Donor CMV serostatus has no impact on CMV viremia or disease when prophylactic granulocyte transfusions are given following allogeneic peripheral blood stem cell transplantation

In the neutropenic phase following allogeneic peripheral blood stem cell (AlloPBSC) transplantation, recipients are at risk for life-threatening infections. The ability to collect an increased number of granulocytes from donors stimulated with granulocyte colony-stimulating factor (G-CSF) has rekindled interest in granulocyte transfusions (GTs) as a possible approach to this problem.1,2 Preliminary data from our institution have suggested that prophylactic G-CSF–mobilized HLA-matched GTs from the PBSC donor may reduce antibiotic utilization and febrile days in neutropenic AlloPBSC recipients and may improve survival.3,4 

However, granulocytes and peripheral blood mononuclear cells (MNCs) are thought to be reservoirs for cytomegalovirus (CMV).5-7 Despite the introduction of intensive CMV surveillance strategies and prophylactic/preemptive antiviral therapy, CMV infection remains a major cause of morbidity after allogeneic transplantation.8-12 Traditionally, CMV-seropositive individuals are excluded as granulocyte donors, especially for CMV-seronegative recipients.13 With the prevalence of CMV seropositivity in the general population being as high as 70% in certain geographic areas, this exclusion criteria alone results in a substantial diminution of the pool of potential granulocyte donors. The published literature on the effect of GTs on CMV infection is limited, and most antedate the era of G-CSF–mobilized GTs, the use of AlloPBSCs and, most importantly, the availability of effective prophylactic/preemptive therapy for CMV.14,15 We therefore chose to study whether donor CMV serologic status had any effect on CMV infection in a large cohort of AlloPBSC recipients following infusion of G-CSF–mobilized prophylactic granulocyte components obtained from AlloPBSC donors.

Related donor AlloPBSC recipients were scheduled to receive or not receive HLA-matched G-CSF–mobilized prophylactic GTs on a prospective study approved by the Institutional Review Board of Washington University. Randomization was biologic, determined by availability of an ABO-compatible AlloPBSC donor. The same ABO-compatible sibling served as both the AlloPBSC and granulocyte donor. Granulocyte donors were selected without regard to donor CMV serologic status. G-CSF (10 μg/kg) was administered to granulocyte donors 12 hours prior to each leukapheresis. Granulocyte components were collected and irradiated as previously reported.16 GTs were administered on day +3 and +6 or on day +5 and +7 based on the conditioning regimen employed. To avoid human error, all AlloPBSC recipients received CMV-seronegative or filtered red cells and platelets irrespective of recipient CMV serologic status. Recipients were monitored for CMV viremia every 2 weeks (from engraftment to day + 180) by shell vial/tube culture. Preemptive therapy with intravenous gancyclovir 5 mg/kg once daily for 21 days was instituted following initial detection of CMV viremia. Patients received acyclovir for herpes simplex virus and trimethoprim-sulphamethoxazole for Pneumocystis carinii prophylaxis.

A total of 225 patients underwent AlloPBSC transplantation. Conditioning regimens were cyclophosphamide (CY)/total body irradiation (TBI) (129), etoposide/CY/TBI (58), busulfan/CY (15), and other (23). AlloPBSC products were infused without T-cell depletion. Graft-versus-host disease (GVHD) prophylaxis consisted of cyclosporine (CSA) and prednisone in 96 patients and CSA alone in 129 patients. Eighty-three patients received prophylactic GTs (cohort A), and 142 patients did not (cohort B). Data on patient characteristics and leukocyte subsets administered in the 2 cohorts are summarized in Tables 1-2. It is of particular relevance that there were no statistically significant differences with respect to observed rates of acute GVHD in the 2 cohorts. As expected, cohort A received a greater number of neutrophils. This cohort also received a greater number of T cells (CD3⁺), T-cell subsets (CD4⁺ and CD8⁺), and natural killer (NK) cells (CD16⁺CD56⁺).

The overall incidence of CMV viremia was not different between recipients in cohorts A and B: 26 of 83 (31%) (95% confidence interval [CI], 22%, 42%) versus 49 of 142 (34%) (95% CI, 27%, 43%), respectively. The median time to detection of viremia was similar in cohort A (median, 35 days; range, 17-53 days) and cohort B (median, 36 days; range, 17-77 days). There was also no difference in the overall incidence of CMV disease between recipients in cohorts A and B: 6 of 83 (7.2%) (95% CI, 1.2%, 8.0%) versus 5 of 142 (3.5%) (95% CI, 2.7%, 15%), respectively. A detailed analysis of 4 subgroups based on donor and recipient CMV serostatus showed that within each subgroup prophylactic GTs had no effect on the incidence of CMV viremia (Table 3). In particular, the incidence of CMV viremia between cohorts A and B was not different in the CMV-seronegative recipients who had CMV-seropositive donors (subgroup 2).

Two reports in the early 1980s found an increased incidence of CMV infection in seronegative recipients of GTs and formed the basis of the general policy of excluding CMV-seropositive individuals as granulocyte donors for immunocompromised CMV-seronegative recipients. The report of Winston et al was limited by a small and heterogeneous group of patients that received GTs from multiple granulocyte donors, most of whom did not have CMV serologic data available.15 Although seronegative granulocyte recipients did have a statistically higher prevalence of CMV infection, this was mainly accounted for by a higher incidence of asymptomatic seroconversion in the cohort receiving GTs. The rate of CMV disease among patients who did or did not receive GTs was the same. Hersman et al reported a higher incidence of CMV infection in seronegative recipients of GTs from seropositive granulocyte donors when compared with seronegative patients who did not receive GTs or seronegative patients who received GTs from seronegative donors.14 However, it appears that CMV-seronegative patients were permitted to receive red cell and platelet transfusions from donors irrespective of donor CMV serostatus. This potentially confounding variable could explain why an extremely high proportion (35%) of seronegative patients who received GTs from seronegative donors developed CMV infection. Recently, the debate on this issue has been rekindled with Narvios et al suggesting that screening of potential granulocyte donors for CMV antibody is not warranted, whereas Nichols et al disagree.17,18 

As expected, our analysis of the cohort that did not receive prophylactic GTs showed that a seronegative recipient of AlloPBSCs from a seropositive donor had a greater risk of CMV viremia compared with a seronegative recipient who received AlloPBSCs from a seronegative PBSC donor (30.8 vs 2.7%) (Table 3). However, the additional transfusion of granulocyte components from CMV-seropositive donors to CMV-seronegative recipients did not significantly increase the risk of CMV viremia over that observed in similar donor-recipient CMV serostatus pairs given AlloPBSCs alone (33.3% vs 30.8%) (Table3). Although granulocyte components were not tested for the CMV genome and the power of the subgroup analysis based on donor and recipient serostatus is limited by the small sample size, it appears that most of the risk of CMV viremia was derived from the transfusion of the AlloPBSC product from a CMV-seropositive donor. The transfusion of granulocyte components collected from CMV-seropositive donors did not add substantially to this risk. This may be due to passive transfer of immunity to CMV by the large numbers of immune effector cells in the granulocyte components.

Our data show that in an era of effective surveillance and preemptive therapy for CMV, AlloPBSC recipients can safely receive 2 transfusions of prophylactic G-CSF–mobilized granulocyte components from CMV-seropositive AlloPBSC donors. This knowledge may help expand the donor pool in areas with a high prevalence of CMV in the general population.