Central Nervous System Vasculitis (Primary Angiitis of the CNS)
This 32 year old male had a previous admission for a right middle cerebral artery (MCA) stroke which was confirmed by the CAT, MRA, and MRI scans. The entire work-up was otherwise within normal limits except for a CSF WBC of 31 with a differential of 6% neutrophils, 82% lymphocytes, and 11% monocytes. The patient's hemiparesis gradually improved and he was prescribed ECASA 325 mgm daily and discharged home to return for follow-up in the Clinic.
In retrospect, the increased WBCs in the CSF probably warranted further evaluation. Although it could be caused by the stroke itself, the increased WBCs could also result from CNS vasculitis. Since we found no etiology for the stroke, an angiogram might have revealed the cause of the initial stroke. However, the angiogram itself carries a risk of stroke of 1-2%. Furthermore, in up to 40% of the autopsy documented cases of CNS vasculitis, the angiogram shows no abnormality. Thus, it could be argued that no further tests were warranted, and that close follow-up was the most prudent course.
Although the patient's hemiparesis improved, he continued to have chronic headaches, progressive memory difficulties, and personality changes. Three months following the first admission, the patient presented with a severe throbbing headache and a right homonymous hemianopsia. On admission, a CAT scan without contrast demonstrated changes consistent with an acute left posterior cerebral artery (PCA) infarct.
The presence of two different brain infarcts certainly increased the probability of vasculitis. Lumbar puncture on the second admission also revealed an elevated WBC of 17, with a differential of 2% neutrophils, 90% lymphocytes, and 7% monocytes. Given the clinical scenario (multiple strokes in different vascular territories, headaches, memory difficulties, behavioral changes, and CSF pleocytosis) and the absence of systemic signs and symptoms, a diagnosis of central nervous system vasculitis was most likely. Usually arteritides involving the cerebral vessels becomes symptomatic after systemic manifestations in diseases such as polyarteritis nodosa, lupus erythematosus, Wegener's granulomatosis, Behcet's disease, Churge-Strauss syndrome, and Sjogren's syndrome. This patient denied systemic symptoms such as fever, weight loss, joint or muscle pains, skin rash or ulcers; and serologic evidence of inflammation and autoantibodies was absent.
On this second admission, the patient had a cerebral angiogram which was consistent with vasculitis. Five days after admission, a repeat MRI showed the previously mentioned right MCA and left PCA territory infarcts along with new areas of ischemia in the head of the right caudate, right basal ganglia, and right inferior frontal lobe.
In this population of patients, the causes of stroke are numerous. Compared to the typical causes of cerebrovascular disease in older patients, there is a high incidence of treatable disorders. His previous work-up, including normal stroke in the young laboratory tests and a MRA of the head and neck (reported as having a decrease in the number of visualized vessels in the sylvian branches of the right MCA, but normal vessels of the neck) made hematologic disorders, premature atherosclerosis, and arterial dissection less likely.
He had also previously been tested for HIV and CNS infectious causes of stroke (fungal, mycobacterial and spirochetal infections). The HIV test and CNS cultures were repeated and remained negative. Drug abuse (most commonly amphetamines or cocaine) leading to stroke was also considered. The previous toxicology screen had been negative, and he denied any recent drug use, which was confirmed by a toxicology screen.
Cardiogenic emboli are a common cause of stroke in young adults, and cortical ischemic insults in multiple vascular territories would be consistent with an embolic source. However, the patient's EKG was normal on both admissions, and during the previous admission a transthoracic echocardiogram was normal, including no evidence of shunt by saline contrast infusion. There was no previous history of cardiac abnormalities, palpitations, or syncope to suggest an arrhythmia.
Bacterial endocarditis, which is associated with bacterial/fibrin emboli or mycotic aneurysms in the brain, was also considered. However, this patient did not have a history of intravenous drug use and blood cultures from both admissions were negative. If an explanation for the patient's strokes had not been found, we would have considered a transesophageal echocardiogram which provides better visualization of the left atrium and improved detection of other potential stroke risk factors such as, patent foramen ovale, mitral valve prolapse, atrial septal aneurysm, aortic arch atheroma, and mitral valve strands.
The MRI and angiographic findings in association with the CSF abnormalities (suggesting inflammation), the clinical manifestations, and absence of other systemic symptoms, were felt to be consistent with a diagnosis of CNS vasculitis. Although the "gold standard" for all forms of vasculitis is histologic confirmation, a recent review by Calabrese et al., suggests that nearly one in four biopsy procedures may result in a false-negative outcome (3). Therefore, after discussion with the patient and his family, the decision was made to treat for CNS vasculitis and forego the risk involved with brain and leptomeningeal biopsy.
The patient was treated with both intravenous corticosteroids and intravenous cyclophosphamide (cytoxan). This was tolerated without complications and a dramatic improvement of his headache was noted. He was discharged on oral prednisone. The long term plan is to continue monthly cytoxan boosters for 6 months and to continue low-dose oral prednisone throughout this time. The clinical course, as well as the CSF protein and WBC, will be monitored to determine response to treatment and also to help regulate the duration of treatment and the dose of oral prednisone.
CNS vasculitis was first described as a distinct entity by Cravioto and Feigin in 1959 and was named granulomatous angiitis, reflecting the presence of prominent granulomata (4). Granulomata were subsequently found to be a variable and nonspecific finding (3), and in 1983 the term isolated angiitis of the CNS (IAC) was adopted, reflecting the striking restriction of inflammation to vessels within the CNS. More recently, others have used the term primary angiitis of the CNS (PACNS) citing cases in which varying degrees of inflammation have been reported in extracranial vessels, making the term isolated too restrictive (3).
The clinical features of the reported cases are diverse. The mode of onset ranges from acute to insidious, but there often is a prodromal period of six months or longer. Characteristically, the course is chronic and fluctuating. Typically, there are nonfocal symptoms (headaches, confusion, decrease in cognitive function), and a wide range of neurologic deficits have been reported (transient ischemic attacks, strokes, paraparesis, cranial neuropathies, ataxia, seizures, myelopathy) (3,8). Unlike other forms of vasculitis, systemic symptoms such as fever, weight loss, joint or muscle pain are uncommon and serologic evidence of inflammation or autoantibodies are typically absent.
The pathology is an inflammatory process centered around small veins and arterioles that is more prominent in the leptomeninges than in the underlying cortex. The infiltrate is usually granulomatous and may contain multinucleated giant cells. However, 15% of the cases in a recent review showed nongranulomatous pathology and skip lesions to be common (3). The histological appearance of PACNS lacks total diagnostic specificity and it may be difficult to distinguish from other forms of vasculitis affecting the CNS, such as giant cell arteritis, the arteritis of sarcoidosis, and arteritis associated with infection.
The cause of PACNS remains unknown. The presence of granulomata and the absence of antibodies or immune complexes in the vessel walls suggest a disorder of cell-mediated immunity (8). However, the etiology for activation of cell mediated immunity is not clear.
Diagnosis is based on CSF analysis, MRI, cerebral angiography, and leptomeningeal & cortical biopsy. The CSF findings usually include a modest pleocytosis, elevated protein, and normal glucose. Increased IgG synthesis and the presence of oligoclonal bands have rarely been reported (2). It is extremely important that the CSF be cultured, since vasculitis of the CNS has been reported in association with numerous infections involving bacteria, fungi, parasites, spirochetes, mycobacteria, and viruses (5). One recent report suggests that the CSF is abnormal in 90% of histologically confirmed PACNS (2).
MRI changes include evidence of ischemia, involving the cortex and deep white matter, which is often bilateral. Enhancement of these lesions, as well as meningeal enhancement, is variable. There may also be associated hemorrhage and mass effect. It is important to note that a small number of cases have had normal MRI studies; therefore, it is impossible to exclude the diagnosis based on a normal MRI (7). It has been suggested by Calabrese in a recent review (2) that a normal MRI and normal CSF should strongly dissuade against more invasive diagnostic studies, such as angiography or biopsy.
Angiographic features include arterial narrowing and dilatation, occlusion, and alteration of circulation time in the affected vascular distribution. Beading of the arteries and aneurysm formation may also be present. However, these findings are nonspecific and may also be seen in systemic vasculitic disorders and in nonvasculitic conditions, such as neoplastic, infectious, arteriosclerotic, and spastic vascular disease. Unfortunately, angiography is a test of low sensitivity, and results are normal in nearly 40% of patients with histologically proven PACNS (3).
The "gold standard" for diagnosis is histologic confirmation. The results of antemortem biopsy in PACNS suggest that 25% of biopsy procedures may result in false-negative outcome (3). Both technical difficulties and the skip nature of the disease are cited as factors contributing to this high false-negative rate. Stereotactic biopsy is recommended for patients with a mass lesion and biopsy of the non-dominant temporal tip is recommended for patients without focal lesions (8). Both leptomeningeal and cortical biopsies are needed.
Moore has suggested the following criteria to establish the antemortem diagnosis of PACNS:
There are no controlled trials for therapy of PACNS. In general, steroid therapy alone has a transient or no effect on the course of the disease (8). Based on a small series of successfully treated cases in 1983 by Cupps, et al. (6), standard therapy has been a combination of corticosteroids and cyclophosphamide. It is suggested that this combined therapy be continued for 6-12 months after the patient is in remission (2). Monitoring the response to treatment and disease activity can be difficult. Serial monitoring of CSF, MRI's, and angiography have all been reported (3,1).
Email comments: