Rebound acid hypersecretion after long-term inhibition of gastric acid secretion

Authors

  • R. Fossmark,

    1. Department of Cancer Research and Molecular Medicine, Faculty of Medicine, Norwegian University of Science and Technology, Trondheim, Norway
    Search for more papers by this author
  • G. Johnsen,

    1. Department of Surgery, St Olav's Hospital, Trondheim, Norway
    Search for more papers by this author
  • E. Johanessen,

    1. Department of Internal Medicine, Gastroenterological Section, St Olav's Hospital, Trondheim, Norway
    Search for more papers by this author
  • H. L. Waldum

    1. Department of Cancer Research and Molecular Medicine, Faculty of Medicine, Norwegian University of Science and Technology, Trondheim, Norway
    2. Department of Internal Medicine, Gastroenterological Section, St Olav's Hospital, Trondheim, Norway
    Search for more papers by this author

Dr R. Fossmark, Department of Cancer Research and Molecular Medicine, MTFS 5 etasje, Olar Kyrresgate 3, N-7489, Trondheim, Norway.
E-mail: reidar.fossmark@medisin.ntnu.no

Summary

Background : Rebound acid hypersecretion develops after the use of acid inhibitors.

Aim : To estimate the duration of hypersecretion and to elucidate the role of the enterochromaffin-like (ECL) cell in rebound acid hypersecretion.

Methods : Patients waiting for anti-reflux surgery who had used a proton pump inhibitor daily > 1 year were included. All patients discontinued taking acid inhibiting drugs after the operation. Basal and pentagastrin stimulated acid output was measured at 4, 8, 16 and 26 weeks postoperatively. Oxyntic mucosal biopsies were collected before and 26 weeks after the operation for counting of histidine decarboxylase (HDC) immunoreactive cells. Serum chromogranin A (CgA) and gastrin were measured before and at 4, 8, 16 and 26 weeks after the operation.

Results : Pentagastrin stimulated acid secretion was higher at 4 and 8 weeks than at 26 weeks after the operation. Gastrin and CgA were significantly reduced at 4 and 8 weeks, respectively. The number of HDC immunoreactive cells was reduced by 60% at 26 weeks postoperative.

Discussion : Rebound acid hypersecretion lasts more than 8 weeks, but less than 26 weeks after long-term proton pump inhibition.

Conclusion : The findings indicate that not only the parietal cell mass, but also ECL cell mass and activity are involved in the mechanism of acid hypersecretion.

Introduction

Rebound acid hypersecretion was first described in rats more than 15 years ago after treatment with omeprazole.1 In humans, we have previously reported that a 3-month treatment period with omeprazole in patients with reflux oesophagitis resulted in a marked rebound acid hypersecretion,2 a finding which has been confirmed in subsequent studies.3, 4

The mechanism of rebound acid hypersecretion has been discussed and it is agreed upon that the hypersecretion is related to the effects of hypergastrinemia caused by drug-induced hypoacidity.

To fully understand rebound acid hypersecretion hypotheses related to other phenomena previously observed during hypergastrinemia have been introduced. When evaluated, the time course of the known phenomena has been compared with the time course of rebound acid secretion. It has therefore been important to determine the duration of rebound acid hypersecretion to further understand its mechanism. Whereas the hypergastrinemia during administration of acid inhibiting drugs causes the oxyntic mucosal changes preceding the rebound hypersecretion, elevated gastrin levels rapidly return to normal after withdrawal of acid inhibiting drugs5, 6 and persistent hypergastrinemia can therefore not explain rebound acid hypersecretion.

Gastrin has a general trophic effect on the oxyntic mucosa,7–9 but has a more specific effect on the enterochromaffin-like (ECL) cell.10, 11 Increased parietal cell mass, as well as ECL mass, has been suggested to be involved in the mechanism of hypersecretion, but the relative contribution of the two cell types is not fully understood.

The present study aimed at determining the duration of the acid hypersecretion following long-term treatment with a proton pump inhibitor and to relate the rebound acid hypersecretion to parameters of an increased ECL cell mass.

Materials and methods

Patients

Patients with reflux oesophagitis, who had been treated with a proton pump inhibitor for a year or more and were accepted for anti-reflux surgery, were invited to participate. Three female and four male patients aged 31–66 years participated in the study. Two patients had been taking daily doses of 40 mg omeprazole for 3 years, one for 6 years and one for 1.5 years. One patients had taken 40 mg esomeprazole for 2 years, and one patient had taken 40 mg omeprazole for 1 year before taking 40 mg esomeprazole for another 1.5 years. The last patient had been taking 30 mg lanzoprazole for 5 years. All patients underwent a standardised laparoscopic Nissen–Rosetti anti-reflux procedure and the posterior vagal nerve trunk was identified and left undamaged during the dissection in the hiatal area. All patients were able to discontinue the proton pump inhibitor therapy following the operation. The study was approved by the Regional Ethics Committee in Trondheim.

Gastric acid secretion

Basal and maximal acid secretion was determined by a pentagastrin test, as described previously,2 at 09:00 h after a 12-h fast at 4, 8, 16 and 26 weeks after surgery. To avoid influence on the result of the operation we waited until 4 weeks postoperative before the first pentagastrin test.

Blood analyses

Blood samples were collected prior to the pentagastrin test and serum from these samples was stored at −20 °C before analyses. Serum gastrin12 and chromogranin A (CgA)13 concentrations were determined by methods described previously, before operation and at 4, 8, 16 and 26 weeks after the operation. Helicobacter pylori status was evaluated before surgery by measuring IgG against H. pylori in serum samples by the routine RIA method used at our hospital.

Biopsy collection and immunohistochemistry

Tissue samples from the gastric corpus for immunohistochemistry were collected during gastroscopy before and at 26 weeks after the operation. Gastroscopy was performed using a GIF IT20 gastroscope (Olympus, Tokyo, Japan) after an overnight fast and biopsy specimens were taken from mucosa of the corpus region, using a standard biopsy forceps (FB-13K E, Olympus). Biopsies were fixed in 4% phosphate-buffered formaldehyde for 24 h and dehydrated in 80% ethanol before paraffin embedding. Sections 4 μm thick were cut from paraffin blocks and transferred to positively charged slides. The sections were deparaffinised with xylene, rehydrated and treated with 3% hydrogen peroxide for 10 min to block endogenous peroxidase activity. Antigen retrieval was achieved by boiling the sections in a commercial microwave oven at 160 W for 15 min in Tris–EDTA buffer pH 9.0. The histidine decarboxylase (HDC) antiserum (Code B 260–1, Eurodiagnostica, Malmö, Sweden) was diluted 1:15 000 in phosphate buffered saline containing 0.25% Triton X-100 (Calbiochem, San Diego, CA, USA) and 0.25% bovine serum albumin (Sigma, St Louis, MO, USA) and the sections were incubated with primary antiserum for 1 h at room temperature. The EnVision-HRP kit (K5007, DAKO, Glostrup, Denmark) and DAB+ (K4065, DAKO, Glostrup, Denmark) was used to visualise the immunoreaction. Nucleated immunoreactive cells were counted at 400× magnification using a 1 × 1 mm grid with 100 squares placed in the ocular of a microscope. Immunoreactive cells were only counted in areas of the sections where both the luminal surface and the muscularis mucosae layer were visible and are presented as cells/mm2, making the number of counted cells independent of the sectioning angle.

Statistics

Results are presented as mean ± S.E.M. To evaluate differences between the different points of time, the Wilcoxon's paired, signed rank test was used. P < 0.05 was considered significant.

Results

Two patients did not meet for examination at 8 weeks after the operation. One patient was Helicobacter positive before the operation, but did not differ from the other patients as regards the other parameters measured.

Acid secretion

Pentagastrin stimulated acid secretion was unchanged from 4 to 8 weeks after the operation, but there was a significant reduction in secretion from 8 to 26 weeks after the operation (32 ± 6 mmol/h vs. 17 ± 2 mmol/h, P = 0.032) (Figure 1). However, the reduction from 16 to 26 weeks was not significant (24 ± 4 mmol/h vs. 17 ± 2 mmol/h, P = 0.30). Furthermore, the reduction seen in basal acid secretion from 4 to 26 weeks after operation was nearly significant (3.8 ± 0.9 mmol/h vs. 2.4 ± 0.5 mmol/h, P = 0.06) (Figure 2).

Figure 1.

Pentagastrin stimulated acid secretion (mmol/h), serum chromogranin A (CgA) (ng/mL) and serum gastrin (pm) in patients followed from before anti-reflux surgery to 26 weeks after the operation. The data are presented as mean ± S.E.M. The reduction in pentagastrin stimulated acid secretion from 8 to 26 weeks (P = 0.032), the reduction in s-CgA from before to 8 weeks after the operation (P = 0.02) and s-gastrin from before to 4 weeks after the operation (P = 0.032) were significant.

Figure 2.

Basal acid output (mean ± S.E.M.) in patients followed from before anti-reflux surgery to 26 weeks after the operation. The reduction from 4 to 26 weeks was nearly significant (P = 0.06).

Chromogranin A, gastrin and enterochromaffin-like cell density

Serum CgA concentrations were reduced after 4 weeks, and were significantly reduced after 8 weeks compared with preoperative values (53.1 ± 19 ng/mL vs. 14 ± 1.4 ng/mL, P = 0.02) and continued to fall during the observation period. The serum gastrin concentrations followed a similar pattern, with a significant reduction seen after 4 weeks (17 ± 2.5 pm vs. 6.4 ± 1.3 pm, P = 0.03).The number of HDC immunoreactive cells decreased in all patients from preoperative to 26 weeks postoperative samples (33 ± 8 cells/mm2 vs. 13 ± 1 cells/mm2, P = 0.031) (Figure 3).

Figure 3.

The density of histidine decarboxylase immunoreactive cells in biopsies from the oxyntic mucosa were significantly reduced at 26 weeks compared with preoperative samples (P = 0.032).

Discussion

Rebound acid hypersecretion has gained new attention the past year,5, 14–16which seems relevant as an increasing number of patients use proton pump inhibitors as symptomatic treatment for upper gastrointestinal disorders.

Although this study is limited by the low number of patients enrolled, we found rebound acid hypersecretion in patients who had used a proton pump inhibitor for more than a year before anti-reflux surgery. The hypersecretion seems to last more than 8 weeks, but less than 26 weeks. Serum gastrin and CgA concentrations were significantly reduced after 4 and 8 weeks, respectively. Furthermore, the number of ECL cells decreased in all patients from preoperative to 26 weeks after the operation. Whereas the preoperative serum gastrin is influenced by administration of a proton pump inhibitor 24 hours before blood sampling, the reduction in gastrin after 4 weeks is explained by the cessation of acid inhibition. However, it is likely that these patients have a high and prolonged meal stimulated hypergastrinaemia and that a fasting serum gastrin value underestimates the influence of gastrin.2

Previous studies have shown an increase also in basal acid secretion after proton pump inhibition,2, 4, 5 and the nearly significant (P = 0.06) decrease seen in this study most likely represents a true fall in acid secretion.

Gastrin has a general trophic effect on the oxyntic mucosa, but has specific trophic effect on the ECL cell.10, 11, 17 Hypergastrinemia will thus result in an increased parietal cell mass, but a more pronounced increase in the ECL cell mass. Even moderate hypergastrinemia has a trophic effect on the ECL cell.9 It has previously been shown that patients receiving long-term treatment with omeprazole have an increase in corpus argyrophil cell density from 83 to 149 cells/mm2.18 These numbers suggests an even more pronounced increase in ECL cell density, as ECL cells constitute only 35% of the neuroendocrine cells in the normal human corpus mucosa.19 In the present study we used HDC as an ECL cell marker and found a 60% reduction in ECL cell density 26 weeks after cessation of acid inhibition. The previous finding of increased mucosal histamine concentration in patients with rebound acid hypersecretion also indicates ECL cell hyperplasia.2

Although the parietal cell mass will influence the maximal gastrin stimulated acid secretion, there are also findings supporting the hypothesis that an increased ECL cell mass could explain the phenomenon of rebound acid hypersecretion. First, previous studies have shown that the ECL cell mass is critical for the magnitude of gastrin-induced histamine release in the rat20 and that gastrin does not release histamine in quantities great enough to give maximal histamine stimulation of acid secretion.21 Whereas the maximal histamine stimulated acid secretion is regarded as a functional equivalent to the parietal cell mass,22 both parietal cell mass and ECL cell mass seem to limit maximal gastrin stimulated acid secretion.21, 23 Thus, we have previously proposed that the increased ECL cell mass could be responsible for the increased gastric acid secretory capacity seen after treatment with a proton pump inhibitor.2 Secondly, relating parietal cell and ECL cell half-life to the duration of rebound acid hypersecretion could give further indications about their relative importance. Although both parietal cell24 and ECL cell25 kinetics have been studied in mice, the results are not easily compared as different protocols have been used. The half-life of gastric mucosa cells in humans is even more difficult to determine, but parietal cells possibly have a half-life from 1 up to several years.26 The fall in ECL cell density seen after 26 weeks in this study indicates that ECL cells have a shorter half-life than parietal cells. If rebound acid secretion is related to an increase in ECL cell mass, one would expect hypersecretion to persist as long as there is an increased ECL cell mass.

Measuring serum CgA has been suggested to be useful when evaluating ECL cell hyperplasia in patients using acid inhibitors,27–30, and serum gastrin is in these patients correlated to CgA concentrations. In rats dosed with omeprazole, the circulating level of pancreastatin, which is a spilt product of CgA, was found to reflect both the size of the ECL cell population as well as the activity of the ECL cells,31 and this is probably also true for serum CgA concentrations. The fall in gastrin and CgA seen after 4 and 8 weeks preceding the fall in pentagastrin stimulated acid secretion is probably related to a reduction in ECL cell activity. The rebound acid hypersecretion observed after short-term administration of H2-blockers32 lasts only a few days33 is most likely also related to a short-lasting increase in ECL cell activity. It has also been found that patients receiving omeprazole for 90 days have increased mucosal histamine concentration and reduced effect of a H2-blocker caused by increased histamine release from ECL cells.34 We thus find it likely that both an increase in ECL cell activity and ECL cell mass are involved in the mechanism of rebound acid hypersecretion seen after long-term proton pump inhibition.

Potential side effects of prolonged acid inhibition have been discussed since the 1980s,35, 36 and rebound acid hypersecretion is one of them. The clinical impact of rebound acid hypersecretion remains to be shown, but it could have negative influence on the disease or symptoms for which they are prescribed. Resistance towards H2-blockers after the use of omeprazole has already been mentioned above.34 This study shows that rebound acid hypersecretion after long-term proton pump inhibition lasts more than 8 weeks, but less than 26 weeks, and provides further indications that not only the parietal cell mass, but also the ECL cell volume and activity are involved in the mechanism of rebound acid hypersecretion.

Acknowledgements

We thank Britt Schulze, Bjørn Munkvold and Kari Slørdahl for their technical assistance. The study was supported by Hoels Legat, Norway.

Ancillary