Allo-HSCT compared with immunosuppressive therapy for acquired aplastic anemia: a system review and meta-analysis

Acquired aplastic anemia (AA) is a rare hematologic disease characterized by a profound deficit of hematopoietic stem and progenitor cells, bone marrow hypocellularity, and peripheral blood pancytopenia. It mainly affects children, young adults, and those over 60 years of age. The estimated incidence rate of AA ranges from 0.7 to 4.1 per million population each year, and it appears to be two to three times higher in Asia than in Europe and North America [1, 2]. Although the occurrence of AA can be partly explained by some drugs, chemicals, viruses, and other external factors, the majority of cases are idiopathic [3]. The underlying pathophysiology is thought to be an aberrant autoimmune reaction involving the T-cell-mediated destruction of hematopoietic cells [4]. Major symptoms are infections, hemorrhage, and symptoms of anemia. Symptoms may be severe and life-threatening or minor enough to not require transfusion support. The survival rate for AA has markedly improved in the past four decades because of advances in hematopoietic stem cell transplantation (HSCT), immunosuppressive drugs, and supportive care.

According to the 2015 Guidelines for the Diagnosis and Management of AA (GDMAA) of the British Committee for Standards in Haematology (BCSH), first-line immunosuppressive therapy (IST) is a combination of antithymocyte globulin (ATG) and ciclosporin (CsA), indicated for non-severe AA (NSAA) patients who are suffer with transfusion dependency, encountering infections, recurrent bleeding, or hope for improved quality of life; severe AA (SAA) or very severe AA (VSAA) patients in the absence of an HLA-matched sibling donor (MSD); or SAA/VSAA patients > 35–50 years of age [5]. Although IST is effective at alleviating pancytopenia in a number of patients, it is not effective in all cases. In addition, it has been recognized that a part of patients treated with IST develop clonal hematopoiesis or somatic mutations that lead to myelodysplastic syndrome (MDS) or acute myeloid leukemia (AML) [6]. Furthermore, the high-risk recurrence of AA makes this treatment strategy a second choice behind HSCT from an HLA-matched family donor.

The 2015 GDMAA recommends first-line allogeneic hematopoietic stem cell transplantation (allo-HSCT) from the bone marrow of an MSD for treating SAA in young and adult patients who have an MSD. Unrelated donor (URD) HSCT is indicated for SAA after failure to respond to at least one course of nontransplant IST. Alternative donor allo-HSCT using either cord blood (CB) or a haploidentical family donor (HID) may be recommended for patients after failure to respond to IST and in the absence of a MSD or a suitably matched URD [5]. Although matched related donor (MRD) HSCT can be successfully conducted after the failure of IST or after evolution to MDS/AML, overall survival is decreased when transplantation is used as second-line treatment [7]. Moreover, the outcomes after allo-HSCT from an URD have steadily improved over the past three decades. Recent data have revealed similar outcomes for upfront-unrelated and matched sibling HSCT for pediatric AA, which supports the recommendation for first-line treatment with an URD-HSCT for children who lack an MSD [8, 9]. In the past, alternative donor HSCT was another salvage choice for cure in patients with refractory AA after IST, but morbidity and mortality from graft failure and complications of graft-versus-host disease (GVHD) have limited clinical applications for this approach. With the improvements in transplantation technology and management of GVHD, haploidentical-HSCT (HID-HSCT) has become a viable alternative treatment for patients who lack an MRD. The successful application of posttransplant cyclophosphamide (PT-CY) for HLA-haploidentical grafts for AA patients generated high rates of engraftment, low rates of transplant-related mortality, low rates of GVHD, and eradication of pre-existing clonal diseases [10]. The evidence of improved long-term survival after HID-HSCT supports the potential role of HID-HSCT as a first-line therapy.

To compare the efficacy and safety of allo-HSCT with that of IST as a front-line treatment for patients with AA, we performed a meta-analysis of available studies that examined the impact of the two major competing treatment strategies for AA.

Acquired aplastic anemia (AA) is a serious hematologic disorder characterized by peripheral blood pancytopenia caused by bone marrow failure. The pathogenesis of this disease is thought to be the destruction of hematopoietic stem cells by autoimmunity. Allo-HSCT and IST using a combination of ATG and CsA have been the cornerstone of therapy for both SAA and NSAA patients since the 1970s. MRD-HSCT is now recommended as a first-line treatment for young and adult patients who have an MRD. MUD-HSCT is indicated for SAA patients after failure to respond to IST. First-line IST is a therapeutic option for patients in the absence of an MRD or with old age. However, this treatment approach is based on the results of comparative studies conducted mainly in the 1980s. Transplantation success for AA patients with an MRD has improved considerably over the past three decades, with a 75 to 80% chance of long-term cure. Now there is controversy concerning the upper age limit for MRD-HSCT as a first-line treatment because results vary in different series. Data from the European Group for Blood and Marrow Transplantation (EBMT) database revealed similar outcomes for patients in the age ranges of 20 to 30 years, 30 to 40 years, and 40 to 50 years [26]. Although treating a patient with transplantation alone in the case of IST failure is an appealing strategy, outcomes in patients undergoing transplantation after failing IST are worse than those in patients treated with first-line MRD-HSCT [7]. During the last two decades, the outcome of allo-HSCT for AA patients with an MUD has also improved significantly, suggesting that this treatment should be given an increased role in the treatment of children and young adults with AA who are without an MRD. Recent studies have revealed similar long-term OS after transplantation from an MRD compared to an MUD [8]. MUD-HSCT is therefore recommended as the first-line treatment for AA patients eligible for transplant but lacking an MRD. Furthermore, it may even be appropriate for some older adults to proceed with first-line MUD-HSCT. Alternative donor transplantation with PT-CY for AA has shown satisfactory outcomes, with high rates of engraftment, low rates of transplant-related mortality, low rates of GVHD, and eradication of pre-existing clonal diseases [10]. This program allowed expansion of the donor pool to allow use of HIDs and MRDs, and it is also under development as a first-line therapy in appropriate patient circumstances. Altogether, the evidence for improved long-term survival after allo-HSCT from various donor types supports the broader role of allo-HSCT as a first-line therapy. To comprehensively evaluate the efficacy and safety of allo-HSCT compared with IST as a front-line treatment for patients with AA, we performed a meta-analysis that examined the impact of the two major competing treatment strategies.

Our meta-analysis demonstrated significantly longer OS and FFS, as well as a higher response rate, for patients who underwent first-line allo-HSCT compared to first-line IST. However, this outcome should be interpreted with caution. All studies were non-randomized retrospective studies, and there was significant selection bias in the analysis. Young patients with severe disease and with available donors were more likely to receive transplantation, whereas older patients, patients with no available donor, or patients with severe complications tended to receive IST. Moreover, non-transplant centers would probably apply more likely IST while transplant centers would tend to apply allo-HSCT, which would add further bias. Most of the studies reported the outcomes of patients given first-line MRD-HSCT compared to those of patients given IST. Only two and three studies reported the outcomes of first-line MUD-HSCT and HID-HSCT compared to IST, respectively. Evidence for improved long-term survival with first-line allo-HSCT over IST is more robust for patients given first-line MRD-HSCT. Moreover, we found no difference in survival between patients undergoing first-line HID-HSCT and first-line IST in our subgroup analysis. This is consistent with the present controversial treatment situation. In China, HID-HSCT has been advocated as a first-line treatment for children with AA [24], but in the United States and Europe, HID-HSCT is regarded as an experimental treatment for the relatively limited number of cases reported with unknown long-term effects of complicated regimens and a mismatched immune system. In addition, we observed a great disparity in FFS for patients treated with allo-HSCT versus IST. Treatment failure was defined as death, no response, disease progression, or relapse in patients who received IST. However, the time of response evaluation was at 6 months after completion of treatment in most studies. In fact, the number of patients that achieved response increased over time for patients who received IST. Some patients acquired a defined response as late as 24 months after the completion of treatment [19]. In this situation, FFS may have been underestimated for patients who received IST. In addition, although overall survival was reduced when allo-HSCT was used as a second-line treatment compared to a first-line treatment in some studies, the survival advantage of allo-HSCT as a first-line treatment over allo-HSCT as a second-line treatment after the failure of IST or IST as a first-line treatment is still unknown.

Eltrombopag is a synthetic small molecule mimetic of thrombopoietin used for patients who remain pancytopenic after treatment with IST. Recently, it was also under evaluation for use in combination with IST for the treatment of SAA as a first-line therapy. In a recent study, the addition of eltrombopag to IST was associated with markedly higher rates of hematologic response among patients with SAA than in a historical cohort. The overall response rate at 6 months was 94% in one of three cohorts; at a median follow-up of 2 years, the survival rate was 97% [27]. However, to date, there has been no study comparing the efficacy of first-line allo-HSCT and first-line IST in combination with eltrombopag. In our opinion, this nontransplant combination therapy strategy may counterbalance the survival advantage of allo-HSCT in indicated patients with AA.

Some researchers have recommended revision of the guidelines for the first-line treatment of patients with newly diagnosed SAA. Allo-HSCT should be recommended as an initial treatment for newly diagnosed SAA patients. HLA typing should be conducted to identify a marrow donor among family members or in the donor registries at the time of diagnosis. The priority of donor source for allo-HSCT is MRD, MUD, follow by HID if a MUD is not rapidly accessible. Each of these donors may be superior to IST because of the long-term high-risk for disease recurrence and secondary MDS/AML with the use of IST for patients with AA. On the contrary, allo-HSCT is associated with high cure rate, a low risk for disease recurrence or the development of clonal disorders, and a relatively low risk for GVHD [28]. It has been well recognized that a high incidence of somatic mutations that lead to the development of clonal evolution and MDS/AML can be detected after IST treatment. However, it should be noted that clonal hematopoiesis on the situation of bone marrow failure does not reliably predict the development of clinical diagnosed diseases. In one study, only a small minority of AA patients with DNMT3A or ASXL1 mutations developed MDS during the follow-up time [29]. This is consistent with what was found during our meta-analysis. We found that a median of 15% (9.5–45%) of patients in our included studies experienced relapse, which is lower than the 37–38% of patients reported in other studies [30, 31]. This discrepancy may be due to different patient populations, the definition of response, treatment protocols, and the limited follow-up time in some studies. It has been suggested that IST should not be discontinued after response to therapy in patients with NSAA and SAA due to the high risk of relapse. Indefinite administration of low dose CsA may reduce the incidence of relapse [32]. For patients who received allo-HSCT, we found a higher incidence of GVHD, but the incidence of grade II–IV aGVHD, extensive cGVHD, and death caused by GVHD was relatively low.

Given that we saw no difference in all-cause mortality between patients who received allo-HSCT and IST, the quality of life assessment is a matter of cardinal significance for comparing these two treatment strategies. The late adverse events of allo-HSCT are of major consideration for patients who survive long-term after transplantation, but they are often left out of the comparison of the two treatment strategies. For example, fertility has been shown to be reduced when alkylating agents are combined with total-body irradiation during conditioning. Children born from patients after allo-HSCT are at increased risk for developing genetic diseases or congenital anomalies [33]. There is a methodology termed the quality-adjusted time without symptoms and toxicity (Q-TWiST) which allows for evaluate quality of life by retrospective analysis the time a patient spent in different health states, supposing that quality of life is reliance on different health state [34]. This assessment procedure can provide a detailed view on the result of the two treatment modalities by integrating quality of life parameters into the comparison, for instance, transfusion requirement, drug demand, adverse events, GVHD or clonal evolution. Studies used Q-TWiST revealed that patients treated with allo-HSCT take more time cured from AA, whereas IST patients have more transfusion requirements, medication demand, close medical care and spend more time in cost-intensive periods, in spite of similar overall survival and event-free survival were observed [21].

Although survival is significantly longer among AA patients undergoing first-line allo-HSCT compared to first-line IST, the selection of initial treatment for patients with newly diagnosed AA requires a comprehensive evaluation of donor availability, age, expected quality of life, risk of disease relapse or clonal evolution after IST, and the use of adjunctive eltrombopag. Our meta-analysis highlights the need for prospective studies to examine the role of these two treatment modalities.