CHICAGO – Early endoscopic follow-up within 24 months detected dysplasia in nearly one in 10 patients with nondysplastic or low-grade Barrett’s esophagus in a retrospective study at the Mayo Clinic.
Initial endoscopy missed four cases of high-grade dysplasia or esophageal adenocarcinoma (1.9%) and 16 cases of low-grade dysplasia (7.6%) for an overall miss-rate of 9.5%.
Patients on proton pump inhibitors were less likely to have dysplasia missed than were those off PPIs (20% vs. 52.6%, P = .008).
Those with long- versus short-segment Barrett’s esophagus were more likely to have dysplasia overlooked (85% vs. 53.6%; P = .008; mean 6 mm vs. 4 mm; P = .006), Dr. Kavel Visrodia said at the annual Digestive Disease Week.
Current American College of Gastroenterology (ACG) guidelines recommend early repeat esophagogastroduodenoscopy (EGD) to exclude the presence of missed dysplasia in newly diagnosed nondysplastic Barrett’s esophagus (BE), while the ACG and American Society for Gastrointestinal Endoscopy call for repeat EGD within 6 months for those with low-grade dysplasia.
The yield for repeat EGD has not been established, and only one study exists in the literature, said Dr. Visrodia of the department of medicine, Mayo Clinic, Rochester, Minn.
That study (Dis. Esophagus 2012 Sept. 28. [doi:10.1111/j.1442-2050.2012.01431.x]) showed a miss-rate of 8.2% among 146 patients with newly diagnosed nondysplastic BE. Long-segment BE was the only significant predictor of dysplasia on follow-up (odds ratio, 9.18; P = .008).
The cohort was relatively small and had no long-term follow-up, and with an interval to follow-up of 36 months, "it’s possible that some of these were actually incident cases of dysplasia and not prevalent cases," he said.
To address these gaps, Dr. Visrodia and his colleagues identified 488 BE cases from 1977 to 2011 in the Rochester Epidemiology Project in Olmsted County, Minn. A total of 278 patients were excluded because of high-grade dysplasia (HGD) or esophageal cancer on index endoscopy or repeat endoscopy after 24 months, leaving 181 patients with nondysplastic BE and 29 with low-grade dysplasia (LGD).
Repeat endoscopy within 24 months revealed 2 cases of HGD or cancer and 16 cases of LGD in the nondysplastic BE group, and 2 cases of HGD or cancer in the LGD group, Dr. Visrodia said.
Three of the four HGD/cancer cases were in patients with long-segment BE, defined as at least 3 cm of columnar mucosa.
Biopsies were insufficient in 63% of patients with missed dysplasia, compared with 55% in the group without missed dysplasia. Biopsies were considered adequate if the number of biopsies divided by the BE length was at least 2, indicating that samples were taken every 2 cm in accordance with guidelines. This risk factor is noteworthy, although the difference between groups was not statistically significant, possibly because of the small sample size, he said.
Finally, after a median of 6.8 years of follow-up, 30 asymptomatic, prevalent HGDs or cancers were detected within 24 months, compared with 22 incident cases detected after 24 months. This suggests that "a greater number of high-grade dysplasias and cancers were detected up front rather than during long-term careful surveillance," Dr. Visrodia said.
During a discussion of the study, one attendee asked whether the results make a better case for aggressive ablation up front rather than for surveillance, while others expressed surprise at the high miss rate at an institution such as the Mayo Clinic.
Dr. Visrodia replied that the results do give them pause, and suggested that tighter early endoscopic surveillance may be warranted, particularly in those with long-segment BE.
Dr. Visrodia and his coauthors reported no financial disclosures.