Article Text

Download PDFPDF

Natalizumab-associated progressive multifocal leucoencephalopathy: a practical approach to risk profiling and monitoring
  1. David Hunt1,
  2. Gavin Giovannoni2
      1. 1Department of Clinical Neurosciences, University of Edinburgh, Edinburgh, UK
      2. 2Department of Neuroscience and Trauma, Blizard Institute of Cell and Molecular Science, Barts and The London School of Medicine and Dentistry, London, UK
      1. Correspondence to Dr David Hunt, Clinical Lecturer in Neurology, Department of Clinical Neurosciences, University of Edinburgh, Chancellor's Building, Little France, Edinburgh, Midlothian, EH16 4SB, UK; david.hunt{at}


      Natalizumab reduces relapse frequency, delays onset of disease progression and improves disease outcomes in relapsing–remitting multiple sclerosis (MS) and is a cost-effective treatment for rapidly evolving severe relapsing–remitting MS. However, it is associated with the development of progressive multifocal leucoencephalopathy (PML), a serious opportunistic brain infection caused by a neurotropic strain of the JC virus (JCV). Until May 2011, 83 300 patients had received natalizumab for MS. One hundred and twenty-four patients had developed PML, of whom 23 (19%) died. In order to maximise the benefit–risk ratio of natalizumab for MS patients it is important to develop a strategy for risk profiling and monitoring for PML. Central to this is an understanding of the biology of the JCV and the emerging clinical picture of natalizumab-associated PML. This paper reviews the evidence for managing the risk of PML in natalizumab-treated patients and the authors propose an algorithm for risk profiling and risk management. Key features of this algorithm include risk stratification based on emerging risk factors, heightened clinical vigilance for the clinical features of natalizumab-associated PML and considerations for temporary and permanent cessation of natalizumab dosing.

      Statistics from

      Request Permissions

      If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


      Natalizumab (Tysabri; Biogen-Idec, Elan Pharmaceuticals, Cambridge, MA, USA)is a humanised monoclonal antibody directed against the α4 subunit of the adhesion molecule α4β1 integrin. Natalizumab binds α4 integrin on leucocytes, reducing their ability to migrate across the blood–brain barrier.1 2 It has proved to be a cost-effective treatment for highly active or rapidly evolving severe relapsing–remitting multiple sclerosis (MS).3 Among patients with highly active MS enrolled into the AFFIRM (Natalizumab Safety and Efficacy in Relapsing-Remitting MS) and SENTINEL (Safety and Efficacy of Natalizumab in Combination with Avonex in Patients with Relapsing-Remitting MS) phase III trials, there was an 81% reduction in annualised relapse rate compared with placebo over 2 years and a 64% reduction in the risk of disability progression over 2 years.1 4 5

      However, these impressive efficacy data were tempered by three patients, treated with natalizumab in clinical trials for MS and Crohn's disease, who subsequently developed progressive multifocal leucoencephalopathy (PML).6 PML is an opportunistic viral demyelinating disease caused by a neurotropic mutant strain of the JC virus (JCV), most often in the context of immunodeficiency.7 Regulatory authorities suspended natalizumab use in 2005 pending a detailed independent review. The review suggested the PML risk was approximately 1 in 1000 for patients treated for 18 months.8 Following the implementation of a risk management strategy, the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA) remarketed natalizumab as monotherapy for MS in July 2006, with recommendations for monitoring for new cases of PML.9

      Until May 2011, 83 300 patients had been treated with natalizumab in the post-marketing setting, representing 148 800 patients-years of natalizumab exposure. There were 124 proven cases of PML, resulting in 23 deaths (19% mortality). The overall incidence of PML in patients treated with natalizumab is 1.44 per 1000 patients (95% CI 1.20 to 1.72), similar to that in clinical trials (risk=1.00 per thousand (95% CI 0.2 to 2.8) (Biogen-Idec Medical Information Service, May 2011)).

      In the UK, natalizumab was approved by the National Institute for Health and Clinical Excellence to treat ‘rapidly evolving severe relapsing–remitting MS’,3 defined by two or more disabling relapses in 1 year and one or more gadolinium-enhancing lesions on brain MRI or a significant increase in T2 lesion load compared with a previous MRI.3 The risk of PML increases with duration of therapy; the majority of cases present after at least 12 months of therapy.

      As a result of concerns regarding emerging PML cases and evidence of increasing risk after 24 months of treatment, the EMA Committee for Medicinal Products for Human Use reviewed natalizumab's safety and recommended that the benefits still outweigh the risks for patients with highly active relapsing–remitting MS.9 The FDA made similar recommendations. The updated prescribing information reflects the following:

      1. Patients should be informed about the risk of PML via a new treatment initiation/continuation form.

      2. The clinician and patient together should reconsider the benefits and risks of the treatment after 2 years.

      3. Patients should undergo MRI scans within 3 months before starting natalizumab, with scans repeated annually.

      4. Natalizumab should be promptly stopped on suspicion of PML, with subsequent appropriate evaluation, including a standardised MRI scan and lumbar puncture. These patients should be monitored closely for the development of immune reconstitution inflammatory syndrome (IRIS).

      5. Patients should be treated preferably as part of national registries or post-marketing studies.

      There is a clear need to develop and implement a risk profiling strategy for patients with MS receiving natalizumab. This paper reviews the scientific rationale for such a strategy and we propose an algorithm for risk management that is being implemented in several UK centres.

      Natalizumab-associated PML: the emerging clinical picture

      PML and JCV: overview and pathogenesis

      The JCV is a 40 nm non-enveloped double stranded DNA polyomavirus encoding six regulatory and structural proteins.10 It has a 70% homology with BK virus (BKV).10 JCV infection is prevalent in most people, with infection rates of 30–80% in different populations, varying with age and geographical area.10 11 Higher seroprevalence rates, reported in earlier studies, are probably due to less specific assays with cross-reactivity between JCV and BKV.11 More recent studies with specific assays that distinguish between BKV and JCV report adult seroprevalence rates of 30–60%.11 12

      Figure 1 outlines the proposed stages of PML pathogenesis. Sites of JCV persistence may include the kidney, bone marrow and the lymphoreticular system, for example, tonsils and spleen.10 Adaptive mutations in the non-coding transcriptional control region and VP1 region strongly associate with the development of neurotropism and PML; these probably occur many months or years before PML develops. The mechanism whereby natalizumab contributes to or accelerates the pathogenesis of PML is unknown. Natalizumab undoubtedly reduces immune surveillance of the central nervous system (CNS).13 However, it remains speculative whether or not natalizumab has other effects, for example, mobilisation of bone marrow cells harbouring latent virus.10

      Figure 1

      Key steps in proposed pathogenesis of natalizumab-associated progressive multifocal leucoencephalopathy (PML). The initial event is primary infection with JC virus (JCV). There is limited evidence for a latency programme, but persistent sites of infection include bone marrow, lymphoreticular system, kidney and gut. The central nervous system (CNS) is the likely site of persistent infection. JCV subsequently develops adaptive mutations in the non-coding transcriptional control region and VP1 region, allowing the development of pathogenic neurotropic virus with specificity for glial cells. A further failure of CNS immune surveillance is required before the patient develops overt PML. Natalizumab may lower the threshold for development of PML at various levels. For example, natalizumab leads to a mobilisation of B cells in the peripheral blood, thereby facilitating viral replication. Reducing leucocyte migration across the blood–brain barrier, a known effect of natalizumab, may contribute to the risk of developing PML by inhibiting CNS immune surveillance. NCCR, non-coding control region; PBMC, peripheral blood mononuclear cell.

      Clinical features

      The records of 28 sequential cases of natalizumab-associated PML identify several important clinical features which may allow its early recognition.14 The most common presenting symptoms were cognitive and behavioural changes as well as motor deficits and dysphasia.14 Box 1 summarisesthe clinical features, imaging and cerebrospinal fluid (CSF) findings in natalizumab-associated PML. In all cases it was the clinical symptoms that alerted clinicians to the possible diagnosis. Typically, patients showed multiple neurological features that deteriorated over weeks. Several cases also showed seizures and persistent fever.14 No patient had optic nerve or spinal cord involvement. There was a trend to earlier recognition of PML as more cases occurred, with a median duration from symptom onset to PML diagnosis of 1 month.14 This almost certainly reflects the heightened clinical vigilance to early PML.

      Box 1 Recognition of natalizumab-associated progressive multifocal leucoencephalopathy

      1. Clinical features

      Evolving over weeks

      Frequently multiple focal symptoms

      • Cognitive/behavioural (50%)

      • Motor (37%)

      • Language (31%)

      • Visual (26%)

      • Ataxia (17%)

      Seizures 17%


      Optic nerve and spinal cord syndromes have not been described

      2. MRI appearances

      Hypointense lesions on T1-weighted images, hyperintense on T2 and FLAIR

      Lesions often monofocal and located in the subcortical white matter, frequently involving U-fibres

      30–40% of the lesions showed some degree of gadolinium enhancement at diagnosis

      3. CSF findings

      JCV DNA levels were low (<500 copies/ml) in approximately half of patients

      • CSF, cerebrospinal fluid; FLAIR, fluid attenuation inversion recovery; JCV, JC virus.

      There are several differences between natalizumab-associated PML compared with PML in the contexts of immunodeficiency (HIV) or malignancy. The MRI appearance differ slightly from those in other clinical settings.8 10 14 In particular, 30–40% of lesions showed patchy punctate gadolinium enhancement, making differentiation from MS lesions more difficult than previously anticipated.14 15 CSF analysis in proven cases showed low or zero levels of JCV DNA (<500 copies/ml) in 16/28 patients, highlighting a high false negative rate for CSF JCV PCR. Some PCR results with less sensitive assays were negative but subsequent analysis with quantitative PCR in the National Institutes of Health reference laboratories were positive.14 16 Repeat CSF and, in one case brain biopsy, was sometimes needed for diagnosis.


      The management of natalizumab-associated PML is based upon the principles of:

      1. Drug removal with plasma exchange and/or immunoabsorption.

      2. Identification and treatment of PML associated IRIS.

      3. Adjunctive antiviral treatment.

      All except one patient underwent either plasma exchange or immunoabsorption on diagnosis of PML.14 17 In MS patients treated with natalizumab but without PML, plasma exchange rapidly reduced natalizumab levels with ex-vivo evidence of reconstitution of normal CNS surveillance by leucocytes.2

      PML associated IRIS is a well-recognised clinical syndrome, most commonly encountered with HIV-associated PML after starting highly active antiretroviral therapy (HAART).18 IRIS, characterised by worsening PML symptoms and enlarging or gadolinium-enhancing lesions on MRI, occur in most patients within a few days to weeks of plasma exchange.14 PML-IRIS also developed in a patient who stopped the drug and did not receive plasma exchange, suggesting that it is not unique to the abrupt immune reconstitution caused by drug removal.14 Experts feel that the frequency and severity of IRIS reactions is greater than in other situations where PML is found. Intravenous corticosteroids, often in repeat doses, are typically used to treat IRIS reactions and more recently, as prophylaxis against the development of severe IRIS, which typically occurs with posterior fossa involvement and extensive hemispheric disease.14 18 Corticosteroid treatment is associated with clinical improvement, particularly if there is mass effect, though there are no controlled trials regarding optimal management of natalizumab-associated PML-IRIS.

      Most patients received antiviral agents, although there is no proven effective antiviral agent for PML.14 The two agents most commonly used, mirtazapine and mefloquine, were chosen on the basis of in vitro activity.19 A trial of mefloquine for the treatment of PML recently closed ( identifier NCT00746941).


      Of the 124 patients reported so far, 23 died (19%). Of the survivors, 35% are left with severe neurological impairment (as assessed by Karnofsky Performance Scale, a measure of functional status), 50% with moderate impairment and 15% with mild impairment, with some able to return to work.14 20 It is too early to report the final outcomes of the survivors with confidence. No surviving patients have had recurrent PML, although serial CSF analysis from PML survivors suggests that JCV is not always cleared by immune reconstitution.21

      Factors that do not correlate with survival from PML include JCV DNA load in CSF at PML diagnosis, gender, MS duration, gadolinium enhancement on MRI and previous immunosuppressant therapy, although the latter may predispose to PML.14 Patients with MRI evidence of multilobar involvement have poorer outcomes, particularly if the lesions are in critical areas such as the brainstem. Other adverse factors are older age at diagnosis, higher pre-PML expanded disability status scale, and longer time from first symptoms of PML to diagnosis. PML-IRIS was the cause of neurological deterioration in almost all patients with natalizumab-associated PML.14 In the setting of HIV-associated PML, survival is associated with cytotoxic T lymphocyte responses in the periphery, HAART and the development of IRIS.22

      These outcomes contrast with the near universal fatal outcomes seen with rituximab-associated PML.23 Perhaps the trend towards earlier diagnosis, with the more recent cases of natalizumab-associated PML, has led to better outcomes.14 However, early recognition of PML and minimisation of damage by early reconstitution of immune surveillance are likely to be important to clinical outcome.

      PML risk profiling and monitoring in natalizumab-treated patients

      Key steps in the reduction of risk associated with PML include (1) identifying patients at increased risk of PML, (2) close monitoring of patients at risk of PML and heightened clinical vigilance for evidence of natalizumab-associated PML, (3) promptly stopping natalizumab where PML is suspected (4) obtaining rapid and accurate diagnosis of PML (5) developing strategies for temporary or permanent discontinuation of natalizumab in patients where risk of PML is deemed to be too high.

      The cases identified so far suggest that the following factors are associated with increased PML risk.

      Duration of treatment

      Over the first 2 years, the duration of natalizumab exposure clearly relates to the development of PML (figure 2). PML incidence may peak at 25–36 months, but there are few data beyond 3 years upon which to draw firm conclusions24 (Biogen Medicines Information Service), and only limited clinical experience beyond 5 years.

      Figure 2

      Incidence of progressive multifocal leucoencephalopathy in multiple sclerosis patients treated with natalizumab until May 2011. Data updated from Sandrock et al.24

      Prior cytotoxic immunosuppressant use

      After diagnosis of two early PML cases in the combined β-interferon/natalizumab treatment group of the SENTINEL trial, it was suggested that the combination may predispose to the development of PML. However, there has been no further supportive evidence for this hypothesis from the SENTINEL and GLANCE (Glatiramer Acetate and Natalizumab Combination Evaluation) trials,in which beta-interferon and glatiramer acetate (GA) were co-administered with natalizumab.5

      However the prior use of cytotoxic immunosuppressant agents has emerged as a risk factor for the development of PML.24 Previous immunosuppressive medication use is more common in patients with natalizumab-associated PML than the overall natalizumab-treated MS population, and perhaps confers a fourfold increased risk.24 These immunosuppressants include mitoxantrone, azathioprine, methotrexate, cyclophosphamide, mycophenolate mofetil and cladribine (but not β-interferon, GA or corticosteroids). In total, 46% of patients with PML received an immunosuppressant before receiving natalizumab, most commonly mitoxantrone.24 This risk does not appear to be associated with timing or duration of previous immunosuppressant use.24 These data are consistent with observations extrapolated from PML associated with other monoclonal antibodies, particularly rituximab, where it develops in patients exposed to other immunosuppressant medication.23

      Geographical region also seems to be a risk factor for PML. More cases occurred within the European Economic Area (59 PML cases out of 29 658 patients treated with natalizumab, Biogen medical information service, April 2011) than in the USA (46 PML cases out of 44 200 patients treated). The approximate twofold increase in European risk may reflect the higher use of cytotoxic immunosuppressive medications in Europe.24 However, other epidemiological factors associated with JCV may also contribute to these differences; it is too early to draw firm conclusions about the geographical variation of risk.

      JCV seropositivity

      JCV specific antibodies can be measured using different assays. Gorelik et al recently developed a two-stage optical ELISA using recombinant VP1 protein to detect JCV specific antibodies, combined with a secondary confirmatory ELISA.25 This was evaluated in 800 MS patients treated with natalizumab in the STRATA study (Safety of Tysabri, redosing and treatment).25 Approximately 54% of this natalizumab-treated MS population were seropositive, with a 2% annual seroconversion rate. The ELISA has a false negative rate of 2.5%.25

      Gorelik et al tested blood from 17 patients who subsequently developed PML. All were seropositive when tested 16–180 months from PML onset,25 suggesting that seropositivity at baseline is a significant predictive risk factor for PML development; further validation of this is underway. The HIV literature supports this observation, with baseline seronegative cases of PML being rare. These few seronegative patients may have either lost their antibody response or were unable to mount an antibody response due to severe immunosuppression that typically occurs with HIV-AIDS-related PML. This is unlikely to occur in natalizumab-treated patients as they are not as systemically immunosuppressed and appear to have a preserved immune response to JCV as witnessed by the universal occurrence of IRIS in reported cases.14 The report of three recent seronegative/borderline seropositive natalizumab-associated PML cases is difficult to interpret, as serum for antibody testing was taken during plasma exchange and when retested in the two-step assay were found to be positive.21 26 The implications of the results, needing validation in prospective studies, is that JCV seronegative patients are at very low risk of developing PML but the risk is increased in seropositive patients. Interestingly, a rising JCV antibody titres is associated with natalizumab-associated PML, raising the possibility of monitoring antibody titres to assess PML risk.27 This warrants further prospective evaluation and hopefully an answer will emerge from the ongoing longitudinal studies.

      Blood, urine and CSF JCV PCR as predictors of PML

      There is no convincing evidence that monitoring JCV PCR in either blood (plasma or peripheral blood mononuclear cells), urine or CSF provides useful information about the risk of developing PML. One study suggested that subclinical reactivation of JCV occurs frequently in natalizumab-treated patients and that JCV shedding could be identified in urine JCV DNA samples.28 This has not been replicated in larger studies.29,,31 JCV DNA has also been detected by PCR in the CSF of 8/200 MS patients receiving natalizumab. Subsequent PCR was negative after stopping natalizumab in these 8 patients; none developed PML.32 However, the clinical relevance of positive JCV PCR results is unclear and may well be a normal finding in MS patients not receiving natalizumab.33 34 These observations raise the possibility that the CNS may be a nidus for persistent JCV infection and suggest that persistent JCV infection of the CNS may provide the environment for the selection of the neurotropic mutants which cause PML.

      MRI monitoring

      The current recommendation is that all patients undergoing natalizumab therapy have a baseline MRI scan within 3 months before starting treatment, and annually thereafter while on therapy. Several patients who developed PML had MRI scans in the months before symptom onset but retrospective evaluation of these scans did not show PML.14 This suggests that yearly MRI scans are unlikely to detect PML in the presymptomatic phase, but rather fulfil an important role in disease monitoring; a baseline study therefore allows one to assess whether any new lesions are suspicious of PML or another disease process. The guidelines recommend standardised protocols including gadolinium enhancement and fluid attenuation inversion recovery (FLAIR),15 but this is not essential. Given the potential difficulties in differentiating between different white matter pathologies, it is best if experienced neuroradiologists review the MRI scans.

      Clinical vigilance

      The cornerstone of the PML risk management strategy remains clinical vigilance for the clinical features of PML (box 1). Clinicians need a high index of suspicion for events that do not behave like typical MS relapses. The aim of this approach is to identify and treat PML as early as possible to limit progressive CNS damage. It is critical to assess all neurological events while on natalizumab carefully, followed by appropriate MRI and CSF analysis—repeated where necessary—and for experienced medical and nursing teams who know the patient well to review the patient regularly. It is important to have heightened clinical vigilance, not just for PML but for other potential infectious or neoplastic complications that may occur in the setting of decreased CNS immune surveillance. For example, there are reports of herpes simplex and zoster infections in patients receiving natalizumab therapy; such infections may present atypically. For example, herpes simplex virus and other viral encephalitides may present subacutely as a pseudotumour, similar to its presentation in advanced HIV 1-AIDS.35 CNS lymphoma occurred in three patients on natalizumab, although two of these were likely to be pre-existing and the third case was EBV-negative, which is unusual for isolated CNS lymphomas in the context of immune suppression.36 Like the reported cases of melanoma in natalizumab-treated patients, it is not clear that these events were attributable to natalizumab therapy or simply represented independent comorbidity.

      Issues to consider if natalizumab is paused or discontinued

      Rationale for temporary or permanent cessation of dosing

      Patients and doctors may wish to stop natalizumab temporarily or permanently for various reasons including concerns about PML, planned pregnancy, lack of efficacy and difficulty tolerating the drug. The decision to stop natalizumab needs careful discussion with the patient and counselling regarding the risks of either continuing or stopping natalizumab.

      Disease recurrence on stopping natalizumab

      Pharmacokinetic and pharmacodynamic data demonstrate that plasma exchange accelerates natalizumab clearance from patients with MS who do not have PML.2 Plasma exchange reduces serum natalizumab concentrations by 92% after three plasma exchange sessions. At natalizumab concentrations below 1 µg/ml, desaturation of α4 integrin occurs and peripheral blood mononuclear cell transmigratory capacity, assessed using an in vitro blood–brain barrier, increases.2 Stopping natalizumab may allow a return of MS disease activity to a level similar to that before treatment. Small observational studies have reported variable results, some suggesting little return of disease activity, based on clinical, MRI and immunological criteria.37 Others suggest that patients rapidly return to prenatalizumab relapse rates, possibility even showing excess ‘rebound’ activity.27 The number of gadolinium-enhancing lesions on MRI 3–6 months after stopping therapy in the MS 231 and AFFIRM trials was similar to pretreatment levels.1 4 Higher ‘rebound’ MS activity occurred in the subgroups with higher pretreatment disease activity. This may help clinicians plan further disease modifying therapy after stopping natalizumab.

      Some clinicians suggest ‘drug holidays’ (temporarily stopping natalizumab or increasing the dosing interval, eg, to every 2 months), allowing time for CNS immune reconstitution, and so reducing the risk from opportunistic infections.38 There is no evidence to guide this, although small prospective studies have suggested that temporary dosage suspensions might, as predicted, lead to re-emergence of MS activity.27 Preliminary results from a randomised natalizumab interruption study (RESTORE) showed clinical and radiological recurrence of disease activity following natalizumab dosage suspension.39 Treatment with β-interferon, GA or methylprednisolone following treatment interruption did not seem to prevent this recurrence, though numbers in each group were small.21 Given the lack of evidence, we cannot recommend drug holidays or altered dosing schedules except in the context of a clinical study.

      It is possible that presymptomatic PML will become symptomatic on stopping natalizumab, with the development of PML-IRIS; this occurred in HIV-associated PML after initiating HAART.18 40 The similarity between PML-IRIS and MS relapse makes MRI scanning mandatory in the event of new neurological events during the discontinuation period and it may also be necessary to analyse CSF to exclude PML. Seizures in this context, being relatively rare in MS relapse, should increase the suspicion of PML-IRIS.14 Gadolinium-enhanced MRI should help distinguish MS disease activity from PML-IRIS.8 14

      Switching to alternative immunotherapy

      If, after stopping natalizumab, it is decided not to restart it, then the clinician needs to consider alternative disease modifying agents. Pretreatment disease activity appears to predict disease activity on stopping natalizumab and so this may help to guide the decision. Potential therapeutic options in this scenario were discussed at recent meetings in the USA and Europe41 with a consensus that the choice of subsequent drug would depend on the individual patient's treatment history, previous disease activity and clinical and imaging markers of disease activity 3–4 months after stopping natalizumab.

      When considering switching to β-interferon or GA it is worth noting that there are no significant interactions between either of these drugs and natalizumab, based on data collected from the SENTINEL (natalizumab plus β-interferon) and GLANCE (natalizumab plus GA) trials. Therefore, there is no reason why treatment should not start soon after stopping natalizumab. However, many patients treated with natalizumab have experienced treatment failure with β-interferon and GA and alternative disease modifying agents may need to be considered. Given the potential for a cumulative PML risk with the combination of cytotoxic immunosuppressant drugs and natalizumab, a washout period of at least 3–6 months seems prudent to try to minimise the risk of ‘carry-over’ PML risk.41 Clinical and radiological assessment during this period may also help to decide the activity of the patient's MS.

      It is worth noting that while CNS immune reconstitution rapidly follows removal of circulating natalizumab by plasma exchange, the mechanisms of action of other immunosuppressive medications such as mitoxantrone and the emerging therapies cladribine, alemtuzumab, daclizumab and anti-CD20 (rituximab, ocrelizumab and ofatumumab), may not be so amenable to rapid reversal.

      An algorithm for risk profiling and management

      Based on the studies reviewed above and the clinical experience with natalizumab and the management of PML, we propose a risk profiling and monitoring algorithm (figure 3). The algorithm was designed for subjects already on natalizumab therapy and also those yet to start treatment. We envisage clinicians and patients using this algorithm as an aid to decide on whether or not to start natalizumab and as a guide the duration of therapy. This algorithm will undoubtedly be refined as further safety data emerge. The EMA recommended that the treatment preferably occurs in the context of a registration system, so that data can be further monitored. A UK national register is currently being designed for this purpose. Such national registers will provide information and allow appropriate alterations to the algorithm as observational data emerge.

      Figure 3

      Algorithm for natalizumab PML risk profiling and monitoring. See text for details. CNS, central nervous system; CSF, cerebrospinal fluid; EDSS, expanded disability status scale; GA, glatiramer acetate; IRIS, immune reconstitution inflammatory syndrome; MS, multiple sclerosis; PML, progressive multifocal leucoencephalopathy.

      1. The first step of the algorithm is the defining of the PML risk, based upon current known risk factors and JCV serology status. At present, UK clinicians use the JCV serology assay provided by the Health Protection Agency, Colindale. This was developed and validated for general clinical use and has not specifically been studied in the context of MS. It is likely that the two-stage ELISA developed by Biogen-Idec, and which has been more extensively validated in MS populations with the specific question of PML risk stratification, will be available to clinicians later this year.11 We propose that when patients at initial screening should also undergo urine analysis for JCV DNA using PCR, since there is a false negative rate of at least 2.5% associated with the antibody testing in both assays.11 JCV seronegative patients should undergo annual repeat JCV serology testing. The estimated adult seroconversion rate, based on published data, is 0.5–2% annually.11

        The risk of PML appears to be very low (<1 in 10,000) in patients who are JCV seronegative.24 26 Patients who have had over two years exposure to natalizumab and are JCV seropositive without prior immunosuppression are at intermediate risk of developing PML (∼3 in 1,000). Patients at greatest risk of developing PML (∼1 in 100) are JCV seropositive, have received previous immunosuppressive treatment and have been exposed to natalizumab for over 2 years.15 24

      2. The second step is the heightened clinical vigilance for natalizumab-associated PML, based on clinical features (box 1). On suspicion of PML, natalizumab should be stopped, and MRI and CSF examination undertaken. If these do not confirm PML but the clinical history is progressive and remains suggestive, then these investigations should be repeated and there should be discussions with the virology laboratory about repeat testing at a central reference laboratory. If PML is diagnosed, plasma exchange should start immediately to remove the drug. Five plasma exchanges of 1.5–2.0 plasma volumes over 5–10 days have been successful in case series. Based on observations to date, you should expect patients to develop PML-IRIS 7–14 days after the last plasma exchange.14 There is insufficient evidence to recommend routine prophylactic corticosteroids to prevent the complications of PML-IRIS, although corticosteroids are used as PML-IRIS arises.18 However, based on an analysis of published cases14 we would consider starting prophylactic corticosteroids in patients with brainstem involvement and/or large hemispheric lesions; the aim being to reduce the likelihood of potentially life threatening complications of oedema and mass effect in these situations.

      3. The third step applies to most patients being monitored and remaining well with no evidence of PML. After 2 years, they should be re-consented to continue treatment, as recommended by the EMA, and should understand that are entering a period of higher risk for PML. Annual MRI scans in patients at risk of PML may help to identify suspicious lesions, but more importantly may provide an up-to-date baseline scan should there be atypical neurological events. Patients who are at higher risk of PML should be counselled about the risks of ongoing treatment. Clinicians should discuss the options of temporarily or permanently stopping natalizumab, weighing the risks of continued treatment against the risk of return of MS activity, and also the risks and benefits of alternative treatment strategies.

      Gadolinium-enhanced MRI at 12–14 weeks, to assess recrudescence of MS disease activity and to look for PML-IRIS, is discretionary. However, we would strongly recommend MRI if patients develop new features or before patients switch to one of the ‘high-risk’ therapies discussed above. Careful neurological assessment of patients in the months after stopping natalizumab is important. New neurological events during this period of CNS immune reconstitution are likely to represent MS reactivation, but clinicians should also consider IRIS (usually against PML but may conceivably occur against other opportunistic infections). Assessment of MS activity using clinical and MRI factors during the months after stopping natalizumab may help to guide further treatment decisions, which include restarting natalizumab or switching to other immunotherapies of varying intensity.


      We propose an algorithm for risk management of natalizumab-associated PML. This is based upon the following key points:

      1. In the post-marketing setting, the overall risk of PML approaches 2 in 1000, with 124 reported cases out of 83 300 treated patients.

      2. Natalizumab-associated PML presents with cognitive impairment and other focal deficits progressing over weeks. There are occasionally seizures and fever. MRI often shows gadolinium enhancement but JCV PCR in CSF may be negative. Therefore, clinicians must carefully assess MRI and CSF and these may need repeating.

      3. Risk factors for developing natalizumab-associated PML include JCV seropositivity, previous immunosuppressant use and duration of treatment, peaking after 2 years of therapy.

      4. Patients on natalizumab may be risk stratified based on JCV serology and prior history. High-risk patients should undergo annual MRI scanning. Clinicians must be vigilant for PML in all patients to facilitate early diagnosis and treatment of PML by removal of circulating natalizumab.

      5. Patients who develop natalizumab-associated PML should undergo urgent plasma exchange to allow immune reconstitution of the CNS. An immune reconstitution syndrome against PML (PML-IRIS) should be anticipated in these patients.

      6. Re-emergence of MS disease activity should be anticipated 3–6 months after stopping natalizumab and this may guide the need for further disease modifying therapies.


      View Abstract


      • Please see the end of the article for the list of collaborators and their affliations.

      • Competing interests GG has received personal compensation for participating on Advisory Boards in relation to clinical trial design, trial steering committees and data and safety monitoring committees from Bayer-Schering Healthcare, Biogen-Idec, Eisai, Elan, Fiveprime, Genzyme, Genentech, GSK, Ironwood, Merck-Serono, Novartis, Pfizer, Roche, Sanofi-Aventis, Synthon BV, Teva, UCB Pharma and Vertex Pharmaceuticals. DPJH has performed consultancy work for Biogen-Idec with all fees donated to local neurological research charities.

      • Provenance and peer review Not commissioned; externally peer reviewed. This paper was reviewed by Neil Scolding, Bristol, UK

      Other content recommended for you