A review of the treatment of esophageal atresia (EA) was undertaken to examine current methods of management and to ascertain the influence of prematurity and associated medical defects on survival. One hundred patients with EA presented at The Children's Orthopedic Hospital and Medical Center, Seattle, between 1967 and 1979. Eighty-two percent had a blind proximal esophageal pouch and distal tracheoesophageal fistula (TEF). Ninety-two patients were initially treated by gastrostomy. Fifty-nine patients were born at term and had no major medical problems. They underwent either primary or delayed (up to 1 week) repair of the esophagus. The survival rate was 93% and surgical mortality was 3%. Twenty-two patients born prematurely or having major medical problems underwent a staged repair. Survival in the staged group was 55% with a surgical mortality of 27%. Overall survival for the entire series was 79% (79/100). The use of air rather than contrast material for the radiographic diagnosis of EA was associated with fewer subsequent pulmonary abnormalities. There was an 18% incidence of minor anastomotic leaks, regardless of technique. The route of approach (transpleural, retropleural) did not influence mortality or morbidity in this series. There were more than three times as many significant strictures among patients who had double-layer anastomoses (18%) as among those who had single-layer repairs (5%). The surgical treatment of infants with EA has reached a level in which associated illness, anomalies, or prematurity are now the most significant determinants of survival.