|Home | About | Journals | Submit | Contact Us | Français|
Although recent advances in therapy offer the promise for improving survival in patients with severe aplastic anemia (SAA), the small size of the patient population, lack of a mechanism in North America for longitudinal follow-up of patients and inadequate cooperation among hematologists, scientists, and transplant physicians remain obstacles to conducting large studies that would advance the field. In order to address this issue, the Blood and Marrow Transplant Clinical Trials Network (BMT CTN) convened a group of international experts in March 2010 to define the most important questions in the basic science, immunosuppressive therapy (IST), and bone marrow transplantation (BMT) of SAA and propose initiatives to facilitate clinical and biologic research. Key conclusions of the working group were: 1) new patients should obtain accurate, expert diagnosis and early identification of biological risk; 2) a population-based SAA outcomes registry should be established in North America to collect data on patients longitudinally from diagnosis through and after treatment; 3) a repository of biologic samples linked to the clinical data in the outcomes registry should be developed; 4) innovative approaches to unrelated donor BMT that decrease graft vs. host disease are needed, and 5) alternative donor transplantation approaches for patients lacking HLA-matched unrelated donors must be improved. A partnership of BMT, IST and basic science researchers will develop initiatives and partner with advocacy and funding organizations in order to address these challenges. Collaboration with similar study groups in Europe and Asia will be pursued.
Aplastic anemia is a marrow failure syndrome with an lincidence of 2 per million in Western countries and 4-6 per million in Asia.(1, 2) In the vast majority of patients, the disease results from T-cell mediated autoimmune destruction of marrow elements leading to life-threatening cytopenias. The preferred therapy for younger patients with severe aplastic anemia (SAA) is HLA-matched sibling allogeneic bone marrow transplantation (BMT), which results in long-term survival in 85-90% of recipients.(3-5) Only 20-30% of patients will have HLA- matched siblings, and some will not receive an upfront BMT approach due to patient choice, physician preference, or BMT access issues. Hence most patients with SAA receive initial treatment with immunosuppressive therapy (IST), most commonly with a combination of anti-thymocyte globulin (ATG) and cyclosporine (CsA). Although 60-75% of patients respond with a decrease in or elimination of transfusion requirements, 10-35% of patients will relapse (require transfusions again) and the majority of patients will require long-term (5 year) therapy with cyclosporine.(6, 7) Others are at risk of clonal evolution to hemolytic paroxysmal nocturnal hemoglobinuria (PNH), myelodysplasia (MDS) or acute myeloid leukemia (AML).(8, 9) Well-matched unrelated donor (URD) BMT can be successful in patients failing immunosuppression, but because transplant related mortality (TRM) and graft versus host disease (GVHD) are higher than with HLA-matched sibling BMT, there has been limited enthusiasm for this approach in the past.
With recent improvements in survival after URD BMT,(10-12) SAA experts from the United Kingdom published guidelines recommending matched URD BMT if patients fail to respond to IST after 4-6 months (Figure 1).(13) This approach is being adopted more widely in Europe and Japan, supported by a prospective study in Japanese pediatric patients.(14) Prospective validation of these guidelines in older adults is needed, and several key questions about URD transplantation in SAA remain: 1) It is not clear when to offer URD transplantation to patients who relapse after an initial response to IST, as most patients will respond to further treatment with IST. 2) Half of SAA patients will not have a well-matched URD, and the role of alternative donor procedures such as unrelated cord blood (UCB) or haploidentical related donor transplantation is unclear. 3) Finally, there are inadequate data regarding long-term quality of life after URD BMT for SAA. If survival is improved after transplant, does it come at a high cost?
Answering these questions is challenging, as it requires comprehensive tracking of patients from diagnosis through all therapies. Long-term follow up of SAA patients is especially important because many adverse events (MDS/AML) can occur 1-2 decades after diagnosis. In the United States, except in a few centers, long-term outcomes of patients with this rare disorder are not followed. The Center for Blood and Marrow Transplant Research (CIBMTR) collects long-term outcome data on the minority of patients who undergo BMT, but data collected regarding therapies prior to transplantation is often inadequate to address many issues. More importantly, there is no effective mechanism to compare outcomes of BMT recipients with comparable patients receiving IST approaches.
In view of these challenges, an ad hoc SAA Committee was formed by the Steering Committee of the Blood and Marrow Transplant Clinical Trials Network (BMT CTN), a program sponsored by the National Heart, Lung, and Blood Institute and the National Cancer Institute, to consider potential research strategies in this disease. The Committee convened a working group of international experts in March 2010 in Bethesda, Maryland, in conjunction with an educational and scientific meeting sponsored by the Aplastic Anemia and Myelodysplastic Syndrome International Foundation. The purpose of the working group was to:
The conclusions of this working group are summarized below.
Idiopathic SAA patients have immune-mediated oligoclonal expansion of cytotoxic T-cells targeting hematopoietic stem and progenitor cells. These T-cells have a Th1 profile and secrete interferon-γ (15); potentially relevant polymorphisms in genes associated with an increased immune response have been identified.(16) Regulatory T-cells are decreased in almost all patients with SAA,(17) and infusion of T-regs abrogates lymphocyte-induced marrow dysplasia in mouse models.(18)
A notable observation in a portion of patients with SAA is the presence of shortened telomeres.(19, 20) Mutations of the telomerase enzyme complex (TERT, TERC, DKC1, NOP10, or NHP2) or in the shelterin telomere protection complex (TINF2) form the basis for the inherited marrow failure disorder, dyskeratosis congenita. Just under 10% of SAA patients will have a mutation in either TERT or TERC. A smaller percentage of patients with SAA and no other clinical phenotype will have a mutation in TINF2. Genetic variants in TERF1 may also contribute to risk of SAA, although to a lesser extent.(21) All of these genes are thought to contribute to telomere erosion, increasing risk of marrow failure and malignant transformation. Although telomere length does not predict response to immunosuppression in SAA patients (as opposed to dyskeratosis congenita patients who do not respond to IST), retrospective studies show that SAA patients with shorter telomeres at diagnosis are at higher risk of relapse after IST and are also more likely to undergo clonal evolution to MDS or AML.(22)
The impact of telomere dysfunction on BMT outcomes in SAA is not known. Patients with dyskeratosis congenita have a high incidence of organ toxicity, most notably hepatic and pulmonary, after BMT.(23-25) In telomerase knock-out mice (Terc-/-), short and dysfunctional telomeres preclude appropriate engraftment of donor wild-type hematopoietic stem cells, possibly due to poor stromal function.(26) A large study correlating telomere length with engraftment, toxicity, and survival in patients who received unrelated donor BMT for SAA over the past decade is currently underway through the National Marrow Donor Program (NMDP) and CIBMTR. Although this study may define putative risks associated with shorter telomeres during URD BMT, prospective studies will be needed to test the applicability of these associations with modern BMT therapy.
Until recently, laboratory based predictive biomarkers for IST response in SAA were lacking. Scheinberg and the NIH group correlated absolute reticulocyte count (ARC) and absolute lymphocyte count (ALC) at initial diagnosis with response, identifying groups at low and higher risk of failure and early mortality (Figure 2).(27) Further investigation showed that ARC combined with telomere length had better predictive power than either biomarker alone. Patients with both high ARC and longer telomeres appear to have excellent outcomes, whereas those with low ARC and shorter telomeres do poorly; patients with only one of the two adverse factors had intermediate outcomes.(28) Important follow-up questions to address include: 1) Does the prognostic ability of these assays hold up in a prospective multicenter cohort; and, 2) can intervention with URD BMT improve survival of patients with low ARC and shorter telomeres compared to IST? Other important goals for future trials are discovery of additional biologic factors with prognostic value (cytokine polymorphism profiles, single nucleotide polymorphism (SNP-A) genotypes, etc.), or identification of genetic aberrations that contribute to the pathophysiology of SAA.
Initial therapy of SAA with horse anti-thymocyte globulin (ATG) and cyclosporine (CsA), standard for more than two decades, results in response rates of 50% at three months and 60-75% at six months.(29-31) A second course of rabbit ATG given after a minimum of three months may lead to response in about a third of patients who do not respond to the first course.(32) Among patients who respond initially but later relapse, most will have some response to subsequent courses of immunosuppressive therapy. Slowing the rate of taper of CsA appears to decrease the likelihood or delay the onset of relapse.(33)
Over the past decade, researchers sought to increase initial response rates by increasing the intensity of IST through the addition of mycophenolate mofetil (MMF), sirolimus, or other agents to ATG/CsA.(34, 35) These efforts were not successful, suggesting that even intense IST is insufficient to abrogate the auto-immune aggression in some patients, or that, some of patients have more severe destruction of hematopoietic progenitors resulting in worse marrow reserve and insufficient stem cells to support renewed blood cell production after abrogating the auto-immune response. The possibility that we have reached a ceiling in the percentage of patients with the capacity to respond to immunosuppression was raised. Consistent with this idea, the EBMT group reported that although significant improvements in survival after IST occurred over each decade between the 1970s and the 1990s. Unfortunately, survival of patients treated between 2000 and 2007 has remained unchanged compared to those treated between 1990 and 2000 (Figure 3).(36)
In the context of this lack of improvement in response to IST, what agents or approaches might improve survival or patient quality of life in the future? A randomized study of 120 patients in the United States comparing horse ATG with rabbit ATG (NCT00260689) has completed accrual and will soon offer insights into the quality and length of response with these two agents. A few pilot studies show responses to alemtuzumab, although this highly immunosuppressive agent requires attentive supportive care measures to avoid life-threatening infectious complications.(37) Other new immunosuppressive agents will be tested in patients with relapse of SAA to establish efficacy and toxicity. Finally, investigators at Johns Hopkins University using high dose cyclophosphamide without stem cell rescue have demonstrated a high response rate with relatively low toxicity in newly diagnosed patients.(38) The use of high dose cyclophosphamide by other groups was associated with high rates of early and late toxicities, leading to closure of randomized trials examining this approach.(39) There is recent renewed interest in this agent as investigators in China have shown high rates of response with manageable toxicity using lower doses of cyclophosphamide than used by the Hopkins group.(40)
Survival after HLA-matched sibling BMT in patients with SAA less than 30 years old has exceeded 80% for the past 20 years, making this the preferred approach for these patients. In the last decade, survival of older BMT recipients improved significantly. Several factors likely contributed to this improvement. An EBMT analysis of HLA-matched sibling BMT outcomes in patients older than 30 years showed a statistically significant improvement in survival when a fludarabine (Flu)/ cyclophosphamide (Cy)/ATG preparative regimen was used, compared with traditional Cy/ATG approaches. Five-year survival in the Flu/Cy/ATG cohort was 77%, compared to 60% in the Cy/ATG group, and patients between the ages of 30 and 40 years had a survival probability exceeding 80%.(41)
Survival after URD BMT also improved dramatically in recent years (from 30-40% in the 1990s (42) to 70-80% currently(11)). EBMT data using Flu/Cy/ATG ± low dose TBI showed that improvement was especially notable after 2004, and that patients have the best chance of survival after BMT when they undergo the procedure within two years of diagnosis (Figure 4). Unpublished data from the CIBMTR using similar approaches show that two-year survival rates after 8/8 HLA-matched (using high resolution typing) URD BMT for SAA exceeds 80% (personal communication, M. Eapen, CIBMTR). There are many possible reasons for these improvements: the advent of molecular HLA typing resulting in better HLA matching, modern supportive care, and optimization of reduced intensity conditioning approaches.(12) Patient selection is also a factor. In early studies, BMT was only offered to high-risk patients who had failed multiple rounds of IST. There is now a tendency to offer BMT earlier in the course of therapy.(11, 12) Patients undergoing transplantation earlier in their disease course are more likely to begin the procedure with a history of fewer infections and with a lower likelihood of iron overload, renal dysfunction from long-term cyclosporine, transfusion-induced alloimmunization (which can increase risk of rejection), or platelet-refractoriness.
The major focus of recent clinical trials in URD BMT for SAA has been optimizing preparative regimens to allow sustained engraftment while minimizing regimen-related toxicity. A study published by Deeg et al evaluated de-escalation of TBI doses and demonstrated better survival in patients receiving cyclophosphamide (200mg/kg)/ATG plus 200 cGy TBI compared to higher TBI doses.(12) Five year survival probabilities after HLA-matched URD BMT using the regimen containing 200 cGy of TBI were 78% for patients 20 years of age or younger and 50% for older patients. A second optimization trial is currently underway under the auspices of the BMT CTN. This trial (BMT CTN 0301; NCT00326417) is designed to determine the optimal dose of cyclophosphamide (0, 50mg/kg, 100mg/kg, 150mg/kg) when given in combination with fludarabine, ATG, and a single dose of TBI (200 cGy). The 0 mg and 150 mg levels have been closed because of rejection and toxicity, respectively. The trial continues to accrue patients at the 50 mg and 100 mg dose levels. Thus, this type of conditioning regimen should be considered investigational, and caution should be exercised when selecting the cyclophosphamide dose in this setting.
There is some concern about TBI-based regimens increasing the risk of malignancies after URD BMT for SAA. Studies demonstrating an increased risk of second malignancies after related donor BMT with TBI-based regimens were published in the early 1990s. Those studies involved patients treated in the 1970-80s using TBI doses >900cGy or total abdominal irradiation (TAI) doses of 5-600cGy.(43, 44) Whether current approaches using a single dose of 200cGy of TBI increase the risk of post-transplant malignancies is unknown. Long-term follow-up of these patients is very important.
The dramatic improvement in survival rates after URD BMT that has occurred over the past decade has raised an important question. In the context of current therapy, when should patients with SAA be offered URD transplantation? While many groups now support offering this approach after failure of first IST, when should it be offered after subsequent failures? Can biological risk factors for failure of IST help determine the timing of URD BMT? What level of HLA typing and matching is required? When can alternative graft sources (cord blood, haploidentical donors) be used? What is the quality of life of survivors after URD or alternative donor transplantation? Some insight into URD transplantation of younger patients failing to respond to their initial round of IST was provided by Kosaka et al.(14) In this study, 201 pediatric patients with SAA lacking HLA-matched sibling donors were treated initially with IST. Of 60 patients who failed to respond at 6 months, 31 underwent URD BMT if they had matched URDs (25 patients), single antigen mismatched related donors (4 patients), or single antigen mismatched cord blood units (2 patients). Patients who did not have donors received subsequent rounds of IST. While overall survival at 5 years was not different between the transplantation and IST groups, failure free survival was dramatically better in the BMT group at 84% vs. 9% (p=0.001), and the majority of patients treated with subsequent courses of IST had continuing marrow failure.
An additional issue is the availability of suitable HLA-matched donors. Only about 70% of Caucasian patients will find a fully matched and available URD; patients from ethnic groups such as Hispanic, Black, or Asian-Pacific islander will find a fully matched and available donor less than half of the time.(45) Cord blood transplantation, which allows greater degrees of donor-recipient HLA-match, might be considered for patients without a suitable adult donor, but published experience from Japan and unpublished CIBMTR and European Group for Blood and Marrow Transplantation (EBMT) data show high rates of rejection and survival rates of less than 50% after cord blood transplantation for SAA.(46) Some small studies using combinations of flu/cy/ATG/TBI for conditioning show more promising survival rates after cord blood transplantation,(47) but larger validation studies are required. Some groups have explored the feasibility of haploidentical transplants, but reports to date are anecdotal.(48) Novel approaches that improve survival with the use of cord blood or haploidentical donors are needed to allow all patients who do not have good options with IST to undergo transplantation.
Age is a significant factor in both IST and BMT outcomes, with higher chances of failure and mortality, especially in patients >40 years of age. An analysis published in 1999 showed that the response rates to ATG/CsA IST among patients aged >60, 50–59 and <50 years were 37%, 49% and 57%, respectively; corresponding 5-year survival rates were 50%, 57% and 72%.(49) In the decade since this analysis was published, response rates to IST have not changed substantially, but there has been a steady improvement in supportive care leading to increased survival after both IST and BMT. Rates of GVHD are higher in older patients, however, and remain a major barrier to successful outcomes. Whether URD BMT can offer a survival advantage over IST in older patients is unclear, however, older patients failing IST may benefit from BMT approaches aimed at reducing GVHD while maintaining donor engraftment.
There are several barriers to advancing care of patients with SAA. First, the disease is rare. Only about 600 new cases/year are expected in the United States. This makes all types of studies difficult because any given center will only have a handful of patients. Second, the natural history of the disease plays out over more than a decade, with some patients failing multiple courses of therapy, but still surviving for 5-10 years, and others developing late-onset secondary MDS/AML. Third, referral to centers with specific expertise in SAA is sporadic and varies by patient location. Referral is important in order to have the latest testing performed to rule out inherited bone marrow failure syndromes, hypoplastic MDS, and other conditions that mimic SAA, and to enroll patients in studies. Fourth, timing and indications for referral for transplant vary considerably among centers, with many patients delayed excessively and referred to BMT only after they have developed active infections, significant transfusion burden, alloimmunization, and/or platelet refractoriness. Finally, because patients with SAA may be seen by local physicians or hematologists, referred to a regional academic center, and then referred a second time for BMT evaluation, it is difficult to follow patients from diagnosis through all of their therapeutic courses with an observation period sufficient to understand their long-term outcomes. In the US, particularly, a mechanism to track patients through multiple care-providers is lacking.
The SAA working group agreed that advances in biology and BMT survival warrant a series of initiatives to better understand the appropriate roles and timing of IST and BMT in treating SAA. One important effort that would greatly assist in moving the field forward would be to create a mechanism for identifying a high percentage of SAA patients at diagnosis, obtaining blood specimens to evaluate prognostic biomarkers, and following these patients through their treatment courses. Biologic samples for later studies could be obtained, with appropriate consent, and an SAA repository established. Patients could be offered participation in studies of related donor BMT, IST, URD BMT and alternative donor BMT as they became eligible for such studies. BMT timing would be determined by patient age, the availability of well-matched related or unrelated donors, response to IST and risk as determined by biologic markers. All patients would be followed long-term, and quality of life studies could carefully compare outcomes of surviving patients receiving URD BMT vs. IST.
The working group felt that the most appropriate way to improve care and enroll patients onto a large registry study would be to designate regional centers of excellence, where comprehensive evaluations, including key biologic assessments (telomere studies, etc.) could be performed at diagnosis and other key time points. Vital to this process is early assurance that the diagnosis of SAA is correct. A subcommittee headed by Dr. Richard Harris was appointed to assemble comprehensive diagnostic guidelines including tests to rule out inherited BM failure syndromes and other non-SAA diseases. Because the therapy of SAA patients sometimes involves transplantation and the CIBMTR currently has a large database of information on patients receiving BMT, the CIBMTR is a possible choice to manage an SAA registry or longitudinal observational study. Trials of biology, IST and BMT could be facilitated by a population-based SAA outcomes registry by identifying potential study subjects; patients would benefit by having wider access to studies of relevance to them. Patients in the registry could also be targeted for distribution of educational and support materials. The working group plans to seek funding from patient advocacy groups, private corporations and governmental funding sources in order to established a population-based outcomes registry and accompanying biologic sample repository to facilitate studies to address many of the issues discussed in this document.
The important questions in the therapy of acquired SAA can be addressed most effectively with collaboration among transplantation physicians, hematologists interested in IST, and scientists studying the biology of marrow failure. The relevant issues are interdependent. For instance, can telomere length and telomerase complex mutations help predict whether patients will fail IST and should, therefore, seek early transplantation therapy? Can selected biological factors (cytokine profiles, etc.) better define clinical risk groups? If we can define patients at high risk of IST failure, do the same or different factors predict outcomes after BMT? If a patient either fails to respond to IST or relapses afterwards, can new agents induce or prolong responses (alemtuzumab, etc.)? Can cyclophosphamide, given at a modified dose, improve the durability of initial responses without excessive early morbidity? Finally, we need to know whether novel strategies for URD BMT that decrease GVHD and maintain engraftment can be developed and performed safely with reasonable survival in older patients. Can alternative donor grafts be used successfully so that more patients are able to go to transplantation if immunosuppression is unsuccessful?
These questions can be directly addressed through the proposed SAA registry/biology study, since at given time points (first or subsequent failure of IST, etc.) patients who have appropriate donors and go to BMT can be compared with similar patients who undergo IST. Questions regarding access to BMT (inability to get BMT due to insurance, etc) can be addressed by the registry study as well. Transplantation studies could be performed through the BMT CTN in cooperation with international study groups as needed to allow for sufficient accrual or to rapidly test highly promising ideas.
Treatment for SAA has advanced in the past decade, most notably with 1) the development of biological measures that may allow clinical risk classification and 2) improvement in survival after HLA-matched URD BMT. Creation of an SAA outcomes registry and specimen repository would allow investigators to follow patients from diagnosis through all of their therapies, would facilitate better studies comparing specific therapies, and thus may help optimize timing of URD BMT for patients failing IST. Cooperation among hematologists, transplant physicians, and basic scientists in the study and treatment of SAA patients should speed advances in clinical care and improve outcomes.
This International Working Group was supported by the Blood and Marrow Transplant Clinical Trials Network, the National Heart, Lung, and Blood Institute, the National Cancer Institute, and a grant from Genzyme. The authors gratefully acknowledge the participating centers and co-investigators of the BMT CTN 0301 study.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.