- Email: [email protected]
Dysplasia and Cancer in a Large Multicenter Cohort of Patients With Barrett’s Esophagus
CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2006;4:566 –572
ORIGINAL ARTICLES Dysplasia and Cancer in a Large Multicenter Cohort of Patients With Barrett’s Esophagus PRATEEK SHARMA,* GARY W. FALK,‡ ALLAN P. WESTON,* DEAN REKER,* MARK JOHNSTON,§ and RICHARD E. SAMPLINER¶ *University of Kansas School of Medicine & Veterans Affairs Medical Center, Kansas City, Missouri; ‡Cleveland Clinic Foundation, Cleveland, Ohio; §Bethesda Naval Medical Center, Bethesda, Maryland; and ¶University of Arizona Health Science Center and Southern Arizona VA Health Care System, Tucson, Arizona
Background & Aims: The exact incidence of adenocarcinoma in patients with Barrett’s esophagus (BE) is not known and is reported to vary from 0.2%–2% per year. Published series of patients with BE have included relatively small numbers of patients with limited duration of follow-up. The goal of this study was to define the prevalence and incidence of dysplasia and cancer and evaluate the paths of progression in a large multicenter cohort of BE patients. Methods: The BE study is a multicenter clinical and endoscopic outcomes project involving a single large database of patients with BE. Data from each of the participating centers were merged into the main study database. Cancers and HGD occurring within 12 months of the index endoscopy were regarded as prevalent cases. Results: One thousand three hundred seventy-six patients met the study criteria (95% white, 14% women); 91 patients had cancer at the initial endoscopy (prevalent cases, 6.7%; 95% confidence interval [CI], 4.8%– 8.7%). Six hundred eighteen patients were followed for a total of 2546 patient-years; mean follow-up was 4.12 years. Twelve patients developed cancer during follow-up, a cancer incidence of 1 in 212 patient-years of follow-up (0.5% per year; 95% CI, 0%–1.1%). The combined incidence of HGD and/or cancer was 1 in 75 patient-years of follow-up or 1.3% per year (95% CI, 0%–2.2%). Of the 34 patients developing HGD and/or cancer, 18 patients (53%) had at least 2 initial consecutive endoscopies with biopsies revealing nondysplastic mucosa. The incidence of LGD was 4.3% per year (95% CI, 2.8%– 6.0%). In the 156 patients with LGD, regression to no dysplasia occurred in 66%, persistent LGD in 21%, and progression to HGD/cancer in 13%. The incidence of cancer in patients with LGD was 1 in 156 patient-years of follow-up or 0.6% per year (95% CI, 0%–1.3%). Conclusions: Preliminary results from this trial define the prevalence and incidence of dysplasia and cancer in a multicenter cohort of patients with BE. At least half the patients who developed HGD and/or cancer had 2 consecutive initial endoscopies with biopsies revealing nondysplastic mucosa. The majority of patients with LGD regressed and had a cancer incidence similar to all BE patients.
arrett’s esophagus is a known complication of chronic gastroesophageal reflux disease and is a premalignant lesion for adenocarcinoma of the esophagus and the esophagogastric junction.1,2 Initial Surveillance, Epidemiology & End Results (SEER) program data (1980s) demonstrated a rapidly rising incidence for esophageal adenocarcinoma, and subsequent updated reports through 2001 have confirmed a continuing increase in the incidence of this cancer in the United States.3–5 The rise in incidence has been most marked among white men. The exact incidence of adenocarcinoma in patients with Barrett’s esophagus is not clear and has been reported to vary from 1 in 50 patient-years of follow-up to 1 in 285 patient-years of follow-up.6 –12 The published incidence data in cohorts of patients with Barrett’s esophagus have been influenced by, among other factors, the small number of patients being evaluated, short duration of follow-up, as well as by the demographics of the population being studied.13,14 Other biases have been reported, including publication bias that shows that smaller studies have reported a higher incidence of cancer, whereas relatively large studies show a lower incidence of cancer.15 To date, the maximum number of patients included in a prospective study determining the incidence of adenocarcinoma has been 327 patients.12 Surveillance endoscopy in patients with Barrett’s esophagus has been recommended for the early detection of dysplasia and cancer16; however, the time interval of progression to high-grade dysplasia (HGD) or cancer is not known, and the surveillance intervals are not clearly defined.17,18 The burden of endoscopic surveillance of patients with Barrett’s esophagus is enormous, and the role of surveillance is very important but remains controversial.19,20
B
Abbreviations used in this paper: CI, confidence interval; HGD, highgrade dysplasia; LGD, low-grade dysplasia. © 2006 by the American Gastroenterological Association Institute 1542-3565/06/$32.00 doi:10.1016/j.cgh.2006.03.001
May 2006
DYSPLASIA AND CANCER IN PATIENTS WITH BARRETT’S ESOPHAGUS
Moreover, the cost-effectiveness of surveillance endoscopy in patients with Barrett’s esophagus is very sensitive to the incidence of cancer.21 To address these gaps in our knowledge of the epidemiology of Barrett’s esophagus–associated dysplasia and carcinoma, we developed a consortium of investigators and centers interested in this disease entity to develop a large cohort of Barrett’s esophagus patients followed prospectively. The goals of this project were (1) to create a single, large database of patients with Barrett’s esophagus based on standardized criteria and (2) to calculate the prevalent and incident cases of adenocarcinoma, HGD, and low-grade dysplasia (LGD) in this group.
Methods Patients Patients diagnosed with Barrett’s esophagus at each of the participating centers, Veterans Affairs Medical Center, Kansas City, Mo, Southern Arizona Veterans Affairs Health Care System, Tucson, Ariz, Bethesda Naval Medical Center, Bethesda, Md, and Cleveland Clinic Foundation, Cleveland Ohio, were identified from the database of each center. From each center, the following information regarding all patients with Barrett’s esophagus was obtained: age, gender, ethnicity, length of Barrett’s esophagus, date of each endoscopy, and histologic diagnosis at each endoscopy. The duration of follow-up of each patient was calculated from the time of initial diagnosis of Barrett’s esophagus (ie, from their first endoscopy) to their most recent endoscopy with biopsy. The time of occurrence of LGD, HGD, and adenocarcinoma was documented in the database, thus allowing for the recording of prevalence and incidence cases of HGD and adenocarcinoma. The patients reported in this analysis met the following criteria: (1) a standardized definition of Barrett’s esophagus, the presence of columnar mucosa in the distal esophagus of any length with intestinal metaplasia documented on histology; (2) follow-up for at least 12 months from the time of initial diagnosis (ie, after first endoscopy and biopsy); and (3) adenocarcinoma and HGD patients detected within the first 12 months of their initial Barrett’s esophagus diagnosis were defined as prevalence cases. Patients were not included in the analysis if (1) there was columnar mucosa in the distal esophagus but no intestinal metaplasia on biopsy, or (2) patients had undergone endoscopic ablation therapy.
Endoscopy, Biopsy, & Dysplasia Recording At each center, patients with Barrett’s esophagus were defined by using a combination of endoscopic and histologic criteria.9,10 Once Barrett’s esophagus was documented, patients were enrolled in a surveillance program. Although endoscopic surveillance intervals were not standardized as part of the study protocol, each center followed the recommendations of the published guidelines.16 During surveillance visits, each patient underwent endoscopy with biopsies of the Barrett’s segment. Again, the biopsy protocol had not been standardized between the cen-
567
ters; each center followed a rigorous biopsy protocol of at least 4 quadrant biopsies every 2 cm with either a standard or jumbo biopsy forceps. The patient demographics, length of Barrett’s esophagus, and dysplasia grade from their baseline/initial endoscopy were recorded, and at subsequent endoscopies the grade of dysplasia for each visit was also noted. At each center, the length of Barrett’s esophagus was recorded in centimeters and judged by measuring the distance from the most proximally displaced squamocolumnar junction to the anatomic gastroesophageal junction.22–24 Dysplasia in the histology specimens was recorded according to established criteria for LGD, HGD, and adenocarcinoma and was reported by specialized gastroenterology pathologists at each site.25 Central reading of pathology was not performed for this analysis.
Data Entry, Management, and Analysis The study was approved by the Institutional Review Board of each institution. Each patient was provided with a unique identification number, and all patient identifiers were deleted in compliance with Health Information Portability and Accountability Act regulations. Data were generated directly from the existing data sets into the main study database. The database structure was designed by using Microsoft Access for Windows 1997 (Microsoft Corp, Redmond, Wash). The study coordinators at each center were responsible for collecting patient information and data entry as outlined above. Missing information was obtained by chart abstraction at each site. All collected and merged data were double entered, compared, and reconciled for accuracy. Data analysis was performed by using SAS version 8.2 (SAS Institute, Inc, Cary, NC), and survival curves were constructed by using the Kaplan-Meier method. Data collection, management, and analysis were supervised by a health services researcher (D.R.) at the Veterans Affairs Medical Center, Kansas City, Mo. The prevalent cases of dysplasia were defined as the percentage of patients with Barrett’s esophagus who had dysplasia at initial endoscopy. The follow-up interval for each patient was calculated as the time between the first endoscopy revealing Barrett’s esophagus and the most recent follow-up examination. The incidence was calculated from the number of patients lacking dysplasia at the baseline endoscopy who developed dysplasia during followup. In the calculation of the incidence of Barrett’s esophagus– related adenocarcinoma, patients found to have Barrett’s esophagus at the same time as the diagnosis of an adenocarcinoma or within 12 months of the initial diagnosis of Barrett’s esophagus were excluded and considered as prevalent cases.
Results Demographics and Prevalence Data From the 4 participating centers, 1376 patients met the study criteria and had at least one endoscopy with biopsy documenting Barrett’s esophagus. The mean age of this cohort was 59 ⫾ 12.3 years (standard deviation) and included 83% men and 17% women. The majority of the
568
SHARMA ET AL
CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 5
Table 1. Prevalence and Incidence of Dysplasia and Adenocarcinoma in Patients With Barrett’s Esophagus
Total number of patients Patients with LGD Patients with HGD Patients with adenocarcinoma aAnnual
Prevalence cases
Incidence cases
Incidence (per year)
1376 101 (7.3%) 42 (3%) 91 (6.6%)
618 100 22 12
4.3%a (95% CI, 2.8%–6.0%) 0.9%a (95% CI, 0%–1.7%) 0.5%a (95% CI, 0%–1.1%)
incidence.
patients were white (97%). This cohort was used to calculate the prevalent cases. At the first recorded endoscopy, esophageal adenocarcinoma was diagnosed in 91 patients, HGD in 42, and LGD in 101 patients (Table 1). Incidence of High-Grade Dysplasia and Adenocarcinoma Of 1376 patients meeting the initial study criteria, only 618 patients had follow-up of at least 1 year with a repeat endoscopy. One hundred sixty-five patients had a follow-up of less than 1 year and therefore were not included in the analysis of the incidence follow-up cases. Moreover, 549 patients had only a single endoscopy, and they either were not due for their repeat endoscopy or were lost to follow-up. These 618 patients were followed for a total of 2546 patient-years with a mean follow-up period of 4.12 years (range, 1–22.5 years). Twelve patients developed esophageal adenocarcinoma during follow-up, a cancer incidence of 1 in 212 patient-years of follow-up or 0.5% per year (95% confidence interval [CI], 0%–1.1%). The mean time to development of cancer was 5.3 years (range, 2.4 – 11.2 years). Seven of 12 patients had HGD before cancer development, 2 had only LGD (ie, were never detected with HGD during follow-up), whereas 3 developed cancer from nondysplastic Barrett’s esophagus (ie, were never detected with LGD or HGD at any instance) (Figure 1). Twenty-two patients developed HGD during follow-up, resulting in an HGD incidence of 1 in 116 patient-years of follow-up or 0.9% per year (95% CI, 0%–1.7%). The mean time to HGD development was 3.8 years (range, 1.2–7.9 years). Eleven of 22 patients developed HGD from Barrett’s esophagus without dysplasia, whereas the remaining 11 progressed from LGD to HGD. Thus, the total number of patients developing HGD and/or cancer during follow-up were 34, and the combined incidence of HGD and/or cancer was 1 in 75 patient-years of follow-up or 1.3% per year (95% CI, 0%–2.3%) (Table 1). The rate of cancer development (ie, annual incidence of adenocarcinoma in Barrett’s esophagus patients) in the different participating centers varied from none at the Bethesda Naval Medical Center (total number of patients, 75) to 0.4% at the Cleveland Clinic Foundation (total number of patients, 241), 0.6% at the Southern Arizona Veterans
Affairs Medical Center (total number of patients, 135), and 0.8% at the Kansas City Veterans Affairs Medical Center (total number of patients, 167) (not significantly different). The time periods of enrollment and database entry periods were as follows: Bethesda 1993–2000, Cleveland Clinic Foundation 1979 –2000, Kansas City Veterans Affairs Medical Center 1984 –1993, and Southern Arizona Veterans Affairs Medical Center 1975–1998. The annual rate of HGD development (incidence) was 0.7% each at the Kansas City Veterans Affairs Medical Center and the Cleveland Clinic Foundation and was 1.1% each at the Bethesda Naval Medical Center and the Southern Arizona Veterans Affairs Medical Center. The overall incidence of either cancer and/or HGD development varied from 1.0% per year to 1.7% per year at the different study sites. As shown in the Kaplan-Meier survival graphs (Figure 2), at 5 years of follow-up (N ⫽ 157), 97% of patients were free of cancer, and at 8 years 95% were free of cancer (N ⫽ 66). This indicates that at least for the initial 5– 8 years of follow-up after a diagnosis of Barrett’s esophagus, the cancer incidence was approximately 5% (95% CI, 3.3%– 6.7%), although the number of patients at 8 years in this series was only 66.
Figure 1. Progression to adenocarcinoma in patients with Barrett’s esophagus during follow-up. Twelve patients developed cancer during a mean follow-up of 4.12 years. The grades of dysplasia at baseline, during follow-up, and at the final endoscopy are also shown. The number of patients is shown in parentheses.
May 2006
DYSPLASIA AND CANCER IN PATIENTS WITH BARRETT’S ESOPHAGUS
569
Figure 2. Barrett’s esophagus and development of cancer: event distribution. Kaplan-Meier curve showing the percentage of patients free of esophageal adenocarcinoma during follow-up. There was a gradual decrease in the number of patients as the duration of follow-up increased.
Progression to High-Grade Dysplasia and Adenocarcinoma: Findings at Initial Endoscopy Of the 34 patients developing either HGD and/or cancer during follow-up, 18 patients (53%) had at least 2 initial consecutive endoscopies documenting nondysplastic tissue, intestinal metaplasia without dysplasia. The mean time interval between these 2 consecutive endoscopies was 2 years (range, 1 month–3 years). After the second endoscopy that documented intestinal metaplasia without dysplasia, HGD/cancer was detected after a mean of 3 years (range, 3 months–7.8 years). The pathways of progression from nondysplastic Barrett’s esophagus to HGD and cancer are shown in Figures 1 and 3.
patients developing cancer during follow-up. Many patients with LGD (42%) had regression to nondysplastic epithelium (ie, intestinal metaplasia) at their subsequent endoscopy, whereas 32% had intermittent LGD (ie, went back and forth between LGD and nondysplastic Barrett’s esophagus during their follow-up period).
Incidence and Outcome of Low-Grade Dysplasia Of the initial cohort of 618 patients with at least 1 year of follow-up, the diagnosis of LGD during follow-up was confirmed in 100 patients, an LGD incidence of 4.3% per year (95% CI, 2.8%– 6%). Besides these 100 patients who developed LGD during follow-up, an additional 56 patients who had LGD at their baseline endoscopy also had further surveillance and follow-up. Thus, a total of 156 patients with LGD were followed for at least 1 year with at least 2 assessments for a mean duration of 5 years (range, 1–15.5 years). The outcome of LGD in these patients at their most recent endoscopy is shown in Figure 4. The incidence of cancer in patients with LGD was 1 in 156 patient-years of follow-up or 0.6% per year (95% CI, 0%–2%), with a total of 5
Figure 3. Even patients with nondysplastic Barrett’s esophagus progressed to HGD and cancer during follow-up. The number of patients is shown in parentheses.
570
SHARMA ET AL
Figure 4. Outcome of patients with LGD during follow-up. The majority of the patients regressed to nondysplastic Barrett’s esophagus.
Cancer Incidence Estimates for Patients Lost to Follow-Up There were 549 cases that were excluded from the “incidence data analysis” because of a single endoscopy or loss to follow-up. If we assume that none of these cases would have developed cancer during a mean follow-up of 4 years, the numerator would remain as 12 incidence cancer cases during 4732 patient-years of follow-up, giving an incidence of 0.2% per year (95% CI, 0%– 0.4%). This would represent the most conservative cancer rate estimate for the entire cohort. However, if we assume an additional 12 cases of cancer to develop during a mean follow-up of 4 years for these 549 patients, this would provide 24 cases during 4732 patient years of follow-up, giving an annual incidence of 0.5% per year (95% CI, 0.1%– 0.9%).
Discussion The number of incidence cases of dysplasia (both LGD and HGD) and cancer in this study was much lower than the prevalent cases, indicating that the majority of the neoplastic lesions are detected at the initial endoscopy with biopsy. On the other hand, HGD and adenocarcinoma can develop within 1–2 years of the initial endoscopy revealing nondysplastic Barrett’s esophagus. These data highlight the need for better risk stratification of patients with Barrett’s esophagus, given that only the minority will develop HGD or cancer and patients can progress to cancer from nondysplastic tissue. Moreover, at least half the patients who developed HGD and/or cancer had 2 consecutive initial endoscopies revealing intestinal metaplasia without dysplasia (mean time to progression, 3 years). Some authors have suggested that patients with an initial endoscopy lacking dysplasia might not require follow-up.26 Even patients with 2 consecutive findings of intestinal metaplasia without dysplasia can develop neoplasia changes. Factors other than a
CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 5
negative initial dysplasia grade need to be evaluated for risk stratification for HGD/cancer development.27 The majority of the HGDs and cancers were detected as prevalent cases. Although the exact indication for endoscopy at the index time period was not recorded in the patients and available in the database, one can speculate that the greatest advantage of diagnosing and recognizing lesions in patients suspected to harbor Barrett’s esophagus or dysplasia cancer occurs at the index endoscopy. This might be due to a number of factors including referral bias, alarm symptoms at the time of diagnosis, and selection bias. Surveillance of patients with known Barrett’s esophagus is recommended for the early detection of dysplasia and cancer, a practice followed by the majority of practicing gastroenterologists in the United States.16,28,29 Patients undergoing surveillance endoscopy for Barrett’s esophagus have been noted in retrospective studies and in a recent population-based study to have better survival when cancer is detected, as opposed to patients presenting with dysphagia and a prevalent adenocarcinoma with Barrett’s esophagus.20,30 –33 There are no prospective studies indicating a survival benefit of surveillance versus no surveillance. Surveillance of patients with Barrett’s esophagus has raised a number of issues including cost-effectiveness of surveillance, given the relatively low risk of cancer.34 A recent study with a computer model found that if the risk of cancer in Barrett’s esophagus patients is 0.4% per year, the most cost-effective surveillance strategy would be endoscopy every 5 years.21 Another recent model determined that surveillance only for patients with dysplasia would be a costeffective strategy.26 Both models are very sensitive to the incidence of dysplasia and adenocarcinoma in patients with Barrett’s esophagus, and any surveillance program has to take the incidence estimate into account. Therefore, determining the accurate incidence of dysplasia and cancer is of critical importance to the cost-effectiveness of a surveillance program in patients with Barrett’s esophagus. The incidence of adenocarcinoma in patients with Barrett’s esophagus has ranged from 0.2%–2% per year.6 –11 In a study by Hammeeteman et al6 in the 1980s, the annual risk was 1.9%, whereas in 1997 Drewitz et al9 reported an annual incidence of 0.5%. A number of factors including study design, retrospective nature of study, short duration of follow-up, few patients, lead time bias, length time bias, and healthy volunteer bias might have contributed to this discrepancy in the literature. A recent report also cited publication bias as a potential cause for this, with small studies reporting a higher incidence of cancer and larger studies reporting a lower incidence.13 The authors estimated the true incidence of cancer to be approximately
May 2006
DYSPLASIA AND CANCER IN PATIENTS WITH BARRETT’S ESOPHAGUS
0.5% per year, similar to our results in a large cohort of patients with Barrett’s esophagus. Another unresolved issue is whether cancer incidence changes over time in patients with Barrett’s esophagus. The observations in our study need to be extended so that the long-term cancer risk of younger patients with Barrett’s esophagus can be estimated. Finally, risk factors for development of adenocarcinoma in patients with Barrett’s esophagus are not fully defined. Although factors such as obesity, smoking, and gastroesophageal reflux disease symptoms have been identified as risks for esophageal adenocarcinoma,35–38 Barrett’s length, hiatal hernia, aneuploidy, and p53 loss of heterozygosity are the only factors identified as risks in patients with Barrett’s esophagus.27,39 High-risk patient groups with Barrett’s esophagus could be identified, and surveillance strategies could be focused on this patient group, saving health care dollars. Another important finding was the incidence and outcome of LGD in this large cohort. The incidence of LGD was 4.3% per year, with a cancer incidence in patients with LGD of 0.6% per year. This was not significantly different than the incidence of cancer in the entire cohort. The majority of patients with LGD regressed to nondysplastic epithelium, 66% at their last surveillance endoscopy. This confirms data from previous studies.27,40 If the cancer risk in patients with LGD is indeed similar to those without dysplasia, they could potentially be managed similarly to Barrett’s esophagus patients without dysplasia and could probably forego more frequent endoscopic surveillance once it has been documented that LGD is the worst lesion detected with endoscopic biopsies. However, the major problems with LGD include variable pathologic classifications and the interobserver disagreement in the reading of LGD.41 This is an important issue because the finding of dysplasia, especially LGD, drives the cost of surveillance endoscopy.42,43 There are a number of limitations of this study. It is a database study, and surveillance intervals and biopsy protocols were not standardized between the centers, central pathology readings were not performed, and referral bias might also be a factor. Jumbo biopsies during surveillance were not routinely performed. The databases at each center were not initially created for such a study but rather as single center initiatives. The cohort is composed mainly of white men, emphasized by the 2 Veterans Affairs centers, and is not reflective of the general population but of the majority of patients with Barrett’s esophagus. Both Barrett’s esophagus (2:1, male to female) and esophageal adenocarcinoma (7:1, male to female) are male-predominant diseases, and the population in this data set is reflective of those at risk for esophageal adenocarcinoma. The large number of
571
patients in the database and the relatively long patient-years of follow-up are the major strengths. The results pertaining to the outcome of LGD might also be more generalizable to the community setting, given the varied nature of the participating centers and their geographical location. Another limitation of this study is the lack of information on patients who were lost to follow-up in the surveillance program. These patients were not contacted to determine whether they developed cancer or had any other medical health issues, or what their causes of mortality were. Given that there is no link to a cancer registry in these areas, this information has not been documented. These differences in follow-up in the groups could therefore impact the cancer incidence estimates; the cancer rate might be much lower than presented in this report. It should also be noted that the majority of the patients in this study were from a veterans’ population, which might affect the generalizability of these results. This Barrett’s esophagus study is an effort on the part of multiple investigators at different institutions to collaborate to establish the natural history of this disease and to evaluate endoscopic outcomes. This study has made it possible for the first time in the United States to integrate databases involving a large number of patients with Barrett’s esophagus (who have undergone extensive prospective follow-up) and offers an opportunity to enhance our understanding of this premalignant lesion in accordance with recommendations from a recent Program Review Group on esophageal and stomach cancers and the National Institute of Health.44,45 In conclusion, preliminary results from this multicenter effort provide estimates for the prevalence and incidence of dysplasia and cancer in a large combined cohort of patients with Barrett’s esophagus. These data can be an important resource for future research as well as for defining the risk factors for progression to HGD and adenocarcinoma in patients with Barrett’s esophagus.
References 1. Winters C, Spurling TC, Chobanian SJ, et al. Barrett’s esophagus: a prevalent occult complication of gastroesophageal reflux disease. Gastroenterology 1987;92:118 –124. 2. Hamilton SR, Smith R, Cameron JL. Prevalence and characteristics of Barrett esophagus in patients with adenocarcinoma of the esophagus or esophagogastric junction. Hum Pathol 1988;19:942–948. 3. Blot WJ, Devesa SS, Kneller RW, et al. Rising incidence of adenocarcinoma of the esophagus and gastric cardia. JAMA 1991; 265:1287–1289. 4. Devesa SS, Blot WJ, Fraumeni JF. Changing patterns in the incidence of esophageal and gastric carcinoma in the United States. Cancer 1998;83:2049 –2053. 5. Brown LM, Devesa SS. Epidemiologic trends in esophageal and gastric cancer in the United States. Surg Oncol Clin N Am 2002; 11:235–256. 6. Hammeeteman W, Tytgat NJ, Houthoff HJ, et al. Barrett’s esoph-
572
7.
8.
9.
10.
11.
12.
13. 14.
15.
16.
17. 18. 19. 20.
21.
22.
23.
24.
25.
26.
27.
28.
SHARMA ET AL
agus: development of dysplasia and adenocarcinoma. Gastroenterology 1989;96:1249 –1256. Robertson CS, Mayberry JF, Nicholson DA, et al. Value of endoscopic surveillance in detection of neoplastic change in Barrett’s oesophagus. Br J Surg 1988;75:760 –763. Miros M, Kerlin P, Walker N. Only patients with dysplasia progress to adenocarcinoma in Barrett’s oesophagus. Gut 1991; 32:1441–1446. Drewitz DJ, Sampliner RE, Garewal HS. The incidence of adenocarcinoma in Barrett’s esophagus: a prospective study of 170 patients followed 4.8 years. Am J Gastroenterol 1997;92:212–215. O’Connor JB, Falk GW, Richter JE. The incidence of adenocarcinoma and dysplasia in Barrett’s esophagus. Am J Gastroenterol 1999;94:2037–2042. Spechler SJ, Lee E, Ahnen D, et al. Long-term outcome of medical and surgical therapies for gastroesophageal reflux disease: follow-up of randomized controlled trial. JAMA 2001;285:2331–2338. Rudolph RE, Vaughan TL, Storer BE, et al. Effect of segment length on risk for neoplastic progression in patients with Barrett esophagus. Ann Intern Med 2000;132:612– 620. Heading RC. Barrett’s oesophagus: epidemiology comes up with a surprise. Gut 2003;52:1079 –1080. Anderson LA, Murray LJ, Murphy SJ, et al. Mortality in Barrett’s oesophagus: results from a population based study. Gut 2003; 52:1081–1084. Shaheen NJ, Crosby MA, Bozymski EM, et al. Is there publication bias in the reporting of cancer risk in Barrett’s esophagus? Gastroenterology 2000;119:333–338. Sampliner RE, Practice Parameters Committee ACG. Updated guidelines for the diagnosis, surveillance and therapy of Barrett’s esophagus. Am J Gastroenterol 2002;97:1888 –1895. Falk GW. Barrett’s esophagus. Gastroenterology 2002;122: 1569 –1591. Fennerty MB. Endoscopic diagnosis and surveillance of Barrett’s esophagus. Gastrointest Endosc Clin N Am 2003;13:257–267. Jankowski JA, Harrison RF, Perry I, et al. Barrett’s metaplasia. Lancet 2000;356:2079 –2085. Corley DA, Levin TR, Habel LA, et al. Surveillance and survival in Barrett’s adenocarcinomas: a population-based study. Gastroenterology 2002;122:633– 640. Provenzale D, Schmitt C, Wong JB. Barrett’s esophagus: a new look at surveillance based on emerging estimates of cancer risk. Am J Gastroenterol 1999;94:2043–2053. Weston AP, Krmpotich PT, Cherian R, et al. Prospective long-term endoscopic and histological follow-up of short segment Barrett’s esophagus: comparison with traditional long segment Barrett’s esophagus. Am J Gastroenterol 1997;92:407– 413. Lipscomb SJ, Schoenfeld P, Johnston M, et al. The incidence of high grade dysplasia and adenocarcinoma among patients with Barrett’s esophagus: a cohort of 154 patients followed for a total of 471 patient-years. Gastroenterology 1999;116:A238. McClave SA, Boyce HW, Gottfried MR. Early diagnosis of columnar-lined esosphagus: a new endoscopic criterion. Gastrointest Endosc 1987;33:413– 416. Riddell RH, Goldman H, Ransohoff DF, et al. Dysplasia in inflammatory bowel disease: standardized classification with provisional clinical applications. Hum Pathol 1983;14:931–968. Inadomi JM, Sampliner R, Lagergren J, et al. Screening and surveillance for Barrett esophagus in high-risk groups: a cost utility analysis. Ann Intern Med 2003;138:176 –186. Reid BJ, Prevo LJ, Galipeau PC, et al. Predictors of progression in Barrett’s esophagus II: Baseline 17p (p53) loss of heterozygosity identifies a patient subset at increased risk for neoplastic progression. Am J Gastroenterol 2001;96:2839 –2848. Falk GW, Ours TM, Richter JE. Practice patterns for surveillance of Barrett’s esophagus in the United States. Gastrointest Endosc 2000;52:197–203.
CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 5
29. Gross CP, Cruz-Correa M, Canto MI, et al. The adoption of ablation therapy for Barrett’s esophagus: a cohort study of gastroenterologists. Am J Gastroenterol 2002;97:279 –286. 30. van Sandick JW, van Lanschot JJB, Kuiken BW, et al. Impact of endoscopic biopsy surveillance of Barrett’s oesophagus on pathological stage and clinical outcome of Barrett’s carcinoma. Gut 1998;43:216 –222. 31. Streitz JM, Andrews CW Jr, Ellis FH Jr. Endoscopic surveillance of Barrett’s esophagus: does it help? J Thorac Cardiovasc Surg 1993;105:383–388. 32. Peters JH, Clark GWB, Ireland AP, et al. Outcome of adenocarcinoma arising in Barrett’s esophagus in endoscopically surveyed and nonsurveyed patients. J Thorac Cardiovasc Surg 1994;108: 813– 822. 33. Fitzgerald RC, Saeed IT, Khoo D, et al. Rigorous surveillance protocol increases detection of curable cancers associated with Barrett’s esophagus. Dig Dis Sci 2001;46:1892–1898. 34. Spechler SJ. Barrett’s esophagus: an overrated cancer risk factor. Gastroenterology 2000;119:587–589. 35. Lagergren J, Bergstrom R, Lindgren A, et al. Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N Engl J Med 1999;340:825– 831. 36. Lagergren J, Bergstrom R, Adami HO, et al. Association between medications that relax the lower esophageal sphincter and risk for esophageal adenocarcinoma. Ann Intern Med 2000;133: 165–175. 37. Gammon MD, Schoenberg JB, Ahsan H, et al. Tobacco, alcohol, and socioeconomic status and adenocarcinomas of the esophagus and gastric cardia. J Natl Cancer Inst 1997;89:1277–1284. 38. Lagergren J, Bergstrom R, Nyren O. Association between body mass and adenocarcinoma of the esophagus and gastric cardia. Ann Intern Med 1999;130:883– 890. 39. Weston AP, Badr AS, Hassanein RS. Prospective multivariate analysis of clinical, endoscopic, and histological factors predictive of the development of Barrett’s multifocal high-grade dysplasia or adenocarcinoma. Am J Gastroenterol 1999;94:3413–3419. 40. Schnell TG, Sontag SJ, Chejfec G, et al. Long-term nonsurgical management of Barrett’s esophagus with high-grade dysplasia. Gastroenterology 2001;120:1607–1619. 41. Montgomery E, Bronner MP, Goldblum JR, et al. Reproducibility of the diagnosis of dysplasia in Barrett esophagus: a reaffirmation. Hum Pathol 2001;32:368 –378. 42. Ofman JJ, Lewin K, Ramers C, et al. The economic impact of the diagnosis of dysplasia in Barrett’s esophagus. Am J Gastroenterol 2000;95:2946 –2952. 43. Sharma P, Morales TG, Sampliner RE. Short segment Barrett’s esophagus: the need for standardization of the definition and of endoscopic criteria. Am J Gastroenterol 1998;93:1033–1036. 44. Report of the stomach/esophageal cancers progress review group: National Cancer Institute. Available at: http://prg.nci.nih. gov/stomach/stomach_esophageal.pdf Accessed July 2005. 45. Lieberman D, Hamilton F. NIH-ADHF Workshop on Endoscopy Priorities: workshop statement and recommendations. Gastrointest Endosc 1999;49:S3–S4.
Address requests for reprints to: Prateek Sharma, MD, Gastroenterology Section (111), Department of Veterans Affairs Medical Center, 4801 E Linwood Blvd, Kansas City, MO 64128-2295. e-mail: [email protected]; fax: 816-922-4692. Supported by the Veterans Affairs Medical Center, Kansas City, Mo, and the American Society for Gastrointestinal Endoscopy (ASGE) Career Development Award. Presented in part at the plenary sessions of Digestive Diseases Week 2000 and 2001 and the American College of Gastroenterology Annual Meeting 2001.
ORIGINAL ARTICLES Dysplasia and Cancer in a Large Multicenter Cohort of Patients With Barrett’s Esophagus PRATEEK SHARMA,* GARY W. FALK,‡ ALLAN P. WESTON,* DEAN REKER,* MARK JOHNSTON,§ and RICHARD E. SAMPLINER¶ *University of Kansas School of Medicine & Veterans Affairs Medical Center, Kansas City, Missouri; ‡Cleveland Clinic Foundation, Cleveland, Ohio; §Bethesda Naval Medical Center, Bethesda, Maryland; and ¶University of Arizona Health Science Center and Southern Arizona VA Health Care System, Tucson, Arizona
Background & Aims: The exact incidence of adenocarcinoma in patients with Barrett’s esophagus (BE) is not known and is reported to vary from 0.2%–2% per year. Published series of patients with BE have included relatively small numbers of patients with limited duration of follow-up. The goal of this study was to define the prevalence and incidence of dysplasia and cancer and evaluate the paths of progression in a large multicenter cohort of BE patients. Methods: The BE study is a multicenter clinical and endoscopic outcomes project involving a single large database of patients with BE. Data from each of the participating centers were merged into the main study database. Cancers and HGD occurring within 12 months of the index endoscopy were regarded as prevalent cases. Results: One thousand three hundred seventy-six patients met the study criteria (95% white, 14% women); 91 patients had cancer at the initial endoscopy (prevalent cases, 6.7%; 95% confidence interval [CI], 4.8%– 8.7%). Six hundred eighteen patients were followed for a total of 2546 patient-years; mean follow-up was 4.12 years. Twelve patients developed cancer during follow-up, a cancer incidence of 1 in 212 patient-years of follow-up (0.5% per year; 95% CI, 0%–1.1%). The combined incidence of HGD and/or cancer was 1 in 75 patient-years of follow-up or 1.3% per year (95% CI, 0%–2.2%). Of the 34 patients developing HGD and/or cancer, 18 patients (53%) had at least 2 initial consecutive endoscopies with biopsies revealing nondysplastic mucosa. The incidence of LGD was 4.3% per year (95% CI, 2.8%– 6.0%). In the 156 patients with LGD, regression to no dysplasia occurred in 66%, persistent LGD in 21%, and progression to HGD/cancer in 13%. The incidence of cancer in patients with LGD was 1 in 156 patient-years of follow-up or 0.6% per year (95% CI, 0%–1.3%). Conclusions: Preliminary results from this trial define the prevalence and incidence of dysplasia and cancer in a multicenter cohort of patients with BE. At least half the patients who developed HGD and/or cancer had 2 consecutive initial endoscopies with biopsies revealing nondysplastic mucosa. The majority of patients with LGD regressed and had a cancer incidence similar to all BE patients.
arrett’s esophagus is a known complication of chronic gastroesophageal reflux disease and is a premalignant lesion for adenocarcinoma of the esophagus and the esophagogastric junction.1,2 Initial Surveillance, Epidemiology & End Results (SEER) program data (1980s) demonstrated a rapidly rising incidence for esophageal adenocarcinoma, and subsequent updated reports through 2001 have confirmed a continuing increase in the incidence of this cancer in the United States.3–5 The rise in incidence has been most marked among white men. The exact incidence of adenocarcinoma in patients with Barrett’s esophagus is not clear and has been reported to vary from 1 in 50 patient-years of follow-up to 1 in 285 patient-years of follow-up.6 –12 The published incidence data in cohorts of patients with Barrett’s esophagus have been influenced by, among other factors, the small number of patients being evaluated, short duration of follow-up, as well as by the demographics of the population being studied.13,14 Other biases have been reported, including publication bias that shows that smaller studies have reported a higher incidence of cancer, whereas relatively large studies show a lower incidence of cancer.15 To date, the maximum number of patients included in a prospective study determining the incidence of adenocarcinoma has been 327 patients.12 Surveillance endoscopy in patients with Barrett’s esophagus has been recommended for the early detection of dysplasia and cancer16; however, the time interval of progression to high-grade dysplasia (HGD) or cancer is not known, and the surveillance intervals are not clearly defined.17,18 The burden of endoscopic surveillance of patients with Barrett’s esophagus is enormous, and the role of surveillance is very important but remains controversial.19,20
B
Abbreviations used in this paper: CI, confidence interval; HGD, highgrade dysplasia; LGD, low-grade dysplasia. © 2006 by the American Gastroenterological Association Institute 1542-3565/06/$32.00 doi:10.1016/j.cgh.2006.03.001
May 2006
DYSPLASIA AND CANCER IN PATIENTS WITH BARRETT’S ESOPHAGUS
Moreover, the cost-effectiveness of surveillance endoscopy in patients with Barrett’s esophagus is very sensitive to the incidence of cancer.21 To address these gaps in our knowledge of the epidemiology of Barrett’s esophagus–associated dysplasia and carcinoma, we developed a consortium of investigators and centers interested in this disease entity to develop a large cohort of Barrett’s esophagus patients followed prospectively. The goals of this project were (1) to create a single, large database of patients with Barrett’s esophagus based on standardized criteria and (2) to calculate the prevalent and incident cases of adenocarcinoma, HGD, and low-grade dysplasia (LGD) in this group.
Methods Patients Patients diagnosed with Barrett’s esophagus at each of the participating centers, Veterans Affairs Medical Center, Kansas City, Mo, Southern Arizona Veterans Affairs Health Care System, Tucson, Ariz, Bethesda Naval Medical Center, Bethesda, Md, and Cleveland Clinic Foundation, Cleveland Ohio, were identified from the database of each center. From each center, the following information regarding all patients with Barrett’s esophagus was obtained: age, gender, ethnicity, length of Barrett’s esophagus, date of each endoscopy, and histologic diagnosis at each endoscopy. The duration of follow-up of each patient was calculated from the time of initial diagnosis of Barrett’s esophagus (ie, from their first endoscopy) to their most recent endoscopy with biopsy. The time of occurrence of LGD, HGD, and adenocarcinoma was documented in the database, thus allowing for the recording of prevalence and incidence cases of HGD and adenocarcinoma. The patients reported in this analysis met the following criteria: (1) a standardized definition of Barrett’s esophagus, the presence of columnar mucosa in the distal esophagus of any length with intestinal metaplasia documented on histology; (2) follow-up for at least 12 months from the time of initial diagnosis (ie, after first endoscopy and biopsy); and (3) adenocarcinoma and HGD patients detected within the first 12 months of their initial Barrett’s esophagus diagnosis were defined as prevalence cases. Patients were not included in the analysis if (1) there was columnar mucosa in the distal esophagus but no intestinal metaplasia on biopsy, or (2) patients had undergone endoscopic ablation therapy.
Endoscopy, Biopsy, & Dysplasia Recording At each center, patients with Barrett’s esophagus were defined by using a combination of endoscopic and histologic criteria.9,10 Once Barrett’s esophagus was documented, patients were enrolled in a surveillance program. Although endoscopic surveillance intervals were not standardized as part of the study protocol, each center followed the recommendations of the published guidelines.16 During surveillance visits, each patient underwent endoscopy with biopsies of the Barrett’s segment. Again, the biopsy protocol had not been standardized between the cen-
567
ters; each center followed a rigorous biopsy protocol of at least 4 quadrant biopsies every 2 cm with either a standard or jumbo biopsy forceps. The patient demographics, length of Barrett’s esophagus, and dysplasia grade from their baseline/initial endoscopy were recorded, and at subsequent endoscopies the grade of dysplasia for each visit was also noted. At each center, the length of Barrett’s esophagus was recorded in centimeters and judged by measuring the distance from the most proximally displaced squamocolumnar junction to the anatomic gastroesophageal junction.22–24 Dysplasia in the histology specimens was recorded according to established criteria for LGD, HGD, and adenocarcinoma and was reported by specialized gastroenterology pathologists at each site.25 Central reading of pathology was not performed for this analysis.
Data Entry, Management, and Analysis The study was approved by the Institutional Review Board of each institution. Each patient was provided with a unique identification number, and all patient identifiers were deleted in compliance with Health Information Portability and Accountability Act regulations. Data were generated directly from the existing data sets into the main study database. The database structure was designed by using Microsoft Access for Windows 1997 (Microsoft Corp, Redmond, Wash). The study coordinators at each center were responsible for collecting patient information and data entry as outlined above. Missing information was obtained by chart abstraction at each site. All collected and merged data were double entered, compared, and reconciled for accuracy. Data analysis was performed by using SAS version 8.2 (SAS Institute, Inc, Cary, NC), and survival curves were constructed by using the Kaplan-Meier method. Data collection, management, and analysis were supervised by a health services researcher (D.R.) at the Veterans Affairs Medical Center, Kansas City, Mo. The prevalent cases of dysplasia were defined as the percentage of patients with Barrett’s esophagus who had dysplasia at initial endoscopy. The follow-up interval for each patient was calculated as the time between the first endoscopy revealing Barrett’s esophagus and the most recent follow-up examination. The incidence was calculated from the number of patients lacking dysplasia at the baseline endoscopy who developed dysplasia during followup. In the calculation of the incidence of Barrett’s esophagus– related adenocarcinoma, patients found to have Barrett’s esophagus at the same time as the diagnosis of an adenocarcinoma or within 12 months of the initial diagnosis of Barrett’s esophagus were excluded and considered as prevalent cases.
Results Demographics and Prevalence Data From the 4 participating centers, 1376 patients met the study criteria and had at least one endoscopy with biopsy documenting Barrett’s esophagus. The mean age of this cohort was 59 ⫾ 12.3 years (standard deviation) and included 83% men and 17% women. The majority of the
568
SHARMA ET AL
CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 5
Table 1. Prevalence and Incidence of Dysplasia and Adenocarcinoma in Patients With Barrett’s Esophagus
Total number of patients Patients with LGD Patients with HGD Patients with adenocarcinoma aAnnual
Prevalence cases
Incidence cases
Incidence (per year)
1376 101 (7.3%) 42 (3%) 91 (6.6%)
618 100 22 12
4.3%a (95% CI, 2.8%–6.0%) 0.9%a (95% CI, 0%–1.7%) 0.5%a (95% CI, 0%–1.1%)
incidence.
patients were white (97%). This cohort was used to calculate the prevalent cases. At the first recorded endoscopy, esophageal adenocarcinoma was diagnosed in 91 patients, HGD in 42, and LGD in 101 patients (Table 1). Incidence of High-Grade Dysplasia and Adenocarcinoma Of 1376 patients meeting the initial study criteria, only 618 patients had follow-up of at least 1 year with a repeat endoscopy. One hundred sixty-five patients had a follow-up of less than 1 year and therefore were not included in the analysis of the incidence follow-up cases. Moreover, 549 patients had only a single endoscopy, and they either were not due for their repeat endoscopy or were lost to follow-up. These 618 patients were followed for a total of 2546 patient-years with a mean follow-up period of 4.12 years (range, 1–22.5 years). Twelve patients developed esophageal adenocarcinoma during follow-up, a cancer incidence of 1 in 212 patient-years of follow-up or 0.5% per year (95% confidence interval [CI], 0%–1.1%). The mean time to development of cancer was 5.3 years (range, 2.4 – 11.2 years). Seven of 12 patients had HGD before cancer development, 2 had only LGD (ie, were never detected with HGD during follow-up), whereas 3 developed cancer from nondysplastic Barrett’s esophagus (ie, were never detected with LGD or HGD at any instance) (Figure 1). Twenty-two patients developed HGD during follow-up, resulting in an HGD incidence of 1 in 116 patient-years of follow-up or 0.9% per year (95% CI, 0%–1.7%). The mean time to HGD development was 3.8 years (range, 1.2–7.9 years). Eleven of 22 patients developed HGD from Barrett’s esophagus without dysplasia, whereas the remaining 11 progressed from LGD to HGD. Thus, the total number of patients developing HGD and/or cancer during follow-up were 34, and the combined incidence of HGD and/or cancer was 1 in 75 patient-years of follow-up or 1.3% per year (95% CI, 0%–2.3%) (Table 1). The rate of cancer development (ie, annual incidence of adenocarcinoma in Barrett’s esophagus patients) in the different participating centers varied from none at the Bethesda Naval Medical Center (total number of patients, 75) to 0.4% at the Cleveland Clinic Foundation (total number of patients, 241), 0.6% at the Southern Arizona Veterans
Affairs Medical Center (total number of patients, 135), and 0.8% at the Kansas City Veterans Affairs Medical Center (total number of patients, 167) (not significantly different). The time periods of enrollment and database entry periods were as follows: Bethesda 1993–2000, Cleveland Clinic Foundation 1979 –2000, Kansas City Veterans Affairs Medical Center 1984 –1993, and Southern Arizona Veterans Affairs Medical Center 1975–1998. The annual rate of HGD development (incidence) was 0.7% each at the Kansas City Veterans Affairs Medical Center and the Cleveland Clinic Foundation and was 1.1% each at the Bethesda Naval Medical Center and the Southern Arizona Veterans Affairs Medical Center. The overall incidence of either cancer and/or HGD development varied from 1.0% per year to 1.7% per year at the different study sites. As shown in the Kaplan-Meier survival graphs (Figure 2), at 5 years of follow-up (N ⫽ 157), 97% of patients were free of cancer, and at 8 years 95% were free of cancer (N ⫽ 66). This indicates that at least for the initial 5– 8 years of follow-up after a diagnosis of Barrett’s esophagus, the cancer incidence was approximately 5% (95% CI, 3.3%– 6.7%), although the number of patients at 8 years in this series was only 66.
Figure 1. Progression to adenocarcinoma in patients with Barrett’s esophagus during follow-up. Twelve patients developed cancer during a mean follow-up of 4.12 years. The grades of dysplasia at baseline, during follow-up, and at the final endoscopy are also shown. The number of patients is shown in parentheses.
May 2006
DYSPLASIA AND CANCER IN PATIENTS WITH BARRETT’S ESOPHAGUS
569
Figure 2. Barrett’s esophagus and development of cancer: event distribution. Kaplan-Meier curve showing the percentage of patients free of esophageal adenocarcinoma during follow-up. There was a gradual decrease in the number of patients as the duration of follow-up increased.
Progression to High-Grade Dysplasia and Adenocarcinoma: Findings at Initial Endoscopy Of the 34 patients developing either HGD and/or cancer during follow-up, 18 patients (53%) had at least 2 initial consecutive endoscopies documenting nondysplastic tissue, intestinal metaplasia without dysplasia. The mean time interval between these 2 consecutive endoscopies was 2 years (range, 1 month–3 years). After the second endoscopy that documented intestinal metaplasia without dysplasia, HGD/cancer was detected after a mean of 3 years (range, 3 months–7.8 years). The pathways of progression from nondysplastic Barrett’s esophagus to HGD and cancer are shown in Figures 1 and 3.
patients developing cancer during follow-up. Many patients with LGD (42%) had regression to nondysplastic epithelium (ie, intestinal metaplasia) at their subsequent endoscopy, whereas 32% had intermittent LGD (ie, went back and forth between LGD and nondysplastic Barrett’s esophagus during their follow-up period).
Incidence and Outcome of Low-Grade Dysplasia Of the initial cohort of 618 patients with at least 1 year of follow-up, the diagnosis of LGD during follow-up was confirmed in 100 patients, an LGD incidence of 4.3% per year (95% CI, 2.8%– 6%). Besides these 100 patients who developed LGD during follow-up, an additional 56 patients who had LGD at their baseline endoscopy also had further surveillance and follow-up. Thus, a total of 156 patients with LGD were followed for at least 1 year with at least 2 assessments for a mean duration of 5 years (range, 1–15.5 years). The outcome of LGD in these patients at their most recent endoscopy is shown in Figure 4. The incidence of cancer in patients with LGD was 1 in 156 patient-years of follow-up or 0.6% per year (95% CI, 0%–2%), with a total of 5
Figure 3. Even patients with nondysplastic Barrett’s esophagus progressed to HGD and cancer during follow-up. The number of patients is shown in parentheses.
570
SHARMA ET AL
Figure 4. Outcome of patients with LGD during follow-up. The majority of the patients regressed to nondysplastic Barrett’s esophagus.
Cancer Incidence Estimates for Patients Lost to Follow-Up There were 549 cases that were excluded from the “incidence data analysis” because of a single endoscopy or loss to follow-up. If we assume that none of these cases would have developed cancer during a mean follow-up of 4 years, the numerator would remain as 12 incidence cancer cases during 4732 patient-years of follow-up, giving an incidence of 0.2% per year (95% CI, 0%– 0.4%). This would represent the most conservative cancer rate estimate for the entire cohort. However, if we assume an additional 12 cases of cancer to develop during a mean follow-up of 4 years for these 549 patients, this would provide 24 cases during 4732 patient years of follow-up, giving an annual incidence of 0.5% per year (95% CI, 0.1%– 0.9%).
Discussion The number of incidence cases of dysplasia (both LGD and HGD) and cancer in this study was much lower than the prevalent cases, indicating that the majority of the neoplastic lesions are detected at the initial endoscopy with biopsy. On the other hand, HGD and adenocarcinoma can develop within 1–2 years of the initial endoscopy revealing nondysplastic Barrett’s esophagus. These data highlight the need for better risk stratification of patients with Barrett’s esophagus, given that only the minority will develop HGD or cancer and patients can progress to cancer from nondysplastic tissue. Moreover, at least half the patients who developed HGD and/or cancer had 2 consecutive initial endoscopies revealing intestinal metaplasia without dysplasia (mean time to progression, 3 years). Some authors have suggested that patients with an initial endoscopy lacking dysplasia might not require follow-up.26 Even patients with 2 consecutive findings of intestinal metaplasia without dysplasia can develop neoplasia changes. Factors other than a
CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 5
negative initial dysplasia grade need to be evaluated for risk stratification for HGD/cancer development.27 The majority of the HGDs and cancers were detected as prevalent cases. Although the exact indication for endoscopy at the index time period was not recorded in the patients and available in the database, one can speculate that the greatest advantage of diagnosing and recognizing lesions in patients suspected to harbor Barrett’s esophagus or dysplasia cancer occurs at the index endoscopy. This might be due to a number of factors including referral bias, alarm symptoms at the time of diagnosis, and selection bias. Surveillance of patients with known Barrett’s esophagus is recommended for the early detection of dysplasia and cancer, a practice followed by the majority of practicing gastroenterologists in the United States.16,28,29 Patients undergoing surveillance endoscopy for Barrett’s esophagus have been noted in retrospective studies and in a recent population-based study to have better survival when cancer is detected, as opposed to patients presenting with dysphagia and a prevalent adenocarcinoma with Barrett’s esophagus.20,30 –33 There are no prospective studies indicating a survival benefit of surveillance versus no surveillance. Surveillance of patients with Barrett’s esophagus has raised a number of issues including cost-effectiveness of surveillance, given the relatively low risk of cancer.34 A recent study with a computer model found that if the risk of cancer in Barrett’s esophagus patients is 0.4% per year, the most cost-effective surveillance strategy would be endoscopy every 5 years.21 Another recent model determined that surveillance only for patients with dysplasia would be a costeffective strategy.26 Both models are very sensitive to the incidence of dysplasia and adenocarcinoma in patients with Barrett’s esophagus, and any surveillance program has to take the incidence estimate into account. Therefore, determining the accurate incidence of dysplasia and cancer is of critical importance to the cost-effectiveness of a surveillance program in patients with Barrett’s esophagus. The incidence of adenocarcinoma in patients with Barrett’s esophagus has ranged from 0.2%–2% per year.6 –11 In a study by Hammeeteman et al6 in the 1980s, the annual risk was 1.9%, whereas in 1997 Drewitz et al9 reported an annual incidence of 0.5%. A number of factors including study design, retrospective nature of study, short duration of follow-up, few patients, lead time bias, length time bias, and healthy volunteer bias might have contributed to this discrepancy in the literature. A recent report also cited publication bias as a potential cause for this, with small studies reporting a higher incidence of cancer and larger studies reporting a lower incidence.13 The authors estimated the true incidence of cancer to be approximately
May 2006
DYSPLASIA AND CANCER IN PATIENTS WITH BARRETT’S ESOPHAGUS
0.5% per year, similar to our results in a large cohort of patients with Barrett’s esophagus. Another unresolved issue is whether cancer incidence changes over time in patients with Barrett’s esophagus. The observations in our study need to be extended so that the long-term cancer risk of younger patients with Barrett’s esophagus can be estimated. Finally, risk factors for development of adenocarcinoma in patients with Barrett’s esophagus are not fully defined. Although factors such as obesity, smoking, and gastroesophageal reflux disease symptoms have been identified as risks for esophageal adenocarcinoma,35–38 Barrett’s length, hiatal hernia, aneuploidy, and p53 loss of heterozygosity are the only factors identified as risks in patients with Barrett’s esophagus.27,39 High-risk patient groups with Barrett’s esophagus could be identified, and surveillance strategies could be focused on this patient group, saving health care dollars. Another important finding was the incidence and outcome of LGD in this large cohort. The incidence of LGD was 4.3% per year, with a cancer incidence in patients with LGD of 0.6% per year. This was not significantly different than the incidence of cancer in the entire cohort. The majority of patients with LGD regressed to nondysplastic epithelium, 66% at their last surveillance endoscopy. This confirms data from previous studies.27,40 If the cancer risk in patients with LGD is indeed similar to those without dysplasia, they could potentially be managed similarly to Barrett’s esophagus patients without dysplasia and could probably forego more frequent endoscopic surveillance once it has been documented that LGD is the worst lesion detected with endoscopic biopsies. However, the major problems with LGD include variable pathologic classifications and the interobserver disagreement in the reading of LGD.41 This is an important issue because the finding of dysplasia, especially LGD, drives the cost of surveillance endoscopy.42,43 There are a number of limitations of this study. It is a database study, and surveillance intervals and biopsy protocols were not standardized between the centers, central pathology readings were not performed, and referral bias might also be a factor. Jumbo biopsies during surveillance were not routinely performed. The databases at each center were not initially created for such a study but rather as single center initiatives. The cohort is composed mainly of white men, emphasized by the 2 Veterans Affairs centers, and is not reflective of the general population but of the majority of patients with Barrett’s esophagus. Both Barrett’s esophagus (2:1, male to female) and esophageal adenocarcinoma (7:1, male to female) are male-predominant diseases, and the population in this data set is reflective of those at risk for esophageal adenocarcinoma. The large number of
571
patients in the database and the relatively long patient-years of follow-up are the major strengths. The results pertaining to the outcome of LGD might also be more generalizable to the community setting, given the varied nature of the participating centers and their geographical location. Another limitation of this study is the lack of information on patients who were lost to follow-up in the surveillance program. These patients were not contacted to determine whether they developed cancer or had any other medical health issues, or what their causes of mortality were. Given that there is no link to a cancer registry in these areas, this information has not been documented. These differences in follow-up in the groups could therefore impact the cancer incidence estimates; the cancer rate might be much lower than presented in this report. It should also be noted that the majority of the patients in this study were from a veterans’ population, which might affect the generalizability of these results. This Barrett’s esophagus study is an effort on the part of multiple investigators at different institutions to collaborate to establish the natural history of this disease and to evaluate endoscopic outcomes. This study has made it possible for the first time in the United States to integrate databases involving a large number of patients with Barrett’s esophagus (who have undergone extensive prospective follow-up) and offers an opportunity to enhance our understanding of this premalignant lesion in accordance with recommendations from a recent Program Review Group on esophageal and stomach cancers and the National Institute of Health.44,45 In conclusion, preliminary results from this multicenter effort provide estimates for the prevalence and incidence of dysplasia and cancer in a large combined cohort of patients with Barrett’s esophagus. These data can be an important resource for future research as well as for defining the risk factors for progression to HGD and adenocarcinoma in patients with Barrett’s esophagus.
References 1. Winters C, Spurling TC, Chobanian SJ, et al. Barrett’s esophagus: a prevalent occult complication of gastroesophageal reflux disease. Gastroenterology 1987;92:118 –124. 2. Hamilton SR, Smith R, Cameron JL. Prevalence and characteristics of Barrett esophagus in patients with adenocarcinoma of the esophagus or esophagogastric junction. Hum Pathol 1988;19:942–948. 3. Blot WJ, Devesa SS, Kneller RW, et al. Rising incidence of adenocarcinoma of the esophagus and gastric cardia. JAMA 1991; 265:1287–1289. 4. Devesa SS, Blot WJ, Fraumeni JF. Changing patterns in the incidence of esophageal and gastric carcinoma in the United States. Cancer 1998;83:2049 –2053. 5. Brown LM, Devesa SS. Epidemiologic trends in esophageal and gastric cancer in the United States. Surg Oncol Clin N Am 2002; 11:235–256. 6. Hammeeteman W, Tytgat NJ, Houthoff HJ, et al. Barrett’s esoph-
572
7.
8.
9.
10.
11.
12.
13. 14.
15.
16.
17. 18. 19. 20.
21.
22.
23.
24.
25.
26.
27.
28.
SHARMA ET AL
agus: development of dysplasia and adenocarcinoma. Gastroenterology 1989;96:1249 –1256. Robertson CS, Mayberry JF, Nicholson DA, et al. Value of endoscopic surveillance in detection of neoplastic change in Barrett’s oesophagus. Br J Surg 1988;75:760 –763. Miros M, Kerlin P, Walker N. Only patients with dysplasia progress to adenocarcinoma in Barrett’s oesophagus. Gut 1991; 32:1441–1446. Drewitz DJ, Sampliner RE, Garewal HS. The incidence of adenocarcinoma in Barrett’s esophagus: a prospective study of 170 patients followed 4.8 years. Am J Gastroenterol 1997;92:212–215. O’Connor JB, Falk GW, Richter JE. The incidence of adenocarcinoma and dysplasia in Barrett’s esophagus. Am J Gastroenterol 1999;94:2037–2042. Spechler SJ, Lee E, Ahnen D, et al. Long-term outcome of medical and surgical therapies for gastroesophageal reflux disease: follow-up of randomized controlled trial. JAMA 2001;285:2331–2338. Rudolph RE, Vaughan TL, Storer BE, et al. Effect of segment length on risk for neoplastic progression in patients with Barrett esophagus. Ann Intern Med 2000;132:612– 620. Heading RC. Barrett’s oesophagus: epidemiology comes up with a surprise. Gut 2003;52:1079 –1080. Anderson LA, Murray LJ, Murphy SJ, et al. Mortality in Barrett’s oesophagus: results from a population based study. Gut 2003; 52:1081–1084. Shaheen NJ, Crosby MA, Bozymski EM, et al. Is there publication bias in the reporting of cancer risk in Barrett’s esophagus? Gastroenterology 2000;119:333–338. Sampliner RE, Practice Parameters Committee ACG. Updated guidelines for the diagnosis, surveillance and therapy of Barrett’s esophagus. Am J Gastroenterol 2002;97:1888 –1895. Falk GW. Barrett’s esophagus. Gastroenterology 2002;122: 1569 –1591. Fennerty MB. Endoscopic diagnosis and surveillance of Barrett’s esophagus. Gastrointest Endosc Clin N Am 2003;13:257–267. Jankowski JA, Harrison RF, Perry I, et al. Barrett’s metaplasia. Lancet 2000;356:2079 –2085. Corley DA, Levin TR, Habel LA, et al. Surveillance and survival in Barrett’s adenocarcinomas: a population-based study. Gastroenterology 2002;122:633– 640. Provenzale D, Schmitt C, Wong JB. Barrett’s esophagus: a new look at surveillance based on emerging estimates of cancer risk. Am J Gastroenterol 1999;94:2043–2053. Weston AP, Krmpotich PT, Cherian R, et al. Prospective long-term endoscopic and histological follow-up of short segment Barrett’s esophagus: comparison with traditional long segment Barrett’s esophagus. Am J Gastroenterol 1997;92:407– 413. Lipscomb SJ, Schoenfeld P, Johnston M, et al. The incidence of high grade dysplasia and adenocarcinoma among patients with Barrett’s esophagus: a cohort of 154 patients followed for a total of 471 patient-years. Gastroenterology 1999;116:A238. McClave SA, Boyce HW, Gottfried MR. Early diagnosis of columnar-lined esosphagus: a new endoscopic criterion. Gastrointest Endosc 1987;33:413– 416. Riddell RH, Goldman H, Ransohoff DF, et al. Dysplasia in inflammatory bowel disease: standardized classification with provisional clinical applications. Hum Pathol 1983;14:931–968. Inadomi JM, Sampliner R, Lagergren J, et al. Screening and surveillance for Barrett esophagus in high-risk groups: a cost utility analysis. Ann Intern Med 2003;138:176 –186. Reid BJ, Prevo LJ, Galipeau PC, et al. Predictors of progression in Barrett’s esophagus II: Baseline 17p (p53) loss of heterozygosity identifies a patient subset at increased risk for neoplastic progression. Am J Gastroenterol 2001;96:2839 –2848. Falk GW, Ours TM, Richter JE. Practice patterns for surveillance of Barrett’s esophagus in the United States. Gastrointest Endosc 2000;52:197–203.
CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 4, No. 5
29. Gross CP, Cruz-Correa M, Canto MI, et al. The adoption of ablation therapy for Barrett’s esophagus: a cohort study of gastroenterologists. Am J Gastroenterol 2002;97:279 –286. 30. van Sandick JW, van Lanschot JJB, Kuiken BW, et al. Impact of endoscopic biopsy surveillance of Barrett’s oesophagus on pathological stage and clinical outcome of Barrett’s carcinoma. Gut 1998;43:216 –222. 31. Streitz JM, Andrews CW Jr, Ellis FH Jr. Endoscopic surveillance of Barrett’s esophagus: does it help? J Thorac Cardiovasc Surg 1993;105:383–388. 32. Peters JH, Clark GWB, Ireland AP, et al. Outcome of adenocarcinoma arising in Barrett’s esophagus in endoscopically surveyed and nonsurveyed patients. J Thorac Cardiovasc Surg 1994;108: 813– 822. 33. Fitzgerald RC, Saeed IT, Khoo D, et al. Rigorous surveillance protocol increases detection of curable cancers associated with Barrett’s esophagus. Dig Dis Sci 2001;46:1892–1898. 34. Spechler SJ. Barrett’s esophagus: an overrated cancer risk factor. Gastroenterology 2000;119:587–589. 35. Lagergren J, Bergstrom R, Lindgren A, et al. Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N Engl J Med 1999;340:825– 831. 36. Lagergren J, Bergstrom R, Adami HO, et al. Association between medications that relax the lower esophageal sphincter and risk for esophageal adenocarcinoma. Ann Intern Med 2000;133: 165–175. 37. Gammon MD, Schoenberg JB, Ahsan H, et al. Tobacco, alcohol, and socioeconomic status and adenocarcinomas of the esophagus and gastric cardia. J Natl Cancer Inst 1997;89:1277–1284. 38. Lagergren J, Bergstrom R, Nyren O. Association between body mass and adenocarcinoma of the esophagus and gastric cardia. Ann Intern Med 1999;130:883– 890. 39. Weston AP, Badr AS, Hassanein RS. Prospective multivariate analysis of clinical, endoscopic, and histological factors predictive of the development of Barrett’s multifocal high-grade dysplasia or adenocarcinoma. Am J Gastroenterol 1999;94:3413–3419. 40. Schnell TG, Sontag SJ, Chejfec G, et al. Long-term nonsurgical management of Barrett’s esophagus with high-grade dysplasia. Gastroenterology 2001;120:1607–1619. 41. Montgomery E, Bronner MP, Goldblum JR, et al. Reproducibility of the diagnosis of dysplasia in Barrett esophagus: a reaffirmation. Hum Pathol 2001;32:368 –378. 42. Ofman JJ, Lewin K, Ramers C, et al. The economic impact of the diagnosis of dysplasia in Barrett’s esophagus. Am J Gastroenterol 2000;95:2946 –2952. 43. Sharma P, Morales TG, Sampliner RE. Short segment Barrett’s esophagus: the need for standardization of the definition and of endoscopic criteria. Am J Gastroenterol 1998;93:1033–1036. 44. Report of the stomach/esophageal cancers progress review group: National Cancer Institute. Available at: http://prg.nci.nih. gov/stomach/stomach_esophageal.pdf Accessed July 2005. 45. Lieberman D, Hamilton F. NIH-ADHF Workshop on Endoscopy Priorities: workshop statement and recommendations. Gastrointest Endosc 1999;49:S3–S4.
Address requests for reprints to: Prateek Sharma, MD, Gastroenterology Section (111), Department of Veterans Affairs Medical Center, 4801 E Linwood Blvd, Kansas City, MO 64128-2295. e-mail: [email protected]; fax: 816-922-4692. Supported by the Veterans Affairs Medical Center, Kansas City, Mo, and the American Society for Gastrointestinal Endoscopy (ASGE) Career Development Award. Presented in part at the plenary sessions of Digestive Diseases Week 2000 and 2001 and the American College of Gastroenterology Annual Meeting 2001.