We have previously reported that the incidence of acute graft-versus-host disease (GVHD) after myeloablative peripheral blood stem cell transplantation correlates more strongly with the blood donor CD34+ cell count after G-CSF mobilization than with the dose of CD34+ cells infused to the recipient (Dhedin et al. Exp Hematol 2006). The aim of this study was to confirm this finding in non myeloablative transplantation. We studied 92 patients transplanted with HLA-identical sibling donor after a conditioning regimen associating 2GY total body irradiation and fludarabine. The median blood donor CD34+ cell count on day 5 of G-CSF administration was 70/μl (range 9–210/μl). The median infused CD34+ and CD3+ cell doses per kilogram of recipient bodyweight were respectively 6.4 × 106 (range 1.3–22.6) and 275 × 106 (range 88–839). The cumulative incidence rates of grade II–IV aGVHD and grade III–IV aGVHD were respectively 34.4% (se=0.05) and 9.9% (se=0.03) on day 100. The cumulative incidence rates of all GVHD and extensive cGVHD at 24 months were respectively 60.5% (se=0.06) and 36.2% (se=0.06). Donor and recipient characteristics, the doses of infused CD34+ and CD3+ cells, and the donor CD34+ cell count after G-CSF mobilization were analyzed as potential risk factors for GVHD. No correlation was found between these parameters and the occurrence of acute GVHD. As in several previous studies, the CD34+ cell dose tended to influence the incidence of extensive cGVHD, but the correlation was not statistically significant (p=0.14). In contrast, the blood donor mobilized CD34+ cell count, divided into quartiles, was the only factor significantly associated with extensive chronic GVHD: In univariate analysis, the incidence of extensive cGVHD was significantly higher in the quartile of patients (N=24) whose donors had the highest mobilized CD34+ cell counts (quartile n°4: >109 CD34+ cells/ml at day 5 of G-CSF mobilization), compared to the three other quartiles (cumulative incidence of extensive cGVHD in quartile n°4 was 58.9% compared to respectively 20.8, 37.5 and 25% in the third, second and first quartiles (p=0.03)). In multivariate analysis, the donor CD34+ cell count was the only variable associated with the risk of extensive cGVHD: the hazard ratio associated with the highest quartile was 2.59 (95% CI [1.30–5.18], p=0.007).In conclusion, we confirm in non myeloablative transplants, that donor response to G-CSF is a better prognosis factor of GVHD, than the dose of CD34+ cell infused. “Outstanding mobilizers” seemed to have greater alloreactive potential. These findings do not support limiting the dose of CD34+ cells infused to prevent GVHD. They have also direct implications for transplant management, and especially for the modalities of GVHD prophylaxis.

Author notes

Disclosure: No relevant conflicts of interest to declare.