This study determined utility of preoperative spirometry for prediction of postoperative pulmonary complications (PPC) defined as pneumonia, ventilator dependence greater than 48 hours, and adult respiratory distress syndrome in 147 patients undergoing vascular surgery from June 1988 through March 1990 [39 aortic aneurysm repairs, 21 carotid procedures, and 87 operations for occlusive disease including aorto-ileofemoral, infra-inguinal, and visceral]. The incidence of PPC was 12.9 per cent, while cardiac complications (myocardial infarction, congestive heart failure, and ventricular arrhythmias) were present in 9.8 per cent. Prior or current smoking, which was present in 80 per cent, was not predictive of PPC. FEV1 was 2.2 +/- 0.7 L/s (mean +/- 1 SD). Abnormal FEV1 (2.0 or less L/s) was present in 42 per cent (n = 62). For FEV1 of 2.0 or less, PPC rate was 22.5 per cent versus 5.8 per cent for FEV1 greater than 2.0 L/s (P less than 0.005, Fisher exact). The incidence of PPC was 30.7 per cent for aortic aneurysm repair, 8.0 per cent for occlusive disease, and 4.7 per cent for carotid procedures. Abdominal aortic procedures (performed in 67 patients: 39 for aortic aneurysm repair and 28 for aortoiliac occlusive disease) were associated with a PPC rate of 22.4 per cent versus 5.0 per cent for "nonabdominal" procedures (P less than 0.002, Fisher exact). Life table analysis after surgery demonstrated decreased survival for patients with PPC (P = 0.031, Mantel-Haensel) during follow-up (250 +/- 165 days). PPC are associated with abnormal FEV1 and abdominal vascular procedures. In conclusion, preoperative spirometry is useful for the prediction of PPC after vascular surgery.