Lesson learned in pediatric haploidentical transplantation in a low-resource environment: delivering melphalan IV and using low dose radiation reduce graft failure

ABSTRACT

Objectives: Primary graft failure (pGF) after hematopoietic stem-cell transplant is associated with considerable morbidity and mortality. The incidence in haplo-HSCT has been reported to be between 0% and 30%. In 2018, we identified a pGF incidence of 35% in our pediatric haplo-HSCT recipients with hematologic malignancies, which motivated us to enact changes to the conditioning regimen.

Methods: We performed a single-center prospective, pre–post study of consecutive patients under 16 years with hematologic malignancies, from January 2015 to December 2022 who received a haplo-HSCT. Twenty-six pediatric patients received a haplo-HSCT before September 2018 (G1) and 36 patients after (G2). The main conditioning regimen for G1 was myeloablative with Flu/Cy/Bu, and for G2 the main regimen was reduced intensity Flu/Cy/Mel/TBI2.

Results: Nine patients (35%) in G1 had primary graft failure, while in G2 there were no patients with pGF. The median follow-up for G1 was 15.9 months, and for G2 was 24.8 months, with an estimated overall survival at 12 months of 63% (95% CI 47–76) versus 85% (95% CI 73–93), and at 24 months of 47% (95% CI 31–64) versus 70% (95% CI 54–82) respectively (p = .007).

Conclusion: After September 2018 conditioning regimen modifications were implemented with the objective of reducing primary failure, consisting mainly of switching from busulfan to melphalan as the alkylating agent of choice, and adding, when clinically possible TBI. Primary failure has been significantly reduced in our institution since then.

KEYWORDS:

• Haploidentical
• Primary graft failure
• Pediatric
• Conditioning regimen
• Melphalan
• TBI
• Reduced intensity
• Low-and-middle-income countries

Introduction

Haploidentical hematopoietic cell transplants (Haplo-HSCT) in low-and-middle-income countries (LMICs) represent an opportunity for patients who require an HSCT, but who do not have an HLA-compatible, familiar or unrelated donor, a situation that occurs in approximately 75% of cases. Thus this obstacle to performing the procedure has been virtually eliminated since approximately 95% of patients have a potential haploidentical donor [1, 2].

The fundamental clinical obstacle to haplo-HSCT arises from intense, bi-directional responses from T cells to allogeneic HLA molecules resulting in graft versus host disease (GVHD), graft failure, and non-relapse related mortality (NRM) [3]

Primary graft failure (pGF) is associated with considerable morbidity and mortality, ranging from 91% to 100% if a second HSCT is not performed [4]. Multiple factors intervene to ensure a successful engraftment; they have been described as modifiable (stem-cell source, conditioning regimen, anti-HLA antibodies) and non-modifiable (microenvironment, disease, HLA disparity). The incidence of graft failure in the setting of post-transplant cyclophosphamide (PTCy) based haplo-HSCT, using either a myeloablative conditioning (MAC) or nonmyeloablative (NMA) conditioning regimen, ranges from 0% to 30% [5]. Donor-specific anti-human leukocyte antigen antibodies (DSAs) have been described as one of the key factors in HSCT which can increase graft failure rates up to 32% [6].

In 2018, we identified a pGF incidence of 35% in pediatric haplo-HSCT recipients with malignancies treated in our institution, which motivated us to enact changes to the conditioning regimen. The main objective of our study was to determine the pre–post difference in pGF and after the year 2018, when conditioning regimen modifications aimed at intensifying conditioning regimen immunosuppression were implemented in our institution.

Patients and methods

Study design and participants

We performed a single-center prospective, pre–post study of consecutive patients under 16 years with hematologic malignancies, from January 2015 to December 2022 who received a haplo-HSCT. The study was approved by the Institutional Review Board and was performed according to the Declaration of Helsinki. Before September 2018, pediatric patients with malignancies underwent peripheral blood stem cell (PBSC) related haplo-HSCT with busulfan based conditioning and PTCy and did not have routine donor specific antibody testing performed in all cases Group 1. The primary conditioning regimen in this period was fludarabine 25 mg/m² plus cyclophosphamide at 350 mg/m² were administered intravenously from days −7 to −5. Busulfan dosage was calculated based on age and weigh from 6 to 12 mg/kg, without trough level measurements due to lack of availability, and was administered from day −4 to day −2 (FluCyBu). During this time, a graft failure rate of 35% was identified and an intervention designed to reduce it and improve patient outcomes.

Intervention

After September 2018, all patients had DSA testing and received melphalan plus 2Gy of total body irradiation (TBI) (Group 2). For Group 2, the regimen fludarabine and cyclophosphamide at the same doses on days −5 to −3, melphalan was administered at a dose between 140 and 200 mg/m² on day −2, and total body irradiation (TBI) at 2 Gy on day −1. Also, after 2018 anti-HLA DSA determination was routinely performed.

Donor eligibility criteria and transplant procedure

The donor selection process involved initial typing of potential family members at the HLA-A, HLA-B, and HLA-DRB1 loci at an intermediate resolution level. Family members identified as potential donors underwent further typing at the HLA-C locus at an intermediate resolution level. Confirmation studies are then conducted on both recipients and donors, involving high-resolution typing. Age, gender disparity, blood group, CMV serostatus, and donor comorbidities were also considered.

Anti-HLA DSA determination was routinely performed after 2018 in patients in the month prior to HSCT; its presence was studied using LSA kit (LIFECODES Single Antigen, Immucor, Norcross, GA) following the manufacturer's instructions. Briefly, 10 μL of serum was incubated with 40 μL beads for 30 minutes. After a wash, the diluted antihuman IgG PE conjugate was added to the beads. After a final 30-minute incubation, wash buffer was added to the wells, the plate was placed in the Luminex instrument (Luminex Corporation, Texas, USA), The signal intensity of each bead was compared to the intensity of the negative internal control bead to score the anti-HLA DSA test result as positive or negative.

All donors received granulocyte colony stimulating factor (G-CSF) at a dose of 10 μg/kg/d for 4 consecutive days. Stem cell collection was performed between 12 and 24 hours after the last dose of G-CSF. Peripheral blood collection was performed, and a fresh graft was infused on day 0 for all patients.

Transplantation and graft-versus host-disease prophylaxis

The conditioning and stem-cell infusion procedures are carried out on an outpatient basis. Graft-versus-host disease prophylaxis consisted of cyclophosphamide at 50 mg/kg/day with Mesna on days +3 and +4, mofetil mycophenolate (MMF) was given at 15 mg/kg/dose every 8 hours until day +35, starting at day +5, and a calcineurin inhibitor was administered for at least 100 days after infusion. Patients are admitted on day +2 for hydration before the administration of PtCy.

They received antimicrobial prophylaxis for bacterial, fungal, herpetic, and Pneumocystis jiroveci infection according to standard institutional practices. Quantitative polymerase chain reaction or serologic assay for cytomegalovirus was performed when the attending physician deemed necessary.

Discharge is considered once evidence of the first increase in neutrophils >0.5 × 10⁹/L is observed, followed by continuous monitoring and evaluation. If the patient showed no sign of acute GVHD, immunosuppression was tapered every week beginning at day +100 after transplantation. Routine donor chimerism analysis of a bone marrow aspirate was performed on day +30.

Outcomes and definitions

The primary endpoint of this study was pGF, defined as absolute neutrophil count (ANC) <0.5 × 10⁹/L or as <5% donor chimerism in peripheral blood or bone marrow by day +30 without detected bone marrow disease. Secondary graft failure was defined as loss of donor chimerism to less than 5% after confirmed engraftment, in the absence of progressive malignancy affecting the marrow [7]. Secondary outcomes included the cumulative incidence of neutrophil and platelet recovery, overall survival, event-free survival, graft-versus-host relapse-free survival and non-relapse mortality. Neutrophil recovery time was defined as the number of days from haplo-HSCT to the first of 3 consecutive days with an absolute neutrophil count above 0.5 × 10⁹/L. Platelet recovery time was defined as platelet count greater than 20 × 10⁹/L without platelet transfusion in the preceding 7 days. Overall survival (OS) was the time from HCST to death from any cause or last follow-up. Progression-free survival (PFS) was the time from haplo-HSCT to disease relapse, progression, or death from any cause. Non-relapse mortality (NRM) was defined as death without disease relapse. Diagnosis and severity of acute and chronic GVHD were diagnosed using standard criteria [8,9]. Graft-versus-host relapse-free survival was defined as the duration from transplantation until death, relapse, development of grade III–IV acute GVHD, or development of chronic GVHD that required systemic treatment.

Other known factors associated with pGF were assessed in both groups: DSA, gender disparity, donor age, and types of conditioning. Conditioning regimens were considered reduced intensity if patients received busulfan <9 mg/kg and melphalan at 140 mg/m² [7].

Statistical analysis

A sample size of 23 patients per group was estimated to detect a 25% difference in the pGF proportion from the historical baseline of 36% with 80% power and a 2-sided p-value of .05, by using the formula n = (Z_α/2+ Z_β)² * (p₁(1–p₁) + p₂(1–p₂))/(p₁–p₂)². Categorical variables were compared between G1 and G2 were compared using the chi² test or Fisher's exact test. Continuous variables that were normally distributed were compared with Student's t test or with the Mann–Whitney U-test for non-normally distributed variables. Survival probabilities of OS, EFS, and GRFS were estimated with the Kaplan–Meier method and compared with the log rank test using the SPSS software. The cumulative incidence of NRM and neutrophil and platelet engraftment were calculated by considering death without relapse or recovery as competing risks and compared using Gray's test using R software and the tidycmprsk statistical package.

Results

Patient and donor characteristics

Patient characteristics and demographics are summarized in Table 1. Twenty-six patients received haplo-HSCT before September 2018 (G1) and 36 patients after (G2). After 2018, one patient with elevated DSA who experienced failure of the desensitization protocol and disease relapse could not undergo transplantation. The median age was 9 years in both groups. Ninety percent of patients had acute leukemia as the primary diagnosis, and more than two-thirds underwent transplantation with a high Disease Risk Index (DRI) in second remission. The father was the donor in 50% of the transplants, with a median age of 33 (range: 4–51 years).

Table 1. Patient and donor characteristics

Variable		2015–2018	%	2019–2022	%	P value
		N = 26		N = 36
Recipient Age in years	Median (range)	9	(3-15)	9	(1-16)	.05
Sex	Men	17	65	24	67	.91
	Women	9	35	12	33
Diagnosis	ALL	18	69	20	55	.42
	AML	7	27	11	30
	CML	1	4	1	3
	LCH			1	3
	HL			2	6
	NHL			1	3
DRI	Low	1	4	1	3	0.46
	Intermediate	5	19	11	28
	High	20	77	24	69
Status at HCT	CR1	3	12	12	34	.03
	CR2	16	62	20	56
	CR3>	6	22	1	3
	RC			1	3
	Advanced	1	4	2	3
Conditioning	MAC	20	77	12	33	.001
	RIC	6	23	24	67
Conditioning regimen	Flu/CY/Bu	23	88			.001
	Flu/CY/Mel	2	8	5	14
	Flu/CY/ATG	1	4
	Flu/CY/Mel/TBI			31	86
Donor	Father	12	46	21	58	.35
	Mother	11	42	9	25
	Brother	3	12	6	17
Donor Age in years	Median (range)	35	(10–51)	31	(4-51)	.12
CMV status	D+/R+	14	54	19	52	.35
	D-/R-	1	4	5	14
	D-/R+	4	15	2	6
	D+/R-	7	27	10	28
CD34 cells	× 10^6/kg	14	(3.9–15)	10	(5–15)	.09

Abbreviations: HSCT: Allogeneic Stem Cell Transplantation; ALL: Acute Lymphoblastic Leukemia; AML: Acute Myeloid Leukemia; CML: Chronic Myeloid Leukemia; LCH: Langerhans cell histiocytosis; HL: Hodgkin´s disease; NHL: Non-Hodgkin Lymphoma; ; MAC: myeloablative conditioning; RIC: reduced intensity conditioning; Flu: fludarabine; CY: cyclophosphamide; Bu: busulfan; Mel: melphalan; ATG: antithymocyte globulin; TBI: total body irradiation; CMV: Cytomegalovirus.

Engraftment and characteristics of the second HSCT

In Group 1 (G1), an incidence of failure was observed at 35% (N = 9), in contrast to G2 where it was 0%. Additionally, of the 17 patients that engrafted, neutrophil recovery occurred more slowly in Group 1, with a median of 19 days (range: 15–36 days), as opposed to Group 2 with 15 days (range: 15–20), (p = <.0001). Platelet recovery was similar, with no significant difference observed Table 2. Cumulative incidence of ANC recovery was 3.8% at 15 days in G1 (95% CI 0.26–17%) versus 69% in G2 (95% CI 50–81), 62% at 30 days (95% CI 40–78%) and 65% at 100 days in G1 (95% CI 43–81) versus 100% in G2 at 30 and 100 days (95% CI non-calculable) (p < .001) (Figure 1A). Cumulative incidence of Platelet recovery was 27% at 15 days in G1 (95% CI 12–45%) versus 47% in G2 (95% CI 30–63), 65% at 30 days (95% CI 43–81%) and 65% at 100 days in G1 (95% CI 43–81) versus 100% in G2 at 30 and 100 days (95% CI non-calculable) (p < .001) (Figure 1 B).

Figure 1. Cumulative incidence of neutrophil (A) and platelet (B) engraftment.

Table 2. Post-HSCT dynamics.

Variable		2015–2018	%	2019–2022	%	P value
		N = 26		N = 36
ANC recovery	Days	19	(15–36)	15	(12–20)	<.001
PLT recovery	Days	18	(11–25)	16	(12–30)	.33
pGF		9	35	0	0	<.001
Acute GVHD grades	II–IV	5	29	9	25	.75
Acute GVHD grades	III–IV	2	12	5	14	.6
Chronic GVHD	M–S	7	44	9	26	.21
Relapse events		11	42	9	25	.12

Abbreviations: ANC: absolute neutrophil count; PLT: platelet count; pGF: Primary graft failure, GVHD, Graft Versus Host Disease; M: moderate, S: severe.

Anti-HLA antibodies were analyzed in 10/26 (38%) in G1 and 100% of G2 patients and detected in one patient from G2 (3%) without an alternative donor available. He received a desensitization protocol with Rituximab and Bortezomib, successfully reducing the titers and achieving an adequate engraftment of neutrophils and platelets on days 17 and 26.

Among the nine patients who experienced graft failure in G1, n = 2 did not have DSA determination, but they achieved engraftment after a second transplant, one had the same donor and the other one had a different haploidentical sibling donor. Of the nine patients (35%) who experienced graft failure, only six underwent a second HSCT with a median time-to-transplant of 70 days (range: 50–280). The patient with the longest time, 280 days between the first and second transplant, showed autologous recovery after the first transplant, enabling the search for a second donor and received myeloablative conditioning; Currently, with a 5-year follow-up, the patient is GVHD-free, and in remission. Only four out of six patients who received a second transplants achieved engraftment; the previously mentioned patient, one patient with neutrophil recovery who died due to sinusoidal obstructive syndrome before chimerism assessment, and two patients who died from acute and chronic graft-versus-host disease. Of the two patients who did not engraft after the second transplant, one received a third HSCT after changing donor and conditioning, achieving engraftment, and the second patient died due to disease progression. The three patients who did not undergo a second transplant died. All patients in G2 achieved complete donor chimerism by day 30 of haploidentical hematopoietic stem cell transplantation (haplo-HSCT).

Graft-versus-host disease and HSCT

Grades II and IV aGVHD occurred in 5 of 17 evaluable patients (29%) in G1 and 9 of 36 evaluable patients (25%) in G2, and grades III and IV in two patients (12%) in G1 and 5 patients (14%) in G2 with no significant difference observed (p = 0.6). There were also no significant differences in the incidence of moderate or severe chronic GVHD occurred in 7 patients (44%) of G1 and 9 patients (26%) of group 2 (p = 0.2) (Table 2).

NRM, relapse, and survival

The median follow-up for G1 is 15.9 months and for G2 is 24.8 months (Table 3), with an estimated overall survival at 12 months of 63% (95% CI 47–76) versus 85% (95% CI 73–93) and at 24 months of 47% (95% CI 31–64) versus 70% (95% CI 54–82) respectively (p = .007) (Figure 2A). EFS and GRFS were higher in the post-2018 group, as shown in Figure 2B and C.

Figure 2. Overall survival in the two groups (A), event-free survival in the two groups (B), and graft-versus-host-free disease-free relapse-free survival in the two groups (C).

Table 3. Survival outcomes.

Variable		2015–2018 N = 26	95%CI	2019–2022 N = 36	95%CI	P value
		%			%
Median follow up	Months (min–max)	15.9	(1.9–75)	24.8	(2.5–38.8)	.053
OS	1 month 3 months 6 months 12 months 24 months	96 87 76 63 47	84–99 72–94 61–87 47–76 31–64	100 97 94 85 70	95–100 90–100 84–98 73–93 54–82	.007
EFS	12 months 24 months	60 44	44–74 28–61	80 64	67–89 48–77	.012
GRFS	12 months	33	22–52	50	36–64	.07
NRM	3 months	16	4.8–33	2.8	0.2–13	.4
	6 months	16	4.8–33	2.8	0.2–13
	12 months	16	4.8–33	5.6	0.97–17
	24 months	20	7.1–38	9.2	2.2–33

Abbreviations: OS: Overall Survival, EFS: Event-free survival, GRFS: Graft-versus-host-free disease-free relapse-free survival, NRM: non-relapse-mortality.

NRM was 16% at 1 year in G1 (95% CI 4.8–33%) versus 5.6% (95% CI 0.97–17%) in G2, NRM was 20% at 2 years (95% CI 7.1–38%) in G1 versus 9.2% (95% CI 2.2–23%) in G2 (p = 0.4) (Figure 3).

Figure 3. Non-relapse mortality in the two groups.

Discussion

Haploidentical HSCT in LMICs is not an easy task. Technology and access to transplants are not evenly distributed worldwide. In Latin America, over 600 million people reside, but the transplant frequency is 20–40 times lower compared to that in Europe and North America [2]. The emergence of PtCy has made this option more available in our region. We have adjusted our HSCT program to reduce its cost, minimize failure and transplant related mortality. In this study, we analyzed one of the modifiable risk factors for primary failure, the conditioning regimen, in a prospective cohort. This change from busulfan to melphalan and total body irradiation was made after detecting a high incidence of primary failure with the former.

In pediatric reports of patients undergoing Haplo-HSCT with PTCy, the incidence of pGF is reported to be 8–15% [10,11]. Due to the alarming graft failure rate of 35%, and a 12-month OS of 63% observed in Group 1, changes to the conditioning protocol were decided on 2018. The most relevant challenges for the program were the lack of infrastructure and personnel for measuring Bu levels, access to DSA, as well as a shortage of hospital beds. Since Busulfan (Bu) was key component in the conditioning regimen used in 88% of G1 patients, and Xinying et al. demonstrated that optimal systemic exposure (expressed as the area under the concentration-time curve [AUC]) can determine the outcome of hematopoietic stem cell transplantation (HSCT), with values <900 μM × min showing a significant increase in the incidence of graft failure, our team decided that a modification to the alkylating agent of the conditioning regimen was needed [12].

We required a more stable alkylating agent that was not limited by the need of level measurements that was accessible in the region and yielded similar results in OS: melphalan [13]. This was combined with low-dose radiation, a widely used method to reduce primary failure in both benign and malignant diseases [14,15]. Additionally, other strategies were implemented by the program, such as outpatient follow-up, and a previously published reduced-intensity conditioning, contributing to improved outcomes [16].

The other modifiable factor was the determination of DSA. Meng et al. describe the prevalence of DSA in 486 pediatric patients on the waiting list for haplo-HSCT with malignant hematologic diagnoses and aplastic anemia. The distribution was variable, with DSA against HLA-A being more frequent (7.8%) with a median MFI of 1678, and DSA against HLA-DP (2.9%) with a median MFI of 1372. The diagnosis was identified as an independent risk factor for HLA sensitization, with ALL patients having a lower incidence of anti-HLA class II antibodies and antibodies against specific loci [17]. The correlation between the existence of DSA before mismatched allogeneic HSCT and its adverse effects on primary engraftment and survival is now firmly established. Specifically, the presence of DSA against DPB1 in a cohort of 591 patients increased the incidence of graft failure to 37%, compared to 3% in the antibody-negative group [18]. Since 2010, Spellman et al. found that 24% of a group of 37 patients with primary graft failure had DSA against HLA-A, -B, or –DP, compared to only 1% of 78 successful transplants in the control group [19].

In our center, where access to unrelated HSCT is very limited, and virtually 100% of pediatric patients have an available haploidentical family donor, it is essential to ensure that all patients are evaluated for the presence of DSA, and that this factor is considered in donor selection. As mentioned in our results, only 34% of patients prior to 2018 had undergone this assessment, representing a potential cause for the high percentage of primary graft failures observed. Of the 9 patients with primary failure, 2 patients had negative pre-transplant DSA evaluation, and 5 patients had post-failure DSA screening, at day +30, finding only one patient with positive DSA. The two patients without DSA measurement and primary failure had a second haplo-HSCT performed, with successful engraftment, suggesting that in our group improvement in the primary graft failure rate was mainly motivated by conditioning regimen modification. After 2018, one patient had elevated DSA, and desensitization was achieved with a bortezomib protocol, leading to a transplant with successful engraftment.

In group 1, six patients received a second haplo-HSCT. The 1-year overall survival following a retransplantation for pGF is 11%, without relapse. Factors such as donor change, and conditioning contribute to success but come with a high cost of conditioning-related toxicity and complications [20]. One of the advantages for patients who underwent retransplantation was the immediate access to another donor. Given that pediatric patients often have parents as readily available donors, this facilitates the prompt execution of the second transplant. In our group, this occurred with a median of 70 days between the first and second transplants, resulting in a 50% overall survival at 12 months.

The incidence of acute GVHD grades III and IV was similar in both groups, with 2% in G1 and 5% in G2. In contrast, Srinivasan et al. reported incidences in pediatric patients undergoing Haplo, MSD, and MUD HSCT of 11.5% (95% CI, 0–23%), 3.2% (95% CI, 0–9.2%), and 8.7% (95% CI, 0.2–16.5%), respectively. While Fierro et al. reported a 0% incidence in 32 patients with acute leukemia who underwent haplo-BMT with PTCy, prompting us to explore the possibility of using bone marrow as a source, which is currently not logistically feasible in our center. Our incidence of moderate to severe cGVHD was higher than that reported by other groups, ranging from 9% to 15%. This could be explained by the use of peripheral blood as a source [11].

While the implemented strategies to reduce primary graft failures proved highly effective, the success of a transplant program, especially in pediatric patients, is determined by the long-term ability to keep our patients alive, in disease remission, and free from transplant-related complications that may interfere with their quality of life. Currently, one of the best ways to assess this is through GRFS, which, although increased from 35% to 50% at 12 months, is lower than that reported by other pediatric groups undergoing haplo-HSCT, ranging from 61% to 65%.

Progress in transplants requires interventions that provide feasible solutions with the available resources, and ideally, do not limit access to transplantation. The practice of prospective data reporting in CIBMTR, coupled with annual analyses of program results and strategic interventions addressing identified areas of enhancement is essential for continuous improvement of HSCT program results. A melphalan/2Gy TBI-based conditioning regimen could be a practical solution for LMICs that don't have Bu level measurements, and wish to implement an haplo-HSCT program with low graft failure rates.

Authors contributions

VJ: Conceived and designed the project; VJ and JC: Performed data collection, VJ, AG, and EG: Analyzed and interpreted the data; VJ Wrote the manuscript; AG, JC, EG, DG, and OG: Edited the manuscript.

Acknowledgments

The authors would like to thank the patients, their families, and their caregivers.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

There are no relevant financial relationships to disclose.