The prevalence and characteristics of Barrett esophagus in patients with adenocarcinoma of the esophagus or esophagogastric junction are uncertain. We studied 61 consecutive esophagogastrectomy specimens with adenocarcinoma, which were subjected to extensive histopathologic examination. Barrett esophagus was found in 64% of the cases (39 of 61), but had been recognized in only 38% of the patients with Barrett-associated carcinoma who had undergone preoperative endoscopy with biopsy (13 of 34). The median extent of Barrett esophagus with adenocarcinoma was 5 cm (range, 1 cm to 12 cm), and distinctive-type ("specialized") Barrett mucosa predominated (35 of 39; 90%). The Barrett adenocarcinomas were centered in the distal esophagus 2 cm +/- 0.3 cm above the esophagogastric junction. The patients with Barrett adenocarcinoma showed a striking predominance of white men (34 of 39; 87%) in contrast to gastric adenocarcinoma cases (21 of 69; 30%) and to Barrett patients without carcinoma or dysplasia (75 of 149; 50%), but similar to patients having adenocarcinoma of the esophagus or esophagogastric junction without demonstrable Barrett esophagus (16 of 22; 73%). Our findings suggest that most adenocarcinomas of the esophagus or esophagogastric junction are Barrett carcinomas, rather than gastric cardiac cancers or other types of esophageal adenocarcinoma; most Barrett adenocarcinomas occur in short segments of Barrett esophagus, which may be difficult to detect at endoscopy; and white men with Barrett esophagus may constitute a clinically identifiable at-risk group suitable for surveillance.