Barrett 食管指南

PRACTICE GUIDELINES Updated Guidelines for the Diagnosis, Surveillance, and Therapy of Barrett’s Esophagus Richard E. Sampliner, M.D., and The Practice Parameters Committee of the American College of Gastroenterology Southern Arizona VA Health Care System, Tucson, Arizona PREAMBLE Guidelines for the diagnosis and treatment of Barrett’s esophagus were published by the American College of Gas- troenterology in 1998 (1). These and other guidelines un- dergo periodic review. Significant advances have been made in the area of Barrett’s esophagus over the past several years, leading us to review and revise our previous guide- lines statements. These advances have included more infor- mation on the natural history of high-grade dysplasia and the chronic use of proton pump inhibitors. These and the orig- inal guidelines are intended to apply to all physicians who address Barrett’s and are intended to indicate the preferable, but not only acceptable, approach. Physicians must always choose the course best suited to the individual patient and the variables that exist at the moment of the decision. These guidelines are intended to apply to adult patients with the diagnosis of Barrett’s esophagus defined in the guidelines. Both these and the original guidelines were developed under the auspices of the American College of Gastroenter- ology and its Practice Parameters Committee and approved by the Board of Trustees. The world literature was reviewed extensively for the original guidelines and was once again reviewed using the National Library of Medicine database. All appropriate studies were reviewed, and any additional studies found in the reference list of these papers were obtained and reviewed. Evidence was evaluated along a hierarchy, with randomized controlled trials given the great- est weight. Abstracts presented at national and international meetings were only used when unique data from ongoing trials were presented. When scientific data were lacking, recommendations were based on expert consensus. During preparation, the guidelines were reviewed by the American Gastroenterological Association and the American Society for GI Endoscopy. Recommendations and comments ob- tained from these reviews were incorporated into the final document whenever possible. DEFINITION OF BARRETT’S ESOPHAGUS (No change) Barrett’s esophagus is a change in the esoph- ageal epithelium of any length that can be recognized at endoscopy and is confirmed to have intestinal metaplasia by biopsy of the tubular esophagus and excludes intestinal metaplasia of the cardia. The current definition of Barrett’s identifies individuals in whom the abnormal epithelium can be identified at endos- copy and in whom surveillance endoscopy and biopsy are appropriate. The definition of Barrett’s esophagus has evolved over the last 2 decades from the columnar-lined esophagus (2), to 3 cm of columnar lining or intestinal metaplasia in the esophagus (3), to the requirement for intestinal metaplasia in the esophagus without specification of length. Adenocarcinoma of the esophagus continues to be the most rapidly rising incidence cancer in the United States (4, 5). Intestinal metaplasia of the esophagus is the premalig- nant lesion for adenocarcinoma of the esophagus and the esophagogastric junction. The vast majority of adenocarci- nomas of the esophagus are accompanied by intestinal meta- plasia (6–10), and many adenocarcinomas of the esophago- gastric junction are associated with esophageal intestinal metaplasia (11–13). There has been a recent focus on “short-segment” Bar- rett’s esophagus—intestinal metaplasia in the distal esoph- agus �3 cm in length (14–16). There is increasing evidence that the physiological ab- normalities—esophageal pH exposure and lower esopha- geal sphincter pressure—in short-segment Barrett’s are less severe than in long-segment Barrett’s but are qualitatively similar (17, 18). Short-segment Barrett’s was arbitrarily excluded in older definitions of Barrett’s esophagus that did not include histological criteria. Short-segment Barrett’s esophagus needs to be distinguished from intestinal meta- plasia of the gastric cardia, a lesion of the stomach that cannot be seen on routine endoscopy, with less well-defined epidemiology and significance (19–21). Importantly, dys- plasia is more common in short-segment Barrett’s than intestinal metaplasia of the gastric cardia (22). SCREENING FOR BARRETT’S ESOPHAGUS IN PATIENTS WITH GASTROESOPHAGEAL REFLUX DISEASE (GERD) (Old) Patients with long-standing GERD symptoms, partic- ularly those�50 yr of age, should have upper endoscopy to detect Barrett’s. THE AMERICAN JOURNAL OF GASTROENTEROLOGY Vol. 97, No. 8, 2002 © 2002 by Am. Coll. of Gastroenterology ISSN 0002-9270/02/$22.00 Published by Elsevier Science Inc. PII S0002-9270(02)04267-3 (New) Patients with chronic GERD symptoms are those most likely to have Barrett’s esophagus and should undergo upper endoscopy. The major reason to evaluate patients with chronic symp- toms of GERD is to recognize Barrett’s esophagus (see American College of Gastroenterology guidelines for GERD) (23). Unfortunately, the epidemiology of Barrett’s is incompletely described. However, we know from series of surgical resections of adenocarcinoma of the esophagus associated with Barrett’s esophagus that white men over- whelmingly predominate (6–10). Additionally, there are data suggesting that the longer the duration of reflux symp- toms, the higher the prevalence of Barrett’s esophagus at the time of upper endoscopy (24, 25). In one population-based study, the severity of reflux symptoms was correlated with the risk of esophageal adenocarcinoma (26). Some experts recommend a one-time endoscopy to exclude Barrett’s dur- ing the lifetime of a patient with GERD (1, 27). The specific timing of this endoscopy by patient age and duration of symptoms is not yet defined. The age threshold for initial endoscopy is being reevaluated (28). The factors of gender, race, and age can be used to determine the threshold for endoscopy in patients with GERD to screen for the presence of Barrett’s. The highest yield of Barrett’s esophagus would be expected in white men with chronic symptoms of GERD. However, the spe- cific criteria to select patients to screen for Barrett’s are not yet defined. The recognition of asymptomatic individuals with Bar- rett’s esophagus remains a problem. Such people may ac- count for the higher number of patients presenting with cancer and Barrett’s (prevalence cases) than patients with Barrett’s developing adenocarcinoma (incidence cases) in published series (29). Asymptomatic Barrett’s highlights the need to assess the distal esophagus carefully in all patients undergoing upper endoscopy for any indication. THE DIAGNOSIS OF BARRETT’S ESOPHAGUS (No change) The diagnosis of Barrett’s esophagus requires systematic biopsy of the abnormal-appearing esophageal mucosa to document intestinal metaplasia and to detect dysplasia. At the time of endoscopy, when “gastric-appearing mu- cosa” or apparent “columnar-lined esophagus” is evident, multiple biopsies are indicated to detect intestinal metapla- sia. For the recognition of Barrett’s esophagus, it is essential to specifically identify the squamocolumnar junction and the esophagogastric junction. When the squamocolumnar junc- tion is displaced proximal to the esophagogastric junction, then Barrett’s esophagus may be present. Erosive esophagi- tis or erythema of the esophagus may be confused visually with Barrett’s esophagus. This is one reason that biopsy is so essential in making a diagnosis of Barrett’s esophagus. Erosive esophagitis may impair the recognition of Barrett’s esophagus, necessitating re-endoscopy to rule out Barrett’s (30, 31) after treating with acid suppression therapy. This scenario provides the rationale for evaluating a patient after reflux symptoms are controlled with therapy. As a group, patients with Barrett’s esophagus have a very low lower esophageal sphincter pressure (17, 32). This may make recognition of the end of the esophagus difficult. The proximal margin of the gastric folds within a hiatal hernia when the distal esophagus is deflated serves as a marker for the end of the esophagus and the location of the esophago- gastric junction (1, 33). A variety of endoscopic staining techniques have been used to enhance the recognition of Barrett’s esophagus. Vital stains used include Lugol’s iodine (34), toluene blue (35), indigo carmine (36), and methylene blue (37, 38). Methylene blue may be the most promising technique with evidence that Barrett’s can be diagnosed with fewer biopsies with targeting of the stained mucosa and in increased yield of short-segment Barrett’s (39). Staining can be tedious, can prolong the procedure time, and may not be reproducible by all endoscopists. Magnification endoscopy may improve the recognition of Barrett’s (40, 41) by enhancing the recognition of mucosal detail and identifying high-yield sites for biopsy. The number of biopsies necessary to detect intestinal metaplasia has not been defined. The more biopsies taken, the greater the likelihood of recognizing intestinal metapla- sia. The greater the length of abnormal-appearing esopha- gus, the higher the yield of intestinal metaplasia by biopsy (39, 42). The recognition of intestinal metaplasia on biopsy, i.e., specifically goblet cells, is increased by the use of alcian blue stain at pH 2.5 (43). This will decrease the chance of missing the presence of goblet cells or of misinterpreting cells with cystic structures in the cytoplasm as goblet cells. If the endoscopist suspects Barrett’s esophagus, biopsy of the lesion at the time of endoscopy is essential for the recognition of dysplasia. Dysplasia is a change in the cy- tology and architecture of the metaplastic glands that is the first step in the neoplastic process. SURVEILLANCE OF BARRETT’S ESOPHAGUS (Old) Patients with Barrett’s esophagus should undergo surveillance endoscopy and biopsy at an interval deter- mined by the presence and grade of dysplasia. (New) The grade of dysplasia determines the endoscopy interval, and an abnormal epithelial surface such as a nodule or ulcer requires special sampling attention. Sur- veillance endoscopy intervals are lengthening in the ab- sence of dysplasia on two consecutive endoscopies with biopsy—a 3-yr interval is appropriate. The rationale for surveillance in Barrett’s esophagus is based on the increased risk of developing adenocarcinoma and the fatal nature of this cancer. Preliminary data suggest that esophageal adenocarcinoma detected by surveillance is at an earlier stage with a more favorable survival than carcinoma detected at the time of the diagnosis of Barrett’s, 1889AJG – August, 2002 Practice Guidelines for Barrett’s Esophagus typically when patients present with dysphagia (44–46). In contrast, the low incidence of adenocarcinoma is used to support an approach of not surveying patients with Barrett’s esophagus. A cohort study of patients with Barrett’s esoph- agus not undergoing surveillance demonstrated that esoph- ageal cancer was an uncommon cause of death—2.5% of the deaths of 155 patients followed a mean of 9 yr (47). The follow-up of a smaller cohort of Barrett’s patients comes to a similar conclusion, although 9% of the patients died from esophageal cancer (48). Patients with short-segment Barrett’s esophagus can de- velop dysplasia (49, 50) and cancer (51–53), the incidence of the development of cancer in these patients is not fully defined. In a series that may not have been adequately powered to detect a difference, patients with short-segment Barrett’s esophagus had the same incidence of cancer as patients with long-segment Barrett’s esophagus (53). This provides a rationale for surveillance of patients with short- segment Barrett’s esophagus. The goal of surveillance in patients with Barrett’s esoph- agus is the detection of dysplasia and early cancer. Dyspla- sia occurs on the background of metaplasia—a fundamental and distinctive change in the epithelium of the esophagus from one differentiated cell type to another. Dysplasia rep- resents the final step of neoplasia and is characterized by cytological and architectural changes. Dysplasia is the best current indicator of the risk of cancer. The grading of dysplasia in Barrett’s esophagus is based on the system developed for ulcerative colitis (54). Observer variation in the grading of dysplasia is a prob- lem. Interobserver agreement is in the range of 85% when separating high-grade dysplasia and intramucosal carcinoma from low-grade dysplasia, indefinite for dysplasia and neg- ative (55). The reproducibility of the diagnosis of dysplasia has been recently researched using kappa (�) statistics (56). Interobserver agreement for no dysplasia, indefinite, and low-grade dysplasia versus high-grade dysplasia and carci- noma was substantial (� � 0.7—a � of 1 means complete agreement) and for four grades—no dysplasia, indefinite, and low-grade dysplasia, high-grade dysplasia, and carci- noma, moderate (� � 0.46). It is important to realize that dysplasia is a first step in the neoplastic process and that any grade of dysplasia may overlie or may be adjacent to a frank carcinoma. Therefore, the reading of dysplasia of any grade on a biopsy warrants a repeat endoscopy and intensive biopsy of the area with dysplasia to rule out coexisting carcinoma, with attention given to maximum acid suppres- sion before rebiopsy. Patients with Barrett’s esophagus are candidates for sur- veillance if there is a potential to prolong life expectancy with a therapeutic intervention for early cancer. Therefore, age and comorbid conditions are important factors to weigh. Documentation of efficacy of promising endoscopic eradi- cation technologies may make more patients eligible for surveillance (57–61) because endoscopic therapy may be applicable to patients whose functional status and/or major cardiopulmonary disease might preclude resectional sur- gery. The appropriate surveillance intervals for patients with Barrett’s esophagus are a function of the grade of dysplasia (Table 1). However, the published database of the natural history of dysplasia is limited to five centers that have performed prospective studies (62–66) and one registry (67). A total of 783 patients have been followed for 2.7–7.3 yr. Nine of 382 (2%) patients have been followed from no dysplasia to cancer. Five of 72 (7%) patients have been followed from low-grade dysplasia to cancer. Recently, it has been demonstrated that low-grade dysplasia is com- monly a “transient” finding. In a series of 34 patients who had low-grade dysplasia, 73% had no evidence of dysplasia on at least one subsequent surveillance endoscopy (68). If two experienced GI pathologists agreed on the diagnosis of low-grade dysplasia, there was a significant association with progression to high-grade dysplasia or cancer (69). In con- trast, 37 of 170 (22%) patients with high-grade dysplasia have progressed to cancer. Unfortunately, it is not always possible to determine the length of follow-up of patients with different grades of dysplasia. Not knowing the percent progressing to cancer over a specific time interval limits the usefulness of the information. Furthermore, uniform criteria and exclusion of prevalence cancers have not been applied. For example, eliminating the cancers that developed in the first year, a common criterion for prevalence (versus incidence) cancers would exclude an estimated 15 of 33 cancers from one series (64). In this same series, the cumulative 5-yr incidence of cancer drops from 59% to 31% by analyzing only the incident high-grade dysplasia patients. The latter analysis presumably reduces the referral bias. By adjusting the data in this manner, the apparently broad range of cancer inci- dence is dramatically narrowed from 16% to 59% to 16% to 24%! This estimate is substantiated by a recently published retrospective series of patients with high-grade dysplasia (69). After eliminating the first 6-month prevalence cancers, 16% of 86 patients developed cancer over an approximate 3-yr interval. Another variable commonly not defined in series of high- grade dysplasia is mucosal nodularity. This feature increases the risk of cancer 2.5 times (p � 0.01) (69). After adjusting for nodularity, patients with diffuse high-grade dysplasia had a 3.7-times increased risk of cancer compared with patients with focal high-grade dysplasia—five crypts or less involved in one biopsy specimen from the entire set of biopsies (p � 0.02). Table 1. Grade of Dysplasia and Development of Esophageal Adenocarcinoma Dysplasia (%) n Cancer (%) None 382 9 (2) Low grade 72 5 (7) High grade 170 37 (22) A total of 783 patients followed a mean of 2.9–7.3 yr (61–65). 1890 Sampliner and ACG Practice Parameters Committee AJG – Vol. 97, No. 8, 2002 Mucosal nodularity offers the opportunity for endoscopic mucosal resection (70, 71). This results in the capability of assessing more tissue than obtained with endoscopic biopsy and to more accurately identify and stage early cancer. Surveillance endoscopy and biopsy intervals are length- ening as the database on the outcomes of dysplasia in- creases. Surveillance of Barrett’s esophagus patients lacking dysplasia with a systematic biopsy protocol at two endos- copies may be extended to a 3-yr interval (Table 2). THE MANAGEMENT OF DYSPLASIA In patients with low-grade dysplasia as the highest grade after a follow-up endoscopy with concentrated biopsies in the area of dysplasia, annual endoscopy is recommended until there is no dysplasia. The finding of high-grade dys- plasia requires a repeat endoscopy with special attention to any mucosal irregularity potentially including endoscopic mucosal resection. An intensive biopsy protocol ideally with a therapeutic endoscopic and large capacity biopsy forceps should be performed. An expert pathologist should confirm the interpretation of high-grade dysplasia. Focal high- grade dysplasia (less than five crypts) may be followed with 3-month surveillance. Intervention may be considered in a patient with confirmed multifocal high-grade dysplasia. The treatment of high-grade dysplasia is controversial, especially with the recognition of the lower risk of progress- ing to cancer, particularly if the high-grade dysplasia is focal. High-grade dysplasia may be present only intermit- tently or may regress to low grade even over a long-term follow-up (64). Although esophagectomy is commonly rec- ommended for patients with high-grade dysplasia (72, 73), given the morbidity of this procedure, the mortality at low- volume institutions (74), and the variability of the natural history of high-grade dysplasia, caution is justified. Esoph- agectomy at a high-volume institution remains a reasonable strategy in the surgically fit patient with recurrent diffuse high-grade dysplasia confirmed by an expert GI pathologist. An intensive biopsy protocol may be successful at endo- scopically differentiating high-grade dysplasia from cancer (75). The surgical literature contrasts with this experience. Of 126 cases with high-grade dysplasia alone by endoscopic biopsy, 41% had cancer at the time of esophagectomy (76, 77). These studies did not follow a uniform endoscopic biopsy protocol, yet the majority of cancers were early stage, with a more favorable patient survival. The goal of surveillance is to decrease the mortality from adenocarcinoma so that intervention before frank cancer, which has the risk of metastasis, is reasonable. The precise threshold for intervention needs to be individualized and agreed upon after a formal discussion of the concerned clinician with the concerned patient. These discussions may include the therapeutic endoscopist and the surgeon. The recent excellent outcomes of expert surgeons resecting early adenoca