Barrett s esophagus (BE), a known complication of chronic

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2011;9: Patients With Nondysplastic Barrett s Esophagus Have Low Risks for Developing Dysplasia or Esophageal Adenocarcinoma SACHIN WANI,* GARY FALK, MATTHEW HALL,* SRINIVAS GADDAM,* AMY WANG, NEIL GUPTA,* MANDEEP SINGH,* VIKAS SINGH,* KENG YU CHUANG, VIKRAM BOOLCHAND, HEMANTH GAVINI, JOHN KUCZYNSKI, PRITI SUD, SAVIO REDDYMASU,* AJAY BANSAL,* AMIT RASTOGI,* SHARAD C. MATHUR,* PATRICK YOUNG,* BROOKS CASH, DAVID A. LIEBERMAN, RICHARD E. SAMPLINER, and PRATEEK SHARMA* *Division of Gastroenterology and Hepatology, Veterans Affairs Medical Center and University of Kansas School of Medicine, Kansas City, Missouri; Department of Gastroenterology and Hepatology, Cleveland Clinic Foundation, Cleveland, Ohio; Division of Gastroenterology and Hepatology, Veterans Affairs Medical Center and Oregon Health and Science University, Portland, Oregon; Department of Gastroenterology and Hepatology, Southern Arizona Veterans Affairs Health Care System and University of Arizona Health Science Center, Tucson, Arizona; Division of Gastroenterology and Hepatology, National Naval Medical Center, Bethesda, Maryland This article has an accompanying continuing medical education activity on page e26. Learning Objectives At the end of this activity, the learner will appreciate that the rate of progression to low-grade dysplasia is much higher than the incident rate per year for esophageal cancer for Barrett s esophagus; appreciate the risk factors for progression to esophageal cancer in patients with Barrett s esophagus; and recognize the wide variability in the previous reporting of progression of Barrett s esophagus to cancer. See editorial on page 194. BACKGROUND & AIMS: The risks of dysplasia and esophageal adenocarcinoma (EAC) are not clear for patients with nondysplastic Barrett s esophagus (NDBE); the rate of progression has been overestimated in previous studies. We studied the incidences of dysplasia and EAC and investigated factors associated with progression of BE. METHODS: The BE study is a multicenter outcomes project of a large cohort of patients with BE. Neoplasia was graded as low-grade dysplasia, high-grade dysplasia (HGD), or EAC. Patients followed up for at least 1 year after the index endoscopy examination were included, whereas those diagnosed with dysplasia and EAC within 1 year of diagnosis with BE (prevalent cases) were excluded. Of 3334 patients with BE, 1204 met the inclusion criteria (93.7% Caucasian; 88% male; mean age, 59.3 y) and were followed up for a mean of 5.52 years ( patient-years). RESULTS: Eighteen patients developed EAC (incidence, 0.27%/y; 95% confidence interval [CI], ) and 32 developed HGD (incidence, 0.48%/y; 95% CI, ). The incidence of HGD and EAC was 0.63%/y (95% CI, ). There were 217 cases of low-grade dysplasia (incidence, 3.6%/y; 95% CI, ). Five and 10 years after diagnosis, 98.6% (n 540) and 97.1% (n 155) of patients with NDBE were cancer free, respectively. The length of the BE was associated significantly with progression (EAC 6 cm, 0.09%/y vs EAC 6 cm, 0.65%/y; P 0.001). CONCLUSIONS: There is a lower incidence of dysplasia and EAC among patients with NDBE than previously reported. Because most patients are cancer free after a long-term follow-up period, surveillance intervals might be lengthened, especially for patients with shorter segments of BE. Keywords: Barrett s Esophagus; Dysplasia; Esophageal Adenocarcinoma; Esophageal Cancer; Screening; Surveillance; Prevention. Barrett s esophagus (BE), a known complication of chronic gastroesophageal reflux disease, is a well established premalignant lesion for esophageal and gastroesophageal adenocarcinoma. 1,2 Approximately 10% to 15% of patients with chronic gastroesophageal reflux disease are diagnosed with BE. In addition, BE has been reported in patients with no reflux symptoms. 3 The risk of esophageal adenocarcinoma (EAC) is increased 30 to 40 times among patients with BE compared with those without this condition. EAC continues to increase at a rate greater than any other cancer in the Western world ( 500% since the 1970s), exceeding that of other more common cancers such as breast, colon, lung, and prostate cancer. 4 In 2009, it is estimated that 16,470 new cases of esophageal cancer will be diagnosed in the United States, of which close to 60% will be adenocarcinomas. 5 Despite all the recent advances in the diagnosis and management of this lethal cancer, the overall 5-year survival rate remains dismal (15% 20%). 6 Although not evaluated in randomized controlled trials, surveillance of patients with BE is recommended by all major gastroenterology societies and published guidelines. 1,7 Multiple observational studies suggest that endoscopic surveillance is associated with detection of EAC at an earlier stage along with improved survival. 8,9 However, the burden of endoscopic surveillance of BE patients is significant and continues to generate a great deal of controversy. 10,11 In addition, there has been a lot of interest in the endoscopic ablation of nondysplastic BE (NDBE). The true incidence of EAC in patients with BE is central to determining the effectiveness of surveillance endoscopy or any intervention strategy. The exact incidence of EAC Abbreviations used in this paper: BE, Barrett s esophagus; CI, confidence interval; EAC, esophageal adenocarcinoma; HGD, high-grade dysplasia; LGD, low-grade dysplasia; NDBE, nondysplastic Barrett s esophagus; NSAID, nonsteroidal anti-inflammatory drug; PPI, proton pump inhibitor; SD, standard deviation by the AGA Institute /$36.00 doi: /j.cgh

2 March 2011 LOW INCIDENCE OF CANCER IN NONDYSPLASTIC BE 221 in BE patients is not clear and has ranged from 0.2% to 3.5% per year This considerable variation has been attributed to the small number of patients being evaluated, short duration of follow-up evaluation, retrospective study designs, lead-time bias, length-time bias, and publication bias, with selective publications of smaller studies reporting a higher incidence of cancer and relatively larger studies reporting a lower incidence of cancer. 13,18 We previously reported an incidence rate of 0.5% per year (95% confidence interval [CI], 0.0% 1.1%; 1 in 212 patient-years of follow-up evaluation) in this large multicenter cohort of 618 patients with NDBE (2546 patient-years; mean follow-up period, 4.12 y) in the United States. 11 In addition, data on the rate of progression of NDBE to the combined end point of high-grade dysplasia (HGD) and EAC are sparse. Given the advent of endoscopic eradication therapies for the management of HGD, rates of progression of NDBE to HGD and mucosal EAC assume greater importance. Although male predominance in patients with EAC is well recognized, the influence of sex on the rate of progression of NDBE to HGD/EAC is less clear and risk factors for progression have been poorly defined. 14,22 25 The vast majority of the natural history studies in BE do not report on demographics and endoscopic characteristics. Hence, the reported incidence rates do not account for potentially important predictors associated with disease progression in BE. Thus, the aims of this study were as follows: (1) to define the incidence rate of dysplasia and EAC in patients with NDBE, (2) to assess for gender differences in the incidence rates of dysplasia and EAC in a large multicenter cohort of NDBE patients, and (3) to identify predictors (demographic and endoscopic) for progression to dysplasia and EAC in patients with BE. Methods Patients The Barrett s esophagus study is a multicenter outcomes project that includes 5 tertiary care referral centers with an interest in BE. These include the Veterans Affairs Medical Center in Kansas City, Missouri; Southern Arizona Veterans Affairs Health Care System in Tucson, Arizona; Cleveland Clinic Foundation in Cleveland, Ohio; Veterans Affairs Medical Center in Portland, Oregon; and the National Naval Medical Center in Bethesda, Maryland. The study was approved by the Institutional Review Board at each institution. Patients diagnosed with BE at each of the participating centers were identified and entered into a standard Microsoft Access database (Microsoft Inc, Redmond, WA) at each center. The following information was recorded regarding all patients with BE: demographics (age, sex, and ethnicity), endoscopy results (date of procedure, presence of hiatal hernia, and length of BE), and histologic diagnosis at each endoscopic procedure. When available, data on use of aspirin, nonsteroidal anti-inflammatory drugs, proton pump inhibitors (PPIs), and history of smoking were recorded. The duration of follow-up evaluation of each patient was calculated from the time of initial diagnosis of BE (ie, from their first endoscopy) to their most recent endoscopy with biopsy. The time of occurrence of dysplasia (low-grade dysplasia [LGD], HGD) and EAC was documented in the database, which allowed for the recording of prevalent and incident cases of dysplasia and EAC. Patients diagnosed with LGD, HGD, and EAC within the first year of their initial diagnosis of BE were defined as prevalent cases. Patients diagnosed with dysplasia and EAC at least 1 year after the first endoscopic evaluation with biopsy were defined as incident cases. The inclusion criteria for this study were as follows: (1) patients who met the standardized definition of BE, that is, the presence of columnar-lined mucosa in the distal esophagus of any length at endoscopy and the presence of intestinal metaplasia documented on histology 1,7 and (2) an endoscopic follow-up period of at least 1 year from the time of initial diagnosis. Exclusion criteria for this analysis were as follows: (1) prevalent cases of dysplasia (LGD and HGD) and EAC, and (2) columnar-lined mucosa in the distal esophagus with no intestinal metaplasia on histology. Endoscopy, Surveillance, and Histopathology At the time of upper endoscopy, the length of BE was recorded in centimeters and calculated by measuring the distance from the anatomic gastroesophageal junction to the most proximally displaced squamocolumnar junction. The presence/ absence and size of hiatal hernia at each endoscopy also was recorded. Although endoscopic surveillance intervals were not standardized a priori as part of the study protocol, each center followed the recommendations of the published guidelines. 7,26 At the time of initial endoscopy and during surveillance visits, each patient underwent biopsies of the Barrett s segment. Again, although the biopsy protocol was not standardized a priori as part of the study protocol between the participating centers, each center followed a rigorous biopsy protocol of at least 4-quadrant biopsies every 2 cm with either a standard or jumbo biopsy forceps. Histopathologic assessment of biopsy specimens was reported using established criteria for NDBE, LGD, HGD, and EAC. 27,28 The worst histologic grade identified was the overall histologic grade for that endoscopy. These were reported by local expert gastroenterology pathologists at each site. Review of dysplasia and EAC slides by a second local expert pathologist was performed as part of routine clinical practice at each site. Central reading of all pathology specimens was not performed in this study. Data Management and Statistical Analysis The study coordinators at each center collected patient information and data entry was performed for all the earlier-stated variables. A database of all BE patients was created at each center and each patient was provided with a unique identification number. All patient identifiers were deleted in compliance with the Health Information Portability and Accountability Act regulations. Data sets from each center then were merged into the main study database at the Veterans Affairs Medical Center in Kansas City. This was performed using Microsoft Access for Windows 2007 (Microsoft Corp, Redmond, WA). All collected and merged data were compared and reconciled for accuracy. Data management and analysis was performed and supervised by a senior outcomes researcher (M.H.). The follow-up interval for each patient was calculated as the time between the first endoscopy revealing BE and the most recent follow-up endoscopy with biopsy. The incidence rates with 95% CIs of LGD, HGD, EAC, and HGD/EAC were calculated from the number of patients lacking dysplasia at the baseline endoscopy who developed these end points dur-

3 222 WANI ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 9, No. 3 and SDs. All analyses were performed using SAS 9.2 (SAS Institute, Cary, NC), and a P value of less than.05 was considered statistically significant. Results Of the 3334 patients with BE, 1204 patients met the inclusion criteria for this analysis. The reasons for exclusions are highlighted in Figure 1. The mean age of the patients in this cohort was 59.3 years (SD, y) and the vast majority were Caucasians (93.7%). This cohort included 88% males and 12% females with a mean follow-up period of 5.52 years (SD, 3.32 y) for a total of patient-years. The mean BE length was 3.8 cm (SD, 3.4 cm). Hiatal hernia was present in 56.8% of the BE patients with a mean size of 3.4 cm (SD, 1.84 cm). The distribution of prevalent cases were as follows: 165 patients with EAC, 164 patients with HGD, and 412 patients with LGD. Figure 1. Patient flow chart. IM, intestinal metaplasia. ing follow-up evaluation. Prevalent cases of LGD, HGD, and EAC were excluded from this analysis. Time-to-event analyses for progression to LGD, HGD, EAC, and HGD/EAC alone were performed. Among patients who progressed, we calculated the means and standard deviations (SDs) for the time (years) to LGD, HGD, EAC, and HGD/EAC from baseline endoscopy. Product-limit estimates to estimate the survival function using the Kaplan Meier method was performed. The log-rank test was used to compare survival functions across levels of covariates. Similarly, incidence rates with 95% CIs for men and women were calculated separately. The effects of important dependent variables such as age, length of BE, smoking, presence of hiatal hernia, and use of aspirin, nonsteroidal anti-inflammatory drugs (NSAIDs), and PPIs in the rates of progression were assessed. This was performed by using bivariate analyses that included computing and comparing incidence rates per 1000 patient-years across the demographic characteristics using Poisson regression. Covariates that were significant at a P value of less than.10 in bivariate analyses were entered into a multivariable Poisson model to adjust the overall incidence rates. By using a combined end point of HGD/EAC, we attempted to dichotomize age and length of BE using the minimum significant P value approach to determine the optimal cut-off point. In the process, we excluded the outer 10% of the distributions from contention as cut-off points, and corrected P values using the Bonferroni correction to the significance level. Categoric covariates were summarized with counts and percentages whereas continuous covariates were summarized with means Incidence of Dysplasia and Esophageal Adenocarcinoma During the follow-up evaluation 18 patients developed EAC, which translated to a calculated incidence rate of 0.27% per year (95% CI, 0.17%/y 0.43%/y). The mean time to development of EAC was 5.29 years (SD, 3.83 y; range, y). Thirty-two patients developed HGD, resulting in an incidence rate of 0.48% per year (95% CI, 0.34%/y 0.68%/y) with a mean time to development of HGD of 5.6 years (SD, 4.17 y; range, y). By using a combined end point of HGD/EAC, there were 42 patients (8 patients progressed from HGD to EAC and hence counted once in the overall tally of patients) who met this end point with an incidence rate of 0.63% per year (95% CI, 0.47%/y 0.86%/y). There were 217 patients who developed LGD; the incidence rate was 3.6% per year (95% CI, 3.19%/y 4.16%/y) for this outcome. The mean time to development of LGD was 4.59 years (SD, 3.2 y; range, y) (Table 1). The incidence rate of EAC development in the participating centers varied from 0.2% to 0.4% per year, whereas that of HGD varied from 0.18% to 0.97% per year. The overall incidence rate of HGD/EAC ranged from 0.3% to 1.2% per year at the different study sites. There was no statistically significant difference between the incidence rates of dysplasia and EAC at the different participating centers. Kaplan Meier survival graphs showed that cancer-free survival was 98.6% at 5 years of follow-up evaluation (n 540), 97.5% at 8 years of follow-up evaluation (n 267), and 97.1% at 10 years of follow-up evaluation (n 155) (Figure 2). Similarly, using an end point of HGD/EAC, at 5 years of follow-up evaluation (n 534) 96.8% were free of HGD/EAC, at 8 years of follow-up evaluation (n 263) 94.7% were free of HGD/EAC, and at 10 years of follow-up evaluation (n 152) 92.7% were free of HGD/EAC. Table 1. Incidence of Dysplasia and EAC in Patients With NDBE Diagnosis No. of incident cases Incidence rate %/y Mean time to development, y (SD) [range] LGD ( ) 4.59 (3.2) [ ] HGD ( ) 5.6 (4.17) [ ] EAC ( ) 5.29 (3.83) [ ] HGD/EAC 42 a 0.63 ( ) 5.41 (4.01) [ ] a Eight patients progressed from HGD to EAC and hence were counted once in the overall tally of patients for the combined end point of HGD/EAC.

4 March 2011 LOW INCIDENCE OF CANCER IN NONDYSPLASTIC BE 223 Figure 2. Kaplan Meier curve showing the percentage of patients with NDBE free of EAC. Incidence of Dysplasia and Esophageal Adenocarcinoma Based on Sex There were a total of 143 females who met the inclusion criteria in this cohort of NDBE patients. The incidence rate of LGD was 0.29% per year (95% CI, 0.18%/y 0.46%/y) and that of HGD/EAC was 0.44% per year (95% CI, 0.14%/y 1.30%/y). No cases of EAC were diagnosed in this group of patients. On the other hand, there were 1045 males with NDBE included in this analysis. The incidence rate of LGD was 0.37% per year (95% CI, 0.32%/y 0.43%/y), HGD was 0.49% per year (95% CI, 0.34%/y 0.71%/y), EAC was 0.3% per year (95% CI, 0.19%/y 0.48%/y), and HGD/EAC was 0.66% per year (95% CI, 0.48%/y 0.90%/y) (Table 2). As shown in the Kaplan Meier survival curve, there was no difference in the rates of progression of NDBE to LGD or HGD/EAC (Supplementary Figure 1) between females and males. Predictors of Progression to Dysplasia and Esophageal Adenocarcinoma Bivariate analyses evaluating the risk factors for progression showed no effect of age, smoking, use of aspirin/ NSAIDs, and the presence of hiatal hernia on overall incidence rates of dysplasia and EAC in patients with NDBE. PPI use was almost universal with 97.8% reporting use of PPI therapy. Threshold analysis to find an optimal cut-off point for age was performed. No comparisons of incidence rates of HGD/EAC across the 30 cut-off points under consideration (age, y) were significant. Hazard ratios for HGD, EAC, and HGD/EAC using age as a continuous variable in 10-year increments were 0.79 (95% CI, ), 1.47 (95% CI, ), and 0.95 (95% CI, ), respectively. Threshold analysis for BE length showed that the optimal cut-off point was at less than 6 cm versus 6 cm or greater. Patients with BE length of 6 cm or greater had significantly higher incidence rates of dysplasia (LGD and HGD) and EAC compared with those with BE length less than 6 cm (LGD: 5.66% per year [95% CI, ] vs 3.2% per year [95% CI, ], P.001; HGD: 1.03% per year [95% CI, ] vs 0.34% per year [95% CI, ], P.003; EAC: 0.65% per year [95% CI, ] vs 0.09% per year [95% CI, ], P.001 and HGD/EAC: 1.47% per year [95% CI, ] vs 0.34% per year [95% CI, ], P.001]; Table 3). Although the optimal cut-off point was at less than 6 cm versus 6 cm or more, no cancers were detected in patients with 2 cm or less. Hazard ratios for progression using length ( 3, 3 6, and 6 cm) have been highlighted in Supplementary Table 1. Discussion There has been significant variation in the reported incidence rates of progression to EAC in patients with NDBE. Results of this large multicenter cohort of NDBE patients that included 1204 patients with a mean follow-up period of 5.52 years ( patient-years) show that the incidence of EAC was 0.27% per year (95% CI, ) with a mean time of progression of 5.29 years. In addition, the vast majority of patients with NDBE were free of cancer at 5 years (n 540; 98.6%) and 10 years (n 155; 97.1%), with the 5- and 10-year cancer rates being low at 1.4% and 2.9%, respectively. Prior reports have reported an incidence rate of EAC that varied from 0.2% to 3.5% per year In a study by Hammeeteman et al 29 (n 50) in the 1980s, the incidence rate was 1.9% per year. Streitz et al 30 (n 149) in the 1990s reported an incidence rate of 1.3% per year. Bani-Hani et al 31 (n 307), in 2000 reported a cancer incidence of 0.9% per year. Reid et al 32 (n 129) during the same time frame reported an incidence rate of 3.5% per year. In a previous publication using this cohort of patients, an annual cancer incidence rate of 0.5% per year was reported, albeit with a smaller number of patients (n 618) and a shorter duration of follow-up evaluation (2546 patient-years). 11 In a recent study, de Jonge et al 33 reported the incidence of EAC in a nationwide cohort of BE patients in The Netherlands. Results of this study, which included 16,333 patients with NDBE and LGD with a total follow-up period of 78,480 patient-years, reported an annual incidence rate of 0.58% per year (95% CI, ) using an end point of HGD/EAC as Table 2. Sex-Based Analysis of Incidence Rates of Dysplasia and EAC in Patients With NDBE Females (n 143) Males (n 1045) Diagnosis Number Incidence rates Mean time, y (SD) Number Incidence rates Mean time, y (SD) LGD ( ) 4.55 (3.11) ( ) 4.59 (3.23) HGD ( ) 5.73 (3.11) ( ) 5.58 (4.31) EAC ( ) 5.29 (3.83) HGD/EAC ( ) 5.73 (3.11) 39 a 0.66 ( ) 5.38 (4.1) a Eight patients progressed from HGD to EAC and hence were counted once in the overall tally of patients for the combined end point of HGD/EAC.

5 224 WANI ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 9, No. 3 Table 3. Predictors of Progression to Dysplasia and EAC (Incidence Rates) in Patients With NDBE Variable LGD HGD EAC HGD/EAC Aspirin/NSAIDs (n 662) Yes, 512 (77.3%) 3.51 ( ) 0.41 ( ) 0.24 ( ) 0.52 ( ) Hazard ratio 0.78 ( ) 2.88 ( ) NA 3.65 ( ) No, 150 (22.6%) 4.25 ( ) 0.13 ( ) 0.13 ( ) Age (n 1190) 65 y, 685 (57.5%) 4.31 ( ) 0.59 ( ) 0.19 ( ) 0.67 ( ) 65 y, 505 (42.4%) 3.22 ( ) 0.42 ( ) 0.32 ( ) 0.62 ( ) Hazard ratio 0.72 ( ) 0.66 ( ) 1.52 ( ) 0.87 ( ) Smoking (n 714) Yes, 524 (73.4%) 3.79 ( ) 0.62 ( ) 0.31 ( ) 0.73 ( ) Hazard ratio 1.10 ( ) 6.42 ( ) 3.32 ( ) 3.85 ( ) No, 190 (26.6%) 3.39 ( ) 0.09 ( ) 0.09 ( ) 0.19 ( ) Hiatal hernia (n 1204) Yes, 684 (56.8%) 3.51 ( ) 0.43 ( ) 0.25 ( ) 0.57 ( ) Hazard ratio 0.96 ( ) 0.83 ( ) 0.95 ( ) 0.85 ( ) No, 520 (43.2%) 3.8 ( ) 0.55 ( ) 0.29 ( ) 0.71 ( ) Barrett s length (n 1182) 6 cm, 1000 (73%) 3.20 ( ) 0.34 ( ) 0.09 ( ) 0.34 ( ) 6 cm, 362 (27%) 5.66 ( ) 1.03 ( ) 0.65 ( ) 1.47 ( ) Hazard ratio 1.78 ( ) 3.02 ( ) 7.14 ( ) 4.29 ( ) malignant progression and an annual EAC risk of 0.43% (95% CI, ). For patients with NDBE, the annual incidence rate of EAC alone was 0.39% per year (95% CI, ), which is very similar to the current study. Given the absence of large cohort studies in BE patients in terms of progression rates to dysplasia and EAC, at least 6 systematic reviews and meta-analyses have been published to provide an estimate on risk of EAC in BE. The initial analysis by Shaheen et al 13 estimated the risk of progression to EAC to be 0.5% per year and cited publication bias as the potential cause for significant variation in the progression rates of BE to EAC, with smaller studies reporting a higher incidence of cancer and larger studies reporting a lower incidence. The most recent study by Sikkema et al 12 that included 50 studies with 14,109 patients followed up for 61,804 patient-years also reported a similar incidence rate of 0.63% per year (95% CI, ). Publication bias again was noted among studies reporting the incidence of EAC. In a recent systematic review and metaanalysis that assessed time trends in the incidence of EAC in BE patients, similar to previous reports, the pooled estimate of cancer incidence was 0.59% per year (95% CI, ). Timetrend analysis showed a significant decrease in the reported incidence of EAC in BE patients and a bubble plot based on study size clearly showed a lower incidence of cancer in larger studies. 18 Thus, these data and results from this study and other contemporary studies with larger numbers of patients and longer follow-up periods suggest that the incidence rates of EAC in NDBE patients may be lower than those previously reported (Figure 3). This study provides the estimate of the rate of progression to the combined end point of HGD and EAC (0.63% per year; 95% CI, ). Because esophagectomy is no longer the only treatment response to HGD, the estimate of this combined end point is of greater value from a health-services perspective. 7 Given the increasing availability of relatively well-tolerated endoscopic ablative therapy and the presence of 2 randomized controlled studies in the literature showing a decreased rate of progression to EAC in subjects with HGD treated with ablative therapy, progression to HGD or superficial cancer often results in the same intervention (endoscopic eradication therapies) ,34 Results from previous meta-analyses using this end point have reported a rate of 1% per year. 12,14,17 Results from this study were similar to those reported by de Jonge et al 33 (0.52% per year; 95% CI, ) and thus appear to be lower than those previously reported. Results from this study showed no difference in the rates of progression to dysplasia and EAC between females and males. The role of sex on the rate of progression of NDBE to HGD/EAC is unclear, with inconsistent results. 14,25,35 The lack of gender differences in cancer progression in this study may be related to a type II error owing to small sample sizes Figure 3. Bar diagram highlighting a lower incidence of EAC in larger contemporary studies and a higher incidence in smaller and older cohort studies.

6 March 2011 LOW INCIDENCE OF CANCER IN NONDYSPLASTIC BE 225 in the female category and needs to be explored further. In this study, BE length was the only predictor associated with progression to dysplasia and EAC in NDBE patients. No effect of age, smoking, use of aspirin/nsaids, or presence of hiatal hernia on overall progression rates was noted. These results are consistent with previous retrospective and prospective studies that have shown BE length to be a risk factor for progression to dysplasia or EAC. 22,36,37 Threshold analysis suggests that a cut-off length of 6 cm significantly increases risk of progression to EAC. In addition, no cancers in patients with BE length of 2 cm or less were identified. Consistent with a recent report and based on these results, there are insufficient data to support age or smoking as a risk factor in the progression of NDBE. 22 Previous cohort studies have indicated an association of PPI use and reduced risk of dysplasia and EAC as well as showed a reduction in EAC incidence in aspirin/nsaid users. 42,43 However, we were unable to show an impact on PPI and aspirin/nsaid use and dysplasia and EAC development. The ongoing Aspirin Esomeprazole Chemoprevention Trial should provide definitive conclusions regarding the use of these medications in the chemoprevention of dysplasia and EAC in NDBE patients. 44 The results from this study have several important implications. The true prevalence of BE in the general US population remains unknown. However, a Swedish population-based study revealed a prevalence of 1.6%. 45 These numbers when applied to the US population suggest that about 2 to 3 million individuals may have this precancerous condition. The cost effectiveness of surveillance endoscopy in patients with BE is very sensitive to the incidence of cancer and determining the accurate incidence of dysplasia and EAC is of critical importance to the cost effectiveness of any surveillance program in patients with BE. A computer model, using an annual incidence rate of 0.4%, suggested that surveillance every 5 years was the only effective strategy and had an incremental cost-utility ratio of $98,000/ quality-adjusted life years. 46 The incidence of EAC in recent studies (including this study) is nearly half the incidence rate used in published cost-effectiveness analyses. 33 Appraisal of current cost-effectiveness models using these incidence rates should be performed in future studies. In addition, because the vast majority of patients are cancer-free at 5- and 8-year intervals and the mean time to progression to EAC was 5.29 years, lengthening of surveillance intervals to possibly every 5 years for NDBE patients should be considered. Accurate estimates are also important for informing patients and conducting chemoprevention trials or ablation trials. Given the low rates of progression to cancer, routine endoscopic ablation of NDBE would be difficult to justify. Assuming that endoscopic ablation decreases the risk of cancer by 50%, the number needed to treat to prevent one case of EAC would be 769. Most importantly, this study provides important estimates in making decisions in the management of this premalignant condition for patients, physicians, and policymakers. Describing the true risk of cancer to patients combined with potential benefits of any intervention (eg, surveillance, chemoprevention, and ablation) would allow for a patient-centered decision regarding management strategies. There were several limitations of this database study that merit consideration. Surveillance intervals and biopsy protocols (including use of jumbo biopsy forceps) were not standardized between the centers. Central pathology readings (especially for all cases of dysplasia) were not performed routinely. It would have been ideal to have all cases reviewed by an expert central gastrointestinal pathologist. However, significant interobserver variability has been reported even among expert pathologists in the interpretation of dysplasia 10,28 and highlight the limitations of dysplasia as a marker for cancer development in BE. Another important limitation of this study was the lack of information on patients who were lost to follow-up evaluation in the surveillance program. The impact of this on overall incidence rates cannot be excluded. Endoscopic eradication therapies for HGD may have impacted the overall incidence of EAC. Hence, results using the end point of HGD/EAC have been presented in this analysis. The cause of mortality could not be assessed in this study because of the lack of a link to a cancer registry. The potential for type II error on the lack of effect of age, smoking, use of aspirin/nsaids, and PPI use on the progression rates to dysplasia and cancer cannot be excluded. Information on body mass index and waist circumference and its role on progression in BE could not be explored in this study. All participating centers were tertiary care referral centers with a special interest in BE and a sizeable population of this cohort included veterans, which may limit the generalizability of these results. The Barrett s esophagus study is an effort in building a clinical consortium for endoscopic research in the field of BE to define the natural history of this disease. The large number of patients with BE in the United States along with relatively long patient-years of follow-up evaluation were the major strengths of this endeavor. Future goals of this consortium include comprehensive collection of risk factors for progression to improve risk-stratification and risk-reduction strategies and define health care use. In conclusion, results from this large multicenter cohort study highlight the low annual incidence rates of dysplasia and EAC in patients with BE (EAC, 0.27%; HGD, 0.48%; and HGD/EAC, 0.63%). Because the vast majority of patients are free of cancer at long-term follow-up evaluation, lengthening of surveillance intervals should be considered. Development of a risk-stratification system in a large multicenter clinical consortium that incorporates clinical and biomarker variables to focus surveillance on patients at highest risk for progression should be a priority. Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at and at doi: / j.cgh References 1. Sharma P, McQuaid K, Dent J, et al. A critical review of the diagnosis and management of Barrett s esophagus: the AGA Chicago Workshop. Gastroenterology 2004;127: Sharma P. Clinical practice. Barrett s esophagus. N Engl J Med 2009;361: Rex DK, Cummings OW, Shaw M, et al. Screening for Barrett s esophagus in colonoscopy patients with and without heartburn. Gastroenterology 2003;125: Pohl H, Welch HG. The role of overdiagnosis and reclassification in the marked increase of esophageal adenocarcinoma incidence. J Natl Cancer Inst 2005;97:

7 226 WANI ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 9, No Jemal A, Siegel R, Ward E, et al. Cancer statistics, CA Cancer J Clin 2009;59: Eloubeidi MA, Mason AC, Desmond RA, et al. Temporal trends ( ) in survival of patients with esophageal adenocarcinoma in the United States: a glimmer of hope? Am J Gastroenterol 2003;98: Wang KK, Sampliner RE. Updated guidelines 2008 for the diagnosis, surveillance and therapy of Barrett s esophagus. Am J Gastroenterol 2008;103: Cooper GS, Kou TD, Chak A. Receipt of previous diagnoses and endoscopy and outcome from esophageal adenocarcinoma: a population-based study with temporal trends. Am J Gastroenterol 2009;104: Corley DA, Levin TR, Habel LA, et al. Surveillance and survival in Barrett s adenocarcinomas: a population-based study. Gastroenterology 2002;122: Wani S, Mathur SC, Curvers WL, et al. Greater inter-observer agreement by endoscopic mucosal resection than biopsy samples in Barrett s dysplasia. Clin Gastroenterol Hepatol 2010;8: Sharma P, Falk GW, Weston AP, et al. Dysplasia and cancer in a large multicenter cohort of patients with Barrett s esophagus. Clin Gastroenterol Hepatol 2006;4: Sikkema M, de Jonge PJ, Steyerberg EW, et al. Risk of esophageal adenocarcinoma and mortality in patients with Barrett s esophagus: a systematic review and meta-analysis. Clin Gastroenterol Hepatol 2010;8: 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: Yousef F, Cardwell C, Cantwell MM, et al. The incidence of esophageal cancer and high-grade dysplasia in Barrett s esophagus: a systematic review and meta-analysis. Am J Epidemiol 2008;168: Thomas T, Abrams KR, De Caestecker JS, et al. Meta analysis: cancer risk in Barrett s oesophagus. Aliment Pharmacol Ther 2007;26: Wani S, Puli SR, Shaheen NJ, et al. Esophageal adenocarcinoma in Barrett s esophagus after endoscopic ablative therapy: a metaanalysis and systematic review. Am J Gastroenterol 2009;104: Chang EY, Morris CD, Seltman AK, et al. The effect of antireflux surgery on esophageal carcinogenesis in patients with Barrett esophagus: a systematic review. Ann Surg 2007;246: Wani S, Cook MB, Hall M, et al. Has the incidence of esophageal adenocarcinoma (EAC) in Barrett s esophagus (BE) increased? A time-trend analysis. Gastroenterology 2009; 136(Suppl 1):A Wani S, Sayana H, Sharma P. Endoscopic eradication of Barrett s esophagus. Gastrointest Endosc 2010;71: Shaheen NJ, Sharma P, Overholt BF, et al. Radiofrequency ablation in Barrett s esophagus with dysplasia. N Engl J Med 2009; 360: Overholt BF, Lightdale CJ, Wang KK, et al. Photodynamic therapy with porfimer sodium for ablation of high-grade dysplasia in Barrett s esophagus: international, partially blinded, randomized phase III trial. Gastrointest Endosc 2005;62: Prasad GA, Bansal A, Sharma P, et al. Predictors of progression in Barrett s esophagus: current knowledge and future directions. Am J Gastroenterol 2010;105: Bani-Hani KE, Bani-Hani BK, Martin IG. Characteristics of patients with columnar-lined Barrett s esophagus and risk factors for progression to esophageal adenocarcinoma. World J Gastroenterol 2005;11: de Jonge PJ, Steyerberg EW, Kuipers EJ, et al. Risk factors for the development of esophageal adenocarcinoma in Barrett s esophagus. Am J Gastroenterol 2006;101: Oberg S, Wenner J, Johansson J, et al. Barrett esophagus: risk factors for progression to dysplasia and adenocarcinoma. Ann Surg 2005;242: Sampliner RE. Updated guidelines for the diagnosis, surveillance, and therapy of Barrett s esophagus. Am J Gastroenterol 2002;97: Riddell RH, Goldman H, Ransohoff DF, et al. Dysplasia in inflammatory bowel disease: standardized classification with provisional clinical applications. Hum Pathol 1983;14: Montgomery E, Bronner MP, Goldblum JR, et al. Reproducibility of the diagnosis of dysplasia in Barrett esophagus: a reaffirmation. Hum Pathol 2001;32: Hammeeteman W, Tytgat NJ, Houthoff HJ, et al. Barrett s esophagus: development of dysplasia and adenocarcinoma. Gastroenterology 1989;96: Streitz JM Jr, Ellis FH Jr, Tilden RL, et al. Endoscopic surveillance of Barrett s esophagus: a cost-effectiveness comparison with mammographic surveillance for breast cancer. Am J Gastroenterol 1998;93: Bani-Hani BK, Sue-Ling H, Johnston D, et al. Barrett s oesophagus: results from a 13-year surveillance programme. Eur J Gastroenterol Hepatol 2000;12: Reid B, Levine DS, Longton G, et al. Predictors of progression to cancer in Barrett s esophagus: baseline histology and flow cytometry identify low- and high-risk patient subsets. Am J Gastroenterol 2000;95: de Jonge PJ, van Blankenstein M, Looman CW, et al. Risk of malignant progression in patients with Barrett s oesophagus: a Dutch nationwide cohort study. Gut 2010;59: Pech O, Behrens A, May A, et al. Long-term results and risk factor analysis for recurrence after curative endoscopic therapy in 349 patients with high-grade intraepithelial neoplasia and mucosal adenocarcinoma in Barrett s oesophagus. Gut 2008;57: Falk GW, Thota PN, Richter JE, et al. Barrett s esophagus in women: demographic features and progression to high-grade dysplasia and cancer. Clin Gastroenterol Hepatol 2005;3: Weston AP, Sharma P, Mathur S, et al. Risk stratification of Barrett s esophagus: updated prospective multivariate analysis. Am J Gastroenterol 2004;99: O Connor JB, Falk GW, Richter JE. The incidence of adenocarcinoma and dysplasia in Barrett s esophagus: report on the Cleveland Clinic, Barrett s Esophagus Registry. Am J Gastroenterol 1999;94: El-Serag HB, Aguirre TV, Davis S, et al. Proton pump inhibitors are associated with reduced incidence of dysplasia in Barrett s esophagus. Am J Gastroenterol 2004;99: Hillman LC, Chiragakis L, Shadbolt B, et al. Effect of proton pump inhibitors on markers of risk for high-grade dysplasia and oesophageal cancer in Barrett s oesophagus. Aliment Pharmacol Ther 2008;27: Nguyen DM, El-Serag HB, Henderson L, et al. Medication usage and the risk of neoplasia in patients with Barrett s esophagus. Clin Gastroenterol Hepatol 2009;7: Vaughan TL, Dong LM, Blount PL, et al. Non-steroidal anti-inflammatory drugs and risk of neoplastic progression in Barrett s oesophagus: a prospective study. Lancet Oncol 2005;6: Corley DA, Kerlikowske K, Verma R, et al. Protective association of aspirin/nsaids and esophageal cancer: a systematic review and meta-analysis. Gastroenterology 2003;124: Nguyen DM, Richardson P, El-Serag HB. Medications (NSAIDs, statins, proton pump inhibitors) and the risk of esophageal adenocarcinoma in patients with Barrett s esophagus. Gastroenterology 2010;138: Jankowski J, Barr H. Improving surveillance for Barrett s oesoph-

8 March 2011 LOW INCIDENCE OF CANCER IN NONDYSPLASTIC BE 227 agus: AspECT and BOSS trials provide an evidence base. BMJ 2006;332: Ronkainen J, Aro P, Storskrubb T, et al. Prevalence of Barrett s esophagus in the general population: an endoscopic study. Gastroenterology 2005;129: 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: Inadomi JM, Sampliner R, Lagergren J, et al. Screening and surveillance for Barrett esophagus in high-risk groups: a costutility analysis. Ann Intern Med 2003;138: Gerson L, Groeneveld PW, Triadafilopoulos G. Cost-effectiveness model of endoscopic screening and surveillance in patients with gastroesophageal reflux disease. Clin Gastroenterol Hepatol 2004;2: Reprint requests Address requests for reprints to: Prateek Sharma, MD, Gastroenterology (111), Department of Veterans Affairs Medical Center, 4801 E. Linwood Boulevard, Kansas City, Missouri psharma@kumc.edu; fax: (816) Acknowledgments Results of this study were presented in part at the AGA Esophagogastroduodenoscopy Plenary Session at Digestive Disease Week 2010, New Orleans, LA. Support provided by TAP Pharmaceutical Product Inc, now part of Takeda Pharmaceutical North America, Inc. Conflicts of interest The authors disclose no conflicts.

9 March 2011 LOW INCIDENCE OF CANCER IN NONDYSPLASTIC BE 227.e1 Supplementary Figure 1. Kaplan Meier curve showing the percentage of patients with NDBE free of high-grade dysplasia or EAC during follow-up evaluation between females and males. Supplementary Table 1. Incidence Rates and Hazard Ratios for Progression of BE to Dysplasia and EAC Based on Length Barrett s length LGD HGD EAC HGD/EAC 3 cm, 596 (55.6%) 2.38 ( ) 0.25 ( ) NA 0.25 ( ) Hazard ratio 3 6 cm, 292 (27.2%) 5.14 ( ) 0.81 ( ) 0.37 ( ) 0.93 ( ) Hazard ratio 2.15 ( ) 3.18 ( ) NA 3.67 ( ) 6 cm, 184 (17.2%) 6.01 ( ) 0.80 ( ) 0.69 ( ) 1.12 ( ) Hazard ratio 2.59 ( ) 3.30 ( ) NA 4.91 ( ) Test for departure from linearity P.057 P.134 NA P.143