We read with interest the recent report by Vij and colleagues,1 in which they suggested that the cytomegalovirus (CMV) serostatus of the donor has no impact on CMV viremia or disease when 2 prophylactic granulocyte transfusions are given to allogeneic stem cell transplant recipients (on day +3 and +6 or +5 and +7). Patients in this study were allocated to receive granulocyte transfusions (or not) according to “biologic” randomization, namely, the availability of an ABO-compatible, HLA-matched sibling donor. Thus, the granulocyte donor for these patients was the same HLA-matched sibling who served as the allogeneic peripheral blood stem cell (PBMC) donor. In the subgroup of primary interest (CMV-seronegative recipients of stem cells from seropositive donors [D+/R–], patients who are at risk for primary infection from this approach), those who received granulocyte transfusions had similar rates of CMV viremia (5 of 15; 33%) when compared to those who did not receive granulocytes (8 of 26; 30.8%). Rates of CMV disease appeared to be higher in the granulocyte arm (2 of 15; 13.3%) than in the standard arm (1 of 26; 3.8%), though not significantly so. The authors conclude that the approach is therefore safe.
Closer inspection of their findings, however, may be needed. It is particularly important to note the limited setting to which these conclusions may apply, namely, the use of 2 prophylactic granulocyte transfusions from the same matched sibling donor who provided the PBSC product. The study does raise the possibility that there simply are blood donors who transmit CMV (via the stem cell allograft, granulocyte, or other blood product) and those who do not, a hypothesis that is supported by the near-identical primary infection rate in the 2 groups. One possible explanation would be that a critical viral load in the stem cell and/or granulocyte product is required for CMV transmission, and this variable is relatively stable in the same donor for 1-2 weeks (the time period between stem cell donation and the 2 prophylactic granulocyte donations). While this may be true, one also must consider that external triggers may make immunocompetent donors transiently viremic,2 which could theoretically place the CMV-seronegative recipient at risk even when the same donor is used for both the stem cell product and the granulocyte products. Unfortunately, the application of current technology (such as highly sensitive polymerase chain reaction for CMV DNA) has not permitted identification of those donors (or those products) who are more likely to transmit.3
The conclusions drawn in this study are thus not generalizable to the population suitable for therapeutic granulocyte transfusions, whose numbers are potentially much larger and who increasingly rely on unrelated granulocyte donors.4 This is because the frequency of granulocyte transfusions and the number of different donors used in the treatment setting are significantly higher than those used for prophylaxis, which increases the probability that the seronegative recipient would receive an “infectious” product. Furthermore, the biology of transfusion-transmitted CMV infection (which is thought to be augmented by alloreactivity between donor and recipient5,6 ) suggests that matched related granulocyte products are likely to be associated with a far lower risk for primary CMV infection when compared with multiple unmatched products from family members or community donors. Thus, the impact of this study on the overall use of granulocyte transfusion therapy is likely to be quite small.
It bears repeating that CMV infection has proven to be a risk factor for overall mortality after stem cell transplantation (SCT) even in the current era of effective antiviral therapy,7,8 and as a result, CMV-seronegative or leukoreduced blood products are recommended by the Center for Disease Control and Prevention to prevent primary CMV infection in seronegative SCT recipients.9 Clearly, multiple granulocyte transfusions from CMV-positive donors do not meet this standard. Furthermore, granulocyte transfusions have not yet been proven to be effective.4 Vij and colleagues have shown that 2 prophylactic granulocyte transfusions may indeed be safe if taken from their matched sibling SCT donor. But because we believe that the results are not generalizable to the setting of therapeutic granulocyte transfusions, the conclusion that “this knowledge may help to expand the donor pool in areas with a high prevalence of CMV in the general population”1 may miss the mark.
Response: Granulocyte transfusions and cytomegalovirus infection
We agree with the letter by Nichols et al and stated explicitly in our original manuscript that our observation that donor cytomegalovirus (CMV) serostatus has no impact on CMV viremia or disease is based on allogeneic peripheral blood stem cell recipients receiving 2 transfusions of prophylactic granulocyte colony-stimulating factor (G-CSF)–mobilized granulocyte components from CMV-seropositive donors.1 In addition, we wish to reiterate that these granulocyte recipients underwent aggressive surveillance and received pre-emptive therapy for CMV viremia.
However, we think that our study, based on a relatively large cohort of patients, is provocative and leads one to question what has been accepted as dogma in this field. The practice of excluding CMV-seropositive individuals as granulocyte donors for CMV-negative recipients is based on 2 reports published in the early 1980s.2,3 Unfortunately, both of these earlier publications had serious inherent flaws in their design, as pointed out in the discussion section of our original manuscript. In contrast, our study was well designed and free of these confounding variables.
The concerns raised by Nichols et al on the potential for variability in viral loads and the increased alloreactivity of unmatched donors are theoretical and need to be studied in the clinical setting. We are of the opinion that our data provide sufficient grounds for a prospective study using aggressive CMV surveillance and a pre-emptive therapy strategy in a population of patients receiving multiple unmatched granulocyte products from family members or community donors. We stand by our conclusion that this knowledge may help expand the donor pool in areas with a high prevalence of CMV in the general population.
Correspondence: Ravi Vij, Section of Bone Marrow Transplantation and Leukemia, Campus Box 8007, Washington University School of Medicine, 660 S Euclid Ave, St Louis, MO 63110
