Approach to the diagnosis of aplastic anemia

The bone marrow failure (BMF) state of aplastic anemia (AA) is marked by cytopenias and ineffective hematopoiesis.¹ AA confers a significant risk for morbidity and death as a result of its progressivenatural history and/or complications related to suboptimal therapy.^2,3 Without definitive treatment, mortalityfrom severe AA (SAA) approaches 70% at 2 years.⁴ Establishing an accurate etiology of the AA ischallenging but also absolutely critical to appropriate management, especially differentiating between acquired and inherited forms of the disease. With ever-evolving diagnostic techniques^5,6 in the field, as well as increasing recognition of genetic underpinnings of disease,⁷ the standard diagnostic algorithm for AA has become more complex. Coupling this diagnostic heterogeneity with the increasingly widespreadpreference to use early related (matched and mismatched) donor bone marrow transplantation (BMT) in the field^2,8 necessitates updated guidelines to ensure that appropriate diagnostic and therapeutic pathsare pursued.

Patient 1 is a 48-year-old female of European ancestry with obesity, obstructive sleep apnea, and history of Hashimoto’s hypothyroidism. There is a strong family history of autoimmunity with a paternal grandmother and sister with rheumatoid arthritis, as well as a father with psoriatic arthritis. Our patient called her internist because of increased bruising; she also noted a 2-month history of decreased exercise tolerance, dyspnea on exertion, and fatigue. Initial laboratory studies revealed that she had a platelet count of 5 × 10⁹/L, hemoglobin of 6 g/dL, and a total white blood cell count of 2.9 × 10⁹/L, with an absolute neutrophil count of 1.1 × 10⁹/L. Initially, immune thrombocytopenia was thought to be the most likely diagnosis. In that context, she was started on prednisone, 100 mg/d for 3 days, which did not augment her counts. Further diagnostic testing was needed.

AA is a diagnosis of exclusion. Thus, there is no single test that can be used to consistently diagnose AA from the myriad other causes of BMF. Consequently, the diagnostic evaluation must assess for and exclude these alternative etiologies. At initial AA presentation, many patients manifest with fatigue, weakness, pallor, and headaches due to anemia. Often, patients have petechiae of the skin and mucous membranes, epistaxis, and/or gum bleeding related to severe thrombocytopenia. Fever and infections can also be seen in these patients as a result of low white blood cell counts and neutropenia. AA patients identified earlier in the course of the disease by abnormalities found on routine laboratory testing may not have any physical manifestations of their disease.

Acquired AA, also called idiopathic AA, represents the majority (∽70%) of all newly diagnosed cases.^9,10 The incidence of acquired AA is 2 per million in Western countries and 4 to 6 per million in Asia.⁹ Across the lifespan, there are 2 incidence peaks: 1 among young adults and a second in the elderly.¹¹ BMF as part of an underlying inherited BMF disorder (IBMFD) presumably represents the remainder of AA diagnoses. The proportion of AA cases accounted for by IBMFD-related BMF is rising as a result of the increasing recognition and the discovery of multiple new genetic syndromes in this context along with improved access to testing for these IBMFDs.¹² More than 25% of pediatric patients and ∽5% to 15% of adults age 40 or younger who present with AA have an inherited etiology.^3,13,14 This proportion is less well defined for patients older than 40 years of age but likely will become increasingly common as we investigate the inherited genetic basis of this disease further.

Acquired AA (called “AA” hereafter) is a hematopoietic stem cell disorder featuring reduced bone marrow cellularity and decreased hematopoiesis. This decreased hematopoiesis may disproportionately affect 1 or 2 lineages in the early stages of disease, but AA is ultimately associated with trilineage hypoplasia.¹⁵ The initial diagnostic work-up focuses on distinguishing AA from other diverse causes of cytopenias.¹⁶ Diagnosis can be challenging in some patients because AA, other immune cytopenias, myelodysplastic syndrome (MDS; cellular or hypocellular), paroxysmal nocturnal hemoglobinuria (PNH), and IBMFD are all considered BMF states.¹⁷ These disorders can have significant overlap in the clinic, as well as in an individual patient. It is not uncommon for autoimmune diseases to track in families, even if not the identical disease; therefore, a compelling family history may be suggestive of an autoimmune phenotype¹⁸ or an inherited predisposition.¹⁹ Patients with acquired disease are often treated with immunosuppressive therapy (IST).^2,20,21 

Three attempts were made to obtain a marrow biopsy for patient 1 but were challenging because of her body size and resulted in insufficient sample. Ultimately, a successful aspirate and core under conscious sedation showed a markedly hypocellular (<5%) marrow consisting predominantly of fat with a few scattered areas of erythroid precursors but no signs maturing of myeloid elements. The CD34 count was <1%. A PNH red blood cell clone size of 0.09% was detected in peripheral blood. This included 0.06% type III cells and 0.03% type II cells. A PNH white blood cell clone was detected in 13.1% of granulocytes and in 12.6% of monocytes. The karyotype was 46,XX. During the 5-week assessment period, the patient received 7 platelet transfusions and 5 units of packed red blood cells; her absolute neutrophil count fell to a nadir of 0.47 × 10⁹/L.

The goal in a thorough and standard work-up is to confirm the diagnosis and exclude other causes of pancytopenia in the setting of a hypocellular bone marrow. At a minimum, the laboratory tests outlined in Table 1 should be performed using peripheral blood.

Detecting the presence of PNH clones is also important at diagnosis. The presence of PNH clones and a copy number–neutral loss of heterozygosity of chromosome arm 6p may suggest a diagnosis of AA.²² Most often, a PNH clone is considered a marker of acquired disease, and its presence alone is not equivalent to a diagnosis of classical PNH. The classical PNH disease requires evidence of a hemolytic anemia or thrombotic event and often is a cellular or hypercellular marrow, distinct from AA.^23-26  The close pathophysiologic relationship between PNH and AA is explained by the lack of particular glycosylphosphatidylinositol anchor proteins on the surface of these mutated stem cells that render them relatively resistant to immune attack; thus, these cells do not have an inherent proliferative advantage but rather have a conditional advantage because they escape immune destruction.²⁷ Up to 70% of patients with acquired AA have a small PNH clone at diagnosis.^25,27,28 The presence of a PNH clone, irrespective of size, was a good predictor of response to IST in several series.^28-30  A PNH clone, in the setting of true marrow failure with AA, usually does not necessitate C5 blockade as primary treatment typical for a patient with more classical PNH.^31,32 However, when a clone is present, reassessing for expansion of the PNH clone over time and immediately prior to BMT is recommended because terminal complement component activation blockade (eg, C5 to C5a conversion by eculizumab) can be used to avoid thrombotic events at the start of conditioning.³³ 

A bone marrow biopsy and aspirate often showing the characteristic “empty” marrow on histology are a prerequisite for a diagnosis of AA. BMF arising from an underlying IBMFD can have distinguishing syndrome-specific features in the marrow,^12,34,35 or it can appear indistinct from acquired AA, as in cases 3 and 4 below. In the case of patchy marrow cellularity, a very low CD34 count quantified by flow cytometry of bone marrow aspirates or immunohistochemistry on the core biopsy favors a diagnosis of AA.^17,36 It is not uncommon in AA for cytogenetics to have a nondiagnostic result due to insufficient metaphases. In that case, fluorescence in situ hybridization (FISH) studies can be added to evaluate for the common aberrations (ie, chromosome 5 and 7 abnormalities) that might be diagnostic for MDS.^37-39  Most often, AA patients have normal cytogenetics; however, some cytogenetic abnormalities seen in AA are not considered adverse or indicative of MDS (in the absence of dysplasia). These include del13q, trisomy 8, and loss of heterozygosity of short arm of chromosome 6, among others.⁴⁰ Single-nucleotide polymorphism microarrays can also be used as an alternative karyotyping tool in these cases to detect most cytogenetic aberrations.^41-43 

AA is further classified clinically by the severity of the depression of the peripheral blood counts. SAA is defined by a decrease in blood counts involving ≥2 hematopoietic lineages (ie, absolute reticulocyte count <60 × 10⁹/L, absolute neutrophil count <0.5 × 10⁹/L, or platelet count <20 × 10⁹/L) and bone marrow hypocellularity (<25% of the normal cellularity). Very severe AA has an absolute neutrophil count <0.2 × 10⁹/L, whereas moderate AA is characterized by the depression of blood counts not fulfilling the definition of severe disease.^44,45 This classification system is closely tied to prognosis and is often used to triage patients meeting Camitta criteria for severe or very severe disease to a therapeutic path including early BMT.² 

In patients with cytopenias, it is no longer standard practice to do bilateral biopsies for initial AA diagnosis. However, repeat marrow examinations over time may be necessary to establish the diagnosis. The interval between marrows should be informed by the severity of cytopenias in patients meeting criteria for moderate disease at the time of diagnosis, any major changes in clinical status or blood counts, and whether the etiology is AA or BMF due to an underlying IBMFD.

Patient 1 was ultimately diagnosed with acquired SAA in the setting of compelling personal/family autoimmune history and presence of the PNH clone. She went on to receive IST with antithymocyte globulin and cyclosporine. She did have immune-responsive disease with improvement in blood counts and transfusion independence by 5 months posttreatment, which has lasted for >4 years and allowed for weaning of her cyclosporine over time.

A 63-year-old male has only a history of type 2 diabetes, hypercholesterolemia, and hypertension. He suffered nosebleeds on 3 consecutive days and went to the emergency department where he presented with a platelet count of 9 × 10⁹/L, hemoglobin of 7.9 g/dL, total white blood cell count of 1.6 × 10⁹/L, and an absolute neutrophil count of 0.9 × 10⁹/L. A bone marrow biopsy was hypocellular at 5% cellularity overall with scattered clusters of fat cells with rare myeloid cells. Karyotype was 46,XY, and a targeted next-generation sequencing panel test revealed a pathogenic BCOR variant at a variant allele frequency of 5.3% and a DNMT3A mutation detected below the lower limit of standard reporting (<2%). His local hematologist wondered whether this BCOR variant or the DNMT3A raised concern for a diagnosis of hypocellular MDS vs AA.

For the patient and their treating clinicians, it is important to use available techniques to ensure adequate explanation of the biology of the marrow failure process. This includes thorough attempts to differentiate AA from hypoplastic MDS and unique subtypes of IBMFD where feasible, because phenotypic presentations of these diseases might be indiscernible.⁴⁶ To help with this diagnostic challenge, the work-up of pancytopenia (even suspected AA) now often incorporates next-generation sequencing of targeted panels of genes commonly acquiring pathogenic variants in various myeloid malignancies.⁴⁷ Multiple studies have applied these techniques to investigate the mutational spectrum in AA; the frequencies of somatic mutations range from 5% to >70% in some series.^48-53  Mutations in DNMT3A, BCOR, BCORL1, and ASXL1 are seen in AA, MDS, aging populations, and inherited marrow failure disorders.⁵⁴ There is considerable overlap between the mutational spectrum in AA and MDS, with mutations in BCOR/BCORL1, DNMT3A, and ASXL1 common to both disease states⁵² at similar variable allele frequencies (VAFs). For example, the prevalence of BCOR and BCORL1 mutations is estimated at 5% in MDS⁵⁵ vs 4%⁴⁹ or 7% to 10% in AA.⁵² Thus, mutations in these genes do not have the discriminating power to help with the diagnosis of hypoplastic MDS vs AA. However, mutational profiles more typical of AA included smaller clonal populations represented by VAFs < 10%.⁵² Mutations in DNMT3A and ASXL1 are frequent in AA and MDS as well and may increase in clonal burden over in time.⁵⁶ A baseline panel at diagnosis also allows assessment of new mutations acquired over tim with changes in disease status (eg, relapse).

Currently, the prognostic implications of somatic mutations in AA are also unclear, because no prospective study has evaluated the outcomes of newly presenting AA patients with adequate patient numbers and follow-up. However, retrospective^49,50 and prospective studies^52,54 suggest that detection of certain somatic mutations by next-generation sequencing (ie, ASXL1, DNMT3A, and BCOR) can enable risk stratification of patients with AA at high risk for progressing to a myeloid neoplasm. Yoshizato et al⁵² did not find any significant difference in overall survival and progression-free survival (evolution to MDS) between all patients with somatic mutations and those without mutations. When analyzed by individual genes, only BCOR/BCORL1 were independently associated with an improved therapeutic response. In the same study, the investigators also found (by exploratory artificial learning analysis) an improved prognosis for AA patients with a favorable set of mutations (BCOR/BCORL1 and PIGA); another aggregated unfavorable set of mutations was associated with worse outcomes. Notably, none of the individual genes (including DNMT3A and ASXL1) was found to be linked to poor outcomes on their own. Other series have also attempted to explain risk in AA, based on various mutations, without clear therapeutic recommendation or necessary action.^56-59  Detection of a less favorable mutation profile at diagnosis may push certain patients toward upfront related or alternative donor BMT over IST where feasible; however, it is important to note that detection of a somatic mutation by itself is not a sufficient indication for BMT. A variety of other parameters, such as peripheral blood counts, VAF, and karyotype, are crucial to take into account when considering the preferred treatment approach.

Patient 2 ultimately went on to receive a related donor transplant for his treatment-naive SAA on a clinical protocol and has a normal hemogram and full donor chimerism without further evidence of clonal hematopoiesis 2 years post-BMT.

Patient 3 is a 16 year-old male with no significant health history who was brought to the pediatrician by his mother. She had noted he was more listless and less engaged over the prior few weeks. On examination, he was pale with ecchymoses throughout his lower extremities. His physical examination was otherwise unremarkable, with appropriate stature for age and no facial or limb abnormalities. His total white blood cell count was 0.4 × 10⁹/L with 0% neutrophils, his hemoglobin was 7.7 g/dL, and his platelet count was 8 × 10⁹/L at presentation. A peripheral blood smear showed a paucity of cells and no circulating malignant cells. The pediatrician noted “irregular bloodwork” a few times in his youth. He had not seen a provider since age 11 years. There was no significant family history.

Patient 4 is 19 year-old otherwise healthy male who was seen in the emergency room for intense migraines, fatigue, and several weeks of gingival bleeding. He had had a significant upper respiratory infection 6 to 8 weeks prior. He had wet purpura in his mouth, thrush, and ecchymoses throughout his lower extremities. His physical examination was otherwise unremarkable. His total white blood cell count was 1.2 × 10⁹/L with only 24% neutrophils, his hemoglobin was 5.7 g/dL, and his platelet count was 2 × 10⁹/L at presentation. A peripheral blood smear showed a paucity of cells and no circulating malignant cells. His family history was unrevealing.

Patients 3 and 4 were referred to hematology/oncology for management. Both had urgent bone marrow evaluations that showed marked hypocellularity (<5%) and no evidence of infiltrating tumor. Karyotype was 46,XY in patient 4 and noninformative in patient 3, but FISH was negative for common MDS abnormalities in both. Patient 3 had an elevated fetal hemoglobin level at 9%; patient 4 did not have this checked.

The clinical and family history can be useful for differentiating AA from other BMF states and to identify those with an IBMFD. This is imperative in patients of all ages but perhaps is more acute in pediatric and young adult assessments. Table 1 outlines medical and family history questions that should be asked to screen for the more frequent signs and symptoms of the known IBMFD. Positive history and physical combinations that are suggestive of an IBMFD should warrant a dedicated work-up by an expert clinician or genetic counselor. For example, the combination of AA with a personal or family history of pulmonary fibrosis is strongly suggestive of an underlying short telomere syndrome (STS) with an expected positive yield of genetic evaluation.^60,61 In contrast, a personal history of extensive cutaneous or anogenital warts and lymphedema in a patient with AA would greatly increase suspicion for GATA2 deficiency syndrome.⁶² Several excellent reviews summarize patterns of medical and family history in the known IBMFD.^13,19,63-65  However, it is important to recognize that the usual signs, symptoms, and family history of an IBMFD are absent in up to 40% of cases.¹³ Thus, a high index of suspicion is necessary. Further, given that the expected frequencies of an IBMFD among children and young adults aged ≤40 years with AA are 25% and 5% to 15%, respectively, and syndrome-specific treatment and BMT regimens are available for many of these disorders, screening for these disorders per the Laboratory and Molecular Diagnostics Studies sections in Table 1 is now warranted for all patients with AA aged 40 years and younger. An upper age cutoff remains challenging, because IBMFDs have been diagnosed in individuals into their sixth decade of life. Clinical judgment is necessary until data exist regarding IBMFD frequency in older adults without other syndromic features. Patients with a high suspicion for or with a documented IBMFD likely do not warrant treatment with IST; thus, early diagnosis is imperative for management decisions.

Careful attention to findings that may suggest the presence of an IBMFD, such as severe monocytopenia (eg, GATA2 deficiency, although not specific) or elevated hemoglobin F (seen in multiple IBMFDs) is necessary. Screening for the presence of an STS via peripheral blood lymphocyte telomere length measurement by Clinical Laboratory Improvement Amendments–certified flow cytometry and FISH (Flow-FISH) and for Fanconi anemia (FA) by chromosome breakage studies on peripheral blood lymphocytes is recommended for all adults with AA aged 40 years and younger or any adults with history findings that are suggestive of either of these disorders. Lymphocyte telomere length shorter than the first percentile in the context of AA is highly sensitive and specific for a diagnosis of STS⁶⁶; however, it is possible for individuals with an STS to have lymphocyte telomere lengths within the low normal range (most often in the first to tenth percentile), especially those older than age 40 years.^66-68  Rarely, other IBMFDs have lymphocyte telomere lengths at or just below the first percentile.⁶⁶ Thus, all patients with a lymphocyte telomere length at or below the first percentile and those with a personal and family history suggestive of an STS, regardless of the telomere length, should have a genetic evaluation as suggested in Figure 1. Polymerase chain reaction–based quantification of telomere length is not a reliable measure and should not be used in clinical settings.^68,69 There are reports of patients with acquired AA with short telomeres; however, these may not have used Flow-FISH methodology and may not be less than the first percentile.^68,70,71 Patients with radial chromosomes or increased chromosome breaks on diepoxybutane (DEB) and/or mitomycin C exposure in standard chromosome-breakage assays that are within the range typical for FA warrant genetic evaluation for a molecular diagnosis. In these scenarios, it is critical that a genetic evaluation is completed prior to donor and preparative regimen selection in patients considering BMT to ensure donor and recipient safety.

Our practice is to use a systematic approach to evaluate a patient with AA (Figure 1). Some investigators advocate a stepwise methodology that allows for initial results to come in prior to initiation of a more extensive work-up. However, this approach can be less efficient, even if resource conscience, and may miss those without a classical IBMFD clinical presentation or family history, as is expected for 40% of cases ultimately diagnosed with an IBMFD. It should be noted that estimates of IBMFD prevalence are highly dependent on the specific patient populations studied. The prevalence of IBMFD in young patients referred for genetic testing for clinical suspicion of IBMFD is higher compared with the prevalence in older adults with previously normal blood counts and newly diagnosed AA.^9,12,63 

Using the approach outlined in Figure 1, all patients with AA would undergo the same clinical/family history, basic laboratory tests, and bone marrow studies outlined in Table 1. Patients aged ≤40 years or those proceeding to BMT as initial therapy would also have peripheral blood lymphocyte telomere length measurement by Flow-FISH⁶⁸ and chromosome breakage with DEB to rule out STS and FA, respectively. If the clinical scenario warrants, these tests should be requested immediately with expected 10- to 14-day turnaround times. The benefit of this approach is that it quickly identifies patients at risk for STS and FA who are at high risk for increased toxicity, as described above. However, it is important to note that, although a lymphocyte telomere length shorter than the first percentile is highly sensitive and specific for a STS diagnosis in young patients with AA, other IBMFDs can be associated with lengths falling within this range, making a broad genetic panel recommended over genes of telomere maintenance alone. Further, lymphocyte telomere lengths in adults, especially those older than age 40 years with a genetically confirmed STS, can be well within the normal range, usually between the first and tenth percentiles for age. These caveats can make this a challenging stand-alone screening test, especially for older adults.

If a patient has findings on any of these tests or screening questions that suggest an IBMFD, an IBMFD-specific genetic evaluation at a pace that is appropriate for the clinical scenario should be undertaken. For example, if the patient has very severe AA and needs treatment initiation immediately, genetic evaluation and its associated testing should be requested urgently. This evaluation may be performed by a geneticist, genetic counselor, hematologist, or other clinician who has experience identifying IBMFD-related signs and symptoms in the clinical history and examination, can aid in family member evaluations if needed, and is able to help coordinate and interpret genetic findings from IBMFD screening and genetic tests. Telemedicine consultations may help if access is limited locally.

Genetic testing in patients with AA requires consideration of the tissue to be used for testing, as well as the genes to be tested. There has not been a head-to-head comparison of the use of peripheral blood vs cultured skin fibroblasts for genetic testing in this scenario. However, in many expert centers, cultured skin fibroblasts are preferred because it is known that acquired events can obscure detection of the germline pathogenic variant(s) in patients with IBMFD.⁶³ For example, a somatic reversion event occurring in a population of hematopoietic precursors can completely correct 1 of the 2 pathogenic germline variants in patients with FA, masking the diagnosis if testing is done using peripheral blood.^72-74  SAMD9/SAMD9L, in which the germline mutation–containing allele is lost as part of acquired monosomy 7, is relevant as well.⁷⁵ In addition, the finding of somatic variants suggestive of clonal hematopoiesis (see Molecular diagnostic studies in AA diagnosis) in the blood of patients with AA can lead to confusion if they overlap with genes causative of IBMFD (eg, RUNX1). With multiple commercial laboratories now offering to ship the appropriate media to the clinic, culture the skin, and perform the genetic testing all as part of a single test, lack of access should be less of an issue. Advanced planning to have the correct supplies and someone to do the skin biopsy when the patient is going to be present in the clinic is necessary. Timely referral to an academic center may aid in access.

Except for rare situations in which there is a known pathogenic variant running in a patient’s family or the patient’s clinical picture is highly suggestive of a genetic disorder for which there is a single causative gene (eg, GATA2 deficiency syndrome), a gene panel–based approach will be necessary. The number of genes known in the field to cause BMF in the setting of an IBMFD continues to increase. Even within each specific IBMFD, new genes are increasingly identified to explain the pathobiology of disease, and we continue to search for more in patients in whom a genetic cause remained elusive.⁶³ IBMFDs that have been identified in patients with AA include STS, FA, Schwachman-Diamond syndrome, Diamond-Blackfan anemia, GATA2-deficiency syndrome, as well as newer syndromes. Other hereditary syndromes less commonly thought of as contributors to BMF, such as Li-Fraumeni syndrome (TP53) and familial platelet disorder (RUNX1), have also been observed.^13,63 Therefore using a panel specifically designed for detection of an IBMFD (through an experienced laboratory) as well as assurances of dedicated testing for any genes of particular relevance to your patient remains of key importance. Attention to the details of the testing ordered is vital because false reassurance could come from a test that does not cover the relevant genes for the inherited syndrome; consultation with a genetic counselor, IBMFD expert, or molecular pathologist prior to testing should be considered to ensure selection of the ideal test, as well as after results are received to aid appropriate interpretation.

The rapid pace of discovery and lack of uniform consensus guidelines on when and how to test for IBMFDs in adult AA populations have led to controversy and confusion about the need for screening. We believe that there are many benefits of real-time differentiation of AA from a BMF due to an IBMFD. First and foremost are the lack of expected benefit and the potential for harm from the use of IST in patients with an IBMFD, thus making a thorough diagnosis imperative. Additionally, some IBMFDs are characterized by extrahematopoietic manifestations that can contribute to significant morbidity and for which screening can change management (Table 1). Recognition of the possibility of these multiorgan system implications and treatment interactions is only the first step. Subspecialist referrals and engagement of additional resources should follow.⁷⁶ Second, knowledge of the etiology provides insight into the expected natural history of the disorder and can impact interpretation of bone marrow findings. For example, dysplasia may be part of the underlying IBMFD-altered hematopoiesis, making differentiation of BMF and hypocellular MDS in this context syndrome specific. Third and most importantly, diagnosis of an IBMFD can facilitate more appropriate therapy in an affected individual. For example, awareness of the potential for severe systemic toxicity in FA and STS alters selection of the BMT preparative regimen. Potential lack of responsiveness to immunosuppression and other noncurative therapies for BMF due to an IBMFD should prompt earlier BMT evaluation. Lastly, prompt recognition of an IBMFD in the patient with BMF allows cascade testing of at-risk relatives who may be willing to serve as a donor for matched sibling or haploidentical BMT.

Given these benefits, at a minimum, screening for STS and FA is needed in patients younger than age 40 years or those proceeding to BMT, because they are at risk for harm from the conditioning for BMT if undiagnosed.^77,78 Notably, these tests include peripheral blood lymphocyte telomere length measurement by Flow-FISH⁶⁸ and chromosome breakage with DEB, as detailed above.⁷⁹ The significant phenotypic overlap and the challenge of somatic events that may mask IBMFD diagnoses using these peripheral blood–based tests alone suggest that all AA patients aged ≤40 years and those proceeding to transplant should have a germline IBMFD genetic panel test. In several academic centers where access to this testing is readily available, this strategy is already in place.

Patient 3 was ultimately diagnosed with acquired SAA in the setting of negative DEB testing, normal telomere lengths by Flow-FISH, and a skin biopsy negative for known genes associated with IBMFDs. He underwent a matched sibling donor bone marrow transplant and is well, with normal blood counts at 5 years postprocedure.

Patient 4 had negative DEB testing and normal telomere lengths by Flow-FISH; however, a skin biopsy grown for fibroblast culture for testing of known genes associated with an IBMFD revealed biallelic mutation of the MPL gene as the presumed etiology for his SAA. He underwent an unrelated donor BMT and is well, with normal blood counts at 3 years postprocedure.

Efficient and accurate diagnosis of AA are critical to proper management of this disease. The severity of the illness also warrants a clear path from accurate diagnosis to the most appropriate treatment option. Germline predisposition to AA, even in teenagers and adults, is more common than previously recognized and important to diagnosis in real time prior to treatment and, especially, prior to BMT. Consistent work-up of all patients with AA, as outlined here, will facilitate high-quality care and foster research.

Contribution: A.E.D. and J.E.C. wrote the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Amy E. DeZern, 1650 Orleans St, CRBI Room 3M87, Baltimore, MD 21287; e-mail: adezern1@jhmi.edu.