To evaluate the incidence and clinical features of cytomegalovirus (CMV) pneumonitis after cardiac transplantation, we identified 27 (16%) of 171 consecutive recipients in whom CMV pneumonitis was confirmed by strict diagnostic criteria. Cytomegalovirus pneumonitis occurred in 6 (30%) of 20 patients treated with azathioprine and prednisone, and 8 (25%) of 32 patients treated with azathioprine, cyclosporine, and prednisone, but only 13 (11%) of 119 patients treated with cyclosporine and prednisone. The incidence of CMV pneumonitis was not related to recipient preoperative CMV titers or to postoperative cardiac rejection, but there was a trend toward increased CMV pneumonitis in patients who received organs from CMV-positive donors. Mean onset of CMV pneumonitis was 2.9 ±1.6 (SD) months after transplantation. In the azathioprine-prednisone group, CMV was always associated with at least one other respiratory pathogen (Aspergillus, n = 5; Pneumocystis carinii, n = 2). In the two cyclosporine groups, CMV was either the sole respiratory pathogen (n = 9), or associated with P carinii (n = 11). Roentgenographically, diffuse bilateral hazy pulmonary opacities were present in 19 (70%) of 27 patients, but focal subsegmental opacity (26%), small pleural effusion (26%), and lobar consolidation (7%) were also observed. When bronchoscopy was performed, bronchoalveolar lavage was the most sensitive technique for detecting CMV (72%), whereas transbronchial biopsy (39%) and combined washings and brushings (33%) were relatively insensitive techniques. Respiratory failure and death occurred in 52% and 44%, respectively, of patients with CMV pneumonitis. In this population of immunocompromised hosts: (1) CMV pneumonitis, alone or with other respiratory pathogens, was a major cause of morbidity and mortality; (2) localized roentgenographic opacity did not exclude CMV pneumonitis; (3) bronchoalveolar lavage was the most sensitive bronchoscopic technique for detecting CMV pneumonitis.
(Arch Intern Med. 1991;151:1118-1124)