This page is an English translation of the Korean original. If there is any discrepancy, the Korean version prevails.
| Item | Details |
|---|---|
| A. Classification codes | ICD-10: D61.1 (drug-induced aplastic anemia) | Special Copayment Reduction: V283 (severe intractable disease, 5 years) |
| B. Synonyms | Drug-induced aplastic anemia, drug-induced bone marrow failure, drug-induced pancytopenia |
| C. Adult onset | Likelihood of adult onset: very high — it can occur at any age, but occurs more frequently in adults and the elderly taking causative drugs (antibiotics, NSAIDs, antiepileptics, anticancer drugs, etc.). |
| 1. History | 1934: first report of chloramphenicol-associated aplastic anemia → 1950s: concept of reactivity established → expansion of the list of causative drugs over decades → introduction of bone marrow transplantation and immunosuppressive therapy. |
| 2. Causes | Chloramphenicol, antiepileptics (carbamazepine, phenytoin), NSAIDs, sulfonamides, gold compounds, D-penicillamine, quinine, ticlopidine, etc.; immune-mediated and direct toxicity mechanisms. |
| 3. Genetics | HLA-B*15:02 (carbamazepine), HLA-B*58:01 (allopurinol), CYP enzyme polymorphisms, GSTM1 deletion; genotypes vulnerable to immune reactions. |
| 4. General epidemiology | Overall incidence of aplastic anemia 2–3 per million per year (West), 3–7 per million per year in Asia; drug-induced causes account for about 10–25% of the total. |
| 5. Pathophysiology | T-cell-mediated autoimmune attack by drugs or reactive metabolites → destruction of hematopoietic stem cells; mechanisms of direct toxicity (anticancer drugs), hapten formation, and oxidative stress. |
| 6. Symptoms | Progressive fatigue and weakness, anemia symptoms (pallor, dyspnea), bleeding due to thrombocytopenia (petechiae, ecchymosis, epistaxis), fever and infection due to neutropenia. |
| 7. Signs | Pancytopenia (decreased hemoglobin, thrombocytopenia, neutropenia), reduced reticulocytes, bone marrow hypoplasia, pale skin and hemorrhagic signs. |
| 8. Screening | Regular CBC monitoring in patients receiving high-risk drugs (chloramphenicol, antiepileptics, gold compounds); immediate discontinuation of the drug at the first cytopenia. |
| 9. Diagnosis | Peripheral blood smear (pancytopenia, reduced reticulocytes), bone marrow aspiration and biopsy (hypocellular marrow, fat replacement), flow cytometry (PNH clone), chromosome testing, confirmation of history of causative drug use. |
| 10. Treatment | Immediate discontinuation of the causative drug; allogeneic hematopoietic stem cell transplantation (age ≤40, first-line with a matched donor); immunosuppressive therapy (ATG + cyclosporine + erythropoietin + G-CSF); addition of eltrombopag. |
| 11. Prognosis | Recovery is unlikely with discontinuation of the causative drug alone; 5-year survival after bone marrow transplantation 75–90%; complete response rate of immunosuppressive therapy about 60–70%; risk of relapse and late clonal disease. |
| 12. Prevention | Regular CBC monitoring when giving high-risk drugs; HLA pharmacogenomic testing (for relevant drugs and ethnicities); preferential use of alternative drugs. |
| 13. References | Young NS. NEJM 2018 and 9 other papers |
The history of drug-induced aplastic anemia is closely tied to the antibiotic chloramphenicol. In 1934, Amundsen in Sweden first reported aplastic anemia occurring after taking a sulfonamide drug, and after chloramphenicol was introduced into clinical practice in 1947, an association with bone marrow toxicity soon began to be reported. By the 1950s, the causal relationship between chloramphenicol and aplastic anemia was epidemiologically established, and it was found that chloramphenicol causes bone marrow toxicity by two mechanisms. The first is dose-dependent reversible bone marrow suppression (inhibition of mitochondrial protein synthesis), and the second is very rare dose-independent irreversible aplastic anemia, which is explained by oxidative toxicity of the nitroso metabolite of chloramphenicol or by an immune-mediated mechanism.
In the 1960s–70s, various drugs were reported as causes of aplastic anemia. Phenylbutazone (a nonsteroidal anti-inflammatory drug), gold compounds (rheumatoid arthritis treatment), D-penicillamine, antiepileptics (phenytoin, carbamazepine), sulfonamides, and antithyroid drugs (carbimazole, propylthiouracil) were identified as the main causative drugs. In 1985, the International Aplastic Anemia Study Group was formed and large-scale epidemiological studies began, and the relative risks of specific drugs began to be quantified.
After the 1990s, an immune-mediated mechanism was established as the main mechanism of drug-induced aplastic anemia. A model gained support in which the drug or its reactive metabolite binds to hematopoietic stem cell surface proteins to form a hapten, and T cells that recognize it as a self-antigen attack the hematopoietic stem cells, resulting in bone marrow failure. At the same time, a pharmacogenomic perspective also developed in which an individual's drug-metabolizing enzyme (CYP family) polymorphisms and HLA genotype determine individual susceptibility to adverse drug reactions.
In terms of treatment, the introduction of allogeneic bone marrow transplantation (allo-BMT) in the 1970s dramatically improved the prognosis of severe aplastic anemia. In 1977, antithymocyte globulin (ATG) was introduced as an immunosuppressive agent, and combination therapy with cyclosporine was subsequently standardized. Since 2014, eltrombopag (a TPO receptor agonist) has been added to immunosuppressive therapy and improved response rates, and treatment research continues to this day.
The causative drugs of drug-induced aplastic anemia are broadly classified into high-risk drugs (high relative risk) and moderate-risk drugs. By mechanism, they are divided into direct toxicity (dose-dependent), immune-mediated (idiosyncratic), and cases in which both mechanisms are mixed.
Chloramphenicol is historically the best-known causative drug. The risk of aplastic anemia in users is estimated at about 1:20,000–1:40,000, which is 2–13 times that of the general population. The nitroso metabolite of chloramphenicol (nitroso-chloramphenicol) directly damages the mitochondrial DNA of hematopoietic stem cells or binds to bone marrow stromal cell surface proteins to cause immune-mediated bone marrow injury. Its use has declined markedly in developed countries, but in developing countries it is still used as an ophthalmic, otic, and topical antibiotic. The sulfonamide class is also associated with aplastic anemia, and trimethoprim-sulfamethoxazole in particular is risky with long-term administration.
Phenylbutazone is the NSAID with the highest risk, and its use has been restricted in most countries. Indomethacin, diclofenac, and ibuprofen are also rarely associated with aplastic anemia. The mechanism of NSAIDs is explained mainly as immune-mediated or oxidative DNA damage by reactive metabolites.
Carbamazepine, phenytoin, and phenobarbital are reported, though rarely, as causes of aplastic anemia. The epoxide metabolite of carbamazepine may be directly toxic to bone marrow cells. Cases of aplastic anemia occurring together with carbamazepine-induced severe cutaneous adverse reactions (SJS/TEN) in Asian patients have been reported and are associated with the HLA-B*15:02 allele.
Gold compounds and D-penicillamine were drugs used to treat rheumatoid arthritis and can cause immune-mediated bone marrow suppression. Ticlopidine and clozapine are also associated with aplastic anemia or granulocytopenia. Antithyroid drugs (methimazole, propylthiouracil), quinine, quinidine, allopurinol, and some antimalarials have also been reported as causative drugs. Anticancer drugs (alkylating agents, anthracyclines, etc.) cause aplastic anemia through dose-dependent direct bone marrow toxicity and may be classified under codes other than D61.1.
Drug-induced aplastic anemia shows a much higher risk of occurrence in individuals with a particular genetic background. This is because polymorphisms of drug-metabolizing enzymes, HLA genotype, and variants of immune regulatory genes are involved in combination.
Particular HLA alleles present drug-protein complexes (haptens) immunogenically and induce T-cell activation. HLA-B*15:02 is associated with carbamazepine-induced severe cutaneous reactions (SJS/TEN) in Southeast Asian, Korean, and Chinese populations, and also raises the risk of bone marrow toxicity. In Taiwan, Hong Kong, and Korea, HLA-B*15:02 testing before carbamazepine administration is recommended (or mandated). HLA-B*58:01 is associated with allopurinol-induced severe cutaneous reactions and bone marrow toxicity. As for HLA associations of aplastic anemia itself, HLA-DR2 (DRB1*15:01) is overexpressed in aplastic anemia patients, suggesting a genetic susceptibility to autoimmune T-cell responses.
Polymorphisms of CYP2C19, CYP2D6, CYP3A4, etc. affect the amount of reactive metabolites of drugs produced. The enzyme activity involved in forming the nitroso metabolite of chloramphenicol differs between individuals, so exposure to toxic metabolites differs even at the same dose. GSTM1 and GSTT1 (glutathione S-transferase) deletion genotypes reduce the ability to detoxify reactive drug metabolites and raise the risk of oxidative bone marrow injury. N-acetyltransferase (NAT2) polymorphism affects toxicity risk in relation to sulfonamide metabolism.
Genes involved in T-cell immune regulation (CTLA-4, IL-2, TNF-α promoter polymorphisms) contribute to susceptibility to drug-induced autoimmune bone marrow injury. Telomerase gene (TERT, TERC, DKC1) variants increase susceptibility to bone marrow failure in some aplastic anemia patients in relation to reduced replicative capacity of hematopoietic stem cells.
The overall incidence of aplastic anemia is about 2–3 per million per year in Western countries and about 3–7 per million per year in Asia (East Asia), higher than in the West. The proportion due to drugs (D61.1) is reported, depending on the study, at about 10–25% of all aplastic anemia. In the International Aplastic Anemia Study (IAAAS), the relative risk of aplastic anemia in NSAID users was about 5–15-fold. For chloramphenicol, irreversible aplastic anemia is estimated to occur in 1 of every 20,000–40,000 users. Reasons proposed for the higher incidence of aplastic anemia in Asia include frequency of chloramphenicol use, high frequency of particular HLA genotypes, and differences in the frequency of aplastic anemia after hepatitis virus infection (hepatitis-associated AA). In Korea as well, according to Health Insurance Review & Assessment Service (HIRA) data, the annual incidence of aplastic anemia is estimated at about 5–6 per million per year; the proportion that is drug-induced has not been precisely counted but is presumed to account for a considerable share.
The pathophysiology of drug-induced aplastic anemia is explained by three main pathways.
This is the most common mechanism: the drug or its reactive metabolite covalently binds to surface proteins of hematopoietic stem cells or stromal cells (hapten formation) or alters protein structure (altered self), and is recognized as a self-antigen. CD8+ cytotoxic T cells (CTL) presented with it via MHC class I molecules are activated and directly destroy hematopoietic stem cells through the Fas/FasL and perforin/granzyme pathways. Activated T cells secrete cytokines such as IFN-γ and TNF-α, inhibiting hematopoietic stem cell proliferation and inducing apoptosis. Reduced function of regulatory T cells (Treg) results in failure to suppress autoimmune T cells.
Nitroso metabolites of chloramphenicol, benzene metabolites, and reactive oxygen species (ROS) directly damage the mitochondrial DNA or nuclear DNA of hematopoietic stem cells and induce cell death. This tends to be dose-dependent, and recovery is mostly reversible when the drug is stopped. However, when progression to irreversible bone marrow failure occurs, the cause is a decrease in hematopoietic stem cell number below a critical threshold due to direct toxicity.
Some drug metabolites induce oxidative stress and accelerate telomere shortening. The shorter the telomeres of hematopoietic stem cells, the lower their self-renewal capacity and the more vulnerable they are to bone marrow failure. In patients with telomerase gene variants, drug-induced oxidative stress may accelerate loss of hematopoietic stem cells.
The clinical symptoms of drug-induced aplastic anemia present as three symptom groups caused by pancytopenia (anemia, bleeding, infection). Symptoms often appear within weeks to months after starting the causative drug, but with chloramphenicol they may appear after several months.
Progressive fatigue, exercise intolerance, dyspnea, palpitations, and pallor are the main symptoms. In moderate or worse anemia, orthostatic dizziness, presyncope, and impaired concentration accompany. In severe anemia (Hb <7 g/dL), dyspnea at rest, angina (in patients with underlying cardiovascular disease), and signs of congestive heart failure may appear.
Bleeding due to thrombocytopenia is characteristic. Cutaneous petechiae, ecchymosis, gum bleeding, and epistaxis are common. In moderate or worse thrombocytopenia (<50,000/μL), menorrhagia (in women), gastrointestinal bleeding (hematochezia, melena), and hematuria may occur. In severe thrombocytopenia (<20,000/μL) there is a risk of spontaneous intracranial hemorrhage, which is the most fatal complication.
Neutropenia (<1,000/μL) markedly increases the risk of bacterial and fungal infection. Fever is the main symptom of infection, and fever in a neutropenic state is a hematologic-oncologic medical emergency. Pneumonia, sepsis, oral mucositis, and perianal infection are the main complications. In severe neutropenia (<200/μL), the risk of invasive fungal infection (aspergillosis, candidemia) increases significantly.
Diagnosis is supported by physical findings and laboratory findings. Pancytopenia is the key laboratory finding, with decreased hemoglobin, thrombocytopenia, and decreased absolute neutrophil count (ANC) occurring together. The peripheral blood smear may show normal red cell morphology (normocytic normochromic anemia) or macrocytosis, and the reticulocyte count is markedly reduced (corrected reticulocyte count <1%), reflecting failure of erythropoiesis in the bone marrow. On physical examination, pale skin, petechiae, ecchymosis, and oral mucosal bleeding are found. Splenomegaly and hepatomegaly are generally absent (important for distinguishing from infiltrative disease), and there is no lymphadenopathy. On bone marrow biopsy, a hypocellular marrow and replacement by fat tissue are characteristic. When bone marrow cellularity is below 25% or residual hematopoietic cells are below 30%, the diagnostic criteria for aplastic anemia are met.
Screening of the general population for drug-induced aplastic anemia is not cost-effective and is not performed. Regular complete blood count (CBC) monitoring in patients receiving high-risk drugs is the key to early detection. Chloramphenicol: avoid use when possible, and when unavoidable, monitor CBC before and during administration (once or twice a week). Antiepileptics (carbamazepine, phenytoin): perform CBC every other week for the first 3 months after starting, then every 3 months. HLA-B*15:02 testing before carbamazepine administration is recommended in Asian patients. Clozapine: in many countries (the US, Europe), weekly ANC monitoring is mandatory before administration and for 18 weeks after. Gold compounds, D-penicillamine: regular CBC monitoring during administration. If a rapid fall in neutrophil count (ANC <1,500/μL) or thrombocytopenia is found on CBC, immediate drug discontinuation and hematology consultation are needed.
Diagnosis of aplastic anemia consists of differentiating the causes of pancytopenia and confirming bone marrow hypoplasia. A drug-induced cause is confirmed by temporal association and drug history.
Peripheral blood tests: On CBC, confirm pancytopenia (decreased Hb, thrombocytopenia, decreased ANC), reduced absolute reticulocyte count (<20,000/μL or corrected reticulocyte percentage <1%), and normal MCV or macrocytosis. Confirm on the peripheral blood smear that there is no dysplasia or increase in blasts, to differentiate myelodysplastic syndrome (MDS). Bone marrow examination: Bone marrow aspiration and biopsy are the standard for diagnosing aplastic anemia. Confirm on bone marrow biopsy tissue cellularity <25% (severe) or <30% residual hematopoietic cells (very severe), fat replacement, and no dysplasia. CD34+ hematopoietic stem cell numbers are markedly decreased. Differential diagnostic tests: To exclude other causes of pancytopenia, flow cytometry (detection of PNH clones: proportion of CD55- and CD59-negative cells), vitamin B12 and folate, liver function, viral tests (EBV, CMV, parvovirus B19, hepatitis A, B, C), immunoglobulins (autoimmune hemolysis), ANA, etc. are performed. Telomere length measurement and Fanconi anemia screening (chromosome breakage analysis) are also considered. The temporal association between starting the causative drug and onset of pancytopenia is important for diagnosis.
The most important first step in drug-induced aplastic anemia is immediate discontinuation of the causative drug. However, only some cases recover spontaneously with drug discontinuation alone, and active treatment is needed in severe or very severe aplastic anemia.
According to the International Aplastic Anemia Study Group (IAAAS) criteria, cases are classified as severe (SAA): bone marrow cellularity <25% + at least 2 of 3 peripheral blood items (ANC <500/μL, platelets <20,000/μL, reticulocytes <20,000/μL), very severe (vSAA): SAA + ANC <200/μL, and non-severe.
It is the first-line treatment for severe/very severe SAA when the patient is 40 years or younger (or 50 or younger in selected patients) and has a fully HLA-matched sibling donor. The 5-year survival after transplantation is about 75–90%. Transplantation from an HLA-matched unrelated donor is also used as second-line treatment after failure of first-line immunosuppressive therapy.
It is the standard treatment for patients who cannot undergo transplantation (older age, no donor). Antithymocyte globulin (hATG or rATG) + cyclosporine + eltrombopag (since 2014) is the current standard first-line IST. In an NIH study, the combination of hATG + cyclosporine + eltrombopag improved the complete response rate compared with hATG + cyclosporine alone (complete response rate about 58% vs. 22%). Response appears within 3–6 months after treatment, and if response is poor, second-line immunosuppressive therapy or HCT is considered.
Supportive treatment is given alongside, including red blood cell transfusion (Hb <7–8 g/dL or when symptomatic), platelet transfusion (platelets <10,000/μL or when bleeding), prophylactic antibiotics and antifungals, and G-CSF (to promote recovery and reduce infection risk).
Spontaneous recovery with discontinuation of the causative drug alone is possible in some mild drug-induced bone marrow suppression, but when it has progressed to severe aplastic anemia, mortality without treatment within 1 year reaches about 80%. In treated severe aplastic anemia, 5-year survival after allogeneic hematopoietic stem cell transplantation is about 75–90% (with young age and a fully HLA-matched donor), and 5-year survival after immunosuppressive therapy (IST) is about 60–80%. The complete response rate to IST is about 30–60%, and the overall response rate including partial response is about 60–80%. Major long-term complications after IST include relapse (about 30–40%), clonal evolution to MDS/AML (about 10–15%), and evolution to paroxysmal nocturnal hemoglobinuria (PNH), so long-term follow-up is needed. Poor prognostic factors include old age, very severe disease (ANC <200), treatment delay, concomitant infection, and HLA-mismatched transplantation.
Primary prevention of drug-induced aplastic anemia lies in careful use of causative drugs and selection of alternatives. Chloramphenicol should be avoided when an alternative antibiotic exists, and when unavoidable used for the minimum period. Topical application to the eye, ear, etc. should also be done with caution considering the possibility of systemic absorption. CBC monitoring when giving high-risk drugs: perform regular CBC in patients receiving carbamazepine, clozapine, gold compounds, D-penicillamine, and NSAIDs (especially the phenylbutazone class) to detect early cytopenia. Pharmacogenomic testing: in Korean and Asian patients, HLA-B*15:02 testing before carbamazepine administration identifies the high-risk group for severe adverse reactions in advance so that an alternative drug can be chosen. Minimizing drug interactions: avoid concomitant use of hepatotoxic or myelotoxic drugs, and when unavoidable, monitor closely. Patient education: educate patients taking high-risk drugs to visit a medical institution immediately when early symptoms such as fatigue, fever, bruising, and bleeding appear.