Objectives. The aim of this study was to describe our experience in the Cytomegalovirus control of active CMV infection following HSCT using two strategies of CMV infection treatment: ganciclovir universal prophylaxis at low doses and pre-emptive therapy with ganciclovir. Methods. The surveillance was based on the monitoring of antigenaemia (AGM) and on a nested potymerase chain reaction (N-PCR) for the detection of CMV in both strategies. Forty-five recipients with malignant diseases and with a risk for CMV disease received universal prophylaxis (Group A). The non-treated group consisted of 24 patients, most of them with non-malignant diseases who did not receive universal prophylaxis (Group B). Results. In Group A, the incidence of positive AGM was 51%, with a positive PCR of 68.9%. In Group B, the AGM positivity was 66.7% and that of N-PCR was 66.7%. CMV disease occurred in 6/55 patients (10.9%), with 2/36 (5.5%) from Group A and 4/19 (21%) from Group B. Two of these six patients (33.3%) died of CMV disease. Conclusions. Our result suggests that AGM and N-PCR can be used as markers for assessing the monitoring and the introduction pre-emptive therapy. This approach could prove to be more cost-effective than ganciclovir universal prophylaxis for treating CMV infection. (C) 2003 The British Infection Society. Published by Elsevier Ltd. ALL rights reserved.502130137