Progressive multifocal leukoencephalopathy associated with brentuximab vedotin therapy: A report of 5 cases from the Southern Network on Adverse Reactions (SONAR) project
Critical review was provided by Eugene O. Major, PhD, chief of Laboratory of Molecular Medicine and Neuroscience at the National Institute of Neurologic Diseases of the National Institutes of Health in Bethesda, Maryland. Zaina Qureshi, PhD, MS; Brian Chen, PhD, JD; and Richard M. Schulz, PhD provided helpful critical review of the policy implications of the study.
Brentuximab vedotin (BV) is an anti-CD30 monoclonal antibody-drug conjugate that was approved in 2011 for the treatment of patients with anaplastic large cell and Hodgkin lymphomas. The product label indicates that 3 patients who were treated with BV developed progressive multifocal leukoencephalopathy (PML), a frequently fatal JC virus-induced central nervous system infection. Prior immunosuppressive therapy and compromised immune systems were postulated risk factors. In the current study, the authors reported 5 patients who developed BV-associated PML, including 2 immunocompetent patients.
Case information was obtained from clinicians (4 patients) or a US Food and Drug Administration database (1 patient).
All 5 patients had lymphoid malignancies. Two patients with cutaneous T-cell lymphomas had not previously received chemotherapy. PML developed after a median of 3 BV doses (range, 2 doses-6 doses) and within a median of 7 weeks after BV initiation (range, 3 weeks-34 weeks). Presenting findings included aphasia, dysarthria, confusion, hemiparesis, and gait dysfunction; JC virus in the cerebrospinal fluid (2 patients) or central nervous system biopsy (3 patients); and brain magnetic resonance imaging scans with white matter abnormalities (5 patients). Four patients died at a median of 8 weeks (range, 6 weeks-16 weeks) after PML diagnosis. The sole survivor developed immune reconstitution inflammatory syndrome.
PML can develop after a few BV doses and within weeks of BV initiation. Clinicians should be aware of this syndrome, particularly when neurologic changes develop after the initiation of BV treatment. The decision to administer BV to patients with indolent cutaneous lymphomas should be based on consideration of risk-benefit profiles and of alternative options. Cancer 2014;120:2464–2471. © 2014 American Cancer Society.
Progressive multifocal leukoencephalopathy (PML) is a rare and often fatal infection of the central nervous system (CNS) that results from reactivation of latent JC polyoma virus (JCV), which typically occurs in immunocompromised patients. In the era of highly active antiretroviral therapy, approximately 80% of new PML diagnoses develop among individuals infected with the human immunodeficiency virus (HIV). However, before the HIV epidemic, the majority of PML cases occurred in patients with lymphoproliferative disorders. Among HIV-negative individuals, PML has been most commonly associated with cytotoxic chemotherapy, immune-suppressing medications, and solid organ and hematopoietic stem cell transplantation.[4-6] PML incidence estimates range from 0.07% in a population-based study of individuals with hematologic malignancies to 0.5% among patients with chronic lymphocytic leukemia treated with fludarabine to 3% to 5% among persons diagnosed with the acquired immunodeficiency syndrome in the post-highly active antiretroviral therapy era. Currently, there is a lack of universally accepted treatments for PML outside of reconstitution of the immune system. Immune reconstitution is often achieved through removal of the offending agent associated with PML, by stimulating an altered immune system, or by initiating antiretroviral therapy in patients with HIV/acquired immunodeficiency syndrome. Some patients with PML survive after the development of an immune reconstitution inflammatory syndrome (IRIS), which is characterized by rapid infiltration of cytotoxic T cells. Although the development of IRIS among patients with PML can be life-saving, CNS inflammation due to IRIS can result in death or permanent neurologic disability.
Most recently, PML has been reported in patients receiving several different immune-modulating monoclonal antibodies. These monoclonal antibodies appear to alter normal immune function and/or immune surveillance. Natalizumab, a monoclonal antibody targeted against the integrin alpha-4, was initially approved for the treatment of multiple sclerosis and Crohn disease, but was withdrawn from the market in 2005 after 3 patients developed PML. After implementation of a global risk management program, natalizumab was reintroduced in 2006. In certain high-risk individuals, natalizumab has been associated with a PML incidence of up to 1 in 85 exposures. Efalizumab, a monoclonal antibody targeting CD11a, was approved in 2003 for the treatment of moderate to severe plaque psoriasis. Marketing of the drug was voluntarily discontinued by the manufacturer in 2009 after 3 confirmed cases of PML developed in patients who had received several years of treatment with efalizumab. Rituximab, an anti-CD20 monoclonal antibody, was approved for the treatment of indolent B-cell non-Hodgkin lymphomas in 1997. Over the next 9 years, the US Food and Drug Administration (FDA) received reports of 10 patients who developed PML after treatment with rituximab. In 2009, we reported 57 rituximab-associated PML cases occurring among patients with autoimmune diseases (5 patients) and lymphoproliferative disorders (52 patients). Subsequently, by mid-2012, the FDA had received reports of 511 patients with rituximab-associated PML.
In 2011, the FDA granted accelerated approval for brentuximab vedotin (BV), a monoclonal antibody targeting CD30-positive cells, which is conjugated to the cytotoxic agent monomethyl auristatin E. The approval was for the treatment of recurrent or refractory systemic anaplastic large cell lymphoma and classic Hodgkin lymphoma. The initial product label included a warning that 1 patient treated with BV had developed and subsequently died of PML. In 2012, the manufacturer added a “Black Box” warning to the label indicating that 2 subsequently reported patients had also developed PML after BV treatment. Long-term clinical follow-up on the 3 patients was not reported. The warning identified potential contributory risk factors as underlying lymphoproliferative disorders, compromised immune systems, and multiagent chemotherapy. Clinicians were advised that among patients treated with BV, signs and symptoms of PML, primarily neurologic changes, may develop over the course of several weeks or months.
For all monoclonal antibodies targeting the immune system, vigilant postmarketing surveillance is essential to detect rare disorders such as PML. When cases of PML arise, assessment of the risks and benefits of the associated agent for the approved indications is warranted. The FDA noted in January 2012 that 3 cases of PML were diagnosed among the first 2000 patients who had received BV, although comprehensive information on presenting findings, prior treatments, PML treatments, and outcome was not provided. In the current study, we summarized the clinical characteristics, interval between BV therapy and onset of disease, response to treatment, pathology/laboratory findings, and long-term outcomes in 5 patients in whom PML developed after the initiation of BV treatment. The identification of these 5 BV-associated PML cases occurred within the first year of marketing of BV.
MATERIALS AND METHODS
The Southern Network on Adverse Reactions (SONAR) project conducts safety initiatives focusing on monoclonal antibody-associated PML. Case information was obtained directly from treating clinicians for 4 patients. We also obtained 6 reports from the FDA's Adverse Event Reporting System (FAERS) of all BV-associated PML cases (search period between August 1, 2011 and December 31, 2012); 1 of these cases was excluded because information confirming the diagnosis of PML was not included in the report and 4 cases were excluded because they represented duplicate reports of cases identified directly by SONAR. Duplicate reports in the FAERS and SONAR databases were confirmed based on cross-referencing patient age, sex, and dates of BV treatment. Information obtained from clinicians was considered preferentially in instances in which duplicate clinical information was available in FAERS and from clinicians, based on previous experience. Inclusion criteria were receipt of BV therapy before PML diagnosis or symptom onset, PML confirmation based on histologic examination of tissue from the CNS or magnetic resonance imaging (MRI) demonstrating lesions consistent with PML, and documentation of JCV DNA in the cerebrospinal fluid (CSF) by polymerase chain reaction. Abstracted data included sociodemographic information, neurological signs and symptoms at the time of presentation, prior therapies, MRI findings, CSF findings, biopsy findings, and clinical outcomes. Two cases have been previously described as case reports (in 3 publications).[18-20]
The 5 patients with PML associated with BV treatment ranged in age from 38 years to 72 years (median, 50 years). (Table 1) Three patients were men. Indications for BV treatment included classic Hodgkin lymphoma (3 patients) that was initially diagnosed 7, 8, and 12 years previously, respectively; primary cutaneous anaplastic large cell lymphoma diagnosed 4 years previously; and transformed mycosis fungoides.
Table 1. Clinical Findings For 5 Patients With Brentuximab Vedotin-Associated PML
|Clinical trial participant||Yes (5th PML case submitted to the FDA and the manufacturer [11/2012])||Yes (1st PML case submitted to the manufacturer and to the FDA [7/04/2011])||Yes (3rd PML case submitted to manufacturer and FDA [11-15-2011])||No (2nd PML case submitted to the FDA [10/2011])||No (4th PML case submitted to the manufacturer and to the FDA [3/29/2012])|
|Age at time of PML diagnosis, y/sex||51/man||47/man||50/man||38/woman||72/woman|
|Speech/mental status||Difficulty with word finding, confusion, forgetfulness over several wk||Dysarthria, encephalopathy||Dysarthria||Mixed nonfluent aphasia, apraxia||Poor short-term memory, confusion|
|Motor||Normal||Left hemiparesis, impaired coordination||Right leg weakness and difficulty writing with right arm weakness||Mildly decreased strength in all extremities||Generalized weakness|
|Gait||Normal||Impaired due to hemiparesis||Impaired due to right leg weakness||Ataxia requiring 1-person assist||Not tested|
|Differential diagnosis at time of PML presentation||PML or viral encephalitis, based on clinical and MRI findings||Ischemic stroke||Subacute CNS ischemia or progression of lymphoma||Metastatic brain lesions||PML, based on clinical findings and MRI findings|
|Detection of JCV||CSF: JCV DNA detected by PCR||CSF: JCV DNA detected by PCR in second LP; no JCV detected in initial LP||Spinal cord lesion biopsy, initially negative for JC virus; later evaluation of the sample detected JC virus; JC virus not detected in CSF from 2 LPs||Brain biopsy- JCV detected by immunohistochemistry.||Brain autopsy with JCV detected by in situ hybridization; JCV not detected in CSF|
|MRI||Multifocal areas of nonenhancing T2 white matter, which were greatest in the anterior/inferior right frontal lobe; additional areas of subcortical T2 hyperintensity in bilateral temporal lobes||T2 hyperintense lesion in right cerebral hemisphere with well-circumscribed enhancing lesions in right parietal lobe||Ill-defined abnormal T2 hyperintensity involving left superior cerebellar peduncle||Multifocal white matter lesions noted in left more than right hemisphere||Asymmetric T1 hypointense and T2 hyperintense lesions of the frontal white matter|
|Other illnesses||Thyroid cancer, thyroidectomy, chronic lymphopenia||Diet-controlled diabetes mellitus, depression||Epilepsy, migraine headaches||Severe eczema||None|
|Prior medications (between 9-12 y prior to PML)|| || ||ABVD; radiation to abdomen and pelvis||Topical steroid creams|| |
|Prior medications (between 5-8 y prior to PML)|| ||Broad-field radiotherapy and ABVD||ICE (for recurrence) with partial response; BEAM; autologous stem cell transplant||Topical steroid creams|| |
|Between 3 and 4 y prior to PML|| ||ESHAP (for recurrence), BEAM, autoSCT,||Localized radiotherapy to right neck (for second recurrence of nodular sclerosing Hodgkin disease)||Topical steroid creams||Topical corticosteroids, nitrogen mustard, phototherapy, interferon-α-2b|
|Between 1 and 2 y prior to PML|| ||gemcitabine, vinorelbine, and liposomal doxorubicin||None||Methotrexate, bexarotene, denileukin diftitox, interferon-gamma-1b; interferon-α-2b||Topical corticosteroids, nitrogen mustard, phototherapy, interferon-α-2b|
|<1 y prior to brentuximab vedotin||ABVD|| || ||Bexarotene, denileukin diftitox, interferon-gamma-1b, interferon-α-2b, vorinostat, pralatrexate||Bexarotene, radiotherapy, romidepsin|
|After brentuximab vedotin||Autologous stem cell transplantation; ICE|| ||Augmented ICE (2 cycles)||Prednisone|| |
|No. of doses of brentuximab vedotin and clinical response||6 doses (1.2 mg/kg every wk); total of 2 cycles of 3 doses; PET scan became negative after last brentuximab vedotin dose.||3 doses (1.8 mg/kg) 21 d apart (PET showed favorable response to therapy)||5 doses (1.8 mg/kg); clinical trial participant was withdrawn due to peripheral neuropathy; 2 additional doses (1.2 mg/kg)||2 doses (1.8 mg/kg every 3 wk) (1 cycle); disappearance of skin tumors||3 doses (1.8mg/kg) (significant response of skin lesions)|
|Subsequent course||Progressive neurological deterioration; died 8 wk after PML diagnosis||Progressive neurological deterioration and worsening MRI; died in hospice 8 wk after PML diagnosis||Progressive neurological deterioration and worsening MRI findings; T1-T2 and T8-T9 intramedullary lesions; radiotherapy (3000 Gy) to T1-T2 lesion and corticosteroids; paraplegia developed and died in hospice 6 wk after PML diagnosis and 20 wk after first brentuximab vedotin treatment||Neurological and radiological improvement; maintained on prednisone at a dose of 50 mg orally/d for 1 y; upon discontinuing corticosteroids, a new enhancing brain lesion developed and the patient restarted prednisone at a dose of 20 mg orally/d with resolution of this lesion and no evidence of new lesions||Progressive neurological deterioration; died in hospice 15 wk after first brentuximab vedotin treatment|
|Time from first brentuximab vedotin to main PML symptoms||36 wk; 6 mo after the patient treated with autoSCT demonstrated confusion, forgetfulness, and difficulty finding words||3 wk; impaired coordination, left >right dysiadokinesis, slurred speech, left-sided hemianopsia and visual neglect, left hemiparesis, and left central facial paresis||24 wk; speech changes, incoordination of right hand (poor penmanship), and right lower extremity weakness||3 wk; mixed aphasia, inability to read, and unsteady gait||7 wk; disorientation, poor short-term memory, and weakness|
The 3 patients with classic Hodgkin lymphoma were negative for HIV and had received standard frontline and salvage combination chemotherapy regimens. Initial therapy was the combination of doxorubicin, bleomycin, vinblastine, and dacarbazine for all 3 patients with Hodgkin lymphoma. The patients also received autologous stem cell transplantation (3 years and 5 years previously, respectively, for 2 patients and immediately after BV for 1 patient) after standard salvage regimens using combination chemotherapy with ifosfamide, carboplatin, and etoposide (2 patients) and etoposide, methylprednisolone, cytarabine, and cisplatin (1 patient). Two patients with Hodgkin lymphoma also received radiotherapy to the abdomen and pelvis. The 3 patients had participated in clinical trials of BV and had responded to BV therapy. In evaluating neurologic dysfunction, initial brain MRI scans revealed multifocal white matter lesions in these 3 patients. JCV DNA was not detected on at least 1 occasion in CSF samples obtained from 2 of these patients, one of whom had an immunostain of a spinal cord lesion biopsy (at the level of the first thoracic vertebra) that was positive for JCV and the second for whom a repeat CSF evaluation was positive.
The 2 patients with cutaneous lymphomas had not participated in clinical trials of BV. They received BV as an off-label treatment for their lymphoma. The patient with transformed mycosis fungoides was negative for HIV and was initially diagnosed with stage IB mycosis fungoides (using TNMB staging system) 4 years before transformation. This patient had received topical corticosteroids, topical nitrogen mustard, phototherapy, interferon-α, bexarotene, localized radiotherapy, and the histone deacetylase inhibitor romidepsin before BV therapy. The patient responded to BV therapy. In evaluating neurologic dysfunction, this patient underwent a brain MRI scan that revealed multifocal white matter lesions. A brain autopsy revealed JCV-positive cells detected by in situ hybridization. JCV had not been detected in CSF samples obtained before the death of this patient. The patient with primary cutaneous anaplastic large cell lymphoma was also negative for HIV. The patient had received topical corticosteroids, methotrexate, bexarotene, denileukin diftitox, interferon-α and interferon-gamma, pralatrexate, and the histone deactylase inhibitor vorinostat previously. The patient also responded to BV treatment. A brain biopsy revealed scattered demyelination and immunohistochemistry demonstrated oligodendrocytes that were positive for JCV.
The most common clinical presentations of PML in the current series were speech dysfunction/aphasia (4 patients), hemiparesis (2 patients), hemianopsia (1 patient), memory loss (2 patients), gait dysfunction (3 patients), and confusion (2 patients) (Table 1). The onset of neurological symptoms occurred days to weeks after the administration of 2 to 6 doses of BV (1.2 mg/kg to 1.8 mg/kg administered intravenously). One patient (with classic Hodgkin lymphoma) underwent an autologous stem cell transplantation 7 weeks after the last of 6 BV treatments and neurologic symptoms developed over the next 4 months. In 4 patients, neurological dysfunction worsened rapidly and progressively over a period of 3 to 8 weeks until death occurred in hospice settings. The 38-year-old patient with primary cutaneous anaplastic large cell lymphoma demonstrated clinical and MRI findings that were suggestive of PML-associated IRIS (PML-IRIS). The diagnosis of PML-IRIS was further supported by infiltrating T cells identified on brain biopsy. The patient improved clinically and radiographically with treatment with daily oral prednisone. When corticosteroids were tapered off several months later, a new contrast-enhancing lesion developed on brain MRI, reflecting reactivation of PML-IRIS. At the time of reinitiation of prednisone, this brain lesion regressed and the patient had no further lesion formation and achieved partial neurologic recovery after several months of follow-up while continuing to receive prednisone therapy.
To the best of our knowledge, the current study is the first comprehensive case series reporting 5 cases of PML among patients treated with BV, 4 of whom died within weeks of the PML diagnosis. In interpreting these findings, several factors should be considered.
First, there are important differences in the clinical features noted in patients who developed PML after BV treatment when compared with PML that developed after exposure to the monoclonal antibodies rituximab, efalizumab, and natalizumab. Time to the onset of symptoms and duration of therapy before PML was much shorter in the patients treated with BV. Three of the patients with PML were symptomatic after only 2 or 3 doses of BV (administered every 3 weeks) and the fourth was symptomatic after the fifth dose. In contrast, PML was diagnosed at a median of 63 weeks after initial exposure to rituximab. For natalizumab, PML symptom onset occurred after a median of 26 months of treatment. For efalizumab, all 3 confirmed cases developed ≥ 3 years after treatment initiation. It is also important to note that the case fatality rate was low for natalizumab at 22% when compared with 80% for BV, 90% for rituximab, and 100% for efalizumab.[7, 13, 22]
Second, in the HIV-negative population, lymphoid malignancies, hematopoietic transplantation, and multiagent chemotherapy are all risk factors associated with the development of PML. Thus, they are potential confounders in the assessment of the association between PML and BV. Three patients in the current series had a history of recurrent classic Hodgkin lymphoma and had received multiple lines of combination chemotherapy and autologous stem cell transplantation. PML onset was 3 years and 5 years after transplantation, respectively, in 2 of these patients, whereas PML occurred 6 months after autologous stem cell transplantation in the third patient. Previous case series have suggested that most hematopoietic transplantation-associated PML cases among individuals with lymphoid malignancies occur within 24 months of the transplant procedure. The remote nature of the transplant procedures in 2 of the 3 patients with classic Hodgkin lymphoma in the current study argue against this as a major contributing risk factor. Similarly, 2 of the 5 patients in the current series had cutaneous lymphomas and had not previously received multiagent chemotherapeutic regimens.
Third, consideration of possible mechanisms of PML causation by BV is warranted. Depletion of CD30-expressing activated T cells could reduce JCV immune surveillance in the CNS, resulting in PML. Related to this, CD30 has been proposed as a therapeutic target on alloreactive activated T cells, causing acute graft-versus-host disease. Just as depletion of CD30-positive T cells might be beneficial in patients with graft-versus-host disease, it may also reduce immune surveillance in patients predisposed to JCV reactivation and PML.
Fourth, to the best of our knowledge, there is an incomplete understanding of what factors may influence the development of and outcomes associated with PML-IRIS among the small number of individuals who develop this syndrome after receiving immune-modulating monoclonal antibody therapies. Among 52 previously reported patients with non-Hodgkin lymphoma with PML who received rituximab, PML-IRIS was not noted. This suggests that patients with lymphoma, perhaps due to immune suppression related to previous treatments or the malignancy itself, may be less able to develop the immune response necessary for PML-IRIS. In contrast, all of the 40 recently reported patients with multiple sclerosis with natalizumab-associated PML developed PML-IRIS (a syndrome that is commonly noted among HIV-infected individuals with PML) after withdrawal of natalizumab and after undergoing plasmapheresis. The development of PML-IRIS shortly after the last dose of BV in the sole survivor of BV-associated PML to the best of our knowledge represents the first case of PML-IRIS in a person with lymphoma who developed PML after therapy with rituximab or BV.[6, 13] As highlighted in a recent report of PML in association with compounded dimethyl fumarate, the development of PML-IRIS after drug discontinuation supports a causal relationship. A prolonged course of corticosteroids may also be required to facilitate neurologic recovery in patients with PML-IRIS, as observed in the survivor in the current case series and suggested in previous literature.
Fifth, the results of the current study have clinical implications. The findings reinforce the importance of maintaining a high index of suspicion for PML when evaluating patients treated with BV who present with speech disturbances, neurocognitive changes, and/or other persistent neurological symptoms. The presence of multifocal white matter lesions on brain MRI scans performed among patients with classic Hodgkin lymphoma should further raise suspicion for PML, given the extreme rarity of multifocal CNS spread of classic Hodgkin lymphoma. Health care providers should also be aware that even when JCV DNA is not detected during CSF sampling, PML remains a possibility due to the low sensitivity of tests used to detect JCV DNA in the CSF. JCV DNA was not detected in CSF specimens obtained from 2 patients, although it was later identified on CNS biopsy specimens from these same patients. JCV DNA was detected in only 1 of 2 CSF samples obtained from a third patient. Thus, in the setting of a high index of suspicion, brain biopsy may be required to confirm or exclude a diagnosis of PML. More sensitive polymerase chain reaction assays currently are being developed to help overcome these challenges.
There are some limitations to the current study. Detailed clinical histories and medical records were available for the PML episode for 4 patients and an FDA safety report was available for the fifth patient. These data sources did not include information regarding prior infectious complications or leukopenia episodes that may have occurred with prior treatments. In addition, data concerning the total number of individuals who have received BV were not available, thereby limiting our ability to estimate the incidence of BV-associated PML.
The findings of the current study have policy implications. One consideration would be to develop a BV-related FDA-approved Risk Evaluation and Mitigation Strategy (REMS) program modeled after that established for natalizumab (Tysabri Outreach: Unified Commitment to Health [TOUCH]). A REMS program is grounded in the concept of risk/benefit tradeoffs, primarily for beneficial drugs associated with important risks in which the REMS ensures that patients and providers understand the risks and benefits before initiating treatment with the pharmaceutical agent.
We conclude that PML is a possible adverse drug reaction that may occur shortly after initiating BV treatment and after a few doses of BV have been administered. For patients with recurrent classic Hodgkin lymphoma, systemic anaplastic large cell lymphoma, or transformed cutaneous lymphoma, the high response rates from BV among those with these diagnoses who otherwise have few treatment options support favorable risk-benefit profiles in these settings.[28, 29] In contrast, the decision to administer BV to patients with indolent cutaneous CD30-positive lymphomas, such as cutaneous anaplastic large cell lymphoma and lymphomatoid papulosis, should be made with careful consideration of alternative options that, such as BV, have high response rates and produce cosmetic benefits, but may have more favorable risk profiles.
Supported in part by the National Cancer Institute (1R01CA165609-01A1), the American Cancer Society (IRG-13-043-01, MSRG-13-077-01-CPHPS and MRSG 3071-81082), the South Carolina SmartState Program, and the Doris Levkoff Meddin Medication Safety Education Program.
CONFLICT OF INTEREST DISCLOSURES
Dr. Carson has acted as a consultant, received research funding, and participated on an advisory board for Millennium Pharmaceuticals. Dr. Newsome has acted as a member of the scientific advisory boards for Biogen Idec and Genzyme. Dr. Craig H. Moskowitz has acted as a member of the Scientific Advisory Board and received research support from Seattle Genetics. Dr. Alison J. Moskowitz has received research funding from Seattle Genetics. Dr. Bookstaver has acted as a member of the advisory boards of Gilead Pharmaceuticals and Durata Pharmaceuticals.