The course of 52 isolated single and double lung transplant recipients who survived more than 2 weeks was reviewed to determine the incidence of cytomegalovirus (CMV) infection. In this group 22 patients were seromatched: 11 donor-recipient pairs were seronegative and 11 were seropositive. Of the remaining 30, 13 were donor-positive, recipient-negative and 17 were donor-negative, recipient positive. Diagnosis of CMV infection was made in the event of (1) seroconversion, (2) clinical symptoms consistent with CMV plus a fourfold rise in CMV titer, (3) isolation of CMV from tissue or body fluid by shell vial monoclonal antibody technique. CMV pneumonitis was diagnosed when shell vial culture from bronchoalveolar lavage fluid was positive in the appropriate clinical setting even if cytopathic changes were not yet present. Clinical CMV infection did not develop in any of the 22 seromatched pairs, and none of the seronegative pairs seroconverted. Of 17 recipient-positive, donor-negative pairs, CMV pneumonitis developed in two; one died. Of 13 recipient-negative, donor-positive pairs, seven seroconverted and pneumonitis developed in two but they did not die. The most recent 10 mismatched pairs received hyperimmune globulin prophylaxis, but this did not prevent the development of infection or clinical disease. Morbidity and mortality were greater in the seromismatched groups than in seromatched groups although this difference could not be directly related to CMV infection in most cases. Our experience suggests that both seronegative and seropositive recipients who receive a mismatched graft, but not the usual high-risk seropositive matched pairs, are at significant risk of clinical CMV disease after isolated single and double lung transplantation.