A multivariate analysis of the risk of cytomegalovirus infection in heart transplant recipients.

Thirty-four consecutive heart transplant recipients were studied over a two-year period for evidence of cytomegalovirus (CMV) infection. Twenty-three episodes of CMV infection were identified; 19 of these occurred within 120 d after transplantation and were statistically analyzed. Fifteen potential risk factors were evaluated, including the following: pretransplant CMV serological status of donor and recipient; recipient's age, sex, race, and cardiac disease; number and type of blood products transfused; type and intensity of immunosuppression; occurrence of rejection; and leukopenia. The Cox proportional hazards model identified both greater than average steroid dosage and positive recipient CMV serology as significant risk factors for CMV infection (P = .014 and .048, respectively). In a separate analysis, however, only greater than average steroid dosage was associated with clinically significant CMV infection (P = .005).