Influence of Donor Type (sibling vs matched-unrelated donor vs haplo-identical donor vs cord blood) on Outcomes after Clofarabine-Based Reduced-Intensity Conditioning Allograft for Myeloid Malignancies

Background: Recently, we have reported the good results of a clofarabine-based reduced-intensity conditioning (RIC) regimen for peripheral blood stem cell (PBSC) allografts, using either matched (Le Bourgeois, ASH 2017) or haploidentical (haplo) donors (Chevallier, ASH 2017). However, the influence of donor type in this setting is unknown.

Patients and Methods: We conducted a retrospective study including all adults (≥18 yo) allografted in our department with a clofarabine-based RIC regimen for a myeloid malignancy. All types of donors were accepted, and patients received allogeneic peripheral blood stem cells (PBSC) or cord blood (CB) transplants. The aim of this study was to assess whether or not the donor type (sibling, matched unrelated (MUD), haplo or CB) impacted outcomes.

Results: Between October 2009 and February 2018, 122 patients met the inclusion criteria. Donors were sibling in 36 cases, "MUD" in 55 (including one 9/10 mismatched donor), haplo in 27 and CB in 4. The sibling+MUD patients received a CloB2A1/A2 RIC regimen consisting of clofarabine 30 mg/m²/day 4 to 5 days (Clo), busulfan 3.2 mg/kg/day 2 days (B2) and 2.5mg/kg/day of rabbit anti-thymocyte globulin 1 or 2 days (A1 or A2). Haplo and CB recipients were conditioned by a Baltimore-like RIC regimen and a Minneapolis-like regimen, respectively, where fludarabine were replaced by clofarabine. All patients received cyclosporine (CsA) + mycophenolate mofetyl as GVHD prophylaxis, except sibling recipients who were given CsA alone. Haplo recipients received also 2 days of post-transplant cyclophosphamide. All non-CB patients received PBSC as source of graft.

The median age of the whole cohort was 61,5 years old (range: 18-73) and the median follow-up 31 months (range: 4,5-106). All patients engrafted except 1 CB patient and 1 haplo recipient who died of infection before engraftment. All CB cases died at 1, 8, 9 and 11,5 months post-transplant, respectively (infection n=1, relapse n=2, chronic GVHD n=1). As a consequence, we thereafter compared outcomes only between the three other donor groups.

There were no significant differences between sibling, MUD or haplo recipient groups in terms of gender, median age, type of disease, ELN 2010 classification for AML, status at transplant (complete remission vs active), disease risk index, donor/recipient CMV status or median graft quantity of CD34+ stem cells. The only differences were the partition of ABO compatibility, less frequent in the MUD group (42% vs sibling 69% vs haplo 67%, p=0,02) and a history of previous allograft (haplo n=7, sibling n=0, MUD n=6, p=0,005).

Neutrophils (>0,5 Giga/L) and platelets (50 Giga/L) recoveries were significantly delayed in the haplo-group (19 days, vs sibling 15 vs MUD 15, p=0.0003; and 31 days vs 11 vs 12, p<0.0001).

Acute grade 2-4 or 3-4 graft versus host disease (GVHD) incidences were similar between the three groups at 50% and 22% for sibling, 34,5% and 14,5% for MUD and 48% and 11% for haplo (p=0,18 for grade 2-4 and p=0,45 for grade 3-4) as well as moderate+severe chronic GVHD incidences (sibling 14%, MUD 16%, haplo 15% p=0,94).

Follow-ups were similar at 34.2 (6-74), 33.7 (4.5-106) and 21.4 (7-53) months respectively for sibling, MUD or haplo donors. This was associated with similar 2-year overall survivals (OS) at 64,7% (50-83), 73,9% (62-87) and 60,2% (43-83) respectively (p=0.4) and 2-year disease-free survivals (DFS) at 57,7% (43-76), 70,9% (59-84) and 53,6% (37-77) respectively (p=0.1). Two-year GVHD-relapse free survival (GRFS) was also similar at respectively 37,9% (24-57), 54,3% (41-70) and 38,9% (23-63) (p=0.2) as well as 2-year non-relapse mortality (NRM) at 14%, 11% and 18,5% (p=0,63).

When considering only AML patients, again similar 2-year survivals were observed between the three groups. OS was 71.4% (54-93) for sibling (n=21), 75.1% (62-90) for MUD (n=40) and 66.6% (46-95) for haplo (n=15, p=0.65) and DFS respectively 66,6% (49-90), 70,8% (57-87) and 66,6% (46-95 p=0.47). The same was observed for GRFS at 38% (22-65), 56% (41-75) and 40% (21-74, p=0.14) and NRM at 14%, 7,5% and 7%, p=0.34).

In multivariate analysis, donor type remained independent of outcomes in this series.

Conclusion: These data confirm that since no difference was observed in terms of outcomes, haplodonors represent a valid alternative for patients receiving a clofarabine-based RIC PBSC allograft for myeloid malignancies who lack a matched sibling donor.