Several endoscopic signs have been reported in eosinophilic esophagitis (EoE). However, associations between these features and EoE have not been critically appraised, nor has their diagnostic utility. Aims: Investigate the association between endoscopic (EGD) features and EoE, & estimate the Dx. sensitivity (SEN), specificity (SPEC), positive & negative predictive value (PPV & NPV) of EGD features suggestive of EoE. Methods: Consecutive patients undergoing EGD with findings suggestive of EoE were entered into the study. Required features included: mucosal rings, white lesions (nodules, specs, and/or exudates), furrows, paper-thin like crêpes & loss of vascular pattern. Histological findings were grouped as: NL: 0-5 eos/HPF, indeterminate: ≥ 6-20 eos/HPF, EoE: ≥ 21 eos/HPF. Biopsies were obtained from the distal, middle and proximal esophagus by a single endoscopist. Statistical analysis: Associations between presence of EGD features and EoE were investigated using Fisher's exact test. SEN, SPEC, PPV, and NPV were estimated for each EGD feature. Results: During the study period, 144 consecutive patients (60 males [42%], median age: 60 [range 20 - 91 years]) with suspected EoE were evaluated. An abnormal esophageal eosinophilic infiltration was noted in 25 (17%) patients: (12 [8%] indeterminate, 13 [9%] EoE). EGD features observed were: rings (84%), white specs (33%), exudates (15%), furrows (8%), loss of vascular pattern (7%) and erosions (6%). There was evidence of an association between EoE and furrows (Normal: 3%, Indeterminate: 33%, EoE: 38%, P<0.001), but there were no statistically significant associations between EoE and any other EGD features (Table). None of the EGD features appears to be useful in diagnosing EoE. Combining one or several EGD signs did not result in noticeable differences when considering pair-wise combinations of features. Conclusions: In the evaluation of patients with endosopic signs suggestive of eosinophilic esophagitis, 17% had an abnormal esophageal eosinophilc infiltrate (9% fulfilling EoE criteria). Esophageal furrows, when found, represent the single best endoscopic feature associated with abnormal esophageal eosinophilic infiltration. Our study findings suggest that the endoscopic signs previously considered diagnostic for EoE are not reliable diagnostic markers. Tabled 1 Endoscopic feature Sensitivity Specificity PPV NPV Rings 12/13 (92%) 22/131 (17%) 12/121 (10%) 22/23 (96%) Furrows 5/13 (38%) 124/131 (95%) 5/12 (42%) 124/132 (94%) Exudates 3/13 (23%) 113/131 (92%) 3/21 (14%) 113/123 (92%) White specs 5/13 (38%) 88/131 (67%) 5/48 (10%) 88/96 (92%) Erosions 0/13 (0%) 123/131 (97%) 0/8 (0%) 123/136 (90%) Loss of vascular pattern 2/13 (15%) 123/131 (97%) 2/10 (20%) 123/134 (92%) Open table in a new tab