The objective of this randomized, prospective study was to compare preemptive to deferred treatment of cytomegalovirus (CMV) infection in high-risk renal transplant recipients. Conducted at a university-affiliated transplant center, the study included 36 renal allograft recipients with donor or recipient CMV-seropositivity who received anti-thymocyte induction therapy. Ganciclovir was administered intravenously for 21 days upon detection of CMV viremia (preemptive, N = 15) or detection of CMV viremia associated with a CMV syndrome (deferred, N = 21). Shell vial culture, conventional culture, and polymerase chain reaction (PCR) were performed upon buffy-coat specimens weekly for 12 to 16 wk. CMV and non-CMV-associated charges were calculated. The comparative sensitivities of PCR, shell vial culture, and conventional culture were 91%, 44%, and 47%, respectively. A delay in specimen processing of > 24 h severely compromised the sensitivity of culture techniques but not that of PCR. Preemptive therapy tended to decrease symptomatic CMV episodes (0.4 versus 0.6 episodes per patient randomized; P = 0.22). One patient in each group had organ involvement, and no patient died. Allograft function and survival were similar. Ganciclovir use was increased in the preemptive group (1.2 versus 0.6 courses per patient randomized; P = 0.02). CMV-associated charges were $10,368 (preemptive) versus $5,752 (deferred); P = 0.13. PCR is superior to conventional monitoring to detect CMV viremia. Culture cannot be considered the "gold standard" for detection of CMV viremia, especially when transport of specimens over distances results in processing delays. Preemptive therapy may reduce symptomatic CMV infections in renal transplant recipients. It was associated with higher CMV-related charges but equivalent overall charges versus deferred treatment with intensive monitoring. Either strategy can achieve control of CMV infection after renal transplantation.