Outcomes after haploidentical hematopoietic cell transplantation with post-transplantation cyclophosphamide: a systematic review and meta-analysis comparing myeloablative with reduced-intensity conditioning regimens and bone marrow with peripheral blood stem cell grafts

Abstract

Haploidentical hematopoietic cell transplantation (haplo-HCT) with post-transplantation cyclophosphamide (PTCy) may be the sole available curative option for several hematologic malignancies. However, the best choice of conditioning regimen and graft source has not been established. This study was conducted to compare myeloablative conditioning (MAC) regimens with reduced-intensity conditioning (RIC) regimens and peripheral blood stem cell (PBSC) grafts with bone marrow (BM) grafts in the haplo-HCT setting with PTCy. We performed a systematic review and meta-analysis of studies comparing MAC with RIC and PBSC with BM in the haplo-HCT. The search was conducted in PubMed and TRIALS on February 2, 2021, without a date limit. We excluded studies with >30% non-PTCy graft-versus-host disease (GVHD) prophylaxis and >30% nonmalignant diseases. We screened 570 abstracts from PubMed and TRIALS and selected 20 for full-text review and 17 for inclusion in the qualitative and quantitative analyses. For PBSC versus BM grafts, we found no difference in overall survival (OS; hazard ratio [HR], 1.05; P = .61; nPBSC = 1983; nBM = 2124), progression-free survival (PFS; HR, 0.95; P = .52; nPBSC = 2663, nBM = 2769), graft-versus-host disease (GVHD)-free relapse-free survival (GRFS; HR, 1.16; P = .07; nPBSC = 1454; pBM = 1647), or nonrelapse mortality (HR, 1.14; P = .13; nPBSC = 1664; nBM = 1862). Relapse was lower with the use of PBSC grafts (HR, 0.84; P = .001; nPBSC = 2663; nBM = 2769). The rates of acute GVHD (aGVHD) and chronic GVHD (cGVHD) were higher with PBSC grafts (aGVHD grade II-IV: HR, 1.67; P < .001; nPBSC = 2663; nBM = 2802; aGVHD grade III-IV: HR, 1.82; P < .001; nPBSC = 1826; nBM = 2000; cGVHD: HR, 1.46; P = .002; nPBSC = 2686; nBM = 2815). Engraftment was higher with PBSC grafts (HR, 1.27; P < .001; nPBSC = 1461; nBM = 1717). Comparing MAC and RIC, the use of MAC was associated with less relapse (HR, 0.70; P < .001; nMAC = 1929; nRIC = 2662), higher nonrelapse mortality (HR, 1.24; P = .002; nMAC = 2016; nRIC = 2790), but better PFS (HR, 0.86; P = .002; nMAC = 1929; nRIC = 2662). There were no differences between the 2 conditioning regimens in OS (HR, .95; P = .32; nMAC = 2123; nRIC = 3155), GRFS (HR, 0.97; P = .67; nMAC = 1182; nRIC = 1330), grade II-IV aGVHD (HR, 1.01; P = .81; nMAC = 2099; nRIC = 3090), or cGVHD (HR, 1.05; P = .44; nMAC=1929; nRIC = 2662). This analysis shows that the use of BM grafts is associated with comparable outcomes as seen with PBSC grafts despite a lower incidence of GVHD and a higher relapse rate. The use of MAC regimens is associated with improved PFS. These results suggest that for fit patients, MAC remains the optimal conditioning regimen in terms of mortality, and that the use of PBSC grafts may further decrease relapse risk and hasten engraftment, provided that further strategies can be incorporated to decrease GVHD. Prospective comparisons are awaited.