Dysplasia detection rate of confirmatory EGD in nondysplastic Barrett's esophagus

Current guidelines for endoscopic surveillance of Barrett's esophagus (BE) recommend that patients with newly diagnosed BE undergo confirmatory esophagogastroduodenoscopy (EGD) to exclude the presence of dysplasia. The extent to which confirmatory endoscopy alters management and detects missed dysplasia in newly diagnosed BE has not been reported. The frequency with which confirmatory endoscopy changed surveillance management in patients with newly diagnosed BE was assessed. A two center cohort analysis was conducted on patients newly diagnosed with BE. The rate of dysplasia on confirmatory endoscopy for patients who had nondysplastic BE was obtained. Demographic and endoscopic variables were assessed for association with dysplasia detection using Firth logistic regression model. Out of the 146 patients newly diagnosed with BE and initially determined to be without dysplasia, 12 had dysplasia on the confirmatory second EGD (8.2%). Eleven of 12 cases with dysplasia on confirmatory endoscopy had long-segment BE (LSBE). Among all the LSBE cases in our cohort, 11 had newly diagnosed dysplasia on confirmatory EGD, 29.7% (11/37). The average number of biopsies obtained from the 11 LSBE cases with dysplasia was comparable with the rest of the LSBE cases without dysplasia (6.73 and 5.42, respectively, P-value 0.205). The rate of dysplasia detection in short-segment BE (SSBE) was much lower, 0.95% (1 out of 105). There were no cases of high-grade dysplasia (HGD) or cancer detected in any SSBE case. HGD was detected on confirmatory EGD in two cases, both were LSBE. Segment length was the only statistically significant factor to predict the presence of dysplasia on confirmatory endoscopy (odds ratio 9.158, P. 0.008). Confirmatory EGD in newly diagnosed LSBE had significant rate of dysplasia detection (29.7%) in this cohort. Among patients with SSBE, there was a low rate of dysplasia detection with confirmatory EGD, less than 1% of cases. No additional cases of HGD or esophageal carcinoma in SSBE cases were detected. This suggests that the yield of confirmatory EGD is greater in patients with LSBE.