To guide the choice of medical versus surgical therapy for patients with descending (type B) aortic dissection (tear in the descending aorta without involvement of the ascending aorta), multivariate survival analysis was applied to 136 patients admitted to two medical centers between 1975 and 1988 with acute (n = 89) or chronic (n = 47) descending dissection: group 1, all 136 patients; group 2, 106 patients without rupture, pulse loss, or visceral organ compromise; and group 3, 56 patients from group 2 without major cardiac or renal disease (23 surgical and 33 medical). Group 3 medical and surgical subgroups were well matched for baseline characteristics and were potential candidates for either mode of therapy. By Cox model analysis, significant predictors of mortality were pleural rupture, other dissection complications, increasing age, and cardiac disease (all p less than 0.01). Surgical versus medical therapy was not an independent determinant of survival in any of the three groups for acute or chronic dissection. Survival probabilities for all group 3 patients at 1, 5, and 10 years were 0.94, 0.87, and 0.32 (medical) and 0.90, 0.80, and 0.50 (surgical). Despite the limitations of this retrospective study (including the possibility of undefined treatment selection biases), these data suggest that medical or early surgical therapy is associated with equivalent outcome in selected patients with uncomplicated acute or chronic descending aortic dissection.