Primary Central Nervous System Vasculitis: Seeing the Invisible

For postgraduate trainees and practicing consultants in internal medicine, neurology, and rheumatology

Dr Neeraj Manikath

 

 

 

KEY POINTS

● Primary CNS vasculitis (PCNSV) is rare (≈2.4 cases per million person-years), affects middle-aged men slightly more than women, and always announces itself on MRI and usually in the CSF — but almost never in the ESR, CRP, or ANCA.

● Angiographic "beading" is a clinical honey trap: reversible cerebral vasoconstriction syndrome (RCVS) looks identical on catheter angiography. The angiogram alone can no longer make this diagnosis.

● The two most important exclusion diagnoses — intravascular lymphoma and infection — can be transiently masked by steroids. Biopsy before you immunosuppress, whenever the patient allows.

● Modern management is stratified and "treat-to-target": glucocorticoids plus cyclophosphamide for severe disease, plus mycophenolate mofetil for less severe disease, with 18–24 months of maintenance.

● Roughly 80% achieve remission today; relapse (~25–30%) is the rule to plan for, not the exception.

 

 

 

Abstract

 

Primary central nervous system vasculitis (PCNSV) — historically called primary angiitis of the CNS (PACNS) — is an inflammatory, vasculitic disease confined to the brain, spinal cord, and leptomeninges, with no systemic involvement. It is simultaneously one of the most feared and most over-diagnosed conditions in neurology. This review synthesises the contemporary evidence base for the general physician: a clinically distilled pathophysiology, the bedside observations that separate the great from the good, the mimics that kill (both from the disease and from its treatment), a practical diagnostic algorithm, and a stratified, target-driven management framework with explicit escalation thresholds. A mnemonic and summary tables are provided for the ward.

 

 

 

1. Introduction: The Tuesday That Teaches

 

Let me begin with a woman I have never fully forgotten.

 

Mrs. KV, 58, a schoolteacher, arrives with ten weeks of "fog". Word-finding difficulty first noticed by her students, then episodic throbbing headaches, then — three days ago — two witnessed focal seizures affecting the right arm. Her husband insists she is "not herself": irritable, slow, occasionally losing the thread of a sentence.

 

Her examination is disarmingly benign: mild anomia, subtle right pronator drift, nothing else. No fever. No rash, no arthritis, no weight loss, no night sweats. ESR 8 mm/h. CRP 2 mg/L. ANCA negative. MRI shows bilateral cortical–subcortical FLAIR lesions, a subacute right frontal infarct, a tiny old left occipital lacune, and faint leptomeningeal enhancement. CSF: 14 lymphocytes/µL, protein 0.88 g/L. Catheter angiography: elegant "beading" of the right M2 and left A2 segments.

 

The ward team, energised, declares: "PACNS — start 1 g methylprednisolone."

 

Pause. That same angiogram, in a 42-year-old woman six weeks postpartum with thunderclap headaches and near-normal CSF, would be RCVS. That same MRI-and-decline picture, in a 66-year-old man with a modestly elevated LDH, might be intravascular lymphoma — a malignancy that melts transiently on steroids and kills when we relax. The angiogram is the least trustworthy witness in the courtroom; it testifies with equal eloquence for three different defendants.

 

Mrs. KV had a navigated brain biopsy (before the pulse steroids — remember that sequencing): granulomatous vasculitis. She received steroids plus cyclophosphamide and returned to her classroom over six months, with a residual naming deficit.

 

Now the epidemiological hook: in Olmsted County, Minnesota — the world's most carefully counted population — PCNSV occurs at ~2.4 cases per million person-years [3]. That means a busy general physician may meet this disease once in an entire career. But the patient who is almost mislabelled with it — or whose true PCNSV is dismissed as "vascular dementia" or "psychiatric" — walks past your clinic every single week.

 

That asymmetry — a disease you will rarely see, and rarely be wrong about at the right time — is why this review exists.

 

 

 

2. Pathophysiology: Only What Changes Your Decisions

 

You do not need the cytokine biology to manage PCNSV; you need four consequences of it.

 

1. Vessel size determines everything diagnostic. PCNSV inflames small and medium-sized leptomeningeal and parenchymal arteries. Medium vessels (M1, A1, basilar) are visible to catheter angiography; arterioles under ~200 µm are not. This single fact explains why ~40–60% of biopsy-proven cases have a normal angiogram [2,7] — "angiography-negative PCNSV" is small-vessel disease, not imagination.

 

2. The inflammation is patchy — "skip lesions." Diseased segments alternate with normal vessel. This is why brain biopsy has a sensitivity of only ~50–75% even in expert hands [6,8]: the surgeon can miss the lesion like a police search missing one house on the street. Target the enhancing lesion and include leptomeninges.

 

3. The inflammation is confined — and quiet — systemically. There is no circulating inflammatory storm, which is why constitutional symptoms are mild or absent and ESR/CRP are usually normal [2]. A normal CRP in a suspected vasculitis is not a red flag against the diagnosis here — it is characteristic. The flip side: fever, weight loss, and a CRP in the hundreds should redirect you toward infection or systemic vasculitis.

 

4. Histology does not dictate therapy. Granulomatous (~50–60%), lymphocytic, and necrotizing patterns coexist and do not reliably stratify prognosis or treatment [4,5]. What does matter is what the pathologist finds instead of vasculitis — lymphoma, infection, amyloid — which is why the biopsy's greatest value is often what it excludes.

 

One cousin deserves separate billing: amyloid-β–related angiitis (ABRA) — CAA with a superimposed vasculitic inflammatory response, seen in older patients (more often women), with rapid cognitive decline and lobar haemorrhagic markers. Mechanistically it is an immune reaction to vascular amyloid, and it behaves differently from PCNSV (more on it in Section 5).

 

 

 

3. Clinical Pearls 🪙 — Counterintuitive, High-Yield, Non-Negotiable

 

🪙 Pearl 1 — The thunderclap veto. PCNSV headache is subacute, progressive, dull, and "confusing" — it is not thunderclap. A thunderclap headache with angiographic beading and near-normal CSF is RCVS until proven otherwise [9,10]. This one sentence prevents more iatrogenic harm in this field than any other.

 

🪙 Pearl 2 — Normal ESR/CRP does not reassure you away from PCNSV. This is the opposite of giant cell arteritis logic. In the Mayo cohort, mean ESR was unremarkable [2]. If systemic markers are high, your prior should shift away from PCNSV toward infection or systemic vasculitis.

 

🪙 Pearl 3 — The gender inversion. Most autoimmune diseases prefer women; PCNSV mildly prefers men (~2:1 in many series), in the sixth decade. A middle-aged man with months of cognitive slippage and multifocal infarcts of different ages on MRI deserves this diagnosis on your differential.

 

🪙 Pearl 4 — A normal MRI is a near-perfect veto. MRI is abnormal in ~97–100% of cases [2,5]. Normal MRI + normal CSF ≈ PCNSV excluded. Stop, rethink, and save the patient from speculative immunosuppression.

 

🪙 Pearl 5 — Strokes of different ages. The signature MRI clue is infarcts in multiple vascular territories at different stages — an acute DWI-bright lesion beside a chronic FLAIR cavitary one. Monophasic large-artery atherosclerosis doesn't do this; vasculitis and cardioembolism do.

 

🪙 Pearl 6 — PCNSV rarely bleeds. Lobar haemorrhage or numerous cortical microbleeds point to CAA/ABRA, not classical PCNSV. Subarachnoid blood on convexity, again, whispers RCVS [9,10].

 

🪙 Pearl 7 — Look at the skin, the joints, the urine. Rash, arthritis, haematuria, or mononeuritis multiplex means the vasculitis is not "primary" to the CNS. Systemic involvement reclassifies everything — ANCA vasculitis, Behçet, PAN — with different treatments.

 

🪙 Pearl 8 — The tempo test. PCNSV evolves over weeks to months. Acute catastrophic stroke-in-evolution in a young patient is more likely arterial dissection, embolism, thrombophilia, or RCVS-related late ischemia. PCNSV is a marathoner, not a sprinter.

 

 

 

4. Oysters 🦪 — The Hidden Gems Most Clinicians Miss

 

🦪 Oyster 1 — The tumour that isn't a tumour. A striking minority of PCNSV presents as a mass lesion — enhancing, oedematous, biopsy-planned as "neoplasm" [5]. Neuropathology then reveals vasculitis. The critical corollary: if the pre-biopsy differential includes lymphoma, withhold steroids until tissue is obtained — glucocorticoids can render lymphoma undiagnosable (the "ghost tumour") while transiently shrinking it.

 

🦪 Oyster 2 — Intravascular lymphoma (IVL): the great imitator. Subacute cognitive decline, multifocal infarcts/lesions, leptomeningeal enhancement, modest CSF pleocytosis — IVL mirrors PCNSV nearly perfectly. Clues: age >60, elevated LDH, skin nodules, anaemia/thrombocytopenia. IVL responds transiently to steroids, inducing false comfort, then explodes. If the picture is atypical or steroid response is partial/relapsing, re-biopsy and search the skin before concluding PCNSV.

 

🦪 Oyster 3 — Steroids cause the syndrome they are being given for. High-dose glucocorticoids are themselves a trigger for RCVS [12]. A patient "treated for vasculitis" who develops a thunderclap headache on day 4 of methylprednisolone has not necessarily relapsed — you may have created a vasoconstriction syndrome. Do not reflexively escalate immunosuppression; re-image and re-think.

 

🦪 Oyster 4 — ABRA, the common cousin. In an older woman with rapid cognitive decline, lobar microbleeds, leptomeningeal enhancement, and vasculitic angiography — think amyloid-β–related angiitis [4,5]. It is steroid-responsive more often than classical PCNSV, cyclophosphamide is less routinely needed, and anticoagulation is hazardous given CAA. You will not think of it unless you know it.

 

🦪 Oyster 5 — The pharmacy history. Systematically ask about SSRIs, triptans, decongestants (phenylephrine), cocaine, amphetamines, heavy cannabis, and the postpartum state — all RCVS triggers [9,10]. Two minutes of medication reconciliation can prevent six weeks of immunosuppression.

 

🦪 Oyster 6 — VZV vasculopathy in the elderly. A shingle-free elderly patient can still have varicella-zoster vasculopathy — small-vessel strokes, basal ganglia predilection, mimicry of PCNSV — weeks after the rash (which the patient and family may have forgotten). CSF VZV PCR and VZV IgG (with parallel serum) should be routine in the workup. Treating "vasculitis" with steroids alone in VZV vasculopathy is harmful; acyclovir is the answer.

 

🦪 Oyster 7 — Susac's syndrome hides in the eye and ear. Encephalopathy plus branch retinal artery occlusions (seen only on fluorescein angiography) and sensorineural hearing loss, with spherical "snowball" lesions in the central corpus callosum. Steroid-responsive but requires IVIG/rituximab thinking. Fundoscopy and audiometry are part of the "vasculitis" examination, not optional extras.

 

🦪 Oyster 8 — Angiogram-negative is not vasculitis-negative. If clinical suspicion remains high after a normal DSA, the disease is small-vessel by definition [7]. Do not stop at the angiogram; biopsy. Conversely, if the angiogram is positive, the CSF normal, and the story thunderclap — stop before the biopsy and diagnose RCVS.

 

🦪 Oyster 9 — "Probable angiographic PACNS" is a concept we have retired. Historical series contain many "probable" cases that later proved to be RCVS, atherosclerotic vasculopathy, or infection. Modern practice demands histology, or a compelling non-invasive picture plus rigorous mimic exclusion — and time as an investigator.

 

🦪 Oyster 10 — Remote radiation, remote cocaine, remote everything. Radiotherapy decades ago causes a slowly progressive vasculopathy indistinguishable on angiography from PCNSV. Ask the question; read the old records.

 

 

 

5. Diagnostic Nuances: What Separates Good from Great

 

5.1 The history — five questions that matter more than the antibody panel

 

1. Tempo: weeks-to-months decline (PCNSV) vs. thunderclap (RCVS) vs. stuttering over years (CADASIL, atherosclerosis)?

2. Headache character: dull/progressive (PCNSV) vs. thunderclap in showers (RCVS)?

3. Constitutional symptoms: their absence supports PCNSV; fever/weight loss argues for infection or systemic disease.

4. Drug and obstetric history: vasoactive triggers, postpartum status, remote radiation.

5. The family history: early strokes, dementia, migraine — CADASIL (anterior temporal pole lesions, NOTCH3), Fabry, moyamoya.

 

5.2 The examination — three things beyond the neuro screen

 

● Fundoscopy ± fluorescein angiography: branch retinal artery occlusions → Susac; retinal vasculitis → Behçet or systemic disease (both argue against PCNSV).

● Hearing: unexplained SNHL → Susac.

● The "systemic sweep": skin, nails (splinter haemorrhages → endocarditis), joints, urine dipstick, pulses. PCNSV is silent outside the neuraxis.

 

5.3 The CSF — a gatekeeper, not a formality

 

Abnormal in ~90% of cases [2]: mild lymphocytic pleocytosis (usually <50 cells/µL) and protein elevation (usually <1.5 g/L). Oligoclonal bands may appear — do not anchor on multiple sclerosis. A striking neutrophilic pleocytosis or protein >2.5 g/L should make you hunt for infection. CSF plus MRI both normal = PCNSV effectively excluded; this combination is the single most useful negative test in the field.

 

5.4 The MRI — read it yourself

 

You are looking for: bilateral, multifocal, supratentorial, cortical and subcortical lesions; DWI infarcts of differing ages; deep grey matter involvement; leptomeningeal enhancement (~30%); and occasionally mass-like lesions. Lobar microbleeds (SWI/T2*) reorient you to CAA/ABRA. Diffuse white matter disease in a younger patient with migraine/family history is CADASIL territory.

 

5.5 The angiogram — handle with gloves

 

Catheter DSA remains more sensitive than CTA/MRA for beading, but its specificity for vasculitis is poor (~30%) [4,9,11]. RCVS, atherosclerosis, infection, and even technical artefact produce "beading." If you must interpret: beading in multiple vessels/territories with a compatible MRI and CSF raises the posterior probability; in isolation, it proves nothing. A high-quality 6-vessel 3D rotational study by an experienced neuroradiologist is worth the logistics.

 

5.6 Vessel-wall MRI — the promising adjutant

 

High-resolution intracranial vessel-wall MRI (pre- and post-contrast, dedicated protocols) can show multifocal, circumferential, often nodular wall enhancement in PCNSV, versus eccentric enhancement in atherosclerosis [13]. Caution: early RCVS can also enhance the wall, so VWMRI is adjunctive, not adjudicative — best used for pattern recognition and, prospectively, treatment monitoring.

 

5.7 The biopsy — still the referee

 

Indications: suspicion persisting after non-invasive workup, angiogram-negative disease, diagnostic uncertainty where lymphoma/infection are live possibilities, or before committing to long-term immunosuppression. Target cortex + leptomeninges at/near an enhancing lesion (stereotactic navigation helps). Sensitivity ~50–75%, specificity high; serious complication rates in modern series are low (1–2%, up to ~6% historically) [6,8]. Tell the pathologist about steroid exposure — treated lymphoma and partially treated vasculitis are pathologic mirages.

 

5.8 The mimic showdown

 

Feature	PCNSV	RCVS	ABRA	Intravascular lymphoma	Neurosarcoid	VZV/TB vasculopathy
Typical patient	Man ~50s	Woman ~40s; postpartum/triggers	Woman ~65+	Man/woman ~65+	Any adult	Elderly / immunosuppressed
Headache	Dull, progressive	Thunderclap, recurrent	Variable	Variable	± chronic meningitic	Variable
Tempo	Weeks–months	Peaks ~3 wks; reverses by 3 mo	Months (fast)	Weeks–months	Months–years	Weeks
CSF	Lymphocytes <50, protein <1.5	Normal/mild	Abnormal	Abnormal; LDH high	Lymphocytes, protein high	PCR/serology diagnostic
MRI signature	Multifocal infarcts of different ages, leptomeningeal enhancement	Convexity SAH, PRES, watershed infarcts late	Lobar microbleeds, CAA	Multifocal lesions, enhancement	Basal leptomeningeal, cranial nerves, spinal cord	Basal ganglia (VZV); basal meninges + hydrocephalus (TB)
DSA	Beading ± normal	Beading, reversible	Beading ± normal	Often normal	Often normal	Beading, progressive
The tell-tale	Biopsy: vasculitis, no amyloid/infection	8–12-week reversibility; triggers	Amyloid on biopsy	LDH, sIL-2R; CD20 on tissue	ACE, chest CT, extraneural signs	CSF PCR/IGG/AFB
Fatal mistake	Over-diagnosing from DSA	Starting steroids	Anticoagulating CAA	"Treating" with steroids and relaxing	Steroids without infection screen	Steroids alone

 

 

 

 

6. State of the Art: What Has Actually Changed Practice

 

1. The angiogram-only diagnosis is dead. The historical category of "probable PACNS" — diagnosed on angiography alone — has been dismantled by cohorts showing that most such patients have RCVS, atherosclerotic vasculopathy, or other conditions [4,9,10,11]. Modern diagnosis requires clinical + CSF + imaging coherence, with histology as the gold standard when feasible.

 

2. Treat-to-target is the new standard. Salvarani and colleagues demonstrated that a structured strategy — glucocorticoids plus cyclophosphamide for severe disease, plus mycophenolate mofetil for less severe disease — achieved remission in ~80% of patients, with relapse in ~25–30% and acceptable toxicity [15]. Define the target (neurological stability/improvement, objective mRS and imaging) and document it.

 

3. Mycophenolate mofetil entered the algorithm. The Mayo group's MMF experience established steroids + MMF as a legitimate induction choice for less severe disease, sparing cyclophosphamide's toxicity [14].

 

4. Rituximab is knocking. Small series and expert practice now deploy rituximab for refractory or relapsing disease — a logical CD20 approach (and one that would also cover IVL if the diagnosis were wrong). No randomised data; enthusiasm is tempered by evidence.

 

5. Steroid-associated RCVS changed early decision-making. Singhal's analyses showed glucocorticoids are not only ineffective in RCVS but associated with its development and prolongation [12] — reshaping the "just give steroids while we wait" reflex.

 

6. Vessel-wall MRI matured from novelty to standardised tool [13], increasingly used for baseline characterisation and follow-up — though it still cannot reliably separate early RCVS from PCNSV.

 

7. Honest frontier: there are no randomised trials in adult PCNSV, and none are likely (numbers preclude them). International registries and consensus statements are consolidating practice. Anecdotal JAK-inhibitor and cytokine-directed reports exist; treat them as horizons, not standards.

 

 

 

7. Management Intricacies: Drugs, Doses, Sequencing, Pitfalls

 

7.1 First, stratify severity

 

● Severe: progressive encephalopathy, disabling or multifocal deficits, strokes in evolution, brainstem/cord involvement, or biopsy-proven necrotizing disease → pulse methylprednisolone + cyclophosphamide.

● Less severe: stable deficits, isolated cognitive change, small-vessel pattern without progression → glucocorticoids + mycophenolate [14,15].

● Diagnostic limbo (suspected RCVS or unproven): no immunosuppression until the picture resolves or tissue is obtained.

 

7.2 The induction regimen (severe disease)

 

1. Methylprednisolone 1 g IV daily × 3–5 days → prednisone 1 mg/kg/day (usual maximum 60–80 mg), tapered over 4–6 months (roughly halve every 2–4 weeks once improving).

2. Cyclophosphamide 750 mg/m² IV monthly × 6 doses, adjusted by the day 10–14 leukocyte nadir (reduce dose ~25% for nadir <2,000/µL; omit if severe sepsis/cytopenias). Mesna and generous hydration for cystitis prophylaxis; cumulative dose ~6–10 g.

3. PJP prophylaxis is non-negotiable: co-trimoxazole 480–960 mg daily (or DS three times weekly), or atovaquone/dapsone if intolerant.

 

7.3 Induction for less severe disease

 

● Prednisone 1 mg/kg/day with the same taper, plus MMF 2–3 g/day in two divided doses [14]. Watch GI intolerance and cytopenias.

 

7.4 Maintenance — where relapses are won or lost

 

● Continue MMF 2–3 g/day (or azathioprine 2–3 mg/kg/day after TPMT/NUDT15 screening) for at least 18–24 months of remission. The treat-to-target data show relapse clustering in the first 2–3 years [15]; maintenance is not optional homework, it is the treatment.

● For relapse: confirm it is relapse (new deficits + new imaging/CSF findings — and exclude infection, PML, and steroid-associated RCVS first), then re-induce (cyclophosphamide if unused, or rituximab) and consider extending maintenance.

 

7.5 Adjuncts and supportive care

 

● Aspirin 75–100 mg if infarction has occurred (evidence thin, practice near-universal).

● Vascular risk control, blood pressure, lipids, glycaemia.

● Bone protection (calcium/vitamin D; bisphosphonate for prolonged steroids), PPI during high-dose steroids, inactivated vaccinations (influenza, pneumococcal, COVID-19) before or during therapy; screen HBV/HCV/HIV and IGRA before immunosuppression and rituximab.

● Fertility discussion before the first cyclophosphamide pulse — sperm or oocyte banking takes days you must plan for. MMF is teratogenic: explicit contraception counselling.

 

7.6 Drug cheat sheet

 

Agent	Role	Typical dosing	Monitor	Pitfalls
Methylprednisolone → prednisone	Induction backbone	1 g IV ×3–5 d → 1 mg/kg/day, taper 4–6 mo	Glucose, BP, mood, bone	Can trigger RCVS; masks lymphoma/infection
Cyclophosphamide	Severe induction	750 mg/m² IV monthly ×6	CBC day 10–14 nadir; urinalysis	Infertility, cystitis, malignancy, infections
Mycophenolate mofetil	Less-severe induction; maintenance	2–3 g/day divided	CBC, LFTs	Teratogenicity; GI intolerance
Azathioprine	Maintenance alternative	2–3 mg/kg/day	TPMT/NUDT15, CBC, LFTs	Slow onset; interaction with allopurinol
Rituximab	Refractory/relapsing	375 mg/m² weekly ×4, or 1 g ×2 fortnightly	Immunoglobulins, HBV screen	Hypogammaglobulinaemia; serum-sickness reactions
Co-trimoxazole	PJP prophylaxis	480–960 mg/day (or DS 3×/wk)	CBC, renal	Hyperkalaemia with ACEi
Nimodipine (RCVS context)	Headache/vasospasm	30–60 mg q4–6h ×3–4 wks	BP	Hypotension; no outcome benefit proven

 

 

⚡ The sequencing commandment: tissue before toxicity. Outside fulminant presentations, obtain biopsy (or at minimum a completed mimic-exclusion workup) before the first high-dose steroid — because steroids can sterilise lymphoma, quiet infection into illegibility, and provoke RCVS, all while appearing to "work."

 

 

 

8. When to Escalate, When to Watch: Decision Thresholds with Reasoning

 

Escalate within hours (pulse steroids ± cyclophosphamide, ICU):

● Rapidly progressive encephalopathy or declining GCS, status epilepticus, new brainstem or spinal cord deficits, crescendo strokes with DWI progression, or biopsy-proven necrotizing vasculitis. Reasoning: here, the disease's arithmetic of irreversible infarction beats the diagnostic purity of waiting; treat, and biopsy-informed adjustments follow.

 

Escalate within days (biopsy first if not yet done):

● Subacute progression over weeks without fulminancy, where lymphoma or infection remains genuinely possible. Reasoning: PCNSV deteriorates over weeks, not hours; the diagnostic window is real, and a "failed" biopsy costs less than a missed lymphoma.

 

Watch — do not treat (structured observation):

● Stable deficits, mild cognitive complaints, normal CSF, or a picture favouring RCVS (thunderclap headaches, triggers, near-normal CSF, convexity SAH). Repeat clinical assessment and MRI in 2–4 weeks; repeat DSA at 8–12 weeks — complete reversal = RCVS; persistence/progression = reconsider PCNSV [9,10,11].

 

The 48-hour rule:

● If a patient deteriorates within the first days of high-dose steroids, your default is not "under-treated vasculitis." Consider — in order — wrong diagnosis (IVL, infection), steroid-associated RCVS, or a metabolic/infectious complication. Re-image, repeat CSF, and re-think before stacking more immunosuppression.

 

Relapse thresholds: new or worsening neurological deficits plus new imaging or CSF abnormality, after documented stability — ideally confirmed objectively (mRS change, radiographic progression), because steroid-related fatigue and headache are notorious impostors of true relapse.

 

 

 

9. The One-Page Summary: Mnemonic + Tables

 

The PACNS Mnemonic 🧠

 

Letter	Remember
P	Persistent subacute decline + headache, Paucity of systemic signs, PJP prophylaxis when treating
A	Angiographic beading ≠ vasculitis (RCVS!); Almost always abnormal MRI; ANCA/ESR usually normal
C	CSF abnormal in ~90% (lymphocytes <50, protein <1.5); CSF + MRI normal excludes; CYC + steroids for severe
N	No thunderclap headache; Necrotizing skip lesions → biopsy sensitivity ~50–75%; Never biopsy-empty-handed — exclude lymphoma/infection first
S	Small-vessel disease = angiogram-negative → biopsy; Steroids can cause RCVS; Second agent always, maintenance 18–24 months

 

 

The five-step algorithm, compressed

 

1. Gate: MRI + CSF. Both normal → stop.

2. Exclude: infection, lymphoma, drugs, RCVS, sarcoid, Susac, atherosclerosis, CADASIL (Table in §5.8).

3. Confirm: pattern coherence ± VWMRI ± DSA; biopsy when feasible and diagnosis-changing.

4. Treat-to-target: stratify severity → GC + CYC (severe) or GC + MMF (less severe) → maintenance ≥18–24 months.

5. Surveil: MRI at 3, 6, 12 months, then annually; mRS + MoCA each visit; relapse = new deficits + new objective abnormality.

 

 

 

10. Closing Remarks

 

PCNSV rewards two clinician virtues in tension: suspicion without sedation, and conviction without hurry. The disease is rare enough that most of us will see it once; its mimics are common enough that we will treat them wrongly many times unless the algorithm above is second nature. Let the CSF and the MRI be your gatekeepers, let the angiogram testify but never convict, let the biopsy speak before the steroids deafen it — and when the diagnosis is finally earned, treat boldly, maintain patiently, and follow with intent.

 

Mrs. KV is teaching again. Her residual anomia surfaces only when she can't find "hippocampus" mid-lesson — which, she tells me, is a fine irony to share with her students.

 

 

 

References

 

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(1):20–39.

2. Salvarani C, Brown RD Jr, Calamia KT, et al. Primary central nervous system vasculitis: analysis of 101 patients. Ann Neurol. 2007;62(5):442–451.

3. Salvarani C, Brown RD Jr, Christianson TJH, et al. Incidence and prevalence of primary central nervous system vasculitis in Olmsted County, Minnesota. Mayo Clin Proc. 2015;90(10):1396–1400.

4. Hajj-Ali RA, Singhal AB, Benseler S, Molloy E, Calabrese LH. Primary angiitis of the CNS. Lancet Neurol. 2011;10(6):561–572.

5. Salvarani C, Brown RD Jr. Adult primary central nervous system vasculitis. Lancet. 2012;380(9843):767–777.

6. Miller DV, Salvarani C, Hunder GG, et al. Biopsy findings in primary angiitis of the central nervous system. Neurology. 2009;72(1):34–39.

7. Salvarani C, Brown RD Jr, Calamia KT, et al. Angiography-negative primary central nervous system vasculitis: a phenomenon involving the small vessels only? J Rheumatol. 2013;40(4):383–389.

8. de Boysson H, Zuber M, Naggara O, et al. Primary angiitis of the central nervous system: description of the first four-tier biopsy-based prospective cohort. Cerebrovasc Dis. 2014;37(4):299–306.

9. Ducros A. Reversible cerebral vasoconstriction syndrome. Lancet Neurol. 2012;11(10):906–917.

10. Singhal AB, Hajj-Ali RA, Topcuoglu MA, et al. Reversible cerebral vasoconstriction syndromes: analysis of prospective series and proposed new diagnostic scheme. JAMA Neurol. 2019;76(1):89–101.

11. de Boysson H, Lamy C, Mas JL, et al. Primary angiitis of the central nervous system and reversible cerebral vasoconstriction syndrome: a comparative study. Stroke. 2015;46(6):1487–1491.

12. Singhal AB, Topcuoglu MA. Glucocorticoid-associated reversible cerebral vasoconstriction syndrome. Neurology. 2020;95(2):152–161.

13. Mandell DM, Mossa-Basha M; American Society of Neuroradiology Vessel Wall Imaging Study Group. Intracranial vessel wall MRI: principles and expert consensus recommendations of the American Society of Neuroradiology. AJNR Am J Neuroradiol. 2017;38(2):218–229.

14. Salvarani C, Brown RD Jr, Christianson TJH, et al. Mycophenolate mofetil in primary central nervous system vasculitis. Semin Arthritis Rheum. 2015;45(3):333–336.

15. Salvarani C, Brown RD Jr, Christianson TJH, et al. Treat-to-target approach in primary CNS vasculitis: an outcome study. Arthritis Rheumatol. 2018;70(10):1668–1676.

 

 

 

Conflicts of interest: none declared. This review is for educational purposes; individual patient decisions should integrate local expertise, drug licensing, and the clinical context.