Probabilistic Modeling of Cytomegalovirus Infection Under Consensus Clinical Management Guidelines

Background. Cytomegalovirus (CMV) is the most common viral pathogen after renal transplantation and remains a major therapeutic challenge with important clinical and economic implications from both direct and indirect consequences of infection. Methods. This 5-year study modeled the relationship between CMV infection and biopsy-proven graft rejection, graft loss, or death after renal transplantation in an inception cohort using Canadian consensus guidelines for CMV management as a component of a detailed cost-analysis of viral infection. Results. Probabilities of CMV viremia and syndrome/disease among 270 sequential graft recipients were 0.27 and 0.09, respectively; 91% of cases occurred in the first 6 months. Probability of CMV infection as the first event was 0.29, with a probability of subsequent biopsy-proven acute rejection (BPAR) of 0.05 (mean: 62 +/- 26 days, range: 32-85 days), whereas the probability of BPAR as the first event was 0.18, with a probability of subsequent CMV infection of 0.38 (mean: 63 +/- 31, range: 27-119 days). Probability of freedom from both CMV infection and BPAR throughout the period of observation was 0.53. Time-dependent Cox analysis showed that neither donor/recipient CMV risk stratum nor CMV infection influenced the risks of BPAR (P=0.24; P=0.74) or of graft loss or death (P=0.26; P=0.34). In contrast, BPAR significantly increased the risk of both subsequent CMV infection (hazard ratio=1.77, P=0.03) and of graft loss or death (hazard ratio=8.31, P<0.0001). Conclusions. Although current antiviral therapy seems to mitigate the reported deleterious effects of CMV infection on BPAR or graft survival, BPAR remains a significantly risk factor for both CMV infection and functional graft survival.