Aliment Pharmacol Ther 2010; 32: 1023–1030
Background Acid reflux is often difficult to control medically.
Aim To assess the effect of 40 mg twice daily esomeprazole (high-dose) on gastric and oesophageal pH and symptoms, and biomarkers relevant to adenocarcinoma, in patients with Barrett’s oesophagus (BO).
Methods Eighteen patients, treated with proton pump inhibitors as prescribed by their treating doctor, had their therapy increased to high-dose esomeprazole for 6 months.
Results At entry into the study, 9/18 patients had excessive 24-h oesophageal acid exposure, and gastric pH remained <4 for >16 h in 8/18. With high-dose esomeprazole, excessive acid exposure occurred in 2/18 patients, and gastric pH <4 was decreased from 38% of overall recording time and 53% of the nocturnal period to 15% and 17%, respectively (P < 0.001). There was a reduction in self-assessed symptoms of heartburn (P = 0.0005) and regurgitation (P < 0.0001), and inflammation and proliferation in the Barrett’s mucosa. There was no significant change in p53, MGMT or COX-2 expression, or in aberrant DNA methylation.
Conclusions High-dose esomeprazole achieved higher levels of gastric acid suppression and control of oesophageal acid reflux and symptoms, with significant decreases in inflammation and epithelial proliferation. There was no reversal of aberrant DNA methylation.
In Barrett’s oesophagus (BO), which is commonly associated with gastro-oesophageal reflux disease, the normal oesophageal squamous epithelium is replaced by a columnar epithelium, which contains goblet cells. Approximately 0.5–1% of patients with BO will develop oesophageal adenocarcinoma each year, and BO patients have a 30- to 125-fold increase in the risk of oesophageal adenocarcinoma compared with the general population.1–3 Exposure to reflux contributes to symptoms and is an independent risk factor for BO and adenocarcinoma.4
Proton pump inhibitors (PPIs) are the mainstay of treatment of reflux disease. Esomeprazole at a dose of 40 mg once-daily provides a more effective control of gastric pH at steady state than once-daily omeprazole 20 mg, lansoprazole 30 mg, pantoprazole 40 mg or rabeprazole 20 mg.5 However, in patients with BO, oesophageal acid exposure is often difficult to control with commonly used dosages of PPI therapy, even though symptom control may be adequate.6, 7 In particular, control of oesophageal acid exposure in the second 12-h period after morning dosing is frequently insufficient in BO patients, and better pH control over the full 24 h may require a second dose.
Whilst one goal of treatment is the control of symptoms, another should be the prevention of progression to oesophageal adenocarcinoma. Nonrandomized studies suggest that acid suppression by PPIs may reduce the risk of developing dysplasia and therefore potentially reduce the risk of developing cancer.8–10 However, data on the protective benefit of medical therapy are not clear. Previous studies did not assess the effectiveness of pH control, and very large trials with long follow-up would be required to assess whether PPI therapy reduces the risk of adenocarcinoma.11 In this study, we used 40 mg twice-daily esomeprazole to control acid reflux, and measured molecular markers of inflammation and proliferation, and aberrant DNA methylation as surrogates for the effect of acid suppressant therapy on the potential for development of adenocarcinoma in patients with BO.
Patients and study design
Nineteen patients with histologically confirmed BO ≥3 cm in length, being treated with PPI therapy as prescribed by their treating doctor, agreed to enter the study. The PPIs and the doses at which they were prescribed are detailed in the Results, and, for the purposes of this study, are collectively described as standard PPI therapy. At entry into the study, whilst on their standard PPI therapy, patients had an endoscopy with collection of biopsies, oesophageal manometry, ambulatory 24-h gastric and oesophageal pH monitoring and completed a symptom score questionnaire. Then PPI therapy was increased to 40 mg twice-daily (high-dose) esomeprazole for 6 months. One patient withdrew within 3 days of starting high-dose esomeprazole because of nausea and was not included in the analysis. Whilst on high-dose esomeprazole, gastric and oesophageal pH monitoring and a symptom score questionnaire were repeated at 2 months, and endoscopy was repeated and oesophageal biopsies were taken at 6 months. The study was approved by the Research Ethics Committee of the Royal Adelaide Hospital and all patients gave their written informed consent.
Oesophageal manometry was performed at entry into the study. After an overnight fast, a 3.5 mm multi-lumen assembly that incorporated a sleeve sensor (Dentsleeve Pty, Ltd, Wayville, SA, Australia) was used to measure lower oesophageal sphincter (LES) pressure, gastric pressure, oesophageal body motility and swallowing. After intubation, the proximal margin of the LES was determined by performing a stepwise pull-through of the distal three oesophageal side-hole sensors. Following a 10- to 15-min period of accommodation, oesophageal responses to 10 water swallows were assessed. Basal LES pressure was measured at end-expiration relative to intra-gastric pressure as a 1-min visual mean over the last minute of the accommodation period. Oesophageal peristalsis was assessed for wave amplitude and peristaltic progression as described previously.12
Ambulatory 24-h pH monitoring
Ambulatory oesophageal and gastric pH monitoring were performed on PPI therapy at entry into the study and after 2 months of treatment with high-dose esomeprazole. Oesophageal and gastric pH were measured using a dual antimony catheter with pH electrodes spaced 15 cm apart (Medtronic Functional Diagnostics Zinetics Inc, Salt Lake City, UT, USA). The proximal electrode was positioned 5 cm above the upper border of the LES; accordingly, the distal electrode was positioned in the upper stomach. The electrodes were connected to a portable data-logger and ambulatory pH recordings were made for 24 h.
Ambulatory 24-h pH data were analysed using the esophagram software (Medtronic Functional Diagnostics Zinetics Inc). The time that oesophageal pH was <4 was determined for the total, upright and recumbent periods as well as for the nocturnal period, defined as 22:00–08:00 h.13 Median and mean gastric pH, as well as the duration of gastric pH <4, were determined for each hour as well as the nocturnal period.
Assessment of symptoms
Heartburn and regurgitation that had occurred over the previous 4 weeks were assessed by the Reflux Disease Questionnaire,14 which was completed by the patient without assistance, at entry into the study and after 2 months of high-dose esomeprazole. The frequency and severity of heartburn and regurgitation were assessed separately. The maximum score for heartburn or regurgitation was 20, with a maximum combined score of 40.
Biopsies were taken at endoscopy at entry into the study and after 6 months of high-dose esomeprazole according to a standard protocol: four quadrantic biopsies were taken at the gastro-oesophageal junction (GEJ), defined as at the tops of the gastric folds, and 4 each at 2-cm intervals in the columnar lined oesophagus, beginning in the oesophagus 1-cm above the GEJ. Biopsies were placed in neutral buffered formalin for histological assessment. An additional biopsy was taken from each level in the columnar lined oesophagus, and from macroscopically normal squamous mucosa 5 cm proximal to the maximal extent of the squamo-columnar junction, and preserved either in RNAlater (Ambion, Austin, TX, USA), or snap frozen in liquid nitrogen, and then stored at −80 °C until measurement of DNA methylation.
DNA methylation analysis
Biopsies from the oesophageal columnar and squamous mucosae, taken at entry into the study and after 6 months of high-dose esomeprazole, were used for DNA methylation analysis. The DNA was isolated and bisulfite modified as described previously.15 Bisulfite modified DNA was polymerase chain reaction amplified using primers and conditions described in Table S1. Methylation was quantified by calculating the T50, the temperature at which half the amplicons are melted, as described previously.16 The average T50 was calculated when multiple columnar biopsies were available.
Biopsies from the columnar lined oesophagus taken at entry into the study and after 6 months of high-dose esomeprazole were used for histological assessment. Formalin-fixed paraffin-embedded sections were stained with haematoxylin and eosin (H&E) for routine histology, or incubated with antibodies for Ki67, COX-2, MGMT or p53 for immunohistochemistry. Staining for Ki67 was assessed separately for the crypts and glands, and for the luminal surface epithelium, and scored as 0 (none), 1 (low) or 2 (high). Staining for COX-2 was assessed separately for the crypts and glands, and for the luminal surface epithelium, and scored as 0 (none), 1 (low) or 2 (high). Staining for MGMT was scored as 0 (normal), 1 (borderline reduced expression) or 2 (reduced expression with complete or clonal loss). Staining for p53 was scored as 0 (negative), 1 (borderline positive staining) or 2 (>10% cells with strong nuclear staining). Inflammation was assessed on the H&E sections and scored as 0 (none), 1 (mild chronic), 2 (more intense chronic) or 3 (substantial admixture of neutrophils). The aggregate score of the four biopsies immediately proximal to the GEJ was used in the analysis.
Normally distributed data were summarized using means and ranges and were compared between reflux groups using unpaired t-tests. Non-normally distributed continuous data were summarized using the median and lower and upper quartiles and were compared using either a Wilcoxon signed-rank test for paired data or a Mann–Whitney test for independent samples. The P-values were adjusted for multiple comparisons using the Sidak procedure. All statistics were considered significant when the two tailed P ≤ 0.05.
Eighteen patients (14 men) with a median age of 56 years (range 38–71) and BO ≥3 cm in length (median circumferential length 4 cm, range 3–8 cm) completed the study. At entry into the study, all patients had BO with histologically confirmed intestinal metaplasia, and none had erosive or ulcerative oesophagitis. Patients were taking a range of PPI therapies at entry into the study. Ten were on a once-daily morning dose [omeprazole 20 mg (n = 5), pantoprazole 40 mg (n = 3), lansoprazole 30 mg (n = 1), or esomeprazole 40 mg (n = 1)], and eight on a twice-daily dose (omeprazole 20 mg (n = 6), lansoprazole 30 mg (n = 2).
Oesophageal acid exposure
At entry into the study, whilst on their standard PPI therapy, the median 24-h oesophageal acid exposure was 4.5% (0.9–15.5%) (Figure 1). Nine of the 18 (50%) patients had excessive total 24-h acid exposure (pH <4 for >5% of the total time): one because of excessive upright exposure, five because of excessive recumbent exposure and four because of excessive upright and recumbent acid exposure. There were no significant differences in total, upright or recumbent acid exposure between patients on once- or twice-daily therapy, although all the four highest acid exposure patients were on once-daily therapy (Figure 1).
Whilst on high-dose esomeprazole, the median 24-h oesophageal acid exposure was only 0.2% (0–0.6%). Only two patients had excessive acid exposure, one as a result of excessive upright reflux and the other due to excessive upright and recumbent reflux. These two patients had the highest oesophageal acid exposure at entry into the study (Figure 1).
The gastric pH profiles in Figure 2 show the median pH each hour of the entire 24-h period. At entry into the study, gastric pH remained below 4 for a median of 38% (29–57%) of the 24-h period (Figure 3). The percentage of the time that gastric pH was below 4 did not differ between the upright [33% (26–48%)] and recumbent periods (45% (35–60%). There were no significant differences in any of these variables between patients on once- or twice-daily therapy.
On high-dose esomeprazole, the overall time that gastric pH remained below pH 4 fell to 15% (5–22%; P < 0.001). However, the percentage of the time that gastric pH was below 4 was significantly greater during the recumbent period [17.2% (9–33%)] than during the upright period [9% (2–17%)]. At entry into the study, gastric pH remained above pH 4 for more than 16 h in only eight of 18 patients, compared with 17 of 18 patients on high-dose esomeprazole (Fisher’s exact test: P = 0.0027).
Nocturnal gastric and oesophageal acidity
At entry into the study, the median duration that gastric pH was below 4 was 320 min (208–440 min) or 53% (35–73%) of the nocturnal recording period (Figure 4). There were no differences between patients on once- and twice-daily therapy. On high-dose esomeprazole, the median time that gastric pH was below 4 [100 min (40–171 min) or 17% (7–27%)] was significantly less than at entry into the study (P < 0.001).
At entry into the study, nocturnal oesophageal acid exposure was 5.5% (0.4–36.3%). Patients on once-daily therapy had similar levels to those on twice-daily therapy. Nocturnal oesophageal acid exposure was almost completely eliminated on high-dose esomeprazole (0%; 0–0.2%) (P < 0.004) (Figure 5).
At entry into the study, basal LES pressure was uniformly low (3.5 ± 0.8 mmHg). The rate of peristalsis was high and successful peristaltic responses occurred with 100% (90–100%) of water swallows. Maximum peristaltic pressure wave amplitudes were 54 ± 6.3 mmHg in the distal oesophagus.
At entry into the study, the median combined heartburn and regurgitation score was 12 (7.25–18) over the previous 4 weeks, with 17 of the 18 patients reporting symptoms. The median heartburn score was 5 (0–7.8), with 12 patients reporting symptoms. The regurgitation score was 8 (4.5–10), with 16 patients reporting symptoms. Following 2 months of treatment with high-dose esomeprazole, the median combined heartburn and regurgitation score was 2 (0–4.75) (P < 0.0001), with 11 of the 18 patients reporting symptoms. The median heartburn score was 0 (0–0) (P = 0.0005), with three patients reporting symptoms. The regurgitation score was 1.5 (0–4) (P < 0.0001), with 10 patients reporting symptoms.
Over the 6 months of high-dose treatment, there was a significant decrease in the expression of Ki67 in the crypts and glands [median 6 (5–7) vs. 5 (4–5.25); P = 0.0442] and in the surface epithelium [median 3 (2–4) vs. 1 (0–2); P = 0.0029]. There was also a significant decrease in inflammation score [median 8 (5.75–9.25) vs. 4 (3.75–5.25); P = 0.0016]. Detectable p53 or COX-2 or altered MGMT expression, when present, was low and did not change with high-dose treatment.
The T50, a measure of the amount of DNA methylation in a sample, was significantly greater in the biopsies from the columnar epithelium than in those from the squamous epithelium for all nine genes (Table S2). There was no significant change in methylation in the columnar lining following 6 months of high-dose treatment (Table S3). Significantly increased methylation of MGMT [median 75.69 (75.66–77.27) vs. 77.53 (77.12–77.76); P < 0.0001] was associated with reduced MGMT protein expression in columnar biopsies at the equivalent oesophageal level.
Patients with BO have increased oesophageal acid reflux that is often difficult to control with standard dose PPIs. If acid reflux plays a role in the progression from BO to cancer, then medical treatment must not simply result in symptomatic relief, but ensure effective control of oesophageal acid exposure as well. The aim of this study was to assess prospectively the effect of high-dose (40 mg twice daily) esomeprazole on acid suppression, symptoms and molecular markers in a heterogeneous group of 18 patients selected on the basis of having at least 3 cm of BO with confirmed intestinal metaplasia. Treatment with high-dose esomeprazole achieved high levels of gastric acid suppression and effective control of oesophageal acid reflux in 16 (89%) patients, with a significant improvement in symptoms. In the BO segment, while there was a significant reduction in inflammation and epithelial cell proliferation, there was no change in p53, COX-2 or MGMT protein expression, and, importantly, no reduction in aberrant DNA methylation of the nine genes assessed.
The management of BO is sometimes difficult because control of oesophageal acid exposure by PPIs is often less effective in patients with BO compared with patients with reflux oesophagitis.6, 13 Persistent acid reflux, particularly at night, is common 7 even when reflux symptoms are reported to be controlled.17, 18 Persistent oesophageal acid exposure would most likely undermine any potential benefit of PPI therapy in reducing the risk of dysplasia and adenocarcinoma. The reasons for the difficulty in suppressing acid reflux and secretion in patients with BO remain unclear. The underlying high levels of acid reflux may require greater levels of acid suppression. However, whether acid secretion is increased in BO is controversial.19, 20 On PPI therapy as used in routine clinical practice, whether once-daily or twice-daily dosing, we found that high levels of gastric acidity and excessive levels of oesophageal acid exposure were common among our patients.
Twice-daily 40 mg esomeprazole, however, was highly effective in controlling gastric acidity and oesophageal acid exposure. Gastric pH remained above pH 4 for 85% of the time overall, and 17/18 (94%) patients maintained a gastric pH above 4 for more than 16 h. Oesophageal acid exposure, which had been excessive in half of the patients on their initial doses of PPI, was reduced to normal levels in all but two patients.
Our findings following treatment with high-dose esomeprazole contrast with those of earlier studies which reported a significant minority of patients having persistent symptoms when treated with what was considered a maximal dose of PPIs (but lower than used in our study).7, 21 High-dose esomeprazole was also very effective in controlling heartburn, but had less impact on regurgitation. These findings are consistent with other recent reports and underline the need for twice daily PPI therapy to control acid secretion and acid reflux reliably in BO.22
While symptom relief is one major aim of treatment of reflux disease, in patients with BO, it would also be desirable that treatment should reduce the significant risk of progression to oesophageal adenocarcinoma. There are reports that high-dose PPIs may decrease the length of BO,23–26 but there is no evidence that they can completely reverse the condition.23, 27–30 Whilst there are claims that acid suppression with a PPI alone reduces the risk for development of dysplasia in patients with BO,8–10 these studies were uncontrolled and retrospective and information on the effectiveness of the control of oesophageal acid exposure was not available because pH monitoring was not included. In the current study the high levels of acid inhibition that were achieved by high-dose esomeprazole were documented by detailed gastric and oesophageal pH monitoring.
For reasons of the large cohort size and long follow-up period required to measure cancer as an end-point of treatment, we selected a range of molecular measures as surrogate markers for cancer risk. Treatment with high-dose esomeprazole was associated with a reduction in epithelial cell proliferation, as measured by proliferating cell nuclear antigen Ki67, in both the crypt and glands and the luminal surface cells. In normal tissues, cell proliferation staining is absent from the luminal cells, and it increases there as the tissues progress along the metaplasia-dysplasia-carcinoma sequence of oesophageal adenocarcinoma.31 Our findings are similar to those of other studies which have shown that the increased proliferation and expression of key cell cycle regulatory genes, which occur early in the neoplastic progression associated with BO, are reduced by treatment with PPIs.32–34
Chronic inflammation is thought to be causally associated with the progression from premalignant disease to cancer, including BO and oesophageal adenocarcinoma.35 We noted a significant decrease in the inflammation score over the 6 months of high-dose treatment. One measure of inflammation which has been associated with malignancy, COX-2 expression, remained unchanged at low or undetectable levels over the course of the study. In other studies of the effects of PPI treatment, COX-2 has been variably reported to increase32 or not change36 following acid suppression. It is possible that the reduction in the inflammation score might have resulted from anti-inflammatory effects which may be exerted by PPIs independent of acid inhibition.37 Our results suggest a dose response to PPIs, as inflammation was reduced with the high-dose esomeprazole compared with the doses of PPIs that patients were on before entering the study. We also found absence or low levels of p53 expression and reduced MGMT expression in several patients at the start of the study. No significant changes were measured following treatment, reflecting the fact that most of the participants had normal expression of these markers on entry into the study.
The progression to cancer in a tissue requires the up- or down-regulation of critical genes. Aberrant methylation of CpG islands in the promoter regions of genes, particularly tumour suppressor genes, can lead to gene silencing and is common in cancer, including oesophageal adenocarcinoma.38, 39 We have previously reported that methylation of seven of the genes analysed in this study is as frequent in BO as in EAC.15 Despite the strong link between this and the development of cancer, little is known about changes in aberrant methylation in BO as a result of treatment of reflux. We did not find any significant change after 6 months treatment with high-dose esomeprazole. Methylation and demethylation most commonly occur at cell division. As columnar epithelial cells have a short cell cycle time, even in treated patients, we would expect there to have been more than enough cell divisions to have seen a change in DNA methylation profile over the 6-month period if control of acid would result in its reversion back to normal.
In this study, we have shown that high-dose esomeprazole significantly reduces gastric and oesophageal acid levels, and reflux-associated symptoms, in patients with ≥3 cm of BO. An important finding is that despite this convincing benefit in the control of acid and symptoms, as well as reduced inflammation and epithelial cell proliferation, there was no reversal of aberrant DNA methylation, a biomarker believed to be causally associated with development of cancer. These findings suggest that while high-dose esomeprazole offers an appropriate treatment for difficult-to-control reflux symptoms, it may not prevent the development of Barrett’s-associated oesophageal adenocarcinoma in patients with BO.
Declaration of personal interests: Richard Holloway has served as a speaker and consultant for AstraZeneca and Janssen-Cilag, and has received research funding from Nycomed. John Dent has received consulting fees for advisory committees/review panels, consulting, speaking and teaching, and grant/research support for basic science research, from AstraZeneca. Robert Fraser has served as consultant or speaker for GlaxoSmithKline, Nycomed and Johnson and Johnson, and has received research funding from GlaxoSmithKline. Declaration of funding interests: This study was funded in part by a Grant-in-Aid from AstraZeneca.