Treatment of relapsed or refractory acute myeloid leukemia

INTRODUCTION — The first step in the successful treatment of a patient with acute myeloid leukemia (AML) is the attainment of a complete remission (CR), which is characterized by (table 1):

●Normal values for absolute neutrophil count (>1000/microL) and platelet count (>100,000/microL), and independence from red cell transfusion.

●A bone marrow biopsy that reveals no clusters or collections of blast cells. Extramedullary leukemia (eg, central nervous system or soft tissue involvement) must be absent.

●A bone marrow aspiration revealing normal maturation of all cellular components (ie, erythroid, granulocytic, and megakaryocytic series). There is no requirement for bone marrow cellularity.

●Less than 5 percent of blast cells present in the bone marrow, and none with a leukemic phenotype (eg, Auer rods). The persistence of dysplasia is worrisome as an indicator of residual AML but has not been validated as a criterion for remission status.

●For patients who had an abnormal karyotype in their AML cells, reversion back to a normal karyotype in their post-treatment marrow cells provides further evidence of a good response, but this is not a criterion for CR.

Twenty to 30 percent of young adult patients and 50 percent of older adult patients with newly diagnosed AML will fail to attain a CR with intensive induction chemotherapy due to drug resistance or death. In addition, a percentage of patients who initially attain a CR will relapse. Treatment of the patient with AML who has failed to enter into CR or who has relapsed after attainment of CR will be reviewed here. The prognosis and initial treatment of the patient with AML are discussed separately:

●(See "Acute myeloid leukemia: Induction therapy in medically fit adults".)

●(See "Acute myeloid leukemia: Management of medically unfit adults".)

●(See "Acute myeloid leukemia: Risk factors and prognosis".)

●(See "Acute myeloid leukemia: Induction therapy in medically fit adults", section on 'Introduction'.)

EVALUATION FOR RELAPSE OR RESISTANCE — After attaining a complete remission (CR) with initial induction and consolidation therapy, patients are followed at routine intervals to monitor for treatment-related complications and relapse. (See "Acute myeloid leukemia in younger adults: Post-remission therapy", section on 'Monitoring for relapse'.)

Definitions of refractory disease and relapse — The following definitions pertain to AML:

●Refractory (resistant) disease – Failure to achieve a CR (ie, failure to eradicate all visible leukemia cells and achieve less than 5 percent blasts in the bone marrow and blood) is described as refractory disease (also referred to as primary induction failure) [1].

Patients who require a second cycle of induction therapy to achieve CR are not considered to have refractory disease [1,2].

Achievement of <5 percent marrow blasts, but failure to restore normal hematopoiesis (normal peripheral blood counts) is referred to as CRi (CR with incomplete hematological recovery); this is not considered refractory disease.

Resistance to induction chemotherapy is closely associated with unfavorable cytogenetic and molecular features and poor overall outcome. (See "Acute myeloid leukemia: Induction therapy in medically fit adults", section on 'Introduction'.)

●Measurable residual disease – The term measurable residual disease (MRD; also called minimal residual disease) refers to persistence of leukemia cells that cannot be identified by light microscopy. The sensitivity of techniques for detection of MRD (eg, multiparameter flow cytometry, polymerase chain reaction, DNA sequencing) is greater than that of bone marrow morphology, but the standards for defining MRD depend on the specific methodology that is utilized. (See "Acute myeloid leukemia: Induction therapy in medically fit adults".)

●Relapsed disease – Relapse describes the reappearance of leukemia cells in the bone marrow, peripheral blood, or elsewhere (extramedullary disease) after the attainment of a CR. The significance of the reappearance of AML detected only by polymerase chain reaction (PCR) analysis is uncertain. (See "Acute myeloid leukemia in younger adults: Post-remission therapy", section on 'Monitoring for relapse' and "Acute myeloid leukemia: Induction therapy in medically fit adults".)

Evaluation of suspected relapse — Patients suspected of having relapsed disease should be evaluated with a unilateral bone marrow aspiration. The use of a trephine biopsy for all patients is controversial. A trephine biopsy is required if the marrow is inaspirable and may add useful diagnostic information above that obtained from a bone marrow aspiration alone in other patients. This sample should be sent for pathologic review, immunophenotyping, cytochemistry, and cytogenetics. Mutation analysis for CEBPA, FLT3, TP53, RUNX1, ASXL1, IDH1/2, and NPM1 mutations should be performed, if available, even if they were performed at the time of initial diagnosis since they may have changed at the time of relapse [1,3]. These mutations provide prognostic information and may identify targets for novel agents (table 2). (See "Acute myeloid leukemia: Risk factors and prognosis".)

Once relapsed or refractory disease is identified, further evaluation should include:

●An assessment of performance status (table 3 and table 4).

●Human leukocyte antigen (HLA) typing should be performed for patients who are candidates for hematopoietic cell transplantation (HCT). If no sibling donor can be identified, an alternative donor search (including unrelated volunteer donors and cord blood stem cells) should be initiated. This is extremely important since HCT is the most likely method for curing patients with relapsed AML. (See 'Determining transplant eligibility' below.)

●Other laboratory studies include a complete blood count with differential, chemistries with liver and renal function and electrolytes, glucose, lactate dehydrogenase (LDH), calcium, phosphorus, uric acid, albumin and total protein, serum lysozyme, and serology for hepatitis B, HIV, herpes simplex virus (HSV) and cytomegalovirus (CMV) infection. An elevation in serum or urinary lysozyme levels may explain renal dysfunction in monocytic and myelomonocytic leukemia. Women of childbearing age should undergo a pregnancy test.

●A chest X-ray, pulmonary function tests, an electrocardiogram (EKG) and a study of cardiac function (eg, ejection fraction measured by echocardiogram or MUGA) should be performed.

●Dental evaluation with X-rays for possible infection foci is warranted. CT scans of the sinuses and chest may detect clinically occult sites of infection, although the yield in asymptomatic patients is unknown.

●Patients with neurologic signs or symptoms should undergo imaging studies to evaluate for meningeal disease, or central nervous system bleeding. Lumbar puncture is indicated in patients with neurologic signs or symptoms to examine the cerebrospinal fluid (CSF) for leukemic involvement. Care must be taken to avoid contaminating the CSF specimen with peripheral blood if blasts were present. CSF should be sent for both cytology (examination of stained cytospin slides) and flow cytometry for detection of AML blasts.

SELECTION OF THERAPY — The therapy that provides the best chance to cure a patient with relapsed or refractory AML is an allogeneic hematopoietic cell transplantation (HCT), and the best outcomes appear to be with myeloablative preparative regimens administered after attaining a complete remission (CR). However, some patients may be cured with myeloablative HCT despite residual disease at the time of HCT. In addition, nonmyeloablative preparative regimens may be considered for patients who are not candidates for myeloablative HCT, but have attained a CR. Autologous marrow transplantation is an option for those patients who achieve a second CR. A physician who specializes in transplantation should be consulted at the time relapsed or refractory disease is identified to help guide this decision process.

The decision of whether to attempt a second remission induction prior to HCT depends on several factors, such as age, comorbidities, duration of first remission, cytogenetic features, access to an experienced transplant center, resources for home care, and of course, the patient's own wishes. Important questions include:

●How likely is the patient to attain a CR? Patients who undergo HCT in CR appear to have better outcomes. Patients who previously attained a CR that lasted longer than six months have a greater than 50 percent chance of attaining a second CR with further chemotherapy. In contrast, of patients with initially refractory disease and those with an initial CR lasting less than six months, 20 percent or fewer will attain a second CR with chemotherapy alone. (See 'Prognostic indices' below.)

●Is an HLA-matched donor immediately available? If a matched donor is not available for immediate collection, chemotherapy administered in the attempt to achieve a CR may control the disease while a search for a compatible donor is undertaken. (See "Donor selection for hematopoietic cell transplantation".)

●Is the patient a candidate for a myeloablative preparative regimen? Myeloablative HCT eradicates tumor cells through both a direct cytotoxic effect and a graft-versus-leukemia (GVL) effect. In contrast, nonmyeloablative preparative regimens rely in large part upon an immunological GVL effect alone. Nonmyeloablative HCT is not appropriate for patients with refractory or rapidly progressive AML because it can take several months for a GVL effect to develop post-transplant. (See 'Nonmyeloablative (reduced intensity) HCT' below.)

●How did the patient tolerate induction therapy before? Attempts at a second induction may result in toxicity that limits the ability to proceed with a potentially curative HCT.

●Does the patient have active infections (ie, aspergillosis) that are associated with a high mortality rate at HCT? (See "Evaluation for infection before hematopoietic cell transplantation".)

There have been no randomized trials of HCT preceded or not by chemotherapy administered in an attempt at inducing a second remission; as such, clinical practice varies. HCT outcomes are better if no detectable leukemia is present at the time of transplantation.

Some experts choose to administer or not administer re-induction chemotherapy prior to myeloablative HCT based on the likelihood of achieving a second CR [4,5]. With this approach, a young patient who had an initial prolonged CR may be given further chemotherapy prior to HCT because more than half of such patients would be expected to attain a second CR. In contrast, further chemotherapy may be deferred for a patient with an initial CR lasting less than six months since it may only prove to worsen the functional status and decrease the chance of a successful HCT. Instead, it may be reasonable to proceed to HCT without prior chemotherapy if the patient is a good candidate for a myeloablative HCT and a donor is available. If the patient is not a candidate for a myeloablative HCT, then a CR should be attained prior to nonmyeloablative HCT. (See 'Myeloablative HCT' below.)

A pretransplant scoring system has been devised to predict the survival of patients not in CR who undergo myeloablative HCT based on the assessment of five criteria (maximum score 6 points) [6]:

●Duration of CR: First CR ≤6 months = 1 point

●Cytogenetics prior to HCT: Poor risk = 1 point

●Donor: Mismatched unrelated donor = 1 point; related donor other than HLA identical sibling = 2 points

●Circulating blasts: Present = 1 point

●Karnofsky performance status: <90 = 1 point

Total score predicted long-term survival when applied to 1673 patients with relapsed or refractory AML not in CR at the time of myeloablative HCT. Rates of three-year overall survival were 42, 28, 15, and 6 percent for patients with 0, 1, 2, and 3 points, respectively. Given these data, some clinicians may choose to proceed with myeloablative HCT without further induction chemotherapy if the patient has a score ≤2 [6]. For patients with a score of ≥3, myeloablative HCT is unlikely to be of benefit; further induction chemotherapy may be administered to these patients in an attempt to decrease the tumor burden and improve the performance status prior to reconsideration of HCT. Alternatively, such patients may be considered for a clinical trial of an investigational agent. (See 'Clinical trials' below.)

REMISSION RE-INDUCTION

Selection of therapy — Our approach is to administer the most active and least toxic regimen to achieve disease control and enable prompt allogeneic hematopoietic cell transplantation (HCT). There is no consensus regarding the optimal regimen for achieving remission in patients with relapsed or refractory AML. The choice will vary based on the presence of a targetable mutation (eg, IDH1, IDH2, FLT3), medical fitness, prior treatments, and patient preference. The urgency of achieving disease control may also influence the choice of therapy, because targeted agents may achieve remission more slowly than intensive chemotherapy.

Treatment with an investigational agent should be considered for AML with a highly unfavorable genotype (eg, monosomal karyotype), which is less likely to respond to standard chemotherapy regimens, and for patients who do not have potentially targetable mutations and are not medically fit for intensive therapy. (See 'Clinical trials' below.)

Targeted agents — The following targeted therapies offer an acceptable alternative to intensive chemotherapy for patients with AML that harbors the following mutations:

●IDH1 mutation – Inhibitors of mutated IDH1 (ie, ivosidenib, olutasidenib) are effective for relapsed or refractory AML that harbors mutant IDH1. Both ivosidenib and olutasidenib have been associated with complete responses in more than one-third of patients with relapsed or refractory AML with mutated IDH1; treatment may cause cytopenias, differentiation syndrome (DS), or other adverse effects (AEs). No study has directly compared these two agents.

•Ivosidenib – A multicenter study that included 179 patients with relapsed/refractory AML who were treated with ivosidenib (500 mg daily) reported 42 percent overall response rate (ORR), including 30 percent complete remission (CR) plus CR with partial hematologic response (CRh) [7]. Median duration of response was >8 months, but some responses were durable for more than one year. Transfusion independence was attained in more than one-third of patients. Treatment was well tolerated, with grade ≥3 AEs that included QT interval prolongation in 8 percent and IDH differentiation syndrome in 4 percent.

•Olutasidenib – Olutasidenib was associated with 48 percent ORR (including 35 percent with CR or CRh) among 147 evaluable patients with relapsed/refractory AML and mutated IDH1 in another multicenter study [8]. Median OS was 11.6 months, median time to response was 2 months, median duration of CR/CRh was 26 months, and the response was maintained in 78 percent at 6 months and in 63 percent at 12 months. One-third of patients became transfusion-independent. Response rates were similar in patients who had or who had not received prior venetoclax. Grade ≥3 AEs included febrile neutropenia (20 percent), cytopenias (up to 20 percent), abnormal liver function (15 percent), and DS (9 percent).

In a dose-escalation study of patients with relapsed/refractory AML or higher-risk MDS, olutasidenib monotherapy was associated with 41 percent ORR in 22 patients, while olutasidenib plus azacitidine was associated with 46 percent ORR in 26 patients [8]. As a single agent, grade ≥3 treatment-emergent AEs were thrombocytopenia (28 percent), febrile neutropenia (22 percent), and anemia (22 percent); similar AEs occurred when olutasidenib was used together with azacitidine.

Ivosidenib is approved by the US Food and Drug Administration (FDA) for adults ≥75 years old with AML that has a susceptible IDH1 mutation, detected by an FDA-approved diagnostic test [9] and was approved by the EMA for treatment of adults with newly diagnosed AML with an IDH1 R132 mutation who are not eligible to receive standard induction chemotherapy.

Olutasidenib is approved by the FDA for adult patients with relapsed/refractory AML with a susceptible IDH1 mutation, as detected by an FDA-approved test [10].

FDA labels for both ivosidenib and olutasidenib carry a boxed warning regarding IDH differentiation syndrome, which may require prompt treatment with steroids, other interventions, and/or drug withdrawal until symptoms have resolved, as described separately. (See "Differentiation syndrome associated with treatment of acute leukemia".)

●IDH2 mutation – Enasidenib is effective for relapsed or refractory AML with IDH2 mutation; treatment is associated with IDH differentiation syndrome [11]. A multicenter study of enasidenib (100 mg daily by mouth) in 214 adults reported 40 percent OR (nine month median survival and six month median duration of response), including CR in 19 percent (20 month median survival) [12,13]. Median overall survival (OS) was nine months and median time to first response was two months. IDH differentiation syndrome was reported as grade ≥3 in <10 percent of patients, but two deaths may have resulted from this condition. Other grade ≥3 treatment-related adverse reactions that occurred in >20 percent of patients included elevated bilirubin and nausea, but severe hematologic toxicities and infections occurred in 10 percent and 1 percent of patients, respectively.

Enasidenib is approved by the FDA for treatment of relapsed or refractory AML with an IDH2 mutation (eg, detected by the companion diagnostic, the RealTime IDH2 Assay) [14]. An EMA application is pending. FDA prescribing information includes a boxed warning and instructions on the risk of differentiation syndrome and need for prompt intervention. (See "Differentiation syndrome associated with treatment of acute leukemia".)

●FLT3 mutation – Gilteritinib is effective for patients with FLT3 mutation; important toxicities include cytopenias and differentiation syndrome. In a phase 3 trial, compared with salvage chemotherapy, gilteritinib achieved longer survival and higher rates of remission for patients with FLT3-mutated AML [15]. In this trial, 371 adults with relapsed or refractory FLT3 ITD, D835, or I836 mutated AML were randomly assigned (2:1) to gilteritinib (120 mg daily by mouth) versus salvage chemotherapy (prerandomization selection of low dose cytarabine, azacitidine, or fludarabine/cytarabine/G-CSF/idarubicin). Compared with salvage chemotherapy, gilteritinib achieved superior median OS (9.3 versus 5.6 months; hazard ratio [HR] for death 0.64; 95% CI 0.49-0.83) and CR with full or partial hematologic recovery (34.0 versus 15.3 percent; HR 18.6; 95% CI 9.8-27.4). Grade ≥3 AEs occurred less frequently with gilteritinib; the most common were grade ≥3 febrile neutropenia, anemia, and thrombocytopenia in 46, 41, and 23 percent, respectively.

Gilteritinib is approved by the FDA for AML with a FLT3 mutation as detected by an FDA-approved test, and the label carries a warning about differentiation syndrome; other significant toxicities include posterior reversible encephalopathy syndrome (PRES), prolonged QT interval, and pancreatitis [16]. Approval by EMA is pending. Differentiation syndrome is discussed separately. (See "Differentiation syndrome associated with treatment of acute leukemia".)

Antibody-directed therapy — Antibody-based therapy, including antibody-chemotherapy conjugates and immunotherapeutic agents, can induce a remission of relapsed or refractory AML, but their impact on survival is not well-defined.

●Gemtuzumab ozogamicin (GO) – Gemtuzumab ozogamicin (GO) is a humanized anti-CD33 antibody linked to the cytotoxic agent, calicheamicin, which can be administered as a single agent for CD33-positive AML.

•Administration – Induction therapy with single-agent GO is 3 mg/m² on days 1, 4, and 7; for patients without disease progression, this can be followed by GO 2 mg/m² on day 1 every 4 weeks for ≤8 cycles [17]. Patients should be premedicated with a glucocorticoid, acetaminophen, and diphenhydramine to mitigate infusion reactions.

The FDA label recommends capping the total dose at one vial (4.5 mg) when GO is given for newly diagnosed AML, but it does not specify that dose cap for treating relapsed or refractory AML.

•Toxicity – The package label includes a black-box warning about hepatotoxicity, including severe or potentially fatal hepatic sinusoidal obstruction syndrome (SOS; also called veno-occlusive disease [VOD]), which has been reported in association with GO treatment. Patients should be monitored frequently for signs and symptoms of VOD after treatment. (See "Hepatic sinusoidal obstruction syndrome (veno-occlusive disease) in adults".)

The most common AEs (>15 percent) include hemorrhage (potentially fatal), fever and/or infection, gastrointestinal toxicity (eg, nausea, vomiting, constipation, anorexia, increased AST/ALT), rash, mucositis, and headache. GO may be associated with embryo-fetal toxicity.

•Outcomes – Treatment of 15 patients with single agent GO for first relapse of AML was associated with 26 percent CR; for patients who reached CR, median relapse-free survival (RFS) was 12 months [18].

●Dual affinity retargeting proteins (DARTs) – DARTS are bispecific drugs that bring together effector cells (eg, T lymphocytes) with target cells (eg, AML myeloblasts).

•Flotetuzumab – Flotetuzumab is an investigational DART that engages CD3 epsilon on T cells and CD123 (low-affinity IL3 receptor alpha subunit), which is expressed by myeloblasts in 60 to 80 percent of patients with AML; expression of CD123 is not limited to AML blasts, as it is also expressed by some hematopoietic cell populations [19-21].

In a phase 2 study, treatment of 50 patients with flotetuzumab (500 ng/kg/day) was associated with hematologic response in 24 percent (18 percent CR); patients who achieved CR had a seven month median duration of response [22]. Rates of response were higher in patients with refractory AML, relapse <6 months from diagnosis, <5 lines of prior treatment, and/or evidence of immune infiltration. Infusion-related reactions and/or cytokine release syndrome (CRS) were nearly universal, but they were generally grade ≤2 and readily managed with dose interruptions and/or tocilizumab. Other CD123-based treatments are under development.

Intensive chemotherapy — There is no consensus regarding a preferred intensive chemotherapy regimen for relapsed/refractory AML, and since these regimens have not been directly compared, a choice is primarily based on clinical experience and institutional approach. An individual's chance of responding to a particular regimen is influenced not only by prior exposure to chemotherapy but also by other patient- and leukemia-associated factors. In theory, the preferred regimen would exclude agents at dose levels for salvage which the patient has been exposed to recently. (See 'Prognostic indices' below.)

No intensive regimen stands out as clearly superior. As an example, an international randomized phase III study of compared elacytarabine versus investigator choice in 381 patients with relapsed or refractory AML [23]. In the control arm, investigators were able to choose among seven regimens prior to randomization. When compared with other therapies, elacytarabine provided no additional benefit in terms of OS (3.5 months), response rate (23 percent), or relapse-free survival (5 months). Importantly, there was no evidence in favor of any of the variations of chemotherapy schedules in the control arm. The only long-term survivors were the few patients who achieved a response and proceeded with allogeneic HCT.

The most commonly used chemotherapy regimens are (table 5):

●High dose cytarabine (HiDAC; eg, 2 to 3 g/m² every 12 hours for 8 to 12 doses) may be effective in 35 to 40 percent of patients resistant to conventional dose cytarabine regimens [24]. If HiDAC was used in the initial induction therapy, an alternative regimen should be selected. An anthracycline (eg, daunorubicin) may be added if it was not used during initial induction, or if the cumulative anthracycline dose would not exceed that likely to be associated with cardiotoxicity. (See "Clinical manifestations, diagnosis, and treatment of anthracycline-induced cardiotoxicity" and "Risk and prevention of anthracycline cardiotoxicity".)

The combination of HiDAC plus an anthracycline may produce higher response rates than HiDAC alone [25]. Side effects appear to be similar with the two regimens. The most common nonhematologic toxicities are nausea/vomiting, abnormal liver chemistries, diarrhea, conjunctivitis, rash, and cerebellar dysfunction. Toxicity is prohibitively high in most patients over the age of 60 years.

●Re-induction with cytarabine plus an anthracycline (eg, daunorubicin) may be considered for patients who did not receive this as initial induction therapy, or for those who received it and achieved an initial CR that persisted for more than one year. For such patients with relapsed AML, this regimen will produce CR in approximately 50 percent [26,27]. (See "Acute myeloid leukemia: Induction therapy in medically fit adults".)

●Mitoxantrone (10 to 12 mg/m² per day) and etoposide (100 mg/m² per day) given together for five days is a commonly used regimen to treat refractory or relapsed AML and has demonstrated CR rates of approximately 40 percent [4,28-30]. Nonhematologic toxicities include stomatitis, nausea, infections, and neutropenic fever. Infrequent transient, mild cardiac failure and tachyarrhythmias have also been reported. The MEC regimen uses the former two drugs plus cytarabine.

●Mitoxantrone has also been used in various combinations with high and intermediate-dose cytarabine (the HAM regimen), both with and without etoposide [31-33]. In one series of 47 patients with recurrent or refractory AML, the combination of mitoxantrone (5 mg/m² per day for five days) and intermediate-dose cytarabine (0.5 g/m² IV every 12 hours for six days) resulted in a CR rate of 62 percent, but the median duration of remission was only 3.7 months [34]. Ninety-six percent of those achieving CR eventually relapsed. Three patients ≥60 developed acute cerebellar toxicity.

●Fludarabine, cytarabine, plus G-CSF (FLAG) has reported CR rates of 45 to 55 percent in patients with primary refractory or relapsing AML [35,36]. Studies including older adults have reported mild nonhematologic toxicity, most commonly mucositis. Median survival was >16 months for those with late relapse (ie, >6 months after stopping chemotherapy), but was only three months for those with early relapse or refractory disease [36]. The FLAG-Ida regimen also incorporates idarubicin.

●The combination of cladribine and cytarabine plus G-CSF has been administered with or without mitoxantrone or idarubicin [37,38]. CRs were seen in 40 to 50 percent of patients. Common nonhematopoietic toxicities were infection and bleeding.

Other regimens that have demonstrated efficacy include:

●Etoposide plus cyclophosphamide [39]

●Cytarabine, daunorubicin, and etoposide (ADE), given together or sequentially [40]

●Liposomal daunorubicin with or without cytarabine [41,42]

●Clofarabine plus cytarabine with or without G-CSF (CLAG) [43,44]

Chemotherapy for relapsed or refractory AML is highly toxic, primarily to the hematopoietic system, and most patients will have a prolonged hospital stay and will require blood product support. Supportive care of cytopenias, infections, and other complications in the setting of AML is presented separately. (See "Acute myeloid leukemia: Induction therapy in medically fit adults".)

HEMATOPOIETIC CELL TRANSPLANTATION — Allogeneic hematopoietic cell transplantation (HCT) is the therapy with the greatest chance of attaining a cure for a patient with relapsed or refractory AML [45]. However, autologous HCT may also yield a durable remission after relapse in approximately 15 to 20 percent of patients who attain a second complete remission (CR) with chemotherapy [46]. Most patients should be given induction chemotherapy prior to allogeneic HCT in an effort to achieve a CR. Others may proceed directly to HCT although the post-transplant relapse rate is higher if patients are not in a CR. A physician who specializes in HCT should be consulted at the time relapsed or refractory disease is identified to help guide this decision process. (See 'Selection of therapy' above.)

Determining transplant eligibility — Eligibility for allogeneic HCT varies across countries and institutions. The mortality associated with allogeneic HCT has declined likely related to reductions in organ damage, infection, and severe acute graft-versus-host disease (GVHD) [47,48]. An age cutoff of 65 years is commonly used for myeloablative HCT. However, age alone should not be a determining factor of transplant eligibility. Instead, decisions are made on a case-by-case basis (based on "physiologic age") and vary across institutions. (See "Determining eligibility for allogeneic hematopoietic cell transplantation".)

In most centers in the United States, Europe, and around the world, patients with one or more of the following factors are often not considered eligible for myeloablative HCT:

●Pulmonary function DLCO, FVC <60 percent

●Direct bilirubin >2.0 mg/dL

●Aspartate transaminase (AST; SGOT) or alanine transaminase (ALT; SGPT) >2x upper limit of normal

●Serum creatinine >1.5 mg/dL; creatinine clearance <60 mL/min

●Karnofsky performance status <70 percent (table 3)

●Ejection fraction <50 percent

In most transplant centers, patients are considered eligible for myeloablative allogeneic HCT if they are younger than 65 or 70 years of age, with normal cardiac, liver, and renal function, and in good performance status.

Conditioning regimen — The ideal conditioning (preparative) regimen prior to HCT is unknown and clinical practice differs by institution according to experience. The best outcomes appear to be with myeloablative preparative regimens. Nonmyeloablative preparative regimens may be considered for patients who are not candidates for myeloablative HCT, but have attained a CR. Possible options include:

●Myeloablative regimens – A myeloablative conditioning regimen consists of a single agent or combination of agents expected to destroy the hematopoietic cells in the bone marrow and produce profound pancytopenia within one to three weeks from the time of administration. The resulting pancytopenia is long-lasting, usually irreversible, and in most instances fatal, unless hematopoiesis is restored by infusion of hematopoietic stem cells. (See "Preparative regimens for hematopoietic cell transplantation", section on 'Myeloablative regimens'.)

Examples include:

•Busulfan and cyclophosphamide

•Melphalan, busulfan, and total body irradiation (TBI) [49]

•Cyclophosphamide plus TBI [50]

●Nonmyeloablative (reduced intensity) regimens – A nonmyeloablative regimen is one that combines the anti-leukemia effect of more modest doses of chemotherapy with the objective of achieving significant immunosuppression (to allow engraftment), and thereby causing minimal organ and mucosal damage. "Nonmyeloablative" is probably not the most appropriate term, since the doses of alkylating agents typically used will result in long period of severe cytopenias prior to marrow recovery. Examples of reduced intensity regimens include fludarabine plus melphalan or busulfan [51]. (See "Preparative regimens for hematopoietic cell transplantation", section on 'NMA and RIC regimens'.)

Myeloablative HCT — Survival following myeloablative HCT in patients with relapsed AML is generally ≤35 percent, significantly less than that seen when performed in first CR. Patients are usually sicker after relapse and their leukemia is often more resistant to treatment. As a result, transplantation-related mortality is higher and post-transplant relapses are more frequent. (See "Acute myeloid leukemia in younger adults: Post-remission therapy", section on 'Myeloablative allogeneic transplantation'.)

As with newly diagnosed AML, patients with relapsed or refractory AML represent a heterogeneous population, and outcomes of individual patients may vary based on age, comorbidities, and certain prognostic features. As an example, prognosis is better for patients with favorable molecular features (eg, biallelic CEBPA mutation or core binding factor rearrangements [52]) (table 2).

A study of 383 adult patients with acute leukemia from nine participating Australian transplant centers, in which progressive disease was responsible for 48 percent of the deaths, yielded the following five-year survival rates for patients with AML receiving myeloablative HCT between 1981 and 1997 [53]:

●Transplanted in untreated first relapse or second CR

•First CR duration longer than six months – 35 percent

•First CR duration less than six months – 20 percent

●Transplanted after primary induction failure – 15 percent

●Transplanted in chemotherapy-refractory first relapse or beyond second CR – 5 percent

The greatest experience with myeloablative HCT for relapsed or refractory disease is provided by a retrospective analysis of 2255 patients with relapsed or refractory leukemia reported to the Center for International Blood and Marrow Transplant Research between 1995 and 2004 [6]. This analysis included 1673 patients with primary refractory AML (38 percent), first untreated relapse (19 percent), first refractory relapse (26 percent), first relapse with unknown treatment status (1 percent), or second or additional relapse (17 percent). Overall, the survival rate at three years was 19 percent. Acute GVHD was seen in 48 percent of patients with 23 percent experiencing severe (grade 3/4) acute GVHD. Chronic GVHD was seen in approximately one-quarter of patients.

Survival rates may be higher with HCT performed using more modern supportive care. A single-center retrospective analysis of HCT for hematologic malignancy compared outcomes in 1418 patients undergoing first allogeneic HCT from 1993 to 1997 with those of 1148 patients undergoing first allogeneic HCT from 2003 to 2007 [48]. There was a substantial reduction in nonrelapse death (26 versus 41 percent) and increased long-term survival (53 versus 37 percent) which appeared to be related to reductions in organ damage, infection, and severe acute GVHD. While part of this benefit is likely due to a shift from high dose myeloablative conditioning regimens to standard myeloablative or reduced intensity conditioning regimens in the latter study period, part of this effect may be due to a change in stem cell source, GVHD prophylaxis, and other supportive measures.

Nonmyeloablative (reduced intensity) HCT — Nonmyeloablative allogeneic HCT is a potential treatment option for patients with relapsed AML who are not candidates for myeloablative HCT. Case series using matched related or matched unrelated donors have reported transplant-related mortality at one year of 17 to 55 percent, with three- to four-year progression-free survival rates of 19 to 30 percent, depending on the treatment program used [54-61]. (See "Preparative regimens for hematopoietic cell transplantation", section on 'NMA and RIC regimens'.)

Nonmyeloablative HCT has its main role as consolidation therapy in patients who can be successfully induced into a CR prior to HCT. This strategy relies in large part upon an immunological graft-versus-leukemia effect rather than a direct cytotoxic effect from the preparative regimen. Because it may take several months for the donor immune system to engraft, expand, and activate, nonmyeloablative HCT is not appropriate for patients with refractory or rapidly progressive AML. (See 'Remission re-induction' above.)

Choice of donor — The preferred donor for patients with relapsed or refractory AML undergoing allogeneic HCT is an HLA-matched sibling donor [6]. However, only 25 to 30 percent of potential recipients have HLA-matched siblings who can serve as donors. With the advent of molecular typing techniques, morbidity and mortality associated with the use of HLA-matched unrelated donors has declined. (See "Donor selection for hematopoietic cell transplantation", section on 'Unrelated donors'.)

For patients without an HLA-matched sibling, the following donors may be considered (listed in order of preference) [6,51,62-71]:

●Well matched unrelated donor

●Partially matched unrelated donor (ie, 9/10 or 8/10 matches)

●Single antigen mismatched related donor

●Umbilical cord blood transplantation

Allogeneic HCT using a matched unrelated donor (MUD) is an option for younger adults who lack a sibling donor. The likelihood of finding a MUD in the National Bone Marrow Donor Registry is related in part to the ethnic background of the patient compared with the volunteer donor pool. It also depends on the number of allelic mismatches between donor and recipient that one wishes to allow (eg, a full 10/10 match or 9/10 or 8/10 with mismatches). The overall match rate is approximately 50 percent for White Americans but only 10 percent for minorities who are both under-represented in the donor pool and more polymorphic with respect to HLA [72]. Outcomes after a well-matched unrelated donor (ie, 10/10 match) approach those seen with an HLA-matched sibling donor. (See "Acute myeloid leukemia in younger adults: Post-remission therapy", section on 'Choice of donor'.)

Due to the relative immaturity of the immune system in umbilical cord samples, stem cells from this source allow the crossing of immunologic barriers that would otherwise be prohibitive [73]. As a result, the degree of HLA-disparity that is tolerable is much greater, making this approach more feasible for members of minorities. In addition, cord blood has already been collected and cryopreserved before the time a match is identified, thereby greatly reducing the time required for accessing the donor and completing the transplant procedure; however, there is less clinical follow-up with cord blood transplantations and long-term follow-up will be necessary to determine whether they truly provide equivalent results [67,74]. (See "Hematopoietic cell transplantation (HCT): Sources of hematopoietic stem/progenitor cells".)

For adult patients with leukemia, there appear to be no significant differences in the rate of recurrence, treatment-related mortality, treatment failure, or overall mortality between those receiving HCT with mismatched marrow or mismatched umbilical cord blood [65,75]. As an example, one study examined unrelated cord blood HCT in 14 patients, most of whom had advanced de novo AML (one primary induction failure, seven in relapse, and six beyond first CR) [76]. Although all donors were mismatched at at least one locus, and acute and chronic GVHD were common complications, 17 of the 18 patients had myeloid reconstitution, with a two-year probability of disease-free survival of 77 percent.

As with nonmyeloablative HCT, cord blood HCT has its main role as consolidation therapy in patients who can be successfully induced into a CR prior to transplant. The reduced numbers of stem cells available from cord blood collections may limit the usefulness of this procedure in larger adults. (See "Selection of an umbilical cord blood graft for hematopoietic cell transplantation", section on 'UCB resources'.)

Haploidentical donors are related donors that are mismatched at three of the six loci (HLA-A, -B and -DR). Haploidentical donors are considered investigational in this setting as only limited data in selected patients are available. Protocols for this type of transplant were developed at a few lead centers with highly specialized expertise. A potential advantage of haploidentical HCT is that within a given family there are usually multiple individuals who could serve as potential donors including parents, siblings, and children. Ongoing studies are evaluating novel approaches to induce graft-versus-tumor activity and decrease GVHD in recipients of haploidentical grafts. As with nonmyeloablative HCT and cord blood HCT, successful outcomes are largely dependent upon the attainment of a CR prior to HCT [77,78]. Experimental approaches have also combined cord blood transplantation with infusion of CD34+ selected stem cells from a haploidentical family member [79-81]. (See "Donor selection for hematopoietic cell transplantation", section on 'Haploidentical donors'.)

POST-TRANSPLANT CARE — Following transplantation, patients are followed periodically to monitor for complications of the transplant procedure (eg, graft-versus-host disease [GVHD]) and for disease relapse.

During the initial several months, the patient is usually followed closely by the transplant team with visits at least once or twice a week. If there are no complications after three months, we typically schedule monthly visits for the remainder of the first year and then every two months. The likelihood of relapsing after 12 months is small, although complications of transplant (eg, GVHD) may still occur late.

Details regarding hematopoietic support after HCT, general symptom management after HCT, and an approach to HCT survivorship are presented separately. (See "Early complications of hematopoietic cell transplantation" and "Hematopoietic support after hematopoietic cell transplantation" and "Long-term care of the adult hematopoietic cell transplantation survivor".)

RELAPSE AFTER HCT — The optimal management of hematopoietic cell transplant (HCT) recipients experiencing AML relapse is a controversial issue. The median survival of patients relapsing after an allogeneic HCT is only three to four months if no further therapy were given [82]. The only therapies that have a potential for cure in these patients are a second transplant or donor lymphocyte infusion [83]. Patients who relapse more than one year after HCT may benefit from a second HCT.

Second HCT — A second allograft offers a chance of long-term leukemia free survival in selected patients. Only a limited proportion of patients are well enough for such intensive treatment, and the treatment related mortality is high (25 to 50 percent) [84-86]. The relapse rate is also high. Leukemia-free survival rates at three to five years range from 11 to 42 percent [84-88]. Prognostic variables associated with a poor outcome include a short interval between initial transplant and relapse (especially if the interval is less than one year), occurrence of GVHD after the first transplant, older age, poor performance status and the lack of remission at the time of second allograft [85-88]. In one study, for example, leukemia-free survival at three years was 52 percent for a subset of patients who relapsed >292 days after the first transplant and who were in remission before receiving the second transplant [86].

A French cooperative study investigated the efficacy of a second HCT following graft failure or relapse in patients with AML, ALL, or CML [89]. Transplant-related mortality and five-year overall survival were 45 and 32 percent, respectively. Factors associated with a better outcome after the second HCT were age <16 at second HCT, relapse >12 months after first transplant, female donor, absence of acute GVHD, and occurrence of chronic GVHD.

A German retrospective registry study identified 179 patients undergoing second HCT following relapse after matched related (75 patients) or unrelated (104 patients) HCT [88]. A second CR was achieved in 74 percent, and half of these patients experienced a second relapse. The estimated overall survival rate at two years was 25 percent. Survival was higher for those whose second HCT was from a matched related donor rather than an unrelated donor (39 versus 19 percent). Changing donors for the second HCT did not have a positive or negative impact on survival. As with other studies, shorter remission duration after first HCT and refractory disease at the time of second HCT were associated with worse outcomes.

A number of suggestions have been made to improve outcome in this setting:

●All patients who might be candidates for a second HCT should receive reinduction chemotherapy first. This establishes that they still have chemotherapy-sensitive disease. Only those responding to treatment (usually no more than 50 percent) should proceed to a second transplant [83]. (See 'Remission re-induction' above.)

●GVHD prophylaxis should be reduced after the second HCT to maximize the graft-versus-leukemia (GVL) effect [83]. (See "Biology of the graft-versus-tumor effect following hematopoietic cell transplantation".)

●Relapsing patients may receive a second transplant following reduced intensity conditioning [90,91]. In one series, seven patients with AML relapsing after an allograft were treated with the FLAG regimen (fludarabine, high dose Ara-C, G-CSF) with or without an anthracycline, followed by a second allogeneic HCT from the original donor [90]. CR was attained in six with no treatment-related mortality. Median duration of the second CR was 14 months, which was slightly longer than the first CR of 11 months. Chronic GVHD was extensive in four of the seven.

Donor lymphocyte infusion — Discontinuation of immunosuppression followed by donor lymphocyte infusion (DLI) induces remission in 20 to 35 percent of patients who relapse after allogeneic HCT [83,92-94]. This can only be done if there is no ongoing active GVHD. It has been suggested that DLI is most effective in patients with measurable residual disease (MRD; also referred to as minimal residual disease) in whom the GVL effect would likely be most effective [83]. DLI may be also considered as a means of T cell repletion for the prevention of leukemia recurrence in patients given a T cell depleted HCT from a compatible sibling [83]. (See "Immunotherapy for the prevention and treatment of relapse following allogeneic hematopoietic cell transplantation", section on 'Donor lymphocyte infusion (DLI)'.)

All data on DLI are based on retrospective studies, the largest of which was a cohort of 399 patients with AML in first hematological relapse after HCT who did or did not receive a DLI [95]. At a median follow-up of over 27 months, the following results were reported:

●Two-year overall survival rates were significantly higher in those patients who received a DLI (21 versus 9 percent, respectively).

●Patients more likely to have a good response to DLI included those with favorable cytogenetics, those with a hematological remission prior to DLI, and those with a lower tumor burden (less than 35 percent blasts in bone marrow) at relapse.

Randomized studies are needed to demonstrate a true advantage of DLI. The use of DLI remains an individualized decision based upon patient age, cytogenetics, and tumor burden. DLI is unlikely to be effective in patients with relapsed leukemia that no longer expresses the mismatched HLA antigen thought to be responsible for the graft-versus-leukemia effect [96].

PROGNOSTIC INDICES — As described above, patients with relapsed or refractory AML are a heterogeneous population with long-term survival rates after allogeneic hematopoietic cell transplantation (HCT) ranging from 40 percent to less than 5 percent [6,26,27]. (See 'Selection of therapy' above.)

Several prognostic scoring systems have been developed to predict outcomes in this population.

●First relapse – A European Prognostic Index (EPI) was developed based on results in 667 patients with AML in first relapse among 1540 newly diagnosed patients with AML (age 15 to 60 years) entered onto Dutch-, Belgian-, and Swiss-collaborative group trials [26]. The following four risk factors (with accompanying point scores) were found on multivariate analysis:

•Relapse-free interval from first complete remission (CR; 0, 3, and 5 points for >18, 7 to 18, and ≤6 months, respectively)

•Cytogenetics at the time of diagnosis (0, 3, and 5 points for t(16;16) or inv(16); t(8;21) with or without other abnormalities, and normal, intermediate, unfavorable, or unknown cytogenetics, respectively)

•Age at first relapse (0, 1, and 2 points for ≤35, 36 to 45, and >45 years, respectively)

•Autologous or allogeneic HCT before first relapse (0 and 2 points for no prior HCT or a prior HCT, respectively)

Three risk groups were identified: favorable (1 to 6 points), intermediate (7 to 9 points), and high risk (10 to 14 points), with five-year overall survival rates of 22 to 46, 12 to 18, and 4 to 6 percent, respectively, when the EPI was employed by two different groups [26,27].

●Second salvage therapy – In a retrospective single center study of 594 patients with AML undergoing a second salvage treatment that included conventional or experimental therapies or HCT, the median survival was only 1.5 months and the one-year survival 8 percent [97]. Thirteen percent of patients achieved a CR and their median CR duration was seven months. On multivariate analysis, the following factors adversely affected overall survival:

•First CR duration <12 months

•Second CR duration <6 months

•Serum bilirubin level ≥1 mg/dL

•Serum albumin level <3 g/dL

•Age >60 years

•Bone marrow blasts ≥50 percent

•Year of therapy before 1991

Patients at low-risk (zero to two adverse factors), intermediate risk (three adverse factors), or high-risk (≥4 adverse factors) had estimated one-year survival rates of 22, 6, and 0 percent, respectively. The rates of CR within the same risk groups were 26, 8, and 2 percent, respectively.

●Patients without CR before HCT – A retrospective analysis from the Center for International Blood and Marrow Transplant Research included 1673 patients with relapsed or refractory AML between 1995 and 2004 [6]. This analysis included 1673 patients with primary refractory AML (38 percent), first untreated relapse (19 percent), first refractory relapse (26 percent), first relapse with unknown treatment status (1 percent), or second or additional relapse (17 percent). Overall, the survival rate at three years was 19 percent.

CLINICAL TRIALS — Often there is no better therapy to offer a patient than enrollment onto a well-designed, scientifically valid, peer-reviewed clinical trial. Additional information and instructions for referring a patient to an appropriate research center can be obtained from the United States National Institutes of Health (www.clinicaltrials.gov).

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Acute myeloid leukemia".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient education" and the keyword(s) of interest.)

●Beyond the Basics topics (see "Patient education: Acute myeloid leukemia (AML) treatment in adults (Beyond the Basics)")

SUMMARY AND RECOMMENDATIONS

●Approximately 20 to 50 percent of adult patients with newly diagnosed acute myeloid leukemia (AML) will fail to attain a complete remission (CR) with intensive induction chemotherapy (primary refractory AML). In addition, a large percentage of patients who initially attain a CR will later have relapsed disease. (See 'Definitions of refractory disease and relapse' above.)

●Patients suspected of having relapsed AML should be evaluated with a unilateral bone marrow aspiration: a trephine biopsy provides additional information and is required if the aspiration sample is poor. Once relapsed or refractory AML is identified, further evaluation should include assessments of performance status and organ function. HLA typing should be done to initiate a donor search for hematopoietic cell transplantation (HCT) and consultation with a physician who specializes in HCT should take place at the time refractory or relapsed disease is identified to help with the decision process and timing. (See 'Evaluation of suspected relapse' above.)

●Allogeneic HCT provides the greatest chance of cure for a patient with relapsed or refractory AML, but there have also been some long-term survivors after autologous HCT.

Eligibility for allogeneic HCT varies by institution, but poor overall performance status or moderate to severe impairment of pulmonary, liver, kidney, or cardiac function may make a patient ineligible. (See 'Determining transplant eligibility' above.)

Patients who are not candidates for transplantation should be considered for clinical trials with investigational agents. (See 'Clinical trials' above.)

●Re-induction therapy is not required for all patients prior to myeloablative allogeneic HCT. However, all patients undergoing a nonmyeloablative HCT should undergo re-induction therapy in order to attain a CR prior to HCT, if possible. (See 'Remission re-induction' above.)

●There is no consensus regarding the optimal regimen for achieving disease control prior to allogeneic HCT. Some patients have AML with mutations (eg, IDH1, IDH2, FLT3) that are amenable to treatment with targeted agents. Selection of targeted therapy versus intensive chemotherapy to control disease must be individualized and will vary based on the presence of a targetable mutation, medical fitness, prior therapy, patient preference, and the urgency of achieving disease control. (See 'Selection of therapy' above.)

●For most patients with relapsed or refractory AML undergoing HCT, we recommend the use of a myeloablative conditioning regimen rather than nonmyeloablative/reduced intensity allogeneic HCT, autologous HCT, or chemotherapy alone (Grade 1B). The ideal conditioning (preparative) regimen prior to HCT is not defined and clinical practice differs by institution.

Nonmyeloablative or reduced intensity conditioning regimens should be considered for patients who are not candidates for myeloablative HCT but have attained a marrow remission.

●The preferred donor for patients with relapsed or refractory AML undergoing allogeneic HCT is an HLA-matched sibling donor. If no sibling donor can be identified, an alternative donor search (including cord blood) should be initiated. If no donor is available, an autologous stem cell transplant might be a possibility in highly selected circumstances. (See 'Choice of donor' above.)

●The optimal management of HCT recipients experiencing AML relapse is a controversial issue. The only therapies that have a potential for cure in these patients are a second transplant or donor lymphocyte infusion. (See 'Relapse after HCT' above.)

●Patients with relapsed or refractory AML are a heterogeneous population with long-term survival rates after allogeneic HCT ranging from 40 percent to less than 5 percent. Several prognostic scoring systems have been developed to predict outcomes in this population. (See 'Prognostic indices' above.)