Abstract
Background. Initial reports suggested that CMV disease is delayed after HLA matched-related NM HCT and that unrelated donor NM HCT shows infections risks similar to myeloablative HCTs (Blood99:1978, 2002; BJH123:662, 2003). These reports were hampered by small sample sizes, which limited complex multivariate modeling as well as the analysis of ganciclovir-related neutropenia (GCV-N) and outcome of CMV disease. The high incidence of GCV-N and the poor outcome of CMV disease have been associated with myeloablative conditioning (Blood90:2502, 1997; CID14:831, 1992). The present analysis was performed in a large cohort of recipients of NM HCTs, with contemporaneous myeloablative HCT recipients serving as controls.
Methods. We compared outcomes between 342 recipients of NM HCTs and 2154 myeloablative (M) HCT recipients (median age 52.9 vs. 39.8 yrs), transplanted between 1/94 and 12/03. NM conditioning consisted of 2 Gy TBI with or without fludarabine. Postgrafting immunosuppression consisted of MMF/CSP for NM- and MTX/CSP for M-HCT recipients. CMV surveillance was done by weekly antigenemia (AG) or PCR testing. GCV/VGCV was given for CMV AG/PCR positivity. CMV endpoints included any AG/DNA detection (CMV infection) by day 100; AG > 10/200,000 PBL or PCR > 1000 copies/mL (CMV-high grade) by day 100 and CMV disease by 1 yr after HCT (among CMV R+ and D+/R- patients). GCV-N was defined as non-relapse-related neutropenia (ANC < 1000, < 500/μL) after start of preemptive therapy for G/PCR positivity. Postengraftment neutropenia was also modeled in the entire cohort. Univariate and multivariable models wer formed to assess the risks of all endpoints.
Results. There was a trend toward less CMV infection (49% vs. 55%, adjusted HR 0.9, p=0.22) and high-grade CMV infection (23% vs. 12%, adj. HR 0.7, p=0.09) in NM-HCT compared to M-HCT recipients. In seropositive recipients the difference for high-grade CMV infections was significant (adj. HR 0.6, p=0.02). CMV disease overall was not different between NM- and M-HCT recipients (HR 0.9, p=0.63), although a delay in onset was noted, especially in matched related NM HCT recipients. Overall, a decline of CMV disease incidence was noted after 2000 (adj. HR 0.7, p=0.03); risk factors for CMV disease during the study period were acute GvHD (adj. HR 2.5, p<0.001), chronic GvHD (adj. HR 4.2, p<0.001), and HSV recipient seropositivity (adj. HR 1.6, p=0.01). Survival after CMV disease was not significantly different between NM- and M-HCT recipients. GCV-N was similar between NM- and M-HCT recipients (ANC < 1000, 50% vs. 39%, HR 1.4, p=0.12; ANC < 500, 31% vs. 17%, adj. HR 1.7, p=0.1). Risk factors for severe GCV-N were recipient age (adj. HR 1.2 [10 yr increments], p=0.03) and chronic GvHD (adj. HR 1.9, p<0.01). There was no decline in GCV-N throughout the study period. In a model that included the entire cohort, risk factors for severe post-engraftment neutropenia (ANC<500) included NM-HCT, BM as cell source, HSV serostatus, acute and chronic GVHD, and CMV infection.
Conclusions. NM conditioning does not appreciably reduce the risk or outcome of CMV infection and disease. The risk of GCV-N was similarly high in NM- and M-HCT recipients. Possible explanatory factors for the high incidence of GCV-N in NM-HCT are higher recipient age and use of myelotoxic co-medications (e.g. MMF, TMP-SMX). Improved CMV prevention strategies are needed for both NM- and M-HCT.
High incidence of cytomegalovirus infection after nonmyeloablative stem cell transplantation: potential role of Campath-1H in delaying immune reconstitution