Factors Associated with Platelet Engraftment Following Allogeneic Hematopoietic Cell Transplantation

Introduction. Delayed platelet engraftment (PE) is a common complication of allogeneic hematopoietic cell transplantation (HCT) and has been shown to be associated with increased non-relapse mortality (NRM) and poor overall survival (OS). Factors such as conditioning intensity, donor type, stem cell source, recipient CMV-seropositivity, and presence of grade II-IV acute graft-versus-host disease (GVHD) have been historically reported to impact platelet recovery time (Ramirez et al., 2011), though recent data are lacking. Thrombopoietin-receptor agonists (TPO-RAs) have been increasingly studied and used in the post-HCT setting for poor graft function (Mahat et al., 2019). We sought to identify pre-HCT risk factors associated with post-HCT PE in a contemporary cohort to identify a higher-risk population in which the pre-emptive use of TPO-RAs may be investigated in the future.

Methods. A retrospective analysis was performed on patients age ≥18 years who underwent first allogeneic HCT from 2012-2021. Patients receiving umbilical cord graft and whose platelet count never nadired <20 k/mL were excluded. PE was defined as a count ≥20 k/mL for 7 days without transfusion. Primary objectives were to describe median time to PE and identify pre-transplant risk factors for engraftment. Secondary objectives were to determine impact of PE on NRM and OS.

Results. 628 patients met inclusion criteria. Median age was 58 (range, 18-76). 56% were Male and 91% were White. HCT-comorbidity index (CI) was high (≥3) in 54% of patients. Disease groups included AML (44%), MDS (24%), ALL (14%), CML/MPN/myelofibrosis (13%), and non-malignant diseases (4%). Disease risk index was intermediate/high in 80% of patients. Over half (52%) of patients received reduced intensity conditioning (RIC) and HCT from an unrelated donor (55%), and most (60%) received a peripheral blood stem cell graft. A third of patients (34%) received post-transplant cyclophosphamide (PTCy)-based GVHD prophylaxis. Median platelet count at the time of HCT admission was 123 k/mL (range, 2-1056).

Median time to PE was 26 days (IQR 24-27). Delayed PE within 90 days of HCT was significantly associated (p<0.05 for all) with several factors: pre-HCT platelet count (30-day without-PE: 44% in patients with pre-HCT platelet count ≤65k vs 31% in >65k); Karnofsky performance status (KPS) (45% in KPS <90 vs 30% in 90-100); renal dysfunction (54% in eGFR<70 vs 34% in 70-89 vs 31% in ≥90); age (35% in <60 vs 32% in ≥60); disease (40% in MDS and 39% in non-malignant disease vs 30% in AML), donor type (60% in haploidentical vs 31% in unrelated vs 20% in related); and GVHD prophylaxis (57% in PTCy vs 34% in calcineurin inhibitor [CNI]/MTX vs 11% in CNI/MMF). In multivariate analysis, lower KPS (HR 0.75, P=0.005), lower pre-HCT platelet count (HR 0.71, P=0.002), renal function eGFR <70 versus ≥90 (HR 0.62, P=0.007), age <60 (HR 0.74, P=0.005), MDS vs AML (HR 0.76, P=0.017), and donor source (haploidentical/bone marrow [BM] [HR 0.45, p<0.001], haploidentical/peripheral blood [HR 0.48, p<0.001], and unrelated/BM [HR 0.71, p=0.024] vs related/BM) were associated with delayed PE by day 30. Given a strong association of GVHD prophylaxis and donor source, GVHD prophylaxis was not included in this multivariate model. Delayed PE was associated with higher NRM (HR 4.5, p<0.0001) and worse OS (HR 3.1, p<0.0001).

Discussion. We identified several pre-HCT factors associated with PE and demonstrated that delayed PE was associated with worse survival. Identification of patients at high risk for delayed PE may lead to the ability to prospectively study earlier interventions post-HCT to enhance platelet recovery and improve transplant outcomes.