Increased CMV infection following nonmyeloablative allogeneic stem cell transplantation: a search for the guilty

I read with interest the brief report by Bainton et al on cytomegalovirus (CMV) reactivation following the use of Campath-based nonmyeloablative conditioning regimens.1 The authors found a high incidence of CMV infection, similar to that reported by us.2 However, they suggest that fludarabine rather than Campath was responsible for this, but the existing literature on nonmyeloablative transplants does not seem to support the idea. A recent study reported CMV reactivation in 87% with and only 25% without the addition of alemtuzumab (Campath-1H) (P < .001) to fludarabine-melphalan regimen.3 The Seattle group did not find a difference in the incidence of CMV infection or disease with the addition of fludarabine to the low-dose radiation regimen.4Similar low incidences of CMV reactivation were reported by other groups using fludarabine in combination with busulphan, melphalan, or low-dose total body irradiation (TBI) (21%-42%).5,6 The median time of onset of CMV infection was also beyond 45 days in all these studies. The only other regimen associated with a higher and earlier incidence of CMV infection has been a combination of fludarabine and antilymphocyte globulin.7 Thus, fludarabine used alone, without other antilymphocyte antibodies, does not seem to increase the predisposition to earlier or higher CMV infections. Whether Campath used alone rather than in combination with fludarabine would be associated with a lower incidence of CMV infection remains speculative and is not supported by the existent literature.

I would also like to make a few comments regarding the data presented by Bainton et al.1 Firstly, the patients receiving BEAM (BCNU, etoposide, cytosine arabinoside, melphalan)–Campath (those not receiving fludarabine) received transplants only for lymphoma/chronic lymphocytic leukemia (CLL) and mostly received matched related grafts (14 of 18). On the other hand, those receiving fludarabine, either as a part of the protocol described by us2 or in addition to BEAM-Campath, were mostly recipients of unrelated donor grafts (11 of 18) and received transplants mostly for diseases other than lymphoma/CLL (11 of 18). Although the authors mention that there was no difference between related and unrelated donors (UDs), this comparison would be restricted entirely to the fludarabine group, as there were no transplants from unrelated donors in the other group. Thus, to attribute the increased CMV reactivation to fludarabine alone might not be entirely acceptable given the above differences. Given the small sample size and the heterogeneity, the power of a multivariate analysis taking the donor type or underlying diagnosis into account might not be satisfactory either.

Secondly, Bainton et al stated that there was no difference in the incidence of CMV reactivation between patients receiving Campath-1H (alemtuzumab) (15 of 16) and Campath-1G (13 of 20). In fact, theP value by Fisher exact test (2-tailed) turns out to be .05. Although the conventional cut-off for significance is .05, it might not be entirely acceptable to ignore a P value of .05 and formulate the inferences on a P value of .04 (the Fisher exact P value for CMV reactivation with and without fludarabine), given the small number of patients. Hence, the statistical interpretation indicates a suggestive trend toward significantly more CMV reactivation in the alemtuzumab group. The effect of alemtuzumab, as we had mentioned, was not only on the incidence of reactivation, but also on the recurrence both before and after 100 days. Late recurrences were correlated with slow recovery of CD4⁺ T cell counts. And without analyzing these factors and given the above data, it cannot be claimed with certainty that both of these antibodies have a similar effect on CMV reactivation.

Finally, Bainton et al suggested that halving the dose of alemtuzumab might not result in reduction in reactivation of CMV. Indeed, that might be the case and further dose reduction could be necessary, but Bainton et al have used alemtuzumab to day 1 in the protocols other than the one similar to ours. The existing data suggests that the use of alemtuzumab closer to the time of transplant results in longer persistence of the antibody.8 Thus, only reduction in the dose of alemtuzumab might not suffice, and consideration must be given to its timing in relation to stem cell infusion. Ultimately, how and when to use Campath antibodies in nonmyeloablative conditioning are yet to be perfected, and clinical studies to explore that are ongoing.