High Melphalan Exposure Increases the Risk of Graft-Versus-Host Disease in Pediatric Patients Undergoing Alpha-Beta T-Cell Depleted Haploidentical Transplantation.

Background: Melphalan is often used as the backbone agent for conditioning prior to A/B-T-cell depleted (A/B-TCD) hematopoietic cell transplant (HCT) due to lower rates of organ toxicity compared to busulfan or total-body irradiation, albeit with significant mucosal injury. Traditional dosing based on body-surface-area (BSA) may result in non-optimal melphalan exposure among certain patient subsets.

Objectives: As mucosal injury is linked to initiation of alloreactivity, we hypothesized that high exposure of melphalan predicted via a pharmacokinetic (PK) model would be associated with an increased risk of acute graft-versus-host disease (aGVHD).

Study design: We performed an analysis of 85 patients who underwent A/B-TCD haploidentical HCT on 2 prospective trials using melphalan-based conditioning for treatment of malignancy at 3 centers from 2015 to 2024. Most patients (61.2%) received a total dose of melphalan at 140 mg/m² using actual body weight; others received a dose adjusted for obesity or age <2 years. We analyzed outcomes based on whether melphalan exposure was above or below the median exposure for the group.

Results: The 100-day cumulative incidences of engraftment syndrome (ES), grade II-IV aGVHD, and grade III-IV aGVHD were 34.2%, 24.8%, and 17.1%, respectively. The 3-year cumulative incidence of chronic GVHD (cGVHD), non-relapse mortality (NRM), and relapse were 17.5%, 8.7%, 21.8%, respectively. The 3-year cumulative incidence of disease-free survival (DFS) and severe GVHD-relapse-free survival (GRFS) were 71.4% and 55.6%, respectively. ES was significantly associated with the subsequent development of aGVHD, both grade II-IV (41.4% vs. 17.3% in those with and without ES, P = .01) and grade III-IV (34.5% vs. 8.5% in those with and without ES; P = .003). Chronic GVHD occurred at significantly higher rates in patients with prior Grade II-IV (66.7% vs. 0% for Grade 0-I; P < .001) and Grade III-IV aGVHD (75% vs. 4.3% for Grade 0-II; P < .001). Compared to non-obese patients, the PK model predicted lower melphalan exposure (P = .02) in obese patients where adjusted ideal body weight was utilized, suggesting overcorrection of the dose. There was no impact of melphalan exposure on immunologic rejection. The median melphalan exposure was 6.81 mg*hr/L (range, 4.4-8.8). Compared to a melphalan exposure ≤6.8 mg*hr/L, a melphalan exposure >6.8 mg*hr/L was associated with a higher incidence of ES (48.8% vs. 19.1%; P = .005), grade II-IV aGVHD (39.3% vs. 10.1%; P = .002), and grade III-IV aGVHD (31.5% vs. 2.5%; P < .001). The 3-year incidence of cGVHD was 27.2% in those with high predicted melphalan exposure compared to 7.4% for low exposure (P = .03); with no difference in 3-year NRM incidence (9.2% vs. 7.7%; P = .82) or 3-year relapse incidence (16.8% vs. 27.6%; P = .31) for high compared to low exposure. However, GRFS was significantly worse in patients with high exposure (46.4%) vs. low exposure (64.6%; P = .02).

Conclusions: High melphalan exposure predicted by a validated population PK model is associated with an increased likelihood of developing ES and subsequently acute and chronic GVHD. Given that a substantial number of patients already require adjustment of standard BSA-based dosing for young age or obesity, a prospective trial of model-based dosing to individualize melphalan exposure is warranted to confirm these results.