Introduction

Despite significant advances in management, bone marrow failure in children remains challenging to diagnose and often somewhat frustrating to treat. Specialized diagnostic investigations are now more widely available, including complex molecular and genetic testing, enabling more rapid specific workup based on the clinical presentation. However, in the absence of an inherited syndrome, acquired causes are more difficult to find, and many children are still diagnosed with “idiopathic” aplastic anemia (AA). While the severity is somewhat variable, bone marrow failure remains challenging to manage. The treatment approach for inherited bone marrow failure syndromes (IBMFS) is more straightforward, with hematopoietic stem cell transplantation (HSCT) being widely accepted as the modality of choice. In acquired severe aplastic anemia (SAA), transplant remains the optimal first choice if a matched sibling donor is available. However, if this is not the situation, there is no wide consensus as to the course of treatment. Several options have been proposed for first and second line treatment, and this has changed over the course of the last several years. In this article authors will review the current diagnostic and treatment algorithms, suggesting a more stepwise approach to minimize testing, and provide a rational management approach to maximize treatment benefit with the least risk for morbidity and mortality, keeping in mind that all of the resources needed may not always be available. This review will focus on the diagnostic and therapeutic approach to the management of bone marrow failure. Detailed descriptions of each inherited condition are beyond the scope of this review.

Evaluation of Bone Marrow Failure

The etiology of bone marrow failure is broad and can result from a diverse array of conditions, some inherited and others acquired (Tables 1 and 2). Appropriate evaluation of the etiology is critical to its management since the choice of first line therapy is dependent on it. Broadly, bone marrow failure may result from an inherited syndrome (henceforth called Inherited Bone Marrow Failure Syndrome or IBMFS), in which case stem cell transplantation is the first line treatment of choice, or from an acquired cause (henceforth called Aplastic Anemia or AA), in which case the choice of first line therapy is broader, and includes immunosuppression vs. stem cell transplantation in appropriate situations. The clinical presentation of bone marrow failure is a combination of signs and symptoms related to the pancytopenia as well as any other features which might indicate an underlying cause or inherited syndrome. A stepwise evaluation algorithm (Figs. 1, 2 and 3) should include clinical features, many of which are pathognomonic in inherited syndromes (Table 1), and corroborated with laboratory testing. Examination of the bone marrow is the key initial step, to determine cellularity as well as to look for dysmorphism which might provide specific diagnostic clues. Flow cytometry, cytogenetics and Fluorescent in situ hybridization (FISH) should be considered routine to look for subpopulations/clones which might point to a specific diagnosis, e.g., Paroxysmal nocturnal hemoglobinuria (PNH) or Myelodysplastic syndrome (MDS). If available, viral polymerase chain reaction (PCR) of the bone marrow may be done in instances where an infectious etiology is suggested by the clinical presentation or serologic testing. When an inherited syndrome does not seem to be the clear diagnosis at presentation, depending on the likely duration of progression of the marrow failure, there may or may not be some time available to proceed along this algorithm in a stepwise fashion. In situations where the pancytopenia is likely long standing, treatment may be more urgently indicated, and several steps may need to be combined for efficiency. Much of the testing for the inherited syndromes relies heavily on DNA testing, now accomplished using multiplex gene testing in panels specifically designed for the purpose. As our understanding of the biology of bone marrow failure advances, whole exome, genome or next generation sequencing may yield actionable information, pointing to a likely inherited/genetic cause, or a targetable polymorphism/mutation. If available, these tests may be performed if no clear diagnosis is apparent. Aplastic anemia is classified as moderate, severe, or very severe based on accepted criteria (Fig.2), often with progression over time.

Table 1 Inherited (primary) causes of bone marrow failure, organized by most common presenting cytopenia
Full size table
Table 2 Acquired (secondary) causes of bone marrow failure
Full size table
Fig. 1
figure 1

Diagnostic approach to pancytopenia or unexplained chronic cytopenia(s). The investigations above will assist in determining whether to pursue an aplastic anemia diagnosis (Fig. 2) or further IBMFS work-up (Fig. 3). ALL Acute lymphoblastic leukemia; AML Acute myeloid leukemia; BM Bone marrow; CMV Cytomegalovirus; EBV Ebstein-Barr virus; FISH Fluorescent in situ hybridization; GPI Glycosylphosphatidylinositol; HHV-6 Human herpesvirus 6; HSCT Hematopoietic stem cell transplantation; IBMFS Inherited bone marrow failure syndrome; MDS Myelodysplastic syndrome; PCR Polymerase chain reaction; PNH Paroxysmal Nocturnal Hemoglobinuria

Full size image
Fig. 2
figure 2

Diagnostic approach to pancytopenia or unexplained chronic cytopenia(s), Aplastic Anemia. Diagnostic criteria for acquired aplastic anemia. AA Aplastic anemia; ANC Absolute neutrophil count; BM Bone marrow; SNP Single nucleotide polymorphism

Full size image
Fig. 3
figure 3

Diagnostic approach to pancytopenia or unexplained chronic cytopenia(s), - Inherited Bone Marrow Failure Syndrome Work-Up. BM Bone marrow

Full size image

Treatment

Without definitive treatment, a child with established IBMFS or SAA will invariably succumb to either a severe infection which would occur as a result of the prolonged and severe neutropenia, or from a catastrophic hemorrhage, which would result from the prolonged and severe thrombocytopenia. Supportive therapy with transfusions and prophylactic antimicrobials is only a temporizing measure, and the reader is directed to specific disease guidelines for IBMFS and a published supportive care guideline for AA [1]. The goal of therapy must be to induce “remission” of the marrow failure and allow for its reconstitution, with resulting normalization of peripheral blood counts. As discussed previously, the choice of first line therapy depends on whether the bone marrow failure is as a result of an underlying inherited syndrome (IBMFS) or an acquired cause (AA). In the former situation, the response to non-transplant therapies is sub-optimal, and even if there is a response, it is usually short-lived, and HSCT is the only treatment which will result in a lasting response. However, for acquired causes of aplasia, immunosuppression as the first course of treatment has a fair and improving track record. More recently, other treatments, such as eltrombopag, have shown response even in SAA, and have entered the treatment paradigm. HSCT as upfront treatment is only considered if a matched related donor is available, and has gained more acceptance as the initial management because of improving outcomes. Figure 4 provides a guide to the different treatment options with specific regimens in Table 3.

Fig. 4
figure 4

Management approach to bone marrow failure. A general approach to management of AA or an IBMFS. AA Aplastic anemia; ATG Antithymocyte globulin; CsA Cyclosporine; HLA Human leukocyte antigen; HSCT Hematopoietic stem cell transplant; IBMFS Inherited bone marrow failure syndrome; IST Immunosuppressive therapy; MUD Matched unrelated donor

Full size image
Table 3 Immunosuppressive therapy for acquired aplastic anemia
Full size table

Immunosuppression

First line immunosuppressive therapy (IST) for SAA is a combination regimen of horse antithymocyte globulin (hATG) and cyclosporine (CsA), resulting in an initial response rate of 70–80% in children [8,9,10,11,12]. Specific treatment regimens can be found in Table 3. Horse ATG is preferred to rabbit antithymocyte globulin (rATG) for upfront IST due to superior hematologic response rates (68% vs. 37%) and 3 y overall survival (OS) (96% vs. 76%) [2]. Careful monitoring and supportive care are required while awaiting hematologic recovery, which can take 2–3 mo or longer in some patients [3]. Long-term OS with h-ATG/CsA is 85% in the best case, though response rates vary considerably and may be as low as 37% [9, 13]. The variability in response is not well understood, and possible reasons for a suboptimal result could be delay in initiating treatment, or instances in which the underlying diagnosis is an undiagnosed inherited syndrome. In patients with refractory disease, a second course of IST with rATG and CsA or alemtuzumab yields a 30–40% response [4, 14, 15]. Remission induced by IST may be durable, sometimes requiring long-term CsA to maintain it, which should be continued for a minimum of 6 mo and then tapered very slowly with careful monitoring [12]. However, a substantial proportion (12–35%) will relapse and become pancytopenic again [16,17,18]. About half of these children will respond to a second course of ATG (usually rabbit), and some of these remissions will be durable as well. However, those who do not respond initially, or who relapse, are considered refractory, and alternate treatment regimens are indicated in such situations. Responders who recover their marrow function need long-term follow up not only because late relapses may sometimes occur, but also for surveillance for the evolution of clonal disease (myelodysplasia or a secondary malignancy) which may occur in approximately 15% [19].

Eltrombopag

The thrombopoietin (TPO) receptor agonist, eltrombopag, can be used in patients with SAA and lack of response to IST. Monotherapy with eltrombopag in refractory SAA results in a 40% response rate with recovery of trilineage hematopoiesis seen [5, 6]. Use of eltrombopag in combination with hATG/CsA as upfront therapy has also been tried in a subsequent trial and yielded the best results yet with an overall hematologic response rate of 95% and complete response rate of 58% at 6 mo in one treatment cohort [7]. Longer follow up is required to determine relapse rates and salvageability, safety regarding clonal evolution, and overall survival. At this time, there is no evidence to suggest increased rate of development of clonal disease with the use of TPO receptor agonists over traditional IST, however the risk of this is lower yet with HSCT compared to medical therapy. There is an ongoing Phase 2 trial evaluating the role of eltrombopag monotherapy for previously treated or untreated moderate aplastic anemia (NCT01328587). The dose of eltrombopag is recommended to be reduced to 50% in individuals of East Asian ancestry with ITP, but specific data are not available for Southeast Asia / Indian subcontinent or in aplastic anemia. While evidence-based, specific dosing recommendations cannot be made, it may be prudent to start at a lower dose and titrate up if there are no untoward changes in liver function testing.

Androgens

While androgens have largely fallen out of favor for medical treatment of AA with an early randomized controlled trial failing to show benefit in the addition of oral androgens to ATG, there may still be a role in certain patients [12, 20]. A randomized trial reported by Bacipalugo et al. did show superior response rates with the addition of androgens to ATG in women, and their group later showed a 59% response rate with oral androgen therapy in patients who failed an initial course of IST [21, 22]. Leleu et al. reported a 77% response rate with the addition of androgens to intensified doses of ATG with a 78% 5 y OS [23]. Therefore, in select scenarios with limited options, such as patients with refractory aplasia following IST who may not tolerate HSCT, for patients who do not have a suitable donor, or in limited resource settings, androgens may prove useful. The side effects, particularly virilization in female patients remains a significant consideration. Androgen therapy (danazole, oxandrolone, and oxymetholone) has also been shown to produce improvement in peripheral blood counts in well over 50% in patients with Fanconi Anemia (FA) and bone marrow failure [24, 25]. The response is primarily seen in hemoglobin and platelet counts, and in some patients is even durable. Again, this treatment modality may be favorable in specific scenarios where patients donot have a suitable donor, access to HSCT, or have too many comorbidities to proceed with HSCT. Careful monitoring is required, however, as bone marrow failure may progress to MDS/acute myeloid leukemia (AML) [26].

Hematopoietic Stem Cell Transplantation

HSCT is the only curative therapy for inherited forms of bone marrow failure and may be the treatment of choice in AA depending on donor options. In IBMFS, a matched and unaffected (based on genetic testing) sibling is the ideal donor, but another matched and unaffected family donor may also be suitable. In cases of consanguinity, parents should have HLA testing performed to assess their suitability as a donor. When a matched family donor is not available, an unrelated donor (UD) would have to be considered, and results from such transplants may be suboptimal as a result of more frequent complications such as rejection and graft vs. host disease (GVHD). For SAA, IST is the most commonly used first line therapy, but HSCT is preferred if there is a matched sibling donor, as OS is comparable between patients treated with IST vs. HSCT (around 90%), but there is a significantly higher rate of failure free survival of 87% in patients treated with HSCT compared to 56% in patients treated with IST [27]. These excellent outcomes with HSCT vary with age, with younger age being strongly associated with improved outcomes, with a 5 y survival of 82% in patients less than 20 y of age decreasing to 72% or 53% after age 21 or 40, respectively [28]. In cases of SAA without an HLA matched sibling, medical treatment with IST is still considered by most to be first line, though HSCT with a matched UD has increasingly been considered for up front first line HSCT with improving outcomes and lower risk for GVHD. Alternate donor transplant should only be pursued in the setting of failure of IST or in experimental research trials [12]. A search of the marrow registry for an UD should be initiated without delay, however, to prepare for the possibility of refractory SAA, in which case HSCT may be the best treatment option. Mismatched or UD HSCT is usually the only option for children with SAA who are refractory to all other treatments.

Before HSCT is pursued, risks and benefits should be weighed with each patient and family, as there is risk of substantial morbidity and mortality. Appropriate counseling regarding fertility preservation options should be provided. Specific conditioning regimens are beyond the scope of this discussion and the reader is directed to disease specific guidelines. In general, myeloablative conditioning is used for AA and reduced-intensity conditioning is recommended for IBMFS, while total body radiation should be avoided in either scenario to reduce long-term toxicity including secondary malignancy, with the exception of UD HSCT for SAA. Bone marrow is the recommended stem cell source for both acquired and inherited bone marrow failure due to a proven survival advantage and reduced risk of development of GVHD. In matched sibling HSCT for SAA, the rate of development of chronic GVHD is 11–12% using bone marrow vs. 22–27% with peripheral blood stem cells (PBSC) [29, 30]. Additionally, Bacigalupo et al. reported improved survival advantage (90% vs. 76%) using bone marrow as the source of stem cells with no difference in rates of engraftment (approximately 90% regardless of stem cell source) [30, 31]. In cases where an UD must be used for HSCT for SAA, a similar pattern is seen, albeit with much higher rates of chronic GVHD -- 31% with bone marrow vs. 48% with PBSC [32]. A recent European Group for Blood and Marrow Transplantation (EBMT) analysis showed that in both matched sibling and UD transplants, the use of peripheral blood as the source of stem cells was the strongest negative predictor of survival [31]. In terms of stem cell source for HSCT in IBMFS, a clear association with rate of GVHD has not been seen in FA, however use of PBSC was associated with significantly higher risk of secondary cancer, therefore bone marrow is also the preferred stem cell source for FA and is also recommended for the remaining IBMFS [33, 34]. GVHD prophylaxis is essential, as development of GVHD in non-malignant transplant is the main contributor to long-term morbidity, mortality, and overall quality of life, and may induce even more significant toxicity in patients with IBMFS. Standard post transplant GVHD prophylaxis for SAA includes CsA and short course methotrexate, which was shown to be superior to CsA alone, and when combined with early timing of transplant and an ATG-containing preparative regimen in matched sibling HSCT, Storb reported a 95% engraftment rate and 90% 2 y OS [35, 36]. In general, rates of graft rejection have improved dramatically in SAA with improved supportive care and conditioning regimens and reduced amount of time before proceeding to HSCT and therefore, less exposure to blood products from multiple donors prior to transplant. For IBMFS, reports in the literature are limited to small and primarily retrospective reviews of specific inherited conditions and thus generalizations are difficult to make. Perhaps the most data comes from FA, where recipients of matched related donor HSCT have 89–100% engraftment, 20–29% chronic GVHD, and survival rates above 90% at median follow up times of 2–3.7 y. In patients with IBMFS, matched UD or matched related cord blood are also acceptable options, while alternative options such as haploidentical transplantation should only be performed in experienced centers and/or within the confines of a clinical trial [37].

Long-Term Follow-Up

For patients with AA treated with IST, the disease is never completely cured, as there is significant risk of refractory disease, late relapse of disease, or late progression to clonal disease, including myelodysplastic syndrome (MDS) or acute myeloid leukemia (AML). Relapse of disease, notable by decreasing peripheral blood counts and loss of transfusion independence, may sometimes be rescued with reinstitution of immunosuppression (CsA) or may require more intensive IST or HSCT. Patients require long-term monitoring and supportive care, specifically with regular bone marrow aspirate and biopsy for morphologic and cytogenetic evaluation, perhaps every 6 mo for the first 1–2 y and then annually, as recommended by Bacigalupo, Scheinberg and Young, and the British guideline for adult AA [3, 12, 38]. While the risk of clonal evolution to MDS or AML is approximately 15% as previously mentioned, a recent report indicates that clonal hematopoiesis may occur at a much more frequent rate of 47% with clinical significance that is not yet well understood [19, 39]. Patients with IBMFS or AA treated with HSCT must be monitored long-term for the usual complications of transplant, the most concerning of which being chronic GVHD, loss of engraftment, and/or relapse of disease. These patients should be managed at specialized transplant centers with experience in managing these complications.