Vasculitis that affects the central nervous system (CNS) is one of the most formidable diagnostic and therapeutic challenges for physicians, for several reasons: (1) the clinical manifestations of CNS vasculitis are highly variable; (2) the CNS is a common target of many forms of systemic vasculitis and may also be the sole target of vasculitis; and (3) specific noninvasive tests are lacking and material for pathophysiologic investigation is limited. Correct diagnosis requires a high degree of suspicion coupled with knowledge of other diseases that may masquerade as vasculitis.

A primary vasculitis limited to the CNS is referred to as primary angiitis of the central nervous system (PACNS). A generalized systemic vasculitic process can also involve the CNS, and in such cases, it is referred to as secondary CNS vasculitis. This will not be discussed in this chapter.

The earliest reports of PACNS described the disease as fatal and progressive, limited to the CNS, and characterized by rich granulomatous vasculitis. The disease was named granulomatous angiitis of the CNS (GACNS), and remained a rare diagnostic entity until the 1980s. Increasing reports of successful treatment with cyclophosphamide and glucocorticoids as well as the use of angiography for the diagnosis of PACNS heightened the interest in the diagnosis. In 1988, Calabrese and Mallek proposed criteria for the diagnosis of PACNS (Table 1).¹

In the 1990s, we and others began to question whether PACNS is a homogeneous disease or whether different clinical subsets exist. A subset of patients diagnosed on the basis of angiography and with a predictably more benign outcome requiring less intensive therapy was identified and named benign angiopathy of the central nervous system (BACNS). Furthermore, most patients with either angiographically or histopathologically documented PACNS do not fall distinctly within the categories of either GACNS or BACNS and are thus considered atypical. Most atypical cases of PACNS have clinical features that preclude a ready diagnosis of BACNS, they have a GACNS-like presentation but whose biopsies reveal nongranulomatous pathology or they have unusual clinical presentations.

PACNS was first described in the mid-1950s. By 1986, only 46 cases had been reported in the English-language medical literature. Since 1975, an increasing number of cases have been described, and 99 cases were reported through 1990. Today, although PACNS is still uncommon, its specter is frequently raised in patients with neurologic problems of obscure origin, making its diagnostic approach a relevant clinical issue.

PACNS is a heterogeneous disease with different clinical subsets of unknown etiology and pathogenesis. The pathologically defined entity (GACNS) is primarily a leptomeningeal and cortical vasculitis involving the small and medium leptomeningeal and cortical arteries. Pathologic findings include classic granulomatous angiitis with Langhans' or foreign body giant cells, necrotizing vasculitis, or a lymphocytic vasculitis. The initial event that primes the inflammatory cells is not known. However, the final pathway of inflammation leads to occlusion of the involved blood vessel, thrombosis and, ultimately, ischemia and necrosis of the territories of the involved vessels. Limited data suggest an association with systemic viral illnesses or a state of altered host defense and PACNS. Duna et al² analyzed 168 reported cases of PACNS and found that 29 of these were associated with an illness characterized by an immunosuppressive state, including corticosteroid therapy, lympho- or myeloproliferative disorders, and human immunodeficiency virus (HIV) infection. In addition, a variety of pathogens have been documented in association with CNS arteritis, including varicella-zoster virus (VZV), HIV, and cytomegalovirus.^3-5

It is likely that, in the setting of altered host defense mechanisms or in a predisposed patient, a pathogen will escape immune defense mechanisms and induce arteritis. In support of this hypothesis is the well-defined clinical syndrome of post-herpes zoster ophthalmicus contralateral hemiplegia. In this syndrome, a contralateral hemiplegia occurs weeks to months after VZV infection of the trigeminal ganglion and nerve, apparently resulting from the retrograde spread of VZV to intracranial vessels. Viral particles have been identified in the cytoplasm and nuclei of smooth muscle cells within the walls of affected vessels. HIV infection has been similarly described in such a setting.^6-7

Alternatively, BACNS is diagnosed solely by angiopathy, and therefore has been dubbed "angiopathy" rather than "angiitis," reflecting the uncertainty of the underlying pathology. By virtue of its angiographic diagnosis, few cases have included brain biopsies, and these have been uniformly unrevealing. It has been proposed that BACNS represents a form of reversible vasoconstriction or spasm rather than true arteritis. Supporting this contention, clinical and angiographic presentation of patients with BACNS is often identical to that of the cerebral vascular syndromes seen after exposure to sympathomimetic drugs, and that found in the setting of complex headaches, pheochromocytoma, and, rarely, in the postpartum period. In addition, data from our recent cohort of patients with BACNS implies that cerebral artery vasospasm may be the underlying pathology.⁸ In this series, 10 of 16 patients underwent follow-up cerebral angiography over 4 weeks to 8 months, which revealed total or near-total resolution of the angiographic findings in all patients. The reversibility of the angiographic findings suggests a vasospastic process rather than a true arteritis. However, it is also possible that some patients presenting with the diagnostic criteria of BACNS may have true angiitis, but again, the few biopsies performed in such patients have been unrevealing.

The clinical features of patients with PACNS differ according to their subset or clinical variant. Three different subsets have been described: GACNS, BACNS, and atypical PACNS.⁹ Their frequency of distribution is shown in Figure 1.

Clinical subsets of patients with PACNS. Frequency of nosologic subsets of PACNS at the Cleveland Clinic Foundation.

Figure 1

Granulomatous Angiitis of the
Central Nervous System
GACNS represents about 20% of all patients with PACNS. Epidemiologic data for this disease are limited, but it appears to be male-predominant and occurs at any age. It is characterized by a long prodromal period, usually of 6 months or longer, with few cases presenting acutely.

The most common presenting signs and symptoms are headaches and mental status changes. Transient ischemic attacks, strokes, seizures, ataxia, visual changes, aphasia and, rarely, coma can develop in patients with GACNS (Table 2). Alteration of level of consciousness is frequent. Signs and symptoms of systemic vasculitis, such as peripheral neuropathy, fever, weight loss, or rash are usually lacking. Because the vasculitis may affect any area of the CNS, presentation may vary widely, and no set of clinical signs is specific for the diagnosis. GACNS may be suspected in the setting of chronic meningitis, recurrent focal neurologic symptoms, unexplained diffuse neurologic dysfunction, or unexplained spinal cord dysfunction not associated with systemic disease or any other process.

Benign Angiopathy of the
Central Nervous System
BACNS is a subset of PACNS diagnosed on the basis of a characteristic set of clinical findings and a classic or high-probability angiogram. In 1993, we proposed the term "benign angiopathy of the CNS" to define case reports and clinical series of patients diagnosed on the basis of angiography alone who had far more benign outcomes than those described for GACNS.¹⁰ BACNS is a distinct nosologic subset (Table 3) characterized by clearly distinguishable clinical features, treatment, and outcome, in contradistinction to GACNS. Patients with BACNS are more likely to be female; they generally present with an acute onset of headache or neurologic event, and their clinical course is more often monophasic and benign than in patients with GACNS. The distinction between the two subsets is generally not difficult to recognize because signs and symptoms are sharply different (Table 4). We have recently described a large series of patients with BACNS that included the clinical, laboratory, and radiologic features of 16 patients as well as their treatment and outcomes.⁸ In this cohort, the mean age was 40 years, with female predominance (female/male ratio of 4.3:1). The most common presenting symptom was headache, which occurred in (14/16) 88% of cases, followed by focal symptoms in (10/16) 63% and diffuse symptoms in (7/16) 44% (Table 3). Cerebral angiography studies were highly abnormal in all patients. Abnormalities on magnetic resonance imaging (MRI) were found in (10/13) 77% of the patients. Cerebrospinal fluid (CSF) analysis showed mild or normal results in most patients. Follow-up cerebral angiography, performed in (10/16) 63% of the patients, revealed total and near-total resolution of changes in all patients (Figure 2). This suggests that reversibility of the angiographic findings at the follow-up study is essential to securing the diagnosis.

Primary Angiitis of the
Central Nervous System: Atypical Cases
Most PACNS patients present with atypical PACNS. This category does not fit the diagnostic criteria for either GACNS or BACNS, yet these patients demonstrate angiographic or histopathologic evidence of PACNS. Included in this group are patients with abnormal CSF findings that preclude a diagnosis of BACNS, or patients with GACNS-like presentation but without granulomatous features on CNS biopsies. In addition, patients presenting with PACNS at unusual anatomic sites such as the spinal cord or those presenting with mass lesions are included in this category.

The approach to patients suspected of having PACNS is not uniform and varies depending on the clinical presentations and subsets described above. In most instances, and depending on the clinical setting, the diagnosis of CNS vasculitis is secured by either a positive biopsy or high-probability vascular imaging, such as angiography, while excluding other conditions that may mimic the disease (Figures 3A, 3B). Important conditions that may present as PACNS and should be excluded before making any definitive diagnosis include infections, neoplasms, drug exposure, vasospastic disorders, systemic vasculitides, and vasculopathies. Table 5 summarizes the mimics of CNS vasculitis.

Laboratory Tests
No blood studies are diagnostic for CNS vasculitis. Acute-phase reactants, such as sedimentation rate and C-reactive protein, are usually normal in patients with PACNS. If serum markers of inflammation are elevated, secondary forms of CNS vasculitis should be evaluated. If the history and physical examination point to a systemic vasculitis, testing should proceed accordingly. Testing for a variety of infectious organisms, such as mycobacteria, fungi, and HIV is warranted in patients presenting with chronic meningitis. Other serologic tests are indicated if there is a history of exposure, such as tick bites in Lyme disease. Evaluation for hypercoagulable states, emboli, and investigation of drug exposure, including over-the-counter medications, are essential in patients who present with acute focal or multifocal disease.

Cerebrospinal Fluid Analysis
Analysis of the CSF is essential. It is a helpful tool in patients suspected of having CNS vasculitis and is of great value in ruling out infectious mimics. CSF findings are abnormal in 80% to 90% of pathologically documented cases of PACNS. Findings usually reflect aseptic meningitis, with modest pleocytosis and elevated protein levels. The importance of obtaining appropriate stains, cultures, and serologies to exclude any infectious etiologies cannot be overstressed, especially in patients presenting with chronic meningitis. Patients with BACNS typically have a normal or near-normal CSF analysis.

Neuroimaging
Neuroimaging studies, such as computed tomography and MRI, are important in the diagnosis of PACNS but are not specific or sufficient for diagnosis. MRI is a more sensitive diagnostic imaging technique than computed tomography in CNS vasculitis, and should be the initial study of choice when approaching a patient with unexplained ischemia except when cerebral hemorrhage is suspected. The sensitivity of MRI varies with the standard used for the final diagnosis of CNS vasculitis. In angiographically defined cases, the sensitivity of MRI varies between 77% and 100%.^8 11 In our experience with biopsy-proven cases, the sensitivity of MRI approaches 100%. Findings on MRI are variable and include multiple and often bilateral infarcts in cortex, deep white matter, or leptomeninges, with or without contrast enhancement. Sites of contrast enhancement in the leptomeninges provide ideal places for biopsy and may increase the yield of this technique. Advanced neuroradiographic techniques such as the inclusion of diffusion and fluid attenuated inversion recovery MRI sequences, single photon emission computed tomography, or positron emission tomography increase the sensitivity of finding abnormalities but not the specificity for diagnosis. Combining neuroimaging with lumbar puncture increases the overall sensitivity; a normal MRI and lumbar puncture have a high negative predictive value against a diagnosis of PACNS.

Angiography
Angiography is a vital diagnostic modality in evaluating patients with CNS vasculitis provided that its limited specificity and lack of quantitative and qualitative codification is appreciated. Cerebral angiography is limited by several factors. Its sensitivity decreases with the caliber of the vessel, being most sensitive for disease of larger vessels. In pathologically proven cases, such as GACNS where the pathologic involvement is predominantly in the small penetrating vessels beneath the leptomeninges, cerebral angiography findings are generally normal.¹² Moreover, the interpretation of the angiographic findings is often limited by poor specificity (26%).¹³ Findings on angiography include alteration of the vessel caliber (ie, beading) and absence or cutoff of one or more vessels (Figure 4). These types of changes may be seen in multiple vessels in multiple vascular beds or may be limited to a single vessel. These findings are not diagnostic and can be encountered in vasospastic, infectious, embolic, atherosclerotic diseases and in hypercoagulable status, and should be carefully interpreted. In patients with BACNS presentation, involvement of multiple vessels in multiple vascular beds (high-probability angiogram) is characteristic, and documentation of reversibility of the angiographic abnormalities, along the course of the disease, is essential to secure the diagnosis (Figure 2).

Pathologic Examination
of the Central Nervous System
Histologic confirmation of vasculitis is considered the gold standard for the diagnosis of CNS vasculitis, but this procedure is also limited by several factors. First, the procedure is highly invasive and requires the skills of an experienced neurosurgeon, who may not always be available. Second, the technical aspects of the procedures should be tailored to the individual patient. In patients with suspected GACNS, the procedure of choice is open wedge biopsy of the tip of the nondominant temporal lobe with sampling of the overlying leptomeninges and underlying cortex. Alternatively, directing the biopsy to an area of leptomeningeal enhancement, when present, may increase the sensitivity.¹⁴ In addition, CNS vasculitis is a notoriously patchy disease, which limits the sensitivity of the procedure, and as many as 25% of the biopsies are false negative.¹² Finally, the presence of vasculitis in the biopsy specimen should not preclude performing special stains and cultures for occult infections that may produce secondary vascular inflammation.

Several issues emerge when treating CNS vasculitis. An accurate diagnosis, the most important initial step, is hampered by the protean manifestations of the disease and the lack of specific and sensitive noninvasive studies coupled with the low specificity and test efficiency of the invasive studies, such as biopsy and angiography. At the same time, diagnosis and treatment are urgent if permanent neurologic damage is to be avoided. Finally, a major problem for clinicians is how to monitor disease activity and not confuse irreversible target-organ damage with treatment-resistant disease.

Granulomatous Angiitis of the
Central Nervous System: Therapy
There are no controlled studies of therapy for GACNS and therapeutic guidelines are based largely on experts' consensus opinion. On the basis of the historical literature, GACNS is a progressive and highly fatal disorder; thus, patients are treated with a combination regimen of cyclophosphamide and glucocorticoids. Originally, cyclophosphamide treatment was continued for approximately 1 year after remission. More recently, and similar to the treatment of antineutrophilic cytoplasmic antibody-associated vasculitis, the goal is to limit exposure to alkylating agents by treating with oral cyclophosphamide for 3 to 6 months and, once the patient is in remission, to switch to an antimetabolite such as azathioprine or methotrexate for the duration of therapy. Oral glucocorticoids are used, as in Wegener's granulomatosis, starting with 1 mg/kg/day of oral prednisone and gradually tapering to a small daily dose over 8 to 12 weeks. Assessment of disease activity includes monitoring of any new neurologic event, serial MRI examinations at 3- to 4-month intervals, and evaluation and documentation of clearance of CSF abnormalities. Serial MRI examinations are performed primarily to search for silent progression during tapering of therapy rather than to look for radiographic resolution. Adjunctive therapies, such as prophylaxis for Pneumocystis carinii infection and adequate prophylaxis for osteoporosis, should be implemented to avoid treatment-related toxicities.

Benign Angiopathy of the
Central Nervous System: Therapy
As in GACNS, there are no controlled therapeutic studies in patients with BACNS. In general, BACNS patients are treated less aggressively than those with GACNS. In our experience,⁸ patients are generally treated with glucocorticoids for not more than 6 months, and most patients were treated with adjunctive calcium channel blockers. We usually initiate therapy with prednisone 1 mg/kg/day in divided doses in addition to verapamil 240 mg/day. The dose of verapamil is increased if symptoms are not controlled, particularly headaches. In patients with more catastrophic presentation, like strokes or major focal neurologic deficits, intravenous methylprednisolone pulse therapy is initiated. Occasional patients with mild disease burden are treated with calcium channel blockers alone. Once clinical remission is achieved, a follow-up angiography is performed at 6 to 12 weeks to document total or near-total resolution of the underlying abnormalities. When improvement cannot be documented on angiography or when there is angiographic progression, the diagnosis of BACNS should be reevaluated and other entities should be considered. The absence of dynamic angiographic changes after therapy suggests atherosclerotic disease, embolic disease, a hypercoagulable state or, rarely, that this is an aggressive form of true angiitis that is being undertreated.

Atypical Primary Angiitis of the
Central Nervous System: Therapy
Most cases of PACNS in the atypical category can be initially treated with glucocorticoids alone, with tailoring of treatment according to severity and/or progression of the disease. For those patients with a BACNS-like presentation, the addition of a calcium channel blocker is warranted. The addition of cyclophosphamide may be needed in patients with a severe presentation.

Earlier descriptions characterized PACNS as a fatal disease, with most patients diagnosed postmortem. In the early 1980s, reports of successful treatment emerged and physicians became more interested in the diagnosis of CNS vasculitis. With better understanding of the subsets of the disease, the advancing therapy, and treatment-related toxicities as well as the advent of neuroradiology, more favorable outcomes and improving mortality rate have been reported. We have studied the long-term outcomes of the PACNS cohort at the Cleveland Clinic Foundation (total of 54 patients). Forty-one PACNS patients were interviewed and their outcomes were analyzed using the Barthel Index, a validated scale used for patients with stroke, as well as a new scale developed expressly for this investigation (cognitive index).¹⁵ Overall, the outcome was favorable, with a 29% (16/54) relapse rate and 10% mortality rate (5/54). Eighty percent (33/41) of the interviewed patients demonstrated mild to no impairment on the Barthel Index. Using the cognitive index, mild deficits of concentration and memory and decreased energy level were the most frequent symptoms. The benign outcome of BACNS patients was documented by this series, with 94% (15/16) of these patients experiencing significant recovery and 71% (5/7) showing no evidence of long-term disability.⁸

1. Calabrese LH, Mallek JA. Primary angiitis of the central nervous system. Report of 8 new cases, review of the literature, and proposal for diagnostic criteria. Medicine (Baltimore). 1988; 67:20-39.
2. Duna GF, George T, Rybicki L, Calabrese LH. Primary angiitis of the central nervous system (PACNS): an analysis of unusual presentations [abstract]. Arthritis Rheum. 1995;38:S340.
3. Calabrese LH. Vasculitis of the central nervous system. Rheum Dis Clin North Am. 1995;21:1059-1076.
4. Yankner BA, Skolnik PR, Shoukimas GM, Gabuzda DH, Sobel RA, Ho DA. Cerebral granulomatous angiitis associated with isolation of human T-lymphotrophic virus type III from the central nervous system. Ann Neurol. 1986;20:362-364.
5. Gilden DH, Kleinschmidt-DeMasters BK, LaGuardia JJ, Mahalingam R, Cohrs RJ. Neurologic complications of the reactivation of varicella-zoster virus. N Engl J Med. 2000;342:635-645; erratum 2000;342:1063.
6. Linnemann CC, Alvira MM. Pathogenesis of varicella zoster angiitis in the CNS. Arch Neurol. 1980;37:239-40.
7. Yanker BA, Skolnik PR, Shoukimas GM, Gabuzda DH, Sobel RA , Ho DA. Cerebral granulomatous angiitis associated with isolation of human T-lymphotrohic virus type III from the central nervous systems. Ann Neurol. 1986;20:362-64.
8. Hajj-Ali RA, Furlan A, Abou-Chebel A, Calabrese LH. Benign angiopathy of the central nervous system: cohort of 16 patients with clinical course and long-term followup. Arthritis Rheum. 2002; 47:662-69.
9. Calabrese LH, Hochberg MC, Silman AJ, Smolen JS, Weinblatt ME, Weisman MH. Rheumatology. 3rd ed. New York:Mosby;2003. Chapter 157, Primary angiitis of the central nervous system.
10. Calabrese LH, Gragg LA, Furlan AJ. Benign andiopathy: a distinct subset of angiographically defined primary angiitis of the central nervous systems. J Rheumatol. 1993;20:2046-50.
11. Pomper MG, Miller TJ, Stone JH, Tidmore WC, Hellmann DB. CNS vasculitis in autoimmune disease: MR imaging findings and correlation with angiography. Am J Neuroradiol. 1999; 20:75-85.
12. Calabrese LH, Furlan AJ, Gragg LA, Ropos TJ. Primary angiitis of the central nervous system: diagnostic criteria and clinical approach. Cleve Clin J Med. 1992;59:293-306.
13. Duna GF, Calabrese LH. Limitations of invasive modalities in the diagnosis of primary angiitis of the central nervous system. J Rheumatol. 1995;22:662-67.
14. Parisi JE, Moore PM. The role of biopsy in vasculitis of the central nervous system. Semin Neurol. 1994;14:341-348.
15. Hajj-Ali RA, Villa-Forte A, Abou-Chebel A, et al. Long-term outcomes of patients with primary angiitis of the central nervous system (PACNS) [abstract]. Arthritis Rheum. 2000;43:S162.