A prospective, randomized study was conducted to compare the retroperitoneal versus transperitoneal approach for elective aortic reconstruction. One hundred thirteen patients (transperitoneal = 59, retroperitoneal = 54) were randomized between March 1987 and October 1988. In addition, to assess the changing course of patients undergoing aortic reconstruction similar data were gathered retrospectively on a group of 56 patients undergoing aortic reconstruction by the same surgeons performed via a transperitoneal approach in 1984 to 1985. Randomized patients were identical in age, male to female ratio, smoking history, incidence and severity of cardiopulmonary disease, indication for operation, and use of epidural anesthetics. Details of operation including operative and aortic cross-clamp times, crystalloid and transfusion requirements, degree of hypothermia on arrival at the intensive care unit, and perioperative fluid and blood requirements did not differ significantly for patients undergoing transperitoneal versus retroperitoneal reconstruction. Respiratory morbidity, as assessed by percent of patients requiring postoperative ventilation, deterioration in pulmonary function tests, and the incidence of respiratory complications, was identical in randomized patients. Other aspects of postoperative recovery including recovery of gastrointestinal function, the requirement for narcotics, metabolic parameters of operative stress, the incidence of major and minor complications, and the duration of hospital stay were similar for randomized patients undergoing transperitoneal and retroperitoneal reconstruction. When compared to retrospectively reviewed patients having aortic reconstruction, randomized patients undergoing transperitoneal and retroperitoneal operations had highly significant (p less than 0.001) reductions in postoperative ventilation, transfusion requirements, and length of hospital stay. Such trends were all independent of transperitoneal versus retroperitoneal approach.(ABSTRACT TRUNCATED AT 250 WORDS)