Differentiation Syndrome Associated With Enasidenib, a Selective Inhibitor of Mutant Isocitrate Dehydrogenase 2: Analysis of a Phase 1/2 Study

Key Points

Question  What are the signs, symptoms, and appropriate approaches to management of isocitrate dehydrogenase differentiation syndrome, an unexpected event observed in the first clinical study of enasidenib mesylate, recently approved for relapsed/refractory mutant IDH2 acute myeloid leukemia?

Findings  In a phase 1/2 clinical trial of 281 patients with relapsed/refractory acute myeloid leukemia treated with enasidenib, possible or probable isocitrate dehydrogenase differentiation syndrome was identified in 33 patients and had recognizable signs and symptoms, including dyspnea, unexplained fever, pulmonary infiltrates, and hypoxia. These signs and symptoms could occur months after initiation of enasidenib treatment and could mimic symptoms seen during treatment and progression of myeloid neoplasms; prompt intervention with systemic corticosteroid therapy was an effective management approach.

Meaning  As use of mutant isocitrate dehydrogenase inhibitors increases, awareness of the potential for isocitrate dehydrogenase differentiation syndrome is imperative so that patients can be promptly and effectively treated.

Importance  Enasidenib mesylate, a mutant isocitrate dehydrogenase 2 (IDH2) protein inhibitor that promotes differentiation of leukemic myeloblasts, was recently approved by the US Food and Drug Administration for use in relapsed/refractory (R/R) mutant IDH2 acute myeloid leukemia (AML). During the first study of enasidenib in humans, a minority of patients with advanced myeloid neoplasms experienced unexpected signs/symptoms of a differentiation syndrome (DS), a potentially lethal entity.

Objective  To characterize IDH-inhibitor–associated DS (IDH-DS) and its effective management.

Design, Setting, and Participants  Using data obtained from a multicenter, open-label, pivotal phase 1/2 study of enasidenib, a differentiation syndrome review committee retrospectively evaluated potential cases of IDH-DS in enasidenib-treated patients with R/R AML. Data were collected between August 27, 2013, and October 14, 2016. The committee identified and agreed on signs and symptoms characteristic of IDH-DS and developed an algorithm for identification and treatment. Among 281 patients with R/R AML enrolled in the trial, the committee identified 72 patients for review based on investigator-reported cases of DS (n = 33) or reported adverse events or signs and symptoms characteristic of IDH-DS.

Interventions  Treatment with enasidenib at a dosage of 50 to 650 mg/d was evaluated during the dose-escalation phase, and a dosage of 100 mg/d was used in the phase 1 expansion and phase 2, all in continual 28-day cycles.

Main Outcomes and Measures  Unexpected adverse events of IDH-DS during the phase 1/2 study.

Results  Thirty-three of the 281 patients (11.7%) were identified as having possible or probable IDH-DS. Median age of those 33 patients was 70 years (range, 38-80 years); 20 (60.6%) were male. The most frequent manifestations were dyspnea, fever, pulmonary infiltrates, and hypoxia. Median time to onset was 30 days (range, 7-129 days). Patients who experienced IDH-DS were less likely to have less than 20% bone marrow blasts (6% vs 22%, P = .04) and more likely to have undergone fewer previous anticancer regimens (median, 1.0 [range, 1-4] vs 2.0 [range, 1-14], P = .05) at study entry than those who did not. Thirteen patients (39.4%) had concomitant leukocytosis. Isocitrate dehydrogenase differentiation syndrome was effectively managed with systemic corticosteroids. The enasidenib regimen was interrupted for 15 patients (45.5%), but permanent discontinuation of treatment was not required.

Conclusions and Relevance  Isocitrate dehydrogenase differentiation syndrome is a recognizable and potentially lethal clinical entity, occurring in approximately 12% of enasidenib-treated patients with mutant-IDH2 R/R AML. It requires prompt recognition and management. As use of mutant IDH inhibitors increases, these findings and recommendations are increasingly germane to care of patients with mutant-IDH neoplasms.

Trial Registration  clinicaltrials.gov Identifier: NCT01915498

Genetic and epigenetic alterations that block cellular differentiation underlie the development of myeloid neoplasms.¹ Agents that induce resumption of differentiation may be clinically beneficial; however, drug-induced differentiation of leukemic cells can be accompanied by a potentially serious condition, differentiation syndrome (DS), as seen during treatment of acute promyelocytic leukemia.^2-5 Proliferation of differentiated leukemic cells can alter cytokine balance, leading to tissue damage and inflammation.⁴ Signs and symptoms of DS include unexplained fever, weight gain, respiratory symptoms, and pleural or pericardial effusions.⁴

Enasidenib mesylate (AG-221; IDHIFA; Celgene Corporation) is a first-in-class, oral inhibitor of mutant isocitrate dehydrogenase 2 (mIDH2) proteins. Enasidenib was approved by the US Food and Drug Administration (FDA) in August 2017 for treatment of adult patients with mutant IDH2 (OMIM 147650) (mIDH2) relapsed/refractory acute myeloid leukemia (R/R AML). Mutant IDH2 proteins neomorphically catalyze the oncometabolite, (R)-2-hydroxyglutarate ([R]-2HG).⁶ High (R)-2HG concentrations promote hypermethylation, altered gene expression, and blocked differentiation of hematopoietic cells.⁷ Enasidenib suppresses 2HG and induces differentiation of mIDH2 leukemic cells, producing fully functioning neutrophils that retain the IDH2 mutation.^8,9 Patients who respond to treatment with enasidenib exhibit evidence of hematopoietic recovery, typically without intervening bone marrow aplasia.^3,8

In the first human, phase 1/2 study of enasidenib in patients with mIDH2 malignant hematologic neoplasms, investigators reported adverse events reminiscent of those seen in DS. Similar events have been reported for ivosidenib (AG-120), a mutant IDH1 protein inhibitor.¹⁰ These signs and symptoms of mutant-IDH inhibitor–associated DS (termed IDH differentiation syndrome [IDH-DS])¹¹ are nonspecific and can resemble other clinical conditions common in AML. To characterize IDH-DS and establish practicable diagnostic and therapeutic recommendations, an independent differentiation syndrome review committee (DSRC) retrospectively evaluated potential cases of IDH-DS among patients with R/R AML enrolled in the aforementioned study.

This was a multicenter, open-label, pivotal phase 1/2 study. Study design, conduct, eligibility criteria, and clinical response and safety assessments in the phase 1 dose-escalation and expansion study portions are reported elsewhere.³ Phase 2 enrollment was limited to adult patients with R/R AML, all of whom received 100 mg of enasidenib once daily. This study was conducted in accordance with the Declaration of Helsinki.¹² All patients provided written informed consent. The institutional review boards that approved the study can be found in the eAppendix in the Supplement.

The DSRC independently reviewed cases of investigator-reported DS and those with adverse events suggestive of IDH-DS (eTable 1 in the Supplement). Isocitrate dehydrogenase differentiation syndrome was considered when characteristic signs and symptoms occurred in the absence of significant secondary causes, with concurrent evidence of cellular differentiation and proliferation in the blood or bone marrow, and with subsequent clinical response to corticosteroid therapy. Hematologic response was assessed per International Working Group 2003 AML criteria.¹³ Baseline characteristics of patients who did or did not experience IDH-DS were summarized and compared using the Wilcoxon Mann-Whitney test (continuous variables) or Fisher exact test (categorical variables). P values are 2-sided, with a significance level of .05. Statistics were performed using SAS 9.2 statistical software (SAS Institute Inc).

Two hundred eighty-one patients with R/R AML participated between August 27, 2013, and data cutoff (October 14, 2016). The DSRC reviewed 72 suspected cases of IDH-DS and, of them, identified 33 patients (11.7% of all patients with R/R AML) as having experienced possible or probable IDH-DS. The median age of patients with possible or probable IDH-DS was 70 years (range, 38-80 years). Twenty of the 33 (60.6%) were male. Twenty-five patients had 1 IDH-DS episode, 6 had 2 episodes, and 2 experienced 3 episodes. Reasons for rejecting an IDH-DS diagnosis included alternative causes for symptoms, lack of clinical evidence of cellular differentiation, lack of response to corticosteroid therapy, and clear evidence of disease progression as a more likely explanation for the constellation of clinical signs and/or symptoms.

Baseline characteristics were generally comparable between patients who experienced IDH-DS and those who did not (eTable 2 in the Supplement). Patients with IDH-DS were significantly less likely to have had less than 20% bone marrow blasts (6% vs 22%, P = .04) and more likely to have undergone a lower median number of previous antileukemic regimens (median, 1.0 [range, 1-4] vs 2.0 [range, 1-14], P = .05). Higher peripheral blast counts (median 31% [range, 0%-88%] vs 13% [range, 0%-98%], P = .05) and greater proportion of those with lactate dehydrogenase levels above 1 × upper limit of normal (51% vs 36%, P = .09) were noted for patients with IDH-DS, but these variables did not reach statistical significance. In multivariate analysis, only a lesser median number of previous antileukemic therapies was significantly associated with IDH-DS (odds ratio, 0.66; 95% CI, 0.44-1.00; P = .05).

Median time from beginning of enasidenib treatment to IDH-DS onset was 30 days (range, 7-129 days); onset at day 129 occurred for 1 patient after a treatment interruption of approximately 1 month. The most frequent clinical manifestations of IDH-DS were dyspnea, culture-negative fever, pulmonary infiltrates, and hypoxia (Table 1). Ten patients required intensive care unit admission for related symptoms. Leukocytosis occurred concurrently in 13 patients. No death was attributed to IDH-DS. Clinical and laboratory profiles for 2 illustrative cases in patients who experienced IDH-DS are described in eFigures 1 and 2 in the Supplement.

Dosages greater and less than 100 mg/d were evaluated in small patient cohorts during dose escalation. Events related to IDH-DS occurred in none of 19 patients with R/R AML receiving enasidenib dosages less than 100 mg/d, 26 of 214 patients (12.1%) receiving 100 mg/d, and 7 of 48 patients (14.6%) receiving greater than 100 mg/d. Frequency of IDH-DS was not statistically different among these groups.

Twenty-eight of the 33 patients with IDH-DS (84.8%) received symptomatic treatment with corticosteroids. For 20 patients with available data, median duration of corticosteroid therapy for IDH-DS was 12 days (range, 4-43 days). Eleven of 13 patients with concomitant leukocytosis received hydroxyurea (median, 15 days [range, 4-71 days]). The enasidenib regimen was interrupted for 15 patients (45.5%) until IDH-DS symptoms improved. Two patients (6.1%) had dosage reduced, but none discontinued enasidenib therapy permanently because of IDH-DS.

Isocitrate dehydrogenase differentiation syndrome management guidelines have been incorporated into an amended protocol. At first suspicion of IDH-DS, dexamethasone, 10 mg twice daily, is recommended until signs and symptoms are significantly improved (Figure).

Overall response rate for patients with IDH-DS was 45.5% (15 of 33 patients) and for patients without IDH-DS, 37.5% (93 of 248) (Table 2). Complete remission or complete remission with incomplete platelet or neutrophil count recovery was experienced by 36.4% of patients (12 of 33) in the IDH-DS group and 26.2% (65 of 248) in the non–IDH-DS group (P = .22). Disease remained stable in approximately half of the patients in each group.

Isocitrate dehydrogenase differentiation syndrome is a newly defined clinical entity in patients with mIDH2 AML receiving enasidenib, likely owing to the drug’s mechanism of action, promoting differentiation of myeloblasts into mature cells.⁸ No single pathognomonic clinical sign or laboratory result is diagnostic of IDH-DS. A constellation of findings supports the diagnosis, which can be challenging because signs and symptoms can mimic those of leukemic progression or other acute comorbidities. Diagnosis should be suspected on exclusion of other potential causes of characteristic symptoms.

Our results showed that IDH-DS was significantly associated with fewer previous anticancer therapies and nonsignificantly associated with higher baseline peripheral blast counts (P = .05) and lactate dehydrogenase levels (P = .09), both laboratory markers suggestive of cellular turnover. Why fewer preceding therapies might increase risk is unclear. This IDH-DS study population is small; further study is needed to determine whether reliable biomarkers of the potential to develop IDH-DS exist.

The rate of IDH-DS with enasidenib (11.7%) was lower than DS rates reported with all-trans retinoic acid or arsenic trioxide in acute promyelocytic leukemia .^5,15 Unlike all-trans retinoic acid– or arsenic trioxide–induced DS, which have a typical onset of approximately 7 to 12 days from treatment initiation),⁴ IDH-DS onset can occur from 7 days to 5 months after initiation of enasidenib treatment. In addition, IDH-DS can occur after enasidenib therapy interruption or dose escalation. Onset appears to correspond to enasidenib-induced myeloid cell differentiation and maturation, similar to the “neutrophil surge” associated with DS in patients treated with FLT3 (FMS-like tyrosine kinase 3) inhibitors, such as quizartinib.²

Although potentially life-threatening, IDH-DS can be effectively managed. When alternative secondary causes for characteristic symptoms are excluded or do not respond to treatment, or if symptoms worsen within 48 hours of initiating treatment for a different suspected cause, the event should be managed as potential IDH-DS (Figure). For concomitant leukocytosis, cytoreduction per local standard practice is recommended.

Approximately half of patients with IDH-DS received enasidenib therapy without interruption. If associated with severe pulmonary symptoms and/or renal dysfunction lasting more than 48 hours after corticosteroid treatment initiation, enasidenib therapy should be withheld until symptoms improve. Symptoms may not resolve immediately. Given the prolonged half-life of enasidenib (approximately 137 hours),¹⁶ drug treatment interruption is not an alternative to corticosteroid use.

The studied population was relatively small, and larger studies are needed. In addition, because IDH-DS was not initially anticipated, the protocol did not define proactive monitoring for it; however, IDH-DS monitoring is now part of clinical protocols for enasidenib studies.

As use of mutant-IDH inhibitors increases, these recommendations are increasingly germane to care of patients with mutant-IDH hematologic neoplasms.

Corresponding Author: Amir T. Fathi, MD, Massachusetts General Hospital, Assistant Professor, Harvard Medical School, 55 Fruit St, Boston, MA 02114 (afathi@mgh.harvard.edu).

Accepted for Publication: October 17, 2017.

Published Online: January 18, 2018. doi:10.1001/jamaoncol.2017.4695

Open Access: This article is published under the JN-OA license and is free to read on the day of publication.

Author Contributions: Drs Fathi and DiNardo contributed equally to this work and are co–lead authors. Drs Stein and de Botton contributed equally to this work and are co–senior authors. Drs Fathi and DiNardo had full access to all the data in the manuscript and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Fathi, DiNardo, Kline, Gupta, Attar, Stein, de Botton.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Fathi, DiNardo, Kline, Gupta, de Botton.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Kline, Kenvin.

Obtained funding: Kline.

Administrative, technical, or material support: DiNardo, Gupta, Stein.

Study supervision: Fathi, Kline, Kenvin, Stein, de Botton.

Conflict of Interest Disclosures: Dr Fathi reported serving as a consultant for and receiving clinical trial support from Celgene Corporation and Seattle Genetics; serving on advisory boards for Agios Pharmaceuticals, Inc, Merck, Juno, Torero, and Bexalata; and receiving clinical trial support from Takeda and Exelixis. Dr DiNardo reported receiving compensation for serving on advisory boards for Agios, Daiichi Sankyo, Celgene, and Novartis, speakers fees from AbbVie, and receiving clinical research support from Agios Pharmaceuticals, Inc, AbbVie, Daiichi Sankyo, Novartis, and Celgene Corporation. Drs Kline, Kenvin, and Gupta are employees and hold stock in Celgene Corporation. Dr Attar is an employee of Agios Pharmaceuticals, Inc. Dr Stein reported receiving grants and personal fees from Celgene Corporation and Agios Pharmaceuticals. Dr de Botton reported receiving personal fees from Agios Pharmaceuticals, Inc, Celgene Corporation, Novartis, Pfizer, and Servier. No other disclosures were reported.

Funding/Support: This study was funded by Celgene Corporation and Agios Pharmaceuticals, Inc.

Role of the Funder/Sponsor: The funding organizations participated in study design and conduct, managed the data, collected and analyzed data in conjunction with the authors, reviewed the manuscript for accuracy, and agreed to the decision to submit the manuscript for publication.

Group Information: The AG221-C-001 Study Investigators are Florida Cancer Specialists, Sarasota: Manish Patel (principal investigator [PI]); Sarah Cannon Research Institute, Nashville, Tennessee: Ian Flinn (PI); Massachusetts General Hospital Cancer Center, Boston: Amir Fathi (PI); Memorial Sloan-Kettering Cancer Center, New York, New York: Eytan Stein (PI), Martin Tallman (sub-PI); University of Colorado, Denver: Daniel Pollyea (PI); University of Texas Southwestern Medical Center, Dallas: Robert Collins (PI); Stanford Comprehensive Cancer Center, Stanford University, Palo Alto, California: Bruno Medeiros (PI); Sylvester Comprehensive Cancer Center, University of Miami, Miami, Florida: Ronan Swords (PI); Weill Cornell Medical College, Cornell University; New York Presbyterian Hospital, New York: Gail Roboz (PI); Robert H. Lurie Comprehensive Cancer Center, Northwestern University Feinberg School of Medicine, Chicago, Illinois: Jessica Altman (PI); The University of Texas MD Anderson Cancer Center, Houston: Hagop Kantarjian (PI), Courtney DiNardo (sub-PI); Cleveland Clinic Taussig Cancer Institute, Cleveland, Ohio: Mikkael Sekeres (PI); The Ohio State University Comprehensive Cancer Center, Columbus: Alice Mims (PI); Washington University School of Medicine in St Louis, St Louis, Missouri: Geoffrey Uy (PI); Oregon Health & Science University, Portland: Elie Traer (PI); Moffitt, Tampa, Florida: Jeffrey Lancet (PI); City of Hope, Duarte, California: Anthony Stein (PI); Gustave-Roussy Cancer Campus Grand Paris, Villejuif, France: Stephane de Botton (PI); Hospital Haut-Leveque, Pessac, France: Arnaud Pigneux (PI); Institut Universitaire du Cancer, Oncopole, Toulouse, France: Christian Recher (PI); and Dana-Farber Cancer Institute, Boston: Richard Stone (PI).

Additional Contributions: The authors thank Barbara Anderson, RN, of Celgene Corporation for thoroughly collecting, reviewing, and organizing clinical data and collaborating with investigators to develop procedures to diagnose and manage IDH-DS in current clinical trials. Sheila Truten, BS, and Kelly Dittmore, MS (Medical Communication Company, Inc), provided editorial assistance that was funded by Celgene Corporation.