Western Blot or Similar Assay
Several pathogenic antibodies are commonly present in the serum of patients with autoimmune disease. In these cases, and in cases of low-titer antibodies, confirmatory testing may be performed by Western blot. Western blot is best suited for detecting antibodies that bind to cytosolic or nuclear antigens. The substrate for Western blot is generated from neural tissue solubilized in detergent. The proteins are denatured; coated with a negative charge; and separated electrophoretically by size, charge, and isoelectric point. The proteins are then transferred to nitrocellulose membranes or similar and incubated with preadsorbed serum or CSF to allow the neural antibody to bind to the epitopes. These are then detected with an antihuman secondary antibody.
Ion channel antibodies are typically assessed using a radioimmunoprecipitation assay.9 This indirectly quantifies a pathogenic antibody using radioactive iodine–labeled antigen, which binds the pathogenic antibody and is subsequently precipitated from solution by an antihuman IgG. Quantification of the radioactivity in the sediment allows for a semiquantitative analysis. Small protein toxins from venomous animals, such as snakes, frogs, or snails, bind to many neural antibodies of interest. Thus labeling these toxins with radioactive iodine and mixing them with extracts of brain, muscle, or other tissue of interest is a useful way of generating antibody targets. Note that generating targets by this method may cause misidentification of the target antigen. In the case of voltage-gated potassium channel (VGKC) antibodies, the target was subsequently found to be other proteins that remained complexed with the VGKC.10,11 Positive VGKC results at a low titer outside of the correct clinical context are often of questionable clinical significance.
Cell-based assays offer improved specificity over the previously discussed assays. The target antigen is natively expressed in mammalian cells present on a microscopy slide, and binding of a pathogenic antibody is detected using an antihuman secondary antibody. As with tissue immunohistochemistry techniques, a trained evaluator is required. The technique is limited by the presence of multiple isoforms of a particular antigenic target (eg, M1 and M23 isoforms of aquaporin-4) and the fact that cells are permeabilized, allowing binding to the cytosolic component of the target antigen. The subjective interpretation of assays with titers near the cutoff has led to the development of semiautomated, quantitative flow cytometric techniques. These have the additional benefit of using live cells, in which the cytosolic component of the target antigen is not available for antibody binding, thus increasing the specificity of the technique.
Enzyme-linked Immunosorbent Assay
The enzyme-linked immunosorbent assay (ELISA) technique consists of incubating patient serum or CSF with purified target attached to the walls of a plate well. After washing, the presence of antibody is detected with an antihuman secondary antibody linked to alkaline phosphatase or horseradish peroxidase. Antibody titers are inferred by quantitating the color change and comparing this to a standard curve. ELISA is a widely available and rapid test but has some limitations. The most important is the presence of false-positive results in individual sera when they bind not to the target antigen but to the plastic well of the ELISA plate. Empty control wells should be used, but commercially available assays often do not provide this. This may have led to a high rate of false-positive results in ELISA assays for NMO-IgG.12
TESTING SERUM OR CEREBROSPINAL FLUID
Ideally, paired samples of serum and CSF should be tested in patients with suspected autoimmune neurologic disease. In patients with NMO-IgG, serum titers of 1:250 or less are associated with undetectable NMO-IgG in CSF,13 and NMO-IgG is not detectable in CSF if it is not detectable in serum.14 In contrast, patients with NMDA receptor antibody–associated encephalitis can have a negative serum test in over 8% of cases.15 Data are not available for other neural antibodies, but these studies demonstrate that both serum and CSF should be analyzed in most cases (Case 1-1).
A 24-year-old woman presented with headache and cognitive changes. Brain MRI demonstrated right temporal lobe hemorrhagic changes with diffuse T2 signal abnormality. A CSF pleocytosis was found. She was treated with acyclovir for presumed herpes simplex virus (HSV) encephalitis, which was subsequently confirmed on CSF polymerase chain reaction (PCR) testing. Following treatment, she had persistent behavioral changes and cognitive difficulties. She presented to her neurologist 6 months later with worsening cognition and seizures. CSF analysis was normal, and N-methyl-D-aspartate (NMDA) receptor antibody testing was positive. She was treated with IV immunoglobulin (IVIg) and returned to her baseline cognitive function following the initial event.
Comment. The pathophysiology of many autoimmune neurologic diseases remains elusive. NMDA receptor antibody–associated encephalitis is the most common cause of autoimmune encephalitis. It has recently been demonstrated that HSV type 1 encephalitis can be followed by NMDA receptor encephalitis.16 In these cases, the CSF analysis can be normal, and patients typically respond to immunotherapy. NMDA receptor encephalitis should be considered in patients with a history of HSV encephalitis who present with a clinical worsening.
ASSOCIATIONS BETWEEN NEURAL ANTIBODIES AND MALIGNANCY
Evaluation for malignancy in the case of a suspected autoimmune neurologic condition is typically guided by the antibody detected.
Paraneoplastic Antibodies Can Predict the Presence of a Malignancy
Neural antibodies are sometimes associated with a systemic malignancy, with a neurologic syndrome commonly preceding the diagnosis of malignancy (Case 1-2).17 Paraneoplastic antibodies are more strongly predictive of tumor type than of a particular clinical syndrome.18 The most common malignancy associated with paraneoplastic central nervous system (CNS) syndromes is small cell lung cancer. Other antibody-associated malignancies are outlined in Table 1-1 and Table 1-2.
A 66-year-old right-handed woman with a history of breast cancer 15 years previously developed ataxia, diplopia, and vertigo over 4 weeks, requiring a wheelchair to mobilize. Given her age, history of cancer, and acuity of symptom onset, she was evaluated for recurrent malignancy with a brain MRI; CT of chest, abdomen, and pelvis; and transvaginal pelvic ultrasound, which were all unrevealing. CSF analysis revealed an increase in nucleated cell count. A whole-body fludeoxyglucose positron emission tomography (FDG-PET) scan demonstrated FDG uptake within a complex left ovarian mass. Biopsy of her ovarian mass revealed a high-grade serous carcinoma of fallopian tube origin. The patient was treated with IV methylprednisolone within 5 weeks of symptom onset, resulting in an improvement in gait and the ability to walk with a cane. A paraneoplastic antibody screen subsequently demonstrated Purkinje cell cytoplasmic antibody type 1 (PCA-1, anti-Yo) antibody. Her malignancy was treated with paclitaxel and carboplatin. Despite freedom from subsequent tumor disease, she continued to have periods of neurologic worsening and required treatment with a variety of immunosuppressant agents, culminating with symptom control on cyclophosphamide.
Comment. Early identification and treatment of suspected autoimmune and paraneoplastic conditions is thought to lead to the best chance of a good outcome. Factors that raised the suspicion for a paraneoplastic cause in this case were the patient’s age, history of malignancy, acuity of onset of symptoms, and abnormal CSF findings. Treatment with steroids is often instituted before the results of neural antibody testing are available, serving as both a diagnostic test and a treatment, as a clinical response to steroids suggests a possible autoimmune phenomenon. In patients with antibodies targeting intracellular antigens, such as in this case, the prognosis for complete recovery is often guarded, and patients often require long-term immunosuppression.
Specific Antibody Clusters Can Predict the Presence of a Malignancy
The presence of two or more autoantibodies in an individual patient occurs more frequently than would be predicted by chance.19 Some antibody clusters, when present, should alert the clinician to a high probability of systemic malignancy. For example, muscle acetylcholine receptor (AChR) and striational autoantibodies are associated with tumor in 45% of patients. If a third autoantibody is detected, the cancer frequency is higher.20 Thymoma, in particular, is frequently associated with neurologic syndromes and associated neural antibodies. Antibodies are most commonly directed against muscle AChR but also against ganglionic AChR, voltage-gated Kv1 potassium channel complex, and the AMPA receptor.21 Similarly, P/Q-type and N-type calcium channel antibodies, associated with SOX1 antibodies are associated with small cell lung cancer in over 80% of cases.20 In contrast, detection of neuronal voltage-gated calcium channel autoantibodies has a negative predictive value for the presence of a thymic tumor.22
Evaluation for a Suspected Malignancy
The presence of risk factors for malignancy, such as smoking or a family history, or the presence of a neural antibody with an oncologic association should prompt an evaluation for malignancy. Following a detailed history and clinical examination, CT of the chest, abdomen, and pelvis; mammography; testicular ultrasound; and prostate-specific antigen should be considered. Where neuroblastoma is suspected, chest and abdominal CT or MRI along with urine testing for homovanillic acid metabolites should be performed. Antibodies with a particular specificity for cancer (eg, NMDA receptor antibody and teratoma) may require a more targeted oncologic evaluation. PET imaging increases the diagnostic yield by 20% when all standard evaluations (eg, whole-body CT scan) have been uninformative.23 PET is unable to detect gonadal tumors (ovary or testis), neuroblastoma, or thymoma. MRI has good sensitivity for both ovarian and thymic tumors.
CLASSIFICATION AND NOMENCLATURE
Neurologic diseases related to neural antibodies have been traditionally described using the clinical phenotype. Clinical entities such as stiff person syndrome and limbic encephalitis allow the clinician to conceptualize a typical presentation, with the intent of identifying such a clinical entity in the future. As more disease entities have been discovered, it has become apparent that individual neural antibodies are associated with significant heterogeneity in clinical phenotype, with a wide range of overlap in the clinical presentation of different antibodies. Given the failure of clinical and phenomenologic classification systems to predict neural antibody–associated syndromes or underlying malignancy, some experts have advocated a cytologic or molecular classification system to describe these diseases. Patients with myelin oligodendrocyte glycoprotein (MOG) antibody–associated demyelinating disease may fulfill the diagnostic criteria for seronegative NMO. This may not be a clinically useful classification for several reasons. MOG antibody–associated demyelinating disease targets oligodendrocytes in contrast to the predominantly astrocytic pathology in NMO spectrum disorder. Both diseases result in complement activation; however, disease-specific treatments, such as aquaporumab (a monoclonal antibody targeting aquaporin-4 that lacks an Fc receptor, thus not capable of activating complement, that is a proposed treatment for NMO spectrum disorder), would not be expected to be effective in treating MOG antibody–associated disease.24 Terms such as autoimmune aquaporin-4 channelopathy and MOGopathy may be more appropriate. Similarly, the entity of limbic encephalitis has been separated into many phenotypes, primarily based on the antigenic target, which can differ in oncologic association and clinical phenotype even within an apparently homogenous antibody-associated disease.25 Identifying the antigenic target (eg, NMDA receptor, contactin associated protein-like 2 [CASPR2] receptor) in the nomenclature may help to clarify the diagnostic evaluation and treatment. This parallels the move to classify neurodegenerative diseases by their associated proteinopathies.26
Recognition of the importance of so-called supportive cells of the CNS in the pathogenesis of autoimmune diseases has led to a new wave of discovery and potential treatments.
Discovery of the aquaporin-4 water channel, located primarily on astrocytes, as an immune target in NMO spectrum disorder has led to distinction of this disease from multiple sclerosis (MS) and a divergence of therapies. Although initially thought to be a demyelinating disease, the central cellular pathology in this condition relates to astrocyte dysfunction.27 The most recent iteration of the diagnostic criteria emphasizes the importance of detecting NMO-IgG with a sensitive and specific assay in the correct clinical context (optic neuritis, brainstem or area postrema syndrome, myelitis, symptomatic narcolepsy or diencephalic syndrome with an NMO spectrum disorder–typical brain MRI).28 Identification of one of these typical syndromes in association with the detection of NMO-IgG in serum allows one to make the diagnosis of NMO spectrum disorder. The entity of seronegative NMO spectrum disorder requires a more stringent set of criteria to be filled in the absence of NMO-IgG detection. This may be useful in areas where NMO-IgG testing is not readily available with a sensitive assay but should be used with caution. Prevention of NMO spectrum disorder relapses with agents targeting B cells, plasma cells, plasmablasts, complement, and blockade of NMO-IgG binding is currently being evaluated in randomized clinical trials.29–32 Current treatment varies by region but includes oral steroids, azathioprine, mycophenolate mofetil, and rituximab as the most commonly used agents. The duration of treatment is controversial. No evidence exists of disease quiescence after a long duration of disease, such as is present in MS, prompting some experts to advocate lifelong immunosuppression.
Myelin Oligodendrocyte Glycoprotein Autoimmunity
MOG is a component of myelin. Antibodies directed against this glycoprotein have been postulated to be involved in demyelinating-type diseases for decades. Prior testing strategies, including ELISA assays, were of low specificity, rendering the results uninterpretable. Recently, cell-based assays have been found to be much more sensitive and specific for MOG antibodies.33 In addition, the selection of secondary antibody has been found to be important, with antihuman IgG1–specific secondary antibodies effectively distinguishing a distinct group of patients with non-MS CNS demyelinating disorders from patients with MS.33 Several groups have reproduced findings indicating that MOG-specific antibodies are associated with a distinct phenotype of CNS demyelinating disease, including conus-predominant myelitis and bilateral optic neuritis, often occurring simultaneously, associated with “cotton wool” brain lesions with poorly defined margins.34 The long-term outcome for patients with MOG antibody–associated CNS inflammatory disease is uncertain. Relapsing disease has been described, so a medium-term course of immunosuppression (1 to 2 years) could be considered.
Glial Fibrillary Acidic Protein Autoimmunity
Antibodies to the glial fibrillary acidic protein (GFAP)-α isoform have recently been described as a biomarker of a steroid-responsive autoimmune meningoencephalomyelitis.35 Neurologic manifestations are diverse and include headache, transverse myelitis, cognitive decline, optic neuropathy, and cerebellar ataxia that improve with high-dose corticosteroid treatment. Relapses require long-term immunosuppressive therapy. Some patients have associated neoplasms, such as prostate and gastroesophageal adenocarcinomas, myeloma, melanoma, colonic carcinoid, parotid pleomorphic adenoma, and teratoma. CSF is generally inflammatory. Cranial MRI often reveals linear perivascular enhancement oriented radially to the ventricles.
Autoimmunity to neuronal targets results in both neuronal loss and neuronal dysfunction to varying degrees.
Encephalitides, Neuropsychiatric Disorders, and Dementia
Antibodies directed against targets at or near the NMDA receptor account for the second most common form of autoimmune encephalitis after acute disseminated encephalomyelitis (ADEM)36 and over half of undiagnosed encephalitides previously thought to be viral in nature.37
The clinical phenotype is that of initial agitation with subsequent catatonia and seizures, memory loss, decreased level of consciousness, central hypoventilation, and characteristic orofacial and “piano-playing” dyskinesias. Although the original description of this entity was in women with teratomas, the disease has been found in all age groups, and approximately 40% of patients are found to have a neoplasm, predominantly teratoma.38 The glutamatergic NMDA receptor is thought to be the primary antigenic target of NMDA receptor antibodies. Recent evidence suggests that dysregulation of extracellular cross talk between the GluN2-NMDA receptor subtype with the membrane receptor ephrin (EPHB2) causes dispersion of the GluN2-NMDA receptor away from the synapse and reduces synaptic plasticity, possibly accounting for memory loss in affected patients.39 NMDA receptor encephalitis occurring following HSV type 1 encephalitis has been reported, giving a clue to the underlying etiology of the disorder.16 In these cases, the CSF analysis can be normal, and patients typically respond to immunotherapy and not to antiviral therapy. NMDA receptor encephalitis should be considered in patients with a history of HSV encephalitis who present with a clinical worsening (Case 1-1).
Cognitive impairment secondary to an autoimmune cause is typically associated with other features, such as seizures, autonomic dysfunction, movement disorders, or other abnormal clinical findings. A summary of antibody and T-cell–mediated autoimmune causes of neuropsychiatric and dementia presentations and encephalitis are summarized in Table 1-2. Rapid-onset cognitive impairment, in particular if associated with a personal or family history of autoimmunity or abnormal CSF findings, should prompt consideration of an autoimmune cause. Any treatment of cognitive impairment with immunotherapy should be accompanied by careful objective documentation of cognitive deficits before embarking on an immunotherapy trial to allow an objective demonstration of any treatment response. While some causes of autoimmune cognitive impairment, such as NMDA receptor antibody–associated encephalitis, may require medium- to long-term immunosuppression, other presumed autoimmune causes of cognitive impairment are monophasic. In the absence of a definitively identified antibody, careful monitoring after withdrawal of an initial therapeutic trial of immunosuppressive medication may be warranted.
An autoimmune basis for seizures has long been recognized and is typically associated with other neurologic symptoms, such as in the case of NMDA receptor antibody–associated encephalitis. More recently appreciated is the presence of seizures in the absence of other neurologic symptoms from an autoimmune etiology. Isolated mesial temporal sclerosis has been demonstrated in patients with autoimmune epilepsy.40 Features that should prompt the clinician to consider an autoimmune cause for seizures include a new-onset seizure disorder with frequent events; new-onset refractory status epilepticus (NORSE); multiple event types in one individual; antiepileptic drug (AED) treatment resistance; CSF abnormalities; and a history of malignancy, smoking, or autoimmune disease. Note that CSF abnormalities are not invariable in autoimmune conditions, so the presence of a normal CSF should not dissuade the clinician from considering autoimmune causes. Both intracellular and extracellular antigenic targets have been described in patients with autoimmune epilepsy, indicating diverse mechanisms of epileptogenesis. No trials have been performed comparing immunotherapy alone versus immunotherapy and AEDs in patients with autoimmune epilepsy. Pragmatically, patients with autoimmune epilepsy tend to be resistant to AED treatment so are on several agents by the time a therapeutic trial of immunotherapy has been administered.
Historically, chorea has been described following streptococcal infection (Sydenham chorea), in pregnancy, and in patients with serologic markers of antiphospholipid syndrome, suggesting an immune-mediated cause in these cases. More recently, antibody associations have been described in almost all types of movement disorders. Some, but not all, of these syndromes are associated with malignancy. In these cases, the antibody detected is more strongly predictive of the presence of malignancy than of a particular clinical phenotype. Variability in the clinical presentation is common. Features that should prompt the clinician to consider an autoimmune cause for a movement disorder include a subacute onset and a widespread distribution of symptoms and signs, including involvement of the trunk and head as well as extremities. Early treatment with immunotherapy can halt or reverse the progression of disease. Movement disorders associated with intracellular antigens, such as PCA-1–associated cerebellar ataxia or antineuronal nuclear antibody type 2 (ANNA-2)–associated dystonia, tend to be associated with significant disability unless early immunotherapy is initiated and malignancy identified and treated.
Stiff person syndrome is a condition of limb and paraspinal muscular rigidity characterized by hyperexcitability of the brainstem and spinal motor neurons.41 Limited forms of the condition have been described, including stiff limb and stiff trunk. On the other end of the spectrum, rigidity can be accompanied by other neurologic features such as cerebellar ataxia and seizures. A rapidly progressive syndrome of progressive encephalomyelitis with rigidity and myoclonus (PERM) has been described, which also forms part of this spectrum.42 Glutamic acid decarboxylase 65 (GAD65) antibodies are the most commonly encountered neural antibody in these conditions. Note that these are also detected in patients with type 1 diabetes mellitus at lower titers. Serum GAD65 titers of 20 nmol/L or higher are more commonly associated with an immunotherapy-responsive neurologic condition. Other antibodies that are associated with stiff person syndrome include amphiphysin (associated with small cell lung cancer, breast adenocarcinoma, and melanoma) glycine receptor, and gephyrin antibodies (one case reported). Patients with glycine receptor antibodies may have a more robust response to immunotherapy.43
Diencephalic and Brainstem Disorders
Well before neural antibodies were associated with sleep disruption, the tight linkage of human leukocyte antigen (HLA)-DR2 to narcolepsy suggested an autoimmune basis to this disorder.44 Subsequently, several neural antibodies have been identified with characteristic sleep abnormalities. The key clinical features associated with the most commonly encountered neural antibodies in sleep disruption are summarized in Table 1-4.45
VGKC complex antibodies include CASPR2 and LGI1 and are associated with profound insomnia, nocturnal agitation, and dream reenactment (agrypnia excitata).46 The phenotypic variability of VGKC complex antibody–associated syndromes is still being elucidated. Even antibody subtypes such as CASPR2 appear to target a number of different antigens, giving rise to the phenotypes of neuromyotonia, Morvan syndrome, and limbic encephalitis.25
Aquaporin-4 is distributed throughout the brain but is localized in the periventricular regions, including the floor of the fourth ventricle.47 Brainstem and diencephalic symptoms of hiccups, vomiting, and symptomatic narcolepsy, have been included as core clinical characteristics of the disease in the updated diagnostic criteria for NMO spectrum disorder, allowing for the diagnosis of an NMO spectrum disorder in the presence of one of these symptoms and NMO-IgG antibody.28 In cases of NMO spectrum disorder–associated narcolepsy, bilateral hypothalamic involvement is typical, and no cases of NMO spectrum disorder–associated narcolepsy have been described with associated cataplexy. Other signs of hypothalamic dysfunction are commonly associated, including hypothermia, dysautonomia, and the syndrome of inappropriate secretion of antidiuretic hormone (SIADH).
NMDA receptor antibody–associated encephalitis can be associated with severe insomnia in the early excitatory phase of the disease. Central hypoventilation is present in two-thirds of patients, particularly in patients later in the disease process.48 Even after apparent clinical recovery, 27% of patients continue to have significant sleep disturbance.48
IgLON5 antibodies have been demonstrated in a small number of patients with dysarthria, dysphagia, dysautonomia, gait ataxia, ocular motility abnormalities, and chorea. Obstructive sleep apnea, nocturnal stridor, and abnormal motor behaviors of sleep were key features.49 Curiously, autopsy findings in two patients demonstrated neuronal hyperphosphorylated tau protein in the hypothalamus and brainstem tegmentum. It is unclear whether this condition represents a neurodegenerative proteinopathy associated with nonpathogenic neural antibodies or a neuroinflammatory condition with secondary neurodegeneration.
Hypersomnolence and cataplexy have been reported in association with Ma antibody–associated narcolepsy. Hypothalamic endocrine dysfunction, seizures, and supranuclear gaze palsies have also been described.50 Multiple sleep latency tests show findings typical of narcolepsy, with reduced sleep latencies and multiple sleep-onset rapid eye movement (REM) periods. Seropositivity for both Ma1 and Ma2 antibodies is associated with carcinoma of the lung, gastrointestinal tract, breast, salivary glands, and ovary; non-Hodgkin lymphoma; germ cell tumors; renal rhabdoid tumors; and melanoma. Seropositivity for the Ma2 antibody only is associated with testicular cancer, underscoring the need to perform a testicular examination in men with a diencephalic clinical presentation.51 The nomenclature of Ma1/Ma2 can be confusing. Ma1 and Ma2 are distinct proteins. The entity previously described as Ma antibody described detection of both Ma1 and Ma2 antibodies. Ta antibody described detection of Ma2 alone.
ANNA-2 (anti-Ri) antibody is characteristically associated with opsoclonus-myoclonus syndrome but is also found in association with multifocal brainstem, cerebellar, and spinal cord dysfunction. Stridor is a characteristic feature, occurring frequently during sleep. Jaw-opening dystonia and laryngospasm can occur, sometimes requiring tracheostomy to manage.52 Carcinoma of the breast, lung, and uterine cervix have been described in adult patients in association with ANNA-2. Neuroblastoma is more common in children. Other antibodies associated with opsoclonus-myoclonus syndrome include glycine receptor and human natural killer 1 (HNK-1) antibodies, both associated with lung malignancy.53
The differential diagnosis of myelopathies can be a challenging clinical dilemma. This section focuses specifically on autoimmune myelopathies. Similar to all autoimmune diseases, autoimmune myelopathies are more common in women. NMO spectrum disorder is more common in patients of African, Native American, and Hispanic descent. Paraneoplastic myelopathies are more common in the elderly. MS is more common among Caucasians, and a family history of the disease increases the risk of MS. Similarly, a family history of autoimmune disease, such as autoimmune thyroid disease, lupus, or rheumatoid arthritis, may suggest a predisposition toward NMO spectrum disorder or other antibody-mediated myelopathies.
The clinical course can yield clues to the differential diagnosis, with typical transverse myelitis being of subacute onset over days to weeks and conditions such as neurosarcoidosis and paraneoplastic myelopathies having a progressive course from onset.
Neuroimaging is essential in the differential diagnosis of myelopathies. A schematic of the different patterns of spinal cord abnormalities is shown in Figure 1-354 and Figure 1-4. A common pitfall is to mistake multiple contiguous short high-signal lesions on T2-weighted imaging for a longitudinally extensive lesion. Careful examination of axial T2-weighted images will typically show multiple discrete eccentrically located lesions in the case of MS in contrast to a single long centrally located lesion in NMO or tract-specific abnormalities in paraneoplastic myelopathies. Differentiation from non–immune-mediated causes of myelopathy is equally important, particularly given that compressive and vascular causes are often amenable to surgical correction. Cord enhancement has been described in cases of spondylosis with compressive myelopathy, with the enhancement typically present at or below the site of maximum compression.55 Dural arteriovenous fistula should also be considered in the case of thoracolumbar myelopathy.56
PRINCIPLES OF TREATMENT
Although several trials are under way in the treatment of NMO, no large randomized controlled trials have been performed in patients with the majority of the conditions discussed here. The goal of initial treatment is to determine the maximum response that can be obtained with immunotherapy. Immunotherapy serves both as an initial treatment and a diagnostic test. Patients who have no response to immunotherapy, in the absence of an antibody known to require long-term treatment (such as NMO-IgG or NMDA receptor IgG), should prompt reevaluation for alternative etiologies (Case 1-3). Treatment typically consists of IV methylprednisolone or IV immunoglobulin (IVIg) in patients who may not tolerate steroids. Plasma exchange can be used in patients with treatment-refractory disease or in patients with contraindications to other treatments. The intensity and duration of the initial phase of treatment varies depending on the clinical syndrome and severity of illness. Patients who are clinically unstable, such as those requiring intensive care unit support, may require a rapid escalation of treatment strategies. In the outpatient setting, patients are typically treated with 5 days of IV methylprednisolone 1 g/d or IVIg 0.4 g/kg/d, followed by once weekly methylprednisolone 1 g or onceweekly IVIg 0.4 g/kg, for 6 to 12weeks (Case 1-4). Where possible, objective measures of disability and treatment response should be obtained before and after treatment (eg, EEG, MRI, CSF, video). Clinical response is a more important outcome measure than change in antibody titers. Note that in patients treated with IVIg, false-positive antibody results can be seen due to the transfused immunoglobulin.
Once reversibility of the clinical syndrome has been established by objective improvements following an initial treatment trial and a maximal response is deemed to have been achieved, long-term management should be addressed. The duration of long-term treatment should be tailored based on the clinical syndrome and associated neural antibodies. Autoimmune encephalitis may be monophasic in nature, such as in the case of NMDA receptor encephalitis or LGI1 antibody encephalitis. Conversely, NMO is known to relapse even after 40years of disease quiescence. In the authors’ practice, patients who require long-term treatment are prescribed oral or IV steroids for 6 months while a steroid-sparing agent such as azathioprine or mycophenolate mofetil is initiated. When evaluating treatment response, medication regimen adherence is important to consider. Treatment with azathioprine is typically associated with a rise in mean corpuscular volume of 5 or more femtoliters from the pretreatment value.58 Detecting such a rise is consistent with medication regimen adherence. Serum mycophenolic acid, the active metabolite of mycophenolate mofetil, can be quantified to ensure therapeutic levels. After 6 months, steroids are tapered slowly over a further 6 months, and patients are watched carefully for a symptom relapse. It is important to continue corticosteroid or immunoglobulin treatment for approximately 12weeks after initiation of azathioprine and 8 weeks after initiation of mycophenolate mofetil (Case 1-5). No data exist to guide the duration of long-term immunosuppression. The authors generally start a trial of medication withdrawal after 2 years, with objective monitoring where possible before and after treatment withdrawal.
General precautions should be undertaken for patients on long-term steroid and immunosuppressant use. Apart from the drug-specific monitoring recommendations, the authors recommend that all patients have a tuberculosis test before starting treatment and a baseline chest x-ray. Vaccinations should be up-to-date, including pneumococcal vaccine and seasonal inactivated influenza vaccine, and live vaccines should be avoided for the duration of treatment. Preventive medications for pneumocystis (eg, trimethoprim-sulfamethoxazole, atovaquone, or nebulized pentamidine) should be considered.
A 68-year-old man presented for a second opinion about a diagnosis of stiff person syndrome. He reported progressive pain, stiffness, and gait difficulty associated with cognitive changes. Glutamic acid decarboxylase 65 (GAD65) antibodies were noted to be elevated. He had been treated with steroids, azathioprine, mycophenolate mofetil, IV immunoglobulin (IVIg), and cyclophosphamide. Despite treatment, he did not report an improvement in symptoms. Repeat neural antibody testing demonstrated elevated GAD65 antibodies in the serum (titer 12 nmol/L) but not in CSF. Neuropsychometric testing demonstrated mild cognitive inefficiencies with primarily attentional cognitive deficits. Neurophysiologic testing for agonist and antagonist muscle cocontraction was normal. Immunosuppressive treatments were withdrawn, and a repeat evaluation 6 months later was unchanged.
Comment. Autoimmune syndromes are diagnosed on the basis of a typical clinical syndrome and supportive findings, such as abnormal MRI brain or CSF analysis or the presence of a pathogenic neural antibody. The presence of a neural antibody alone does not constitute a disease and is not a requirement for the diagnosis of an autoimmune neurologic syndrome. Care must be taken when interpreting neural antibody testing results to avoid overdiagnosis of autoimmune neurologic diseases and consequent inappropriate immunosuppression. In this case, the patient presented with symptoms which, although they can be present in stiff person syndrome, were relatively nonspecific. The presence of GAD65 antibodies had led to the diagnosis of stiff person syndrome. GAD65 is commonly elevated in patients with a predisposition to autoimmunity. GAD65 titers of more than 20 nmol/L, detectable antibody in CSF, and the presence of other potentially pathogenic neural antibodies are more likely to be associated with an immunotherapy-responsive neurologic syndrome.57 In patients with a suspected autoimmune neurologic syndrome with no therapeutic response to immunotherapy, the diagnosis should be reevaluated. In some cases, a trial of immunotherapy may be helpful to clarify the diagnosis.
A 66-year-old man presented with subacute cognitive decline, encephalopathy, and a CSF pleocytosis. He returned to normal cognition with a course of IV steroids. His cognition deteriorated again after 1 month, and he was treated with IV immunoglobulin (IVIg) for 5 days and then azathioprine and monthly IVIg. On a return visit, he reported a deterioration in cognition in the week before his IVIg infusion. His mean corpuscular volume (MCV) was noted to have increased by only 3 femtoliters (fL) from his baseline. The dose of azathioprine was increased. At a return visit 6 months later, he no longer reported an end-of-dose phenomenon with IVIg, and his MCV had risen 6 fL from his baseline before azathioprine therapy.
Comment. The half-life of IVIg is 18 to 32 days. The effect of IVIg on autoimmune encephalopathies tends to wear off after 4 to 5 weeks. The presence of an end-of-dose worsening with IVIg suggests that the autoimmune disease continues to be active and require treatment. Patients who have a clinical response when treated with azathioprine tend to have a 5 fL or more elevation in MCV in response to treatment. In this case, the lack of compete clinical response was accompanied by only a 3 fL rise in MCV. The dose of azathioprine was increased, and the patient demonstrated a sustained response to treatment.
A 36-year-old man was diagnosed with N-methyl-D-aspartate (NMDA) receptor encephalitis after an episode of subacute cognitive decline, weight loss, and abnormal hand movements. He was treated with IV immunoglobulin (IVIg) and subsequently with mycophenolate mofetil. On returning for review after starting mycophenolate mofetil, he reported worsening cognition and hallucinations. Serum mycophenolic acid levels were found to be low. He was treated with a brief course of IV steroids and the dose of mycophenolate mofetil was increased, resulting in symptom resolution.
Comment. Mycophenolate mofetil is rapidly hydrolyzed after absorption to its active metabolite, mycophenolic acid. Therapeutic drug monitoring is not routinely recommended in patients treated with mycophenolate mofetil; however, in patients with loss of disease control, mycophenolic acid serum levels are useful to guide treatment toward dose escalation or drug switching.
The identification of autoimmune and paraneoplastic neurologic diseases requires a high degree of clinical suspicion in the setting of subacute-onset symptoms. The clinical response is best if treated early, so a trial of immunosuppression should be considered in the correct clinical context, even in the absence of identified neural antibodies. A favorable response to treatment supports the diagnosis, whereas a lack of treatment response should prompt a reevaluation for alternative etiologies. Comprehensive clinical, neuropsychological, radiologic, electrophysiologic, serologic, and CSF evaluation permits accurate characterization of the neurologic deficits, immunologic abnormalities, and risk for an underlying cancer.
* Unless a high degree of suspicion exists for a single antigenic target in patients presenting with neurologic disorders, such as in neuromyelitis optica, the authors advocate a global screen for a number of potential causative antibodies.
* Indirect tissue immunofluorescence and immunohistochemistry serve as excellent screening tools for the presence of neural antibodies.
* Western blot is best suited for detecting antibodies that bind to cytosolic or nuclear antigens.
* A high rate of false-positive results for neuromyelitis optica IgG exists with use of enzyme-linked immunosorbent assays.
* Ideally, paired samples of serum and CSF should be tested in patients with suspected autoimmune neurologic disease.
* Paraneoplastic antibodies are more strongly predictive of tumor type than of a particular clinical syndrome.
* Some antibody clusters, when present, should alert the clinician to a high probability of systemic malignancy.
* The presence of risk factors for malignancy, such as smoking or a family history, or the presence of a neural antibody with an oncologic association should prompt an evaluation for malignancy.
* Cytologic and molecular classification systems have been proposed to describe antibody-associated diseases.
* The most recent iteration of the diagnostic criteria for neuromyelitis optica spectrum disorder emphasizes the importance of detecting neuromyelitis optica IgG with a sensitive and specific assay in the correct clinical context (optic neuritis, brainstem or area postrema syndrome, myelitis, symptomatic narcolepsy, or diencephalic syndrome with neuromyelitis optica spectrum disorder– typical brain MRI).
* The entity of seronegative neuromyelitis optica spectrum disorder requires a more stringent set of criteria to be filled in the absence of neuromyelitis optica IgG detection.
* Myelin oligodendrocyte glycoprotein–specific antibodies are associated with a distinct phenotype of central nervous system demyelinating disease, including conus-predominant myelitis and bilateral optic neuritis, often occurring simultaneously, associated with “cotton wool” brain lesions with poorly defined margins.
* Antibodies to the glial fibrillary acidic protein-α isoform have recently been described as a biomarker of a steroid-responsive autoimmune meningoencephalomyelitis.
* Antibodies directed against targets at or near the N-methyl-D-aspartate receptor account for the second most common form of autoimmune encephalitis after acute disseminated encephalomyelitis.
* Viral herpes simplex type 1 encephalitis can be followed by N-methyl-D-aspartate receptor encephalitis.
* Rapid-onset cognitive impairment, in particular if associated with a personal or family history of autoimmunity or abnormal CSF findings, should prompt consideration of an autoimmune cause.
* Any treatment of cognitive impairment with immunotherapy should be accompanied by careful objective documentation of cognitive deficits before embarking on an immunotherapy trial to allow an objective demonstration of any treatment response.
* Features that should prompt the clinician to consider an autoimmune cause for seizures include a new-onset seizure disorder with frequent events; new-onset refractory status epilepticus; multiple event types in one individual; antiepileptic drug treatment resistance; CSF abnormalities; and a history of malignancy, smoking, or autoimmune disease.
* CSF abnormalities are not invariable in autoimmune conditions, so the presence of a normal CSF should not dissuade the clinician from considering autoimmune causes.
* Features that should prompt the clinician to consider an autoimmune cause for a movement disorder include a subacute onset and a widespread distribution of symptoms and signs, including involvement of the trunk and head as well as extremities.
* A family history of autoimmune disease, such as autoimmune thyroid disease, lupus, or rheumatoid arthritis, may suggest a predisposition toward neuromyelitis optica spectrum disorder or other antibody-mediated myelopathies.
* The clinical course of a myelopathy can yield clues to the differential diagnosis, with typical transverse myelitis being of subacute onset over days to weeks and conditions such as neurosarcoidosis and paraneoplastic myelopathies having a progressive course from onset.
* The goal of initial treatment of neuromyelitis optica is to determine the maximum response that can be obtained with immunotherapy.
* In patients with a suspected autoimmune neurologic syndrome with no therapeutic response to immunotherapy, the diagnosis should be reevaluated.
* Objective measures of disability and treatment response should be obtained before and after treatment of suspected autoimmune neurologic conditions.
* In patients treated with IVIg, false-positive antibody results can be seen due to the transfused immunoglobulin.
* Patients who have a clinical response when treated with azathioprine tend to have a 5-femtoliter or more elevation in mean corpuscular volume in response to treatment.
* Therapeutic drug monitoring is not routinely recommended in patients treated with mycophenolate mofetil; however, in patients with loss of disease control, mycophenolic acid serum levels are useful to guide treatment toward dose escalation or drug switching.
1. Quek AM, Britton JW, McKeon A, et al. Autoimmune epilepsy: clinical characteristics and response to immunotherapy. Arch Neurol 2012;69(5):582–593. doi:10.1001/archneurol.2011.2985.
2. McKeon A, Pittock SJ. Paraneoplastic encephalomyelopathies: pathology and mechanisms. Acta Neuropathol 2011;122(4):381–400. doi:10.10007/s00401-011-0876-1.
3. Hinson SR, Pittock SJ, Lucchinetti CF, et al. Pathogenic potential of IgG binding to water channel extracellular domain in neuromyelitis optica. Neurology 2007;69(24):2221–2231. doi:10.1212/01.WNL.0000289761.64862.ce.
4. Pittock SJ, Lennon VA, McKeon A, et al. Eculizumab in AQP4-IgG-positive relapsing neuromyelitis optica spectrum disorders: an open-label pilot study. Lancet Neurol 2013;12(6):554–562. doi:10.1016/S1474-4422(13)70076-0.
5. Pittock SJ, Palace J. Paraneoplastic and idiopathic autoimmune neurologic disorders: approach to diagnosis and treatment. Handb Clin Neurol 2016;133:165–183. doi:10.1016/B978-0-444-63432-0.00010-4.
6. McKeon A, Lennon VA, Pittock SJ. Immunotherapy-responsive dementias and encephalopathies. Continuum (Minneap Minn) 2010;16(2 Dementia):80–101. doi:10.1212/01.CON.0000368213.63964.34.
7. Tobin WO, Lennon VA, Komorowski L, et al. DPPX potassium channel antibody: frequency, clinical accompaniments, and outcomes in 20 patients. Neurology 2014;83(20):1797–1803. doi:10.1212/WNL.0000000000000991.
8. Waters P, Pettingill P, Lang B. Detection methods for neural antibodies. Handb Clin Neurol 2016;133:147–163. doi:10.1016/B978-0-444-63432-0.00009-8.
9. Iorio R, Lennon VA. Neural antigen-specific autoimmune disorders. Immunol Rev 2012;248(1):104–121. doi:10.1111/j.1600-065X.2012.01144.x.
10. Irani SR, Alexander S, Waters P, et al. Antibodies to Kv1 potassium channel-complex proteins leucine-rich, glioma inactivated 1 protein and contactin-associated protein-2 in limbic encephalitis, Morvan’s syndrome and acquired neuromyotonia. Brain 2010;133(9):2734–2748. doi:10.1093/brain/awq213.
11. Lai M, Huijbers MG, Lancaster E, et al. Investigation of LGI1 as the antigen in limbic encephalitis previously attributed to potassium channels: a case series. Lancet Neurol 2010;9(8):776–785. doi:10.1016/S1474-4422(10)70137-X.
12. Fryer JP, Lennon VA, Pittock SJ, et al. AQP4 autoantibody assay performance in clinical laboratory service. Neurol Neuroimmunol Neuroinflamm 2014;1(1):e11. doi:10.1212/NXI.0000000000000011.
13. Takahashi T, Fujihara K, Nakashima I, et al. Anti-aquaporin-4 antibody is involved in the pathogenesis of NMO: a study on antibody titre. Brain 2007;130(pt 5):1235–1243. doi:10.1093/brain/awm062.
14. Jarius S, Franciotta D, Paul F, et al. Cerebrospinal fluid antibodies to aquaporin-4 in neuromyelitis optica and related disorders: frequency, origin, and diagnostic relevance. J Neuroinflammation 2010;7:52. doi:10.1186/1742-2094-7-52.
15. Gresa-Arribas N, Titulaer MJ, Torrents A, et al. Antibody titres at diagnosis and during follow-up of anti-NMDA receptor encephalitis: a retrospective study. Lancet Neurol 2014;13(2):167–177. doi:10.1016/S1474-4422(13)70282-5.
16. Hacohen Y, Deiva K, Pettingill P, et al. N-methyl-D-aspartate receptor antibodies in post-herpes simplex virus encephalitis neurological relapse. Mov Disord 2014;29(1):90–96. doi:10.1002/mds.25626.
17. Gultekin SH, Rosenfeld MR, Voltz R, et al. Paraneoplastic limbic encephalitis: neurological symptoms, immunological findings and tumour association in 50 patients. Brain 2000;123(pt 7):1481–1494. doi:10.1093/brain/123.7.1481.
18. Pittock SJ, Kryzer TJ, Lennon VA. Paraneoplastic antibodies coexist and predict cancer, not neurological syndrome. Ann Neurol 2004;56(5):715–719. doi:10.1002/ana.20269.
19. Mackay IR. Clustering and commonalities among autoimmune diseases. J Autoimmun 2009;33(3–4):170–177. doi:10.1016/j.jaut.2009.09.006.
20. Horta ES, Lennon VA, Lachance DH, et al. Neural autoantibody clusters aid diagnosis of cancer. Clin Cancer Res 2014;20(14):3862–3869. doi:10.1158/1078-0432.CCR-14-0652.
21. Zekeridou A, McKeon A, Lennon VA. Frequency of synaptic autoantibody accompaniments and neurological manifestations of thymoma. JAMA Neurol 2016;73(7):853–859. doi:10.1001/jamaneurol.2016.0603.
22. Vernino S, Lennon VA. Autoantibody profiles and neurological correlations of thymoma. Clin Cancer Res 2004;10(21):7270–7275. doi:10.1158/1078-0432.CCR-04-0735.
23. McKeon A, Apiwattanakul M, Lachance DH, et al. Positron emission tomography-computed tomography in paraneoplastic neurologic disorders: systematic analysis and review. Arch Neurol 2010;67(3):322–329. doi:10.1001/archneurol.2009.336.
24. Zamvil SS, Slavin AJ. Does MOG Ig-positive AQP4-seronegative opticospinal inflammatory disease justify a diagnosis of NMO spectrum disorder? Neurol Neuroimmunol Neuroinflamm 2015;2(1):e62. doi:10.1212/NXI.0000000000000062.
25. Joubert B, Saint-Martin M, Noraz N, et al. Characterization of a subtype of autoimmune encephalitis with anti–contactin-associated protein-like 2 antibodies in the cerebrospinal fluid, prominent limbic symptoms, and seizures. JAMA Neurol 2016;73(9):1115–1124. doi:10.1001/jamaneurol.2016.1585.
26. Kovacs GG. Molecular pathological classification of neurodegenerative diseases: turning towards precision medicine. Int J Mol Sci 2016;17(2):189. doi:10.3390/ijms17020189.
27. Howe CL, Kaptzan T, Magan˜a SM, et al. Neuromyelitis optica IgG stimulates an immunological response in rat astrocyte cultures. Glia 2014;62(5):692–708. doi:10.1002/glia.22635.
28. Wingerchuk DM, Banwell B, Bennett JL, et al. International consensus diagnostic criteria for neuromyelitis optica spectrum disorders. Neurology 2015;85(2):177–189. doi:10.1212/WNL.0000000000001729.
33. Waters P, Woodhall M, O’Connor KC, et al. MOG cell-based assay detects non-MS patients with inflammatory neurologic disease. Neurol Neuroimmunol Neuroinflamm 2015;2(3):e89. doi:10.1212/NXI.0000000000000089.
34. Kitley J, Waters P, Woodhall M, et al. Neuromyelitis optica spectrum disorders with aquaporin-4 and myelin-oligodendrocyte glycoprotein antibodies: a comparative study. JAMA Neurol 2014;71(3):276–283. doi:10.1001/jamaneurol.2013.5857.
35. Fang B, McKeon A, Hinson SR, et al. Autoimmune glial fibrillary acidic protein astrocytopathy: a novel meningoencephalomyelitis. JAMA Neurol 2016;73(11):1297–1307. doi:10.1001/jamaneurol.2016.2549.
36. Granerod J, Ambrose HE, Davies NW, et al. Causes of encephalitis and differences in their clinical presentations in England: a multicentre, population-based prospective study. Lancet Infect Dis 2010;10(12):835–844. doi:10.1016/S1473-3099(10)70222-X.
37. Gable MS, Sheriff H, Dalmau J, et al. The frequency of autoimmune NMDA receptor encephalitis surpasses that of individual viral etiologies in young individuals enrolled in the California Encephalitis Project. Clin Infect Dis 2012;54(7):899–904. doi:10.1093/cid/cir1038.
38. Titulaer MJ, McCracken L, Gabilondo I, et al. Treatment and prognostic factors for long-term outcome in patients with anti-NMDA receptor encephalitis: an observational cohort study. Lancet Neurol 2013;12(2):157–165. doi:10.1016/S1474-4422(12)70310-1.
39. Mikasova L, De Rossi P, Bouchet D, et al. Disrupted surface cross-talk between NMDA and Ephrin-B2 receptors in anti-NMDA encephalitis. Brain 2012;135(pt 5):1606–1621. doi:10.1093/brain/aws092.
40. Kotsenas AL, Watson RE, Pittock SJ, et al. MRI findings in autoimmune voltage-gated potassium channel complex encephalitis with seizures: one potential etiology for mesial temporal sclerosis. AJNR Am J Neuroradiol 2014;35(1):84–89. doi:10.3174/ajnr.A3633.
41. Moersch FP, Woltman HW. Progressive fluctuating muscular rigidity and spasm (“stiff-man” syndrome); report of a case and some observations in 13 other cases. Proc Staff Meet Mayo Clin 1956;31(15):421–427.
42. Whiteley AM, Swash M, Urich M. Progressive encephalomyelitis with rigidity: its relation to ‘subacute myoclonic spinal neuronitis’ and to the ‘stiff man syndrome.’ Brain 1976;99(1):27–42. doi:10.1093/brain/99.1.27.
43. McKeon A, Martinez-Hernandez E, Lancaster E, et al. Glycine receptor autoimmune spectrum with stiff-man syndrome phenotype. JAMA Neurol 2013;70(1):44–50. doi:10.1001/jamaneurol.2013.574.
44. Langdon N, Welsh KI, van Dam M, et al. Genetic markers in narcolepsy. Lancet 1984;2(8413):1178–1180. doi:10.1016/S0140-6736(84)92742-9.
45. Silber MH. Autoimmune sleep disorders. Handb Clin Neurol 2016;133:317–326. doi:10.1016/B978-0-444-63432-0.00018-9.
46. Lugaresi E, Provini F. Agrypnia excitata: clinical features and pathophysiological implications. Sleep Med Rev 2001;5(4):313–322. doi:10.1053/smrv.2001.0166.
47. Venero JL, Vizuete ML, Ilundáin AA, et al. Detailed localization of aquaporin-4 messenger RNA in the CNS: preferential expression in periventricular organs. Neuroscience 1999;94(1):239–250. doi:10.1016/S0306-4522(99)00182-7.
48. Dalmau J, Gleichman AJ, Hughes EG, et al. Anti-NMDA-receptor encephalitis: case series and analysis of the effects of antibodies. Lancet Neurol 2008;7(12):1091–1098. doi:10.1016/S1474-4422(08)70224-2.
49. Sabater L, Gaig C, Gelpi E, et al. A novel non-rapid-eye movement and rapid-eye-movement parasomnia with sleep breathing disorder associated with antibodies to IgLON5: a case series, characterisation of the antigen, and post-mortem study. Lancet Neurol 2014;13(6):575–586. doi:10.1016/S1474-4422(14)70051-1.
50. Adams C, McKeon A, Silber MH, Kumar R. Narcolepsy, REM sleep behavior disorder, and supranuclear gaze palsy associated with Ma1 and Ma2 antibodies and tonsillar carcinoma. Arch Neurol 2011;68(4):521–524. doi:10.1001/archneurol.2011.56.
51. Hoffmann LA, Jarius S, Pellkofer HL, et al. Anti-Ma and anti-Ta associated paraneoplastic neurological syndromes: 22 newly diagnosed patients and review of previous cases. J Neurol Neurosurg Psychiatry 2008;79(7):767–773. doi:10.1136/jnnp.2007.118588.
52. Pittock SJ, Lucchinetti CF, Lennon VA. Anti-neuronal nuclear autoantibody type 2: paraneoplastic accompaniments. Ann Neurol 2003;53(5):580–587. doi:10.1002/ana.10518.
53. Armangué T, Sabater L, Torres-Vega E, et al. Clinical and immunological features of opsoclonus-myoclonus syndrome in the era of neuronal cell surface antibodies. JAMA Neurol 2016;73(4):417–424. doi:10.1001/jamaneurol.2015.4607.
54. Flanagan EP. Autoimmune myelopathies. Handb Clin Neurol 2016;133:327–351. doi:10.1016/B978-0-444-63432-0.00019-0.
55. Flanagan EP, Krecke KN, Marsh RW, et al. Specific pattern of gadolinium enhancement in spondylotic myelopathy. Ann Neurol 2014;76(1):54–65. doi:10.1002/ana.24184.
56. Yang HK, Lee JW, Jo SE, et al. MRI findings of spinal arteriovenous fistulas: focusing on localisation of fistulas and differentiation between spinal dural and perimedullary arteriovenous fistulas. Clin Radiol 2016;71(4):381–388. doi:10.1016/j.crad.2016.01.007.
57. Pittock SJ, Yoshikawa H, Ahlskog JE, et al. Glutamic acid decarboxylase autoimmunity with brainstem, extrapyramidal, and spinal cord dysfunction. Mayo Clin Proc 2006;81(9):1207–1214. doi:10.4065/81.9.1207.
58. Constanzi C, Matiello M, Lucchinetti CF, et al. Azathioprine: tolerability, efficacy, and predictors of benefit in neuromyelitis optica. Neurology 2011;77(7):659–666. doi:10.1212/WNL.0b013e31822a2780.
Supplemental Digital Content
© 2017 American Academy of Neurology