Efficacy of tumor necrosis factor α blockade in primary central nervous system vasculitis resistant to immunosuppressive treatment



Primary central nervous system vasculitis (PCNSV) is an uncommon form of vasculitis that is limited to the brain and spinal cord. PCNSV may be associated with serious morbidity and mortality unless it is treated with high-dose corticosteroids and immunosuppressive medications (1–3).

Because of the lack of controlled clinical trials and uniform diagnostic criteria, uncertainties exist regarding the response to treatment and the long-term outcome of PCNSV. Some patients may have a more benign course and may need less intense therapy (4, 5). Others require more aggressive therapy. Still others experience recurrent symptoms or have severe adverse effects, therefore new treatment options are needed (6–9).

We report 2 patients with PCNSV who did not respond to treatment with corticosteroids and immunosuppressive medication. They were treated with tumor necrosis factor (TNF) blockers in an attempt to control the disease. This study was approved by the Mayo Clinic Institutional Review Board.

Report of 2 Cases

Case 1.

A 27-year-old previously healthy woman was admitted with a 2-week history of severe headache and confusion. Results of the physical examination showed bilateral papilledema but were otherwise unremarkable. An emergent magnetic resonance image (MRI) of the brain showed a left frontal and right occipital infarct. Magnetic resonance angiography findings suggested vasculitis; segmental narrowing affected multiple proximal and distal cerebral arteries.

Laboratory test results included a hemoglobin level of 12.8 gm/dl (normal 12.0–15.5 gm/dl), a leukocyte count of 10.7 × 109/liter (normal 3.5–10.5 × 109/liter; 60% neutrophils, 30% lymphocytes, 10% monocytes), and a platelet count of 301 × 109/liter (normal 150–450 × 109/liter). The erythrocyte sedimentation rate (ESR) was 51 mm/hour (normal 0–29 mm/hour; Westergren method). Levels of serum liver enzymes, creatinine, electrolytes, and lipids were normal, and urinalysis findings were normal. Findings from blood coagulation studies, tests for serum antinuclear antibodies, antibodies against double-stranded DNA, antineutrophil cytoplasmic antibodies, and anticardiolipin antibodies, and levels of cryoglobulins, rheumatoid factor, serum lupus anticoagulant, and angiotensin-converting enzyme were negative or normal. In addition, the following tests had normal or negative findings: tuberculin skin test, fluorescent treponemal antibody absorption test, serology test (for Toxoplasma gondii, human immunodeficiency virus, herpes simplex virus, hepatitis B virus, and hepatitis C virus), and a fungal serology survey (tests for antibodies against Coccidioides, Histoplasma, and Blastomyces).

Cerebrospinal fluid (CSF) glucose concentration was normal (67 mg/dl), protein concentration was 15 mg/dl (normal 14–45 mg/dl), erythrocyte cell count was 838/μl, and white blood cell count was 4/μl (70% lymphocytes, 24% polymorphonucleates, 6% monocytes; normal cell count <5 mononuclear cells/ml). A venereal disease research laboratory test and Cryptococcus antigen test showed negative findings. Cytologic findings were normal; polymerase chain reaction assays for herpes simplex and zoster viruses, Epstein-Barr virus, cytomegalovirus, T gondii, and Borrelia burgdorferi showed negative results.

Findings from a chest radiograph, transesophageal echocardiogram, and duplex carotid ultrasonogram were unremarkable. Conventional cerebral angiography was performed on hospital day 2 and showed results highly suggestive of vasculitis; images showed scattered areas of narrowing that involved the large and small arteries with lesions most prominently in the basilar system, bilateral posterior inferior cerebellar arteries, anterior inferior cerebellar arteries, superior cerebellar arteries, and left middle cerebral artery trifurcation proximal to sylvian branches. Additional scattered areas of narrowing occurred in bilateral anterior cerebral arteries, the distal middle cerebral artery, and bilateral posterior cerebral artery territories.

The diagnosis of PCNSV was made on the basis of clinical manifestations and angiographic findings. The patient initially was treated with pulse intravenous methylprednisolone (1 gm/day) and intravenous cyclophosphamide (1.7 gm/month) beginning on hospital day 1. However, her neurologic status continued to deteriorate. On day 3, a new left-sided hemiparesis developed, and she had a marked reduction in visual acuity. A contrast-enhanced MRI of the brain on day 3 showed multiple acute interval infarcts involving both cerebellar hemispheres and both cerebral hemispheres with minimal abnormal enhancement of the lesions. Scattered linear areas of increased fluid-attenuated inversion recovery signal were observed in multiple sulci with minimal enhancement.

Despite treatment with high-dose methylprednisolone, the patient became more confused, aphasic, and blind. Therefore, on day 6 she received a single intravenous infusion of the anti-TNF monoclonal antibody infliximab (5 mg/kg) in addition to the daily dose of intravenous methylprednisolone. Her neurologic state stabilized and progressively improved. She was much less confused within the next 24 hours. Four days later, her visual acuity markedly improved, and she was able to recognize cartoon character faces on balloons in her room. The patient's ESR also normalized (5 mm/hour). Patient characteristics, treatment, and outcome are summarized in Tables 1 and 2. Intravenous methylprednisolone (1 gm/day) was replaced with oral prednisone (60 mg/day) the day after the infliximab infusion. Treatment with aspirin (100 mg, once per day) also was initiated on day 8.

Table 1. Patient characteristics, clinical findings, and indications for anti-TNF therapy*
CaseAgeSexDisease duration before initiating anti-TNF therapyTreatment before initiating anti-TNF therapyIndications for anti-TNF therapyClinical findings and status at initiation of anti-TNF therapyFindings of contrast-enhanced MRI of the brain at initiation of anti-TNF therapy
  • *

    Anti-TNF = anti–tumor necrosis factor; MRI = magnetic resonance imaging; IV = intravenous; MRS = modified Rankin score; FLAIR = fluid-attenuated inversion recovery; CSF = cerebrospinal fluid; WBC = white blood cells.

127Female20 daysIV methylprednisolone (1 gm/day) for 5 days; 1 IV infusion of cyclophosphamide (1.7 gm)Progressive deterioration of neurologic statusLeft-sided hemiparesis, confusion, aphasia, and blindness; MRS = 5Multiple acute interval infarcts involving both cerebellar hemispheres and both cerebral hemispheres with minimal abnormal enhancement of the lesions. Scattered linear areas of increased FLAIR signal in multiple sulci with minimal enhancement.
239Male29 monthsIV methylprednisolone (1 gm/day) for 3 days; oral prednisone (initial dose 60 mg/day) for 29 months; IV cyclophosphamide (1 gm/month) for 17 months; oral mycophenolate mofetil (1 gm/day) for 12 monthsSecond relapseHeadache, confusion, increased CSF cell count (WBC 51/μl; erythrocytes 10/μl), and increased CSF protein (160 mg/dl); MRS = 3New foci of T2-weighted signal change within the centrum semiovale (bilateral), the region of the left thalamus, and the basal ganglia (bilateral). Punctate subtle foci of enhancement within the basal ganglia regions (bilateral) and the region of the genu of the corpus callosum.
Table 2. Treatment and outcome after anti-TNF therapy*
CaseAnti-TNF therapyOther treatment after initiating anti-TNF therapyFindings of contrast-enhanced MRI of the brain during followupStatus at last followup visit
  • *

    See Table 1 for definitions.

1Infliximab (5 mg/kg), a single IV infusionOral prednisone (initial dose 60 mg/day) for 13 months; IV cyclophosphamide (1.7 gm/month) for 3 months; oral azathioprine (100 mg/day) for 12 months30 days: no new interval infarcts. 16 months: stable with multiple bilateral cerebral and cerebellar infarcts. Moderate cerebral and cerebellar atrophy. No new infarcts.34 months: persisting improvement, no relapses or recurrences, suspension of therapy; MRS = 3
2Etanercept (25 mg) twice weekly for 20 months, then 25 mg/kg, once weekly for 8 monthsOral prednisone for 15 months (varying dose)4 months: reduction of T2-weighted signal abnormalities. No abnormal areas of enhancement.60 months: persisting improvement, no relapses or recurrences, suspension of therapy; MRS = 2

At dismissal after 4 weeks, the patient's neurologic examination showed improved visual function in the right hemifield, better memory and cognition, and better motor function. However, the patient had considerable disability from left hemiparesis and impaired gait, and she needed assistance in activities of daily living. A followup cranial MRI 50 days after her initial presentation showed no new infarcts.

The prednisone dose was gradually reduced. Intravenous cyclophosphamide, after the fourth month injection, was replaced with azathioprine (100 mg/day). Nine months later, treatment with oral prednisone was discontinued. A followup MRI of the head 3 months later showed no new lesions. The patient was able to walk independently but had a hemiparetic and unsteady gait. Azathioprine therapy was discontinued. At her final followup visit 18 months later, she no longer received corticosteroids or immunosuppressants, and she had no recurrence of symptoms attributable to vasculitis.

Case 2.

A 39-year-old man presented with a 1-month history of headaches; confusion; progressive weakness in the legs; numbness in the sacrum, scrotum, and bilateral lower extremities; urinary retention requiring catheterization; and fecal incontinence. The neurologic examination showed a T4 sensory level on the left side and T8 on the right side. He had no previous medical problems.

Laboratory test findings were unremarkable. The ESR was 5 mm/hour and the C-reactive protein level was 0.2 mg/ml (normal <0.8 mg/dl). Tests with normal or negative findings included serum antinuclear antibodies, antineutrophil cytoplasmic antibodies, antibodies against double-stranded DNA, cryoglobulins, anticardiolipin antibodies, blood coagulation studies (including lupus anticoagulant), tuberculin skin test, serology tests (for hepatitis B, hepatitis C, and human immunodeficiency virus), and a fungal serology survey (Coccidioides, Histoplasma, and Blastomyces).

CSF glucose concentration was 31 mg/dl and total protein concentration was 265 mg/dl. The erythrocyte cell count was 50/μl and the white blood cell count was 230/μl (91% lymphocytes, 9% monocytes). CSF cytologic immunostaining showed an increased number of small lymphocytes, consistent with reactive lymphocytosis, and no malignant cells or features consistent with leukemia or lymphoma were identified. Additional CSF tests with negative results included the venereal disease research laboratory test, Cryptococcus antigen tests, fungal and bacterial cultures, and polymerase chain reaction assays for herpes simplex and zoster viruses, Epstein-Barr virus, cytomegalovirus, Tgondii, and Bburgdorferi.

Findings from a chest radiograph, transesophageal echocardiogram, and duplex carotid ultrasonogram were unremarkable. Initial contrast-enhanced MRI of the brain showed diffuse leptomeningeal enhancement overlying both cerebral hemispheres and in the posterior fossa. Abnormal increased T2-weighted signal was observed in the right cerebellar hemisphere and the right occipital cortex. Initial contrast-enhanced MRI of the thoracic and lumbar spine showed T2-weighted signal abnormality of the distal thoracic spinal cord and contrast-enhanced cauda equina nerve roots, which were not thickened or clumped. Initial contrast-enhanced MRI of the cervical spine showed normal findings. Cerebral angiography did not suggest vasculitis. An open brain biopsy of the contrast-enhanced lesions was performed. Histologic examination of the biopsy specimens showed granulomatous leptomeningeal and intraparenchymal vasculitis (Figure 1). Stains of biopsy specimens were negative for fungal and mycobacterial organisms.

Figure 1.

Brain biopsy specimen (case 2) showed small leptomeningeal vessels of the cerebellum that were affected by granulomatous vasculitis. Giant cells (arrow) caused considerable narrowing of the residual vessel lumen (arrowheads). (Hematoxylin and eosin stained; original magnification × 200.).

PCNSV with brain and spinal cord involvement was diagnosed, and the patient initially was treated with pulse intravenous methylprednisolone (1 gm/day) for 4 days, followed by oral prednisone (60 mg/day) and monthly pulse intravenous injections of cyclophosphamide (1 gm/month). His neurologic state dramatically improved in the first week. His headache and problems with weakness in the legs and sphincter function almost resolved. He was able to perform many of his previous usual activities but continued to have a neurogenic bladder.

Followup contrast-enhanced MRI of the brain and spine performed 18 days after the initial studies showed nearly complete resolution of the lesions. Fourteen months later, while being treated with prednisone (10 mg/day) and bimonthly injections of cyclophosphamide, the patient had a relapse with headache and confusion. CSF analysis showed increased protein levels (169 mg/dl); a leukocyte cell count was 48/μl (87% lymphocytes) and erythrocyte cell count was 2/μl. Contrast-enhanced MRI of the brain, performed 14.5 months after the initial brain imaging studies, showed a new gadolinium-enhanced lesion in the superior right thalamus and new T2-weighted signal abnormalities in the deep white matter of both cerebral hemispheres (associated with intravascular enhancement), right centrum semiovale, and white matter adjacent to the atrium of the right lateral ventricle.

Although magnetic resonance angiography of the brain showed normal findings and contrast-enhanced MRI of the thoracic and lumbar spine showed no new lesions, the overall findings indicated recurrent active vasculitis. The prednisone dosage was increased (to 60 mg/day) and cyclophosphamide was replaced with mycophenolate mofetil (1 gm/day). Neurologic symptoms remitted in 1 week, and a CSF examination 2 months later showed normal cell counts and protein level. Eight months later, a followup contrast-enhanced MRI of the brain showed nearly complete resolution of the enhanced lesions and the T2-weighted signal abnormalities. Six months later, while continuing treatment with prednisone (4 mg/day) and mycophenolate mofetil (1 gm/day), the patient again developed headache and confusion. CSF had increased levels of leukocytes (51/μl, 71% lymphocytes), erythrocytes (10/μl), and protein (160 mg/dl). MRI indicated active disease and showed new foci of T2-weighted signal change within the centrum semiovale (bilateral), the left thalamus, and the basal ganglia (bilateral). Punctate subtle foci of enhancement within the basal ganglia regions (bilateral) and the region of the genu of the corpus callosum were also observed. MRI of the spine was unchanged. Prednisone dosage was increased (30 mg/day), and mycophenolate mofetil was replaced with etanercept (25 mg twice weekly), a fusion protein combining human p75 TNFα receptor with Fc-IgG1. The patient characteristics and outcome are summarized in Tables 1 and 2.

On examination 2 weeks later, neurologic symptoms had resolved and the prednisone dose was reduced. Four months later, the CSF had reduced levels of white blood cells (7/μl, 88% lymphocytes), erythrocytes (1/μl), and protein (119 mg/dl). MRI of the head and spine showed no signs of active disease. Eleven months later, prednisone treatment was discontinued. The patient received etanercept (25 mg twice weekly) for another 5 months, and the dosage was then reduced to 25 mg/week.

At an 8-month followup visit, remission had persisted, and etanercept therapy was discontinued. Four months later, the patient's neurologic state was stable and he had no symptoms of vasculitis. He had only a minimal disability and conducted normal activities.


During an analysis of 101 patients with PCNSV treated at Mayo Clinic Rochester over a 21-year period (3), we noted that 2 patients (described herein) were treated with anti-TNF agents for relapse while receiving traditional combined immunosuppressive therapy. The first patient received infliximab because her neurologic condition deteriorated, even after treatment with corticosteroids and cyclophosphamide. A single infusion of infliximab rapidly and effectively improved her neurologic status and resolved MRI abnormalities. In the second patient, long-term TNF inhibition with etanercept arrested the relapse and permitted discontinuation of corticosteroid therapy. In this patient, intensive immunosuppressive therapy with cyclophosphamide and mycophenolate mofetil was not sufficient to maintain disease remission.

Because no single finding or test is diagnostic of PCNSV, conditions that could cause similar symptoms need to be excluded. Other diseases that may mimic PCNSV include sarcoidosis (involvement of other organs is often present and central nervous system [CNS] biopsy samples show a granulomatous process, but seldom show vascular lesions and never have necrosis), systemic vasculitis (e.g., polyarteritis nodosa; vasculitic lesions also are present in other organs), connective tissue diseases such as systemic lupus erythematosus (positive for serum antinuclear antibodies and anti-DNA antibodies; musculoskeletal, cutaneous, and other systemic manifestations), Behçet's disease (recurrent oral and genital ulcerations, ocular inflammation, and venous thromboses), and infections that might be more likely to develop after treatment with immunosuppressive medication. Other entities to be considered in the differential diagnosis of PCNSV include Wegener's granulomatosis, giant cell arteritis, use of selected medication such as stimulants, and noninflammatory vasculopathies such as moyamoya disease. In PCNSV, the ESR is elevated in a minority of patients (e.g., case 2) but should alert the physician to consider other disorders.

Several recent studies have reported the efficacy of TNF blockade in vasculitis (6–9). The results are encouraging because they show a prompt response in cases that are refractory to standard treatment. However, no well-designed randomized controlled trials have been performed, and some report that infliximab and etanercept may not be equally effective in the treatment of different forms of vasculitis or in the treatment of different manifestations of the same vasculitic syndrome (10, 11). The efficacy of TNFα inhibitors for CNS involvement in vasculitis has been studied only in Behçet's syndrome. Three reports of patients with neurologic manifestations of Behçet's syndrome (2 resistant to therapy with a combination of corticosteroids and cyclophosphamide) who were treated with infliximab have been published (12–14). All patients showed a rapid and complete resolution of the neurologic findings and MRI abnormalities. Similar to these studies, our findings indicate that TNF blockade may be particularly effective for patients with severe cerebral vasculitic manifestations.

The risks for patients with vasculitis using these new medications are unclear. Possible infectious and neoplastic complications of anti-TNF therapy have been reported (15). Clinicians using these medications to treat PCNSV should watch for the development of new findings unrelated to the vasculitis.

Our cases suggest that TNFα blockade may be a helpful therapy for patients with PCNSV that is refractory to combined immunosuppressive treatment. Although a randomized controlled trial would be ideal, the rarity of this condition makes this improbable. Currently, pending greater experience with these agents in this clinical situation, we recommend reserving TNFα blockade as second-line therapy for refractory cases.


Dr. Brown had full access to all of the data in the study and takes responsibility for the integrity of the data.

Study design. Salvarani, Brown, Meschia, Giannini, Hunder.

Acquisition of data. Salvarani, Brown, Giannini, Miller, Hunder.

Analysis and interpretation of data. Salvarani, Brown, Calamia, Huston, Meschia, Giannini, Miller, Hunder.

Manuscript preparation. Salvarani, Brown, Calamia, Huston, Meschia, Giannini, Hunder.


Editing, proofreading, and reference verification were provided by the Section of Scientific Publications at the Mayo Clinic.