Article Text
Abstract
Introduction Two-thirds of patients with reflux symptoms will have a normal white light endoscopy (WLE). Using high definition narrow-band imaging (NBI), however, we found abnormalities in the distal oesophagus correlating with histologic findings of reflux. A common appearance was noted in patients referred for investigation of reflux (where endoscopy was otherwise normal). This was termed the ‘linear furrow’ sign (LFS) as it comprised of linear, alternating lighter and darker ‘ridges’ of mucosa resembling furrows in a ploughed field. This observational cohort study was conducted prospectively to establish the relevance of the LFS.
Method Consecutive patients referred for investigation of reflux underwent endoscopy (BH or AH) to the institution protocol: a pair of routine, non-targetted distal and mid-esophageal biopsies were taken to examine for microscopic changes of reflux and to exclude eosinophilic esophagitis (EoE).
Prior to biopsies, the oesophagus was examined using NBI. The endoscopist’s opinion as to whether the LFS was present was recorded in the routine report and digital still images captured.
Histology was reported in the routine clinical service, then reviewed by a single experienced Histopathologist (HAA), blinded to endoscopic findings. Positive biopsies were those initially reported as “findings compatible with reflux”, but in each case the reviewing Histopathologist recorded the changes contributing to this overall impression. Sensitivity and specificity of the LFS was calculated using biopsies as the gold standard.
Results n = 31 with normal WLE were examined (13F; mean age 44.3 ± 15.0 yrs). The LFS was present (LFS+) in 25 patients, extending approximately 4–5 cm above the squamocolumnar junction. Histologic examination of biopsies from this region demonstrated changes consistent with reflux in 23/25. 6 patients were LFS-, with normal histology in 5. Mid-esophageal biopsies were positive in only three patients. In one, diagnostic criteria for EoE were met (responding completely to acid suppression).
Magnification endoscopy with NBI demonstrated that the changes seen arise from the impression of a ridged appearance to the mucosal surface – most likely reflecting elongation of papillae and basal cell hyperplasia. Spongiosis (dilated intercellular spaces; DIS) was demonstrable in 75% of LFS+. Although reflux symptoms were not systematically evaluated, a self-reported response to PPI was noted in 70% of such patients (LFS+, DIS+) and only 10% of those without (LFS-, DIS-; p < 0.01).
Conclusion The LFS was detected in 80% of patients, achieving diagnostic accuracy of 90% and appears to relate to self-reported symptomatic response to PPI. A further prospective study is under way to determine the clinical significance of this finding.
Disclosure of interest None Declared.