Survival after hematopoietic stem cell transplantation (HSCT) in acquired aplastic anemia (AA) has been remarkedly improved over the last three decades. Haploidentical (haplo-) HSCT provides easily available donors for patients with AA and guarantees a favorable engraftment rate (1–3). Stable hematopoietic recovery is the key point of successful HSCT for AA. However, hematologists have noticed that even if the patients achieved initial hematopoietic reconstitution and maintain complete donor-originated hematopoietic cells, they may develop intractable multilineage cytopenia afterwards (4, 5). This is defined as secondary poor graft function (sPGF), and it occurs in 5-27% of post-transplantation cases (6). Patients with sPGF lose their initial hematopoietic reconstitution, which leads to increased risks of severe infection, major bleeding events, and other life-threatening complications after transplantation. The therapeutic options for sPGF are limited, and the prognosis remains poor.

Studies have reported several risk factors for sPGF. Cytomegalovirus (CMV) reactivation and graft-versus-host disease (GvHD) are the two most recognized risk factors (6). Other potential risk factors include haplo-HSCT setting, recipient age, conditioning regimen, Epstein-Barr virus (EBV) infection, etc. (7–9) However, these conclusions were limited by heterogeneity in disease categories and transplantation settings. The clinical outcomes and risk factors of sPGF have not been elucidated in haplo-HSCT for AA patients. Based on the largest-scale AA cases receiving haplo-HSCT, we herein retrospectively analyzed the incidence, outcomes, and risk factors of sPGF.

PGF is a type of BM failure syndrome and leads to high morbidity and mortality post-transplantation. Based on the largest-scale AA cases that received haplo-HSCT, we revealed a CIR of 4.62% of sPGF at 3 years post-transplantation. OS was significantly decreased in sPGF patients. Later neutrophil engraftment and a history of refractory CMV viremia were the independent risk factors for sPGF.

Several studies on hematological malignancies suggested that haplo-HSCT can be associated with a greater risk of sPGF (9, 22). However, a previous report from our center demonstrated no association between haplo-HSCT setting and sPGF (23). In line with Liu et al. (24), our results reported an acceptable incidence of sPGF after haplo-HSCT for AA. In contrast, Japanese colleagues reported a higher incidence of sPGF of 15% in 49 pediatric AA patients, which included 3 transplantations from matched sibling donors (20.0%), 3 from unrelated donors (10.3%), and 1 from haploidentical donor (20%) (25). Similar to the study of Kako et al (8), Flu was suggested to be responsible for the increased incidence of sPGF, which was denied by our studies (11). The discrepancy in results can be explained by the difference in conditioning regimen instead of HSCT type, as Liu and we additionally applied 2 days of Bu to ensure successful engraftment and stable full donor chimerism in the haplo-setting (26). Of note, in line with prior reports, patients with high titers of DSA were successfully handled with rituximab desensitization (15, 27) and co-infusion with DSA-free cord blood (28), and as a result, none of the patient experienced primary PGF. Employing intensive conditioning regimens may be beneficial in maintaining stable donor-type chimerism and excellent hematopoietic recovery in AA patients undergoing HSCT. Further research should be conducted to clarify this hypothesis.

Widely proposed is the “seed, soil and climate” model for the pathophysiology of PGF (6, 29). Current literature suggests hematopoietic stem cells (seed) abnormalities have a causative role in PGF. In the present study, patients with sPGF or without PGF received grafts of similar dose and composition. Impressively, the dose of infused CD34+ cells was not associated with sPGF. However, we did observe distinct features of engraftment between the two groups. The univariable and multivariable analyses identified later neutrophil engraftment as an indicator of sPGF. Our results indicate that the development of sPGF is more likely the result of qualitative, rather than quantitative, abnormality of hematopoietic stem cells. Several studies demonstrated that no deficit was found in the cells’ capacity to repopulate the marrow when stored CD34+ cells from donors whose recipients developed PGF were xenografted to mice (30). Moreover, donor-derived CD34+ cells boost is an emerging therapeutic option with promising response rates in patients with sPGF, as presented in this report and others (31–33). Taken together with these findings, these data suggests that the deficits in “seed” are acquired after transplantation. Inducers or enhancers of allo-immunity (climate), such as CMV reactivation and GvHD, may amplify the intrinsic dysfunction of hematopoietic stem cells, ultimately leading to sPGF.

CMV inhibits hematopoiesis directly by infecting bone marrow or suppress hematopoiesis indirectly through the infection of stromal cells (34, 35). Previous studies (7, 23, 36) have identified CMV viremia as an independent risk factor for sPGF. It is reported that recipients undergoing haplo-HSCT have a higher incidence of CMV reactivation, as well as refractory CMV viremia (37, 38). Lv et al. recently revealed that CMV reactivation was the only hazard element for sPGF in haplo-HSCT (9). Nevertheless, one should note that refractory CMV viremia rather than early CMV reactivation increased the risk of sPGF in our study. Patients in the sPGF group had lower B cell levels in the first month after haplo-HSCT, which may explain their greater susceptibility to refractory CMV viremia (38). Our results indicated that the influence of CMV on the BM niche can be time-dependent and may be irreversible under sufficient viral load. Since the clinical course of CMV reactivation is often complicated with GvHD, administration of immunosuppressants and immune reconstitution, further study on the impact of clinical characteristics and kinetics of CMV on graft hematopoietic function is required. On the other hand, antiviral medications, including GCV and foscarnet, can exacerbate the suppression in hematopoiesis (22) and are major players in the development of sPGF. Given the fact that GCV was involved in all CMV-positive patients in this study, it was impossible to separate the influence of CMV itself and anti-CMV pharmacotherapy. Anyway, timely evaluation and initiation of CMV-specific cellular therapy may be of great help to avoid inhibition of hematopoiesis and provide better transplant outcomes (13, 39).

The occurrence of aGvHD has also been accepted as contributing to the development of sPGF (36, 40). In this study, we found a marginal association between grade 3-4 aGvHD and sPGF. In vivo studies corroborated that GvHD can lead to PGF via overactivated T cells and dysregulated cytokines (41, 42). Moreover, severe aGvHD requires intensive immunosuppression and thus always occurs in concert with prolonged viral reactivation. Limited by the number of cases, we were unable to analyze the effect of aGvHD and co-current CMV viremia on sPGF. Large-scale studies can help understand this process. Reducing the incidence of aGvHD is now one of the most important goals of unmanipulated haplo-HSCT for AA patients, but it is noteworthy that we need to develop better strategies to balance the risk of immunosuppression and viral reactivation. In the ATG-based modality of in vivo T cell depletion, the dose of ATG is positively related to delayed immune reconstitution and the risk of viral infection (37, 43, 44). Recent works provide a promising option by optimizing ATG dosing (2.5 mg/kg daily for 3 days) to reduce viral activation while maintaining sufficient GvHD prophylaxis in haplo-HSCT (45, 46). Similar phenomena were observed in haplo-HSCT using a combination of post-transplantation cyclophosphamide and low-dose ATG (47, 48). A lower dose of ATG (2.5 mg/kg) successfully reduces CMV reactivation without compromising the favorable effect of preventing GvHD (49). Studies should be conducted in larger cohorts to determine the optimal dose of ATG to reach maximum immunosuppression, minimum risk of severe infections and in the end the best survival.

In conclusion, sPGF can develop in 4.62% of AA patients after haplo-HSCT and significantly decreases survival. The independent hazard elements for sPGF were later neutrophil engraftment and a history of refractory CMV reactivation. Considering the limited number of sPGF cases in this report, our results warrant investigation in further studies.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the Ethics Review Committee of Peking University People’s Hospital. Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.

LX and XH designed the study. The data analysis and manuscript development were led by FL and TH. All authors contributed to providing clinical data and approved the final version for submission.

This work was supported by the National Key Research and Development Program of China (No. 2017YFA0104500), Innovative Research Groups of the National Natural Science Foundation of China (No. 81621001), Key Program of the National Natural Science Foundation of China (No. 81530046), and National Natural Science Foundation of China (No. 82100227).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.