Background: Unsedated transnasal upper endoscopy has a diagnostic yield comparable with that of sedated conventional upper endoscopy. The ability of transnasal upper endoscopy to detect Barrett's metaplasia or dysplastic change has not been systematically evaluated. The aim of this study was to assess the feasibility of transnasal upper endoscopy for surveillance of patients with Barrett's esophagus and to evaluate its histopathologic yield for Barrett's metaplasia and dysplasia. Methods: Thirty-two patients with Barrett's esophagus were evaluated with conventional upper endoscopy and transnasal upper endoscopy. An effort was made to recruit patients known to have dysplasia. Quadrantic biopsy specimens were taken with standard (conventional upper endoscopy) and pediatric (transnasal upper endoscopy) biopsy forceps at procedures performed at least 1 week apart. Two blinded pathologists evaluated the specimens. Results: Transnasal upper endoscopy detected Barrett's metaplasia histopathologically in 31 of 32 patients. Level of agreement for presence of dysplasia in biopsy specimens obtained between conventional upper endoscopy and transnasal upper endoscopy was excellent (k = 0.79). Interobserver agreement for dysplasia in specimens obtained by conventional upper endoscopy (k = 0.61) and by transnasal upper endoscopy (k = 0.61) were similar. Intraobserver agreement between conventional upper endoscopy and transnasal upper endoscopy biopsy specimens for dysplasia by pathologist 1 (k = 0.73) and pathologist 2 (k = 0.75) were also similar. No significant adverse effects were noted. Conclusions: Transnasal upper endoscopy is feasible and accurate for identification and histopathologic confirmation of Barrett's metaplasia with a histopathologic yield for dysplasia comparable with conventional upper endoscopy. (Gastrointest Endosc 2002;56:472-8.)