BACKGROUND & AIMS: Guidelines recommend periodic endoscopic surveillance of Barrett’s esophagus (BE) patients to detect and treat early esophageal adenocarcinomas; however, no trials or population-based studies exist. We evaluated the association between endoscopic surveillance of BE and survival among esophageal/gastric cardia adenocarcinoma patients. METHODS: We studied a cohort of 23 BE patients, among 589 esophageal or gastric cardia adenocarcinoma patients diagnosed between 1990-1998 at Northern California Kaiser Permanente (a large health maintenance organization). We measured the presence of BE, detection of cancer by endoscopic surveillance, cancer stage, mortality, and potential confounders. RESULTS: BE was diagnosed in 135 of 589 adenocarcinoma patients, with 23 BE patients diagnosed greater than 6 months before cancer was diagnosed. Among these 23 patients, 73% of the surveillance-detected cancer patients (n = 15) were alive at the end of follow-up, compared with none of the patients without surveillance-detected cancers (n = 8; P = 0.001). All surveillance-detected cancer patients had low-stage disease and none died directly from cancer. The surveillance/survival association was not substantially altered by stratification for age at BE diagnosis or other potential confounders. CONCLUSIONS: Surveillance-detected BE-associated adenocarcinomas were associated with low-stage disease and improved survival. Additional studies are needed to evaluate potential biases and whether screening/surveillance programs decrease mortality among all patients in surveillance. Few patients (3.9%) had a BE diagnosed before their cancer. Thus, even if current surveillance techniques are effective, they are unlikely to substantially impact the population’s mortality from esophageal cancer; better methods are needed to identify at risk patients.