- Email: [email protected]
Acid suppression increases rates of Barrett’s esophagus and esophageal injury in the presence of duodenal reflux
Acid suppression increases rates of Barrett’s esophagus and esophageal injury in the presence of duodenal reflux Ayman O. Nasr, MB,a Mary F. Dillon, MD,a Susie Conlon, MB,b Paul Downey, MB,b Gang Chen, PhD,c Adrian Ireland, MD,c Eamon Leen, MB,b David Bouchier-Hayes, MCh,c,d and Thomas N. Walsh, MD,a Dublin, Ireland
Background. The contribution of gastric acid to the toxicity of alkaline duodenal refluxate on the esophageal mucosa is unclear. This study compared the effect of duodenal refluxate when acid was present, decreased by proton pump inhibitors (PPI), or absent. Methods. We randomized 136 Sprague-Dawley rats into 4 groups: group 1 (n = 33) were controls; group 2 (n = 34) underwent esophagoduodenostomy promoting ‘‘combined reflux’’; group 3 (n = 34) underwent esophagoduodenostomy and PPI treatment to decrease acid reflux; and group 4, the ‘gastrectomy’ group (n = 35) underwent esophagoduodenostomy and total gastrectomy to eliminate acid in the refluxate. Esophaguses were examined for inflammatory, Barrett’s, and other histologic changes, and expression of proliferative markers Ki-67, proliferating cell nuclear antigen (PCNA), and epidermal growth factor receptor (EGFR). Results. In all reflux groups, the incidence of Barrett’s mucosa was greater when acid was suppressed (group C, 62%; group D, 71%) than when not suppressed (group B, 27%; P = 0.004 and P < .001). Erosions were more frequent in the PPI and gastrectomy groups than in the combined reflux group. Edema (wet weight) and ulceration was more frequent in the gastrectomy than in the combined reflux group. Acute inflammatory changes were infrequent in the PPI group (8%) compared with the combined reflux (94%) or gastrectomy (100%) groups, but chronic inflammation persisted in 100% of the PPI group. EGFR levels were greater in the PPI compared with the combined reflux group (P = .04). Ki-67, PCNA, and combined marker scores were greater in the gastrectomy compared with the combined reflux group (P = .006, P = .14, and P < .001). Conclusion. Gastric acid suppression in the presence of duodenal refluxate caused increased rates of inflammatory changes, intestinal metaplasia, and molecular proliferative activity. PPIs suppressed acute inflammatory changes only, whereas chronic inflammatory changes persisted. (Surgery 2012;151:382-90.) From the Department of Surgery,a Royal College of Surgeons and the Department of Histopathology,b Connolly Hospital, Blanchardstown; and the Biomedical Research Facilityc and the Department of Surgery,d Royal College of Surgeons, Beaumont Hospital, Dublin, Ireland
FEW TOPICS IN MEDICINE excite greater controversy than the role of duodenal gastroesophageal reflux (DGER) in the pathophysiology of Barrett’s esophagus and esophageal cancer. The question of how duodenal refluxate interacts with gastric acid and Partially funded by the Oesophageal Cancer Fund. Accepted for publication August 18, 2011. Reprint requests: Thomas N. Walsh, MD, Department of Surgery, Royal College of Surgeons in Ireland, Connolly Hospital, Blanchardstown, Dublin, 15, Ireland. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2012 Mosby, Inc. All rights reserved. doi:10.1016/j.surg.2011.08.021
382 SURGERY
whether the interaction is detrimental or protective has proven difficult to resolve. It is undisputed that acid reflux causes esophageal mucosal damage, but it seems that acid alone cannot account for the metaplasia–dysplasia–carcinoma sequence. If this concept was correct, the widespread prescribing of proton pump inhibitors (PPI) should have resulted in decrease reduction in the incidence of esophageal adenocarcinoma. The converse is true, however, with a steeply increasing incidence of esophageal adenocarcinoma observed since the 1970s,1-3 Similarly, studies on the ability of PPIs to control or reverse Barrett’s esophagus have been disappointing.4-7
Surgery Volume 151, Number 3
Conceptually, it makes more sense that intestinal metaplasia is a cytologic response to a stimulus other than gastric acid. As Dixon et al8 observed, gastric metaplasia would be an appropriate response by esophageal epithelium to prolonged acid attack. The histologic hallmark of Barrett’s esophagus, however, is intestinal rather than gastric metaplasia, which is more likely to be a protective response against bile and duodenal juices. Numerous studies demonstrate a greater prevalence of Barrett’s esophagus when DGER is present consistently in the esophagus,9-15 but how this metaplasia relates to the adverse effects of acid on esophageal epithelium is less clear. Some believe that acid and DGER act syngerstically.16,17 Others go further, and suggest that acid has a calming effect on the noxious effects of DGER.18 Bile is most active at an alkaline or normal pH in its conjugated, soluble, ionized form, whereas bile in an acid environment precipitates and is much more inactive. This observation leads to the emergence of a more disturbing hypothesis: That acid suppression treatment may increase the risk of esophageal carcinoma. This concept is not without precedent, because gastric cancer has been linked with hypochlorhydria.19,20 Against this concept, other studies demonstrate that PPIs decrease gastroesophageal reflux disease (GERD),21,22 and possibly DGER, by decreasing the overall volume of gastric juices, although volume of DGER may be less relevant than pH and relative concentrations of the constituents of the esophageal environment. The purpose of this animal study was to observe the effect of acid and acid suppression on DGERmediated changes in the esophageal epithelium at the histologic and molecular levels. Incorporated into the model were groups that underwent acid suppression medically by means of PPI treatment and those that achieved complete acid suppression by means of total gastrectomy. METHODS The study was conducted in the Biomedical Research Facility, Beaumont Hospital, Dublin, Ireland. Ethics approval was obtained from the research ethics committee of The Royal College of Surgeons in Ireland. One author (AN) completed detailed animal handling and microsurgery training prior to commencement of the study. The animals used were Sprague Dawley (SD) outbred rats (Sprague Dawley Sprague Dawley, Harlan UK Ltd. Registered trademarks of Harlan Sprague Dawley, Inc, Indianapolis, IN, and Harlan Sprague Dawley, Inc, United Kingdom). Rats were
Nasr et al 383
maintained under a standard 12-hour dark–light cycle in a warm, balanced climate at a temperature of 22–258C and humidity of about 60%. They were all fed standard chow and water ad libitum. All animal handling was carried out under strict animal license guidelines provided and supervised by a licensed full-time animal handling supervisors. Study groups. The study population consisted of 145 male Sprague-Dawley rats were randomized into 4 study groups. The rats were between 8 and 10 weeks old and weighed 250–300 g. The study was powered on the basis of an expected frequency of columnar metaplastic changes of 33% in the omeprazole group, and it was calculated that 34 rats were required for each group. Provision was made for a presumed operative mortality of 10% based on the literature.18,23 In this study, 9 rats did not complete the study protocol owing primarily to perioperative complications. The 136 rats which completed the study were randomized into 4 groups: Group 1 (n = 33), the control group, were not subject to operation or pharmacologic acid suppression (Fig 1, A). Group 2 (n = 34), the combined reflux group, underwent esophagoduodenostomy to produce combined gastric (acidic) and duodenal (alkaline) reflux (Fig 1, B). Group 3 (n = 34), the PPI group, were ‘alkaline refluxers’ that underwent esophagoduodenostomy. Medical acid suppression with omeprazole therapy promoted isolated duodenal reflux (Fig 1, C). Group 4 (n = 35), the gastrectomy group,underwent esophagoduodenostomy and total gastrectomy to ensure isolated duodenal reflux and were ‘total alkaline refluxers’ with complete acid suppression (Fig 1, D).
Operation. All operations were performed under gaseous general anesthetic using isoflurane 5% and O2. The DGER model was created by performing an esophagoduodenostomy with preservation of the stomach performed by dividing the esophagus 1 cm proximal to the cardia and performing a 4 mm, longitudinal duodenotomy incision 1 cm distal to the pylorus. The esophagus was anastomosed to the duodenum using 6/0 polyprolene and an end-to-side anastomosis. The proximal end of the stomach was closed using a nylon ligature. For the DGER model, total gastrectomy was performed by dividing the stomach 1 cm proximal to the cardia and dividing the duodenum about 1 cm distal to the pylorus. An esophagoduodenostomy anastomosis was then performed making a 4 mm, longitudinal duodenotomy incision 1 cm distal to the pylorus, and the esophagus anastomosed end-to-side to the duodenum with 6/0
384 Nasr et al
Surgery March 2012
Fig 2. Gross specimen of resected esophagus.
Fig 1. A, Control group (Y acid reflux, Y bile reflux). B, Combined reflux group ([[ acid reflux, [[ bile reflux). C, PPI-treated group (Y acid reflux, [[ bile reflux). D, Gastrectomy group (YYacid reflux, [[ bile reflux).
polyprolene. The duodenal stump was closed a using nylon ligature. Management and measurements. Omeprazole, the PPI used in this study, was administered to the PPI group by means of a subcutaneous injection (13.5 mg/kg) on alternate days. All other groups received equivalent volume of 0.9% NaCl injections. After 26 weeks, animals were killed using an overdose of isoflurane gas. Immediately after death, the esophagus was resected from just below the larynx to the esophagogastric junction in the control group (including 2 mm of stomach wall) and to the esophagoduodenal anastomosis in the operated groups to include 2 mm of the duodenum wall (Fig 2). The esophagus was then opened longitudinally from the proximal end to the distal end through the anastomosis/esophagogastric junction. The wet/dry weight of the esophagus was obtained as a measure of tissue edema, reflecting
inflammatory activity due to mucosal injury. One longitudinal third of the resected esophagus was weighed to obtain the wet weight, and was then heated to 1718C for 72 hours in a dry heat oven to obtain the dry weight. The remaining esophagus was preserved in 10% neutral, buffered, formalin solution as the main specimen for histology and immunohistochemistry processing. Pathologic assessment. The specimens were paraffin fixed, sectioned at 4-mm intervals, and stained with hematoxylin and eosin. The slides were reviewed by a consultant histopathologist in a blinded fashion. Histologic grading was reported with reference to similar animal experiment in the literature.24 Acute inflammation was considered present whenever there was a predominance of neutrophils, whereas chronic inflammation was reported when there was a predominance of lymphocytes, plasma cells, or eosinophils. Erosion was defined as a histologic defect in the surface epithelium (mucosa) and ulceration was defined as histologic loss of surface epithelium with extension into the submucosa. Columnar metaplasia was defined as the presence of histologically abnormal glandular mucosa proximal to the anastomotic line or the esophagojejunal junction. Barrett’s metaplasia was defined as the presence of intestinal mucosa with goblet cells proximal to the anastomotic line or the esophagojejunal junction. Mucinous adenomatous changes were defined as the presence of pleomorphism, nuclear enlargement, hyperchromasia, abnormal chromatin pattern, and a high mitotic rate with predominance of mucinous locules. The histologic grading of hyperplasia and hyperkeratosis was defined with reference to previous animal experiment in literature.18 A specimen was defined as having epithelial hyperplasia when there was thickening of the epithelium with papillary elongation or basal cell hyperplasia compared with controls. Basal cell hyperplasia was defined as thickening of the basal cell layer to >10–15% of the epithelial thickness. Hyperkeratosis was defined as thickening of the keratin layer to >10–15% of the epithelial thickness compared with controls.
Surgery Volume 151, Number 3
Sections were dewaxed, rehydrated, and washed with phosphate-buffered saline. Endogenous peroxidase activity was neutralized by immersing the tissue in 0.5% hydrogen peroxide for 10 minutes. Antigen retrieval process was performed by boiling 0.01 mol/L citrate retrieval solution (pH 6.0) 1:100–1:200 in a stainless steel pressure cooker in which slides were immersed for 1 minute, and then cooled. The sections were incubated with primary antibodies as follows: Anti–Ki-67, proliferating cell nuclear antigen (PCNA), and EGFR mouse monoclonal antibodies (Novocastra Laboratories Ltd, Newcastle, United Kingdom), were used at 1:200 dilutions and incubated for 60 minutes at 258C. The sections were incubated subsequently with avidin/biotin complex-horseradish peroxidase for 60 minutes at 258C, after which they were developed in diaminobenzidine tetrahydrochloride for 60 minutes at 258C and counterstained with hematoxylin. The degree of immunohistologic staining of EGFR, Ki-67, and PCNA in the esophagus was determined using recognized scoring systems. Only squamous areas were scored under the microscope. The score given is based on the greatest area of staining present above the basal layer and must be >5% of each high power field. EGFR was scored according to a previously published scoring system25 (0 = Membranous staining in <10% of cells; 1 = partial membranous staining in >10% of cells; 2 = strong, complete membranous staining in >10% of cells). Ki-67 and PCNA staining was also scored using a recognized scoring system25 (0 = no score; 1 = increased staining in <50% of epithelial thickness; 2 = increased staining in >50% of epithelial thickness). A score of $1 was taken as positive. Statistical analysis. Sample size was calculated using PS: Power and Sample Size Calculation software (Version 2.1.31, 2003, Department of Biostatistics; Vanderbilt University School of Medicine, Nashville, TN). Using a power of 0.8 and an expected frequency of columnar metaplastic changes of 33% in the omeprazole group, 34 rats were required in each group. Further statistical analysis was performed using SPSS 17.0 software for Windows (SPSS Inc., Chicago, IL). Chi-square tests, Fisher exact tests, and the Mann–Whitney U test were used where appropriate. RESULTS Level of inflammation. The mean percentage of the length of the inflamed segment to the full length of esophagus of the control group was 0.05%. The percentage of esophagus that was
Nasr et al 385
inflamed for the combined reflux, PPI and gastrectomy groups was 78%, 82%, and 80%, respectively. Esophageal weight. The median weight of the esophagus for the control group was 443 mg. In comparison, all other groups demonstrated a significantly increased esophageal weight. (The median weight in the combined reflux, PPI, and gastrectomy groups was 1445, 1264, and 1707 mg [P < .001 each], respectively.) There was a difference in the median weight of the combined reflux group when compared with the gastrectomy group (P = .039) but not in the combined reflux group when compared with the PPI group (P = .21). The median wet and dry weights of (1/3 sections) of the esophaguses in each group are shown in Table I. The median wet/dry weight ratio of the esophagus for the control, combined reflux, PPI, and gastrectomy groups were 1.70, 1.68, 1.81, and 1.97. Only the gastrectomy group differed from the control group when comparing the wet/dry weight ratio (P = .001). The median wet/dry weight of the control group vs the combined reflux and the PPI groups was not different (P $ .17 each). Histology. Inflammation: Table II summarizes the microscopic findings of the study. There were no acute inflammatory changes noted in the control group. Ninety-four percent (n = 32/34) of the combined reflux group and 100% (n = 35/35) of the gastrectomy group developed acute inflammation. In contrast, only 8% (n = 3/34) of the PPI group developed acute inflammatory changes. Chronic inflammatory changes were evident in 9% of the control, 97% of the combined reflux, and 100% of the gastrectomy groups. Although few esophaguses in the PPI group demonstrated acute inflammation, all (100%) demonstrated chronic inflammation. Ulcerations/erosions: Microscopic ulceration was absent in the control group but was identified at high levels in all other groups (Table II). The presence of ulceration in the PPI and gastrectomy groups were greater than the combined reflux group (P # .042 each). Similarly, erosions were greater in combined reflux, PPI, and gastrectomy groups, and greater in the gastrectomy group compared with the combined reflux group (P = .005) but not when the PPI group was compared with the combined reflux group (P = .17). Barrett’s esophagus: Barrett’s metaplasia was not identified in the control group but was present in 27% of the combined reflux, 62% of the PPI, and 71% of the gastrectomy groups (Fig 3; Table II). There was a greater degree of Barrett’s in the PPI
386 Nasr et al
Surgery March 2012
Table I. Median esophageal weights of the various groups Weight (mg)
Control
Combined refluxers
PPI
Gastrectomy
Median total esophageal Median dry* Median wet*
443
1445
1264
1707
171 292
265 437
255 430
284 550
*Dry and wet weights are determined from a longitudinal section of 1/3 the original esophageal specimen. PPI, Proton pump inhibitors.
group compared with the combined reflux group (P = .004) and in the gastrectomy group compared with the combined reflux group (P < .001). Mucinous adenomatous changes: Mucinous adenomatous changes were evident in combined reflux, PPI, and gastrectomy groups and greatest in the gastrectomy group (26%). Epithelial and basal cell hyperplasia and epithelial hyperkeratosis: All patients in the combined reflux, PPI, and gastrectomy groups exhibited epithelial cell hyperplasia compared with only 36% of the control group (the hyperplasia in the control group was graded as mild). Moderate-to-severe hyperplasia was present in 0%, 88%, 88%, and 100% of the control, combined reflux, PPI, and gastrectomy groups (Table II). There was no basal cell hyperplasia in the control group. Levels of moderate to severe basal cell hyperplasia in the combined reflux, PPI, and gastrectomy groups were 79%, 76%, and 83%, respectively. Mild epithelial hyperkeratosis was present in 39% of the control group, whereas moderate-to-severe epithelial hyperkeratosis was present in 3%, 97%, 100%, and 94% of, respectively, the control, combined reflux, PPI, and gastrectomy groups. Immunohistochemical analysis of expression of EGFR, Ki-67, and PCNA: EGFR expression in the control, combined reflux, PPI, and gastrectomy groups was 6% (2/33), 68% (23/34), 88% (30/34), and 83% (29/35; Fig 4), respectively. The differences between the combined reflux and the PPI groups were significant (P = .041) but not between the combined reflux and the gastrectomy groups (P = .14). Ki-67 expression in the control, combined reflux, PPI, and gastrectomy groups was 3% (1/33), 73% (25/34), 76% (26/34), and 97% (34/35). Higher expression of Ki-67 in the gastrectomy group was greater compared with the combined reflux group (P = .006). PCNA expression in the control, combined reflux, PPI, and gastrectomy groups was 0% (0/33), 67% (23/34), 56% (19/34), and 91% (32/35), respectively. Again, differences between
the combined reflux and the gastrectomy groups were significant (P = .014). Combined expression of proliferation markers: Combined expression in the control, combined reflux, PPI, and gastrectomy groups was 0% (0/33), 32% (10/34), 44% (15/34), and 77% (27/35), respectively. The differences between the combined reflux and the gastrectomy groups were significant (P < .001), but not between the combined reflux and the PPI groups (P = .21). DISCUSSION This study suggests that pure duodenal reflux is a powerful stimulus to the development of Barrett’s metaplasia. More important, acid suppression by medical or operative means worsened the cellular changes caused by DGER. The design of the study was powered specifically to examine the development of Barrett’s esophagus in response to DGER combined with PPI therapy. We demonstrated that although the combination of DGER and acid reflux resulted in a 27% rate of Barrett’s esophagus, a much greater percentage (62%) of rats developed Barrett’s when subjected to DGER combined with PPI therapy. Those rats were subjected to pure DGER and extreme acid suppression (by means of radical gastrectomy) developed the greatest rates of Barrett’s metaplasia (71%). In addition, further histologic and immunohistochemical analysis in this study showed consistently the heightened noxious effect of DGER in the setting of minimal or no acid. In addition to increasing the rate of Barrett’s metaplasia, decreases in acid in the duodenalesophageal refluxate also increased markers of inflammation, including erosions, ulcerations, and edema, as well as provoking an overall increase in measured markers of proliferation. Of particular interest was the finding that while the PPI group demonstrated very low levels of acute inflammation when compared with the other groups, this group maintained a very high level of chronic inflammatory changes. It is possible that PPIs may result in amelioration of acute inflammatory changes and even acute symptomatic GERD, whereas chronic, more insidious changes persist. The ability of PPI therapy to decrease the incidence of acute inflammatory changes when compared with the gastrectomy group may be attributed theoretically to its ability to decrease the volume of reflux or to other intrinsic anti-inflammatory effects of the PPI therapy. However, it is chronic inflammation that is thought to be the main stimulus or risk factor for $18% of cancers.26 Certainly, chronically sustained activation of the inflammatory response is thought
Nasr et al 387
Surgery Volume 151, Number 3
Table II. Summary of histology results Control (n = 33), n (%) No inflammation Acute inflammation Microscopic erosion Microscopic ulceration Chronic inflammation Epithelial hyperplasia Basal cell hyperplasia Barrett’s metaplasia Mucinous adenomatous changes
30 0 0 0 3 21 0 0 0
(91) (0) (0) (0) (9) (64) (0) (0) (0)
Combined (n = 34), n (%) 1 32 27 18 33 All All 9 5
(3) (94) (79) (53) (97) (100) (100) (27) (15)
PPI (n = 34), n (%) 0 3 31 26 All All All 21 5
Gastrectomy (n = 35), n (%)
(0) (9%) (91) (77) (100) (100) (100) (62) (15)
0 All All 32 All All All 25 9
(0) (100) (100) (91) (100) (100) (100) (71) (26)
PPI, Proton pump inhibitors.
80%
120
71%
70%
100
62% % expression
60% 50% 40% 26%
30% 20%
80 60 40 20
10% 0%
0
0% Control
Combined
PPI
Gastrectomy
Fig 3. Distribution of Barrett’s metaplasia among the different groups.
Control
Combined Refluxers
PPI
gastrectomy
EGF-R Ki-67 PCNA All 3 Markers
to produce a protumorigenic microenvironment needed to generate the culmination of changes needed for cell transformation whether it be metaplastic or carcinogenic.27 The alternative metaplastic theory, whereby inflammatory cytokines attract and transform stem cells to become metaplastic,27,28 is also more likely to be supported by a chronically inflamed microenvironment. A wide range of molecular markers have been used to examine proliferative changes in the esophagus.24,26,29 In this study, 3 well-established markers were used to estimate proliferative activity and pathologic progression. EGFR expression was greater in the PPI group than in the combined DGER/acid group. Levels of Ki-67 and PNCA were greater in the total acid suppression (gastrectomy) group compared with the combined DGER/acid group. Ki-67 is a nuclear antigen expressed in proliferating cells, but not in resting cells. Expression of Ki- 67 increases as cells progress along the normal–Barrett’s esophagus–carcinoma sequence.30-34 PCNA is an essential protein for cell proliferation. Like Ki-67, PCNA has been implicated in the transformation of normal squamous epithelial cells to esophageal adenocarcinoma,30,35 with progressively greater
Fig 4. Proportion of rats expressing various proliferation markers.
levels expressed when normal cells transform into Barrett’s esophagus and adenocarcinoma.35,36 Gillen et al37 found that PNCA expression in Barrett’s mucosa located adjacent to adenocarcinoma was greater than in normal Barrett’s mucosa. The third marker, EGFR, has been associated with both squamous cell carcinoma and adenocarcinoma of the esophagus.29,38 EGFR is important, because EGFR inhibitors have been tested clinically in other carcinomas and found to have a therapeutic effect.38-40 Current clinical trials of antibody inhibition of EGFR with cetuximab are under way in esophageal carcinoma40 and there is evidence that EGFR may have prognostic significance. Furthermore, some evidence suggests that Barrett’s change may occur via bile-stimulated cell signaling through the EGFR.41 Taken together, the results of this study further support the histologic findings that DGER combined with acid suppression increases cellular proliferative activity and possibly pathologic progression in esophageal cells.
388 Nasr et al
Our study examined the effects of both PPIinduced acid suppression and total acid suppression on DGER in an animal model. Only 2 other models in the literature studied the role of PPI therapy in esophageal carcinogenesis. Moore et al42 compared a PPI group and a control group, in which they anastomosed the duodenum to the esophagus. They found no difference between the groups. A recent study by Miyashita et al43 used a different model, anastomosing the jejunum to the esophagus and disconnecting the stomach from the esophagus. This technique resulted in both groups having a very alkaline pH in the lower esophagus (pH of 6.5 in both groups). In comparison, the esophageal pH in the present study ranged from 5.4 to 8.4 across the various models. Miyashita et al’s study demonstrated a decrease in esophageal cancer in rats treated with rabeprazole when compared with the control group; the authors hypothesized that rabeprazole, in addition to its acid suppressive effects, may have important anti-inflammatory effects, which in turn may have an anticarcinogenic effect. Other studies have examined manipulation of acid/DGER without the use of PPI. Ireland et al18 treated rats with a nitrosamine-based carcinogen and subjected the rats to various operations to induce varying degrees of acid and DGER. The absence of gastric juice increased the rates of DGER/carcinogeninduced esophageal adenocarcinoma 3-fold. Csendes et al44,45 demonstrated that the incidence of Barrett’s was greater in patients with increased DGER combined with decreased acid secretion as a result of a Billroth II, partial gastrectomy, and selective vagotomy than in patients with Roux-en-Y partial gastrectomy and selective vagotomy. The premise that decreasing acid reflux has a detrimental effect on DGER-mediated esophageal injury has some support from population-based studies.46-48 First, in these studies46-50 medications that decrease acid production did not have a protective effect against esophageal adenocarcinoma. Conversely, others have found an increased risk of esophageal carcinoma with the use of these medications,46-48 although this is not a consistent finding.26 An important confounder in all of these trials is that a proportion of patients on acid decreasing medication are been treated for GERD and are, therefore, potentially at greater risk of esophageal carcinoma. For instance, Garcıa Rodrıguez et al49 demonstrated a 5-fold increase in esophageal cancer when PPIs were prescribed for GERD but not when prescribed for other reasons. Arguably, their observation does not negate the premise that the toxic effects of DGER may be potentiated by PPIs. In contrast,
Surgery March 2012 Duan et al46 found that patients were at an increased risk of esophageal adenocarcinoma if they were taking nonprescription acid-neutralizing tablets without GERD symptoms, and Farrow et al47 reported an increased risk of esophageal adenocarcinoma with over the counter antacids. In a Swedish study, Lagergren et al48 demonstrated an increase in esophageal adenocarcinoma in patients with reflux symptoms who were taking medication for GERD compared with those who took no medication. In conclusion, although our study demonstrated that acid suppression causes detrimental effects in the presence of DGER, it is likely that the processes of progression to Barrett’s and carcinoma have complex causes. Both acid reflux and DGER are, on multivariate analysis, separately and independently associated with the development of Barrett’s esophagus.50 There may, therefore, be more than one potential mechanism of damage to esophageal cells that initiates cell transformation. It is likely that there is a complex interplay between acid and DGER that is affected by the timing and relative concentrations of acid and DGER reflux.51 Alternatively, it may be the case that acid suppression is only detrimental when DGER is present. Nevertheless, the results of this and other studies suggest that in the presence of DGER, acid suppression by medical or other means increases the rate of Barrett’s and potentially increases the risk of esophageal adenocarcinoma. REFERENCES 1. Blot WJ, Devesa SS, Kneller RW, et al. Rising incidence of adenocarcinoma of the esophagus and gastric cardia. JAMA 1991;265:1287-9. 2. Devesa SS, Blot WJ, Fraumeni JF Jr. Changing patterns in the incidence of esophageal and gastric carcinoma in the United States. Cancer 1998;83:2049-53. 3. Brown LM, Devesa SS, Chow WH. Incidence of adenocarcinoma of the esophagus among white Americans by sex, stage, and age. Natl Cancer Inst 2008;100:1184-7. 4. Peters FT, Ganesh S, Kuipers EJ, et al. Endoscopic regression of Barrett’s oesophagus during omeprazole treatment; a randomised double blind study. Gut 1999:45489-94. 5. Faybush EM, Sampliner RE. Randomized trials in the treatment of Barrett’s esophagus. Dis Esophagus 2005;18:291-7. 6. Parrilla P, Martınez de Haro LF, Ortiz A, et al. Long-term results of a randomized prospective study comparing medical and surgical treatment of Barrett’s esophagus. Ann Surg 2003;237:291-8. 7. Rees JR, Lao-Sirieix P, Wong A, et al. Treatment for Barrett’s oesophagus. Cochrane Database Syst Rev 2010;1:CD004060. 8. Dixon MF, Neville PM, Mapstone NP, et al. Bile reflux gastritis and Barrett’s oesophagus: further evidence of a role for duodenogastro-oesphageal reflux? Gut 2001;49: 359-63. 9. Fein M, Maroske J, Fuchs KH. Importance of duodenogastric reflux in gastro-oesophageal reflux disease. Br J Surg 2006;93:1475-82.
Surgery Volume 151, Number 3
10. Koek GH, Sifrim D, Lerut T, et al. Multivariate analysis of the association of acid and duodeno-gastro-oesophageal reflux exposure with the presence of oesophagitis, the severity of oesophagitis and Barrett’s oesophagus. Gut 2008;57: 1056-64. 11. Caldwell MT, Lawlor P, Byrne PJ, et al. Ambulatory oesophageal bile reflux monitoring in Barrett’s oesophagus. Br J Surg 1995;82:657-60. 12. Stein HJ, Kauer WK, Feussner H, et al. Bile reflux in benign and malignant Barrett’s esophagus: effect of medical acid suppression and Nissen fundoplication. J Gastrointest Surg 1998;2:333-41. 13. Marshall RE, Anggiansah A, Owen WJ. Bile in the oesophagus: clinical relevance and ambulatory detection. Br J Surg 1997;84:21-8. 14. Nehra D, Howell P, Williams CP, et al. Toxic bile acids in gastro-oesphageal reflux disease: influence of gastric acidity. Gut 1999;44:598-602. 15. Kauer WK, Peters JH, DeMeester TR, et al. Mixed reflux of gastric and duodenal juices is more harmful to the esophagus than gastric juice alone. The need for surgical therapy re-emphasized. Ann Surg 1995;222:525-31. 16. Champion G, Richter JE, Vaezi MF, et al. Duodenogastroesophageal reflux: relationship to pH and importance in Barrett’s esophagus. Gastroenterology 1994;107: 747-54. 17. Vaezi MF, Singh S, Richter JE, et al. Role of acid and duodenogastric reflux in esophageal mucosal injury: a review of animal and human studies. Gastroenterology 1995;108: 1897-907. 18. Ireland AP, Peters JH, Smyrk TC, et al. Gastric juice protects against the development of esophageal adenocarcinoma in the rat. Ann Surg 1996;224:358-70. 19. Fossmark R, Qvigstad G, Waldum HL. Gastric cancer: animal studies on the risk of hypoacidity and hypergastrinemia. World J Gastroenterol 2008;14:1646-51. 20. McColl KE, el-Omar E, Gillen D. Interactions between H. pylori infection, gastric acid secretion and anti-secretory therapy. Br Med Bull 1998;54:121-38. 21. Netzer P, Gut A, Brundler R, et al. Influence if pantoprazole on oesophageal motility, and bile and acid reflux in patients with oesphagitis. Aliment Pharmacol 2001;15:1375-84. 22. Marshall RE, Anggiansah A, Manifold DK, et al. Effect of omeprazole 20mg twice daily on duodenogastric and gastro-oesophageal bile reflux in Barrett’s oesophagus. Gut 1998;43:603-6. 23. Fein M, Peters JH, Chandrasoma P, Ireland AP, Oberg S, Ritter MP, et al. Duodenoesophageal reflux induces esophageal adenocarcinoma without exogenous carcinogen. J Gastrointest Surg 1998;2:260-8. 24. Atkins D, Reiffen K-A, Tegtmeier CL, Winther H, Bonato MS, Storkel S. Immunohistochemical detection of EGFR in paraffin-embedded tumor tissues: variation in staining intensity due to choice of fixative and storage time of tissue sections. J Histochem Cytochem 2004;52:893-901. 25. Peters CJ, Fitzgerald RC. Systematic review: the application of molecular pathogenesis to prevention and treatment of oesophageal adenocarcinoma. Aliment Pharmacol Ther 2007;25:1253-69. 26. Ohshima H, Tazawa H, Sylla BS, et al. Prevention of human cancer by modulation of chronic inflammatory processes. Mutat Res 2005;591:110-22. 27. Gonda TA, Tu S, Wang TC. Chronic inflammation, the tumor microenvironment and carcinogenesis. Cell Cycle 2009;8:2005-13.
Nasr et al 389
28. Zhang HY, Spechler SJ, Souza RF. Esophageal adenocarcinoma arising in Barrett esophagus. Cancer Lett 2009;275:170-7. 29. Lagarde SM, ten Kate FJ, Richel DJ, et al. Molecular prognostic factors in adenocarcinoma of the esophagus and gastroesophageal junction. Ann Surg Oncol 2007;14:977-91. 30. Kyrgidis A, Kountouras J, Zavos C, et al. New molecular concepts of Barrett’s esophagus: clinical implications and biomarkers. J Surg Res 2005;125:189-212. 31. Polkowski W, van Lanschot JJ, Ten Kate FJ, et al. The value of p53 and Ki67 as markers for tumour progression in the Barrett’s dysplasia-carcinoma sequence. Surg Oncol 1995; 4:163-71. 32. Hong MK, Laskin WB, Herman BE, et al. Expansion of the Ki-67 proliferative compartment correlates with degree of dysplasia in Barrett’s esophagus. Cancer 1995;75:423-9. 33. Binato M, Gurski RR, Fagundes RB, et al. P53 and Ki-67 overexpression in gastroesophageal reflux disease---Barrett’s esophagus and adenocarcinoma sequence. Dis Esophagus 2009;22:588-95. 34. Feith M, Stein HJ, Mueller J, et al. Malignant degeneration of Barrett’s esophagus: the role of the Ki-67 proliferation fraction, expression of E-cadherin and p53. Dis Esophagus 2004;17:322-7. 35. Fujii T, Nakagawa S, Hanzawa M, et al. Immunohistological study of cell cycle-related factors, oncogene expression, and cell proliferation in adenocarcinoma developed in Barrett’s esophagus. Oncol Rep 2003;10:427-31. 36. Kaur BS, Khamnehei N, Iravani M, et al. Rofecoxib inhibits cyclooxygenase 2 expression and activity and reduces cell proliferation in Barrett’s esophagus. Gastroenterology 2002;123:60-7. 37. Gillen P, McDermott M, Grehan D, et al. Proliferating cell nuclear antigen in the assessment of Barrett’s mucosa. Br J Surg 1994;81:1766-8. 38. Arnold D, Peinert S, Voigt W, et al. Epidermal growth factor receptor tyrosine kinase inhibitors: present and future role in gastrointestinal cancer treatment: a review. Oncologist 2006;11:602-11. 39. Ma WW, Hidalgo M. Exploiting novel molecular targets in gastrointestinal cancers. World J Gastroenterol 2007;13: 5845-56. 40. Tew WP, Kelsen DP, Ilson DH. Targeted therapies for esophageal cancer. Oncologist 2005;10:590-601. 41. Avissar NE, Toia L, Hu Y, et al. Bile acid alone, or in combination with acid, induces CDX2 expression through activation of the epidermal growth factor receptor (EGFR). J Gastrointest Surg 2009;13:212-22. 42. Moore KH, Barry P, Burn J, et al. Adenocarcinoma of the rat esophagus in the presence of a proton pump inhibitor: a pilot study. Dis Esophagus 2001;14:17-22. 43. Miyashita T, Shah FA, Marti GP, et al. Rabeprazole impedes the development of reflux-induced esophageal cancer in a surgical rat model. Dig Dis Sci 2011;56:1309-14. 44. Csendes A, Burgos AM, Smok G, et al. Latest results (12–21 years) of a prospective randomized study comparing Billroth II and Roux-en-Y anastomosis after a partial gastrectomy plus vagotomy in patients with duodenal ulcers. Ann Surg 2009;249:189-94. 45. Csendes A, Braghetto I. Development of de-novo Barrett’s esophagus after esophagogastric surgery in human subjects. Eur Surg 2009;41/3:120-5. 46. Duan L, Wu AH, Sullivan-Halley J, et al. Antacid drug use and risk of esophageal and gastric adenocarcinomas in Los Angeles County. Cancer Epidemiol Biomarkers Prev 2009;18:526-33.
390 Nasr et al
47. Farrow DC, Vaughan TL, Sweeney C, et al. Gastroesophageal reflux disease, use of H2 receptor antagonists, and risk of esophageal and gastric cancer. Cancer Causes Control 2000;11:231-8. 48. Lagergren J, Bergstr€ om R, Lindgren A, et al. Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N Engl J Med 1999;340:825-31. 49. Garcıa Rodrıguez LA, Lagergren J, Lindblad M. Gastric acid suppression and risk of oesophageal and gastric
Surgery March 2012
adenocarcinoma: a nested case control study in the UK. Gut 2006;55:1538-44. 50. Koek GH, Sifrim D, Lerut T, et al. Multivariate analysis of the association of acid and duodeno-gastro-oesophageal reflux exposure with the presence of oesophagitis, the severity of oesophagitis and Barrett’s oesophagus. Gut 2008;57:1056-64. 51. Talley NJ, Abeygunasekera S. Bile reflux and Barrett’s oesophagus: innocent bystander or sinister companion? Dig Liver Dis 2002;34:246-8.
Background. The contribution of gastric acid to the toxicity of alkaline duodenal refluxate on the esophageal mucosa is unclear. This study compared the effect of duodenal refluxate when acid was present, decreased by proton pump inhibitors (PPI), or absent. Methods. We randomized 136 Sprague-Dawley rats into 4 groups: group 1 (n = 33) were controls; group 2 (n = 34) underwent esophagoduodenostomy promoting ‘‘combined reflux’’; group 3 (n = 34) underwent esophagoduodenostomy and PPI treatment to decrease acid reflux; and group 4, the ‘gastrectomy’ group (n = 35) underwent esophagoduodenostomy and total gastrectomy to eliminate acid in the refluxate. Esophaguses were examined for inflammatory, Barrett’s, and other histologic changes, and expression of proliferative markers Ki-67, proliferating cell nuclear antigen (PCNA), and epidermal growth factor receptor (EGFR). Results. In all reflux groups, the incidence of Barrett’s mucosa was greater when acid was suppressed (group C, 62%; group D, 71%) than when not suppressed (group B, 27%; P = 0.004 and P < .001). Erosions were more frequent in the PPI and gastrectomy groups than in the combined reflux group. Edema (wet weight) and ulceration was more frequent in the gastrectomy than in the combined reflux group. Acute inflammatory changes were infrequent in the PPI group (8%) compared with the combined reflux (94%) or gastrectomy (100%) groups, but chronic inflammation persisted in 100% of the PPI group. EGFR levels were greater in the PPI compared with the combined reflux group (P = .04). Ki-67, PCNA, and combined marker scores were greater in the gastrectomy compared with the combined reflux group (P = .006, P = .14, and P < .001). Conclusion. Gastric acid suppression in the presence of duodenal refluxate caused increased rates of inflammatory changes, intestinal metaplasia, and molecular proliferative activity. PPIs suppressed acute inflammatory changes only, whereas chronic inflammatory changes persisted. (Surgery 2012;151:382-90.) From the Department of Surgery,a Royal College of Surgeons and the Department of Histopathology,b Connolly Hospital, Blanchardstown; and the Biomedical Research Facilityc and the Department of Surgery,d Royal College of Surgeons, Beaumont Hospital, Dublin, Ireland
FEW TOPICS IN MEDICINE excite greater controversy than the role of duodenal gastroesophageal reflux (DGER) in the pathophysiology of Barrett’s esophagus and esophageal cancer. The question of how duodenal refluxate interacts with gastric acid and Partially funded by the Oesophageal Cancer Fund. Accepted for publication August 18, 2011. Reprint requests: Thomas N. Walsh, MD, Department of Surgery, Royal College of Surgeons in Ireland, Connolly Hospital, Blanchardstown, Dublin, 15, Ireland. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2012 Mosby, Inc. All rights reserved. doi:10.1016/j.surg.2011.08.021
382 SURGERY
whether the interaction is detrimental or protective has proven difficult to resolve. It is undisputed that acid reflux causes esophageal mucosal damage, but it seems that acid alone cannot account for the metaplasia–dysplasia–carcinoma sequence. If this concept was correct, the widespread prescribing of proton pump inhibitors (PPI) should have resulted in decrease reduction in the incidence of esophageal adenocarcinoma. The converse is true, however, with a steeply increasing incidence of esophageal adenocarcinoma observed since the 1970s,1-3 Similarly, studies on the ability of PPIs to control or reverse Barrett’s esophagus have been disappointing.4-7
Surgery Volume 151, Number 3
Conceptually, it makes more sense that intestinal metaplasia is a cytologic response to a stimulus other than gastric acid. As Dixon et al8 observed, gastric metaplasia would be an appropriate response by esophageal epithelium to prolonged acid attack. The histologic hallmark of Barrett’s esophagus, however, is intestinal rather than gastric metaplasia, which is more likely to be a protective response against bile and duodenal juices. Numerous studies demonstrate a greater prevalence of Barrett’s esophagus when DGER is present consistently in the esophagus,9-15 but how this metaplasia relates to the adverse effects of acid on esophageal epithelium is less clear. Some believe that acid and DGER act syngerstically.16,17 Others go further, and suggest that acid has a calming effect on the noxious effects of DGER.18 Bile is most active at an alkaline or normal pH in its conjugated, soluble, ionized form, whereas bile in an acid environment precipitates and is much more inactive. This observation leads to the emergence of a more disturbing hypothesis: That acid suppression treatment may increase the risk of esophageal carcinoma. This concept is not without precedent, because gastric cancer has been linked with hypochlorhydria.19,20 Against this concept, other studies demonstrate that PPIs decrease gastroesophageal reflux disease (GERD),21,22 and possibly DGER, by decreasing the overall volume of gastric juices, although volume of DGER may be less relevant than pH and relative concentrations of the constituents of the esophageal environment. The purpose of this animal study was to observe the effect of acid and acid suppression on DGERmediated changes in the esophageal epithelium at the histologic and molecular levels. Incorporated into the model were groups that underwent acid suppression medically by means of PPI treatment and those that achieved complete acid suppression by means of total gastrectomy. METHODS The study was conducted in the Biomedical Research Facility, Beaumont Hospital, Dublin, Ireland. Ethics approval was obtained from the research ethics committee of The Royal College of Surgeons in Ireland. One author (AN) completed detailed animal handling and microsurgery training prior to commencement of the study. The animals used were Sprague Dawley (SD) outbred rats (Sprague Dawley Sprague Dawley, Harlan UK Ltd. Registered trademarks of Harlan Sprague Dawley, Inc, Indianapolis, IN, and Harlan Sprague Dawley, Inc, United Kingdom). Rats were
Nasr et al 383
maintained under a standard 12-hour dark–light cycle in a warm, balanced climate at a temperature of 22–258C and humidity of about 60%. They were all fed standard chow and water ad libitum. All animal handling was carried out under strict animal license guidelines provided and supervised by a licensed full-time animal handling supervisors. Study groups. The study population consisted of 145 male Sprague-Dawley rats were randomized into 4 study groups. The rats were between 8 and 10 weeks old and weighed 250–300 g. The study was powered on the basis of an expected frequency of columnar metaplastic changes of 33% in the omeprazole group, and it was calculated that 34 rats were required for each group. Provision was made for a presumed operative mortality of 10% based on the literature.18,23 In this study, 9 rats did not complete the study protocol owing primarily to perioperative complications. The 136 rats which completed the study were randomized into 4 groups: Group 1 (n = 33), the control group, were not subject to operation or pharmacologic acid suppression (Fig 1, A). Group 2 (n = 34), the combined reflux group, underwent esophagoduodenostomy to produce combined gastric (acidic) and duodenal (alkaline) reflux (Fig 1, B). Group 3 (n = 34), the PPI group, were ‘alkaline refluxers’ that underwent esophagoduodenostomy. Medical acid suppression with omeprazole therapy promoted isolated duodenal reflux (Fig 1, C). Group 4 (n = 35), the gastrectomy group,underwent esophagoduodenostomy and total gastrectomy to ensure isolated duodenal reflux and were ‘total alkaline refluxers’ with complete acid suppression (Fig 1, D).
Operation. All operations were performed under gaseous general anesthetic using isoflurane 5% and O2. The DGER model was created by performing an esophagoduodenostomy with preservation of the stomach performed by dividing the esophagus 1 cm proximal to the cardia and performing a 4 mm, longitudinal duodenotomy incision 1 cm distal to the pylorus. The esophagus was anastomosed to the duodenum using 6/0 polyprolene and an end-to-side anastomosis. The proximal end of the stomach was closed using a nylon ligature. For the DGER model, total gastrectomy was performed by dividing the stomach 1 cm proximal to the cardia and dividing the duodenum about 1 cm distal to the pylorus. An esophagoduodenostomy anastomosis was then performed making a 4 mm, longitudinal duodenotomy incision 1 cm distal to the pylorus, and the esophagus anastomosed end-to-side to the duodenum with 6/0
384 Nasr et al
Surgery March 2012
Fig 2. Gross specimen of resected esophagus.
Fig 1. A, Control group (Y acid reflux, Y bile reflux). B, Combined reflux group ([[ acid reflux, [[ bile reflux). C, PPI-treated group (Y acid reflux, [[ bile reflux). D, Gastrectomy group (YYacid reflux, [[ bile reflux).
polyprolene. The duodenal stump was closed a using nylon ligature. Management and measurements. Omeprazole, the PPI used in this study, was administered to the PPI group by means of a subcutaneous injection (13.5 mg/kg) on alternate days. All other groups received equivalent volume of 0.9% NaCl injections. After 26 weeks, animals were killed using an overdose of isoflurane gas. Immediately after death, the esophagus was resected from just below the larynx to the esophagogastric junction in the control group (including 2 mm of stomach wall) and to the esophagoduodenal anastomosis in the operated groups to include 2 mm of the duodenum wall (Fig 2). The esophagus was then opened longitudinally from the proximal end to the distal end through the anastomosis/esophagogastric junction. The wet/dry weight of the esophagus was obtained as a measure of tissue edema, reflecting
inflammatory activity due to mucosal injury. One longitudinal third of the resected esophagus was weighed to obtain the wet weight, and was then heated to 1718C for 72 hours in a dry heat oven to obtain the dry weight. The remaining esophagus was preserved in 10% neutral, buffered, formalin solution as the main specimen for histology and immunohistochemistry processing. Pathologic assessment. The specimens were paraffin fixed, sectioned at 4-mm intervals, and stained with hematoxylin and eosin. The slides were reviewed by a consultant histopathologist in a blinded fashion. Histologic grading was reported with reference to similar animal experiment in the literature.24 Acute inflammation was considered present whenever there was a predominance of neutrophils, whereas chronic inflammation was reported when there was a predominance of lymphocytes, plasma cells, or eosinophils. Erosion was defined as a histologic defect in the surface epithelium (mucosa) and ulceration was defined as histologic loss of surface epithelium with extension into the submucosa. Columnar metaplasia was defined as the presence of histologically abnormal glandular mucosa proximal to the anastomotic line or the esophagojejunal junction. Barrett’s metaplasia was defined as the presence of intestinal mucosa with goblet cells proximal to the anastomotic line or the esophagojejunal junction. Mucinous adenomatous changes were defined as the presence of pleomorphism, nuclear enlargement, hyperchromasia, abnormal chromatin pattern, and a high mitotic rate with predominance of mucinous locules. The histologic grading of hyperplasia and hyperkeratosis was defined with reference to previous animal experiment in literature.18 A specimen was defined as having epithelial hyperplasia when there was thickening of the epithelium with papillary elongation or basal cell hyperplasia compared with controls. Basal cell hyperplasia was defined as thickening of the basal cell layer to >10–15% of the epithelial thickness. Hyperkeratosis was defined as thickening of the keratin layer to >10–15% of the epithelial thickness compared with controls.
Surgery Volume 151, Number 3
Sections were dewaxed, rehydrated, and washed with phosphate-buffered saline. Endogenous peroxidase activity was neutralized by immersing the tissue in 0.5% hydrogen peroxide for 10 minutes. Antigen retrieval process was performed by boiling 0.01 mol/L citrate retrieval solution (pH 6.0) 1:100–1:200 in a stainless steel pressure cooker in which slides were immersed for 1 minute, and then cooled. The sections were incubated with primary antibodies as follows: Anti–Ki-67, proliferating cell nuclear antigen (PCNA), and EGFR mouse monoclonal antibodies (Novocastra Laboratories Ltd, Newcastle, United Kingdom), were used at 1:200 dilutions and incubated for 60 minutes at 258C. The sections were incubated subsequently with avidin/biotin complex-horseradish peroxidase for 60 minutes at 258C, after which they were developed in diaminobenzidine tetrahydrochloride for 60 minutes at 258C and counterstained with hematoxylin. The degree of immunohistologic staining of EGFR, Ki-67, and PCNA in the esophagus was determined using recognized scoring systems. Only squamous areas were scored under the microscope. The score given is based on the greatest area of staining present above the basal layer and must be >5% of each high power field. EGFR was scored according to a previously published scoring system25 (0 = Membranous staining in <10% of cells; 1 = partial membranous staining in >10% of cells; 2 = strong, complete membranous staining in >10% of cells). Ki-67 and PCNA staining was also scored using a recognized scoring system25 (0 = no score; 1 = increased staining in <50% of epithelial thickness; 2 = increased staining in >50% of epithelial thickness). A score of $1 was taken as positive. Statistical analysis. Sample size was calculated using PS: Power and Sample Size Calculation software (Version 2.1.31, 2003, Department of Biostatistics; Vanderbilt University School of Medicine, Nashville, TN). Using a power of 0.8 and an expected frequency of columnar metaplastic changes of 33% in the omeprazole group, 34 rats were required in each group. Further statistical analysis was performed using SPSS 17.0 software for Windows (SPSS Inc., Chicago, IL). Chi-square tests, Fisher exact tests, and the Mann–Whitney U test were used where appropriate. RESULTS Level of inflammation. The mean percentage of the length of the inflamed segment to the full length of esophagus of the control group was 0.05%. The percentage of esophagus that was
Nasr et al 385
inflamed for the combined reflux, PPI and gastrectomy groups was 78%, 82%, and 80%, respectively. Esophageal weight. The median weight of the esophagus for the control group was 443 mg. In comparison, all other groups demonstrated a significantly increased esophageal weight. (The median weight in the combined reflux, PPI, and gastrectomy groups was 1445, 1264, and 1707 mg [P < .001 each], respectively.) There was a difference in the median weight of the combined reflux group when compared with the gastrectomy group (P = .039) but not in the combined reflux group when compared with the PPI group (P = .21). The median wet and dry weights of (1/3 sections) of the esophaguses in each group are shown in Table I. The median wet/dry weight ratio of the esophagus for the control, combined reflux, PPI, and gastrectomy groups were 1.70, 1.68, 1.81, and 1.97. Only the gastrectomy group differed from the control group when comparing the wet/dry weight ratio (P = .001). The median wet/dry weight of the control group vs the combined reflux and the PPI groups was not different (P $ .17 each). Histology. Inflammation: Table II summarizes the microscopic findings of the study. There were no acute inflammatory changes noted in the control group. Ninety-four percent (n = 32/34) of the combined reflux group and 100% (n = 35/35) of the gastrectomy group developed acute inflammation. In contrast, only 8% (n = 3/34) of the PPI group developed acute inflammatory changes. Chronic inflammatory changes were evident in 9% of the control, 97% of the combined reflux, and 100% of the gastrectomy groups. Although few esophaguses in the PPI group demonstrated acute inflammation, all (100%) demonstrated chronic inflammation. Ulcerations/erosions: Microscopic ulceration was absent in the control group but was identified at high levels in all other groups (Table II). The presence of ulceration in the PPI and gastrectomy groups were greater than the combined reflux group (P # .042 each). Similarly, erosions were greater in combined reflux, PPI, and gastrectomy groups, and greater in the gastrectomy group compared with the combined reflux group (P = .005) but not when the PPI group was compared with the combined reflux group (P = .17). Barrett’s esophagus: Barrett’s metaplasia was not identified in the control group but was present in 27% of the combined reflux, 62% of the PPI, and 71% of the gastrectomy groups (Fig 3; Table II). There was a greater degree of Barrett’s in the PPI
386 Nasr et al
Surgery March 2012
Table I. Median esophageal weights of the various groups Weight (mg)
Control
Combined refluxers
PPI
Gastrectomy
Median total esophageal Median dry* Median wet*
443
1445
1264
1707
171 292
265 437
255 430
284 550
*Dry and wet weights are determined from a longitudinal section of 1/3 the original esophageal specimen. PPI, Proton pump inhibitors.
group compared with the combined reflux group (P = .004) and in the gastrectomy group compared with the combined reflux group (P < .001). Mucinous adenomatous changes: Mucinous adenomatous changes were evident in combined reflux, PPI, and gastrectomy groups and greatest in the gastrectomy group (26%). Epithelial and basal cell hyperplasia and epithelial hyperkeratosis: All patients in the combined reflux, PPI, and gastrectomy groups exhibited epithelial cell hyperplasia compared with only 36% of the control group (the hyperplasia in the control group was graded as mild). Moderate-to-severe hyperplasia was present in 0%, 88%, 88%, and 100% of the control, combined reflux, PPI, and gastrectomy groups (Table II). There was no basal cell hyperplasia in the control group. Levels of moderate to severe basal cell hyperplasia in the combined reflux, PPI, and gastrectomy groups were 79%, 76%, and 83%, respectively. Mild epithelial hyperkeratosis was present in 39% of the control group, whereas moderate-to-severe epithelial hyperkeratosis was present in 3%, 97%, 100%, and 94% of, respectively, the control, combined reflux, PPI, and gastrectomy groups. Immunohistochemical analysis of expression of EGFR, Ki-67, and PCNA: EGFR expression in the control, combined reflux, PPI, and gastrectomy groups was 6% (2/33), 68% (23/34), 88% (30/34), and 83% (29/35; Fig 4), respectively. The differences between the combined reflux and the PPI groups were significant (P = .041) but not between the combined reflux and the gastrectomy groups (P = .14). Ki-67 expression in the control, combined reflux, PPI, and gastrectomy groups was 3% (1/33), 73% (25/34), 76% (26/34), and 97% (34/35). Higher expression of Ki-67 in the gastrectomy group was greater compared with the combined reflux group (P = .006). PCNA expression in the control, combined reflux, PPI, and gastrectomy groups was 0% (0/33), 67% (23/34), 56% (19/34), and 91% (32/35), respectively. Again, differences between
the combined reflux and the gastrectomy groups were significant (P = .014). Combined expression of proliferation markers: Combined expression in the control, combined reflux, PPI, and gastrectomy groups was 0% (0/33), 32% (10/34), 44% (15/34), and 77% (27/35), respectively. The differences between the combined reflux and the gastrectomy groups were significant (P < .001), but not between the combined reflux and the PPI groups (P = .21). DISCUSSION This study suggests that pure duodenal reflux is a powerful stimulus to the development of Barrett’s metaplasia. More important, acid suppression by medical or operative means worsened the cellular changes caused by DGER. The design of the study was powered specifically to examine the development of Barrett’s esophagus in response to DGER combined with PPI therapy. We demonstrated that although the combination of DGER and acid reflux resulted in a 27% rate of Barrett’s esophagus, a much greater percentage (62%) of rats developed Barrett’s when subjected to DGER combined with PPI therapy. Those rats were subjected to pure DGER and extreme acid suppression (by means of radical gastrectomy) developed the greatest rates of Barrett’s metaplasia (71%). In addition, further histologic and immunohistochemical analysis in this study showed consistently the heightened noxious effect of DGER in the setting of minimal or no acid. In addition to increasing the rate of Barrett’s metaplasia, decreases in acid in the duodenalesophageal refluxate also increased markers of inflammation, including erosions, ulcerations, and edema, as well as provoking an overall increase in measured markers of proliferation. Of particular interest was the finding that while the PPI group demonstrated very low levels of acute inflammation when compared with the other groups, this group maintained a very high level of chronic inflammatory changes. It is possible that PPIs may result in amelioration of acute inflammatory changes and even acute symptomatic GERD, whereas chronic, more insidious changes persist. The ability of PPI therapy to decrease the incidence of acute inflammatory changes when compared with the gastrectomy group may be attributed theoretically to its ability to decrease the volume of reflux or to other intrinsic anti-inflammatory effects of the PPI therapy. However, it is chronic inflammation that is thought to be the main stimulus or risk factor for $18% of cancers.26 Certainly, chronically sustained activation of the inflammatory response is thought
Nasr et al 387
Surgery Volume 151, Number 3
Table II. Summary of histology results Control (n = 33), n (%) No inflammation Acute inflammation Microscopic erosion Microscopic ulceration Chronic inflammation Epithelial hyperplasia Basal cell hyperplasia Barrett’s metaplasia Mucinous adenomatous changes
30 0 0 0 3 21 0 0 0
(91) (0) (0) (0) (9) (64) (0) (0) (0)
Combined (n = 34), n (%) 1 32 27 18 33 All All 9 5
(3) (94) (79) (53) (97) (100) (100) (27) (15)
PPI (n = 34), n (%) 0 3 31 26 All All All 21 5
Gastrectomy (n = 35), n (%)
(0) (9%) (91) (77) (100) (100) (100) (62) (15)
0 All All 32 All All All 25 9
(0) (100) (100) (91) (100) (100) (100) (71) (26)
PPI, Proton pump inhibitors.
80%
120
71%
70%
100
62% % expression
60% 50% 40% 26%
30% 20%
80 60 40 20
10% 0%
0
0% Control
Combined
PPI
Gastrectomy
Fig 3. Distribution of Barrett’s metaplasia among the different groups.
Control
Combined Refluxers
PPI
gastrectomy
EGF-R Ki-67 PCNA All 3 Markers
to produce a protumorigenic microenvironment needed to generate the culmination of changes needed for cell transformation whether it be metaplastic or carcinogenic.27 The alternative metaplastic theory, whereby inflammatory cytokines attract and transform stem cells to become metaplastic,27,28 is also more likely to be supported by a chronically inflamed microenvironment. A wide range of molecular markers have been used to examine proliferative changes in the esophagus.24,26,29 In this study, 3 well-established markers were used to estimate proliferative activity and pathologic progression. EGFR expression was greater in the PPI group than in the combined DGER/acid group. Levels of Ki-67 and PNCA were greater in the total acid suppression (gastrectomy) group compared with the combined DGER/acid group. Ki-67 is a nuclear antigen expressed in proliferating cells, but not in resting cells. Expression of Ki- 67 increases as cells progress along the normal–Barrett’s esophagus–carcinoma sequence.30-34 PCNA is an essential protein for cell proliferation. Like Ki-67, PCNA has been implicated in the transformation of normal squamous epithelial cells to esophageal adenocarcinoma,30,35 with progressively greater
Fig 4. Proportion of rats expressing various proliferation markers.
levels expressed when normal cells transform into Barrett’s esophagus and adenocarcinoma.35,36 Gillen et al37 found that PNCA expression in Barrett’s mucosa located adjacent to adenocarcinoma was greater than in normal Barrett’s mucosa. The third marker, EGFR, has been associated with both squamous cell carcinoma and adenocarcinoma of the esophagus.29,38 EGFR is important, because EGFR inhibitors have been tested clinically in other carcinomas and found to have a therapeutic effect.38-40 Current clinical trials of antibody inhibition of EGFR with cetuximab are under way in esophageal carcinoma40 and there is evidence that EGFR may have prognostic significance. Furthermore, some evidence suggests that Barrett’s change may occur via bile-stimulated cell signaling through the EGFR.41 Taken together, the results of this study further support the histologic findings that DGER combined with acid suppression increases cellular proliferative activity and possibly pathologic progression in esophageal cells.
388 Nasr et al
Our study examined the effects of both PPIinduced acid suppression and total acid suppression on DGER in an animal model. Only 2 other models in the literature studied the role of PPI therapy in esophageal carcinogenesis. Moore et al42 compared a PPI group and a control group, in which they anastomosed the duodenum to the esophagus. They found no difference between the groups. A recent study by Miyashita et al43 used a different model, anastomosing the jejunum to the esophagus and disconnecting the stomach from the esophagus. This technique resulted in both groups having a very alkaline pH in the lower esophagus (pH of 6.5 in both groups). In comparison, the esophageal pH in the present study ranged from 5.4 to 8.4 across the various models. Miyashita et al’s study demonstrated a decrease in esophageal cancer in rats treated with rabeprazole when compared with the control group; the authors hypothesized that rabeprazole, in addition to its acid suppressive effects, may have important anti-inflammatory effects, which in turn may have an anticarcinogenic effect. Other studies have examined manipulation of acid/DGER without the use of PPI. Ireland et al18 treated rats with a nitrosamine-based carcinogen and subjected the rats to various operations to induce varying degrees of acid and DGER. The absence of gastric juice increased the rates of DGER/carcinogeninduced esophageal adenocarcinoma 3-fold. Csendes et al44,45 demonstrated that the incidence of Barrett’s was greater in patients with increased DGER combined with decreased acid secretion as a result of a Billroth II, partial gastrectomy, and selective vagotomy than in patients with Roux-en-Y partial gastrectomy and selective vagotomy. The premise that decreasing acid reflux has a detrimental effect on DGER-mediated esophageal injury has some support from population-based studies.46-48 First, in these studies46-50 medications that decrease acid production did not have a protective effect against esophageal adenocarcinoma. Conversely, others have found an increased risk of esophageal carcinoma with the use of these medications,46-48 although this is not a consistent finding.26 An important confounder in all of these trials is that a proportion of patients on acid decreasing medication are been treated for GERD and are, therefore, potentially at greater risk of esophageal carcinoma. For instance, Garcıa Rodrıguez et al49 demonstrated a 5-fold increase in esophageal cancer when PPIs were prescribed for GERD but not when prescribed for other reasons. Arguably, their observation does not negate the premise that the toxic effects of DGER may be potentiated by PPIs. In contrast,
Surgery March 2012 Duan et al46 found that patients were at an increased risk of esophageal adenocarcinoma if they were taking nonprescription acid-neutralizing tablets without GERD symptoms, and Farrow et al47 reported an increased risk of esophageal adenocarcinoma with over the counter antacids. In a Swedish study, Lagergren et al48 demonstrated an increase in esophageal adenocarcinoma in patients with reflux symptoms who were taking medication for GERD compared with those who took no medication. In conclusion, although our study demonstrated that acid suppression causes detrimental effects in the presence of DGER, it is likely that the processes of progression to Barrett’s and carcinoma have complex causes. Both acid reflux and DGER are, on multivariate analysis, separately and independently associated with the development of Barrett’s esophagus.50 There may, therefore, be more than one potential mechanism of damage to esophageal cells that initiates cell transformation. It is likely that there is a complex interplay between acid and DGER that is affected by the timing and relative concentrations of acid and DGER reflux.51 Alternatively, it may be the case that acid suppression is only detrimental when DGER is present. Nevertheless, the results of this and other studies suggest that in the presence of DGER, acid suppression by medical or other means increases the rate of Barrett’s and potentially increases the risk of esophageal adenocarcinoma. REFERENCES 1. Blot WJ, Devesa SS, Kneller RW, et al. Rising incidence of adenocarcinoma of the esophagus and gastric cardia. JAMA 1991;265:1287-9. 2. Devesa SS, Blot WJ, Fraumeni JF Jr. Changing patterns in the incidence of esophageal and gastric carcinoma in the United States. Cancer 1998;83:2049-53. 3. Brown LM, Devesa SS, Chow WH. Incidence of adenocarcinoma of the esophagus among white Americans by sex, stage, and age. Natl Cancer Inst 2008;100:1184-7. 4. Peters FT, Ganesh S, Kuipers EJ, et al. Endoscopic regression of Barrett’s oesophagus during omeprazole treatment; a randomised double blind study. Gut 1999:45489-94. 5. Faybush EM, Sampliner RE. Randomized trials in the treatment of Barrett’s esophagus. Dis Esophagus 2005;18:291-7. 6. Parrilla P, Martınez de Haro LF, Ortiz A, et al. Long-term results of a randomized prospective study comparing medical and surgical treatment of Barrett’s esophagus. Ann Surg 2003;237:291-8. 7. Rees JR, Lao-Sirieix P, Wong A, et al. Treatment for Barrett’s oesophagus. Cochrane Database Syst Rev 2010;1:CD004060. 8. Dixon MF, Neville PM, Mapstone NP, et al. Bile reflux gastritis and Barrett’s oesophagus: further evidence of a role for duodenogastro-oesphageal reflux? Gut 2001;49: 359-63. 9. Fein M, Maroske J, Fuchs KH. Importance of duodenogastric reflux in gastro-oesophageal reflux disease. Br J Surg 2006;93:1475-82.
Surgery Volume 151, Number 3
10. Koek GH, Sifrim D, Lerut T, et al. Multivariate analysis of the association of acid and duodeno-gastro-oesophageal reflux exposure with the presence of oesophagitis, the severity of oesophagitis and Barrett’s oesophagus. Gut 2008;57: 1056-64. 11. Caldwell MT, Lawlor P, Byrne PJ, et al. Ambulatory oesophageal bile reflux monitoring in Barrett’s oesophagus. Br J Surg 1995;82:657-60. 12. Stein HJ, Kauer WK, Feussner H, et al. Bile reflux in benign and malignant Barrett’s esophagus: effect of medical acid suppression and Nissen fundoplication. J Gastrointest Surg 1998;2:333-41. 13. Marshall RE, Anggiansah A, Owen WJ. Bile in the oesophagus: clinical relevance and ambulatory detection. Br J Surg 1997;84:21-8. 14. Nehra D, Howell P, Williams CP, et al. Toxic bile acids in gastro-oesphageal reflux disease: influence of gastric acidity. Gut 1999;44:598-602. 15. Kauer WK, Peters JH, DeMeester TR, et al. Mixed reflux of gastric and duodenal juices is more harmful to the esophagus than gastric juice alone. The need for surgical therapy re-emphasized. Ann Surg 1995;222:525-31. 16. Champion G, Richter JE, Vaezi MF, et al. Duodenogastroesophageal reflux: relationship to pH and importance in Barrett’s esophagus. Gastroenterology 1994;107: 747-54. 17. Vaezi MF, Singh S, Richter JE, et al. Role of acid and duodenogastric reflux in esophageal mucosal injury: a review of animal and human studies. Gastroenterology 1995;108: 1897-907. 18. Ireland AP, Peters JH, Smyrk TC, et al. Gastric juice protects against the development of esophageal adenocarcinoma in the rat. Ann Surg 1996;224:358-70. 19. Fossmark R, Qvigstad G, Waldum HL. Gastric cancer: animal studies on the risk of hypoacidity and hypergastrinemia. World J Gastroenterol 2008;14:1646-51. 20. McColl KE, el-Omar E, Gillen D. Interactions between H. pylori infection, gastric acid secretion and anti-secretory therapy. Br Med Bull 1998;54:121-38. 21. Netzer P, Gut A, Brundler R, et al. Influence if pantoprazole on oesophageal motility, and bile and acid reflux in patients with oesphagitis. Aliment Pharmacol 2001;15:1375-84. 22. Marshall RE, Anggiansah A, Manifold DK, et al. Effect of omeprazole 20mg twice daily on duodenogastric and gastro-oesophageal bile reflux in Barrett’s oesophagus. Gut 1998;43:603-6. 23. Fein M, Peters JH, Chandrasoma P, Ireland AP, Oberg S, Ritter MP, et al. Duodenoesophageal reflux induces esophageal adenocarcinoma without exogenous carcinogen. J Gastrointest Surg 1998;2:260-8. 24. Atkins D, Reiffen K-A, Tegtmeier CL, Winther H, Bonato MS, Storkel S. Immunohistochemical detection of EGFR in paraffin-embedded tumor tissues: variation in staining intensity due to choice of fixative and storage time of tissue sections. J Histochem Cytochem 2004;52:893-901. 25. Peters CJ, Fitzgerald RC. Systematic review: the application of molecular pathogenesis to prevention and treatment of oesophageal adenocarcinoma. Aliment Pharmacol Ther 2007;25:1253-69. 26. Ohshima H, Tazawa H, Sylla BS, et al. Prevention of human cancer by modulation of chronic inflammatory processes. Mutat Res 2005;591:110-22. 27. Gonda TA, Tu S, Wang TC. Chronic inflammation, the tumor microenvironment and carcinogenesis. Cell Cycle 2009;8:2005-13.
Nasr et al 389
28. Zhang HY, Spechler SJ, Souza RF. Esophageal adenocarcinoma arising in Barrett esophagus. Cancer Lett 2009;275:170-7. 29. Lagarde SM, ten Kate FJ, Richel DJ, et al. Molecular prognostic factors in adenocarcinoma of the esophagus and gastroesophageal junction. Ann Surg Oncol 2007;14:977-91. 30. Kyrgidis A, Kountouras J, Zavos C, et al. New molecular concepts of Barrett’s esophagus: clinical implications and biomarkers. J Surg Res 2005;125:189-212. 31. Polkowski W, van Lanschot JJ, Ten Kate FJ, et al. The value of p53 and Ki67 as markers for tumour progression in the Barrett’s dysplasia-carcinoma sequence. Surg Oncol 1995; 4:163-71. 32. Hong MK, Laskin WB, Herman BE, et al. Expansion of the Ki-67 proliferative compartment correlates with degree of dysplasia in Barrett’s esophagus. Cancer 1995;75:423-9. 33. Binato M, Gurski RR, Fagundes RB, et al. P53 and Ki-67 overexpression in gastroesophageal reflux disease---Barrett’s esophagus and adenocarcinoma sequence. Dis Esophagus 2009;22:588-95. 34. Feith M, Stein HJ, Mueller J, et al. Malignant degeneration of Barrett’s esophagus: the role of the Ki-67 proliferation fraction, expression of E-cadherin and p53. Dis Esophagus 2004;17:322-7. 35. Fujii T, Nakagawa S, Hanzawa M, et al. Immunohistological study of cell cycle-related factors, oncogene expression, and cell proliferation in adenocarcinoma developed in Barrett’s esophagus. Oncol Rep 2003;10:427-31. 36. Kaur BS, Khamnehei N, Iravani M, et al. Rofecoxib inhibits cyclooxygenase 2 expression and activity and reduces cell proliferation in Barrett’s esophagus. Gastroenterology 2002;123:60-7. 37. Gillen P, McDermott M, Grehan D, et al. Proliferating cell nuclear antigen in the assessment of Barrett’s mucosa. Br J Surg 1994;81:1766-8. 38. Arnold D, Peinert S, Voigt W, et al. Epidermal growth factor receptor tyrosine kinase inhibitors: present and future role in gastrointestinal cancer treatment: a review. Oncologist 2006;11:602-11. 39. Ma WW, Hidalgo M. Exploiting novel molecular targets in gastrointestinal cancers. World J Gastroenterol 2007;13: 5845-56. 40. Tew WP, Kelsen DP, Ilson DH. Targeted therapies for esophageal cancer. Oncologist 2005;10:590-601. 41. Avissar NE, Toia L, Hu Y, et al. Bile acid alone, or in combination with acid, induces CDX2 expression through activation of the epidermal growth factor receptor (EGFR). J Gastrointest Surg 2009;13:212-22. 42. Moore KH, Barry P, Burn J, et al. Adenocarcinoma of the rat esophagus in the presence of a proton pump inhibitor: a pilot study. Dis Esophagus 2001;14:17-22. 43. Miyashita T, Shah FA, Marti GP, et al. Rabeprazole impedes the development of reflux-induced esophageal cancer in a surgical rat model. Dig Dis Sci 2011;56:1309-14. 44. Csendes A, Burgos AM, Smok G, et al. Latest results (12–21 years) of a prospective randomized study comparing Billroth II and Roux-en-Y anastomosis after a partial gastrectomy plus vagotomy in patients with duodenal ulcers. Ann Surg 2009;249:189-94. 45. Csendes A, Braghetto I. Development of de-novo Barrett’s esophagus after esophagogastric surgery in human subjects. Eur Surg 2009;41/3:120-5. 46. Duan L, Wu AH, Sullivan-Halley J, et al. Antacid drug use and risk of esophageal and gastric adenocarcinomas in Los Angeles County. Cancer Epidemiol Biomarkers Prev 2009;18:526-33.
390 Nasr et al
47. Farrow DC, Vaughan TL, Sweeney C, et al. Gastroesophageal reflux disease, use of H2 receptor antagonists, and risk of esophageal and gastric cancer. Cancer Causes Control 2000;11:231-8. 48. Lagergren J, Bergstr€ om R, Lindgren A, et al. Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N Engl J Med 1999;340:825-31. 49. Garcıa Rodrıguez LA, Lagergren J, Lindblad M. Gastric acid suppression and risk of oesophageal and gastric
Surgery March 2012
adenocarcinoma: a nested case control study in the UK. Gut 2006;55:1538-44. 50. Koek GH, Sifrim D, Lerut T, et al. Multivariate analysis of the association of acid and duodeno-gastro-oesophageal reflux exposure with the presence of oesophagitis, the severity of oesophagitis and Barrett’s oesophagus. Gut 2008;57:1056-64. 51. Talley NJ, Abeygunasekera S. Bile reflux and Barrett’s oesophagus: innocent bystander or sinister companion? Dig Liver Dis 2002;34:246-8.