Equal outcomes after human leukocyte antigen‐matched unrelated versus sibling donor transplants—Results of the <scp>AML SCT</scp> ‐ <scp>BFM</scp> 2007 trial and consecutive ‘real‐world routine’

In this letter, we report nearly identical outcomes after human leukocyte antigen (HLA)-matched unrelated donor (MUD) and HLA-identical sibling donor (ISD) haematopoietic stem cell transplantation (HSCT) for paediatric acute myelogenous leukaemia (AML). Over the last decade, much progress has been made to define those children who will eventually benefit from an allogeneic HSCT.1, 2 Advances in HSCT, namely by improvements in supportive care and HLA-typing technology, have contributed to better outcomes with survival rates reaching up to 70% in high-income countries. Historically, treatment-related mortality was lower after ISD transplantations than after MUD transplantations for paediatric malignancies.3 The implementation of the European Clinical Directive 2001/20/EC and the related regulatory framework forced the transformation of prospective HSCT studies into good clinical practise-compliant clinical trials. In this context, we had initiated the AML SCT-BFM 2007 trial (ClinicalTrials.gov ID: NCT00606723).4 Parts of this trial comparing outcome data between children transplanted from either ISD or MUD and received identical transplant regimens with the exception of the use of anti-T-lymphocyte globulin in MUDs have not been previously reported. Here, a well-characterized graft as the investigational product, a protocol-stipulated supportive care concept and a standardized indication for transplantation make it now possible to put these data meaningfully in an international context.5-7 Bone marrow as a graft source was utilized in 90 patients and peripheral blood stem cells in 50 patients. Children and adolescents with high-risk de novo AML or relapsed AML were eligible for enrolment. Between May 2010 and February 2016, 140 children (age <12 years, n = 91) and adolescents (age <18 but ≥12 years, n = 49) were recruited in Germany, Austria and the Czech Republic. An HLA-matched donor was either an ISD or an at least 9/10 HLA MUD. Subjects with AML in first complete remission (CR1) (n = 45) or second complete remission (CR2) (n = 47) were conditioned with a myeloablative regiment (MAC) consisting of busulfan (age-adjusted intravenous (iv) dosing: 3.2–4.8 mg/kg body weight on days −7 to −4), cyclophosphamide (60 mg/kg iv on days −3 and −2) and melphalan (140 mg/m2 iv on day −1) (BuCyMel). Graft-versus-host disease (GvHD) prophylaxis was ciclosporin (CSA) and short-term methotrexate (MTX). Patients with MUDs additionally received anti-T-lymphocyte globulin (ATLG, Grafalon®, Neovii Biotech, Switzerland) at a dose of 20 mg/kg/day on days −3 to −1. Patients with poor response to AML induction therapy (n = 48) were stratified to receive a reduced-intensity regimen consisting of a cytoreductive block with fludarabine (30 mg/m2/day iv), amsacrine (100 mg/m2/day iv) and cytarabine (2 g/m2/day iv) (FLAMSA) on days −12 to −9, immediately followed by 4 Gy TBI on day −5 and cyclophosphamide (60 [unrelated]/40 [related] mg/kg/day iv) on days −4 to −3. After early taper of mycophenolate mofetil (MMF) and CSA, three prophylactic donor lymphocyte infusions were scheduled on days 120, 150 and 180. ATLG was added for MUDs (20 mg/kg × day) and ISDs (10 mg/kg × day) on days −4 to −2. The AML SCT-BFM 2007 trial was registered under ClinicalTrials.gov (EudraCT-Number: 2007-004517-34). For data collection in the German National Registry for Paediatric Stem Cell Transplantation (PRST), parents or respective legal representatives have given their written consent. Of the 140 patients (68 female, 72 male), 99 received a graft from an MUD (10/10 n = 70, 9/10 n = 29) and 32 from an ISD. Data on ethnicity were not collected. Nine patients without a matched donor were eligible for haplo-HSCT and excluded from this analysis. Disease and transplant characteristics were published earlier and, except for the graft source (Bone Marrow 91% for ISDs and 59% for MUDs), were similar.4 With a median follow-up of 4 years on the total cohort, overall survival (OS) was 65% (95% confidence intervals [CI] = 0.54–0.76) for patients transplanted from MUD and 62% (95% CI = 0.43–0.76) for those transplanted from ISD (Figure 1). SCT-related causes of death are summarized in Table 1. Similarly, there were no differences between MUD and ISD transplants in the 4-year cumulative incidence rates (CI) of acute GvHD (aGvHD) II–IV with 33% (Standard Error 95% CI = 0.23–0.42) vs. 28% (95% CI = 0.12–0.44), chronic GvHD (cGvHD) 12% (95% CI = 0.006–0.19) vs. 16% (95% CI = 0.03–0.28), relapse 28% (95% CI = 0.19–0.37) vs. 34% (95% CI = 0.18–0.51) and treatment related mortality (TRM) 14% (95% CI = 0.07–0.21) vs. 13% (95% CI = 0.01–0.24) respectively (Figure 1). Event-free survival (EFS) was 57% (95% CI = 0.46–0.66) and 50% (95% CI = 0.32–0.66) respectively. Acute and chronic GVHD-free survival rates were 39% (95% CI = 0.21–048) vs. 41% (95% CI = 0.24–0.57) and 49% (95% CI = 0.36–0.58) vs. 34% (95% CI = 0.19–0.51) respectively (Figure S1). Patients in CR1 or CR2 conditioned with BuCyMel had a comparable OS after MUD and ISD transplants with 76% (95% CI = 0.64–0.84) vs. 62% (95% CI = 0.40–0.78). Rates of aGvHD II–IV 33% (95% CI = 0.22–0.44) vs. 30% (95% CI = 0.11–0.48), relapse 21% (95% CI = 0.12–0.31) vs. 29% (95% CI = 0.10–0.48) and TRM 12% (95% CI = 0.05–0.20) vs. 17% (95% CI = 0.01–0.22) were similar (Figure S2). Patients with poor response to induction who qualified for reduced intensity conditioning using fludarabine, amsacrine, and cytarabine (FLAMSA-RIC) had an OS of 33% (95% CI = 0.16–0.51) compared to 63% (95% CI = 0.23–0.86) respectively. CIs of aGvHD II–IV after FLAMSA-RIC were 31% (95% CI = 0.12–0.49) vs. 25% (95% CI = 0.08–0.58), relapse 46% (95% CI = 0.26–0.66) vs. 50% (95% CI = 0.11–0.89) and TRM 20% (95% CI = 0.04–0.37) vs. 0% (95% CI = 0.0–0.0) (p = 0.18) in MUD and ISD transplants respectively (Figure S3). For the AML SCT BFM 2007 trial, ATLG Grafalon® (Neovii Biotech, Rapperswil, Switzerland) was used for the MUD transplants after BuCyMel and dosed according to its original approval with 60 mg/kg recipient weight. Prospective pharmacokinetic studies measuring either total rabbit anti thymocyte globuline (ATG) or the active ATG as the fraction capable of lymphocyte binding8 were not performed in this trial. Therefore, it is important to acknowledge that these results should not be extrapolated for transplant settings using different ATG preparations since they differ profoundly in their biological activity.9 The active forms of ATG Genzyme, for example, the most commonly used preparation for serotherapy worldwide, have shown to be cleared from the circulation much slower than the respective forms of ATLG Neovii. This likely impacts the incidence of GvHD, infection risk and possibly relapse rates.9-11 Apart from serotherapy, all children treated within our trial received the identical GvHD prophylaxis consisting of CSA plus short-term MTX for the MAC regimen and CSA plus MMF for the RIC transplants. It is noteworthy that 29 of the 99 MUDs were 9/10 HLA-matched only, but for historic reasons considered to be ‘matched’ and all recipients of an ISD received short-term MTX plus CSA. Especially in Europe, there has been a tendency to omit MTX for ISD transplants with the intention to foster graft versus leukemia (GVL) effects.12 This is relevant since we obtained nearly superimposable survival curves for ISDs and MUDs in a disease setting generally considered to be specifically sensitive to GVL effects. It was questionable whether results from a tightly monitored trial would translate into ‘real-world routine’. We, therefore, retrospectively analysed the outcome of paediatric AML transplants (241 MUDs and 70 ISDs) from the national paediatric HSCT registry after the original trial had been closed. It is important to acknowledge that the two cohorts differ in that a significantly higher proportion of children were transplanted in CR1 in the ‘real-world’ cohort and that treosulfan-based myeloablative conditioning regimens were used more frequently. In addition, while GvHD prophylaxis in the trial protocol was standardized (CSA and short-course MTX) and monitored tightly for both MUDs and ISDs, substantially greater heterogeneity exists in real-world practice. This variability primarily involves earlier tapering of CSA or omission of MTX in the ISD setting. In the ‘real-world’ cohort, OS at 4 years was 62% (95% CI = 0.54–0.69) vs. 65% (95% CI = 0.50–0.76), CI of relapse 26% (95% CI = 0.20–0.32) vs. 29% (95% CI = 0.17–0.41) and TRM 13% (95% CI = 0.08–0.17) vs. 9% (95% CI = 0.02–0.16). With CIs of acute (≥II°) and chronic GvHD reaching 21% (95% CI = 0.15–0.26) vs. 27% (95% CI = 0.16–0.37) and 15% (95% CI = 0.01–0.24) vs. 13% (95% CI = 0.08–0.22) (Figure S4A–D), these results seem to reproduce the trial outcomes in a clinical routine setting. Altogether, we can show that both outcomes from a prospective well-controlled paediatric AML HSCT trial and consecutive ‘real-world routine’ are nearly identical between MUDs and ISDs. This may have important ethical and practical implications for the current use of minor donors in paediatric AML. M.G.S. collected and reviewed data, analysed data, wrote and edited the manuscript; A.B. and R.H. wrote and edited the manuscript; M.Z. did the statistical analysis and created the figures; P.J.L., M.H.A., P.B., M.E., B.G., T.K., C.M.-K., R.M., I.M., C.M.N., C.P., B.B., D.R., P.S., B.S., W.W. and R.H. collected data and edited the manuscript. We thank all the participating children and their families. We thank the nursing team for their dedicated and professional work and Dr. Rita Beier for providing access to the PRST registry. This work was supported by grants from the German Federal Ministry of Education and Research (M.G.S.), Verein für krebskranke Kinder E.V. Hannover (M.G.S), a research grant by Neovii Biotech (M.G.S.) and the ‘Katharina-Hardt-Foundation’ (A.B.). M.G.S. received research support from Neovii Biotech. The remaining authors did not declare any conflicts of interest. The remaining authors do not declare any relevant conflict of interests. The trial was approved by both the regulatory authorities and institutional ethics committee of each participating clinical trial centre. Written informed consent was obtained from each patient (≥12 years) and/or from their legal guardians for both the original trial and the consecutively used data registry. The original clinical trial protocol was registered under ClinicalTrials.gov (EudraCT-Number: 2007-004517-34). Deidentified data and the trial protocol are available to qualified researchers who submit an in-scope proposal. The data will be available from the authors of the publication to ensure that data are used in accordance with the informed consent. Data are provided for the sole use of the approved request (tertiary dissemination will not be permitted). To gain access, data requesters will need to sign a data sharing agreement. Figures S1–S4. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.