Primary Angiitis of the CNS

Abstract

Background and Objectives

Primary angiitis of the CNS (PACNS) is a rare disease that has significant morbidity and mortality. Subtypes of PACNS can have different presentations that could be missed with certain diagnostic modalities, further increasing diagnostic complexity. We sought to distinguish the subtypes of PACNS and describe their outcomes in an Indian cohort.

Methods

Adult patients in this retrospective single-center cohort study were reviewed from the PACNS database between 2000 and 2019. Diagnosis was made as per Calabrese and Malleck criteria. Small and medium vessel vasculitis was defined, and their clinical and radiologic profile, treatment, and outcomes were compared. Functional outcomes were noted at 6-month, 1-year, and at last follow-up, while relapses were noted at last follow-up. A poor outcome was defined as modified Rankin Scale >2.

Results

Seventy-two patients fulfilled the inclusion criteria of whom 50 (69.4%) were male. The small vessel vasculitis subtype had a younger age at onset (30.5 vs 40.5 years, p = 0.014), presented less often as a stroke (22% vs 62%, p = 0.001), and had greater delay in diagnosis and treatment initiation (median of 620 days vs 118 days, p = 0.001) compared with medium vessel vasculitis subtype. Although no difference was noted at 6 months, the small vessel vasculitis group had poor outcomes at 1-year and last follow-up (57% vs 20%, p = 0.011 and 72% vs 34%, p = 0.005, respectively) and had more relapses at last follow-up (89% vs 30%, p < 0.001) when compared with the medium vessel vasculitis group. On analyzing the entire cohort, 50 of 72 (69%) and 37 of 53 (69.8%) patients had a good outcome at 6 months and 1 year, respectively. Relapse was noted in 35 of 72 (49%) at final follow-up. The choice of the treatment regimen did not predict outcomes or relapses.

Discussion

The small vessel vasculitis subtype of PACNS is a distinct entity that has diagnostic and treatment delays with poor long-term outcomes and more relapses. Recognizing the different subtypes of PACNS may help to expedite diagnosis and plan treatment.

Introduction

Primary angiitis of the CNS (PACNS) is a very rare disease of unknown etiology that has a high morbidity and mortality especially if not properly diagnosed or treated. Diagnosis is often made through clinical presentation, CNS imaging features, CSF examination, and digital subtraction angiography (DSA), although the gold standard remains histopathologic examination of a meningocortical biopsy (MCB),¹ with exclusion of other secondary causes of vasculitis. The disease can have varied presentations due to predominant involvement of small, medium, or large vessels in the brain. Angiography-negative small vessel vasculitis has been recognized as a distinct subtype that often presents with cognitive decline and inflammatory CSF analysis.² It is important to understand the differences between the subtypes because biopsy-positive small vessel vasculitis is often angiogram-negative due to the smaller size of the affected vessels beyond the resolution of conventional angiogram (<500 μm),³ while the medium and large vessel vasculitis group may be missed by biopsy but is angiogram-positive. This is relevant because the outcomes of patients with PACNS are variable depending on the subtype. Being a rare disease, the data regarding the profile of patients with subtypes of PACNS are largely from 2 major registries, with the Mayo clinic⁴ reporting good outcomes in patients with small vessel vasculitis and the French cohort⁵ reporting no difference in outcomes albeit with a higher relapse rate in patients with small vessel vasculitis. Our primary objective was to describe the differences in clinical profile between the subtypes of isolated small vessel vasculitis and medium/large vessel vasculitis in a cohort of patients with PACNS from India. The secondary objective was to analyze the outcomes of the entire cohort at 6 months and 1 year. Knowledge of these would give us further insights into the disease and to identify any differences between patient populations.

Methods

This is a retrospective review of patients diagnosed with PACNS in a tertiary care academic hospital from the PACNS database from January 2000 to December 2019. The inclusion and exclusion criteria remained the same as those of a previous report from our institute on the initial 45 patients.⁶ Patients were included if they were older than 18 years and were diagnosed to have PACNS based on Calabrese and Mallek criteria⁷: (1) patients with acquired neurologic deficit that is unexplained after clinical and laboratory evaluation. (2) Confirmation of diagnosis through DSA and/or MCB. DSA should reveal alternating areas of stenosis and ectasia of vessels or both in more than 1 vascular bed not explained by other mimics. Biopsy should show evidence of transmural inflammation of vessels and exclusion of alternative etiologies. (3) Exclusion of secondary causes of vasculitis such as systemic vasculitis, infections, drugs, and conditions that could mimic similar angiographic appearances such as reversible cerebral vasoconstriction syndrome and intracranial atherosclerotic disease.

Investigations to rule out secondary causes of vasculitis included anti-nuclear antibody (ANA), anti–double-stranded DNA antibody, anti-neutrophil and cytoplasmic antibodies, antiphospholipid antibodies, anti-Ro and anti-La antibodies, serum Venereal disease research laboratory, viral serologies such as HIV enzyme-linked immunosorbent assay, hepatitis B surface antigen and anti-hepatitis C antibodies, CSF for gram stain and Mycobacterium tuberculosis, and fungal PCR. Patients who had follow-up less than 6 months were excluded (unless the patient died within this period).

The clinical profile, National Institute of Health Stroke Scale (NIHSS), modified Rankin Scale (mRS), and laboratory data were collected through a review of the electronic medical records. Imaging of each patient was independently reviewed by a neuroradiologist. All patients had a MRI of the brain that was reviewed for the number of infarcts (single/multiple) based on the number of arterial territories involved, pattern of infarcts (cortical/subcortical or both) based on the location, hemorrhages (macrobleeds or microbleeds), and small vessel ischemic changes (as per Fazeka grading). Angiograms were reviewed for alternating areas of stenosis and ectasia of vessels or both in more than 1 vascular bed.

Biopsy required the presence of transmural inflammation. Biopsy patterns were classified as granulomatous, lymphocytic, and necrotizing depending on the presence of granulomas, lymphocytic infiltrates, and fibrinoid necrosis, respectively. Biopsy was considered a targeted biopsy if taken from a radiologically abnormal area with or without gadolinium enhancing lesions or a blinded biopsy if taken from the nondominant frontal or temporal pole. CSF was considered abnormal if pleocytosis was >5 cells/micro-L or elevated protein >50 mg/dL. The treatment aspects such as the choice of the initial induction and maintenance regimen and reason for a switch in the regimen, their duration, and adverse effects related to the same were noted. Time to initiation of treatment from the onset of neurologic symptoms was noted for all patients.

The total duration of follow-up was defined as the time from initiation of treatment till the date of last follow-up visit. Poor outcome was defined by an mRS >2 and was noted during follow-up visits at 6 months, 1 year, and the last clinical visit for each patient. Relapse was defined as the occurrence of a new clinical symptom or recurrence of existing symptoms, evidence of a new focal deficit, or corresponding lesion in MRI independent of treatment status. Recurrence of clinical features of headache or seizures without deficits or corresponding new radiologic signs were not considered a relapse. The number of relapses and factors associated with the same was assessed at final follow-up.

We classified vasculitis into small, medium, or large vessel involvement based on the size of the affected vessel. Large vessels consisted of intracranial segment of internal carotid artery and proximal middle, anterior, and posterior cerebral arteries (M1, A1, and P1, respectively). The second divisions of these arteries were considered medium vessels, while the subsequent divisions were considered small vessels.⁶ We segregated patients into 2 groups for further comparisons: isolated small vessel vasculitis who were diagnosed only by biopsy with a normal angiogram; medium/large vessel vasculitis in whom angiograms were positive, and biopsy was either negative or not performed. We compared the differences in the profile of patients between these 2 groups. We also aimed to describe factors associated with good outcomes (at 6 months and 1 year) and relapse in this entire PACNS cohort.

Standard Protocol Approvals, Registrations, and Patient Consents

The study was approved by the Institutional Ethics Committee (IEC number: SCT/IEC/1368-April 2019). As it was a retrospective review, the Institutional Ethics Committee provided a waiver of the requirement for informed consent.

Data Availability

Deidentified patient data not published in this article may be obtained from the corresponding author by qualified researchers on reasonable request.

Statistical Analyses

We reported continuous variables as median with interquartile range and categorical variables as percentages for the baseline characteristics. To assess differences within continuous variables, we used the Mann-Whitney test, and for binary categorical variables, we used the χ² test or Fisher exact test as appropriate. Missing data for final follow-up were analyzed with the last observation carried forward. The χ² or Fisher exact test was used to compare characteristics between isolated small vessel vasculitis and medium vessel vasculitis. A multivariate logistic regression analysis was performed for variables deemed significant in univariate analysis to determine predictors of outcomes at 6 months and 1 year. All p values were 2-sided, with significance being defined by p < 0.05. All statistical analyses was performed using IBM SPSS version 20.

Results

From the database of 146 patients with a diagnosis of probable CNS vasculitis from January 2000 to December 2019, we excluded 87 patients (25 were secondary CNS vasculitis, 38 were lost to follow-up, 10 were diagnosed to have intracranial atherosclerosis, 2 had drug-induced vasculitis, and 12 had no imaging available for review), as shown in the Figure. Among the 25 patients with secondary CNS vasculitis, 9 patients were diagnosed to have anti-phospholipid antibody syndrome, 3 patients had systemic lupus erythematosus, 1 patient had anti-synthetase antibody syndrome, and 3 patients had evidence of peripheral nerve vasculitis concurrent with CNS vasculitis, while 9 patients had strong positive ANA, SSA or SSB, elevated systemic inflammatory markers such as ESR or CRP suggestive of probable secondary CNS vasculitis. Of the 72 patients in the cohort who satisfied the diagnosis of PACNS, 50 were male (69.4%). The most common initial presentation in the cohort was stroke in 37 (51.4%), followed by headache in 17 (23%), seizures in 8 (11%), and cognitive decline in 7 (9.7%) patients, respectively. MCB showed histopathologic evidence of vasculitis in 22 of the 44 patients (50%) who underwent the procedure. A positive yield in biopsy was noted in 18 (82%) patients who underwent a targeted biopsy as compared with 4 (18%) patients who underwent blinded biopsy. The clinical and imaging characteristics of the entire cohort are detailed in Table 1.

Figure. Flowchart Showing the Inclusion of Start Population.

Table 1.

Characteristics of Study Population

Variable	Total N = 72 (%)
Sex, male	50 (69.4%)
Age at onset, median (range), y	38 (19–70)
Clinical features
Transient ischemic attack	5 (6.9)
Ischemic stroke	54 (75)
Intracranial hemorrhage	4 (5.6)
Cognitive decline	25 (34.7)
Neuropsychiatric manifestations	9 (12.5)
Headache	29 (40.3)
Seizures	28 (38.9)
Hemiparesis	51 (70.8)
Paraparesis/quadriparesis	4 (5.6)
Dysarthria	40 (55.6)
Aphasia	19 (26.)
Sensory deficits	20 (27.8)
Cerebellar ataxia	17 (23.6)
Visual symptoms	10 (13.9)
Severity of deficits at admission
NIHSS, median (IQR)	3 (2–7)
mRS, median (IQR)	3 (2–4)
CSF abnormal	45 (62.5)
Abnormal EEG	23 (31.9)
MRI brain, abnormal	72 (100)
Single arterial territory infarct	6 (8.3)
Multiple arterial territory infarct	54 (75)
Small vessel ischemia (Fazeka grade 2 & 3)	39 (54.2)
Intraparenchymal hemorrhage	5 (6.9)
Multiple microbleeds on SWI (≥5)	19 (26.3)
Gadolinium enhancement	29 (40.3)
High-resolution MR vessel wall imaging, abnormal	20/21 (95.2)
Conventional cerebral angiogram, abnormal	54/70 (77.1)
Meningocortical biopsy, abnormal	22/44 (50)
Histopathology
Lymphocytic	14 (64)
Granulomatous	3 (13)
Necrotizing	5 (24)
Induction therapy
Glucocorticoids alone	45 (62.5)
Glucocorticoids with cyclophosphamide	27 (37.5)
Maintenance therapya (n = 71)
Corticosteroids	38 (52.8)
Steroid-sparing agents	33 (45.8)
Duration of follow-up in days, median (IQR)	692 (264–1874)
Good outcomes (mRS ≤2)
At 6 mo	50/72 (69.4)
At 1 y	37/53 (69.8)
Number of patients with relapse at final follow-up	35 (48.6)
Off treatment at final follow-up	19 (26)
Mortality	8 (11.1)

Abbreviations: IQR = interquartile range; mRS = modified Rankin Scale; NIHSS = National Institute of Health Stroke Scale; SWI = susceptibility-weighted imaging.

^aOne patient in the medium vessel vasculitis group who received induction therapy with steroids died during hospitalization before the maintenance regimen could be instituted.

Diagnosis was established through clinical, imaging, and angiogram in 50 patients (69.4%), through biopsy in 18 patients (25%), by both angiogram and biopsy in 4 patients (5.5%).

We identified 18 patients with isolated small vessel vasculitis and 50 patients with medium/large vessel vasculitis. The differences in profile between the 2 groups are given in Table 2. The small vessel vasculitis group had a younger median age at onset of symptoms (30.5 vs 40.5 years, p = 0.014). Headache was the most common presentation among the small vessel vasculitis group (33% vs 20%, p = 0.51), while stroke was the most common presentation in the medium/large vessel vasculitis group (62% vs 22%, p = 0.001). Treatment regimens included an induction regimen of steroids alone or combination of steroids with cyclophosphamide, as given in Table 1. No other induction regimens were used. The maintenance regimen involved steroids alone or steroid-sparing agents such as azathioprine, mycophenolate mofetil, cyclophosphamide, rituximab, and methotrexate.

Table 2.

Comparison Between Subtypes of PACNS (Small Vessel vs Medium/Large Vessel Vasculitis)

Variables	Small vessel vasculitis (biopsy-positive, angiogram-negative) N = 18	Medium/large vessel vasculitis (angiogram-positive, biopsy-negative or not performed) N = 50	p Value
Median age at onset (IQR)	30.5 (24–39)	40.5 (34–49)	0.014
Sex			0.986
Male (%)	13 (72)	36 (72)
Female (%)	5 (28)	14 (28)
Type of presentation(%)			0.041
Headache	6 (33)	10 (20)	0.512
Stroke	4 (22)	31 (62)	0.001
Cognitive decline	2 (11)	4 (8)	0.091
Seizures	4 (22)	4 (8)	0.148
NIHSS ≥5	9 (50)	15 (30)	0.128
EEG abnormal(%)	9 (64)	13 (54)	0.542
CSF abnormal (%)	15 (88)	27 (57)	0.022
Mean cell count, per cu. Mm, (SD)	17(±16)	9.5(±19)	0.110
Mean protein in g/dL (SD)	61(±30.6)	50(±20.8)	0.107
MRI
Multiple infarcts	10 (56)	37 (74)	0.883
Fazeka grade 2 and 3 changes	9 (53)	27 (59)	0.682
Multiple microbleeds in SWI (≥5)	2 (14)	17 (39)	0.109
Gadolinium enhancement	9 (53)	17 (39)	0.345
Abnormal VWI	1 (5)	19 (90)	0.025
Initial treatment (%)			0.272
Glucocorticoids alone	14 (78)	30 (60)
Glucocorticoids with cyclophosphamide	4 (22)	20 (40)
Maintenance therapy, n = 67 (%)a			0.098
Corticosteroids	6 (33)	30 (60)
Steroid-sparing agents	12 (67)	19 (38)
Poor outcomes (mRS >2)
At 6 mo(%)	6/18 (33)	15/50 (30)	0.793
At 1 y	8/14 (57)	7/35 (20)	0.011
At last follow-up	13/18 (72)	17/50 (34)	0.005
Relapse (%)	16 (89)	15 (30)	<0.0001
Mortality at last follow-up (%)	3 (17)	3 (14)	0.667
Median time from symptom onset to treatment in days (IQR)	620.5 (378–801)	118 (35–509)	0.001
Time from first visit to our hospital to diagnosis in days, median (IQR)	127.5 (17–359)	11.5 (6–25)	0.001
Duration of follow-up in days, median (IQR)	1,222 (432–2,409)	608 (243–1,189)	0.173
Off treatment at last visit (%)	3 (17)	15 (30)	0.272

Abbreviations: IQR = interquartile range; mRS = modified Rankin Scale; NIHSS = National Institute of Health Stroke Scale; SWI = susceptibility-weighted imaging; VWI = vessel wall imaging.

^aOne patient in the medium vessel vasculitis group who received induction therapy with steroids died during hospitalization before the maintenance regimen could be instituted.

Medium/large vessel subtype often had abnormal vessel wall imaging (90% vs 5%, p = 0.025). The small vessel vasculitis group often had an abnormal CSF (88% vs 57%, p = 0.022), greater median delay from first visit to our center to diagnosis (128 days vs 12 days, p = 0.001), symptom onset to initiation of treatment (621 days vs 118 days, p = 0.001), and increased relapses (89% vs 30%, p < 0.0001) as compared with the medium/large vessel vasculitis group. Follow-up data at 1 year were available for 49 (68%) patients in these 2 groups. Although the 6-month outcomes were similar, at 1-year and at final follow-up, poor outcomes (mRS >2) were more often seen in the small vessel vasculitis as compared with medium/large vessel vasculitis (57% vs 20%, p = 0.011 and 72% vs 34%, p = 0.005, respectively). The maintenance regimen with steroid-sparing agents was used more often in the small vessel vasculitis group (67 vs 38%, p = 0.098), although the difference was not statistically significant. Mortality was not different between the 2 subtypes.

When we assessed the entire cohort, good outcomes at 6 months were noted in 50 of 72 (69%) patients. The presence of aphasia (50% vs 16%, p = 0.003), altered sensorium at presentation (36% vs 10%, p = 0.007), and NIHSS ≥5 (77% vs 20%, p = 0.0001) were factors associated with poor outcome at 6 months. Fifty-three of 72 (74%) patients had available 1-year outcomes of whom approximately 70% had a good outcome. At 1 year, poor outcomes were noted in those with baseline NIHSS ≥5 (63% vs 22%, p = 0.004), an abnormal CSF (100% vs 49%, p = 0.006), and Fazeka grade 2 and 3 small vessel ischemic changes (81% vs 47%, p = 0.022), while an abnormal DSA predicted a favorable outcome (84% vs 50%, p = 0.022) as detailed in Table 3.

Table 3.

Factors Associated With Outcomes at 6 Months and 1 Year

Variables	Good outcomes at 6 mo (mRS ≤2) N = 50	Poor outcomes at 6 mo (mRS ≥3) N = 22	p Value	Good outcomes at 1 y (mRS ≤2) N = 37	Poor outcomes at 1 y (mRS ≥3) N = 16	p Value
Median age at onset (IQR)	38 (28–44)	41 (36–50)	0.138	36 (28–44)	37 (30–42)	0.909
Sex			0.069			0.912
Male (%)	38 (76)	12 (55)		26 (70)	11 (69)
Female (%)	12 (24)	10 (45)		11 (30)	5 (31)
Headache	25 (50)	5 (23)	0.090	19 (51)	7 (44)	0.735
Stroke	37 (74)	17 (77)	0.768	29 (78)	10 (63)	0.229
Cognitive decline	14 (28)	11 (50)	0.071	10 (27)	7 (44)	0.231
Altered sensorium	5 (10)	8 (36)	0.007	6 (16)	4 (25)	0.467
Seizures	16 (32)	12 (55)	0.071	12 (32)	8 (50)	0.226
Aphasia	8 (16)	11 (50)	0.003	7 (44)	6 (37)	0.177
NIHSS ≥5	10 (20)	17 (77)	0.0001	8 (22)	10 (63)	0.004
EEG abnormal (%)	12 (24)	11 (50)	0.091	7 (37)	6 (60)	0.270
CSF abnormal (%)	29 (63)	16 (73)	0.430	18 (49)	16 (100)	0.001
Mean cell count per cu. Mm (SD)	11.6 (±20.1)	10.1 (±13.6)	0.747	12.67 (±22.26)	15.13 (±16.9)	0.698
Mean protein in g/dL (SD)	51.9 (±22.5)	58 (±28)	0.337	48.6 (±20.9)	70.1 (±18.9)	0.006
MRI
Multiple infarcts	38 (76)	16 (32)	0.768	30 (81)	11 (69)	0.325
Fazeka grade 2&3 changes	25 (64)	14 (67)	0.343	16 (47)	13 (81)	0.022
Multiple microbleeds in SWI (≥5)	11 (28)	8 (38)	0.396	6 (19)	7 (47)	0.082
Gadolinium enhancement	18 (40)	11 (58)	0.189	12 (32)	8 (50)	0.269
Abnormal VWI	15 (94)	5 (100)	0.632	10 (91)	4 (100)	0.733
Abnormal DSA	38 (77)	16 (76)	0.901	31 (84)	8 (50)	0.022
Biopsy-positive	15 (50)	7 (50)	1.000	9 (24)	9 (56)	0.074
Initial treatment (%)			0.141			0.409
Glucocorticoids alone	35 (70)	10 (45)		25 (68)	8 (50)
Glucocorticoids with cyclophosphamide	15 (30)	12 (55)		12 (32)	8 (50)
Maintenance therapy (%)			0.901			0.008
Corticosteroids	27 (54)	11 (50)		22 (59)	4 (25)
Steroid-sparing agents	23 (46)	10 (45)		15 (41)	12 (75)
Median time from symptom onset to treatment in days (IQR)	197 (45–643)	150 (87–679)	0.678	152 (34–591)	211 (101–686)	0.504
Time from first visit to our hospital to diagnosis in days, median (IQR)	16 (7–643)	9 (7–31)	0.311	18 (6–38)	13 (5–140)	0.500
Duration of follow-up in days, median (IQR)	1,074	314 (113–1,058)	0.010	1,172 (669–2,352)	943 (468–1723)	0.342
Off treatment at last visit (%)	15 (30)	4 (18)	0.295	15 (41)	2 (13)	0.045

Abbreviations: DSA = digital subtraction angiography; IQR = interquartile range; mRS = modified Rankin Scale; NIHSS = National Institute of Health Stroke Scale; SWI = susceptibility-weighted Imaging; VWI = vessel wall imaging.

A multivariate logistic regression analysis (not shown) found that NIHSS ≥5 (OR 10.53 [95% CI 2.8–38.4], p < 0.0001) and aphasia (OR 4.19 [95% CI 1.04–16.83], p = 0.043) were predictors of poor 6-month outcomes. Similarly, NIHSS ≥5 (OR 17.35 [95% CI 1.83–163.87], p = 0.013) predicted poor 1-year outcome, but an abnormal DSA (OR 0.057 [95% CI 0.004–0.876], p = 0.04) was predictive of good 1-year outcome.

Relapse was noted in 35 patients (48.6%), with 77% having a single relapse and 20% having 2 relapses. The presence of an abnormal CSF (77% vs 49%, p = 0.008), elevated CSF protein (mean 60.5 g/dL vs 47.6 g/dL, p = 0.028), positive biopsy (71% vs 13%, p < 0.0001), delay from symptom onset to initiation of treatment (median 350 days vs 97 days, p = 0.018), and usage of steroid-sparing agents in maintenance regimen (69% vs 25%, p < 0.0001) were all associated with relapse, while the presence of ≥5 microbleeds in SWI (43% vs 19%, p = 0.043) and abnormal DSA (95% vs 58%, p < 0.0001) were seen in those without a relapse, as given in Table 4. Multivariate logistic regression could not be performed for predictors of relapse due to small numbers.

Table 4.

Factors Associated With Relapse

Variables	No relapse N = 37	Relapse N = 35	p Value
Median age at onset (IQR)	40 (34–49)	36 (26–46)	0.156
Sex			0.504
Male (%)	27 (73)	23 (66)
Female (%)	10 (27)	12 (34)
Headache	12 (32)	18 (51)	0.125
Stroke	29 (78)	25 (71)	0.496
Cognitive decline	11 (30)	14 (40)	0.360
Altered sensorium	5 (14)	8 (23)	0.303
Seizures	13 (35)	15 (43)	0.502
Aphasia	11 (30)	8 (23)	0.508
NIHSS ≥5	12 (32)	15 (43)	0.361
EEG abnormal (%)	9 (24)	14 (40)	0.486
CSF abnormal (%)	18 (49)	27 (77)	0.008
Mean cell count, per cu.mm (SD)	10.1 (±20)	12.2 (±16)	0.648
Mean protein in g/dL (SD)	47.6 (±23)	60.5 (±27.3)	0.028
MRI
Multiple infarcts	26 (70)	28 (80)	0.341
Fazeka grade 2 and 3 changes	17 (52)	22 (65)	0.274
Multiple microbleeds in SWI (≥5)	13 (43)	6 (19)	0.043
Contrast enhancement	16 (50)	13 (41)	0.451
Abnormal VWI	14 (93)	6 (100)	1.000
Abnormal DSA	35 (95)	19 (58)	<0.0001
Biopsy-positive	2 (13)	20 (71)	<0.0001
Initial treatment (%)			0.217
Glucocorticoids alone	20 (54)	25 (71)
Glucocorticoids with cyclophosphamide	17 (46)	10 (29)
Maintenance therapy (%)			<0.0001
Corticosteroids	27 (75)	11 (31)
Steroid-sparing agents	9 (25)	24 (69)
Median time from symptom onset to treatment in days (IQR)	97 (34–557)	350 (113–1,043)	0.018
Time from first visit to our hospital to diagnosis in days, median (IQR)	10 (6–23)	22 (9–230)	0.066
Duration of follow-up in days, median (IQR)	660 (237–1,172)	1,125 (354–2,388)	0.065
Off treatment at last visit (%)	11 (30)	8 (23)	0.508

Abbreviations: DSA = digital subtraction angiography; IQR = interquartile range; mRS = modified Rankin Scale; NIHSS = National Institute of Health Stroke Scale; SWI = susceptibility-weighted imaging; VWI = vessel wall imaging.

Discussion

The subtypes of PACNS have different clinical presentations and treatment responses. Being a rare disease, data regarding the same are limited to few large cohorts.^8,9 Patients who are biopsy-positive, but angiogram-negative represent a distinct subtype of PACNS who behave differently as compared with the angiography-positive subtype. The small vessel vasculitis subtype had younger median age at onset as compared with the medium/large vessel vasculitis subtype (30.5 vs 40.5 years) similar to the French Vasculitis group⁵ but contrary to that reported by Mayo Clinic.² Similar to other cohorts, the small vessel vasculitis subtype often had an inflammatory CSF with greater delays in diagnosis due to the subacute presentation. By contrast, the medium/large vessel subtype often presented as a stroke, similar to those of the French cohort and had abnormal vessel wall enhancement on MRI. Recent systematic reviews have also reported that subacute presentation, cognitive deficits, and seizures were more often seen in small vessel vasculitis subtype as compared with the acute onset focal deficits seen in the medium/large vessel subtype.^10,11 The presence of gadolinium enhancement in MRI was higher in the small vessel vasculitis (53% vs 39%), as reported in the literature,^2,4,5,12 although we did not find it to be statistically significant, possibly because of the smaller numbers.

Although the Mayo clinic series found the small vessel vasculitis subtype to have a good outcome,² our outcomes at 6 months were similar to the French cohort who found no such association.⁵ This histopathologic subtype was predominantly lymphocytic in both our cohorts, as compared with the predominant granulomatous pattern seen in Mayo clinic series. There was a trend towards bad outcome in those with necrotizing and granulomatous vasculitis as compared with the lymphocytic vasculitis.⁴ Moreover, a recent update from the Mayo clinic cohort has shown the lymphocytic subtype to have improved modified Rankin scores and lesser mortality.⁸ However, the long-term outcome of the small vessel vasculitis group in our cohort was poor as compared with the medium/large vessel vasculitis subtype probably because of the increased relapse rates. Most of these patients received steroids alone as the induction treatment similar to other series^4,9 but subsequently required steroid-sparing agents in the maintenance regimen. At 1 year, the small vessel vasculitis group more often received maintenance therapy. The increased relapse rate in the small vessel subtype is also consistent with the results of the French vasculitis group. It is more possible that the medium/large vessel vasculitis group in our cohort often had more severe deficits at presentation due to a stroke and hence were treated upfront with combination immunotherapy.

A high-resolution MR-vessel wall imaging was performed in 21 patients in our cohort that was abnormal in about 95% of them. It was useful in the detection of medium vessel vasculitis subtype, distinguishing PACNS from mimics,¹³ and is another tool that could be used in patients in whom biopsy was not performed or negative.¹⁴ The distinction between small and medium/large vessel vasculitis is not always mutually exclusive but rather represents different ends of a spectrum because the disease process may affect all vessels. Similar patients were also reported by a German cohort.¹² where they had also shown the utility of high-resolution vessel wall imaging in differentiating the medium vessel vasculitis from the small vessel vasculitis. We had 4 such patients in our cohort who were both biopsy-positive and angiogram-positive, which may reflect the spread of the disease during its course.

When assessing the entire cohort, univariate analysis revealed baseline stroke severity, aphasia, and an altered sensorium at presentation to be associated with poor outcomes at 6 months, while baseline stroke severity, inflammatory CSF, Fazeka grade 2 and 3 small vessel ischemic changes, and absence of abnormal angiograms were associated with poor long-term outcomes at 1 year. Although previous studies have reported factors such as advanced age, headache, cognitive decline, infarcts in imaging, and lack of maintenance therapy to be associated with poor long-term outcomes,^8,9 we found only elevated baseline stroke severity to be associated with poor long-term outcomes, whereas an abnormal angiogram was often seen in those with good outcomes. Although the baseline stroke severity can influence the long-term outcomes due to severe deficits at baseline, a normal angiogram in a patient with PACNS may be falsely reassuring and lead to diagnostic delays. Our results emphasize that in patients with strong suspicion of small vessel PACNS, a negative angiogram should not dissuade one from pursuing more invasive diagnostic tests given the higher rate of relapse in this subtype to avert further disability.

Patients who relapsed had an inflammatory CSF and positive biopsy with delay in initiation of treatment and required the maintenance regimen with steroid-sparing immunosuppressants. Those who did not relapse had more microbleeds and an abnormal angiogram. It is possible that these features were clues leading to an earlier diagnosis and treatment in these patients. We did not find gadolinium enhancement on MRI to be associated with relapses, contrary to previous reports^8,9 probably due to the smaller numbers.

We found no differences in short-term outcome or relapse between those who received induction therapy with steroids alone vs those who received combination treatment. This was similar to few other cohorts^8,15,16 where no association of the treatment modality with outcomes was found. Those who relapsed required the maintenance regimen with another alternative immunosuppressive agent that is explained by their aggressive disease course. However, the French cohort had better outcomes in those who received maintenance immunosuppression as compared with those without it.¹⁷

It is important to note that we found poor long-term outcomes in the small vessel vasculitis group, which could be contributed by the increased relapse rates. The delay in initiation of treatment although significantly different between the small vessel and medium/large vessel vasculitis did not influence the outcome within each subtype. The delay in treatment initiation was however associated with increased relapses in the small vessel vasculitis subtype. Another factor to consider would be the existence of an indication bias because the small vessel vasculitis group had higher relapses and hence often received maintenance therapies which if not given could have resulted in further relapses and increased disability.

Treatment recommendations for PACNS are based on extrapolations from trials on anti-neutrophil cytoplasmic antibody (ANCA)–associated vasculitis,¹⁸ with no randomized controlled trials in PACNS to support the same. Conventionally, steroids are the mainstay with additional immunosuppression given with agents such as cyclophosphamide, azathioprine, mycophenolate, rituximab,^15,19,20 and even infliximab and etanercept.²¹

The median duration of follow-up in our cohort was approximately 22 months. At final follow-up, 55% had a good outcome, which is like the French cohort but less than other reported series.^8,15,16 This is probably due to factors such as delay in diagnosis, an aggressive disease course, and a higher relapse rate in our population.

The mortality rate of 11% in our cohort is significantly less than that of Mayo clinic (28%)⁸ and slightly higher than others (8–9%).^9,15,16 These patients had an aggressive disease course with infectious complications in 50% of them due to immunosuppression. Only 26% patients in the entire cohort were off any form of treatment at final follow-up, reflecting the disease severity. The differences in clinical profile and factors predicting the relapses and outcomes between other cohorts and ours serve to emphasize the unknown genetic and environmental influences between different patient populations and the varied presentation of PACNS.

The strengths of our study include a large cohort with clinical and imaging profile and long follow-up data including the treatment given their outcomes that contribute to understanding this rare disease further. The stringent diagnostic criteria including only angiography-proven or biopsy-proven cases with a long follow-up potentially excluded all mimics. The characterization of small vessel vasculitis would enable clinicians to recognize this phenotype and to pursue biopsy more often in such patients.

There are a few limitations to our study. This being a retrospective study, there could be missing data which might influence our findings. Since ours is a tertiary care institute, there could be referral bias and the cohort may not be truly representative of patients with PACNS in the community. MCB, the gold standard diagnostic test, was not performed in all patients. This entails the risk of inclusion of vasculopathy of other kinds, including monogenic inflammatory vasculopathies and noninflammatory vasculopathies. Most investigations were tailored for the individual patient as over the years, we noticed a trend toward ordering less invasive investigations to diagnose PACNS consistent with a systematic review.²²

Being a rare disease, protocols for the treatment of PACNS have evolved over the years in our institute, and currently, all patients with an aggressive presentation especially if they are biopsy-positive are treated with an induction regimen of 5 days of IV pulse steroids followed by oral steroids and concurrent immunosuppression with IV cyclophosphamide for 6 months if there are no contraindications. The maintenance regimen involves oral steroids if patients have achieved remission. Alternative oral immunosuppressants are added as steroid-sparing agents similar to the management of ANCA-associated vasculitis to minimize the adverse effects of long-term steroid exposure. When compared with our initial cohort of 45 patients,⁶ subsequently more patients were treated with the combination of steroids and cyclophosphamide, in part because of standardization of treatment policies.

PACNS is an aggressive disease that has significant morbidity especially if there are diagnostic or treatment delays. The small vessel vasculitis group is a distinct subtype of PACNS with a younger age at onset and presents less often as a stroke, with greater delays in diagnosis and initiation of treatment, poor long-term outcomes, and greater relapse rates. Recognition of this subtype is crucial so that aggressive immunosuppression can be instituted upfront in such patients.

Glossary

ANA

anti-nuclear antibody

ANCA

anti-neutrophil cytoplasmic antibody

DSA

digital subtraction angiography

MCB

meningocortical biopsy

mRS

modified Rankin Scale

NIHSS

National Institute of Health Stroke Scale

PACNS

primary angiitis of the CNS

Appendix. Authors

Name	Location	Contribution
Naveen K. Paramasivan, MBBS, MD	Department of Neurology, Sree Chitra Tirunal Institute for Medical Sciences and Technology	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data
Dev P. Sharma, MD, DM	Imaging Sciences and Interventional Radiology, Sree Chitra Tirunal Institute for Medical Sciences and Technology	Major role in the acquisition of data; analysis or interpretation of data
S.M. Krishna Mohan, MD	Department of Neurology, Sree Chitra Tirunal Institute for Medical Sciences and Technology	Major role in the acquisition of data
Soumya Sundaram, DM	Department of Neurology, Sree Chitra Tirunal Institute for Medical Sciences and Technology	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data
Sapna E. Sreedharan, MD, DM	Department of Neurology; Comprehensive Stroke Care Program, Sree Chitra Tirunal Institute for Medical Sciences and Technology	Drafting/revision of the manuscript for content, including medical writing for content; study concept or design
P Sankara Sarma, PhD	Achutha Menon Centre for Health Science Studies, Sree Chitra Tirunal Institute for Medical Sciences and Technology	Analysis or interpretation of data
PN Sylaja, MD	Department of Neurology; Comprehensive Stroke Care Program, Sree Chitra Tirunal Institute for Medical Sciences and Technology	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data

Study Funding

The authors report no targeted funding.

Disclosure

The authors report no relevant disclosures. Go to Neurology.org/NN for full disclosures.