The role of flexible endoscopy in the diagnosis of esophageal trauma remains undefined. This study evaluates the use of immediate flexible fiberoptic esophagogastroduodenoscopy (EGD) as the primary diagnostic tool for detection of esophageal injury in trauma patients. Flexible EGD was performed on 31 patients for this purpose from August 1991 through January 1994. There were 28 males and 3 females with a mean age of 24.3 years (range, 16-54 years). Twenty-four of 31 patients (77%) were intubated at the time of the examination. Mechanism of injury was penetrating in 24 patients (20 gunshot wounds, four stab wounds) and blunt (motor vehicle crash) in seven patients. Penetrating injuries were located in the neck in 5 of 24 patients, in the chest in 15 of 24 patients, and in both the neck and chest in 4 of 24 patients. Upper gastrointestinal contrast studies were performed for 3 of 31 patients (10%), computed tomography was performed for eight patients (26%), bronchoscopy was performed for 13 patients (42%), angiography was performed for 17 patients (55%), and rigid esophagoscopy and laryngoscopy were each performed for one patient (3%). Evidence of esophageal trauma during EGD was seen in 5 of 31 patients. True-positive studies occurred for four patients, false-positive results occurred for one patient, true-negative results occurred for 26 patients (as demonstrated by exploration in five and clinical follow-up in 21), and no false-negative examinations occurred. Sensitivity of flexible EGD was 100%, specificity was 96%, and accuracy was 97%. No complications occurred related to the performance of EGD. Flexible fiberoptic endoscopy seems to be a safe and effective method for both detection and exclusion of esophageal trauma.