Thanks to IST combined with hATG and cyclosporine A (CsA), the 5-year OS of patients under 16 years of age increased from 50% before 1990 to approximately 80% today (62–64). In a Polish study, the 10-year OS rate reached 78% in children treated mainly with hATG (65). According to a German study, children with vSAA had a higher rate of complete response (CR) than children with SAA (68% vs 45%; P = 0.009). They also showed better survival (93% vs 81%; P < 0.001) (66).

Different ATG preparations were used in IST with various biological effects. Lymphoglobulin^® (hATG, Genzyme) had been widely used for IST in Europe, but it was withdrawn from the market in 2007. In the absence of licensed hATG, rabbit ATG (rATG) sera were used. However, the effectiveness of rATG treatment was worse compared with hATG. In a randomized study by Scheinberg et al., the 3-year OS after rATG was inferior to that of hATG (76% vs 96%) (67). Similarly, the EBMT Severe Aplastic Anemia Working Party (EBMT-SAA-WP) reported that the 2-year OS after r-ATG was only 68% vs 86% for hATG (68). Some studies (mainly retrospective ones) showed similar effectiveness of hATG and rATG (68–74). Chang et al. reported the 4-year OS of 75% without statistical difference between patients treated with hATG (n = 29) and rATG (n = 33) (75). In contrast, the EWOG-SAA-2010 interim analysis showed a lower overall response rate (ORR) to IST in patients treated with rATG compared with hATG (22% vs 42%, P = 0.03) but without differences in OS (88% vs 93%) (76). Similarly, Pawelec et al. demonstrated the 3-year OS of 78% and the 10-year survival of 67% for hATG and rATG, respectively (77). However, in a meta-analysis by Hayakawa et al., the ORR was significantly higher for hATG vs rATG (risk ratio [RR], 1.27; 95% CI, 1.05-1.54; P = 0.015) (78). In another meta-analysis by Yang et al., the beneficial effect of hATG on the ORR at 6 months in treatment-naïve patients with AA was also shown (RR, 1.86) with comparable safety profiles (79). A summarized comparison of studies with hATG and rATG in IST of AA patients is shown in Table 1 .

A third source of ATG – porcine serum (pATG) – was studied in a few Chinese trials. In a study by Bing et al., the ORR was 83.3% (CR, 54.2%; partial response [PR], 29.2%) with a median time of 90 days (range, 23-380), and 10.4% of patients died of infection within 30 days (80). A pediatric study by Zhu et al. revealed a similar therapeutic effect and safety of pATG or rATG combined with CsA (24-month ORR, 81.6% vs 78.6%) as first-line therapy for pediatric patients with AA (81). In retrospective studies by Liu et al. and Chen et al., no significant differences in OS between pATG and rATG were observed (82, 83).

Factors increasing the ORR to IST in children include severity (ORR better in vSAA than in SAA), younger age, higher pretreatment reticulocyte count and lymphocyte count, male sex, and leukocyte count (84, 85). In a meta-analysis by Gu et al., the earlier start of IST, higher neutrophil count at IST, and rate of lymphocyte clearance during IST were associated with improved response rates (86).

The addition of granulocyte-colony-stimulating factor (G-CSF) may reduce the rate of early infectious episodes and days of hospitalization in vSAA patients, but has no effect on the response rate to IST and survival (87, 88). The safety of G-CSF after IST might be compromised. Japanese studies from 1998 and 2002 suggested a close relationship between the long-term use of G-CSF and secondary MDS in nonresponders to IST (89, 90). In a study on 802 SAA patients, a multivariate analysis revealed increasing age, higher disease severity, and increasing number of G-CSF therapy days to be risk factors for MDS/AML transformation. However, the prolonged G-CSF administration in these studies was typical for patients with suboptimal response, and this factor can be associated with other, nonimmune mediated mechanisms of SAA pathogenesis and, in turn, with a higher risk of clonal disease (91). A meta-analysis by Ding et al. and a follow-up study of a randomized, prospective trial by Tichelli et al. did not confirm any impact of G-CSF on the risk of clonal evolution in patients with SAA (92, 93). Regardless of the use of G-CSF, SAA patients treated with IST are at risk of several late malignant and nonmalignant complications. There is a significant risk of relapse (10% at 10 years) and 8%–16% risk of clonal disease (MDS/AML), which does not reach plateau (64, 85, 93).

In the pediatric EWOG strategy for SAA treatment if hATG is not available, rATG may be used as an alternative therapy. Currently, G-CSF should only be used in patients with severe neutropenia (ANC <0.5 × 10⁹/L) and prolonged administration of G-CSF should be avoided. Children not responding to the IST therapy or relapsing afterwards must seek alloHCT from MUDs or alternative donors (ADs) (59). According to the Pediatric Haemato-Oncology Italian Association, patients without MUD can receive the second course of IST (94). A switch of serotherapy from hATG to rATG and vice-versa should be considered, but the expected ORR is only 30% after failure of the first IST and 60% after relapse (95). The remaining options include transplantation from ADs and immunotherapy with alemtuzumab and ELT. The overview of the SAA management in the pediatric population based on the current recommendations is presented in Figure 1 .

According to EBMT recommendations, patients younger than 40 years should undergo alloHCT from MSD. Patients above this age limit and patients without MSD are candidates for IST (60). Best outcomes are reported with hATG combined with CsA and ELT, and this triple combination is currently recommended in adult patients. The current treatment schedule for patients with SAA based on EBMT recommendations and data from the available literature is shown in Figure 2 .

A standard total dose of hATG (ATGAM^®) is 160 mg/kg. Most studies evaluated the dose given in a 4-day schedule (40 mg/kg daily for 4 days). The dose was evaluated in 2 prospective randomized studies by Scheinberg et al.: one study evaluating the use of hATG plus CsA with or without sirolimus and the other comparing hATG with rATG (67, 96). Cyclosporine A was given orally at an initial dose of 5 mg/kg for several months (34, 61). In earlier studies, ATGAM^® was given at lower daily doses for 8 or 10 days, but at the same total dose of 160 mg/kg. Champlin et al. evaluated hATG given for 8 days (20 mg/kg daily) in monotherapy or with androgens (50). Doney et al. used hATG for 10 days in combination with oxymetholone given orally for 3 months (97). In both studies, CsA was not incorporated in IST protocols, so it is impossible to compare the outcomes with those obtained in modern studies. A lower total dose of hATG was evaluated in a retrospective trial by Scott et al. Two schedules of IST based on CsA and hATG were compared: one group (n = 14) received hATG at a standard dose of 40 mg/kg daily for 4 days, and the other group (n = 17) received 15 mg/kg daily for 5 days (98). The efficacy of both treatment regimens was comparable, although a low number of patients and the retrospective trial design should be considered.

Best outcomes were obtained when hATG was combined with CsA, resulting in 60%-65% of responses, including 10% of CR (96, 99). The addition of other immunosuppressive agents, including sirolimus and mycophenolate mofetil (MMF), did not improve the response rate in prospective randomized trials (52, 96, 100). A retrospective multicenter real-life study included all patients diagnosed with AA between 2000 and 2011 in Sweden and treated in the first line with ATG plus CsA. The study evaluated 158 patients, 128 of whom received rATG. Response to treatment was reported in 47% of patients. The 5-year OS was significantly higher in patients achieving PR or CR as compared with nonresponders (89% vs 52%, P <0.001). The 5-year OS was 77.3% and 58.5% for SAA and vSAA, respectively (P = 0.002). The age of patients did not influence the efficacy of treatment (74). The EBMT performed a retrospective study evaluating the efficacy of rATG in Europe and Asia. The study included 955 patients receiving the first-line treatment with ATG and CsA between 2001 and 2012. The ORR was 37% at 3 months and 52% at 6 months, with a subsequent increase up to 65% at 12 months after treatment. The OS at 10 years was 70% (101).

The prospective randomized trial by Scheinberg et al. revealed superior efficacy of hATG vs rATG (67). A total of 120 patients were evaluated, 60 patients per group. The ORR at 3 months of IST was 62% in the hATG group and 33% in the rATG group (P = 0.0017). The response evaluated at 6 months was also superior for hATG compared with rATG (68% vs 37%; P <0.001). The 3-year OS was also significantly higher for hATG than for rATG (96% vs 76%, P = 0.04). Real-world data from France concerning the efficacy of hATG in SAA were also encouraging (102).

The safety issues concerning the infusion of either hATG or rATG are similar, as the most common complications include skin reactions, headache, chills, and fever. In rare cases, anaphylactic reactions or systemic inflammatory reactions due to cytokine release may occur such as cytokine release syndrome or systemic inflammatory response syndrome. To reduce the incidence of immediate adverse reactions, corticosteroids, antipyretics, and antihistamines are recommended (60). To reduce the risk of infusion-related side effects, the preinfusion skin testing should be considered, especially in patients with an increased risk of anaphylaxis (e.g. patients with atopic eczema). However, the negative results of the skin testing do not exclude anaphylactic reactions or development of serum sickness. To reduce the risk of thrombophlebitis, ATG administration via central line is recommended. The prolonged infusion time (minimum 12 hours) is linked to better infusion tolerance (60). Several days after ATG infusion, serum sickness may occur including rash, arthralgia, and fever. The symptoms respond well to steroid treatment. A temporary increase in liver function tests may be observed after ATG administration. The most frequent severe adverse events were infections and hemorrhage (67, 96, 102).

Best response to IST should be evaluated 4 to 6 months after hATG administration (1, 60, 61, 103). The criteria of CR used in recent clinical trials are as follows: ANC >1 × 10⁹/L, hemoglobin >10 g/dL, and platelet count >100 × 10⁹/L (96). Most patients achieve PR when they no longer meet criteria for severe type of AA and become transfusion independent but do not meet CR criteria (ANC >0.5 × 10⁹/L, Hb >8 g/dL, platelet count >20 × 10⁹/L). Long-term responses to IST are observed in 60% of patients. However, some patients require more prolonged CsA treatment exceeding the minimal recommended 12 months and continuation of CsA for a longer time with subsequent slow tapering to minimize the risk of relapse (34). In 10% to 30% of patients, SAA may recur after initial response (1, 60, 61, 103). In such cases, alloHCT should be performed if possible or an alternative strategy with repeating IST might be considered, depending on the individual patient risk assessment and donor access. The second IST pathway should incorporate the option of the change of ATG origin or the use of alemtuzumab and addition of TPO-RA; however, insufficient stem cell reserve might limit the efficacy of IST (34, 52, 60, 61, 95, 104).

In adult patients, ELT proved to be an effective agent in patients with AA refractory to previous IST. Desmond et al. evaluated 43 patients unresponsive to IST and observed long-lasting responses in 40% of cases (56). In the prospective randomized RACE trial conducted by the EBMT-SAA-WP, 101 patients in the hATG + CsA arm and 96 patients in the hATG + CsA + ELT arm were evaluated (33). The median patient age was 53 years, and only 9 patients were children. The dose of ELT was 150 mg daily from day 14. The results were in favor of ELT addition in the first-line treatment with a significantly higher CR in the group receiving ELT at 3 months (22% vs 10%; P = 0.01) and a higher ORR at 3 months (59% vs 31%) and at 6 months (68% vs 41%). The ELT group responded faster to treatment and had no additional toxicity, including hepatotoxicity. The 2-year OS was comparable in both arms. The incidence of somatic mutations was similar in both groups and increased at 6 months after the administration of hATG, with clonal evolution observed in 10% to 15% of cases (33). In a Chinese study, the beneficial effect of prolonged ELT administration was demonstrated in patients without CR at 6 months (105). Based on data from these pivotal studies, the current standard approach to the therapy of SAA in adults includes the combination of IST with hATG with addition of ELT.

However, the effects of ELT on pediatric patients with SAA are less well documented and remain controversial, and the evidence is conflicting. Among the pivotal study population, 21% of patients were children aged 3-18 years, but in contrast to adults, in the pediatric sub-cohort analysis by Groarke et al., in children treated with ELT plus IST there was no significant difference in either the ORR or CR rate at 6 months (ORR: 70% in the ELT group and 72% in the historical group; P = 0.78) (57). In a study by Fang et al., better responses at 6 months were seen in children treated with ELT plus IST (CR: 50% vs 17.9%, P < 0.05; ORR: 94.4% vs 69.2%; P < 0.05) (106). Goronkova et al. reported a similar ORR in the ELT + IST and IST groups (65% vs 53%; P = 0.218). However, the CR rate was significantly higher in the ELT + IST group (31% vs 12%; P = 0.027) (107). Of note, the greatest benefit from ELT combined with IST was observed in patients with SAA with ORR 89% vs 57% (P = .028) but not in those with vSAA (52% vs 50%; P = .902). Other studies conducted in a small number of patients confirmed more rapid hematological responses and CR rates but failed to show ORR and survival advantage in the pediatric population (108). As there are no sufficient data to support the advantage of combined treatment of IST with ELT, the EWOG does not generally recommend this therapy in children (59).

Apart from ELT, also other TPO-RA are investigated in the SAA setting (109). Romiplostim, an Fc-peptide fusion protein administered subcutaneously, activates intracellular transcriptional pathways via the TPO receptor (110). In the dose-finding and long-term treatment phase 2 trial in SAA patients refractory to IST, 10 patients (30%) maintained a platelet response at 2 and 3 years, including 9 and 5 patients with erythroid and neutrophil response, respectively, without significant toxicity. A romiplostim dose of 10 μg/kg once weekly was suggested as a recommended starting dose (111). In a phase 2/3 study in 31 patients with SAA refractory to IST, a romiplostim dose was titrated by platelet response from 10 μg/kg once weekly up to a maximum of 20 μg/kg, and any hematological response at week 27 and 53 was 84% and 81%, respectively (112). In a retrospective study of 10 patients switched to romiplostim (20 μg/kg/week) after failure of the maximum ELT dose, 7 patients achieved neutrophil, erythroid, or platelet response, including one CR at 12 months (113). There is an ongoing phase 2 clinical trial in patients with SAA (either treatment-naïve or relapsed/refractory) with avatrombopag, a second-generation oral TPO-RA, administered from day 1 until day 180 at a dose of 60 mg/day, with dose adjustment according to platelet count, in combination with IST (hATG and CsA) or without IST (114).

Androgens have been used in AA for decades. Their mechanism of action in BMF is complex. Androgens induce erythropoiesis by increasing erythropoietin secretion and response to erythropoietin; they also stimulate the expansion of hematopoietic progenitors (50). Androgens also act by ameliorating telomerase activity in HSCs, resulting the in prevention of telomere attrition both in acquired SAA and in constitutional BMF syndromes (15, 115, 116). Some androgens have additional immunosuppressive mechanisms. For example, danazol inhibits interleukin-1 and TNF-α secretion by human monocytes in a dose-dependent manner (117). Several studies, including one randomized trial, showed the benefit of adding androgens to IST (34, 51, 118). Combined therapy with an immunosuppressive agent, ELT, and androgens in refractory SAA was also reported, with an ORR of 42% (119). The recommended doses for androgens in SAA are 2.5 mg/kg/d for oxymetholone and methenolone and 1 mg/kg/d for methandrostenolone and norethandrolone (46, 120).

Alemtuzumab (a humanized monoclonal antibody directed against CD52 protein) is one of the lymphocytotoxic serotherapeutic options for the treatment of autoimmune diseases. However, in SAA the best results for alemtuzumab were obtained in the relapse and refractory settings, with an ORR of 56% and a 3-year OS of 86% (121).

AlloHCT in AA is associated with several challenges that have been overcome during the last 40 years. The major concern in AA was the high risk of graft rejection, which was related to the presence of stem cell–eliminating T-lymphocytes and high sensitization towards HLA antigens in heavily transfused patients. The benefit of nonmyeloablative but intensive immunosuppressive transplant protocols with hATG and cyclophosphamide (Cy) was proved in transplantations from MSD in the 1990s (122, 123). Similarly, Cy with rATG-based serotherapy showed very good outcomes with high engraftment rates and mortality associated mostly with invasive fungal infections (124). Improved survival after alloHCT from MUDs was associated with improvements in donor typing, less toxic conditioning regimens with omission of total body irradiation (TBI), and use of leukodepleted blood products (125–127). Advances in supportive care with the availability of new antifungals and effective management of viral infections were pivotal in the reduction of transplant-related mortality.

One of the unresolved controversies in the field of pediatric AA is the problem of upfront HCT from MUDs. Some of the previous evidence supports the benefit of an upfront MUD-HCT in SAA, but randomized clinical trials are needed to confirm such an approach. The analysis of studies conducted between 1998 and 2019 that compared the outcomes of alloHCT and IST as an upfront therapy in SAA showed improved OS in upfront transplant recipients (128). In another study, the outcomes of upfront MUD-HCT in pediatric SAA were similar to those observed for MSD-HCT and superior to those for IST and MUD-HCT performed after IST failure (129). Conducting a randomized controlled trial in the setting of SAA poses a significant challenge. The North American Pediatric Aplastic Anemia Consortium and the Pediatric Transplantation and Cellular Therapy Consortium presented the results of a pilot study randomizing patients to upfront MUD-HCT or IST (130). In this study, 23 of the 57 patients with confirmed SAA underwent randomization and received therapy with a median follow-up of 18 months. This can be seen as a promising herald of high-quality evidence for the upfront therapy of SAA.

Transplantation options in adult patients with AA include “upfront” or “delayed” alloHCT after failed IST. The outcomes of alloHCT in AA improved in the last decade with long-lasting OS reaching 70% to 90% (131). On the other hand, despite the improved response to the new immunosuppression protocols with hATG and ELT, some patients are not cured or remain at risk of aplasia relapse or clonal evolution (33). About 3 months after successful IST, in about 19% of patients, chromosomal aberrations, including chromosome 7, were detected. There is also a significant risk of progression to MDS or AML within a median time of 4 to 6 years (27, 31, 55, 103, 132). Therefore, according to the available evidence, the decision to proceed to alloHCT should be based on the benefit-risk analysis of IST vs alloHCT. Accessible factors increasing the probability of response to IST and the risk of clonal evolution should be considered in decision-making.

Evaluated biomarkers potentially predictive of better response to IST are as follows: baseline absolute lymphocyte count ≥1 × 10⁹/L; absolute reticulocyte count ≥25 × 10⁹/L; PIGA, BCOR, and BCORL mutations. There are also potential biomarker candidates requiring clinical validation: predominance of memory/activated phenotype of T regulatory cells; increased intracellular IFN-γ levels in circulating T cells; IFN-γ +874T/A gene polymorphism; bone marrow enrichment of CD8⁺ T cells and T cell–activating intercellular interactions; and miR-150-5p expression (31, 133–138). Poor response to IST might be expected with an undetectable PNH clone and a shorter telomere length, and detection of adverse genetic mutations: ASXL1, DNMT3A, RUNX1, TP53, but also with TPO plasma levels of >1,796.7 pg/ml (32, 139, 140). Telomere shortening and the above adverse mutations have also unfavorable impact on relapse incidence and clonal evolution (14, 140). If germline mutations were detected, the family history should be carefully collected to determine the malignant hematologic disease germline predisposition because for these patients alloHCT is usually needed.

Factors determining the choice of the therapeutic pathway include the patient’s age, comorbidities, availability of an optimal donor, AA severity at diagnosis as well as access to key medical agents effective in nontransplant therapies: hATG and ELT (26). In a study published by the EBMT, the identified risk factors for alloHCT that should be considered were as follows: PB as the source of hematopoietic cells, time interval to alloHCT exceeding 6 months from diagnosis, patient’s age ≥20 years, no ATG use in conditioning regimen, and the cytomegalovirus (CMV) donor/recipient serostatus other than negative. The presence of risk factors determined inferior survival assessed as 77% and 67% in the intermediate and high-risk groups, respectively, compared with 90% in the low-risk group (141). However, the last risk factor is of a lower value due to a broader access to letermovir prophylaxis in CMV seropositive patients. In another study conducted in patients who underwent transplantation with the use of fludarabine (Flu), the risk factors leading to worse OS included age ≥20 years, shorter time interval (3 months), transfusion burden (>20 units of packed red cells, >50 platelet concentrates), and previous alloHCT. However, in a multivariate analysis, all factors but age were associated with poor survival prognosis (142). Therefore, with the use of Flu, age as a sole risk factor should not be treated as an absolute contraindication to alloHCT. In the study by Sheth et al., the outcome of patients transplanted with FCC conditioning protocol (Flu, Cy, and alemtuzumab) was summarized, and no significant differences were found between patients younger and older than 50 years, but alloHCT outcomes were much worse for patients with a HCT–comorbidity index score (HCT-CI) higher than 3 (143).

For adult patients at the age of 40 years or younger without significant comorbidities, alloHCT from MSDs remains the best option. For older patients, due to higher mortality rates, IST is recommended (60, 144, 145). Nevertheless, some recently published papers showed no significant differences in OS between the age groups when conditioning regimens are based on Flu (143, 146). However, it should be noted that the risk of complications related to graft-versus-host disease (GVHD) is much higher in elderly patients; therefore, a decision to perform upfront alloHCT from MSD in selected patients should be individualized (147). The transplant pathway in the older age group may be recommended for patients with a good performance status and a low HCT-CI in the case of SAA or vSAA and a lack of factors increasing the probability of response to IST (26). Considering disease severity, according to the published data, patients with SAA and vSAA more rarely respond to IST compared with those with nonsevere (moderate) AA. Moreover, they more often experience life-threatening infections and bleeding complications (133, 148).

In the last decade, we observe improved outcomes of alloHCT from MUDs, also in patients with AA (149). It may be explained by better high-resolution donor typing, improved supportive care, and the use of Flu and rATG or alemtuzumab in conditioning protocols with the avoidance of higher doses of TBI and PB as a source of hematopoietic cells (26, 34). However, transplant centers without extensive experience with alloHCT from MUDs and ADs should transfer patients to more experienced centers if such a procedure is planned.

For patients younger than 40 years without MSD, IST is the preferred frontline option; however, in young adults (<20 years) with urgent indications to alloHCT (e.g. recurrent infections) and an available MUD within 8 to 12 weeks, serious consideration should be given to this option (26). In retrospective analyses, the results of alloHCT from MSDs and MUDs were comparable in the population of children and teenagers (129). Another factor to determine opting for upfront alloHCT from MUDs would be the detection of a genetically unfavorable clone increasing the risk of progression to MDS or AML. Early alloHCT from MUDs is the option in patients older than 40 years with failed IST, especially in the case of clonal evolution. However, in the older age group, worse outcomes can be expected, including lower OS and disease-free survival resulting from transplant-related mortality, GVHD, and graft failure (GF) (1, 60, 61). The studies on transplantation techniques and alloHCT outcomes for the treatment of SAA are summarized in Table 2 .

The patient’s workup before alloHCT should include the assessment of the HCT-CI score, collection of data on previous and current infections (including tuberculosis, viral and fungal infections), and laboratory screening for latent infections. For transfusions, blood products must be irradiated and leukodepleted to avoid alloimmunization against HLA antigens. In patients with significant transfusion burden, serum ferritin should be evaluated. To decrease the risk of organ toxicity, infections, and GF, iron chelation therapy should be considered if the ferritin concentration is higher than 1000 ng/ml.

Fertility preservation issues should be discussed with patients when potentially gonadotoxic conditioning regimen is planned (26, 161–165). The severity of conditioning toxicities can be calculated using a cyclophosphamide equivalent dose (CED) (166). The risk of infertility depends on age and summarized exposure to gonadotoxic chemotherapeutics and radiotherapy (162, 163). The risk of irreversible gonadotoxicity is very high in men receiving a CED higher than 4000 mg/m² and in women receiving a CED higher than 6000 mg/m² (163–165, 167),.

The choice of conditioning regimen in pediatric SAA has evolved over the last 40 years. Among other factors, the omission of initially popular irradiation techniques was associated with progress of HCT and reduced complication rates in this population (168, 169).

Another pivotal step forward in the conditioning was related to the addition of Flu to the MUD-HCT (170, 171). The introduction of Flu into the conditioning regimens of adults with SAA in the last 2 decades was associated with improved outcomes (172). The reduction in the intensity of the conditioning protocol is associated with improved survival but graft-versus-host-free/relapse-free survival (GRFS) might not exceed 80% (152, 153). This issue is especially important in the pediatric population due to long life expectancy of posttransplant children, and the risk of GVHD-associated organ complications and metabolic consequences of life-long steroid immunosuppression that can arise (153).

Twenty years ago, Flu-Cy-ATG greatly improved the outcome of MUD-HCT in children with SAA (173, 174). In the case of MUD-HCT, the optimal conditioning protocol was not evident, but the 5-year OS probability did not exceed 80%, which was inferior to the newer results observed in children. In a retrospective study by the Pediatric Hematopoietic Stem Cell Transplant Working Group of the Brazilian Bone Marrow Transplantation Society and the Brazil-Seattle Consortium (Gedeco), the 4-year OS for MRD-HCT was lower than expected and did not differ significantly from that for MUD-HCT (82% vs. 69%, respectively; P = 0.08). The authors highlighted the need for standardized pediatric protocols (154).

Despite the existing variability of transplant protocols in SAA, in 2023 most studies agree that the Flu and Cy chemotherapy backbone is the standard of care for pediatric population. The original Cy dose of 200 mg/kg used in SAA for MSD-HCT can be safely decreased to 100 mg/kg (or 3000 mg/m²) in combination with Flu, which results in improved survival, as showed by numerous studies (152, 153, 175). However, at the lower end of the Cy dosing range, in children receiving the dose of 60 mg/kg, an increased risk of graft rejection was observed (176).

A retrospective cohort of 56 children with SAA after Flu-Cy-ATG conditioning showed favorable outcomes (153). The overall incidence of acute GVHD (aGVHD) was 41.5%, and that of grade III-IV aGVHD – 14.3%. Chronic GVHD (cGVHD) was diagnosed in 14.2% of children. The probability of the 2-year GVHD-free survival was 76.1%. In the univariate analysis, a higher dose of Cy and previous IST were significant risk factors for worse OS. Episodes of viral replication occurred in 33 of 56 patients (58.9%), but it did not influence OS. The study showed improvement over the previous treosulfan-based protocol for MUD-HCT (177).

In a Korean study, Yoo et al. described a low rate of secondary GF, higher failure-free survival (FFS), and manageable GVHD regardless of HLA compatibility (155). In 134 children with SAA transplanted with HLA-matched BM transplantation (BMT; n = 24), HLA-matched PB stem cell transplantation (PBSCT; n = 66), and HLA-mismatched PBSCT (n = 44), the estimated 5-year OS, FFS, and GVHD-free FFS of the total cohort were 93.0%, 89.5%, and 77.5%, respectively (155). The incidence of grade II-IV aGVHD was significantly higher in PBSCT than in BMT, but the incidence of grade III-IV aGVHD and cGVHD was similar across all the 3 groups.

In 2023, Vissers et al. described a single-center experience with 228 HCT procedures in pediatric BMF (n = 103) and SAA (n = 125) (178). Patients with SAA transplanted after 2017 had a superior 5-year OS and event-free survival (EFS) of 97% and 85%, respectively, compared with 68% and 59%, respectively, in the cohort transplanted before 2017 (P = 0.0011 and P = 0.017). In the case of BMF, OS and EFS were 89% for those transplanted after 2017 compared with 62% and 59%, respectively, for the other cohort (P = nonsignificant). The authors noticed the typical evolution of the conditioning regimen in SAA, with a change from more toxic to an immune-ablative regimen consisting of Flu/Cy without radiation.

A recent analysis of GRFS by the EBMT-SAA-WP included 479 patients with idiopathic SAA. The probabilities of 5-year GFRS in patients transplanted upfront or in relapsed/refractory cohorts were 77% and 61%, respectively (156). Approximately one-third of all patients were below the age of 20 years, and the group consisted mainly of children. Age was the main risk factor of death (hazard ratio [HR], 1.04; 95% CI, 1.02-1.06; P <0.001), aGvHD (HR, 1.03; 95% CI, 1.00-1.07; P = 0.041), and cGvHD (HR, 1.04; 95% CI, 1.01-1.08; P = 0.032) as the causes of inferior GRFS. The use of ATG/alemtuzumab was associated with a reduced risk of aGvHD (yes vs no: 4% vs 16%; P = 0.006), and the use of TBI – with a reduced risk of GF (yes vs no: 4% vs 11%; P = 0.039). There were no significant differences in GRFS and in the causes of GRFS failure according to graft source, donor type, or time from diagnosis to alloHCT. These results suggest that although cured in terms of hematopoiesis recovery, many patients can suffer from detrimental health issues due to long-term sequelae.

The use of MSD-HCT after conditioning with low-dose Cy (80 mg/kg), Flu (175 mg/m²), and Thymoglobulin (10 mg/kg) was studied in 32 children with BMF below 14 years of age. While the treatment led to primary engraftment in all patients, 2 patients later developed GF or MDS and only 7 patients sustained donor chimerism above 99% (179). The study was heterogenous and contained 9 patients with constitutional BMF (two ERCC6L2, two ANKRD26, two TINF2, one LZTFL1, one RTEL1, and one DNAJC21), for which Flu/Cy–based conditioning is not optimal.

The chemotherapy backbone recommended by the EWOG for alloHCT in children with SAA consists of Flu (30 mg/m²/day, × 4 days) and Cy (25 mg/kg/day, × 4 days) (59). Some small studies indicated that other reduced toxicity protocols, such as Flu with melphalan ± thiotepa, might be used in BMF syndromes (180, 181).

In adult patients, the choice of conditioning regimen is based on the patient’s age and comorbidities, a donor type, evaluation of the risk for graft rejection, access to alemtuzumab, and local experience (26). In the latest evaluation of 955 SAA patients transplanted from HLA-MSD, Flu/Cy/ATG, and Cy/ATG conditioning regimens were associated with the highest survival rates, as compared with Cy ± Flu and busulfan/Cy, with the 5-year probabilities of survival equal to 91%, 91%, 80%, and 84%, respectively (P = 0.001). In the case of 409 recipients of 8/8 and 7/8 HLA allele-matched unrelated donor transplantation (n = 409), the 5-year probabilities of survival with Cy/ATG/TBI-2Gy, Flu/Cy/ATG/TBI-2Gy, Flu/Cy/ATG, and Cy/ATG were 77%, 80%, 75%, and 72%, respectively (P = 0.61) (182). In the MSD setting, Cy-based protocols (200 mg/kg) are recommended in patients younger than 30 years, while in older age groups, Flu (120-150 mg/m²) plus Cy (120 mg/kg) are preferred, with the addition of serotherapy in both cases (26, 142, 151, 182, 183). However, in young adult patients at risk of cardiotoxicity, a Flu/Cy-based regimens should be considered. In the 10/10 MUD and 9/10 mismatched unrelated donor settings, low-dose TBI of 2 Gy should be added to the Flu/Cy/serotherapy backbone to decrease the risk of graft rejection (26, 61). In the haploidentical (haploHCT) setting, an increase in the TBI dose to 4-6 Gy may be considered to achieve sustained engraftment (184–186).

The choice of transplant material in alloHCT for SAA is under debate, and resource availability can affect the choice of stem cell source. Bone marrow is considered as the stem cell source for the first alloHCT in SAA because of higher survival probability and lower GVHD incidence (141, 187, 188). In SAA patients, the use of PBSC in MSD was associated with a high incidence of cGVHD of up to 43% (189). The retrospective analysis of 1886 patients with SAA by Bacigalupo et al. showed that BM was superior to PBSC in all age groups and was associated with a lower incidence of aGVHD and cGVHD as well as lower mortality rates (190). However, for technical reasons, the popularity of alloHCT from PB has increased in the recent years (129, 191). In a study by Kumar et al., a multivariate analysis showed the highest OS with BM as graft source in high-income countries compared with PBSCs in all countries or BM in not-high-income countries (P <0.001) (192). The authors concluded that PBSCs may be an acceptable alternative in countries with limited resources when treating patients at high risk of GF and infective complications (193). The transplant outcome in MSD and MUD-HCT for SAA is also improved by the shorter interval from diagnosis to transplantation (127, 128, 141, 190, 194, 195).

To reduce the complication rates, there is a search for ways to improve GVHD prophylaxis. The ex-vivo immunomagnetic T-cell depletion of MSD PBPC is challenging and requires a specialized graft-engineering facility, but it is associated with faster neutrophil engraftment, lower aGVHD incidence, and similar OS as the standard BM (157). However, the T-cell depletion techniques in haploHCT for SAA are associated with an increased risk of GF (196). With the increased data on favorable outcomes with posttransplant Cy (PTCy) as GVHD prophylaxis not only in haploHCT, but also in the MUD and mMUD settings, this option should also be considered (197).

In adult patients, BM is a recommended source of hematopoietic cells in the MSD and MUD settings, but PB might be accepted when alemtuzumab is used (141, 143, 187, 190). The cellularity of the BM graft should be higher than in hematologic malignancies, with the minimal quantity established as 3 × 10⁶ CD34/kg (3 × 10⁸ total nucleated cells [TNC]/kg) (61). In haploHCT, there are no clear recommendations concerning the source and an optimal dose of cells, with a suggestion for a higher dose to diminish the risk of GF. There are different protocols including the preferential use of BM or G-CSF-mobilized BM and PB (186, 198). It must be noted that in the EBMT study no difference in GVHD-free survival was reported when PB was used in haploHCT, with superior platelet engraftment and lower GF rates with a medium graft cellularity of 3.5 × 10⁸ TNC/kg for BM and 7.1 × 10⁶ CD34/kg for PB (158). An option to perform syngeneic HCT should be used irrespective of the patient’s age because of the very good outcomes (long-term survival of about 90%), and in this case, PB is the preferable HSC source (199).

The conditioning regimen including serotherapy with rATG, hATG, or alemtuzumab in SAA was shown to be associated with a survival advantage (150, 200). The inclusion of serotherapy with rATG is associated with both reduced risk of cGVHD and increased risk of opportunistic infections in alloHCT recipients (201). In some studies, the benefit of ATG in MSD-HCT was unclear. A multicenter prospective trial by the Center for International Blood & Marrow Transplant Research reported a cGVHD incidence of 21% and 32% in patients conditioned with either Cy alone or Cy/ATG, respectively, while the OS was 74% and 80%, respectively (200). Compared with hATG, rATG was associated with a lower risk of grade II to IV aGVHD (HR, 0.39; P <0.001) but not of cGVHD (182). In a study by Atta et al., rATG was protective against aGVHD and cGVHD, but a higher incidence of invasive fungal disease and mixed chimerism was noted in the rATG vs hATG cohorts (202). In another study, 21 patients with SAA, including children, underwent BMT from MSDs following a standard conditioning regimen with Cy (200 mg/kg) and hATG (30 mg/kg/day × 3 days) (203). None of the patients rejected the graft, all had sustained engraftment, and cGVHD developed in 24% of patients. The currently recommended cumulative doses of rATG in children undergoing HCT are 40 to 60 mg/kg for Grafalon and 8 to 10 mg/kg for Thymoglobulin, depending on the donor and grafting material (59).

A strategy to reduce the incidence of GVHD in SAA is the replacement of ATG with alemtuzumab (150, 204, 205). The protocol containing Flu (120 mg/m²) and Cy (120 mg/m²) with alemtuzumab (40-100 mg) was studied in 50 SAA patients transplanted from MSDs (n = 21) or MUDs (n = 29) (206). Acute GVHD was observed only in 13.5% of patients (all grade I-II), and only 2 patients (4%) developed cGVHD, but GF was present in 9.5% of patients after MSD and 14.5% of those after MUD.

In adult patients, the standard total doses of rATG for SAA in MSD and MUD-HCT are 5 to 10 mg/kg (Thymoglobulin) or 60 mg/kg (Grafalon), but in patients at high risk of graft rejection, a higher Thymoglobulin dose of 10-20 mg/kg might be considered (26, 142). A total dose of alemtuzumab ranges from 40 to 100 mg, and usually 0.2 mg/kg/day is administered for 5 consecutive days (151, 206). In patients referred for alloHCT with alemtuzumab with a high PNH clone associated with SAA, a lowered CD52 expression is possible on the PNH clone-derived lymphocytes as CD52 is a glycosylphosphatidylinositol-anchored protein. Despite theoretical concerns there have been no reports of increased graft rejection or graft failure rates in AA patients with PNH clones using alemtuzumab-containing regimens (207).

Over the last 20 years, developments in haploHCT with ex-vivo depletion techniques or PTCy gained a lot of attention in multiple indications, including SAA. Nevertheless, SAA HCT protocols should acknowledge the fact that Cy induces cardiotoxicity in 7% to 28% of patients with a mortality rate of 11% to 43% at the therapeutic dose of 170 to 180 mg/kg intravenously (IV). Therefore, caution should be exerted when increasing the cumulative (pre-HCT and post-HCT) Cy dose (208). In patients with a significant risk of cardiotoxicity, a Flu/Cy-based conditioning regimen should be considered. There are also data on the reduction of a standard total PTCy dose of 100 mg/kg to 80 mg/kg (209).

Initial reports on haploHCT showed an engraftment rate of 67% and a 2-year OS of 78% in 33 patients (39% children) transplanted between 2011 and 2017 with PTCy (158). The outcomes of related haploidentical donors after reduced-intensity conditioning with PTCy were presented in 37 patients with SAA: 20 relapsed/refractory and 17 treatment-naïve (186). The 2-year OS probability was 94%, and the cumulative incidence of grade II-IV aGVHD at day 100 and cGVHD at 2 years was 11% and 8%, respectively. The GF incidence was 11% including 1 case in a relapsed/refractory group and 3 in a treatment-naïve patients receiving 2 Gy of TBI while no GF was observed in patients who received 4 Gy of TBI. The role of irradiation cannot be generalized because low-dose TBI/total lymphatic irradiation reduces the risk of GF in adult SAA and is deemed safe, but in children the irradiation techniques, even at a low dose, are not recommended due to excellent transplant outcomes in pediatric SAA and historic reports showing the risk of malignancies in the radiation field (59, 127, 156, 210, 211).

Haploidentical HCT with CD45RA (4 patients), TCRαβ-CD19 depletion or PTCy (5 patients) was evaluated in 11 patients with SAA or MDS-refractory cytopenia of childhood (212). The conditioning regimens were heterogenous and consisted of Flu combined with irradiation techniques (9), Cy (9), busulfan, treosulfan, or thiotepa. Despite engraftment, 2 patients died, and 4 of 9 long-term survivors developed cGVHD.

The clinical protocols for alloHCT in SAA in some countries show fundamental differences, especially by including a high number of partially matched donors and more myeloablative conditioning protocols. Therefore, a direct comparison is not feasible. Older studies that did not implement PTCy after haploHCT revealed an elevated incidence of aGVHD and worse OS probability (213–215). In a study by Qin X et al., in an analysis of 115 patients (mostly children), a conditioning regimen without low-dose irradiation or busulfan was associated with an inferior FFS in a multivariate analysis, but 60% of donors were partially matched in HLA (216). In an Indian study, children and young adults (aged 21 years or younger) with SAA underwent haploHCT after Flu-Cy-ATG/TBI with or without thiotepa/melphalan and PTCy (217). At a median follow-up of 48 months, the OS and EFS were 61.6% and 58.1%, respectively. Chinese protocols in pediatric SAA contain a more intensive chemotherapy backbone, such as Flu (200 mg/m²), Cy (30-100 mg/kg), and busulfan (6.4 mg/kg) (160). In a study by Gong et al., 61 of the 71 patients with SAA below the age of 20 years were transplanted from haploidentical donors (160). At a median follow-up of 42 months (95% CI, 33-50 months), the OS and EFS were 93.0% (95% CI, 85.3%-97.3%) and 81.7% (95% CI, 71.5-89.3%), respectively. A more intensive myeloablative protocol resulted in complete donor chimerism in all long-term surviving HCT recipients. In addition, the frequencies of grade II–IV aGVHD, overall cGVHD, and moderate-severe cGVHD were 4.5% (95% CI, 1.3-11.6), 17.2% (95% CI, 9.5-27.8), and 3.1% (95% CI, 0.7-9.6), respectively.

In adult patients, retrospective analyses show minor differences between the outcomes of alloHCT from MUDs and MSDs compared with those of haploidentical donors with the use of PTCy as GVHD prophylaxis. It was suggested that haploHCT should be reserved for younger patients without MSD and MUD after an unsuccessful outcome of IST (34, 104, 218). However, the recent study in mixed age cohort showed encouraging outcomes with haploBMT approach in SAA refractory to IST or relapsed after initial response with 1-year OS 81% and the PTCy-based GVHD prophylaxis. The authors concluded that haploBMT with PTCy could now be considered a standard approach for salvage treatment of SAA (219). Moreover, another prospective study by DeZern et al. enrolling 37 SAA patients with median age 25 (range, 3-63 years) to haploBMT as an upfront approach with PTCy showed OS of 92% at 1, 2 and 3 years with minimal GVHD, while in last 20 consecutive patients receiving a TBI dose increased from 2 Gy to 4 Gy, OS of 100% was achieved (159).

Based on increasing available data, the haploBMT-PTCy platform as the salvage or even initial treatment is of special interest in pediatric and adult patients originating from the ethnical minorities but should be also considered as an alternative option to the MUD/mMUD platform when beneficial outcomes over IST are expected and a delayed alloHCT procedure poses a risk of life-threatening complications.

When donor-specific antibodies (DSA) are identified in pretransplant evaluation of the patient, the preferred haploidentical donor is without antigens corresponding to the detected DSA in the recipient. In the case of a lack of a donor choice and DSA above 1000 MFI, there are 2 options to prevent GF: the search for AD or desensitization of the recipient (220).

The role of cord blood (CB) in HCT for SAA remains to be determined, although the number of CB transplantations has been on the wane in recent years (221). Increased experience with haploHCT and suboptimal historic outcomes of CB transplantation in SAA suggest that this tendency is becoming stronger (222–224). According to the Chinese study, Haplo-BMT (35 patients) or haploidentical CB (56 patients) HCT based on the Beijing Protocol, with Flu at a dose of 45 mg/m²/day on days −7 to −4, busulfan at a dose of 3.2 mg/kg/day IV on days −7 to −6, Cy at a dose of 40 mg/kg/day on days −5 to −2, and rATG (Thymoglobulin) at a dose of 2.5 mg/kg/day IV on days −5 to −2, and GVHD prophylaxis with CsA/tacrolimus with MMF and methotrexate (MTX) was feasible in 91 children (225, 226). The Haplo and Haplo-CB groups had similar 3-year OS (92.8% ± 4.9% vs 94.3% ± 3.2, P = nonsignificant), with a lower incidence of aGVHD and cGVHD in the Haplo-CB HCT recipients. In a retrospective analysis of 240 children with SAA transplanted at single center, CB recipients had inferior FFS and GRFS compared with Haplo and MUD cohorts (227). Additionally, CB was associated with a higher GF rate and inferior platelet engraftment. In a study from 2021, HCT in 60 pediatric patients with SAA used unmanipulated haploHCT with third-party 1-5/6 HLA antigen–matched umbilical CB after myeloablative conditioning with busulfan, Cy, and Flu (228). The 5-year transplant-related mortality, OS, and FFS were 2.8%, 97.2%, and 96.2%, respectively. Considering that patients with SAA constituted only a subgroup of 109 patients in the study cohort, the detailed analysis was not provided. It is of note, however, that at +30 days after transplantation, the engrafted HSCs were derived from CB in 5 of the 109 patients. In unique cases, a successful autologous CB transplantation was reported in SAA, but there were differences regarding the use of conditioning therapy before the transplantation (229–232).

The use of CB for transplantation in adults is limited by graft cellularity. Longer time to engraftment is another concern. However, the published outcomes of CB transplantation (CBT) are promising, with an OS of 88% and a low rate of aGVHD and cGVHD of about 15% (233). The APCORD protocol (FluCyATG+2GyTBI) also showed very good results in a prospective trial (222). The recommended minimum TNC dose is 4 × 10⁷ TNC/kg. In a retrospective study of the Japanese Society for Transplantation and Cellular Therapy comparing haploHCT (n = 24) with PTCy vs CBT (n = 59) in adult SAA patients, no significant differences were noted in OS, the cumulative incidence of grade II-IV aGVHD, and cGVHD. However, patients after PTCy-haploHCT experienced significantly higher rates of engraftment in neutrophils (95.8% vs 78%; P <0.001) and platelets (83.3% vs 72.9%; P = 0.025). Moreover, a higher 1-year FFS rate was observed in patients younger than 40 years after haploHCT with PTCy compared with the CBT group (92.9% vs 63.9%; P = 0.047) (218).

The advantages of MUD-HCT include a short time to neutrophil engraftment and high survival rates, but they are counterbalanced by the need for donor search and arrangement of graft donation, which is associated with an up to 4-month delay in therapy. The waiting time for an HSC donor can be used as a window for pharmacotherapy involving IST with or without additional drugs (e.g. TPO-RA or androgens). In addition, HCT can result in treatment-related organ toxicities, infections, GF, poor graft function, acute and chronic GVHD, infertility, and secondary neoplasms, supporting the need for a judicious referral for HCT (234, 235).

The factor with the highest influence on the quality of life is GVHD, and GFRS is the best composite endpoint measuring the success of alloHCT. A risk of GF is another key issue after alloHCT in AA (236). Therefore, the suitable and often individualized posttransplant IST has the most important impact on alloHCT outcome.

Posttransplant GVHD prophylaxis in children typically consists of MTX and CsA, since Locatelli et al. proved this combination to be superior to CsA alone in a randomized controlled trial (237). Due to increased popularity and low cost, PTCy as the sole GVHD prophylaxis after PBSCT from MSDs can be considered, as it is associated with low rates of aGVHD and cGVHD (238).

In adult patients, in the MSD and MUD alloHCT platforms with rATG use, calcineurin inhibitors (CNI) with a short course of MTX is recommended, while alemtuzumab administration is followed by sole CNI-based IST. CNI should be continued for 9 to 12 months with CNI concentrations maintained on higher levels than in hematological malignancy to avoid GF (CsA, 200-300 ng/ml; tacrolimus, 10-15 ng/ml) and very careful tapering under the control of donor chimerism and evidence of GF (61). In haploHCT, there are protocols based on PTCy ± rATG with MMF and tacrolimus given after PTCy administration and the Chinese protocol based on rATG with CsA+MMF starting on day+1 and MTX administrations on days +1, +3, +6, and +11 (186, 239, 240). The APCORD protocol for umbilical CBT comprises CsA at a dose of 200-400 ng/ml for 3 months, with gradual tapering up to 1 year after alloHCT (222).

Donor chimerism is monitored at least 1, 3, 6, and 12 months after alloHCT for the assessment of engraftment kinetics and prediction of GF with progressing mixed chimerism that requires intensification of post-HCT IST (26). Transient mixed chimerism is a common finding after rATG and in the FCC protocol with alemtuzumab, stable mixed chimerism in T lymphocytes indicates that immunotolerance has been achieved (241). In the case of poor graft function with complete donor chimerism, ELT administration or an additional boost of hematopoietic cells without additional conditioning should be considered (242, 243).

In case of alloHCT failure in SAA, second transplants are associated with very high success rates, even with the retransplantation from an original donor (244). In 55 children who experienced GF after the first alloHCT, the 5-year OS and FFS were 82.9% and 81.2%, respectively (245).

The long-term survival of children with SAA after HCT is very high, and the OS probability at 5 and 10 years is 96% and 94%, respectively (246). Changes in the conditioning protocols and improved genetic testing over the last 40 years positively affect the survival outcomes, but caution is warranted. Despite excellent survival probability, the freedom from long-term complications in HCT survivors have not been determined (247). According to Buchbinder et al., the cumulative incidence of late effects among survivors of MSD-HCT was less than 3% and did not increase over time, but after MUD-HCT, the incidence of late effects exceeded 3% by 5 years: gonadal dysfunction, 10.5% (95% CI, 7.3-14.3); growth disturbance, 7.2% (95% CI, 4.4-10.7); avascular necrosis, 6.3% (95% CI, 3.6-9.7); hypothyroidism, 5.5% (95% CI, 2.8-9.0); and cataracts, 5.1% (95% CI, 2.9-8.0). However, the majority of these transplantations included irradiation as part of conditioning regimen (248).

Among the consequences of HCT, the incidence of subsequent malignancies should be considered, estimated at 1.1% among the SAA patients (249). Improved differential diagnosis and genetic testing reduced the misdiagnosis of clonal disorders as SAA and, consequently, the occurrence of myeloid malignancies after HCT (84, 250). It can be suggested that HCT results would further improve in the future due to eliminated or reduced exposure to TBI or alkylating agents. However, long-term surveillance is necessary, because the majority of nonleukemic malignancies developed at least 5 years after transplantation.