Benefit of Additional Screening for Progressive Multifocal Leukoencephalopathy in Patients With Multiple Sclerosis Taking Natalizumab: A Decision Analysis

of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information Report
Category:

Religious & Philosophical

Published:

Views: 38 | Pages: 7

Extension: PDF | Download: 0

Share
Description
Benefit of Additional Screening for Progressive Multifocal Leukoencephalopathy in Patients With Multiple Sclerosis Taking Natalizumab: A Decision Analysis
Tags
Transcript
  Bene  󿬁 t of Additional Screening for Progressive MultifocalLeukoencephalopathy in Patients With Multiple SclerosisTaking Natalizumab: A Decision Analysis  David C. Landy, PhD, MPH and Eric M. Hecht, MD, MSPH  Objective:  Patients with multiple sclerosis (MS) taking natalizumabare at risk for progressive multifocal leukoencephalopathy (PML). Wesought to describe the outcomes of discontinuing natalizumab on the ba-sis of PML risk and those of obtaining additional screening across arange of scenarios using decision tree models. Methods:  Health state probabilities and values, measured as the pro- portion of quality  –  adjusted life years (PQALY) relative to baselinehealth, were based on literature review. Probabilities of worsening MSwhile continuing and discontinuing natalizumab were set to 0.23 and 0.44. For discontinuing therapy, PML risk, worsening MS value, and PML value were varied. For additional screening, the probability of discontinuing natalizumab without screening, PML risk, worseningMS value, and PML value were set to 33%, 1.1%, 0.88, and 0.09, re-spectively, with test sensitivity and speci 󿬁 city varying from 0.50 to 1. Results:  Discontinuing natalizumab provided no bene 󿬁 t until PMLrisk reached 2.9%, assuming an MS relapse value of 0.88 and a PMLvalue of 0.09. Additional screening changed the PQALY by  − 0.3% to1.5%, largely in 󿬂 uenced by speci 󿬁 city. Assuming a sensitivity of 80%and a speci 󿬁 city of 99%, screening increases the PQALY by 1.2%. Conclusions:  The highest PML risk identi 󿬁 ed by strati 󿬁 cation is be-low 2.9%, suggesting that continuing natalizumab outweighs PML risk for most patients on the basis of theoretical calculations. However,decisions based on additional screening with high-speci 󿬁 city tests, in-cluding polymerase chain reaction cerebrospinal  󿬂 uid tests for JohnCunningham virus, may provide bene 󿬁 t and should be clinically tested.Increased precision of probabilities and quality-of-life values are alsoneeded to improve decision making. Key Words:  multiple sclerosis, progressive multifocalleukoencephalopathy, natalizumab( Clin Neuropharm  2014;37: 45  –  51) N atalizumab was approved for the treatment of relapsingmultiple sclerosis (MS) in 2004 based in part on the Natalizumab Safety and Ef  󿬁 cacy in Relapsing-Remitting Mul-tiple Sclerosis (AFFIRM) trial, which showed that natalizumabsigni 󿬁 cantly decreased MS relapses, decreased disease progres-sion, and increased quality of life compared with placebo. 1,2 Ayear later, natalizumab was withdrawn after several patients de-veloped progressive multifocal leukoencephalopathy (PML), arare and severe neurological condition caused by infection or re-activation of the John Cunningham (JC) virus. 3  –  5 Because of its ef  󿬁 cacy compared with alternative therapies,natalizumabwasthenbroughtbacktothemarketin2006,accom- paniedbytheTysabrioutreach:uni 󿬁 edcommitmenttohealthpro-gram, which restricted access to registered health care providers. 6 The program also required evaluating patients 3 and 6 months af-ter starting natalizumab and every 6 months thereafter. The pro-gram relied on clinical evaluation and did not specify a role for risk factor screening.Under this model, more than 350 patients developed PML. 7 Review of these cases identi 󿬁 ed several risk factors for develop-ing PML including duration of natalizumab therapy, prior immu-nosuppression therapy, and serum JC virus antibodies. 8 Althoughthese factors discriminate patients by relative PML risk, their ab-solute discrimination is poor, with a difference of only 1.1% sep-arating the highest- and lowest-risk patient groups. These factorsalso fail to identify all at-risk patients. 9,10 Currently, more than 100,000 patients take natalizumabwith most having one or more risk factors. 8 More than half haveJC virus antibodies in their serum, more than half have takennatalizumab for 2 or more years, and nearly half have received  prior immunosuppressive therapy. 8,11,12 If these patients re-main on natalizumab, some are expected to develop PML.Recommendations for clinical practice now include stratifying patients by risk factors and discontinuing therapy in patients positive for all 3 risk factors or even fewer. 13  –  17 For some patients, natalizumab is the last-line therapy and the only mechanism by which to control their disease. 18 For these patients, the discontinuation of natalizumab and the likelyonset of worsening MS must be compared with their risk for de-veloping PML. For these patients, it is unclear what absoluterisk for developing PML justi 󿬁 es discontinuing natalizumab.Patient-centered data suggest that many are willing to toleratePML risk to avoid worsening MS. 19,20 Although PML can bea devastating complication in patients with MS, recent evidencesuggests that natalizumab-induced PML may be less severe thanthat more commonly seen in patients with HIV. 21,22 Additional screening may identify patients where the deci-sion to continue or discontinue natalizumab is better justi 󿬁 ed.We sought to determine the risks and bene 󿬁 ts of continuingnatalizumab across increasing levels of PML risk and those of obtaining additional screening through decision analysis. METHODS Decision tree models were created to compare possiblechoices of patients with MS taking natalizumab with modelsdepicting paths from choices to the health states that could re-sult. The outcome of each choice was de 󿬁 ned as the sum of the probabilities of the resulting health states after that choicemultiplied by the values of the health states. Choices were com- pared by both probabilities of the resulting health states, Department of Public Health Sciences, University of Miami Miller School of Medicine, Miami, FL.Con 󿬂 icts of Interest and Source of Funding: The authors have no con 󿬂 ictsof interest to declare.Supported by an investigator-initiated grant from Seedlings Life Sciences,LLC. The authors had full access to all data and independence withrespect to any and all publication decisions.Address correspondence and reprint requests to David Cooper Landy, PhD,MPH, Department of Public Health Sciences, University of Miami Miller School of Medicine, Miami, FL; E-mail: DCLandy@med.miami.eduCopyright © 2014 by Lippincott Williams & WilkinsDOI: 10.1097/WNF.0000000000000018 O RIGINAL  A RTICLE Clinical Neuropharmacology   •  Volume 37, Number 2, March/April 2014 www.clinicalneuropharm.com  45 Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.  expressed per 1000 patients, and outcomes that incorporated thevalues of these states. Decisions were examined across a rangeof health state probabilities and values so that choices could  be compared across this spectrum of scenarios. These probabil-ities and health state values were selected on the basis of litera-ture review.The  󿬁 rst model was for patients taking natalizumab decid-ing to continue therapy on the basis of concerns over PML risk versus discontinuing therapy (Fig. 1A). It was assumed that these patients showed no clinical signs of PML and that their MS was well controlled on therapy. For patients choosing tocontinue therapy, 1 of 3 potential health states could result: de-velop PML, have worsening MS, or remain at baseline health.Worsening MS was considered to be a hybrid health state and represented deterioration from baseline greater than expected as compared with a more indolent disease course. For patientschoosing to discontinue therapy, 1 of 2 potential states could re-sult: have worsening MS or remain at baseline health.The second model was for patients taking natalizumab de-ciding whether to pursue additional screening on which to basetheir decision to continue therapy versus basing this decision onrisk factor strati 󿬁 cation alone (Fig. 2A). It was assumed that these patients show no clinical signs of PML and that their MS is well controlled on therapy. Further, it was assumed that these patients have all 3 risk factors currently proposed for strat-ifying patients: have been on natalizumab for 2 years, received immunosuppressive therapy, and have serum JC virus anti- bodies. For patients choosing to receive additional screening,1 of 3 screening test results could result: false negative, truenegative, or positive, either false or true. For patients with a FIGURE 1.  Decision tree model of the decision to continue natalizumab (A) with analysis of a speci  󿬁 c scenario provided (B). Pts indicatespatients; QAL, quality-adjusted lives.  Landy and Hecht   Clinical Neuropharmacology   •  Volume 37, Number 2, March/April 2014 46  www.clinicalneuropharm.com  © 2014 Lippincott Williams & Wilkins Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.  FIGURE 2.  Decision tree model of the decision to receive additional screening for PML upon which to base the decision to continuenatalizumab (A) with analysis of a speci  󿬁 c scenario provided (B). Pts indicates patients; QAL, quality-adjusted lives. Clinical Neuropharmacology   •  Volume 37, Number 2, March/April 2014  Decision Analysis of Continuing Natalizumab  © 2014 Lippincott Williams & Wilkins  www.clinicalneuropharm.com  47 Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.  false-negative result, it was assumed that all developed PML. For  patients with a true-negative result, it was assumed that they con-tinued therapy and either had worsening MS or remained at base-line health. For patients with a positive test, it is assumed that they discontinued therapy and either had worsening MS or remained at baseline health.In the  󿬁 rst model, the risk for PML while taking natalizumabwas  󿬁 rst allowed to vary from 0% to 1.5% to represent the rangeof PML risks predicted by the currently proposed method of risk strati 󿬁 cation. Second, the risk for PML varied from 1% to 10%to assess a wider range of PML risks that might be predicted  by incorporating additional screening results. The risk for wors-ening MS while taking natalizumab and not taking natalizumabwere estimated on the basis of the MS relapse rates for the ther-apy and placebo groups from AFFIRM trial at 23 and 44%, re-spectively. The probability of baseline health was then de 󿬁 ned as 1 minus both the probability of PML and worsening MS.In the second model, the risk for PML while takingnatalizumab was set equal to 1.1% because this model was for  patients positive for all 3 risk factors. For patients deciding onthe basis of risk factor strati 󿬁 cation alone, the probability of choosing to discontinue therapy wasset at 33% on the basis of re-cent recommendations by some that these patients discontinuetherapy. For patients pursuing additional screening, the probabil-ity of a false-positive test result was set equal to the risk for PML multiplied by 1 minus screening test sensitivity. The proba- bility of a true-negative result was set equal to 1 minus PML risk multiplied by screening test speci 󿬁 city. The probability of a positive-testresultwassetequaltoPMLriskmultipliedbyscreen-ing test sensitivity plus 1 minus PML risk multiplied by 1 minusscreening test speci 󿬁 city. Both speci 󿬁 city and sensitivity wereevaluated across the range from 0.50 to 1 to capture the range of values that could be associated with additional screening tests.The PML valuewas set to 0.09; theworsening MS value, to 0.88.For the analysis of both models, health state values weremeasured as the proportion of quality  –  adjusted life years(PQALY) associated with having that health state for 2 yearsrelative to remaining at baseline health for those 2 years, de- 󿬁 ned as the expected life of a patient with controlled MS. Thequality-adjusted life years associated with remaining at baselinehealth for those 2 years were set equal to 1. The PQALYassoci-ated with other health states could be reduced because of either a decreased duration of life or a decreased quality of life. De-creased quality of life associated with a health state was consid-ered to be both the quality during the 2-year period and anydeferred impact on quality. For worsening MS and PML, evi-dence suggests that decreased duration and quality of life likely both impair the PQALY.For PML, short-term mortality rates vary signi 󿬁 cantly withstudies of PML in HIV-positive patients reporting mortality at 50% andhigher. 23 However, for PML associatedwith natalizumab,mortality seems to be lower. 21 For surviving patients withnatalizumab-associated PML, quality of life is signi 󿬁 cantly im- paired, but most patients stabilize several months after diagnosis. 21 Tore 󿬂 ect this heterogeneity and uncertainty, the PQALYassociated with PML varied from 0.03 to 0.15. The lower bound of this rangealso re 󿬂 ects the long-term effect of developing this health state rel-ative to remaining at baseline health.For worsening MS, long-term survival is impaired and highly correlated with accumulating disability. 24 In addition,although the long-term effects of worsening MS over a period such as 2 years are not possible to predict, it is this perceptionof risk that patients are forced to grapple with when weighingthe risks of bene 󿬁 ts of continuing therapy. These patients mayhave a decreased life-span relative to patients with MSremaining at baseline health. 25 Patients with worsening MS alsohave a reduced quality of life as measured in cohort studies and the AFFIRM trial. 26  –  28 To re 󿬂 ect this heterogeneity and un-certainty, the PQALY associated with worsening MS duringa 2-year span varied from 0.80 to 0.96. The upper bound of thisrange also re 󿬂 ects the potentially modest de 󿬁 cit of having wors-ening MS during a limited 2-year period.Although the health state probabilities and values varied and the outcomes of each choice measured across these FIGURE 3.  Change in the PQALY associated with discontinuing compared with continuing natalizumab by PML risk and worsening MSvalue with PML value set at 0.09 across PML risks consistent with risk factor strati  󿬁 cation levels from 0% to 1.5% (A) and a broader range from 1% to 10% (B).  Landy and Hecht   Clinical Neuropharmacology   •  Volume 37, Number 2, March/April 2014 48  www.clinicalneuropharm.com  © 2014 Lippincott Williams & Wilkins Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.  different scenarios, examples analyzing speci 󿬁 c scenarios arealso provided (Figs. 1B, 2B). RESULTS For every 1000 patients discontinuing natalizumab inscenarios with PML risk levels consistent with risk factor strati 󿬁 -cation, there were 1 to 15 fewer PML cases but 207 to 210 moreworsening MS cases. Those discontinuing therapy had a change intheir PQALY from a loss of 4.1% to a gain of 0.6% across thesescenarios. When the PML value was held at 0.09, the change inPQALY after discontinuing therapy varied from a loss of 4.1%to a gain of 0.5%, depending on MS worsening value (Fig. 3A).For every 1000 patients discontinuing natalizumab inscenarios with PML risk levels from 1% to 10%, there were10 to 100 fewer PML cases but 189 to 207 more worseningMS cases. Compared with patients continuing natalizumab,those discontinuing therapy had a change in their PQALY froma loss of 3.4% to a gain of 8.8% across these scenarios. Whenthe PML value was held at 0.09, the change in PQALY after discontinuing therapy varied from a loss of 3.3% to a gain of 8.3%, depending on worsening MS value (Fig. 3B). When thevalues of PML and worsening MS were held at 0.9 and 0.88, re-spectively, the PML risk at which patients discontinuing versuscontinuing therapy had greater PQALY was 2.9%.For every 1000 patients with all 3 stratifying risk factorsreceiving additional screening versus basing their decision onstrati 󿬁 cation alone in the scenarios examined, there were between7 and 2 fewer cases of PML and from 67 fewer to 38 more casesof worsening MS. Compared with patients basing their decisionon strati 󿬁 cation alone, those receiving additional screening had a change in their PQALY from a loss of 0.3% to a gain of 1.5% across these scenarios, and this varied on the basis of thesensitivity and speci 󿬁 city of the additional screening test withspeci 󿬁 city appearing to have a larger in 󿬂 uence (Fig. 4). DISCUSSION For patients without other therapeutic options and PMLrisk at levels consistent with risk factor strati 󿬁 cation, 0% to1.5%, the bene 󿬁 ts of continuing natalizumab likely outweighthe possible effects of PML. For patients assigning values rela-tive to baseline of 0.09 for PML and 0.88 for worsening MSduring a 2-year period, their risk for PML would need to be at least 2.9% for the effect of PML to outweigh the bene 󿬁 t of con-tinuing natalizumab and having a reduced probability of worsen-ing MS. As PML risk exceeds 5%, few patients would likely bene 󿬁 t from continuing therapy, although this level is roughly5 times greater than the highest level predicted by risk factor strati 󿬁 cation. Identifying patients with such a high predicted risk will require additional screening, which our results suggest should be highly speci 󿬁 c to provide the greatest bene 󿬁 t.The manner in which patients actually weigh the risks for PML and worsening MS is not well understood. Some patientsmay be willing to tolerate more PML risk, especially if the dis-ease course is less severe than that reported in other populationssuch as HIV patients. 18  –  20 Given the complexity of the sce-narios and the lack of data on the precise probabilities involved,decision making of the patient is challenging and inexact. Eventhe quality of life associated with these health states is dif  󿬁 cult to estimate and varies across individual patients, further compli-cating patient counseling.In a review of more than 2000 patients with MS enrolled inclinical trials of natalizumab, patients experiencing a relapsehad signi 󿬁 cantly worse physical functioning. 2 The progressivenature of these relapses and disease progression over increased follow-up are not known. In a review of 35 patients withnatalizumab-induced PML and a median follow-up of 5 months,25 survived, although nearly half of these patients were severelydisabled. 21 A physician-guided, patient-centered approach willhopefully maximize decision quality. Additional data on healthstate probabilities and quality-of-life measures aswell as decisionsupport tools are also likely to help the decision-making process.For patients at the highest PML risk level identi 󿬁 ed bystrati 󿬁 cation, our data suggest that additional screening might  provide additional bene 󿬁 t. Of note, the model in this study eval-uated 1-time screening and assessed the outcomes of different choices on the basis of the effects of having a health statefor just 2 years. In actual practice, sequential screening would  FIGURE 4.  Change in the PQALY associated with receiving additional screening for PML upon which to base the decision to continuenatalizumab compared with basing the decision on risk factor strati  󿬁 cation alone. The proportion discontinuing natalizumab withoutadditional screening, PML risk, PML value, and worsening MS value were set at 33%, 1.1%, 0.09, and 0.88, respectively. Clinical Neuropharmacology   •  Volume 37, Number 2, March/April 2014  Decision Analysis of Continuing Natalizumab  © 2014 Lippincott Williams & Wilkins  www.clinicalneuropharm.com  49 Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Recommended
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x