Critical evaluation of current classification systems for reflux esophagitis?

H.W. Boyce Jr (Tampa)

Accurate endoscopic diagnosis of esophagitis depends upon the existence of visually identifiable changes in the esophageal mucosa that can be confirmed upon biopsy in all cases. Any mucosal alteration that does not precisely correlate with histology is not reliable for a diagnosis that is to be used as a basis for therapeutic decisions, or for evaluation of response to any one or combination of therapies. Many authorities agree that capillary congestion and erythema/hyperemia are visible mucosal changes that do not correlate with histology and therefore should not be used as criteria for endoscopic diagnosis of esophagitis. Prior grading systems for severity of esophagitis that included erythema/hyperemia as endoscopic criteria should not be used.

In past years, the lack of a suitable standard grading system for the lesions of reflux esophagitis has led to confusion in patient care and clinical research studies. The documentation of unequivocal epithelial defects typical for reflux is sufficient for an accurate diagnosis of reflux esophagitis. However, neither the absence of visible mucosal defects nor the complete healing of previously documented mucosal injury confirmed by endoscopy can be considered indicative of normal histology. When either clinical or research requirements dictate the need for documentation of complete mucosal healing, only a biopsy can provide confirmation. The fact that

typical symptoms of reflux esophagitis can occur with either microscopic abnormalities or totally normal mucosal histology has been well established. In the absence of unequivocal mucosal defects (erosions), an adequate biopsy is the only method to determine presence or absence of esophagitis.

The accurate diagnosis of reflux esophagitis is possible by endoscopy alone only if there are unequivocal signs of mucosal erosion. The severity of esophagitis by endoscopy or histology does not correlate well with symptoms. An endoscopic grading system for reflux esophagitis can be used only as a measure of the visible extent of mucosal injury, starting at the squamocolumnar mucosal junction and extending proximally into the esophagus either as continuous or "stepping stone" erosions (islands of mucosal erosion surrounded entirely by squamous epithelium). The complications of erosive esophagitis (stricture, ulcer and columnar-lined esophagus) are included in the highest grade of severity. The assumption has been, and it probably is true, that the more extensive the area of mucosal erosion circumferentially about the squamocolumnar mucosal junction, the greater the risk of evolution into stricture, deep ulceration or a columnar-lined esophagus.

A major question is how best to define the degree of erosive disease so that the interpretation of various grades is consistent between endoscopists. The answer may be somewhere between the simple Savary-Miller system proposed in 1977 (Table 1) and the "MUSE" system (M = metaplasia; U = ulcer; S = stricture; E = erosions) proposed in 1992 (Table 3) [1,2].

The need for a grade 0 indicative of normal or no esophageal erosions is uncertain. The only apparent advantage seems to be that it serves to require a definitive response from the endoscopist when grading is a component of the data required for clinical research. Grade 0 serves no purpose for general clinical use in endoscopy reports. Several investigators have also included a grade I indicating hyperemia/erythema with friability (Tables 1 and 2). Interpretation of hyperemia/ erythema or friability is subject to wide interobserver variation, and in many cases is overinterpreted. Hence, this definition of grade I does not seem proper for an objective endoscopic grading system.

Most will agree that there is a need for uniformity and precision with any grading system used for evaluating response to therapy in a clinical research protocol. However, is it necessary to apply the same detailed criteria needed for research purposes to the routine clinical diagnosis of esophagitis? Is there a need for general agreement on separate grading systems for clinical and research purposes? It is probably best to have a single grading language to be used between countries, clinicians and investigators. The most objective and simple system would be the best.

The mild, moderate, severe and complicated grades of the original Savary-Miller system seem quite easy to apply for clinical purposes with no obvious need for modification. The proposers of the "MUSE" system (Table 3) state that one of its advantages is that it may be used to predict the response to treatment, the risk of hemorrhage or malignancy, and the need for long-term surveillance [2]. The original Savary-Miller system can be used similarly [1].The "MUSE" system is also touted as helpful by allowing one to distinguish between the reversible and irreversible sequelae of reflux, i.e., erosions ("E") and complications ("MUS") respectively. Many

Table I. .Grading systems for reflux esophagitis

	Scales	Savary-Miller [1] (1977)	Sonnenberg [3] (1982)	Knuff [6] (1984)	Hetzel [4] (1988)
	Grade 0		No esophagitis	Normal	Normal mucosa with no abnormalities
	Grade I	Single or isolated erosive lesion(s), oval or linear but affecting only longitudinal fold	Mild, isolated round or linear erosions	Mucosal hyperemia, patchy and/or linear	Erythema or hyperemia of the esophageal mucosa with no macroscopic erosion
	Grade II	Multiple erosive lesions, non-circumferential, affecting more than one longitudinal fold, with or without confluence	Severe, confluence of the erosions involving the total esophageal circumference	Hyperemia, granularity and/or friability	Superficial ulceration or erosions involving <10% of the last 5 cm of esophageal squamous mucosa
	Grade III	Circumferential erosive lesions	Complicated, erosions as described, deep ulcers, peptic stenosis and/or columnar epithelium-lined esophagus	Erosions	Superficial ulceration or erosions involving > 10-50% of the last 5 cm of esophageal surface
	Grade IV	Complications: columnar epithelium, ulcer(s), stricture(s) and/or short esophagus. Alone or with grades I- III		Stricture or frank ulcer	Deep ulceration anywhere in the esophagus or confluent erosion of >50% of last 5 cm of esophageal squamous mucosa
	Grade V

will disagree that ulcers are irreversible, since healing is often achieved with adequate therapy.

Any detailed system of grading must provide the user with precise definitions of each entity to be graded in order to minimize the degree of interobserver error. The "S" of the "MUSE" mnemonic stands for stricture, but we are not provided with a

Table 2. .Grading systems for reflux esophagitis

	Scales	Johnson [7] (1989)	Bate [9] (1990)	Tytgat [5] (1990)	Bytzer [8] (1993)
	Grade 0	Normal mucosa	Normal esophageal mucosa	No evidence of reflux-induced damage: crisp sharply delineated SCMJ, smooth shiny squamous distal mucosa	No mucosal abnormalities
	Grade I	One or more nonconfluent lesions with erythema or exudate above the GE junction	Erythema and friability with spontaneous contact bleeding	Mild, patchy, diffuse erythema at level of SCMJ minor friability, loss of shininess, no mucosal break	One or more of the following: diffuse erythema, edema, mucosal friability isolated round or linear erythema without fibrin
	Grade II	Confluent, non-circumferential erosive and exudative lesions	Isolated round or linear erosions affecting the 2 cm of the esophagus and not the entire circumference	Superficial erosions of red dots or streaks with or without exudate. Usually small and on tip of folds. They involve <10% of mucosal surface of the 5 cm above GE junction	Fibrin-covered erosions not involving the entire circumference
	Grade III	Circumferential erosive and exudative lesions	Erosions extending above 2 cm or affecting the entire circumference	Confluent, non-circumferential erosions, may be additional exudate covering defects or slough formation. <50% of overall mucosa of distal 5 cm	Confluent erosions extending for the entire circumference or ulceration but without stenosis
	Grade IV	Chronic mucosal lesions, i.e. ulceration, stricture or Barrett's esophagus	Frank benign ulcer	Circumferential erosions or exudative lesions at level of SCMJ, regardless of the extent along distal esophagus	Complicated esophagitis: stricture or columnar-lined epithelium >3 cm above the GE junction
	Grade V		Stricture	Deep ulceration anywhere or various degrees of stricture

Table 3. ."MUSE" esophagitis classification system

	Degree of severity Metaplasia		Ulcer	Stricture	Erosions
	0. Absent	M0 Absent	U0 Absent	S0 Absent	E0 Absent
	1. Mild	M1 One fold	U1 Junctional	S1 > 9 mm	E1 One fold
	2. Moderate	M2 > Two folds	U2 Barrett's ulcer	S2 < 9 mm	E2 > Two folds
	3. Severe	M3 Circumferential	U3 Combined	S3 Stricture and short esophagus	E3 Circumferential

Metaplasia, Ulceration, Stricturing and Erosions are assessed and graded independently according to the degree of severity: 0 - absent, 1 - mild, 2 - moderate, 3 - severe. The absence or presence, along with location, of a hiatus hernia is also to be recorded (Adapted from ref. [2]).

precise definition of stricture (due to fibrosis), and how it is to be differentiated from a stenosis which is defined as lumenal narrowing of any etiology (i.e., inflammation, fibrosis or neoplasm). Both inflammatory and cicatricial stenosis occur often with reflux disease, with the former being reversible when due to inflammation alone and the latter, due to intramural fibrosis, is not reversible. This differentiation is difficult to impossible in some cases and will detract from the precision and prognostic potential of the "MUSE" system, if specific criteria for a fibrotic stricture can not be included.

In 1982 Sonnenberg et al. (Table 1) used a modification that incorporated aspects of several previously reported grading systems, including the original Savary-Miller system [4]. They included grade 0 as indicative of no erosions, grade I as mild erosions, grade II as severe erosions and grade III as complicated, which included erosions plus deep ulcers, peptic stenosis and columnar-lined esophagus. Here the generic term stenosis was used rather than the specific term stricture, again creating uncertainty as to the actual pathology present.

The descriptive term "deep" is used in some systems to describe ulcers but is not otherwise defined, thereby reducing the objectivity of assessment [3-5]. How is the endoscopist to differentiate a shallow ulcer and a deep ulcer? Is a shallow ulcer the same as an erosion?

It appears then that the original Savary-Miller system is quite adequate for clinical purposes. When more precision is required for clinical research purposes, as it should be, the MUSE system will provide the best possibility for accurate grading with the least interobserver variation.

References

1. Savary M, Miller G. L'oesophage. Manuel et Atlas d'endoscopie. Solothurn (Switzerland): Verlag Gassmann. 1977.

2. Armstrong D, Monnier Ph, Nicolet M, Blum AL. Savary M. Endoscopic assessment of oesophagitis. Gullet 1991;l:63-67.

3. Sonnenberg A, Lepsien G, Muller-Lissner SA, Koelz HR et al When is esophagitis healed? Esophageal endoscopy, histology and function before and after cimetidine treatment Dig Dis Sci 1982;27(4):297-302.

4. Hetzel DT, Dent J, Reed WD, Narielvala FM et al. Healing and relapse of severe peptic esophagitis after treatment with omeprazole. Gastroenterology 1988;95:903-912.

5. Tytgat GNJ Endoscopy of the oesophagus. In: Colton PB, Tytgal GNJ, Williams CB (eds) Annual of Gastrointestinal Endoscopy. London: Curr Sci Ltd, 1990; 15-26.

6. Knuff TE, Benjamin SB, Worsham GF, Hancock JE. Castell DO Histologic evaluation of chronic gastroesophageal reflux: an evaluation of biopsy methods and diagnostic criteria. Dig Dis Sci 1984;29(3):194-201.

7. Johnson NJ, Boyd EJS, Milis JC, Wood JR. Acute treatment of reflux oesophagitis: a multicenter trial to compare ranitidine 150 mg b.d. with ranitidine 300 mg q.d.s. Aliment Pharmacol Ther 1989;3:259-266.

8. Bytzer P, Havelund T, Hansen JM. Interobserver variation in the endoscopic diagnosis of reflux esophagitis. Scand J Gastroenterol 1993;119-125.

9. Bate CM, Keeling PWN, O'Morain C, Wilkinson SP et al Comparison of omeprazole and cimetidine in reflux oesophagitis: symptomatic, endoscopic and histological evaluations Gut 1990;31:968-972