A retrospective comparative study of haploidentical versus HLA-matched allogeneic hematopoietic stem cell transplantation in patients with relapsed/refractory acute myeloid leukemia

Patients with relapsed/refractory (R/R) acute myeloid leukemia (AML) undergoing allogeneic hematopoietic stem cell transplantation (allo-HSCT) may receive transplants from either human-leukocyte-antigen identical (HLA-id) sibling donors, unrelated donors (UDs) or HLA-mismatched (Haplo) family donors. In the R/R AML patient population, allo-HSCT may allow patients to achieve a complete response (CR), with long-term control achieved through other immunomodulatory strategies. Since HLA-id sibling donors are only available in approximately 30% of cases; and unrelated donors (UD) take time to locate, transplants from Haplo donors have increased.¹

Giorgia Battipaglia of Hôpital Saint Antoine, Paris, and Federico II University, Italy, and colleagues conducted a retrospective study comparing the outcomes of a large cohort of patients with R/R AML receiving HLA-id or Haplo HSCTs, utilizing the European Society of Blood and Marrow Transplantation (EBMT) Acute Leukemia Working Party (ALWP) registry.²

Patient characteristics and study design

All results are given as Haplo versus HLA-id

• Adult patients (≥18 years old) with R/R AML undergoing first allo-HSCT from:
• HLA-id sibling donor (N = 1654)
• Haplo family donor with ≥2 antigen mismatches (N = 389)
• Graft versus host disease (GvHD) prophylaxis in the Haplo group was:
  • Post-transplant cyclophosphamide (PTCy) (N = 278)
  • In vivo T-cell depletion (TCD) (N = 95)
  • Both PTCy and TCD (N = 16)
  • Majority of TCD: rabbit anti T-lymphocyte immune globulin (ATG)
• Median age at HSCT: 52 (18–75) years
• Median follow-up: 22 vs 16 months (p = 0.12)
• Peripheral blood stem cell source (PBSC) as the main source: 53% vs 92% (P < 0.01)
• Bone marrow (BM) grafts as the main source: 47% vs 8% (P < 0.01)
• Refractory disease at the time of allo-HSCT: 42% vs 56% (P < 0.01)
• Primary end-point: overall survival (OS) by donor type
• Secondary end-points: leukemia-free survival (LFS), non-relapse mortality (NRM), neutrophil engraftment, rates of acute GvHD (aGvHD) and chronic GvHD (cGvHD), refined GvHD/relapse-free survival (GRFS) and performance status (PS)
• Cumulative incidence (CI) estimator was used to calculate neutrophil engraftment, aGvHD, cGvHD, relapse incidence (RI) and NRM to allow for competing risks

Engraftment and GvHD:

• Neutrophil engraftment CI: 83% vs 93% (P < 0.01)
  • Rate of engraftment lower in Haplo donors: hazard ratio (HR): 0.75 (CI: 0.65–0.86, P < 0.01)
• Time to engraftment: 18 vs 15 days (P < 0.01)
• Grade II–IV aGvHD CI: 28% vs 27% (P = 0.37)
• cGvHD CI: 27% vs 42% (P < 0.01)
• Severe cGVHD CI: 10% vs 12%, (P < 0.01)
• No difference in CI of aGvHD or cGvHD by donor groups
• Factors associated with a lower risk of grade II–IV GvHD:
  • Using BM as stem cell source
  • Using TCD

Relapse incidence and non-relapse mortality

• NRM CI (2-year follow-up): 31% vs 22% (P < 0.01)
  • RI: 50% vs 51% (P = 0.60)
• Both confirmed by multivariate analysis
  • NRM: HR: 1.39 (95% CI, 1.07–1.81, P = 0.02)
  • RI: HR: 1.06 (95% CI, 0.86–1.30, P = 0.58)
• Main causes of mortality:
  • Disease progression: 48% vs 62% (P < 0.01)
  • Infections: 41% vs 25% (P < 0.01)
• CR achieved at day 100 after allo-HSCT: 78% vs 76%
• Patients in relapse at time of allo-HSCT had higher risk of progression vs patients with refractory disease: HR: 1.33 (95% CI, 1.16–1.53, P < 0.01)
• Factors associated with NRM:
  • Ten-year age increase
  • Poor PS
  • Use of a reduced intensity conditioning regimen

OS, LFS and GRFS

• Median follow-up: 17.5 (1–119) months
• Two-year LFS: 19% vs 27% (P < 0.01)
• OS: 25% vs 32% (P < 0.01)
• GRFS: 18% vs 26% (P < 0.01)
• Multivariate analysis, Haplo:
  • Inferior OS: HR: 1.23 (95% CI, 1.03–1.47, P = 0.02)
  • Inferior LFS: HR: 1.19 (95% CI, 1.00–1.41, P = 0.05)
  • Inferior GRFS: HR: 1.22 (95% CI, 1.03–1.44, P = 0.02)
• Patients with relapsing/progressive AML had a lower OS, LFS and GRFS
• LFS higher in patients that did not receive in vivo TCD (P = 0.05)

Inferior outcomes were observed with the Haplo-HSCT group in terms of LFS, OS, NRM compared with HLA-id sibling transplants and the engraftment rate was higher in the HLA-id group. The primary cause of mortality in both groups was disease progression, followed by infections, which were more frequent in the Haplo recipients. Another factor shown to contribute to poorer outcomes was transplanting patients with relapsing disease when compared to those with refractory AML.

There was no exhibited difference in the rates of aGvHD or cGvHD when comparing donor type; however using ATG as GvHD prophylaxis conferred a poorer outcome. This is compared to using PTCy as prophylaxis to GvHD in Haplo HSCTs, which showed no difference when compared to HLA-id HSCTs indicating that this regimen may provide similar outcomes to HLA-id sibling transplants.

The time-critical element of finding a suitable donor for patients with R/R AML means that utilizing a readily available Haplo donor should be considered where a HLA-id donor is not available, especially considering the difficulty of locating UDs in certain ethnicities. This has only become feasible in recent years through improvements in conditioning regimens, immunosuppressive strategies and supportive care. However, whilst HSCTs from Haplo family-matched donors represent a viable alternative option when HLA-id sibling matches are unavailable, HLA-matched siblings are the ideal donors.