Improving Outcomes with Haploidentical Stem Cell Transplantation [HaploSCT] in Children Using Post-transplant Cyclophosphamide: a Single Center Experience

Abstract

Haplo-identical stem cell transplant using post-transplant cyclophosphamide is increasingly being used in children without a matched sibling donor. Between 2010 and June 2021, 127 children underwent 138 transplants with a median age of 7.1 years for malignant and non-malignant disorders. Conditioning regimens included both myeloablative and reduced intensity regimens with peripheral blood stem cells as the main graft source. Engraftment occurred in 113 [81.9%] at a median of 16 days [range: 10–32] with primary graft failure in 10.2%. Cumulative incidence of grade II–IV acute graft versus host disease (GVHD) was 49.5% and chronic GVHD in 40.7%. Majority [92.7%] had at least one infection with 31% incidence of bacterial infection, 76% incidence of viral and 16% incidence of fungal infection. The 2-year overall survival (OS) is 54.9 ± 4.6% with a lower survival among young children aged 0–5 years [28.2 ± 6.4%] compared to 5–10 years [71.3 ± 6.8%] and 11–15 years [55.7 ± 8.8%] [p = 0.032]. 2-year OS has gradually improved from 25.0 ± 2.1% for 2010–2013 to 47.5 ± 6.2% for 2014–2017 and 67.1 ± 6.6% for 2018–2021 [p = 0.049]. On multivariate analysis, bacterial infection [p = 0.017], invasive fungal disease [p = 0.002] and graft failure [p = 0.029] negatively impacted overall survival. Haplo-identical SCT with post-transplant cyclophosphamide is a reasonable option for children who do not have a matched sibling donor. Strategies to reduce graft failure, infection related mortality and GVHD needs to be explored.

Introduction

Hematopoietic stem cell transplantation [SCT] is a curative option for children with both malignant and non-malignant disorders but its widespread use is limited by the availability of a fully matched sibling donor [MSD] in only 30–40%. Population based studies have suggested that the chance of finding a 10/10 HLA identical unrelated donor [MUD] within small donor registries in India is about 14% though it could increase to 60% with larger registries [1]. Because of the difficulty in finding a MUD donor, haplo-identical transplants [HaploSCT] using post-transplant cyclophosphamide [PTCy] are increasingly being used in India similar to patterns across the world [2–5]. Though there are no randomized trials, data suggests that outcomes of T replete and T depleted transplants are similar [6]. In addition, the present outcomes of haploSCT seem to match the outcomes seen after MSD and MUD transplants [7–9]. PTCy based haplo-SCT is a very attractive option for developing countries because of the low cost involved but infective challenges remain a major problem in developing countries [10]. We describe our experience with haplo-SCT using the PTCy platform in a large cohort of children from India.

Materials and Methods

This analysis included all children [≤ 15 years] who underwent Haplo-SCT using the PTCy platform at Christian Medical College between June 2010 and June 2021. We excluded children who received Haplo-SCT after either CD34 selection or αβ T cell depletion or patients with Fanconi anaemia who received lower doses of PTCy. Donor specific antibodies [DSA] was routinely tested for all patients from 2013. An MFI value of > 1000 was considered positive and treatment with either Rituximab, Bortezomib, Iv Immunoglobulin ± plasmapheresis was given if MFI was > 5000. Donors were chosen based upon absence of DSA, blood group match and donor age. Written informed consent was obtained from all prior to Haplo-SCT. This study was approved by the Institutional Review Board of the hospital. One hundred children were included in a previous analysis which looked at combined data on adult and paediatric patients¹⁰.

Transplant: Conditioning regimens, myeloablative or reduced intensity conditioning [RIC], were used based upon disease type. Children [≥ 2 years] received a single dose of total body irradiation [TBI] (200 CGy) on day -1 while few children with aplastic anemia [AA] received 400 cGy. Nine children with ALL received total marrow lymphoid irradiation [TMLI] as part of a study. Graft source was initially unmanipulated bone marrow [BM] but from 2012, peripheral blood stem cells [PBSC] are being used aiming for CD34 dose between 6 and 10 × 10⁶ CD34/Kg. Graft-versus-host-disease (GVHD) prophylaxis consisted of PTCy [50 mg/kg/day] on Day + 3 and + 4 followed by tacrolimus [Tac] and mycophenolate mofetil [MMF] from Day + 5. MMF was stopped by Day 35 while Tac was tapered and stopped over 3–12 months depending upon type of disease and disease risk. Standard definitions of neutrophil and platelet recovery were used to define engraftment. Chimerism analysis using VNTRs was performed on day 30 post-HSCT and repeated at 3, 6 and 12 months.

Antimicrobial prophylaxis and monitoring—All patients were nursed in HEPA filtered rooms and received antimicrobial prophylaxis with fluconazole and acyclovir starting on Day + 1. Oral penicillin and co-trimoxazole-trimethoprim prophylaxis was started on count recovery. For febrile neutropenia, antibiotics were administered as per prevalent institutional guidelines. Children with persistent fever were evaluated for invasive fungal disease [IFD] using high-resolution CT scan [HRCT] of chest and serum galactomannan assay. Weekly monitoring for cytomegalovirus reactivation was performed using DNA PCR from peripheral blood and intravenous ganciclovir was started if copy numbers were > 1000 copies/ml.

Statistical analysis: Data was censored for analysis on 30th December 2022. A modified disease risk index [DRI] was used as described previously [10]. Overall survival (OS) included all patients who were alive at last follow up. For comparison of dichotomous variables, a χ² test was done while continuous variables were compared using either a student's t-test or a Mann–Whitney U-test as was deemed appropriate. The probability of OS and DFS were estimated using Kaplan–Meier method. The prognostic relevance of clinical and biological variables was studied using univariate and multivariate Cox regression analysis. For all tests, a two-sided p-value of ≤ 0.05 was considered statistically significant. Statistical analysis used the IBM SPSS 24.0 Software.

Results

Baseline characteristics: One hundred and thirty-eight transplants were performed in 127 children between June 2010 and June 2021 (Table 1). This included 90 males and 48 females with a median age of 7.1 years [range: 9 m–15 years]. Non-malignant indications were higher [61.5%] and included aplastic anemia [AA] and primary immune deficiency disorders [PID]. Malignant indications [38.5%] included acute lymphoblastic leukemia [AML] and acute myeloid leukemia [AML] followed by pediatric myelodysplastic syndromes [MDS]. Using the modified DRI, patients were classified into mainly intermediate disease [51.4%] and late disease [37.7%] while only 10.9% had early disease. A recent history of bacterial or fungal infection [< 4–6 weeks] was obtained in 19 children [13.7%]. Fifteen children [10.9%] were undergoing a second HSCT.

Table 1.

Baseline characteristics of patients undergoing haplo-identical SCT

Parameters	Patients
	(N = 138)
Median age of patient (in years)	7.1
(range)	(9 m–15)
Age distribution
< 5 years	51 (36.9%)
6–10 years	49 (35.5%)
11–15 years	38 (27.6%)
Gender N (%)
Male	90 (65.2%)
Female	48 (34.8%)
Disease indications for SCT
Malignant	53 [38.5%]
Acute myeloid leukemia	15
Acute lymphoblastic leukemia	30
Myelodysplastic syndrome	5
Non-Hodgkins lymphoma	1
Hodgkins lymphoma	1
JMML	1
Non-malignant	85 [61.5%]
Aplastic anemia	46
Primary immune deficiency [CHS, CGD, GD, HLH, Ipex, SCID, WAS]	21
Adrenoleukodystrophy	7
Sickle cell anemia	3
Diamond Blackfan anemia	4
Gauchers disease	2
Osteopetrosis	2
Disease status
Early disease [ED]	15 [10.9%]
Intermediate disease [ID]	71 [51.4%]
Late disease [LD]	52 [37.7%]
Previous Bacterial or invasive fungal infection [within 4–6 weeks prior to HSCT	19
	[13.7%]
Number having a second transplant	15
	[10.9%]

CGD chronic granulomatous disease, CHS Chediak Higashi syndrome, GD Griscelli syndrome, HLH hemophagocytic syndromes, SCID severe combined immunodeficiency, WAS Wiskott Aldrich syndrome, Ipex Ipex syndrome

Transplant: Donors were mainly parents [90.5%] followed by siblings [9.5%] with median age of 33 years [range: 9–52] (Table 2). High resolution HLA typing showed an HLA match of 5/10 in 75.4% while a match of 6/10–8/10 was seen in the rest. DSA was positive in 10.2% [n = 14] of which 5 children with an MFI > 5000 received treatment prior to HSCT. Myeloablative conditioning regimens [Fludarabine with either Busulfan or Treosulfan] were more commonly used [57.9%] while RIC regimens [Fludarabine with either Melphalan or Cyclophosphamide] were used in 42.1%. TMLI was given to 9 children with ALL as part of an institutional study protocol. PBSC was the main graft source [94.9%] used.

Table 2.

Pre-transplant characteristics of patients undergoing Haplo-identical SCT

Parameters	Patients
	(N = 138)
Median donor age [years] range	33
	[9–52]
Relationship of donor
Parent	125 [90.5%]
Father	72[52.1%]
Mother	53 [38.4%]
Sibling	13 [9.5%]
Brother	7 [5.1%]
Sister	6 [4.4%]
Degree of HLA match
05-Oct	104 [75.4%]
06-Oct	17 [12.3%]
07-Oct	14 [10.2%]
08-Oct	3 [2.1%]
Blood group match
Match	80 [57.9%]
Major mismatch	32 [23.2%]
Minor mismatch	26 [18.9%]
Donor specific antibodies [DSA]
Present	14 [10.2%]
Absent	121 [87.6%]
Not done	3 [2.2%]
Pre-transplant CMV serostatus
Recipient + /Donor + [R + /D +]	134 [97.1%]
Recipient-/Donor + [R−/D +]	4 [2.9%]
Conditioning regimens
Myeloablative	80 [57.9%]
Fludarabine + Busulfan	40
Fludarabine + Treosulfan	31
Cyclophosphamide + TMLI	9
Reduced intensity	58 [42.1%]
Fludarabine + Melphalan	37
Fludarabine + Cyclophosphamide	18
Fludarabine + Cyclophosphamide [low dose]	3
Graft source
PBSC	131 [94.9%]
BM	7 [5.1%]

Engraftment, toxicity and graft-versus-host-disease: Median cell dose infused was 10 × 10⁶ CD34/kg [range: 1.4–39]. One hundred and thirteen transplants [81.9%] achieved an ANC > 0.5 × 10⁹/L at median of 16 days [range: 10–32] (Table 3). Primary graft failure [PGF] occurred in 14 [10.2%] while 11 [7.9%] died before engraftment due to gram negative sepsis. PGF was more common with non-malignant [15.2%] compared to malignant disease [1.9%; p = 0.01] but not different between DSA positive [23%] or DSA negative [10.4%; p = 0.173] children. Sinusoidal obstruction syndrome [SOS] was seen in 10 [7.24%] children while none developed hemorrhagic cystitis due to a judicious use of hyperhydration, forced diuresis with frusemide and mesna. Overall incidence of acute GVHD [Grade II–IV] was 49.5% with a 25.6% incidence of grade III–IV acute GVHD [mostly steroid refractory]. Chronic GVHD occurred in 40.7% and was extensive in 12.4% and limited in 28.3%.

Table 3.

Transplant outcomes of patients undergoing Haplo-identical SCT

Parameters	Patients
	(N = 138)
Median cell dose infused –	10
× 106 CD34/kg /(range)	(1.4–39)
Number of patients who engrafted [%]	113 (81.9%)
Number of patients who expired prior to engraftment	11 (7.9%)
Graft failure	20 (14.4%)
Primary	14 (10.2%)
Secondary	6 (4.2%)
Median time to ANC > 0.5 × 109/L	16
[days]/(range)	(10–32)
Median time to Platelet count > 20 × 109/L	19
[days]/range	(9–65)
Cumulative incidence of grade II–IV [aGVHD] [n = 113]
Grade II–IV	56 [49.5%]
Grade III–IV	29 [25.6%]
Cumulative incidence of chronic graft versus host disease
[cGVHD] [n = 81]	33 [40.7%]
Limited	23 [28.3%]
Extensive	10 [12.4%]
Documented infections
Documented Bacterial infection [DBI]	44 [31.8%]
Gram negative bacteria	75%
Gram positive bacteria	25%
Documented viral infection [DVI]	105 [76.1%]
CMV viremia	63.70%
Invasive fungal disease [IFD]	40 [28.9%]
Proven + Probable IFD	15.90%
2-year overall survival based on disease status
Overall	54.9 ± 4.6%
Early disease [n = 15]	100%
Intermediate disease [n = 71]	66.10%
Late disease [n = 52]	32.60%
2 year 2-year overall survival based on year of HSCT
2010–2013 [n = 4]	25 ± 2.17%
2014–2017 [n = 73]	47.5 ± 6.2%
2018–2021 [n = 61]	67.1 ± 6.6%

Infections: All developed fever after HaploSCT with 128 transplants [92.7%] having one documented bacterial, viral, or fungal infection. Blood stream bacterial infections were seen during the period of neutropenia in 5 patients. None of the patients received Tocilizumab for cytokine fever. Forty-four [31.8%] had blood stream bacterial infections with mainly gram-negative pathogens [75%]. Viral infections were most common [76.1%] and included CMV [63.7%] and BK virus reactivation [36.2%] followed by respiratory viruses. IFD occurred in 28.9% with a 15.9% incidence of proven and probable IFD.

Survival: At median follow up of 19 months [range: 1–104], 79 children are alive while 15 [28.4%] had disease relapse. The median time to relapse was 7 months [range: 3–46]. Relapse rates were similar between intermediate [34.6%] and late disease [33.3%; p = 0.91] with no relapses in the early disease group. Relapses were similar between ALL [36.6%] and AML [21%; p = 0.31]. Two children subsequently achieved remission and underwent a second transplant. Six patients [4.2%] had secondary graft failure—of these 4 subsequently underwent a second Haplo-SCT. The 1-year overall survival [OS] for the entire cohort is 61.6 ± 4.3% while the 2-year OS is 54.9 ± 4.6% (Fig 1). The 2-year OS in children aged 0–5, 6–10 and 11–15 years was 39.3 ± 7.5%, 71.3 ± 6.8% and 55.7 ± 8.8% respectively [p = 0.03] (Fig 2). The 2-year OS was better with early [100%] and intermediate disease [66.1%] compared to late disease [32.6%] but not statistically significant [p = 0.85] (Fig. 3). The 2-year OS has improved with time—25.0 ± 2.1% for 2010–2013, 47.5 ± 6.2% for 2014–2017 and 67.1 ± 6.6% for 2018–2021 [p = 0.04] (Fig. 3). Overall Day 100 non-relapse mortality [NRM] was 25.3% and has reduced from 50% in 2010–2013 to 30.1% in 2014–2017 and 18% from 2018 to 2021.

Fig. 1.

Overall survival in children undergoing haplo-SCT

Fig. 2.

Overall survival among different age groups of children undergoing haplo-SCT

Fig. 3.

Overall survival of children following haplo-SCT based on year of HSCT

On univariate analysis, age 6–10 years [p = 0.01], myeloablative conditioning [p = 0.02] and absence of CMV reactivation [p < 0.05] was associated with better survival while presence of bacterial infection [p < 0.05], viral infection [p < 0.05], IFD [p < 0.05] and graft failure [0.01] were associated with poorer survival (Table 4). Gender, disease type, age and type of donor, degree of HLA match, cell dose infused, neutrophil and platelet engraftment, acute GVHD or chronic GVHD did not impact on overall survival. On multivariate analysis, bacterial infection [p = 0.01], invasive fungal disease [p < 0.05] and graft failure [p = 0.02] continued to remain significant.

Table 4.

Univariate and Multivariate analysis of factors affecting overall survival

Variables	Univariate analysis	p value	Multivariate analysis	p value
Age
0–5 yrs	1		1
6–10 yrs	0.462	0.01	0.690 [0.341–1.397]	0.3
11–15 yrs	0.651	0.18
Conditioning
Nonmyeloablative	1		1
Myeloablative	0.546 [0.326–0.912]	0.02	1.169 [0.621–2.200]	0.62
DBI
No	1	< 0.05	1	0.01
Yes	3.148 [1.874–5.287]		2.243 [1.159–4.344]
IFD
No	1	< 0.05	1	< 0.05
Yes	3.255 [1.938–5.466]		3.265 [1.561–6.831]
DVI
No	1		1
Yes	0.373 [0.206–0.674]	< 0.05	0.786 [0.240–2.575]	0.69
CMV reactivation
Yes	1		1
No	0.379 [0.216–0.664]	< 0.05	0.962 [0.395–2.343]	0.93
Graft failure
No	1	0.01	1	0.02
Yes	3.021 [1.302–7.009]		4.070 [1.151–14.388]

Bold is for significant p values

Outcomes of HSCT for malignant and non-malignant disease were also compared (Table 5). Though disease status, graft source and cell dose infused were similar, myeloablative conditioning was more common with malignant disease [84.9% vs 41.2%; p < 0.05]. Presence of active infection at HSCT was higher in non-malignant disease [20% vs 3.8%; p < 0.05]. Primary engraftment was significantly higher with malignant disease [94.3% vs 74.1%; p < 0.05] with a higher incidence of PGF with non-malignant disease [15.2% vs 1.9%; p = 0.01]. Rates of acute and chronic GVHD were similar between both groups. The Day 100 NRM was significantly higher with non-malignant disease [31.7% vs 15.1%; p = 0.02]. The 2-year overall survival was similar [54.1% with non-malignant and 62.2% with malignant disease; p = 0.34].

Table 5.

Comparison of HSCT outcomes between children have malignant disease versus non-malignant disease

Variables	Malignant (n = 53)	Non-malignant (n = 85)	p value
Disease status
Early disease [ED]	6 [11.3%]	9 [10.6%]
Intermediate disease [ID]	26 [49.1%]	45 [52.9%]	NS
Late disease [LD]	21 [39.6%]	31 [36.5%]
Previous Bacterial or invasive fungal infection [within 8 weeks prior to HSCT	2 [3.8%]	17 [20%]	< 0.05
Conditioning
Myeloablative	45 [84.9%]	35 [41.2%]	< 0.05
Reduced intensity	8 [15.1%]	50 [58.8%]
Graft source
PBSC	52 [98.1%]	79 [92.9%]	NS
Bone marrow	1 [1.9%]	6 [7.1%]
Cell dose	11.3	11.4	NS
	(1.4–22)	(5–39)
% engraftment	94.30%	74.10%	< 0.05
Primary Graft failure	1.90%	15.20%	0.01
Death prior to engraftment	3.80%	10.70%	0.15
Grade II–IV Acute GVHD	50%	49.20%	0.93
Chronic GVHD	47.20%	35.50%	0.28
Day 100 NRM	15.10%	31.70%	0.02
Secondary graft failure relapse	0	4.70%
	15 (28.3%)	–
2 year OS	62.20%	54.10%	0.34
2 year OS based on year of HSCT
2010–2013	0%	50%
2014–2017	50%	46.60%
2018–2021	82.60%	63.10%	0.1
2 year OS based on disease status
Early disease [ED]	100%	100%	0.14
Intermediate disease [ID]	76.90%	60%
Late disease [LD]	33.30%	32.20%

Discussion: This study describes the outcomes of a large cohort of children undergoing Haplo-SCT using the PTCy platform with a reported 2-year survival of 55%. These outcomes are similar to data from the Spanish group GETH involving 232 transplants that reported a 5-year OS of 48.5% with no difference between malignant and non-malignant disease³. In another study involving high-risk hematological malignancies in children, the 4-year OS was 55.6% [11]. There has been an encouraging improvement in OS from 25% in 2010–13 to 67% for 2018–2021 related to reduction in early mortality from 50 to 18%. This reduction is multifactorial and includes early identification of a child for HaploSCT, improvement in supportive care and better antibacterial and antifungal drugs. Survival was lowest in the youngest age group [39.3 ± 7.5%] compared to 6–10 years [71.3 ± 6.8%] and 11–15 years [55.7 ± 8.8%]. A study on 55 infants undergoing HSCT from Greece reported a mortality of 30% due to graft failure, acute GVHD and infective complications [12]. In our cohort, this age group mainly had children with either PID or AA with many having active infection at the time of HSCT. Infection related mortality and graft failure both contributed to early mortality as has been reported in other studies [13].

The PTCy platform is an important cost-effective measure in performing HaploSCT in countries where resource constraints play a vital role in determining both access to transplant and its outcomes. Reduction in non-relapse mortality [NRM] is vital to improving outcomes and the 3 important causes of mortality include graft failure, infections and acute GVHD. Preventing PGF is important since most of our patients are unable to proceed for a second transplant either due to financial implications or presence of life-threatening infections. Primary graft failure was mainly seen in children with AA. Strategies to reduce PGF include adding ATG to conditioning, increasing TBI dose to 400 cGy or using mesenchymal stromal cells prophylactically [14, 15].

Infective complications were common with 93% reporting atleast 1 infection similar to a study involving 104 patients undergoing Haplo-SCT with PBSC where 89% experienced one infection [16]. The incidence of bacterial infections was 32% similar to data from GETH who reported gram-negative infection rates of 46% at 37 months [17]. A study from China involving 1847 children including 1380 haplos showed a higher incidence of pre-engraftment BSI after haplo-transplants [18]. The overall incidence of viral infections was 76% in our series with 63% reporting CMV reactivation similar to data from other studies wherein 60–70% show viral infections with 40–60% showing CMV reactivation [19–21]. IFD was reported in 29% with proven/probable IFI in 18%. Few observational studies from China, Italy and Taiwan have reported similar rates of 15–17% though data from US did not show any effect of donor source [16, 22–24]. Both bacterial and fungal infection were independent risk factors predicting mortality in our series. So how can we try and reduce this mortality? We have previously shown that infections with multidrug resistant organisms [MDRO] were common after HSCT and associated with higher mortality [25]. Our present strategy is to start higher antibiotics in patients with previous history of infection with MDRO or have presence of MDRO on fecal surveillance prior to HSCT. We need to analyze whether this differential strategy has been useful in reducing MDRO related mortality.

The high incidence of fungal and viral infections is mainly related to the high incidence of acute GVHD and though acute GVHD was not identified as an independent risk factor for mortality, we need better strategies to reduce GVHD. Acute GVHD was seen in 50% with 25% having grade III–IV acute GVHD which are similar to reports with the use of PBSC [25–43%] though use of bone marrow has lower rates [26–29]. Strategies to reduce acute GVHD could include the use of mini-Methotrexate or adding rabbit ATG to the regimen. In a study involving 52 children, the replacement of MMF with mini methotrexate was associated with a lower incidence of acute GVHD while low doses of ATG [5 mg/kg] in haplo-identical transplants again was associated with lower rates of GVHD [30, 31].

Relapse was seen in 28% of children with malignant disease. In a study of 180 paediatric patients with ALL, Ruggeri et al. reported a 2-year cumulative incidence of relapse [CIR] of 25, 37, and 50% for patients transplanted in CR1, CR2, and CR3, respectively [32]. Similar rates of relapse have been seen in other small series involving 30–50 patients [27, 28, 30, 33, 34]. Identifying patients early for HSCT and inducing MRD negativity prior to HSCT are important strategies to reduce the risk of relapse.

Outcomes of children with non-malignant disease needs to be improved further. Infectious complications and high incidence of PGF both contribute to a higher mortality following Haploidentical HSCT. In patients with active infection, the judicious use of granulocyte transfusions both prior to and during HSCT may help in reducing mortality [35]. Alternative conditioning regimens need to be explored in children with active infection at HSCT which is common with PID and AA. As mentioned previously, it is important to consider new strategies to reduce graft failure including addition of ATG and/or increasing the dose of TBI [14, 15].

In conclusion, this large series from India demonstrates that T replete haplo-identical transplantation using the PTCy platform is feasible in children and outcomes have been improving over the years. Newer strategies to reduce graft failure, infections and GVHD need to be explored to improve survival further.

Declarations

Conflict of interest

All authors declare no competing financial interests that are directly or indirectly related to the work submitted for publication.