrees well with tumour development in general, with tumours starting as rather benign tumours and gradually being transformed into more malignant ones, reflecting the multistep carcinogenesis.70
It is therefore indisputable that hypergastrinaemia results in ECL cell hyperplasia and ECL cell carcinoids. Moreover, it is clear that the ECL cell plays an important role in human gastric carcinogenesis. There is therefore reason to be concerned about long-term iatrogenic hypo/anacidity, especially in young individuals. Besides the risk of the development of neoplasms in the long term, the trophic effect of drug-induced hypergastrinaemia results in an increased ECL cell mass and increased histamine release71 with a rebound acid hypersecretion.72–74 The rebound acid hypersecretion in man73, 74 reduces the efficacy of H2-blockers,75 and possibly worsens the reflux disease.76 Long-term omeprazole treatment has also been shown to inhibit gastric emptying in the rat71 as well as man.77 The mechanism for this inhibitory effect on gastric emptying is not known. Neither is it known whether this delay in gastric emptying leads to any clinical side-effects.
GASTRIN RECEPTORS ON CELLS OUTSIDE THE OXYNTIC MUCOSA, AND POSSIBLE CONSEQUENCES OF LONG-TERM HYPO/ANACIDITY AND HYPERGASTRINAEMIA ON EXTRAGASTRIC ORGANS
Håkanson and co-workers failed to show any effect of long-term hypergastrinaemia on any other organ than the oxyntic mucosa in the rat.78 Nevertheless, patients with pernicious anaemia have been reported to have an increased risk not only of ECL cell carcinoids and gastric carcinomas, but also pancreatic neoplasms,20, 79 and even some malignant haematological diseases.20 The mechanism for the increased frequency of pancreatic neoplasia may be a trophic effect of gastrin on the pancreas, where gastrin receptors have been localized to both normal acinar cells80 and pancreatic cancer cell lines.81 If, on the other hand, there is a real increase in haematological malignancies in patients with pernicious anaemia,20 this may be related to growth disturbances secondary to B12 deficiency, or to an increase in infections due to the lack of acidic gastric juice.
Since the report of Johnson and collaborators in the 1970s describing a stimulatory effect of pentagastrin on 14C-leucine incorporation into rat colon cells,82 the role of gastrin in colorectal carcinogenesis has been discussed. This has been supported by the increase in blood gastrin found in patients with colon cancer and adenomatous colorectal polyps,83 and the stimulation of the growth of human colon cancer cells by gastrin.84 Moreover, and of great importance, chronic endogenous hypergastrinaemia in patients with Zollinger–Ellison syndrome showed an increase in colonic mucosal cell proliferation compared with normal.85 However, no increase in colorectal tumours was found in patients with Zollinger–Ellison syndrome.85 Because patients with Zollinger–Ellison syndrome were treated with different agents such as omeprazole, ranitidine and somatostatin, and patients with symptoms attributed to irritable bowel syndrome were used as controls,85 the differences between the patients with Zollinger–Ellison syndrome and the controls could be caused by other factors than just hypergastrinaemia. In a recent prospective study, hypergastrinaemia was also reported to be increased in patients with colorectal cancer, although the median gastrin value was similar in those with, and those without, colorectal carcinoma.86
To summarize, although there are many indications for a role for gastrin in colorectal carcinogenesis,83–86 there is a lack of really hard evidence for a causal role of gastrin in colorectal carcinogenesis despite decades of research.87 It should also be added that, to our knowledge, the gastrin (CCK-B) receptor has not been demonstrated on any cell type occurring in the normal colorectal mucosa.
The lungs are derived from the foregut, like the stomach, and there are many similarities between the Kultschitzky and the ECL cells.88, 89 To our knowledge, gastrin receptors have not been demonstrated on normal neuroendocrine cells in the lung. However, gastrin has been reported to stimulate the growth of small cell lung cancer cell lines, and the gastrin/CCK-B specific receptor antagonist L365.260 was found to prevent this stimulation.90 Thus, the oatcell lung cancer cell line has a gastrin receptor, indicating that some of the normal neuroendocrine cells in the lung may also possess such receptors. However, we do not know of any study indicating that long-term hypergastrinaemia in any animal induces lung tumours. On the other hand, it cannot be ruled out that iatrogenic hypergastrinaemia provoked by potent inhibitors of gastric acid secretion in patients with increased risk of oatcell carcinoma due to tobacco smoking91 could have an additional effect on lung tumorigenesis. This is, however, highly speculative.
It has been long known that (penta)gastrin stimulates calcitonin release from the C-cells in the thyroid gland.92 No report, either in animals or man, has described any C-cell hyperplasia or neoplasia induced by hypergastrinaemia. Nevertheless, a high proportion of medullary thyroid carcinomas have gastrin/CCK-B receptors,93 and accordingly iatrogenic hypergastrinaemia could increase the risk of C-cell derived tumours, particularly in those individuals genetically prone to develop such tumours.
Gastrin/CCK-B receptors are also found in the brain.94 However, because most endogenous gastrins with biological activity do not pass the blood/brain barrier, endogenous gastric gastrins probably do not induce any cerebral side-effects.
To conclude, although it cannot be ruled out that gastrin can induce long-term side-effects outside the stomach, the main problems related to profound acid inhibition of gastric acid secretion are related to the stomach, as shown by animal studies6, 10, 11 and indirect18, 24, 55–57 and direct59–61 human studies.
What then should be the practical consequences of the probable long-term serious gastric consequences of profound acid inhibition? Based upon the experience with patients with pernicious anaemia, the risk of gastric hypo/anacidity appears to be low in patients above the age of 50 years. Younger individuals should not be treated with potent inhibitors of gastric acid secretion for mild oesophagitis, and especially not for reflux symptoms without oesophagitis. Young individuals with moderate to severe oesophagitis should, after a short-term treatment period with a proton pump inhibitor, be offered anti-reflux surgery and those refusing operation and continuing proton pump inhibitors should be followed by regular determination of chromogranin A (CgA),73 because CgA in blood reflects the ECL cell mass.