Cardiac mucosa in the remnantesophagus after esophagectomy is anacquired epithelium with Barrett’s-likefeaturesReginald V. N. Lord, MBBS, MD, Kumari Wickramasinghe, MD, Jan J. Johansson, MD,Steven R. DeMeester, MD, Jan Brabender, MD, and Tom R. DeMeester, MD, Los Angeles, Calif

Background. The cervical esophagus is normally lined by squamous epithelium and is usually notexposed to gastroesophageal reflux. The aims of this study were, first, to investigate whether cardiacmucosa can be acquired in the remnant cervical esophagus after esophagectomy and cervicalesophagogastrostomy and, second, to characterize this mucosa if present.Methods. The medical records of 100 patients who had undergone esophagectomy with gastric pull-upreconstruction were studied retrospectively to identify those who had biopsies from the cervical esophagusproximal to the gastroesophageal anastomosis during postoperative follow-up. The histopathology andimmunohistochemical stains were reviewed to assess similarity to Barrett’s mucosa (cytokeratins [CK] 7and 20 and DAS-1), cellular proliferation (topoisomerase 2a), and the potential for dysplasia (cyclo-oxygenase 2 [COX-2] and ornithine decarboxylase [ODC]).Results. Supra-anastomotic biopsies were performed in 20 patients. Cardiac mucosa was present in 10 of20 (50%) patients in whom biopsies were performed. Four patients had areas of intestinal metaplasia,and dysplasia, and adenocarcinoma developed in 1 patient. The CK7/20 and DAS-1 staining of thecolumnar mucosa showed a pattern similar to Barrett’s mucosa. Topoisomerase 2a protein expressionwas present in 50% of patients with cardiac mucosa. DAS-1 protein was expressed in cervical columnarmucosa but not in normal squamous esophagus mucosa. The cardiac mucosa stained weakly for COX-2and ODC.Conclusions. Cardiac mucosa can be acquired. Its expression profile is similar to cardiac mucosa andintestinal metaplasia found in Barrett’s esophagus, and different from normal esophageal or gastricmucosa. The development of cardiac mucosa is likely to be related to reflux of acid into the remnantcervical esophagus as the first step in the development of Barrett’s esophagus. These findings areapplicable to the development of similar changes at the gastroesophageal junction. (Surgery2004;136:633-40.)

From the Departments of Surgery and Pathology, University of Southern California Keck School of Medicine,Los Angeles, Calif

CARDIAC MUCOSA IS A SIMPLE, MUCINOUS COLUMNAR

mucosa with foveolar hyperplasia and no parietalcells. It is found in the region of the gastroesoph-

Accepted for publication January 17, 2004.

Supported by grants from the American Cancer Society, theInternational Society for Diseases of the Esophagus, and theSTOP Cancer Foundation (R.V.N.L.).

Presented at the Society of University Surgeons 62nd AnnualMeeting, February 8-10, 2001.

Reprint requests: Reginald V. N. Lord, MBBS, MD, HCC 514,1510 San Pablo St., Los Angeles CA 90033.

0039-6060/$ - see front matter

� 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.surg.2004.01.009

ageal junction in most adults in Western society.When present, it is almost invariably accompaniedby an infiltrate of chronic or acute inflammatorycells and may thus be termed ‘‘carditis.’’1 Cardiacmucosa is distinguished from the intestinal meta-plasia (IM) that characterizes Barrett’s esophagusonly by the absence of goblet cells.

In the past, it was believed that up to 2 cm ofcardiac mucosa was normally present in the mostproximal section of the stomach, where it separa-tes the parietal cell-containing gastric oxyntic mu-cosa from the esophageal squamous mucosa.2,3

This prevailing view was challenged by a studysuggesting that cardiac mucosa, rather than being anormally occurring mucosa, might be an acquired,

SURGERY 633

SurgerySeptember 2004

634 Lord et al

metaplastic epithelium that develops in response toexposure of esophageal squamous epithelium togastric acid.4 According to this hypothesis, thehistology of the normal gastroesophageal junctionconsists of squamous mucosa abutting the parietalcell-containing gastric oxyntic mucosa. Subsequentstudies have confirmed that this histologic pattern,with squamous epithelium directly abutting oxyntic,does occur.5,6 Further observations suggest thatthe development of cardiac mucosa is induced byexposing squamous epithelium to refluxed gastricacid. The stimulus for the development of car-diac mucosa may be refluxed gastric acid.7–9 Thisintroduced the possibility that the formation ofcardiac mucosa may be the first step in thedevelopment of Barrett’s esophagus.10,11 Othersreject this possibility, leading to controversy re-garding the nature and etiology of cardiac mu-cosa.12–14 Unfortunately, limitations as to theaccurate location of endoscopic biopsies13,15 andthe rapid autolysis of the mucosa of the gastro-esophageal junction in autopsy specimens5 havemade it difficult to resolve this controversy.14

In this study, we used the cervical esophagusafter esophagectomy reconstructed with a gastroeso-phagostomy as a model for de novo gastroesoph-ageal reflux. In this situation, it can be confirmedhistologically that the squamous-lined cervical eso-phagus is anastomosed to the gastric fundus linedwith oxyntic mucosa because any previously presentcardiac mucosa was removed with the surgicalspecimen. Further, the cervical esophagus is notnormally exposed to gastric juice. Even in patientsdiagnosed with gastroesophageal reflux disease(GERD), acid exposure in the cervical esophagusis relatively infrequent—less than 1% of a 24-hourperiod.16 Consequently, even in patients withsevere GERD, the cervical esophagus is normallylined by squamous epithelium. In contrast, afteresophagectomy and gastric pull-up, free reflux intothe cervical esophagus occurs.17 Esophagectomyand cervical esophagogastrostomy is thus a uniquein vivo human model for de novo reflux disease.

This study was undertaken to test the hypothesisthat cardiac mucosa is an acquired, metaplasticepithelium that arises from squamous mucosa inresponse to exposure to gastric acid. If this hypo-thesis is correct, reflux into the cervical esophagusafter esophagectomy should result in the develop-ment of cardiac mucosa in the remnant esophagus,as reported in other studies.18–20 We also sought tocharacterize with immunohistochemistry the ac-quired mucosa and to compare it with cardiacmucosa at the gastroesophageal junction andBarrett’s mucosa in the distal esophagus.

MATERIAL AND METHODS

After obtaining approval for this study from theInstitutional Review Board of the University ofSouthern California Keck School of Medicine, themedical records of 100 patients with esophagealadenocarcinoma or squamous cell carcinoma whohad undergone esophagectomy with gastric tubereconstruction were retrospectively reviewed.Patients who had received chemotherapy or radi-ation therapy were excluded because of the re-ported association between these treatments andthe development of Barrett’s esophagus andesophageal cancer. Of the 100 patients, 20 had anendoscopic evaluation, which included a biopsy ofthe remnant cervical esophagus taken from abovethe gastroesophageal anastomosis 9 months ormore after esophagectomy. The endoscopies wereperformed to investigate symptoms, includingregurgitation, dysphagia, chest pain, and eitherloss or failure to gain weight; thus, these patientswere somewhat selected and may not be represen-tative of the entire group of 100 patients reviewed.The supra-anastomotic biopsies were performed toconduct studies such as the present one. Allbiopsies were performed by members of the sur-gery faculty. The hematoxylin and eosin (H&E)stained slides of all postoperative endoscopicbiopsies from these 20 patients were reviewed,and patients with cardiac mucosa in the cervicalesophagus were identified. Formalin-fixed, paraffin-embedded blocks of the esophageal biopsiesand the gastric biopsies, if obtained, from thepatients with cardiac mucosa in the cervicalesophagus were retrieved from the pathologyarchives. In these selected patients, the pathologyreport on the surgical margin before anastomosiswas reviewed.

Histopathology. The H&E-stained slides of thesupra-anastomotic cervical esophageal biopsieswere examined using the criteria of Paull et al,21

as modified by Chandrasoma et al.22 The diagnosisof pure cardiac mucosa was made when the glandswere composed of mucous cells only, with noparietal cells. Pure cardiac mucosa is thus similarto the junctional epithelium of Paull et al.21

Oxyntic mucosa was recognized by the presenceof glands containing both mucous and parietalcells, equivalent to the fundic epithelium describedby Paull et al.21 The diagnosis of IM required thepresence of definitive goblet cells. When visiblemacroscopically, IM (or specialized epithelium)21 isreferred to as Barrett’s esophagus. Information inthe medical records was insufficient to be certain ofthe endoscopic appearance of the cervical esoph-

SurgeryVolume 136, Number 3

Lord et al 635

agus in some patients, and therefore, macroscopicdata were not collected.

Immunohistochemistry. Immunohistochemicalstains were performed to assess similarity toBarrett’s mucosa (cytokeratins [CK] 7 and 20 andDAS-1), cellular proliferation (topoisomerase 2a),and the potential for dysplasia (cyclo-oxygenase 2[COX-2], and ornithine decarboxylase [ODC]).

Archival formalin-fixed, paraffin-embeddedblocks were cut into 6-lm sections, mounted ontopolylysine-coated slides, dewaxed in xylene, andrehydrated through graded alcohol steps at roomtemperature. Pretreatment by immersion in 10mmol/L citrate buffer pH 6.0 with microwave,pressure cooker heating was performed for all theantibodies used. A 0.05M Tris-HCl buffer solution(pH 7.6) was used to prepare solutions and forwashes between steps. The sections were peroxi-dase-blocked using 3% hydrogen peroxide in 0.05mol/L TRIS-hydrochloric acid buffer, incubatedfor 15 minutes with normal horse serum, andincubated with primary antibody (all overnightat room temperature). The primary antibodiesused were: topoisomerase 2a Mab (diluted 1:100;Neomarkers, Clone JH2.7, Fremont, Calif), cyto-keratin (CK) 7 and CK 20 (both 1:100; DAKO,

Fig 1. Endoscopic photograph showing a tongue ofintestinal metaplasia (Barrett’s esophagus) in the cervicalesophagus above the gastroesophageal anastomosis.

Carpinteria, Calif), ornithine decarboxylase (ODC-29, 1:50; Sigma Chemical, St. Louis, Mo), DAS-1(1:5, kindly provided by Dr Kiron M. Das, Universityof Medicine and Dentistry of New Jersey, RobertWood Johnson Medical School), and cyclo-oxygenase 2 Mab (COX-2 clone 33, 1:50;Transduction Laboratories, Lexington, Ky).Biotinylated horse antimouse secondary antibody(1:200 dilution for 40 minutes; Vector Labs, Burlin-game Calif), peroxidase-conjugated-streptavidincomplex reagent (1:100 dilution, 30 minutes,VectaStain Elite ABC Kit, Vector Labs), and 3,39-diaminobenzidine (DAB, 10 mg in 10 ml tris bufferfor 20 minutes) were used to visualize binding ofthe first antibody. Positive controls includedsections of colon cancer (CK-20), breast cancer(CK-7), lymph node (topoisomerase 2a), andnormal colon (DAS-1). Negative controls used thestudy sections without primary antibody. Immuno-reactivity was graded as positive when there wasmoderate or strong staining of at least 5% of themucosal cells of interest.

RESULTS

Cardiac mucosa was present in the cervicalesophagus in 10 of the 20 (50%) patients whounderwent biopsy. The indications for esopha-gectomy in these 10 patients were adenocarci-noma in 7, squamous cell carcinoma in 2, andstricture in 1. In 4 of the 7 patients with adeno-carcinoma, IM was identified in the esophagectomyspecimen. In all patients, the surgical margins priorto anastomosis showed no IM or cardiac mucosa.Figure 1 shows the endoscopic appearance ofan area of IM in the cervical esophagus in 1patient.

Seven of the 10 patients who showed cardiacmucosa on biopsy of the remnant cervical esoph-agus were males, and the median interval betweenesophagectomy and biopsy of cardiac mucosa was36 months (range, 9 months–42 years). Cardiacmucosa was found on the first endoscopy afteresophagectomy in 9 patients. These are the onlybiopsies that were available for these patients. Onepatient had 2 post-esophagectomy endoscopieswith biopsy. At the first endoscopy, performed 15months after esophagectomy, only squamous epi-thelium was present in the biopsy of the cervicalesophagus. At the second endoscopy, performed 9months later, cardiac mucosa was found in thecervical esophagus.

Four of the 10 patients with cardiac mucosa alsohad goblet cells characteristic of IM in the supra-anastomotic biopsies. One of the 4 patients, a

SurgerySeptember 2004

636 Lord et al

Table. Immunohistochemistry findings in different tissue types from the 10 patients withcardiac mucosa in the remnant cervical esophagus

Immunohisto-chemistrystain Purpose

Normalsquamousmucosa

Normalgastric antral

mucosaCardiacmucosa

Cardiac mucosawith IM

Cardiac mucosa withIM, dysplasia,adenocarcinoma

Number ofpatients

10 5 10 4 1

CK 7 Similarity toBarrett’sesophagusin the distalesophagus

No superficialstaining.Focal deepglandularstaining.

Superficialstainingin 2/5.Deep glan-dularstainingin 5/5.

Superficialstainingin all.Deep glan-dularstainingin all.

Superficialstaining in all.Deep glandularstaining in all.

Superficialstaining. Deepglandularstaining.

CK 20 No superficialstaining.No deepglandularstaining.

Superficialstainingin 2/5.No deepglandularstaining.

Superficialstainingin all. Noor little deepglandularstaining

Superficialstaining in all.No deepglandularstaining

Superficialstaining in all.No deepglandularstaining

DAS-1 No staining. Staining ofoxynticcells.

Weak cytoplasmicstaining ofcolumnarcells in 2/10.

Staining ofmucin withingoblet cells.

Staining ofmucin withingoblet cells.

Topoiso-merase2a

Cellularproliferation

Staining inbasal layer.

Staining ofbasal layerof cryptsand super-ficial glands.

Staining in allwith moreintensestainingin 5/10.

Staining in allwith moreintense stainingin goblet cellsthan in cardiacmucosa.

Intense stainingin dysplasticand tumor cells.

COX-2* Potential fordysplasia

No staining. Staining ofoxyntic cells.

Stainingin 2/10.

No staining. Staining.

ODC No staining. Staining ofoxyntic cells.

Weak stainingin 6/10.

Stainingin 3/4.

Staining.

CK, Cytokeratin; COX, cyclo-oxygenase; ODC, ornithine decarboxylase.*Excludes staining of inflammatory cells.

57-year-old man who had undergone esopha-gectomy at age 15 for an esophageal stricturesecondary to ingesting a coin at 9 months of age,had Barrett’s esophagus with dysplasia and anintramucosal adenocarcinoma in the remnantcervical esophagus 42 years after esophagectomy.

Immunohistochemistry. The immunohisto-chemistry results are shown in the Table. Rep-resentative images are shown in Figures 2, 3, 4, 5,and 6. CK7 staining of the cervical esophagus withcardiac mucosa, IM, and dysplasia was similar toBarrett’s esophagus in the distal esophagus and wasseen consistently in both the surface epitheliumand the deep glandular cells (Fig 2, A). Thestaining was typically more intense in areas of IMand dysplasia. In contrast to the columnar meta-plasia pattern, CK7 staining of squamous mucosashowed focal staining in the deep glandular cellswithout staining in the superficial squamous cells.

Also similar to the pattern seen in Barrett’sesophagus, CK20 staining of the cervical esophaguscontaining columnar mucosa occurred in thesurface cells, but there was little or no staining ofthe deep glandular cells (Fig 2, B). In contrast,squamous mucosa did not stain. Again as found inBarrett’s esophagus, DAS-1 antibody in cervicalcardiac mucosa stained intensely the mucin withingoblet cells and faintly the cytoplasm of some of thecolumnar cells (Fig 3). Squamous epithelium didnot stain. All the cervical cardiac mucosa, includingsections with IM and dysplasia, showed positivetopoisomerase 2a staining, which was typicallystronger at the bases of crypts and in glands thanin the surface epithelium (Fig 4). The staining wasstronger in goblet cells and dysplastic cells than incardiac columnar cells. Normal squamous epithe-lium showed topoisomerase 2a staining in the basallayer.

SurgeryVolume 136, Number 3

Lord et al 637

Fig 2. Immunohistochemistry images showing the ‘‘Barrett’s’’ cytokeratin 7/20 (CK 7/20) stainingpattern in a section of cardiac mucosa with IM biopsied from the supra-anastomotic cervical esophagus. A,CK 7 staining is seen in both the surface epithelium and the deep glandular cells. B, CK 20 staining, incontrast, is seen in the surface cells but not the deep glandular cells.

Fig 3. Immunohistochemistry image showing intenseDAS-1 staining of the mucin in goblet cells in a section ofcardiac mucosa with intestinal metaplasia in the supra-anastomotic cervical esophagus.

Fig 4. Immunohistochemistry image showing moder-ately intense topoisomerase 2a staining in a section ofcardiac mucosa with intestinal metaplasia in the supra-anastomotic cervical esophagus.

The staining pattern for COX-2 and ODC wassimilar. In the cervical esophagus, cytoplasmicCOX-2 epithelial staining was faint and was presentin only 2 of 10 patients with cardiac mucosa (Fig 5).It was not present in those with IM but was presentin the tumor cells in the patient with cancer.Intense ODC cytoplasmic epithelial staining waspresent in the cervical esophageal mucosa withdysplastic and tumor cells (Fig 6). Weak ODCcytoplasmic staining was present in 6 of 10 patientswith cervical cardiac mucosa. It also was present ingoblet cells in 3 of the patients with IM. Squamousmucosa did not stain for both antibodies, except

COX-2 stained inflammatory cells in the laminapropria.

DISCUSSION

Our study shows that cardiac mucosa can bean acquired, metaplastic epithelium.18,19 In theremnant cervical esophagus of patients who hadesophagectomy with esophagogastrostomy, an op-eration that provides a unique human model for denovo reflux disease, we found cardiac mucosa inthe supra-anastomotic biopsy in half the patientswho had biopsy at this site. This indicates that thecolumnar mucosa in these selected patients was

SurgerySeptember 2004

638 Lord et al

acquired after the operation because at the time ofsurgery, oxyntic mucosa was anastomosed to squa-mous mucosa. The stimulus for this metaplasticprocess is almost certainly acid reflux into the rem-nant esophagus. This is supported by a study fromSweden by Oberg et al in which pH probesmeasuring 24-hour acid exposure were placedin the cervical esophagus 1 cm above the eso-phagogastric anastomosis in patients who hadesophagectomy.20 All patients with columnar mu-cosa in the remnant esophagus had abnormal acidexposure, and there was a direct correlation be-tween the length of the metaplastic segment andthe percentage of time the cervical esophagus wasexposed to a pH less than 4.0.20 Interestingly, therewas no association between the presence of cervicalcolumnar metaplasia and exposure to bilirubin,a marker for non-acid reflux, although the fewpatients with IM in the cardiac mucosa hadabnormal esophageal exposure of both acid andbilirubin. These results are similar to those foundin the distal esophagus, as indicated by a study thatfound that similar proportions of patients withcardiac mucosa and IM had abnormal esophagealacid exposure (79% and 83%, respectively), butabnormal esophageal bilirubin exposure was morefrequent in the patients with IM.9 Based on theseresults, the hypothesis may be advanced that acidreflux is sufficient to stimulate the development ofcardiac mucosa, but non-acid reflux may beparticularly important for the development of IM.

We found that cardiac mucosa in the remnantesophagus shares some definitive characteristics

Fig 5. Immunohistochemistry image showing only faintcytoplasmic COX-2 epithelial staining in a section ofcardiac mucosa in the supra-anastomotic cervical esoph-agus. Stronger COX-2 staining is seen in some in-flammatory cells in the lamina propria.

with cardiac mucosa and Barrett’s esophagus in the

distal esophagus. Most importantly, we found a CK7/20 expression pattern similar to that of Barrett’sesophagus23,24 in a majority of the patients withcardiac mucsosa with or without IM and dysplasia.In contrast, a ‘‘non–Barrett’s-like’’ CK pattern waspresent in specimens of normal squamous and gas-tric mucosa. The similarity between supra-anasto-motic cardiac mucosa, distal esophageal cardiacmucosa, and Barrett’s esophagus25 supports thelikelihood that IM may result from the develop-ment of goblet cells within the cardiac mucosa.Further, the ‘‘Barrett’s’’ CK staining pattern alsohas been shown to be associated with a refluxetiology.26 This study further supports the likeli-hood of a reflux etiology for cardiac mucosa in theremnant esophagus after esophagectomy.

The DAS-1 MAb reacts against colonic epithelialcells, but not with normal small-bowel enterocytesor esophageal mucosa.27 The antibody does reactintensely with an unknown epitope in Barrett’sesophagus, particularly the incomplete (II and III)type of IM.27,28 In the present study, intense DAS-1staining of the mucin in goblet cells was observed incervical esophagus with intestinalized cardiac mu-cosa. Further, the pattern of DAS-1 staining in boththe IM and cardiac columnar cells was similar tothat found in the lower esophagus.24

Cellular proliferation was assessed withtopoisomerase 2a immunohistochemistry. As ex-pected, the proliferative zones in normal gastric andesophageal mucosa, dysplastic Barrett’s cells, andcancer cells were strongly positive for topoisomerase2a. There was also evidence of increased prolifera-tion in some of the cardiac mucosa, suggesting thepossibility of disease progression in some patients

Fig 6. Immunohistochemistry image showing moderatelystrong ODC cytoplasmic staining in an area of low gradedysplasia in the supra-anastomotic cervical esophagus.

SurgeryVolume 136, Number 3

Lord et al 639

with non-IM columnar metaplasia. The proteinexpressions of ODC and COX-2 were also exam-ined. Both of these genes have putative roles intumorigenesis. ODC is the initial and rate-limitingenzyme in the biosynthetic pathway of polyamines,which have essential roles in cell growth and dif-ferentiation. Increased ODC protein and mRNAexpression have been reported in Barrett’s esoph-agus and adenocarcinoma.29 Similarly, the pros-taglandin synthesis enzyme COX-2, which hasbeen implicated as a fundamental factor in manytumorigenic processes,30 is upregulated in someBarrett’s tissues.31,32 The low ODC and COX-2expressions found in cardiac mucosa in this studysupport the clinical observation that cardiac mucosahas very little malignant potential.

Cardiac mucosa is frequently present at thegastroesophageal junction in adults in Westernsociety, raising the possibility that we have merelytaken biopsies of long-standing cardiac mucosafrom the distal, gastric side of the anastomosis.This possibility is extremely unlikely because weincluded only patients in whom it was noted thatthe biopsies were from above the anastomosis,and because at least 2 cm of proximal stomach, andthus the gastroesophageal junction and all cardiacmucosa, was resected at the time of esophagectomy.The possibility that columnar mucosa was presentin the cervical esophagus before esophagectomy isalso excluded because only normal squamousepithelium was seen at the proximal resectionmargin at the time of operation. Furthermore, thesupra-anastomotic biopsies were performed for thespecific purpose of conducting studies such as thepresent one. In this respect, although the methodsof data collection and specimen retrieval make thisa retrospective study, the biopsies were collectedprospectively.

In summary, cardiac mucosa can be an acquired,metaplastic epithelium. It is likely that it developscommonly after esophagectomy with gastric re-construction and its presence very likely signifies atleast some reflux into the esophagus. This obser-vation supports the hypothesis that cardiac mucosaat the gastroesophageal junction in unoperatedindividuals, despite the prevalence of this finding,is also an acquired epithelium. CK 7/20 character-ization shows that the cardiac mucosa in theremnant esophagus is similar to Barrett’s in thedistal esophagus. This supports the possibility thatIM could arise from cardiac mucosa. DysplasticBarrett’s and adenocarcinoma developed in 1 pa-tient in our study 42 years after esophagectomy.This observation does not indicate the need forroutine post-esophagectomy surveillance of the

remnant esophagus, except perhaps in long-termsurvivors.

REFERENCES

1. Der R, Tsao-Wei DD, DeMeester T, Peters J, Groshen S, LordRV, Chandrasoma P. Carditis: a manifestation of gastro-esophageal reflux disease. Am J Surg Pathol 2001;25:245-52.

2. Hayward J. The lower end of the oesophagus. Thorax1961;16:36-55.

3. Lord RV. Norman Barrett, ‘‘doyen of esophageal surgery’’.Ann Surg 1999;229:428-39.

4. Oberg S, Peters JH, DeMeester TR, Chandrasoma P, HagenJA, Ireland AP, et al. Inflammation and specialized intestinalmetaplasia of cardiac mucosa is a manifestation of gastro-esophageal reflux disease. Ann Surg 1997;226:522-30.

5. Chandrasoma PT, Der R, Ma Y, Dalton P, Taira M. Histologyof the gastroesophageal junction: an autopsy study. Am JSurg Pathol 2000;24:402-9.

6. Zhou H, Greco MA, Kahn E. Origin of cardiac mucosa.Ontogenic considerations. Mod Pathol 1999;12:499A.

7. Csendes A, Maluenda F, Braghetto I, Csendes P, HenriquezA, Quesada MS. Location of the lower oesophagealsphincter and the squamous columnar mucosal junctionin 109 healthy controls and 778 patients with differentdegrees of endoscopic oesophagitis. Gut 1993;34:21-7.

8. Csendes A, Smok G, Burdiles P, Sagastume H, Rojas J,Puente G, Q, et al. �Carditis�: an objective histologicalmarker for pathologic gastroesophageal reflux disease. DisEsoph 1998;11:101-5.

9. Oberg S, Peters JH, DeMeester TR, Lord RV, Johansson J,DeMeester SR, et al. Determinants of intestinal metaplasiawithin the columnar-lined esophagus. Arch Surg 2000;135:651-5.

10. Chandrasoma P. Pathophysiology of Barrett’s esophagus.Semin Thorac Cardiovasc Surg 1997;9:270-8.

11. Csendes A, Smok G, Flores N, Rojas J, Quiroz J, HenriquezA. Comparison of clinical, endoscopic and functionalfindings in patients with intestinal metaplasia at the cardia,carditis and short-segment columnar epithelium of thedistal esophagus with and without intestinal metaplasia. DisEsoph 2000;13:61-8.

12. Chen YY, Antonioli DA, Spechler SJ, Zeroogian JM, GoyalRK, Wang HH. Gastroesophageal reflux disease versusHelicobacter pylori infection as the cause of gastric carditis.Mod Pathol 1998;11:950-6.

13. Spechler SJ. The role of gastric carditis in metaplasia andneoplasia at the gastroesophageal junction. Gastroenterol-ogy 1999;117:218-28.

14. Kilgore SP, Ormsby AH, Gramlich TL, Rice TW, Richter JE,Falk GW, et al. The gastric cardia: fact or fiction? Am JGastroenterol 2000;95:921-4.

15. Kim SL, Waring JP, Spechler SJ, Sampliner RE, Doos WG,Krol WF, et al. Diagnostic inconsistencies in Barrett’sesophagus. Department of Veterans Affairs Gastroesopha-geal Reflux Study Group. Gastroenterology 1994;107:945-9.

16. Dobhan R, Castell DO. Normal and abnormal proximalesophageal acid exposure: results of ambulatory dual-probepH monitoring. Am J Gastroenterol 1993;88:25-9.

17. Johansson J, Johnsson F, Groshen S, Walther B. Pharyngealreflux after gastric pull-up esophagectomy with neck andchest anastomoses. J Thorac Cardiovasc Surg 1999;118:1078-83.

18. Hamilton SR, Yardley JH. Regeneration of cardiac typemucosa and acquisition of Barrett mucosa after esophago-gastrostomy. Gastroenterology 1977;72:669-75.

SurgerySeptember 2004

640 Lord et al

19. Lindahl H, Rintala R, Sariola H, Louhimo I. CervicalBarrett’s esophagus: a common complication of gastric tubereconstruction. J Pediatr Surg 1990;25:446-8.

20. Oberg S, Johansson J, Wenner J, Walther B. Metaplasticcolumnar mucosa in the cervical esophagus after esoph-agectomy. Ann Surg 2002;235:338-45.

21. Paull A, Trier JS, Dalton MD, Camp RC, Loeb P, Goyal RK.The histologic spectrum of Barrett’s esophagus. N Eng JMed 1976;295:476-80.

22. Chandrasoma PT, Lokuhetty DM, DeMeester TR, BremmerCG, Peters JH, Oberg S, et al. Definition of histopathologicchanges in gastroesophageal reflux disease. Am J SurgPathol 2000;24:344-51.

23. Ormsby AH, Goldblum JR, Rice TW, Richter JE, Falk GW,Vaezi MF, et al. Cytokeratin subsets can reliably distinguishBarrett’s esophagus from intestinal metaplasia of thestomach. Hum Pathol 1999;30:288-94.

24. Glickman JN, Wang H, Das KM, Goyal RK, Spechler SJ,Antonioli D, et al. Phenotype of Barrett’s esophagus andintestinal metaplasia of the distal esophagus and gastro-esophageal junction: an immunohistochemical study ofcytokeratins 7 and 20, Das-1 and 45 MI. Am J Surg Pathol2001;25:87-94.

25. DeMeester SR, Wickramasinghe KS, Lord RV, Friedman A,Balaji NS, Chandrasoma PT, et al. Cytokeratin and DAS-1immunostaining reveal similarities among cardiac mucosa,

CIM, and Barrett’s esophagus. Am J Gastroenterol 2002;97:2514-23.

26. Couvelard A, Cauvin JM, Goldfain D, Rotenberg A,Robaszkiewicz M, Flejou JF, et al. Cytokeratin immunoreac-tivity of intestinal metaplasia at normal oesophagogastricjunction indicates its aetiology. Gut 2001;49:761-6.

27. Das KM, Prasad I, Garla S, Amenta PS. Detection of a sharedcolon epithelial epitope on Barrett epithelium by a novelmonoclonal antibody. Ann Intern Med 1994;120:753-6.

28. Griffel LH, Amenta PS, Das KM. Use of a novel monoclonalantibody in diagnosis of Barrett’s esophagus. Dig Dis Sci2000;45:40-8.

29. Brabender J, Lord RV, Danenberg KD, Metzger R, SchneiderPM, Uetake H, et al. Upregulation of ornithine decarboxylasemRNA expression in Barrett’s esophagus and Barrett’s-associated adenocarcinoma. J Gastrointest Surg 2001;5:174-82.

30. Turini ME, DuBois RN. Cyclooxygenase-2: a therapeutictarget. Annu Rev Med 2002;53:35-57.

31. Wilson KT, Fu S, Ramanujam KS, Meltzer SJ. Increasedexpression of inducible nitric oxide synthase and cyclo-oxygenase-2 in Barrett’s esophagus and associated adeno-carcinomas. Cancer Res 1998;58:2929-34.

32. Shirvani VN, Ouatu-Lascar R, Kaur BS, Omary MB, Triada-filopoulos G. Cyclooxygenase 2 expression in Barrett’sesophagus and adenocarcinoma: ex vivo induction by bilesalts and acid exposure. Gastroenterology 2000;118:487-96.