Screening for cytomegalovirus (CMV) infection in allogeneic bone marrow transplantation using a quantitative whole blood polymerase chain reaction (PCR) method: analysis of potential risk factors for CMV infection.

Abstract

Potential risk factors for CMV infection and the use of quantitative CMV PCR screening to guide pre-emptive anti-CMV therapy were reviewed retrospectively in 32 allogeneic bone marrow transplant patients accrued over a 2-year period. Significant CMV PCR positivity (an indicator of CMV infection) developed in 34% of patients. When analysed by recipient CMV IgG serostatus, 69% of seropositive recipients developed significant CMV PCR positivity while none of the seronegative recipients did so (P = 0.00007). Considering only the seropositive recipients, 100% of those who received the low intensity campath-1H/fludarabine/melphalan 'mini-allograft' conditioning regimen developed significant CMV PCR positivity, while only 44% of those who had received cyclophosphamide/TBI did so (P = 0.0337). The mean time to first episode of significant CMV PCR positivity for those who had received campath/fludarabine/melphalan was 25 days while for those who had received cyclophosphamide/TBI, this was 66 days (P = 0.0372). For the first episode of significant CMV PCR positivity, the mean index and peak CMV PCR counts for those who had received campath/fludarabine/melphalan were 4.54 and 5.22 log copies/ml respectively, while for cyclophosphamide/TBI, the corresponding figures were 3.85 and 4.12 log copies/ml respectively (P = 0.2986 and P = 0.0472 for index and peak values). 85% of those who had significant CMV PCR positivity with the campath/fludarabine/melphalan regimen developed more than one such episode, while 50% of those receiving cyclophosphamide/TBI regimen did so (P = 0.491). Significant CMV PCR positivity was associated with symptoms in a proportion of patients (pyrexia 45%, cough 18%, rise in AST 72%). No patient developed overt CMV disease. CMV PCR is useful for guiding pre-emptive anti-CMV therapy and for monitoring response.