Seventeen RCTs (n=1,980) were included: 11 of universal prophylaxis (n=1,582) and 6 (n=398) of pre-emptive therapy.
All the trials scored 2 or 3 on the Jadad scale. Three of the 17 trials were double-blind (all were trials of universal prophylaxis); allocation concealment was adequate in 5 (all were trials of universal prophylaxis); length of follow-up was assessed as adequate in 7 trials (3 trials of pre-emptive therapy and 4 trials of universal therapy). In general, the trials were of a modest size.
Universal prophylaxis.
There was a statistically significant reduction in the development of CMV organ disease with universal prophylaxis compared with the control group (OR 0.20, 95% CI 0.13, 0.31). There was also a statistically significant reduction in the secondary end points of rate of acute allograft rejection, bacterial and fungal infections, non-CMV viral infections and mortality, but not for graft loss. There was no evidence of statistically significant heterogeneity. The removal of individual trials from the analysis, including the trial with a treatment duration of less than 60 days, did not affect the overall efficacy.
In the subgroup analyses, there was a statistically significant decrease in CMV organ disease compared with controls for donor-positive/recipient-negative CMV serostatus patients, in immunosuppressed patients who received induction therapy with antilymphocyte antibodies, in kidney transplant and liver transplant recipients, and for acyclovir at more than 2 g/day, acyclovir 3.2 g/day and glanciclovir 3 g/day.
Pre-emptive therapy.
There was a statistically significant reduction in the development of CMV organ disease with pre-emptive therapy compared with the control group (OR 0.28, 95% CI 0.11, 0.69). There was also a statistically significant reduction in the secondary end points of rate of acute allograft rejection, but not for graft loss, bacterial and fungal infections or mortality. Non-CMV viral infections could not be assessed. There was no evidence of statistically significant heterogeneity.
In the subgroup analyses, there was no statistically significant decrease in CMV organ disease compared with controls for donor-positive/recipient-negative CMV serostatus patients, nor in immunosuppressed patients who received induction therapy with antilymphocyte antibodies, nor in liver transplant recipients. There was a statistically significant decrease in CMV organ disease for kidney transplant recipients compared with controls.
There was no evidence of publication bias as assessed by the Egger regression method.