Chromo-zoom endoscopy has been demonstrated to be valuable in assessing the degree of intestinal villous atrophy in patients with suspected celiac disease. To evaluate the diagnostic accuracy of chromo-zoom endoscopy in patients with difficult diagnosis because of nonconcordant test results and/or the confounding of a gluten-free diet initiated before an appropriate diagnosis of celiac disease and to compare the findings to a recent reference standard, the in vitro gliadin challenge test. Prospective, case-control study. Tertiary-care referral hospital. Patients without celiac disease (negative control group, n = 9), patients with celiac disease (positive control group, n = 41), and patients with difficult diagnosis (n = 27). Chromo-endoscopy with indigo carmine and endoscopic zoom-magnification were performed. Duodenal fragments were collected for the in vitro gliadin challenge test. The area under the receiver operating characteristic curve (ROC) was used for statistical analyses on accuracy. Diagnostic accuracy of chromo-zoom endoscopy for detection of mucosal abnormalities in patients with difficult diagnosis. Chromo-zoom endoscopy had a high accuracy for celiac disease diagnosis in analyses on negative controls and positive controls (area under roc = 0.99). In the difficult diagnosis group, the accuracy of chromo-zoom endoscopy was lower (area under roc = 0.83), but it increased after exclusion of patients with celiac disease on gluten-free diet (area under roc = 0.88). There was a 4% failure rate in the ability to cultivate biopsies. Also, the study was done at an academic medical center. Chromo-zoom endoscopy has high accuracy for cases of difficult diagnosis of celiac disease but only in untreated patients with celiac disease.