BACKGROUND: Cytomegalovirus (CMV) disease is one of the major challenges of lung transplantation that may determine outcome. The benefits of ganciclovir prophylaxis seem indisputable, but no consensus has been reached on the optimal duration of therapy. Results with different protocols suggest that efficacy is related to the duration of treatment. MATERIALS AND METHODS: To evaluate the additional effect of a prolonged regimen throughout the maximal immunosuppression phase, we conceived a protocol administering ganciclovir, 5 mg/kg/d for 20 weeks from the first postoperative day, to all CMV-seropositive patients undergoing lung transplantation or receiving the lung from a seropositive donor. Virus shedding was routinely measured in body fluids including through BAL. Costs and outcomes are compared with those in shorter prophylaxis protocols from previous reported studies. RESULTS: Of 30 lung transplant recipients, 22 patients at risk for CMV reactivations were observed for (mean SD) 22.9 +/- 13.2 months. CMV infections were detected in eight patients 8.6 +/- 5.1 months after transplantation. CMV pneumonitis developed in one patient 9 months following the transplant event. Prolonged IV ganciclovir prophylaxis was, in general, well tolerated. However, five patients had bacteremia and one had a local thrombosis, with no long-term consequences. A prescription for 8 additional weeks of prophylaxis to cover the whole period of enhanced immunosuppression decreased the cumulative incidence of first CMV infections by 29% 1 year after transplantation compared to 12-week regimens reported in other studies that indicated a 50% reduction in the incidence of first CMV infection. The total cost of 20 weeks of IV ganciclovir prophylaxis was $6,010 (US dollars) per patient more expensive than 12 weeks of IV ganciclovir therapy. This was not offset by the reduced requirement for treatment of infections. Indeed, extrapolating to our cohort of patients, the additional cost per patient was seven times greater than the treatment for the infections that were reported after the 12-week prophylaxis protocol. CONCLUSION: Prolonging ganciclovir prophylaxis to 20 weeks decreased by half the rates of CMV infection when compared to reports from centers using a shorter protocol of 12 weeks for ganciclovir prophylaxis. Additionally, a delay in the onset of the first infection was observed. Nevertheless, the increase in costs and the discomfort of long-term use of venous catheters are important factors that may favor a shorter regimen of 12 weeks followed by preemptive therapies each time CMV infections occur.