Reported incidences of bacteremia after endoscopy with esophageal variceal sclerotherapy are conflicting. A prospective controlled study was conducted to determine the frequency of bacteremia after emergency endoscopy with esophageal variceal sclerotherapy compared with frequency after elective esophageal variceal sclerotherapy and after emergency endoscopy in patients with upper gastrointestinal bleeding from nonvariceal sources. A total of 126 endoscopies were studied in 72 patients. Groups consisted of (a) emergency endoscopy without esophageal variceal sclerotherapy, 37 sessions with 36 patients; (b) elective esophageal variceal sclerotherapy, 33 sessions with 14 patients; and (c) emergency esophageal variceal sclerotherapy, 56 sessions with 36 patients. Blood cultures were obtained before and 5 and 30 minutes after endoscopy. There was a higher frequency of preendoscopic bacteremia in emergency esophageal variceal sclerotherapy (13%) than in emergency endoscopy alone (0%) (P = 0.02). Clinically significant bacteremia in emergency esophageal variceal sclerotherapy was observed in 7 of 56 (13%) sessions, compared with 0 of 33 in elective esophageal variceal sclerotherapy (P = 0.03) and 1 of 36 (3%) in emergency endoscopy alone (P = 0.45). Of these cases, 3 (5.4%) were potentially caused by emergency esophageal variceal sclerotherapy, but not clinically significant postendoscopic bacteremia was attributable to the procedure in the other groups.