MSD HSCT has been the standard first-line treatment for children with SAA/VSAA 10,17, with nearly 90% OS 18,19. In our retrospective study of 87 pediatric AA patients undergoing MSD peripheral blood stem cell transplantation, we noted 98% OS. Two patients experienced GR, 4 PGF, and 2 died of infection, resulting in an FFS of 91%, which is comparable to previous Japanese reports 20,21. Even with PBSC, the incidences of aGVHD and cGVHD in this cohort were only 9% and 15%, respectively, and all were controlled by IS, reaching a GFFS of 89%. This result is in line with previous reports of ex vivo T cell-depleted PBSC or bone marrow MSD HSCT 22,23. Such a high survival rate can be attributed to an intensified conditioning regimen and timely MC intervention. To the best of our knowledge, this is the first study to explore the EW IS strategy for early-onset MC in AA HSCT.
In 2010, EBMT SAAWP recommended the combination of CY and ATG as a standard conditioning regimen for MSD HSCT 8,24. In the early clinical practice of our center and other studies, CY + ATG was found to be insufficient to obtain engraftment. FLU is considered one of the choices to supplement CY + ATG to improve stable engraftment 10,25. To decrease GR, FLU 150 mg/m2 was routinely used to intensify the CY + ATG conditioning regimen for MSD HSCT at our center. Due to concerns about the cardiotoxicity 26, for patients with heavy transfusion or potential cardiac insufficiency before HSCT, we tried to increase the dose of FLU to 180–200 mg/m2 and tapered CY to 120–150 mg/kg. With this intensified conditioning regimen, the rate of MC in this study still reached 26%. We reported an FCA + TBI/BU based conditioning regimen for alternated donor HSCT for AA in 2021 3. Until now, according to the severity of bone marrow failure, the number of transfusion, and degree of HLA matching, the risk-adapted conditioning regimen has been performed in 580 pediatric AA HSCT patients with a complete engraftment rate of 98% (data not shown). Thus, FLU is necessary to intensify CY + ATG and it is important to maintain sufficient engraftment for pediatric acute onset SAA with MSD HSCT.
This retrospective study shows that low-dose CY (CY 120–150 mg/kg) is an independent risk factor for MC (P = 0.0002), whereas increase dosage of FLU did not decrease the MC. Although some studies have shown that CY 150 mg/kg has higher mortality than CY 50 mg/kg or 100 mg/kg combined with FLU + ATG + TBI in old patients 27,28. CY 200 mg/kg is still the main option for pediatric AA 10. In this group, CY 200 mg/kg was used in 38 (57%) patients, with slightly lower than 8% MC and no patient developed cardiac dysfunction. Two days of BU was added to further intensify FCA conditioning for 14 patients, though half of the TD-NSAA patients still developed MC. Even with DLI, three patients exhibited SMC finally (cases 10, 14, and 20). For case 10, who had a 4-year history of TD-NSAA and heavily transfused before HSCT, more intensified conditioning with CY 200 mg/kg + FLU 200 mg/m2 + BU (2 days) was used, but she still had early-onset MC. Even with active DLI, the valley chimerism fluctuated approximately 69% with SMC to the present (post-HSCT 5 years). Although future studies with larger cohorts are needed, we recommend 200 mg/kg CY + 150 mg/m2 FLU as a standard conditioning regimen for pediatric SAA patients undergoing MSD HSCT.
MC is linked to GR and treatment failure in AA 29–31. Compared to late-onset MC, early-onset MC has a greater risk of GR 32. For patients with early-onset MC, reducing IS after non-myeloablative HSCT is a feasible and safe strategy for prophylaxis relapse 33,34. However, EBMT guidelines suggested not reducing or withdrawing IS to protect against later-onset MC 1. This suggestion was based on little clinical data 11. Coexistence of donor and recipient hematopoietic cells in the recipient’s body imply the existence of bidirectional immune tolerance. Until now, how to manage IS for early-onset MC has been unclear 35. In this study, we tried EW IS for 6 out of 17 early-onset MC patients. Except for one case of GR, five patients returned to CC in a median 63 days (range 36–91 days) and 2 completely stopped IS within 70 days post-HSCT (cases 6 and 7). Although most patients underwent DLI, the ratio of DLI was consistent in two cohorts. Thus, the EW IS strategy is helpful for quickly improving the CC transition, which will benefit stable engraftment (P = 0.008). Case 8 achieved CC by EW IS alone and more clearly reflects that early reduction of IS will benefit the recovery to CC. By contrast, Zhang reported that increased immunosuppressant intensity is only effective in 1 of 4 (25%) MC patients32. And our results also showed that timely monitoring and early reduction of IS is key to saving early MC. T cell chimerism is more sensitive than that of whole blood (P = 0.001), giving us more precise data to monitor the donor chimerism and detect MC early. In active EW IS and DLI, 88% of patients recovered to CC (12/17), and only 12% (2/17) experienced GR, which is significantly better than the 30% GR reported by Zhang 36,37. Active intervention and T cell chimerism monitor for early-onset MC achieved an OS of 100%.
Although EW IS and DLI may increase the risk of aGVHD 36,37, these iatrogenic GVHD cases are relatively easy to control. In this EW IS cohort, 50% (3/6) experience aGVHD, two cases related to EW IS and DLI, and one related to EW IS alone. Finally, all GVHD cases resolved, and only case 16 experienced grade III aGVHD and moderate cGVHD. The overall responsiveness to GVHD treatment did not seem to be affected in the patients who developed aGVHD after EW IS 33,34. This may be because the donor's immunity does not prevail in the MC state.
This study has several limitations. This retrospective study was based on real-world treatment data for pediatric AA patients. Thus, the design of the conditioning regimen lacked prospective considerations. The dose of ATG may affect engraftment. In addition, DLI also affects the MC outcome. To find the best treatment for patients, future prospective studies with larger cohorts are needed.