Primary angiitis of the central nervous system: Clinical profiles and outcomes of 45 patients
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.253578
Source of Support: None, Conflict of Interest: None
Keywords: Brain biopsy, CNS vasculitis, cognitive dysfunction, outcome, primary angiitis of central nervous system, relapse, PACNS, stroke
Primary angiitis of central nervous system (PACNS) is a rare inflammatory disorder of uncertain etiology, affecting blood vessels of the central nervous system (CNS) and often resulting in considerable morbidity and mortality., PACNS can affect any age group with the median age of onset being 50 years and with a male predominance in some studies.,, Patients with PACNS commonly present with headaches, recurrent infarcts and rapidly progressive cognitive impairment; they present rarely with intracranial hemorrhage, tumor-like mass lesions or myelopathy.,,,,, Confirmation of diagnosis requires conventional four-vessel cerebral angiogram (DSA) and/or meningo-cortical biopsy (biopsy), and all other mimicking conditions must have been ruled out.
Outcome in PACNS is variable, with remission, relapsing-remitting course or progressive decline and death, despite therapy. Treatment is challenging, with no randomized controlled trials to guide it, and the current evidence is only based upon retrospective studies and expert opinions. To date, only a few cohorts have been reported, with the largest adult ones being from the Mayo clinic in the US and from France that shed light on the various factors determining treatment responses and outcomes.,, Glucocorticoids have been the mainstay of treatment, often combined with cyclophosphamide (CYP), azathioprine, methotrexate, or mycophenolate mofetil.,, We report herein a new cohort of patients with PACNS, the first from India. We describe their clinical presentation, response to treatment, and analyze the predictors of outcomes and relapses.
This study was a retrospective analysis of 45 consecutive patients diagnosed with PACNS at the Comprehensive Stroke Care Unit in our center from January 2000 to December 2015. Patients with age of onset ≥18 years meeting Calabrese et al.,'s criteria  for the diagnosis of PACNS were included: (1) clinical presentation suggestive of PACNS such as stroke, rapidly progressive cognitive decline or any other neurological deficits unexplained after clinical and laboratory evaluation; (2) evidence of arteritic process either by DSA and/or biopsy; and, (3) exclusion of systemic vasculitis, reversible cerebral vasoconstriction syndrome, CADASIL, large vessel atherosclerotic disease, or any other condition in which the angiographic or pathologic features could be secondary. Investigations done routinely to exclude secondary vasculitis included anti-nuclear antibody (ANA), antidouble-stranded deoxyribonucleic acid (anti-dsDNA), lupus anticoagulant, anticardiolipin antibody, antineutrophil cytoplasmic antibodies (ANCA), anti-SSA (anti-Sjögren's-syndrome-related antigen A), anti-SSB (anti-Sjögren's-syndrome-related antigen B cryoglobulins, human immuodeficiency virus enzyme linked immunosorbent assay (HIV ELISA), hepatitis B surface antigen (H BsAg), anti hepatitis C virus (HCV) antibodies and CSF gram stain, mycobacterium PCR (polymerase chain reaction), venereal disease research laboratory (VDRL) test and culture. Patients were also excluded if their follow up period was less than 6 months except those who died within this period.
Demographic data, clinical details and clinical scales [modified Rankin scale (mRS), NIH stroke scale scores (NIHSS)] were recorded; these were performed by certified examiners for the last 14 patients evaluated after 2010. All DSA and MRI data were reviewed by a neuroradiologist. DSA was considered to be suggestive of vasculitis when alternating areas of narrowing and dilatation of cerebral arteries or arterial occlusions (beaded appearance) were present affecting many cerebral vessels in the absence of proximal vessel atherosclerosis or other recognized abnormalities., The type of vessel involvement was classified as large, medium or small, based on the size of the affected vessels. Intracranial segment of internal carotid artery, proximal middle cerebral artery (M1), anterior and posterior cerebral arteries (A1 and P1) were considered large, second divisions were considered medium, and subsequent divisions were considered small vessels.,, Patterns of infarcts were also classified into single and multiple, based on the number of arterial territories involved. Infarcts were also sub-classified into cortical and subcortical (periventricular and deep white matter, basal ganglia infarcts and small vessel ischemic changes) locations.
On biopsy, the diagnosis of vasculitis required the presence of transmural inflammation. The pattern of vasculitis was further classified into granulomatous (presence of granulomas), lymphocytic (lymphocytic infiltration alone) or necrotizing (presence of fibrinoid necrosis). Biopsies were differentiated into either targeted biopsy, if performed at a radiologically abnormal area with or without gadolinium enhancement, or blinded biopsy (non-dominant frontal or temporal pole).
Cerebrospinal fluid (CSF) findings were considered abnormal if total leukocyte count was >5/mL and/or total protein level was ≥50 mg/dL. Electroencephalogram (EEG) was considered abnormal when there was persistent focal or diffuse delta slowing or epileptiform abnormalities. Detailed treatment histories, including the drugs received and their doses, durations and adverse effects, were also recorded.
The follow-up period was calculated from the time of initiation of treatment to last visit or death. Poor outcome was defined as the 6-month mRS ≥3. Relapses and mortality were recorded. Relapse was defined as recurrence or increase in existing symptoms of PACNS, new clinical deficits or evidence of new infarction on subsequent MRI independent of treatment status. Recurrence of seizures or headache without any other clinical or radiological evidence of active CNS vasculitis were not considered to be relapses.
Standard protocol approvals, registrations and patient consents
The study was approved by the Institutional Ethics Committee.
Continuous variables were reported as median (range) and categorical variables as proportions. Differences in continuous variables were assessed by Mann Whitney test and binary categorical variables by chi-square test or Fisher's exact test when appropriate. For variables with more than two categories, statistical tests were not attempted due to the small cell values.
We identified 70 patients with a diagnosis of probable PACNS, (and excluded 6 patients who were lost to follow-up), and 19 patients due to secondary vasculitis. The remaining 45 patients had 36 years as the median age of onset (range, 19-70) years, and 31 (68.9%) were males. The median delay from symptom onset to the presentation to our hospital was 48 days (range, 1-2840). The initial presentation was stroke in 19 (42.2%), headaches in 11 (24.4%), seizures in 8 (17.8%) and rapidly progressive cognitive dysfunction in 5 (11.1%) patients [Table 1]. Among the 19 patients who presented with stroke, four had intracranial hemorrhage. The rarest presentations of PACNS in our cohort were myelopathy (n = 1) and a tumor-like mass lesion (n = 1). Diagnoses were established by DSA in 26 (57.8%), biopsy in 15 (33.3%), and by both biopsy and DSA, in 4 (8.9%) patients.
MRI brain was available for all patients except one. About 35 patients (79.5%) had multiple infarcts and 5 patients (11.4%) had a single infarct [Table 2]. Subcortical, cortical and both subcortical and cortical infarcts were seen in 11 (25%), 10 (22.7%) and 7 (15.9) patients, respectively. Multiple micro-bleeds (≥5) in gradient sequences were seen in 14 patients (31.8%). Spine imaging was performed in 12 patients, one of whom had cervical cord hyperintensity spanning three segments.
Among 22 patients with time of flight sequence MRA (TOF-MRA), 10 (45.5%) showed features of vasculitis. Among the 42 patients who had DSA, 30 were abnormal, showing isolated large vessel, medium vessel and small vessel involvement in 4 (13.3%), 3 (10%) and 8 (26.7%) patients, respectively; the rest had multiple types of vessel involvement. MR vessel wall imaging in two patients showed concentric vessel wall thickening and enhancement suggestive of vasculitis [Figure 1].
Among 30 patients who were DSA positive, 14 underwent a biopsy. Biopsy was not done in 16 patients as these patients refused the procedure. The 12 DSA negative patients and 3 patients without a DSA underwent biopsies. Thus, biopsy was done in 29 patients, and 19 (65.5%) had histopathological evidence of PACNS [Table 1] and [Figure 2]. Lymphocytic, necrotizing and granulomatous patterns were observed in 11, 5 and 3 patients, respectively. Among 18 patients who underwent targeted biopsy, 16 (88.9%) patients had evidence of CNS vasculitis, as compared to 3 (27.3%) of 11 patients with a blinded biopsy. Two patients (7.1%) developed surgical complications, one with worsening of hemiparesis and the other with post-operative fatal meningitis.
The median delay from onset of symptoms to start of treatment was 103 days (range 12-1893); and, with normal and abnormal DSA was 212 (range 24-742) and 60.5 (range 3-1893) days, respectively. All patients received glucocorticoids (prednisolone 1 mg/kg) as their first treatment, and in addition, 14 received monthly intravenous cyclophosphamide (CYP; n = 10), oral CYP (n = 1) or azathioprine (n = 3). The median duration of treatment was 6 months (range 2-36 months). CYP was subsequently initiated in eleven patients due to a relapse (7 had new deficits and 2 had new lesions on MRI and the other 2 had clinical progression). Treatment-related complications were observed in 14 (31.1%) patients, including 11 directly related to steroids (Cushingoid features, hyperglycemia, hypertension, cataract and avascular necrosis of neck of femur), one with Steven Johnson syndrome, one with abdominal tuberculosis and one with septicemia while on CYP.
The median duration of follow-up was 33.1 months (range 0.7-356). About 12 patients (26.7%) had a poor 6-month outcome which included 3 patients who died. Between the initial diagnosis and last follow up, 25 (55.6%) patients had relapsed and 7 (15.6%) patients died. Initial features of cognitive dysfunction (P =0.014), NIHSS ≥5 (P < 0.0005) and an abnormal EEG (P = 0.046) were more frequently associated with a poor outcome [Table 3] and [Table 4]. Higher NIHSS at admission (P = 0.032) and normal DSA at diagnosis (P = 0.002) were also more frequent in patients who had a relapse [Table 5]. Among the seven patients who died, all had received glucocorticoids alone as the initial treatment and two patients subsequently received CYP on relapse. The causes of death were Steven Johnson syndrome, sepsis with multi-organ dysfunction (while on combination therapy), recurrent stroke and aspiration pneumonia (two patients), pulmonary thromboembolism, and intracranial infection after biopsy; the cause of death in one patient was unavailable.
This cohort, the first from India, of 45 patients with PACNS adds to the current, limited knowledge of this rare disease. Stroke was the commonest clinical presentation followed by headaches, seizures and cognitive decline. A poor outcome was seen in 26.7%, relapse in 55.6% and mortality in 15.6% patients, reinforcing the significant morbidity and mortality associated with this disease.
The median age of onset of PACNS in our cohort was 36 years, much younger than reported in earlier studies.,, This observation may be specific to India. Other stroke subtypes in India have also been shown to have a lower mean age compared to data from the western population. Clinical manifestations were similar to previous reports with persistent motor deficits, stroke, headache and cognitive impairment being the most frequent., CSF abnormalities were seen in more than half of our patients, similar to the French cohort. However, in the Mayo-clinic series, more than 90% had CSF abnormalities.
MRA is reported to show abnormal vessels in at least half of the patients, as was observed in our study as well., The sensitivity can be further improved with TOF-MRA at 3T which detected vessel stenosis in 47% of patients diagnosed with PACNS, as compared to 14% at 1.5T. Multiple blooming lesions on gradient echo sequences in patients with focal deficits and headache were observed in 14 patients ultimately diagnosed with PACNS in our study, suggesting that this unique pattern of findings in a young patient should raise the suspicion of PACNS.
High resolution intracranial vessel wall MRI is a novel diagnostic imaging technique used since 2016 at our Institution and was helpful in supporting the diagnosis of PACNS. Two patients in this cohort had classical vessel wall imaging features of concentric thickening with enhancement in the intracranial vessels suggestive of vasculitis. Currently, MR vessel wall imaging is used to differentiate vasculitis from intracranial atherosclerotic disease, to assess vasculitic activity and to select the intracranial target for biopsy.
DSA was positive in 71.4% of patients who underwent the procedure, a figure which was lower than that reported in the Mayo-clinic (88%) series. This may be related to a predominance of smaller vessel vasculitis, since vessels less than 500 μm in diameter are known to be beyond the resolution of DSA., Among those whose vessels were abnormal, we observed more medium than small vessel involvement, unlike the findings in prior studies., This disparity may be related to the variable criteria used for defining large, medium or small arteries in the literature.
The most common histological pattern in our study was lymphocytic, similar to the findings of the previously published cohort from Australia. However, other studies including a small case series from India (n = 8), showed granulomatous vasculitis as the commonest pattern.,, Patients who had rapidly progressing symptoms with histopathological evidence of granulomatous or necrotizing pattern were previously reported to be more likely to have a fatal outcome, but this finding was not observed in our study, possibly due to the inadequate numbers in these subgroups.
The median delay to reach a definitive diagnosis in our study was 103 days, which was markedly longer compared to the previous studies., Treatment was started later in patients diagnosed by biopsy compared with those diagnosed by DSA alone (the median duration being 212 and 60.5 days, respectively). Diagnostic biopsies were delayed if the patient was on empirical steroids before their referral or due to scheduling issues; and, in patients with positive DSA findings, biopsy was pursued only when there was relapse or progression of the disease. The French series showed a comparable delay in patients who underwent biopsy, and in the US series, biopsy was done sooner., An early targeted biopsy including tissues both from the parenchyma and the overlying leptomeninges, to determine treatment strategies and exclude other causes, may be a better approach.
The combination therapy of CYP and glucocorticoids were started in 24.4% of our study population, which was much less as compared to the French cohort (85%). This was associated with lesser relapse (27% vs 55.6%) and mortality (6%% vs 15.6%), but more adverse events (54% vs 31%). Aggressive immunosuppression was often delayed in our cohort, if the DSA was normal, until the diagnosis was a biopsy-proven one, as empiric cytotoxic therapy poses a significant risk for PACNS mimics.
A normal DSA predicted a relapse, which was an interesting finding observed in our study. A favorable prognosis in DSA positive cases compared to biopsy proven cases has been reported in one prior study, but later studies have not replicated this finding., In the French cohort, relapse were more common in patients with gadolinium enhancing lesions on MRI and in those having seizures. In the Mayo clinic series, relapses were more frequent in those with large-vessel involvement, while other studies have reported more frequent relapses in patients with small-vessel involvement.,,
The treatment choices and their timing were not shown to be associated with outcome or relapse in our study. In the recent update from Mayo clinic series, relapses were more frequent in patients treated with glucocorticoids alone initially when compared to the combination therapy of glucocorticoids and CYP. Some groups propose that oral prednisone alone could be prescribed in patients who are stable with minor deficits, and a combination therapy (glucocorticoids and CYP) may be more appropriate for more severe cases. The European League Against Rheumatism recommends a combination therapy with CYP and glucocorticoids as first-line therapy in small and medium vessel systemic vasculitis. The role of newer therapies with fewer adverse effects, such as rituximab, needs to be explored. Randomised controlled trials are needed to identify the optimal initial treatment approach.
The major limitation was that the study was retrospective, and thereby bound to limitations in data collection that can influence the findings. The sample size, while small, encompassed a 15-year period and is among the largest data-set published to date and the first from India. We identified a few factors which could determine the outcome and relapse. Identification of these cases may help in guiding treatment. This study is also limited by a referral bias, so that more treatment resistant or severe cases were sent to our center. Biopsy, the gold standard for diagnosis, was not undertaken in all the cases. There was lack of uniformity in the treatment of these patients which might have influenced the outcome. The majority of our patients received glucocorticoids as the first-line treatment, and the current expert opinion suggests an early initiation of concurrent CYP especially in rapidly progressive course and large vessel vasculitis.
This additional cohort adds to existing literature and may help better identify those patients with PACNS who have poor prognostic factors and may benefit from more aggressive treatments.
The authors acknowledge the contributions of Dr. Rajalakshmi P, Assistant Professor, Department of Pathology of our Institute for providing the histopathological images of meningo-cortical biopsy.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]