Busulfan–fludarabine‐ or treosulfan–fludarabine‐based conditioning before allogeneic HSCT from matched sibling donors in paediatric patients with sickle cell disease: A study on behalf of the EBMT Paediatric Diseases and Inborn Errors Working Parties

How important is choice of conditioning regimen in allogeneic haematopoietic stem cell transplantation (HSCT) for sickle cell disease (SCD)? We compared HSCT outcomes by conditioning regimen in paediatric patients with SCD from the EBMT registry. In 2010–2020, 251 patients aged <18 years underwent a first matched sibling donor (MSD) HSCT with conditioning based on busulfan–fludarabine (bu–flu; n = 89) or treosulfan–fludarabine (treo–flu; n = 162). In the bu–flu and treo–flu groups, 51.7% and 99.4% of patients, respectively, received thiotepa. Median follow-up was 2.7 years. Two-year overall survival (OS) was 98.7% (95% confidence interval [CI]: 90.9–99.8) with bu–flu and 99.3% (95% CI: 95.2–99.9) with treo–flu (p = 0.63). Grade III–IV acute graft-versus-host disease (GVHD) at 100 days was 2.4% (95% CI: 0.4–7.5) and 0.6% (0.1%–3.2%) for bu–flu and treo–flu respectively (p = 0.25). The 2-year incidence of extensive chronic GVHD was 1.5% (95% CI: 0.1–7.3) with bu–flu and 8.0% (95% CI: 4.1–13.3) with treo–flu (p = 0.057). These multinational data confirm the excellent curative capacity of MSD HSCT with myeloablative conditioning. Both conditioning regimens yielded excellent OS, low rates of acute and chronic GVHD, and low rates of graft failure. Sickle cell disease (SCD) is the most common monogenic disorder worldwide, occurring in over 300 000 live births annually.1 The prevalence is highest in Sub-Saharan Africa.2 In high-income countries, mortality from SCD has greatly reduced in recent decades through supportive strategies. Nevertheless, SCD significantly shortens median life expectancy.3 Allogeneic haematopoietic stem cell transplantation (HSCT) remains the only established curative treatment for SCD.4 The best results have been obtained with matched sibling donors (MSDs), for which myeloablative conditioning (MAC) is most commonly used.5, 6 Although options for alternative donors have increased, data from comparative prospective studies are scarce.7-9 Patient age at transplantation is a critical outcome parameter.10, 11 In this retrospective, multicentre study using the European Society for Blood and Marrow Transplantation (EBMT) registry data, we assessed the impact of MAC regimen on HSCT outcome in paediatric patients with SCD. We include patients who underwent first MSD allogeneic HSCT following conditioning based on either busulfan (bu) and fludarabine (flu) or treosulfan (treo) and flu in 2010–2020 when aged <18 years. All patients consented to data reporting and collection. The primary end-point was the 2-year overall survival (OS) calculated using the Kaplan–Meier method. Secondary end-points were the 2-year incidence of second HSCT, graft failure, disease-free survival (DFS) and chronic GVHD (cGVHD); acute graft-versus-host disease (aGVHD) at 100 days; and neutrophil engraftment (the first of three consecutive days with neutrophil count >0.5 × 109/L) at 30 and 60 days. Graft failure was defined as neutrophil count <0.5 × 109/L after initial engraftment not related to infection or drug toxicity, and categorized as either early (≤100 days) or late (≥100 days). Due to the low number of events, only univariable analyses were performed. Statistical methodology is provided in the Supporting Information. A total of 251 patients were transplanted at a median age of 9.1 years (interquartile range: 6.4–12.7 years) (Table 1). Median age at HSCT was 10.3 years and 8.3 years with bu–flu and treo–flu respectively (p < 0.001). Bone marrow was the most common cell source (88.8%). Treo–flu was used in 162 patients and bu–flu in 89 patients. The conditioning regimens used in the bu–flu group included additional cyclophosphamide in 11 (12.4%) patients and thiotepa (thio) in 46 (51.7%) patients (Table 1). Bu–flu-based conditioning was given more often in recent years, with a median administration year of 2018 versus 2017 for treo–flu (p = 0.002). A higher proportion of in vivo T-cell depletion using serotherapy was observed with treo–flu versus bu–flu (93.2% vs. 50.6% respectively; p < 0.001). Two-year OS was 98.7% (95% confidence interval [CI]: 90.9–99.8) with bu–flu and 99.3% (95% CI: 95.2–99.9) with treo–flu (p = 0.63; Table 2). All patients achieved neutrophil engraftment by day 60. Severe (grade III–IV) aGVHD at day 100 occurred in 2.4% (95% CI: 0.4–7.5) of patients receiving bu–flu and 0.6% (95% CI: 0.1–3.2) receiving treo–flu (p = 0.25). The 2-year incidence of cGVHD was 5.5% (95% CI: 1.8–12.6) with bu–flu versus 15.6% (95% CI: 9.9–22.4) with treo–flu (p = 0.04). The 2-year cumulative incidence of extensive cGVHD was 1.5% (95% CI: 0.1–7.3) with bu–flu versus 8.0% (95% CI: 4.1–13.6) with treo–flu (p = 0.057). There was no primary graft failure. The 2-year incidence of secondary graft failure was 9% (95% CI: 3.6–17.5) and 5.7% (95% CI: 2.6–10.5) for bu–flu and treo–flu respectively (p = 0.75). The incidence of second HSCT was 1.8% (95% CI: 0.1–8.3) with bu–flu and 2.3% (95% CI: 0.6–6.0) with treo–flu (p = 0.39). Two-year DFS was 89.7% (95% CI: 79.4–95.0) and 93.6% (88.0%–96.7%), respectively, for bu–flu and treo–flu (p = 0.64). In the bu–flu group, one patient died following first HSCT due to infection-associated multiorgan failure. One patient had early graft failure and five patients experienced late graft failure, with one consecutive second HSCT from the same donor; all six were alive at last follow-up. In the treo–flu group, one patient died following first HSCT due to multiorgan failure. Four patients experienced early graft failure. Among these, two had a second HSCT from the same MSD. Seven patients had late graft failure, four of whom underwent second HSCT. Two underwent HSCT from the same MSD, one from another MSD and one from a mismatched family member. At last follow-up, all seven were alive (Table S1). Thio use increased over time in the treo–flu group; although not comparatively analysed, this may have contributed to improved engraftment stability over time. Prospective studies are needed to determine whether non-bu protocols are required for the preparation of modified cells prior to gene therapy for SCD in patients at risk of hepatic complications.12 Although the incidence of severe aGVHD was higher with bu–flu, the incidence of cGVHD and extensive cGVHD was higher with treo–flu. This might be due to different GVHD prophylaxis, duration and type of immunosuppression, and other local treatment standards.13 Gonadal dysfunction following allogeneic HSCT for non-malignant diseases is of great concern; myeloablative bu–flu, especially in adolescent girls, is responsible for high infertility rates.14 Faraci et al. published a retrospective analysis of 137 children and adolescents transplanted with bu–flu- (n = 118) or treo–flu-based conditioning (n = 19).15 They observed a significantly lower frequency of gonadal damage in the treo–flu group. Despite using the largest European transplant registry, our analysis has some limitations. The excellent OS and low number of events rendered statistical comparison between groups inadequately powered. Moreover, descriptive statistics were more appropriate because the bu–flu group contained a variety of regimens while the treo–flu group contained mainly treo–flu–thio. We could not assess long-term organ and gonadal toxicities related to vascular or transfusion-related complications before or related to HSCT. Details on in vivo and in vitro T-cell depletion were lacking, and granular chimerism data were not available. In summary, using one of the largest datasets of MSD HSCT for SCD, we confirm the excellent curative potential of MSD HSCT following MAC. Both bu–flu- and trio–flu-based conditioning yielded excellent OS and low rates of acute and chronic GVHD and graft failure. Prospective studies are needed to investigate long-term graft function and the late toxicity profile of bu–flu- and treo-based regimens. Jacques-Emmanuel Galimard performed the statistical analysis. Christina Peters, Arjan C. Lankester, Selim Corbacioglu, Annamária Cseh and Jacques-Emmanuel Galimard wrote the letter draft. Jacques-Emmanuel Galimard, Christina Peters, Annamária Cseh, Arjan C. Lankester and Selim Corbacioglu revised the draft critically. All other co-authors contributed to the project with patient recruitment. All authors approved the submitted letter. This study was funded by the St. Anna Children's Cancer Research Institute, Vienna, Austria. A.C. has no competing interests. C.P. declares research grants from Neovii, Riemser, Medac and Amgen (to institution); is a member of advisory boards for Novartis, Amgen, Medac and Pierre Fabre (personal and institutional); and has received speaker fees from Neovii, Medac, Jazz, Riemser, Novartis and Amgen. S.C. is advisory board member for Vertex. All participating centres have received approval to perform allogeneic HSCT for patients with SCD in accordance with national legislation. All participating sites were approved for patient data registration. Data S1. 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.