Negative surveillance endoscopy occurs frequently in patients with short-segment non-dysplastic Barrett's esophagus.

Surveillance endoscopy of non-dysplastic Barrett's esophagus (NDBE) that fails to detect intestinal metaplasia (IM), or negative surveillance, is known to occur in clinical practice, although the frequency and possible outcomes in a large cohort in clinical practice is not well described. The goals of this study were to define frequency in which negative surveillance occurs and endoscopic outcomes in a screening cohort of short segment NDBE. A retrospective cohort (n = 184) of patients newly diagnosed with short segment NDBE at an outpatient academic tertiary care center between 2003 and 2011 were reviewed. Only those with one or more surveillance endoscopies were included to define a frequency of negative surveillance. Included patients were further assessed if they had two or more surveillance endoscopies and were classified into groups as sampling error or negative IM on consecutive surveillances based on the results of their surveillance endoscopies. The frequency of a negative surveillance endoscopy in all short-segment NDBE patients was 19.66% (92 endoscopic exams were negative for IM of 468 total surveillance exams). A negative surveillance endoscopy occurred in 40.76% (n = 75) patients. Sampling error occurred in 44.12% and negative IM on consecutive surveillance endoscopies in 55.88% of those with ≥ 2 surveillance endoscopies and an initially negative surveillance exam. The frequency of negative IM on consecutive surveillances was 19.00% of all patients who had two surveillance endoscopies. When the index diagnostic Barrett's esophagus segment length was < 1 cm, 32.14% (18/56) of all patients (with ≥ 2 surveillance endoscopies) had negative IM on consecutive surveillance endoscopies. Negative surveillance occurs frequently in short-segment NDBE. When an initial negative surveillance endoscopy occurs, it may be due to either a sampling error or lack of detectable IM on surveillance exam. When a <1 cm segment of NDBE is diagnosed, a significant proportion of patients may go on to have continuously undetected IM on consecutive surveillance endoscopic exams without intervention.