Cytomegalovirus infection is a major cause of morbidity and rehospitalization after heart transplantation. To assess its incidence and risk factors in the current era of heart transplantation, we analyzed cytomegalovirus infection data in 1553 patients from 26 institutions (Cardiac Transplant Research Database Group) undergoing primary heart transplantation between Jan. 1, 1990, and June 30, 1992. There were 230 treated cytomegalovirus infections in 200 patients, of which 16 were fatal (6%; 70% confidence limits 5% to 9%). Actuarial freedom from cytomegalovirus infection was 98% 1 month, 88% 3 months, and 83% 24 months after transplantation, with a peak incidence of initial infection at 2 months. Twenty-five (12%) of 200 patients with cytomegalovirus infection had recurrent cytomegalovirus infection during a mean follow-up of 13.9 months. The primary location of cytomegalovirus infection was blood in 100 infections (43%), lung in 69 (30%), gastrointestinal tract in 54 (23%), and other sites in seven patients (3%). Cytomegalovirus pneumonia exhibited the highest mortality rate (13%). Risk factors by multivariate analysis for earlier development of cytomegalovirus infection included pretransplantation cytomegalovirus serology (positive donor, negative recipient [p < 0.0001]; positive donor, positive recipient [p = 0.0002]; and negative donor, positive recipient [p = 0.02]) and cytolytic induction therapy (p = 0.05). A cytomegalovirus- positive recipient with a cytomegalovirus-positive donor had a 15% chance of having cytomegalovirus infection, whereas a cytomegalovirus-negative recipient with a cytomegalovirus-positive donor had a 24% chance. Ganciclovir treatment was administered in 227 (99%) of 230 infections. By multivariable analysis, the likelihood of a fatal cytomegalovirus infection was increased with a higher number of infections of any type during the first post transplantation month (p < 0.0001). There was no increased mortality rate in cytomegalovirus infections associated with cytomegalovirus-positive donor and cytomegalovirus-negative recipient (6% mortality rate) versus all other cytomegalovirus infections (6% mortality rate) (p = 0.9) or with OKT3 induction therapy (0% mortality rate) versus all others (noninduction and induction with other than OKT3) (1.4%) (p = 0.03). In conclusion, the biggest determinant of cytomegalovirus infection is donor and recipient pretransplantation cytomegalovirus serologic results with cytolytic induction therapy adding a small additional risk. The overall mortality rate from cytomegalovirus infections is low (7%) in the current era with rapid culture techniques and ganciclovir therapy. Cytomegalovirus infections are more likely to be fatal if there are more frequent preceding infections of any type, but mortality rates are not increased by OKT3 induction or with a cytomegalovirus-positive donor organ transplanted into a cytomegalovirus- negative recipient.