Lessons Learned From Fatal Progressive Multifocal Leukoencephalopathy in a Patient With Multiple Sclerosis Treated With Natalizumab

Objective To describe the clinical, radiological, and histopathological features of a fatal case of progressive multifocal leukoencephalopathy (PML) in a patient with multiple sclerosis treated with natalizumab. We will use this case to review PML risk stratification and diagnosis.

Design Case report.

Setting Tertiary referral center hospitalized care.

Patient A 55-year-old, JC virus (JCV) antibody–positive patient with multiple sclerosis who died of PML after receiving 45 infusions of natalizumab.

Main Outcome Measures Brain magnetic resonance imaging and cerebrospinal fluid JCV DNA polymerase chain reaction results.

Results The patient developed subacute onset of bilateral blindness following his 44th dose of natalizumab. Ophthalmologic examination was normal, the brain magnetic resonance imaging was not suggestive of PML, and cerebrospinal fluid analysis did not reveal the presence of JCV DNA. The patient was subsequently treated for a presumed multiple sclerosis relapse with high-dose corticosteroids. Two weeks after his 45th dose of natalizumab, he developed hemiplegia that evolved into quadriparesis. Repeated magnetic resonance imaging and cerebrospinal fluid studies were diagnostic for PML. Postmortem histopathological analysis demonstrated PML-associated white matter and cortical demyelination.

Conclusions The risks and benefits of natalizumab must be reassessed with continued therapy duration. When there is high clinical suspicion for PML in the setting of negative test results, close clinical vigilance is indicated, natalizumab treatment should be suspended, and JCV polymerase chain reaction testing and brain magnetic resonance imaging scans should be repeated.

Progressive multifocal leukoencephalopathy (PML) is a demyelinating infection of the central nervous system caused by JC virus (JCV) and a feared adverse event associated with natalizumab treatment in patients with multiple sclerosis (MS). The availability of the JCV antibody assay and an understanding of PML susceptibility allow more accurate risk assessment for individual patients. Herein, we describe the clinical, radiological, and histopathological features of a fatal case of PML in a patient with MS treated with natalizumab. We will use this case to review PML risk stratification and diagnosis.

A 49-year-old white man was diagnosed with relapsing-remitting MS following symptoms of right-leg monoparesis and a brain magnetic resonance image (MRI) suggestive of white matter (WM) demyelination. Initially, he was treated with subcutaneous high-dose interferon beta-1a, 3 times weekly, for 2 years. He was unable to tolerate self-injections and he switched disease-modifying therapy to natalizumab. While taking interferon beta-1a and later natalizumab, he remained relapse free with low disability (Expanded Disability Status Scale score, 2). He was found to be JCV antibody positive during natalizumab therapy but opted to continue treatment.

In October 2011, after his 44th infusion, he developed blurry vision. He was found to have 20/400 visual acuity bilaterally. Concern for PML was raised. A brain MRI (Figure 1, scan 2) was initially read as unchanged from 1 year earlier by a neuroradiologist. Results of JCV polymerase chain reaction (PCR) testing in the patient's cerebrospinal fluid (CSF) were negative. The patient was evaluated by an ophthalmologist and “no organic cause” for his blurry vision was identified. Presumed to be having an MS attack, he was treated with 5 doses of high-dose intravenous corticosteroids, followed by his 45th dose of natalizumab in November 2011. His vision did not improve and he developed new right hemiparesis 3 weeks after his last infusion.

He was referred for admission to our hospital's neurology service from the community for a presumed steroid-refractory MS relapse. Total plasma exchange was initiated. Following his first exchange, he developed encephalopathy and quadriparesis and was transferred to the medical intensive care unit. The MS service was consulted and repeated testing for PML was recommended. Repeated MRI scans were consistent with PML (Figure 1, scan 3) and CSF JCV DNA PCR results were positive. The patient died 5 days into his hospital stay. Retrospective review of his MRI at symptom onset (Figure 1, scan 2) revealed extension of T2 signal abnormality into the right occipital and parietal lobes consistent with PML.

His brain was obtained at autopsy 3 hours post mortem. Macroscopic and microscopic features of the brain were diagnostic for PML (Figure 2). Several well-demarcated, smooth lesions were observed, consistent with MS plaques.

The brain MRI from September 2010 (Figure 1, scan 1) after onset of bilateral vision changes demonstrated abnormal increased T2 and fluid-attenuated inversion recovery signal involving the corpus callosum (Figure 1, scan 1 [arrow]) and radiating away from the ventricles perpendicularly, consistent with MS. The MRI from October 2011 after onset of bilateral visual symptoms demonstrated extension of diffuse WM abnormalities into the subcortical WM and reaching into cortical gray matter in the right occipital and parietal lobes (Figure 1, scan 2 [arrows]), consistent with PML. There was no contrast enhancement noted. The MRI obtained 1 month later during hospital admission demonstrated extension of diffuse WM abnormalities into the subcortical WM and reaching into cortical gray matter in the right occipital, right frontal, and right parietal lobes (Figure 1, scan 3 [arrows]). Again, there was no contrast enhancement present. These findings were consistent with extension of PML.

In 11 examined tissue blocks derived from the occipital and temporal lobes, demyelinating lesions were widespread in WM and adjacent cortex. Separate isolated patchy areas of demyelinated cortex unconnected to WM demyelination were also observed (Figure 2A). White matter lesions were dominated by infiltration with myelin-laden (foamy) CD68⁺ macrophages (Figure 2B). Myelin-laden macrophages were abundant in cortical lesions (Figure 2C). Immunolabeling with glial fibrillary acidic protein revealed reactive, bizarrely shaped astrocytes within WM and cortical lesions (Figure 2E and F). In contrast, myelin-phagocytosing macrophages and reactive astrocytes are usually absent in cortical MS plaques.^1,2

The JCV capsid protein VP-1, indicative of productive, lytic infection, was present in WM lesions within astrocytes and less frequently in oligodendrocytes (Figure 2H). In cortical demyelinating lesions, VP-1 was observed in the cytoplasm of astrocytes and neurons (Figure 2I).³ In the perimeter of white and cortical gray matter lesions and less frequently in remote WM, VP-1 was detected in oligodendrocytes (Figure 2J). Thus, JCV-induced demyelination was widespread in WM and cortex and was characterized by intense microglia/macrophage–dominated infiltration and astrogliotic reaction. This strong inflammatory reaction in the cortex is usually not seen in demyelinating MS lesions.

This case highlights the importance of using clinical judgment to remain suspicious for PML in the setting of negative CSF test results in patients with MS taking natalizumab. Furthermore, MRI changes from PML can be subtle and may be missed even by an experienced neuroradiologist.

Progressive multifocal leukoencephalopathy is a demyelinating disease of the central nervous system caused by JCV. Its manifestations may be confused with an acute MS relapse. The most common presentations of PML include bilateral cortical blindness, hemiplegia, and encephalopathy. These are not common presentations of MS relapses. The subacute onset of bilateral blindness in this patient taking natalizumab with a normal ophthalmologic examination and unchanged brain MRI scan was highly concerning for cortical blindness secondary to PML. Presentation of simultaneous bilateral optic neuritis in MS is rare and has been reported in 1 cohort to occur in only 0.42% of cases.⁴ Similarly, acute-onset hemiparesis is an uncommon presentation of an MS relapse.⁵ This case provides an important clinical lesson. When the clinical presentation is suspicious for PML and atypical for an MS relapse, a negative initial CSF JCV PCR result does not exclude a PML diagnosis. Magnetic resonance imaging should be carefully reviewed to look for subtle changes concerning for PML. These lesions are most apparent on T2 and fluid-attenuated inversion recovery sequences as hyperintense and on T1 sequences as hypointense mainly within subcortical WM. In such cases, it appears most appropriate to hold natalizumab treatment and repeat testing. Holding natalizumab treatment for less than 3 months generally does not place the patient with MS at risk of increased disease activity.⁶

Additionally, the appropriate use of this therapy and the continued evaluation of the risk vs benefit ratio throughout duration of therapy are essential to caring for this patient population. Because of the risk of PML, natalizumab is often used as second-line therapy. In the present case, disease-modifying therapy was switched to natalizumab because of poor tolerability of interferon beta-1a rather than lack of efficacy, given the absence of relapses. Even though the patient was JCV antibody positive, the risk of developing PML was presumed to be low during the first year. The incidence of PML during the first 12 doses of natalizumab is 0.04 per 1000 patients.⁷ There have been 4 reported cases of PML during the first 12 doses of natalizumab. Because the patient had no history of prior immunosuppression, he carried a risk for PML of less than 1 in 1000 between dose 1 and 24. However, after 24 doses, his risk of PML increased to 2.5 in 1000.⁷ Because the patient's disease was previously well controlled with interferon beta-1a with no further MS relapses and a low Expanded Disability Status Scale score at 5 years, the risk vs benefit ratio for natalizumab treatment could have been considered too high after the 24th dose.

A positive DNA test result in CSF is diagnostic for PML whereas DNA testing in urine or blood is not useful. JC viremia does not necessarily occur concomitantly with PML and has conversely been seen in 1.3% of healthy immunocompetent individuals as well as in 10% to 15% of HIV-positive patients without PML.⁸ Likewise, JCV DNA can be detected in the urine of 18% to 30% of healthy individuals.⁹ Importantly, CSF JCV DNA copy numbers were less than 500/mL in about half of patients diagnosed with PML while taking natalizumab, which is below the detection limit of most standard assays.¹⁰ Substantial variability exists with regard to detection levels and consistency between laboratories even in testing the same sample set. Thus, care should be taken to use a high-sensitivity test (<50 copies/mL).

The CSF samples of the patient discussed herein were sent to Mayo Medical Laboratories, which offers a JCV PCR DNA test for spinal fluid with a sensitivity of approximately 50 copies/mL. Focus Diagnostics also offers testing capable of detecting JCV DNA at a sensitivity of 50 copies/mL. The assay with the best sensitivity is offered as a free diagnostic service by the Laboratory of Molecular Medicine and Neuroscience at the National Institutes of Health (eAppendix), with a sensitivity of 10 copies/mL.¹⁰ More than 50% of natalizumab-related PML cases have been analyzed by this laboratory. European laboratories that use the ultrasensitive real-time quantitative PCR include the University of Düsseldorf (Germany), Karolinska Institute (Sweden), and Erasmus Medical University Centre in Rotterdam (the Netherlands).

The risk vs benefit ratio of natalizumab treatment should be regularly reevaluated in each patient as it changes with duration of therapy. When clinical suspicion for PML is high in the setting of negative test results, it may be reasonable to hold natalizumab treatment and to repeat CSF and MRI testing after several weeks. The brain MRI should be carefully reviewed by an experienced neuroradiologist and by the treating neurologist for subtle changes concerning for PML. The JCV DNA copy number in the CSF may be absent or low in many cases of natalizumab-related PML and appropriate testing with a sensitivity of less than 50 copies/mL is recommended.

Correspondence: David Pitt, MD, Departments of Neurology and Neuroscience, The Ohio State University, Biomedical Research Tower, 460 W 12th Ave, Room 688, Columbus, OH 43210 (david.pitt@osumc.edu).

Accepted for Publication: July 17, 2012.

Published Online: January 21, 2013. doi:10.1001/jamaneurol.2013.1960

Author Contributions:Study concept and design: Boster, Nicholas, Morgan-Followell, and Pitt. Acquisition of data: Boster, Topalli, Kisanuki, Pei, Morgan-Followell, Kirsch, and Pitt. Analysis and interpretation of data: Boster, Nicholas, Kirsch, Racke, and Pitt. Drafting of the manuscript: Boster, Nicholas, Topalli, Pei, Kirsch, and Pitt. Critical revision of the manuscript for important intellectual content: Boster, Nicholas, Topalli, Kisanuki, Morgan-Followell, Kirsch, Racke, and Pitt. Obtained funding: Pitt. Administrative, technical, and material support: Boster, Nicholas, Topalli, and Pei. Study supervision: Boster, Nicholas, Kirsch, Racke, and Pitt.

Funding/Support: This was supported in part by a Sylvia Lawry Physician Fellowship (Dr Nicholas) from the National Multiple Sclerosis Society. Dr Morgan-Followell was supported by a Clinical Care Physician Fellowship Award from the National Multiple Sclerosis Society.

Additional Contributions: We thank Eugene Major, PhD (National Institutes of Health), for helpful discussion and R. Chaudhury, MD (The Ohio State University), for help with the pathological specimen.