endoscopic evaluation of gastro-esophageal reflux disease

YALE JOURNAL OF BIOLOGY AND MEDICINE 72 (1999), pp. 93-100.Copyright © 2000. All rights reserved.

Endoscopic Evaluation of Gastro-EsophagealReflux Disease

David ArmstrongaDivision of Gastroenterology, McMaster University Medical Centre, Hamilton, Ontario,Canada

Endoscopy is, currently, the initial investigation ofchoicefor the investigation ofgastroesophagealreflux disease (GERD) in clinical practice and clinical research. Erosion severity is predictive ofapatient's response to therapy and of the likelihood of relapse after therapy. It is, therefore, impor-tant to grade the severity oferosive reflux esophagitis, particularly in the context of clinical trials.

The Savary-Miller endoscopic classification system is used widely but usage and interpretation arevery variable. The "MUSE" (metaplasia [M1, ulceration [U], stricturing [S] and erosions [EJ) clas-sification provides clear definitions of the relevant endoscopic features, and it is based on a stan-dardized report form, which allows the endoscopist to make a clear record of esophagitis severity.Recent studies confirm that endoscopists can identify erosions or mucosal breaks, ulcers, strictures,and metaplasia reproducibly. The "L.A." (Los Angeles) classification describes four grades ofesophagitis severity (A to D), based on the extent ofesophageal lesions known as "mucosal breaks,"but it does not record the presence or severity ofother GERD lesions. Thus, for patients with "com-plicated" reflux disease, the "MUSE" classification offers a more comprehensive description ofesophagitis severity.

Endoscopy is not universally applicable: 40 to 60 percent ofpatients with typical reflux symptomsdo not have esophageal erosions and are now considered to have "endoscopy negative reflux dis-ease" (ENRD). Thus, endoscopy is not the final arbiter as to a diagnosis of reflux disease, and it isnot, therefore, a necessary prerequisite to therapy. Endoscopy is indicated at first presentation forpatients with alarm symptoms referable to the upper gastrointestinal tract. It has also been proposedthat all patients with chronic GERD should have a "once-in-a-lifetime" endoscopy; in the absenceof Barrett's esophagus or other complications, no follow-up is required unless the patient's symp-toms change significantly. A surveillance program with multiple biopsies should be instituted ifthereis evidence ofBarrett's esophagus. Endoscopic evaluation should document the presence and extentof esophageal erosions using the L.A. or MUSE classification systems; complications should alsobe documented and may be recorded using the MUSE classification. Non-erosive changes such aserythema may be ignored on the basis ofpresent evidence, and there are no clear data to supportthe use ofendoscopic biopsies for the diagnosis ofGERD.

a To whom all correspondence should be addressed: David Armstrong, Division ofGastroenterology, Room HSC-4W8, McMaster University Medical Centre, 1200 MainStreet West, Hamilton, Ontario, L8N 3Z5. Tel.: 905-521-2100, Ext. 76404; Fax: 905-521 -5072; E-mail: [email protected] Abbreviations: GERD, gastroesophageal reflux disease; ENRD, endoscopy negativereflux disease; M, metaplasia; U, ulceration; S, stricturing; E, erosions.

93

94 Armstrong: Endoscopic evaluation of GERD

INTRODUCTION

For many years, esophageal erosionshave been considered the sine qua non ofreflux esophagitis and, hence, of gastro-esophageal reflux disease (GERD)b. Atone end of the spectrum, lesions such asulceration, stricture formation, andBarrett's esophagus are recognized to bemore severe manifestations or complica-tions of GERD, whereas "minor" changessuch as erythema, edema, and friabilityare considered to be markers of minoresophageal damage.

However, since the introduction ofhistamine H2-receptor antagonists, theresolution of erosive esophagitis has beenthe most common primary outcome mea-sure used for therapeutic trials in GERD.Patients with more severe lesions, or com-plications of GERD, are rarely enrolled intherapeutic trials, while patients withminor endoscopic changes may be

enrolled, but their results are often exclud-ed from the final analysis.

DIAGNOSIS OF EROSIVE REFLUXESOPHAGITIS

Erosive esophagitis is now consid-ered, by many, to be synonymous withGERD, and the diagnosis of GERD isoften, therefore, predicated on the obser-vation of esophageal erosions. Of thediagnostic techniques available for inves-tigating GERD, upper gastrointestinalradiology has low sensitivity and speci-ficity for the detection of erosions andBarrett's esophagus, although it is some-what better for the diagnosis of ulcerationor stricturing. Esophageal acid exposure,quantified by ambulatory pH monitoring,correlates well with the degree ofesophageal damage, but this techniquecannot be used to diagnose erosive

% patients with recurrent esophagitis60

No Yes

Recurrent SymptomsFigure 1. Recurrence of erosive esophagitis is rare in the absence of recurrent refluxsymptoms [6].

Armstrong: Endoscopic evaluation of GERD 95

Metaplasia Ulcer Stricture Erosions

0 ofEl U05E o E0Q1:absent absent absent absent

'C _ __ _ _ .... __ _ __ _ . .. ....

1i1z ~~~Ul SI13 ElE

*E "ifingers" I discrete ulcer > 9 mm (standard only on peaks

0 +1- "islands" endoscope passes) of folds

SC&SI ~~IrtP~40t22 °" U20* S2 °0 E2a

uzcircumferential 22 discrete or 9adm (standard confluent: on &confluent ulcers enot pass)de between folds

Hiatus Hernia YesO Nor )

Figure 2. A pictorial representation of the "MUSE" classification system, which canbe used as the basis for an endoscopic report form. Metaplasia, Ulceration, Strictureformation and Erosions (M.U.S.E.) are assessed and graded independently according totheir degree of severity (0: absent, 1: mild, 2: severe). For each lesion type, the appropriatebox is ticked and, if relevant, the extent of a lesion such as columnar metaplasia may bemarked with reference to the diaphragmatic hiatus. Examples of some MUSE classifica-tions and the corresponding endoscopic appearances are: M2UoS1 E,: active (with ero-sions) peptic stricture with a diameter greater than 9 mm, situated at e upper pole of acircumferential area of columnar metaplasia; M2U2S E large, confluent ulcers associatedwith a circumferential area of columnar metaplasia; 1AJOSOES : erosions affecting only onefold. (Courtesy of Ref. [12]). 0a0i 0 V

esophagitis per se, and it certainly cannotbe used to document resolution of erosiveesophagitis. Upper gastrointestinalendoscopy has, therefore, become the pre-dominant diagnostic modality for GERDin clinical and research practice.

In the context of managing erosiveesophagitis, endoscopy has significantadvantages: it can be used to document thepresence and extent of esophageal ero-sions as well the presence of ulceration,stricturing, and Barrett's esophagus orcolumnar metaplasia. It must be supple-mented by biopsy for the diagnosis of

metaplasia and the potential sequelae ofdysplasia and neoplasia; endoscopic biop-sy, with or without brushings, also facili-tates the diagnosis of other infectiveesophagitides and malignancies. However,although changes such as elongation of thestromal papillae [1], a neutrophil infiltrate,an eosinophil infiltrate [2, 3] and "bal-loon" cells [4] have all been described inassociation with gastroesophageal refluxdisease, the sensitivity, and specificity ofthese findings for the diagnosis of GERDhave not been defined, and the role of his-tology, therefore, remains undefined.


In general, the complications ofGERDare relatively infrequent, and the most fre-quent identifiable lesion remains theesophageal erosion or "touche peptique"described by Savary [5]. For patients witherosive esophagitis, the recurrence ofsymptoms after therapy correlates well withthe recurrence of esophageal erosions(Figure 1) [6]. Neither the extent nor theseverity of erosions correlates directly withsymptom severity in an individual patient;however, erosion severity is predictive of apatient's probable response to therapy andalso of the likelihood of relapse after cessa-tion of therapy, unlike the presence ofminor, non-erosive changes [7, 8]. It is,therefore, important to grade the severity oferosive reflux esophagitis, particularly inthe context of clinical trials.

ENDOSCOPIC GRADING OFESOPHAGITIS SEVERITY

The most widely used classificationsystem is that described first by Savary &Miller, in which disease severity is gradedaccording to the extent of esophageal ero-sions [9]. The Savary-Miller classificationrecognizes also that complications ofGERD are important but, in both versionsdescribed by the authors [9, 10], they aregrouped together as high-grade esophagi-

tis (Grade 4 or Grade 5), regardless ofwhether or not there are concomitant ero-sions. In consequence, it is very difficult touse the Savary-Miller classification to doc-ument the healing of erosive esophagitis inpatients with complicated disease. Therecognition that metaplasia (M), ulceration(U), stricturing (S) and erosions (E) canoccur independently of each other led tothe formulation of the "MUSE" classifica-tion; in this case, the classification of ero-sion severity (Eo to E2) is based on theSavary-Miller classification, but it is sup-plemehted by independent gradings ofseverity for ulcers (UO to U2), strictures (SOto S2) and metaplasia (Mo to M2) (Figure2). This classification provides clear defin-itions of the different endoscopic featuresof GERD and is based on a pictorial reportform, which allows the endoscopist tomake a clear record of esophagitis severity[11, 12].

In recent years, it has become clearthat the profusion of classification systemshas made it increasingly difficult to com-pare the results of different therapeutic tri-als in GERD. This prompted the formationof an International Working Groupcharged with developing a standardizedendoscopic esophagitis classification sys-tem, which would be generally acceptablein clinical or research practice. In an ini-tial step, the group identified endoscopic

Table 1. Diagnostic agreement between expert endoscopists (kappa values) for theidentification of "mucosal" breaks [13].

"Mucosal break"Length (mm) Exudate kappa value

<5mm No 0.49<5mm Yes 0.88>5mm No 0.29>5mm Yes 0.55

Affecting only one fold 0.84Circumferential 0.59


kappa scores1-

0.8

0.6 Vt0.4

0.2

Metaplasia Ulcer Stricture

E N E N E N

Figure 3. Kappa scores for evaluation of endoscopic features of complicated refluxesophagitis by 29 experienced endoscopists {E) and 13 novices (N), based onreporting of endoscopic video recordings and slide material. A kappa score > 0.70(dashed, horizontal line). is considered to indicate god agreement between observers. Withpermission from Ref. [13].

features, associated with GERD, whichcould be recognized reproducibly byendoscopists. These studies, based onhigh-quality 35 mm slides and videorecordings of esophageal endoscopies,indicate that endoscopists can identifyreliably, erosions or mucosal breaks, par-ticularly if they have an overlying exudate(Table 1) as well as ulcers, strictures andmetaplasia (Figure 3). This study suggest-ed also that experienced endoscopistscould identify reliably, some of the minorchanges (erythema, edema, friability)attributable to reflux [13]. In conjunctionwith this, the group formulated a new clas-sification system - the "L.A." classifica-tion (first presented at the 1992 WorldCongress of Gastroenterology in LosAngeles) - which described four gradesof esophagitis severity (Grade A to GradeD) based on the extent of mucosal breaks.The term "mucosal break" was coined to

encompass erosions as well circumscribedareas of erythema since it was felt therewere no clear criteria for distinguishingbetween a "denuded" erosion and an ero-sion covered by "slough" or an "exudate."The term "mucosal break" is also intendedto include ulcers since there are no definedcriteria for determining when an erosiondeepens sufficiently to become an ulcer.Validation studies with the current versionof the L.A. classification system indicatethat the higher grades of esophagitis areassociated with lower healing rates andhigher relapse rates. Since it is based onendoscopic features, which can be identi-fied reproducibly by endoscopists, theL.A. classification has the potential to be agenerally-accepted, standardized classifi-cation system for erosive esophagitis.However, the L.A. classification does notrecord the presence or severity of otherGERD lesions, which are, themselves,


identified reliably by endoscopists; thus inclinical practice and in therapeutic trialsthat include patients with "complicated"reflux disease, the MUSE classificationoffers a more comprehensive descriptionof esophagitis severity.

LIMITATIONS OF ENDOSCOPY

Despite its importance in clinical andresearch practice, endoscopy is not, how-ever, universally applicable. It is inappro-priate for the diagnosis of hiatal herniationor esophageal dysmotility. More impor-tantly, there is an increasing recognition of"endoscopy negative reflux disease"(ENRD): recent studies suggest that 40 to60 percent of patients with typical refluxsymptoms do not, in fact, have esophagealerosions [14, 15]. For some physicians,this represents proof final that thesepatients do not have GERD and, in somejurisdictions, that these patients do notmerit therapy with proton pump inhibitors.Despite this, many such patients will haveevidence of a temporal correlationbetween reflux episodes and symptomsduring esophageal pH monitoring studies,and their symptoms will often be relievedby medical anti-reflux therapy [14, 15].Thus, it is becoming clear that the clinicalspectrum of GERD does include ENRDand that endoscopy is not, therefore, thefinal arbiter as to a diagnosis of reflux dis-ease.

ROLE OF ENDOSCOPY FOREVALUATION OF GERD

What then is the place of endoscopyin clinical practice? Endoscopic confirma-tion of esophageal erosions is highly spe-cific for a diagnosis of GERD, but it haslow sensitivity, and there is, therefore, thedanger that endoscopy alone will underes-timate greatly the prevalence of GERD.

An outcomes study has shown [16] thatendoscopy leads to a significant change intherapy for patients with reflux symptoms,but the change was most marked inpatients who had erosive esophagitis; it isnot clear that the failure to alter therapy inpatients with ENRD was appropriate. Ithas been argued that endoscopy is not nec-essary to make a diagnosis of GERD in apatient with typical symptoms unless thereis a suspicion of serious underlying dis-ease based on the patient's age or alarmsymptoms such as weight loss, dysphagia,anemia or gastrointestinal blood loss. Ifreimbursement policies mandate a con-firmed diagnosis of erosive esophagitisbefore effective antisecretory therapy canbe started, endoscopy should be performedearly, and, if possible, the patient shouldstop acid antisecretory or prokinetic thera-py at least one to two weeks prior toendoscopy. If erosive esophagitis is con-firmed, there seems to be no reason torepeat endoscopy simply to confirm heal-ing since there is a good correlationbetween symptoms and the recurrence oferosions [6]. In the majority of patients,GERD is a chronic, relapsing condition,and the only indication for repeatendoscopy would be a significant changeor worsening of the patient's symptoms.

For the present, the other major indi-cation for endoscopy is the exclusion ofBarrett's esophagus. There is continuingcontroversy over the advisability of sur-veillance follow-up for patients withBarrett's esophagus, but a cost-modelingstudy suggests that it is cost-effective, ifthe incidence of carcinoma in Barrett'sexceeds 1 percent [17]; this provides sup-port for current recommendations thatendoscopic surveillance be conductedevery two years in patients with confirmedBarrett's esophagus. The natural history ofBarrett's esophagus is still unclear but datafrom Olmsted County [18] suggest that thearea of involved mucosa does not progresssignificantly with time over follow-up


periods of three to eight years. Thus, onpresent evidence, it would seem reason-able to suppose that a patient who does nothave Barrett's esophagus at the time ofdiagnosis is unlikely to develop metaplasiasubsequently. This has led to the conceptof a "once-in-a-lifetime" endoscopy forpatients with GERD. Under this model,there would be no need to perform anendoscopy before embarking on therapybut one examination should be performed- during ongoing therapy - to excludeBarrett's and to minimize the likelihood ofconfusion between the histologicalchanges of acute inflammation and thoseof metaplasia or dysplasia.

CONCLUSION

In the long run, the role of endoscopywill change as more is learned about thenatural history of symptomatic GERD andits relationship to erosive esophagitis,about the natural history of Barrett'sesophagus and its progression toesophageal adenocarcinoma and about theimportance of carditis and its relationshipto the increasing incidence of adenocarci-noma of the cardia.

Endoscopic evaluation of a patientwith reflux symptoms reassures the patientand the physician that no complications ofGERD or other potentially serious condi-tion has been overlooked. It also providessome indication as to the most appropriatetherapy, the likelihood of response to ther-apy, and the likelihood of subsequentrelapse. Against these advantages must beweighed the potential financial cost of theprocedure to the patient and the health caresystem. Endoscopy is relatively costly, butthe absolute cost varies considerably;complications from upper gastrointestinalendoscopy are very infrequent, but theyhave been recorded. The major disadvan-tage of endoscopy arises from the notion

that GERD is synonymous with erosiveesophagitis; this may lead to a failure todiagnose GERD and, hence, to inappropri-ate or inadequate therapy for ENRD.

For the present, endoscopy is not anecessary prerequisite to therapy for typi-cal reflux symptoms, but it is indicated atfirst presentation for patients with alarmsymptoms referable to the upper gastroin-testinal tract. Whenever it is performed,endoscopic evaluation should documentthe presence and extent of esophageal ero-sions using the L.A. or MUSE classifica-tion systems; the presence and extent ofcomplications should also be documentedand may be recorded using the MUSEclassification. Non-erosive changes suchas erythema may be ignored on the basisof present evidence and there are no cleardata to support the acquisition of endo-scopic biopsies for the diagnosis ofGERD. However, a diagnosis of Barrett'sesophagus must be confirmed histological-ly and endoscopic biopsies are, therefore,essential if the endoscopic appearances ofthe distal esophagus suggest the develop-ment of columnar metaplasia. An endo-scopic surveillance program with multiplebiopsies should be instituted if there is his-tological evidence of Barrett's esophagus,provided that the patient would be willingand able to undergo further therapy in theevent that dysplastic or neoplastic lesionswere identified. Endoscopic follow-up isnot required for uncomplicated erosivereflux esophagitis or ENRD unless therehas been a significant change in thepatient's symptoms.

References

1. Ismail-Beigi, F., Horton, P.F., and Pope,C.E. Histological consequences of gastroe-sophageal reflux in man. Gastroenterology58:163-174, 1970.

2. Weinstein, W.M., Bogoch, E.R., andBowes, K.L. The normal human


esophageal mucosa: a histological reap-praisal. Gastroenterology 68:40-44, 1975.

3. Seefeld, U., Kreijs, G.J., Siebenmann, R.E.,and Blum, A.L. Esophageal histology ingastroesophageal reflux. Morphometricfindings in suction biopsies. Am. J. Dig.Dis. 22:956-964, 1977.

4. Jessurun, J., Yardley, J.H., Giardiello, F.M.,and Hamilton, S.R. Intracytoplasmic plas-ma proteins in distended esophageal squa-mous cells (balloon cells). Mod. Pathol.1:175-181, 1988.

5. Savary, M. Les hernies hiatales non com-pliquees. Endoscopie, maladie peptiqueoesophagienne et gastrites hemiares.Medicine et Hygiene 26:789-791, 1968.

6. Blum, A.L., Adami, B., Bouzo, M.H.,Brandstatter, G., Fumagalli, I., Galmiche,J.P., Hebbeln, H., Hentschel, E.,Huttemann, W., and Schutz, E. Effect ofcisapride on relapse of esophagitis: Amultinational, placebo-controlled trial inpatients healed with an antisecretory drug.The Italian Eurocis Trialists. Dig. Dis. Sci.38: 51-560, 1993.

7. Schule, A., Brandli, H., Pelloni, S., Koelz,H.R., Pirozynski, W.J., and Blum, A.L.[Endoscopic diagnosis of oesophagitis:problems of differentiation from normal(German)]. Deut. Med. Wochenschrift.102:606-609, 1977.

8. Leu, H., Schiile, A., Brandli, H., Pelloni, S.,and Blum, A.L. [Are loss of glistening ofnormal colour and increased friability nor-mal aspects of the oesophagus in old age?(German)]. Zeitschrift. fur. Gastro-enterologie 16:417-421, 1978.

9. Savary, M. and Miller, G. The oesophagus:Handbook and atlas of endoscopy.Solothurn: Verlag Gassmann AG; 1978.

10. Ollyo, J.-B., Monnier, P., Fontolliet, C., andSavary, M. The natural history, prevalence

and incidence of reflux oesophagitis. Gullet3(suppl):3-10, 1993.

11. Armstrong, D., Monnier, P., Nicolet, M.,Blum, A.L., and Savary, M. Endoscopicassessment of oesophagitis. Gullet 1:63-67,1991.

12. Armstrong, D. and Blum, A.L. Endoscopicclassification of reflux esophagitis by theMUSE system. Progress in Surgery 23:43-50, 1997.

13. Armstrong, D., Bennett, J.R., Blum, A.L.,Dent, J., De, D.F., Galmiche, J.P., Lundell,L., Margulies, M., Richter, J.E., Spechler,S.J1 Tytgat, G.N., and Wallin, L. The endo-scopic assessment of esophagitis: aprogress report on observer agreement.Gastroenterology 111:85-92, 1996.

14. Lauritsen, K. Management of endoscopy-negative reflux disease: progress withshort-term treatment. Aliment. Pharmacol.Ther. l1 (suppl 2):87-92, 1997.

15. Lundell, L. New information relevant tolong-term management of endoscopy-nega-tive reflux disease. Aliment. Pharmacol.Ther. 11(suppl 2):93-98, 1997.

16. Ellis, K.K., Oehlke, M., Helfand, M., andLieberman, D. Management of symptomsof gastroesophageal reflux disease: doesendoscopy influence medical manage-ment? Am. J. Gastroenterol. 92:1472-1474,1997.

17. Provenzale, D., Kemp, J.A., Arora, S., andWong, J.B. A guide for surveillance ofpatients with Barrett's esophagus. Am. J.Gastroenterol. 89:670-680, 1994.

18. Cameron, A.J. and Lomboy, C.T. Barrett'sesophagus: age, prevalence, and extent ofcolumnar epithelium. Gastroenterology103:1241-1245, 1992.

endoscopic evaluation of gastro-esophageal reflux disease

Documents