Abstract
Cytomegalovirus (CMV) reactivation is associated with increased morbidity and mortality after allogeneic stem cell transplantation (SCT). The incidence of “late” CMV reactivation occurring after post-SCT day 100 has increased over time. We analyzed clinical factors associated with the occurrence of late CMV antigenemia in a group of consecutive recipients of allogeneic SCT for hematological malignancies at the M.D. Anderson Cancer Center surviving at least 100 days after SCT. Recipients of T cell depleted, syngeneic or cord blood grafts were excluded. 269 subjects at risk for CMV reactivation were studied with a median follow-up of 14 months after SCT; of these, 144 subjects (54%) experienced early CMV reactivation while 89 subjects (33%) experienced late CMV reactivation. In univariate analyses, factors associated with a significantly increased risk of late CMV reactivation included prior early CMV antigenemia, SCT for a lymphoid diagnosis (HR=2.4, p=0.002), use of a matched unrelated donor (MUD) or mismatched-related (MMR) donor (HR 1.8 vs. recipients of matched-related (MR) allografts, p=0.003), day 100 lymphocyte count<900/μl (HR 1.9, p=0.01), prior acute GVHD (HR=1.5, p=0.003) and the occurrence of chronic GVHD considered as a time-dependent variable (HR=4.2, p<0.001). As expected, early CMV reactivation was a strong risk factor for late CMV reactivation. 65 of the 144 patients (45%) with early CMV reactivation also experienced late reactivation; only 19 of 125 patients (15%) without antecedent early reactivation experienced late antigenemia (HR=4.1, 95% CI 2.4–7). In 132 subjects experiencing early CMV reactivation, we performed a Cox multivariate analysis of risk factors for late CMV reactivation, after excluding subjects who experienced an episode of early CMV reactivation that continued through day +100. We examined the rate of CMV reactivation in groups stratified according to three clinical factors that were independently associated with increased risk in the multivariate analysis: 1) history of acute GVHD; 2) the use of a MMR or MUD donor rather than a MR donor; and 3) transplantation for a lymphoid rather than myeloid malignancy. The rate of late CMV reactivation was lowest for those subjects transplanted for a myeloid malignancy, who had a MR donor and had no history of acute GVHD (n=33, cumulative incidence 10%). In comparison to this group, we observed an increasing rate of late CMV reactivation in subjects with a MR donor, no history of acute GVHD transplanted for lymphoid malignancy (n=18, incidence 22%, HR=2.3, p=0.1); and in subjects with MMR or MUD donor, positive history of acute GVHD transplanted for a myeloid malignancy (n=46, incidence 30% HR=2.8, p=0.04). The rate was highest in lymphoid malignancy patients who received a MMR or MUD allograft and who experienced acute GVHD (n=35, incidence 50%, HR=7.3, p<0.001). If confirmed, the risk assessment model developed here should be useful for the optimization of monitoring and therapeutic strategies designed to minimize the morbidity and mortality associated with late CMV reactivation after allogeneic SCT.
Figure
