From May 1, 1991 to December 31, 2018, 43 patients underwent allogeneic HSCTs at the Departments of Pediatrics, Chonnam National University Hospital (CNUH) and Chonnam National University Hwasun Hospital (CNUHH). And their medical records were reviewed retrospectively. Baseline clinicopathologic data from patients, such as age, sex, initial blood counts, cellularity of BM, and etiology were recorded. Characteristics regarding to transplantation, such as, donor type, stem cell source, the degree of HLA matching, interval from diagnosis to transplantation, number of transfusions before transplantation, the presence of preceding IST, conditioning regimen, GvHD prophylaxis, engraftment, chimerism status, the presence of major viral infections, the presence of acute and chronic GvHD, transfusion-dependency, the need for second transplantation, and survival were analyzed.

All HSCTs were performed in patients with SAA or very severe aplastic anemia (VSAA). The diagnosis of SAA required at least two of the following criteria: an absolute neutrophil count (ANC) of less than 0.5 × 10⁹/L, a platelet count (PLT) of less than 20 × 10⁹/L, or a corrected reticulocyte count of less than 1% in the presence of hypocellular BM.12 VSAA was reserved to patients when the neutrophil count was below 0.2 × 10⁹/L.1 Chromosome breakage test was performed to exclude congenital/inherited AA. Patients with clonal cytogenetic anomalies were excluded in the study population.

The conditioning regimen has been modified over time based on the donor type and the degree of HLA match. Initially, patients were conditioned with intravenous cyclophosphamide 50 mg/kg/d given for 4 consecutive days with horse ATG (Atgam; Pfizer, New York, NY, USA) 30 mg/kg/day for 3 days.13 Additional procarbazine (12.5 mg/kg/day) was used in very early phase of the study. Since 2005, fludarabine (Flu)-based preparative regimen was used in most cases. The most prevalent preparative regimens in this study consisted of Flu 25 or 40 mg/m² and cyclophosphamide (Cy) 750 mg/m² for 4 days plus thoracoabdominal irradiation (TAI) 3 Gy the day before transplantation in matched related transplantation, and rabbit ATG (1.25 mg/kg/day for 4 days) was added in unrelated transplantation.

GvHD prophylaxis was continuous intravenous infusion of CsA (5 mg/kg/day) for matched related transplantation, or tacrolimus (0.03 mg/kg/day) plus short course of methotrexate (MTX) 15 mg/m² at D+1, 10 mg/m² at D+3, D+6, and D+11 for unrelated transplantation.

The patients were managed in a protective isolation room until the recovery of neutrophils. Immunoglobulin was administered every 3 weeks until 100 days after transplantation. Antifungals and pentamidine nebulizer were used to prevent infections. Donor chimerism was monitored at 1, 3, 6, 9, and 12 months after transplantation. Other routine measures of supportive care have been modified over the years. HLAs were initially serologically typed for Class I, but all HLA-A, -B, -C and -DRB1 were DNA typed since 2002.

For patients with SAA or VSAA frontline HSCT was offered if there were MRDs. Salvage unrelated HSCTs were offered to the refractory/relapsed cases without MRDs after IST, while MRD transplants were performed in small number of cases of IST failure who initially presented with moderate diseases.

The patients were subdivided into subgroups by transplant strategies, mainly by conditioning regimens. The survival was evaluated as of December 31, 2018. Patients were censored at the time of the events or at time of last follow-up.

Complete response (CR) after IST or transplantation was defined as simultaneous fulfillment of the following: hemoglobin > 10 g/dL, ANC > 1.5 × 10⁹/L and PLT > 100 × 10⁹/L with evidence of normal hematopoiesis. Partial response (PR) was defined when the blood counts no longer met the diagnostic criteria of SAA, but still did not meet the criteria for CR. Patients who did not achieve CR or PR, or were transfusion-dependent at 3 months after transplantation were defined as non-responders.13

OS was defined as the time between transplantation and death from any cause or time of last contact. FFS was calculated from the date of transplantation to last follow-up or first event (death due to any causes, lack of response, relapse, clinical paroxysmal nocturnal hemoglobinuria, secondary malignancy, HSCT after IST whichever occurred first).7 GFFS was defined as the absence of grade 3–4 acute GvHD or chronic GvHD with systemic treatment in addition to the status of FFS.9

Engraftment failure was defined by failure to achieve neutrophil engraftment, as evidenced by ANC < 0.5 × 10⁹/L on D+28 post-transplant.14 Graft failure after initial recovery of hematopoiesis was considered to be secondary graft failure. Donor-type aplasia was considered when patients had hypocellular BM with persistent cytopenia (hemoglobin < 8 g/L, neutrophils < 1.0 × 10⁹/L, or platelets < 50 × 10⁹/L) in full donor chimerism for > 6 months after achieving engraftment.15

Major acute complications including acute GvHD, cytomegalovirus (CMV) reactivation and disease, Epstein-Barr virus (EBV)-lymphoproliferative disease, hepatic veno-occlusive disease (VOD), and hemorrhagic cystitis after transplantation were documented and their incidences were compared among subgroups.

Continuous variables were compared using the Student's t-test and categorical variables were compared using the χ² test of Fisher's exact test. Probabilities of survival were estimated using the Kaplan-Meier (K-M) method and were compared using the log-rank test. Prognostic variables were evaluated by multivariable analyses using a Cox regression proportional hazard model. P value < 0.05 was considered statistically significant. The software package SPSS version 21.0 (SPSS Inc., Chicago, IL, USA) was used for all statistical analyses.

This study was conducted as a retrospective, observational, descriptive study of clinical aspects of pediatric acquired AA, which was approved by the Institutional Review Board of the CNUHH (CNUHH-2019-118). Informed consent was obtained from legal guardians of the patients.

A total of 43 acquired AA patients underwent allogeneic transplantations and were enrolled in this study. The characteristics of the patients involved in this study are shown in Table 1. Of these, 24 (55.8%) were females with male/female ratio of 1:1.26. Median age at diagnosis was 8.2 years, median age at transplantation was 11.3 years, and the median time from diagnosis to transplantation was 1.3 years (0.1–19.6 years). The average length of follow-up was 7.1 years. The etiology was idiopathic in all but 1 case of EBV infection. Paroxysmal nocturnal hemoglobinuria (PNH) clones (> 1%) at the time of diagnosis were seen in only one case, but the patient did not have clinical symptoms. At the time of transplantation, all patients were either SAA (n = 30, 69.8%) or VSAA (n = 13, 30.2%).

The details of transplantation characteristics, such as, type of donor, source of stem cells, median dose of harvested CD 34+ cells, conditioning regimen, GvHD prophylaxis and ABO incompatibility were described in Table 2. In related transplantations, all the donors were siblings. BM was the main source of donor cells throughout the study. There were two main backbones of conditioning regimens: Cy-based regimen followed by Flu-based one.

The 10-year K-M estimated OS, FFS, and GFFS for 43 allogeneic transplants were 86.0%, 60.5%, and 51.2%, respectively (Fig. 1). The OS, FFS and GFFS were compared between matched related and UDs (Fig. 2). MRD showed significantly better OS (96.4% vs. 66.7%; P = 0.006), FFS (71.4% vs. 40.0%; P = 0.014), and GFFS (60.7% vs. 33.3%; P = 0.031) than UD, respectively.

For stem cell source, BM was used in most cases (n = 37), while PB stem cells were used in 4 (all in unrelated settings), and umbilical cord in 2. The OS, FFS and GFFS of BM group were superior to those of PB, but not reaching the significant level, probably because of small numbers of PB group (OS, 89.2% vs. 75.0%, P = 0.468; FFS, 64.9% vs. 25.0%, P = 0.111; GFFS 54.1% vs. 25.0%, P = 0.171) in this study (Fig. 3).

With regards to conditioning regimen, the K-M 10-year OS for 14 patients who received Cy-based conditioning was 92.9%, while that of 29 patients with Flu-based conditioning was 82.1% (P = 0.367) (Fig. 4A). But Flu-based conditioning tended to have better FFS and GFFS than those with Cy-based conditioning (Fig. 4B and C). Table 2 also compared characteristics of transplants by conditioning regimens. In Cy-group, all the cases (n = 14) were MRDs, while only 41.4% (12/29) of cases were MRDs in Flu-group (P = 0.019). As a frontline treatment, transplantation was performed in 22 patients (51.2%) for whom all had MRDs, and there was no up-front transplantation using UD. The remaining 21 patients received IST using ATG and CsA as a frontline treatment, which included SAA cases without MRDs (n = 8) and non-severe AA cases evolving to SAA later (n = 13). More patients received previous IST in Flu-group than in Cy-group (62.0% vs. 21.4%; P = 0.012). Ten patients (23.3%) received more than 40 transfusions before transplantation without any significant difference between the two groups. Irradiation was incorporated in the conditioning more commonly in Flu-group (26/29, 89.7%) than in CY-group (3/14, 21.47%; P < 0.001), but there were no considerable differences in FFS (66.5% vs. 50.0%, P = 0.324) and GFFS (55.2% vs. 42.9%, P = 0.433) between irradiation (n = 29) and non-irradiation group (n = 14), respectively. ATG was frequently used in the conditioning regimen in both groups (P = 0.198). The use of ATG in the conditioning regimens had a tendency of lower survival rates in OS (82.1% vs. 93.3%, P = 0.320), FFS (53.6% vs. 73.3%, P = 0.194), and GFFS (50.0% vs. 53.3%, P = 0.726), but they were not statistically significant. Also, the use of ATG was not associated with survival within Cy-based and Flu-based conditioning subgroups (data not shown).

Fig. 5 illustrates the survival according to the donor type and the conditioning regimen. In related donor transplantation, OS was similar between Flu-group (92.9%, n = 14) and Cy-group (100.0%, n = 14), while FFS (100.0% vs. 42.9%; P = 0.001) (Fig. 5B) and GFFS (85.7% vs. 35.7%; P = 0.006) (Fig. 5C) were significantly better in Flu-group than in Cy-group. In UD transplantation (n = 15) including 2 cord blood transplants, all of whom received Flu-conditioning had OS of 66.7%, which was significantly lower than related donors with Flu-conditioning. However, the FFS and GFFS were quite similar between related donors with Cy-conditioning and UDs with Flu-conditioning.

The frontline transplantation group was superior to the frontline IST followed by transplantation group in OS, FFS, and GFFS (Fig. 6). Among them, FFS was statistically different (P = 0.041). Additionally, other factors such as the incorporation of irradiation in the conditioning, and method of GvHD prophylaxis (CsA- vs. tacrolimus-based) were compared for survival analyses, but they were not different.

Engraftment failure was seen in 13 patients with 30.2% of cumulative incidence: 8 after matched related transplant (8/28, 28.6%), and 5 after unrelated transplant (5/15, 33.3%; P = 0.746) (Fig. 7A). The median CD34+ cell doses of patients with stable engraftment and those with graft failure were not different (4.32 × 10⁶/kg vs. 3.99 × 10⁶/kg; P = 0.947). Primary graft failure was seen in only one patient, who was transplanted with 2 allele-mismatched BM of 2.3 × 10⁶/kg of CD34+ cell dose. Secondary graft failure after achieving initial engraftment was seen in 12. Among the secondary graft failure, 10 patients showed evidence of donor-type aplasia, which was seen in 13.8% (4/29) after Flu-based conditioning, and in 42.6% (6/14) after Cy-based (P = 0.035). As a rescue, a 2nd transplantation was performed in 8 cases, resulting in survival in all 8. However, a 3rd transplantation was needed in 2, resulting in complete response and transfusion dependence in each. Among 13 patients with engraftment failure, 2 patients died: one with primary graft failure died of CMV disease, and the other with secondary graft failure after mismatched transplant refused further treatment. For 10 patients with donor-type aplasia, 7 were salvaged by 2nd transplant, 1 responded to IST, 1 on thrombopoietin receptor agonist, and the last is still dependent on transfusions.

Major complications were documented in Table 3 and their cumulative incidences were shown in Fig. 7. The cumulative incidences of acute GvHD (≥ Grade II) was 9.3% (Fig. 7B), whereas chronic GvHD requiring systemic treatment was 18.6% (2 in Cy-group and 6 in Flu-group (14.3% vs. 20.7%; P = 0.613) (Fig. 7C). The use of ATG was not associated with the incidence of acute GvHD (P = 0.388) and chronic GvHD (P = 0.401) (data not shown). The cumulative incidence of CMV reactivation was 20.9% (9/43), but CMV diseases including pneumonia and encephalitis were seen in 3 (7.0%), which were associated with fatalities. They were all after UD transplants, and ATG was used in conditioning in 2 cases. EBV lymphoproliferative disease was observed in one patient with MUD BM transplantation, but was successfully treated with rituximab. The cumulative incidence of hepatic VOD was 14.0% (6/43; 3/14 in Cy-group versus 3/29 in Flu-group; P = 0.326). The median onset of hepatic VOD was 9 days after transplantation. In one case, hepatic VOD developed 25 days after the second PBSCT after graft failure, which was 150 days after the first BMT. Severe hemorrhagic cystitis was seen in one case each from two conditioning regimens.

The characteristics of deaths after transplantation were described in Table 4. Six patients died (1 in Cy-group and 5 in Flu-group; P = 0.371). Except for one case (Pt. #1) who died of late graft failure following matched related transplant, the remaining 5 cases were transplanted from donors other than related donors and were pre-treated with IST. Four deaths were attributed to pneumonia in which two of them were due to CMV pneumonia. CMV encephalitis with intracranial bleeding was documented in a case.

Prognostic factors were evaluated for OS, FFS, and GFFS by univariable and multivariable analyses (Table 5). Type of donor (odds ratio [OR], 10.975; 95% confidence interval [CI], 10.647–11.303; P = 0.029) and viral reactivation (OR, 10.266; 95% CI, 10.006-10.526; P = 0.008) after transplantation were found to be the independent factors in OS. For FFS and GFFS, number of prior transfusions and donor type were significant prognostic factors. The source of stem cells and the period of transplant before and after 2005 were not associated with OS, FFS, and GFFS (data not shown). Second transplantation was more commonly needed in patients conditioned with Cy (OR, 6.500; 95% CI, 6.256–6.744; P = 0.022). Likewise, donor-type aplasia was more frequently encountered in patients conditioned with Cy (OR, 4.687; 95% CI, 4.459–4.915; P = 0.043). Multivariable analyses showed that donor type was the prognostic factor in FFS, and GFFS. Viral reactivation in OS and the number of transfusions in FFS and GFFS were additional prognostic factors, respectively.