Differentiation syndrome associated with treatment of acute leukemia

INTRODUCTION — Differentiation syndrome (DS; originally called "retinoic acid syndrome") is a potentially fatal complication of treatment of acute promyelocytic leukemia (APL) with all-trans retinoic acid and/or arsenic trioxide, treatment of acute myeloid leukemia (AML) with inhibitors of isocitrate dehydrogenase (IDH; ie, enasidenib, ivosidenib, olutasidenib), and treatment of AML with mutant FLT3 (Fms-related tyrosine kinase 3) with gilteritinib.

The clinical presentation, diagnosis, and treatment of differentiation syndrome will be discussed here.

Clinical manifestations, diagnosis, and treatment of APL and an overview of AML are presented separately. (See "Clinical manifestations, pathologic features, and diagnosis of acute promyelocytic leukemia in adults" and "Initial treatment of acute promyelocytic leukemia in adults" and "Acute myeloid leukemia in adults: Overview".)

EPIDEMIOLOGY AND RISK FACTORS — DS occurs in roughly one-quarter of patients with acute promyelocytic leukemia (APL) treated with all-trans retinoic acid (ATRA) and/or arsenic trioxide (ATO), up to one-fifth of patients treated with IDH inhibitors (ie, ivosidenib, olutasidenib, enasidenib) for acute myeloid leukemia (AML), and a small fraction of those treated with gilteritinib for AML. Factors that predict development of DS are not well defined.

●Acute promyelocytic leukemia – DS is reported in approximately one-quarter of patients who are treated with ATRA for APL, but the incidence has ranged from 2 to 48 percent; the wide range in incidence reflects differences in induction regimens, preventive measures, and different diagnostic criteria [1-10]. For patients treated with ATO for APL, the incidence of DS ranged from 7 to 31 percent [5,11-14]. DS can also occur in patients with relapsed APL who were treated with ATRA and/or ATO [15]. The incidence of DS was similar in both arms of a trial in which patients were randomly assigned to ATRA plus ATO versus ATRA plus chemotherapy [16]. DS has not been reported with ATO treatment of non-APL malignancies or during consolidation or maintenance phase of APL, because the syndrome depends on the presence of leukemic blasts and/or promyelocytes [17].

●Acute myeloid leukemia – DS has been reported in 10 to 19 percent of adults with IDH-mutated relapsed or refractory AML treated with the IDH inhibitors, enasidenib, ivosidenib, or olutasidenib [18-23], and in 17 percent of patients with newly diagnosed IDH-mutated AML treated with ivosidenib and azacitidine [24]. DS was reported in 3 percent of patients treated with gilteritinib, an inhibitor of mutant FLT3 [25].

●Risk factors – Elevated white blood cell (WBC) count is frequently present at the onset of DS, but DS can also occur without leukocytosis. In one study of APL, development of DS was associated with elevated levels of WBC, creatinine, lactate dehydrogenase, and peripheral blood blast count during treatment, according to multivariate analysis [17]. Other reported associations include elevated WBC count at presentation, rapidly rising WBC count, expression of CD13 or CD11b on APL blasts, and elevated body mass index (BMI) [2,6,7,26-33]. The incidence of DS does not appear to be related to the dose of ATRA or ATO, concurrent use of chemotherapy, or drugs that affect cytochrome P450 activity. In one study of APL, multivariate analysis indicated that fatal outcomes with DS were associated with poor performance status and low serum albumin [34]. No factors have been shown to predict development of DS in patients treated with IDH inhibitors.

PATHOGENESIS — The etiology of DS is incompletely understood, but DS has been linked to a large burden of maturing myeloid cells and their production of inflammatory cytokines.

In acute promyelocytic leukemia (APL), all-trans retinoic acid (ATRA) and arsenic trioxide (ATO) can induce maturation of promyelocytes and promote tissue infiltration [35]. Other factors that may contribute to DS include a systemic inflammatory response associated with increased cytokine expression, endothelial damage with capillary leak syndrome, and occlusion of the microcirculation. Inflammatory vasoactive cytokines that have been implicated in DS include interleukin (IL)-1, IL-6, IL-8, tumor necrosis factor alpha, and CCL2 [36,37]. Release of cathepsin G, a serine protease that enhances capillary permeability, may contribute to capillary damage [38]. ATRA and ATO act through the aberrant retinoid receptor PML-RARA to induce differentiation of APL promyelocytes, increase chemokine production, and upregulate leukocyte integrins, which increase adherence to capillary endothelium and enhance organ infiltration [39-44]. The molecular pathology of APL is discussed separately. (See "Molecular biology of acute promyelocytic leukemia", section on 'Retinoic acid and the retinoic acid receptor'.)

Less is understood about the cause of DS associated with inhibitors of mutant isocitrate dehydrogenase (IDH; ie, ivosidenib, enasidenib, olutasidenib) and the FLT3 inhibitor, gilteritinib. However, these agents can promote differentiation of myeloid blasts in patients with acute myeloid leukemia (AML) [45-48].

PATHOLOGY — Lung histopathology in patients with DS reveals interstitial infiltration with maturing myeloid cells [3,26]. Patients with fatal DS may also have septal edema, diffuse alveolar damage, intra-alveolar hemorrhage, and capillaritis (small vessel inflammation) [4,39]. Post-mortem studies have also revealed extensive infiltrates of myeloid cells in other organ systems, including lymph nodes, spleen, liver, and pericardium [3].

CLINICAL PRESENTATION

Onset of DS — The onset of clinical manifestations of DS can range from days to weeks after beginning treatment, but findings generally arise earlier in patients with severe disease. The type of leukemia and induction therapy influence the timing of DS onset.

●DS associated with acute promyelocytic leukemia (APL) generally occurs 7 to 12 days after starting induction therapy, but can appear as early as the first day of treatment [8,10,49]. In one study, DS occurred a median of 12 days after starting all-trans retinoic acid (ATRA; range 0 to 46 days) but followed a bimodal pattern; DS began in 46 percent of patients within the first week, while symptoms developed during the third week or beyond for 38 percent [50]. Severe DS typically begins earlier (median six days) than moderate DS (median 15 days).

●Reported median onsets of DS complicating treatment of acute myeloid leukemia (AML) with enasidenib are 19 to 30 days from initiation (range 1 to 129 days) [19,21]. Median times of onset of DS reported with ivosidenib treatment are 20 to 29 days (range 1 to 78 days) [21,51,52]. DS following treatment with gilteritinib for AML with mutant FLT3 occurred 2 to 75 days after beginning treatment [25]. The following studies illustrate typical DS rates and onset patterns in AML:

•Among 153 IDH1 inhibitor-naïve patients who received monotherapy with olutasidenib for relapsed or refractory IDH1-mutant AML, DS adverse events occurred in 22 (14 percent) patients, with 14 (9 percent) grade ≥3 (table 1) and 1 fatal case reported [22]. For most affected patients, DS occurred within the first two cycles of treatment, with a median time to first event of 18 days (range, 1 to 561).

•In a separate trial, 146 newly diagnosed IDH1-mutant AML patients were randomized 1:1 to receive azacitidine and either ivosidenib or placebo [52]. The percentage of patients with DS of any grade (table 1) was 14 percent with ivosidenib and azacitidine (with no grade ≥4 events) and 8 percent with placebo and azacitidine (including one grade 4 event). The median time to onset of DS of any grade in the ivosidenib-and-azacitidine group was 20 days (range, 3.0 to 33).

Clinical manifestations — Typical clinical findings of DS include dyspnea, fever, peripheral edema, hypotension, weight gain, pleuro-pericardial effusion, acute renal failure, musculoskeletal pain, and hyperbilirubinemia [3,5,26,50,53,54]. Less commonly, DS may be associated with pulmonary hemorrhage or Sweet syndrome (acute febrile neutrophilic dermatosis) [4,39,55-57]. No symptoms or signs distinguish DS caused by treatment of APL from that caused by treatment of AML. (See "Sweet syndrome (acute febrile neutrophilic dermatosis): Pathogenesis, clinical manifestations, and diagnosis".)

The frequency of clinical findings varies with the severity of the syndrome. Among 739 patients with APL, the following findings were noted in moderate and severe DS: dyspnea (59 and 95 percent, respectively), edema (53, 81 percent), unexplained fever (53, 74 percent), weight gain >5 kg (38, 68 percent), and hypotension (12, 39 percent) [50,58]. An erythematous rash may be present [18,41].

Headache can be a manifestation of pseudotumor cerebri, which is a potential complication of treatment with ATRA and/or arsenic trioxide (ATO). Patients with headache and/or visual disturbance should have a funduscopic examination promptly to identify papilledema or other manifestations of pseudotumor cerebri, which is a medical emergency. (See "Initial treatment of acute promyelocytic leukemia in adults", section on 'High intracranial pressure'.)

Laboratory — Abnormal hematologic parameters and chemistries are common in DS but differ slightly between APL and AML.

Elevated white blood cell (WBC) count, anemia, thrombocytopenia, and coagulopathy are common in DS associated with APL, but some of these abnormalities may be due to the underlying leukemia and/or concomitant chemotherapy [5,6,26,50]. In one series of DS in APL, the median WBC count was 31,000/microL (range 6800 to 72,000/microL) at the onset of the syndrome [26].

Among patients with AML treated with enasidenib, approximately 40 percent had concomitant leukocytosis [19], while only 10 percent of those treated with ivosidenib developed leukocytosis [18].

Renal failure was reported in 9 and 46 percent of patients with moderate and severe DS, respectively, in a large series of patients with APL [50,58].

Imaging — Chest radiograph abnormalities are common in DS, but findings may differ based on the severity of the syndrome. Pleural effusions appear to be more common in association with ATRA/ATO than with isocitrate dehydrogenase (IDH) inhibitors.

Chest radiograph findings attributable to DS include increased cardiothoracic ratio (87 percent), increased vascular pedicle width (87 percent), septal lines and peribronchial cuffing (60 percent), diffuse ground glass opacity (60 percent), parenchymal consolidation (47 percent), nodular opacities (47 percent), air bronchogram (33 percent), and pleural effusion (73 percent) [4,50]. Approximately 40 percent of patients with DS have a clear chest radiograph at presentation, although only 11 percent of patients with severe disease have a clear chest radiograph [50]. In a large series of patients with APL, patients with moderate and severe DS had pulmonary opacities on chest radiograph in 38 and 81 percent, respectively, pleural effusion in 27 and 58 percent, and pericardial effusion in 11 and 23 percent [50,58].

Pleural effusions were reported in 12 percent of patients treated with ivosidenib and in 42 percent treated with enasidenib [18,19].

EVALUATION — DS should be suspected in patients with any of the characteristic features (eg, unexplained fever, hypotension, weight gain, dyspnea, radiographic pulmonary opacities, pleural/pericardial effusion, and/or renal insufficiency) in the setting of acute promyelocytic leukemia (APL) or acute myeloid leukemia (AML) treated with differentiation agents. (See 'Prompt initiation of glucocorticoids' below and 'Additional testing for more severe or progressive disease' below.)

Initial steps in all patients — Importantly, treatment of DS should begin at the time that DS is first suspected, even as diagnostic testing is being performed. (See 'Prompt initiation of glucocorticoids' below.)

Daily weights and fluid balance should be monitored. In addition, the following studies should be obtained in all patients suspected of DS:

●Hematology – Complete blood count (CBC) with differential count, prothrombin time (PT), partial thromboplastin time (PTT)

●Chemistry – Serum electrolytes, blood urea nitrogen, creatinine

●Oxygenation – Pulse oximetry or arterial blood gas

●Microbiology – Blood cultures

●Imaging – Chest radiograph

Additional testing for more severe or progressive disease — Additional testing is appropriate for patients who present with more severe respiratory insufficiency, fever, or hypotension, and those who do not respond to initial glucocorticoid treatment within 24 hours. (See 'Prompt initiation of glucocorticoids' below.)

As the diagnosis of DS is based on a constellation of nonspecific clinical features, the main goals of additional testing are to identify processes in the differential diagnosis of DS, such as infection, heart failure, pulmonary hemorrhage, or thromboembolism, realizing that patients may have more than one process contributing to symptoms. (See 'Differential diagnosis' below.)

●Microbiologic testing – Specific laboratory studies to evaluate infection in leukemic patients with fever are informed by the clinical findings and are described separately. (See "Diagnostic approach to the adult cancer patient with neutropenic fever", section on 'Patient evaluation'.)

●Computed tomography (CT) can clarify the pattern, extent, and location of radiographic abnormalities, if needed to guide bronchoscopic evaluation and can help identify pleural and pericardial effusions. The use of CT to evaluate respiratory insufficiency is described separately. (See "Approach to the patient with dyspnea", section on 'Chest computed tomography'.)

●Bronchoscopy/bronchoalveolar lavage (BAL) can be helpful for evaluation of pulmonary infection or pulmonary hemorrhage but is generally reserved for patients who require intubation and mechanical ventilation. BAL can usually be performed safely, even with thrombocytopenia and/or coagulopathy [50]. When possible, the area of lung with the greatest degree of radiographic abnormality should be lavaged. The specimens should be processed for special stains and culture for bacterial and fungal infection, and also PCR and culture for viral infections (table 2). BAL may also detect alveolar hemorrhage as a complication of severe differentiation syndrome (picture 1). (See "Approach to the immunocompromised patient with fever and pulmonary infiltrates", section on 'Lung sampling'.)

Endobronchial or transbronchial biopsy should be avoided because of the increased risk of hemorrhage in the setting of thrombocytopenia and/or coagulopathy. (See "Flexible bronchoscopy in adults: Indications and contraindications", section on 'Contraindications'.)

●Cardiac evaluation – Patients with dyspnea, edema, and/or hypotension that may be caused by reduced cardiac output or excess fluid accumulation should be evaluated with brain natriuretic peptide (BNP), electrocardiogram (EKG), and/or echocardiogram to detect heart failure, pericardial effusion, or dysrhythmias, as described separately. (See "Heart failure: Clinical manifestations and diagnosis in adults".)

Patients with DS are at risk for acute kidney injury due to capillary leakage and hypotension from intravascular depletion. Close monitoring of renal function and intravascular volume is prudent. (See "Evaluation of acute kidney injury among hospitalized adult patients", section on 'Evaluation'.)

Avoidance of invasive procedures — It is important to avoid nonessential invasive procedures (eg, thoracentesis, lung biopsy, kidney biopsy) because patients with APL or AML have a high risk of bleeding due to associated coagulation defects and bleeding disorders. (See "Clinical manifestations, pathologic features, and diagnosis of acute promyelocytic leukemia in adults", section on 'Coagulopathy and APL' and "Acute myeloid leukemia: Overview of complications", section on 'Bleeding'.)

DIAGNOSIS — A prompt clinical response to dexamethasone is strong evidence for uncomplicated DS and may obviate the need for more invasive or expensive investigations. The diagnosis of DS is based on the clinical presentation and setting (differentiation therapy for leukemia) and is supported by a response to treatment and exclusion of alternate diagnoses.

In the setting of treatment of acute promyelocytic leukemia (APL) with all-trans retinoic acid (ATRA) and/or arsenic trioxide (ATO) or treatment of acute myeloid leukemia (AML) with enasidenib or ivosidenib, a presumptive diagnosis (based on ≥1 clinical features) enables initiation of glucocorticoid treatment while pursuing evaluation of other potential causes. The presence of three or more features is sufficient for a confident clinical diagnosis of DS in the absence of another cause [1,3,18,19,59]:

●Fever ≥38° C

●Weight gain >5 kg

●Hypotension

●Dyspnea

●Radiographic opacities

●Pleural or pericardial effusion

●Acute renal failure

The exact criteria for the diagnosis of DS have not been formally established. Other sets of diagnostic criteria include combinations of these features, but vary in the number required to establish the diagnosis [26,50,60].

Of note, features of DS overlap with other processes that are common in this setting, and more than one disease process (eg, DS and infection) can contribute to symptoms and signs. (See 'Differential diagnosis' below.)

In general, disease severity correlates with the number of clinical features that are present [50]:

●Severe: ≥4 clinical features

●Moderate: 3 clinical features

●Indeterminate: 1 to 2 clinical features

DIFFERENTIAL DIAGNOSIS — Clinical features of DS overlap substantially with other processes, such as infection/sepsis, heart failure, and pulmonary thromboembolism that occur in this setting. Treatment of DS should begin at the time that DS is first suspected, even as other causes are being evaluated. (See 'Prompt initiation of glucocorticoids' below.)

●Infection/sepsis − Both infection/sepsis and DS can present with fever, hypotension, respiratory distress, and/or radiographic pulmonary opacities. Neither condition can be conclusively distinguished on the basis of routine initial laboratory testing and chest radiograph. The evaluation for infection should include peripheral blood cultures, sputum stains and cultures, nasal swab for viral PCR, and serum and urine tests for specific bacterial, fungal, and viral infections. Empiric antibiotic therapy is often initiated pending study results.

If the patient has respiratory failure or does not respond to empiric antibiotics, flexible bronchoscopy with bronchoalveolar lavage may help to identify an organism. Evaluation of fever and empiric antibiotic therapy are described separately. (See "Diagnostic approach to the adult cancer patient with neutropenic fever" and "Treatment of neutropenic fever syndromes in adults with hematologic malignancies and hematopoietic cell transplant recipients (high-risk patients)", section on 'Empiric therapy' and 'Treatment' below.)

●Thromboembolism − Clinical features of pulmonary embolism (PE)/deep vein thrombosis (DVT; eg, dyspnea/hypoxia, hypotension, peripheral edema, leg swelling, hemoptysis) may be indistinguishable from those of DS. Furthermore, there is an increased risk of thrombosis in DS [17]. Pulmonary embolism should be suspected in patients with evidence of deep vein thrombosis (lower extremity or related to central venous catheter) and in those with hypoxemia that is out of proportion to the extent of the radiographic opacities. Evaluation of DVT/PE is described separately. (See "Clinical presentation, evaluation, and diagnosis of the nonpregnant adult with suspected acute pulmonary embolism".)

●Heart failure − Heart failure can be associated with dyspnea, hypotension, edema, and abnormalities on chest radiograph that may be indistinguishable from findings of DS. Furthermore, patients with acute leukemia may have ventricular dysfunction from prior chemotherapy. Clinical evaluation, measurement of brain natriuretic peptide (BNP), echocardiography, and/or computed tomography (CT) should distinguish heart failure due to left ventricular function, pericardial effusion, or dysrhythmia from DS. (See "Natriuretic peptide measurement in heart failure" and "Heart failure: Clinical manifestations and diagnosis in adults", section on 'Echocardiography'.)

●Alveolar hemorrhage − Alveolar hemorrhage is a well-described complication of acute leukemia in patients with diffuse intravascular coagulation, although the frequency has markedly decreased since introduction of differentiation therapy [55]. Alveolar hemorrhage is also described in patients with DS and normal clotting studies and those with lung infection [61]. In a patient with DS who develops diffuse pulmonary opacities, it can be difficult to determine whether the findings are caused by DS, alveolar hemorrhage, pneumonia, or a combination. However, a specific diagnosis of alveolar hemorrhage (eg, by bronchoalveolar lavage [BAL]) is generally not needed. (See "The diffuse alveolar hemorrhage syndromes".)

●Drug allergy − Allergy to antibiotics or other medications given prior to the onset of DS can cause fever, rash, capillary leak, and dyspnea. (See "Drug reaction with eosinophilia and systemic symptoms (DRESS)".)

●Acute kidney injury − Acute kidney injury is one of the criteria of DS, but can also be a consequence of hypotension, decreased cardiac output, infection, or drug toxicity. It is important to avoid nonessential kidney biopsy to reduce the risk of hemorrhage in this setting. The evaluation of acute kidney injury is discussed separately. (See "Evaluation of acute kidney injury among hospitalized adult patients", section on 'Evaluation'.)

TREATMENT — DS is a potentially fatal complication of acute promyelocytic leukemia (APL) treated with all-trans retinoic acid (ATRA) and/or arsenic trioxide (ATO), and acute myeloid leukemia (AML) treated with inhibitors of mutant isocitrate dehydrogenase (IDH; ie, enasidenib, ivosidenib, olutasidenib) or mutant FLT3 (gilteritinib). It is important to have a high degree of suspicion for the diagnosis, and to initiate treatment of DS concurrently with the diagnostic evaluation (algorithm 1).

Prompt initiation of glucocorticoids — For all patients suspected of having DS (one or more clinical features), we recommend prompt treatment with systemic glucocorticoids concurrently with the diagnostic evaluation, rather than waiting to treat until completion of the diagnostic evaluation [59,62]. This recommendation is based on the favorable balance of modest toxicity of therapy weighed against a generally prompt and robust response; in contrast, DS can progress rapidly if treatment is delayed. (See 'Diagnosis' above.)

●Initial dosing – We typically initiate treatment with dexamethasone 10 mg, orally or intravenously, every 12 hours [59]. Most patients improve substantially within 12 to 24 hours.

●Inadequate response to initial treatment – If the patient has not improved substantially within 24 hours, the dosing frequency can be increased to every six hours, while reassessing the possibility of an alternate diagnosis. (See 'Differential diagnosis' above.)

●Duration and tapering – Dexamethasone (or the equivalent) is continued at full dose for at least three days or until fever, dyspnea, and hypoxemia have resolved, followed by a tapering course (eg, reducing by 50 percent every two to three days). Occasionally, DS may recur as the glucocorticoids are tapered or discontinued. (See 'Recurrent DS' below.)

Potential toxicities of glucocorticoid treatment include hyperglycemia, insomnia, depression, emotional instability, and immunosuppression, but severe toxicity is rare. (See "Major adverse effects of systemic glucocorticoids".)

●Efficacy – There are no randomized trials of early treatment with systemic glucocorticoids versus delayed therapy for DS in APL. However, retrospective and prospective studies of ATRA and/or ATO report that prompt treatment with steroids has reduced DS-associated mortality to approximately 1 percent [1,9,26,50,59,60,63,64]. As an example, one study reported 0.2 percent DS-associated mortality among 240 patients with APL who received ATRA plus idarubicin induction therapy; DS was diagnosed in 15 percent of patients in this study based on the presence of five clinical criteria [9]. In contrast, two studies of APL that were performed prior to the routine early administration of steroids were each associated with 9 percent DS-associated mortality [3,6].

In a series of 281 patients treated with enasidenib for relapsed or refractory AML, there were no DS-associated deaths among the 12 percent of patients who developed DS and were promptly treated with steroids [19]. There were no DS-associated deaths among the 4 percent of 179 patients who developed DS during treatment with ivosidenib for relapsed or refractory AML; all patients received prompt initiation of glucocorticoids [51].

Our approach is consistent with that proposed for DS by the National Comprehensive Cancer Network (NCCN) and the European Leukemia Network [62,65,66].

Management of differentiation agents — We suggest continuing differentiation agents in the majority of patients who develop DS. As an exception, differentiation agents may need to be held temporarily in patients with severe manifestations of DS, such as life-threatening organ dysfunction (eg, hypoxemia requiring ventilatory support, hypotension requiring vasopressor therapy, acute renal failure). The suggestion to continue differentiation therapy in most patients is based on favorable outcomes with prompt initiation of glucocorticoids and the desire to achieve optimal outcomes for the leukemia by continuing therapy.

The decision to interrupt the differentiation therapy is also informed by the specific agent(s) used:

●ATRA and/or ATO – For patients with severe DS receiving ATRA and/or ATO, we suggest withholding these agents while continuing glucocorticoid therapy. ATRA/ATO are withheld until findings of severe DS have resolved (ie, ≤3 diagnostic criteria or resolution of organ failure), at which point they can be resumed and then continued until achievement of complete remission (CR) or completion of the minimum duration of remission induction therapy.

●Ivosidenib, olutasidenib, enasidenib, and gilteritinib – For patients receiving ivosidenib, enasidenib, or gilteritinib for AML, withdrawal of these agents is unlikely to be effective in the immediate term because of the prolonged half-lives of ivosidenib (estimated 97 hours), olutasidenib (estimated 67 hours), enasidenib (estimated 137 hours), and gilteritinib (113 hours) [67,68]. If needed, hydroxyurea can be added to control leukocytosis, as discussed separately.

Empiric antibiotics — Prompt administration of broad-spectrum, empiric antibiotics is prudent in many patients with DS, especially those with neutropenia, fever, or progressive radiographic opacities. Empiric antibiotics are accompanied by rigorous evaluation for superimposed infection. Antibiotic therapy can be adjusted or discontinued, depending on microbiologic test results and response to therapy. (See "Treatment of neutropenic fever syndromes in adults with hematologic malignancies and hematopoietic cell transplant recipients (high-risk patients)", section on 'Empiric therapy'.)

Supportive care — Other supportive measures are important in optimal management of DS, including treatment for fluid overload, hypotension, respiratory or renal insufficiency, heart failure, bleeding, and other complications. Because the clinical findings of DS may be indistinguishable from those of other disorders that can complicate treatment of acute leukemia, patients should receive treatment for other conditions in the differential diagnosis, as described above. (See 'Differential diagnosis' above.)

●Fluid management and renal replacement therapy – For patients with weight gain or dyspnea due to DS, cautious diuresis is given as tolerated by the patient’s hemodynamic and renal status. Approximately 87 percent of patients with differentiation syndrome require diuretic therapy; in some cases this has led to rapid improvement [50,69]. For patients with kidney dysfunction related to hypotension, renal perfusion should be supported with intravenous fluids and/or vasopressors. Approximately 12 percent of patients need at least short-term renal replacement therapy [50]. Patients with hypotension that is not immediately responsive to fluid resuscitation should be evaluated for other causes of hypotension, as described separately. (See "Evaluation of and initial approach to the adult patient with undifferentiated hypotension and shock", section on 'Initial diagnostic evaluation'.)  

●Treatment of coagulopathy – For patients with fluid overload and high requirements for blood products to control coagulopathy, cryoprecipitate, fibrinogen, and other coagulation factor concentrates are preferred over fresh-frozen plasma, to limit the fluid burden [59]. The management of coagulopathy in APL is described separately. (See "Initial treatment of acute promyelocytic leukemia in adults", section on 'Control of coagulopathy'.)

●Oxygenation and mechanical ventilation – Many patients will require supplemental oxygen, and some will require noninvasive or invasive mechanical ventilation to achieve adequate oxygenation [17,26]. In the PETHEMA (Programa Español de Tratamientos en Hematología) studies, mechanical ventilation was needed in 26 percent of patients [17]. (See "Noninvasive ventilation in adults with acute respiratory failure: Benefits and contraindications" and "Acute respiratory distress syndrome: Fluid management, pharmacotherapy, and supportive care in adults" and "Acute respiratory distress syndrome: Ventilator management strategies for adults".)

Recurrent DS — A small percentage of patients who respond to the initial course of dexamethasone will develop recurrent manifestations of DS upon tapering or discontinuation of dexamethasone [10,26]. For these patients, dexamethasone (10 mg intravenously, every 6 to 12 hours) is resumed until complete resolution of these manifestations [59]. As with initial therapy, other possible explanations, such as infection, heart failure, renal insufficiency, must be considered and/or treated. (See 'Treatment' above.)

PREVENTION OF DS — We suggest not routinely treating with preventive steroids or other prophylactic therapy during induction therapy with all-trans retinoic acid (ATRA), arsenic trioxide (ATO), ivosidenib, olutasidenib, enasidenib, or gilteritinib. The suggestion against routine DS prophylaxis reflects our preference to avoid unnecessary toxicity and the lack of evidence that prophylaxis can improve outcomes. Rather, we suggest prompt administration of steroid therapy upon development of clinical findings consistent with DS, as described above. (See 'Prompt initiation of glucocorticoids' above.)

A potential role for prophylactic therapy for some or all of the duration of induction therapy, or for particular subsets of patients (eg, based on white blood cell [WBC] count, serum creatinine) has been examined in case series, but no randomized trials have been performed [59]. It remains unclear whether DS-preventive approaches result in meaningfully improved outcomes [59]. As an example, a review of three studies from Programa Español de Tratamientos en Hematología (PETHEMA) reported that steroid treatment was associated with a modest reduction of DS incidence, but no differences in DS-related mortality for acute promyelocytic leukemia (APL) treated with ATRA [34]. However, it is difficult to draw conclusions from this retrospective review that compared three different nonrandomized studies that administered steroids by various doses, routes, and schedules only for patients with WBC ≥5,000/microL.

The role of DS prophylaxis during treatment with isocitrate dehydrogenase (IDH) inhibitors has not been specifically examined and preventive therapy for DS was not included in studies of enasidenib or ivosidenib for acute myeloid leukemia (AML) [19,51]. Given the lower frequency of DS and relatively late onset, we do not prescribe preventive therapy in these patients.

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 promyelocytic leukemia" and "Society guideline links: Acute myeloid leukemia".)

SUMMARY AND RECOMMENDATIONS

●Epidemiology and risk factors – Differentiation syndrome (DS) is a potentially fatal complication of treatment with differentiation agents in patients with acute promyelocytic leukemia (APL) or acute myeloid leukemia (AML). DS occurs in approximately one-quarter of patients with APL during induction therapy with all-trans retinoic acid (ATRA) and/or arsenic trioxide (ATO), up to one-fifth of patients with AML treated with isocitrate dehydrogenase (IDH) inhibitors (eg, enasidenib, ivosidenib, olutasidenib), and in smaller numbers (estimated 3 percent) of patients with AML treated with gilteritinib. (See 'Epidemiology and risk factors' above.)

●Pathogenesis – The cause of DS is incompletely understood, but contributing factors include a systemic inflammatory response associated with increased cytokine expression, endothelial damage with capillary leak syndrome, occlusion of microcirculation, tissue infiltration, and other effects of maturation of large numbers of myeloblasts/promyelocytes. (See 'Pathogenesis' above.)

●Initial evaluation – Routine initial studies include a complete blood count with differential count, prothrombin time, partial thromboplastin time, serum electrolytes, blood urea nitrogen, creatinine, pulse oximetry and/or arterial blood gas, blood culture, and a chest radiograph. Further evaluation depends on the response to initial glucocorticoid therapy. (See 'Evaluation' above.)

●Diagnosis – In the setting of treatment of APL or AML with a differentiation agent, a clinical diagnosis of DS can be made based on ≥3 of the following features, although one or more features not attributable to another cause is sufficient for a presumptive diagnosis (see 'Diagnosis' above):

•Fever ≥38° C

•Weight gain >5 kg

•Hypotension

•Dyspnea

•Radiographic opacities

•Pleural or pericardial effusion

•Renal failure

●Differential diagnosis – Clinical features of DS overlap substantially with other processes (eg, infection/sepsis, heart failure, pulmonary thromboembolism) that occur in this setting. Evaluation for other causes should proceed during initial treatment for presumptive DS. (See 'Differential diagnosis' above.)

●Treatment

•Prompt initiation of glucocorticoids – For patients clinically suspected of having DS, we recommend prompt treatment with systemic glucocorticoids rather than delaying treatment until completion of the diagnostic evaluation (Grade 1B). The typical initial dose of dexamethasone is 10 mg every 12 hours. (See 'Prompt initiation of glucocorticoids' above.)

•Ongoing management of differentiation therapy – For most patients with DS, we suggest continuing differentiation agents while monitoring the response to glucocorticoid therapy (Grade 2C). Most patients will recover with prompt initiation of glucocorticoids, and continuation of differentiation agents enables optimal treatment of the leukemia. (See 'Management of differentiation agents' above.)

For patients with severe DS (eg, respiratory or renal insufficiency requiring intensive care) or an inadequate response after 24 to 48 hours of glucocorticoid therapy, temporary interruption of ATRA and/or ATO is appropriate. Interruption of IDH or FLT3 inhibitors is less likely to be of benefit given their long half-life. (See 'Management of differentiation agents' above.)

●Supportive measures, including empiric antibiotics – Patients with fever in the setting of immunosuppression require prompt administration of broad-spectrum antibiotics and rigorous evaluation for superimposed infection. Supportive care for fluid overload, coagulopathy, bleeding, and organ failure is appropriate. (See 'Empiric antibiotics' above and 'Supportive care' above.)

●Prevention – We suggest not routinely administering systemic glucocorticoids as attempted preventive or prophylactic treatment of DS during induction therapy with ATRA, ATO, ivosidenib, enasidenib, or gilteritinib (Grade 2C).(See 'Prevention of DS' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Steven Weinberger, MD, who contributed to earlier versions of this topic review.