The authors evaluated all respiratory complications of cardiac transplantation in a 10-year study of 94 consecutive recipients. Mean follow-up time was 20 +/- 17 months. The initial 20 patients were treated with azathioprine and prednisone, while the subsequent 74 patients received cyclosporine and prednisone. In the azathioprine group, respiratory infections accounted for 24 of 60 (40%) infections. Two-thirds of the respiratory infections occurred in the first 3 postoperative months and were generally localized processes (focal pneumonitis, nodule(s), abscess, or empyema). Gram-positive and gram-negative bacteria (8/30) and aspergillus (8/30) were the predominant pathogens. Respiratory failure occurred in 29% of infectious episodes. In the cyclosporine group, there were significantly fewer respiratory infections. There was also a reduction in the number of nonrespiratory infections; hence, the percentage of total infections due to respiratory causes, 26 of 50 (52%), was not significantly different. In contrast, however, nearly two-thirds of the respiratory infections in cyclosporine-treated patients occurred after the first 3 postoperative months, and were usually diffuse processes. Despite diffuse disease, respiratory failure was observed with similar frequency (19%). Pneumocystis carinii (9/31) and cytomegalovirus (CMV) (7/31) were the predominant pathogens. CMV pneumonitis tended to occur earlier than that due to P. carinii (2.9 +/- 1.9 mo vs. 9.8 +/- 11.2 mo, respectively), but there was considerable overlap. In comparison with infectious processes, there were 50% fewer noninfectious respiratory complications in both groups. These were primarily pleural (46%) or thromboembolic (18%) disorders. Four of five pulmonary emboli occurred in patients with intercurrent cardiorespiratory illness, and were detected only at autopsy. The authors conclude that respiratory infections account for one-half of all infections observed in cardiac transplant recipients, despite the reduced infection rate associated with the use of cyclosporine. Furthermore, respiratory infections in cyclosporine-treated patients exhibit different clinical and etiologic features than those seen in azathioprine-treated patients. Finally, occult thromboemboli may be difficult to recognize in cardiac transplant recipients because of the high incidence of coexisting cardiorespiratory disease.