Transplantation from alternative donors has been used since several years for patients with hematological malignancies lacking of HLA identical donors. The Baltimora group pioneered the use of cyclophosphamide (Cy) after haploidentical unmanipulated (BM) stem cells infusion. BM was chosen to reduce the risk of acute and chronic graft versus host disease (GVHD). PBSC could replace BM, fastening engraftment and immunological reconstitution. However, the risk of GVHD could be higher.

We retrospectively analyzed a cohort of patients from two institutions, receiving haploidentical transplantation with nonmyeloablative conditioning regimen (NMA) and PSBC or BM, with post-infusion Cy.

Patients and Methods

From April 2009 to April 2013, 72 patients with poor prognosis hematological malignancies received haploidentical transplantation. Conditioning regimens consisted of Cy 14.5 mg/kg d -5 and -6, fludarabine 30 mg/m2 d-6 to d-2, and low dose TBI (2 Gy) at d-1. GVHD prophylaxis consisted of Cy 50 mg/kg day +3 and +4, tacrolimus (FK, 1 mg total dose, in continuous infusion) until days +180 (Milan cohort) or cyclosporine (CsA, 3 mg/kg) (Marseille cohort) and MMF (15 mg/kg 3 per day) until day +35. FK, CsA and MMF were started at d +5. G-CSF was started at d +5 in all patients. Donors underwent bone marrow harvest under general anesthesia and a total of 4 x 10e8 nuclear cells per kg of recipient was targeted. Unmanipulated bone marrow was used as stem cell support at d0. In Marseille, donors were mobilized using 5 to 6 days of subcutaneous G-CSF (Neupogen®) (10 mcg/kg/day). A minimum of 4 x 10e6 CD34/kg was harvested.


The median follow-up was 12 months (range: 1-48). For the population as a whole, the median time to ANC more than 0.5 x 10e9/L was 20 days (14-32) and the median time to transfusion-independent platelet (PLT) count was 29 days (14-52). Engraftment results in the two cohorts of patients (BM vs PBSC) were not significantly different [ANC 21 days (14-32) vs 20 days (14-27), and PLT 29 days (16-46) vs 27 days (14-52)]. Overall, aGVHD 2-4 incidence was 27% and cGVHD was 13%. No difference was founded in the two cohorts: aGVHD 2-4 24% vs 33% and cGVHD 11% vs 17%. The 1-year non relapse mortality was 18% overall, and it was not statistically different even if numerically lower in the PBSC group (22% vs 9%).


This retrospective study showed that there was not significant differences in terms of hematological reconstitution and both acute and chronic GVHD, using unmanipulated BM and PBSC after NMA conditioning regimen and post-infusion Cy. The 1-year NRM, even if not statistically different, was lower with PBSC. These data warrant confirmation with more patients and longer follow-up.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.