Stroke Snapshot: Reversible Cerebral Vasoconstriction Syndrome

History and Nomenclature

Reversible cerebral vasoconstriction syndrome (RCVS) is a clinicoradiographic syndrome characterized by abrupt onset of severe headache with dynamic segmental cerebral vasoconstriction that reverses within 3 months.^1,2 The syndrome has a mean age of onset in middle age and a slight female preponderance.^2,3 People with RCVS often present with thunderclap headache (TCH), which may recur over a period of 1 to 2 weeks. Diagnosis is made on clinical and radiographic grounds and requires a high index of suspicion because of the dynamic and transient nature of radiographic findings. Although long-term prognosis is generally favorable, complications include transient or permanent neurologic deficits caused by ischemic and hemorrhagic strokes, convexity subarachnoid hemorrhage, (SAH) or seizures.

In 1988, Call and Fleming et al described 4 previously healthy young people (ages 19-47) with acute-onset severe headache and transient reversible cerebral vasoconstriction.⁴ Along with 12 previously published cases that clinically and radiographically resembled the 4 described, all had the defining features of sudden-onset severe headache associated with reversible cerebral vasoconstriction. This syndrome became known as the Call-Fleming syndrome, which has been called by various names (Box 1) since been unified under the term RCVS.²

Epidemiology

RCVS most commonly, but not exclusively, affects persons age 20 to 50 with peak occurrence at age 42 and more commonly affects women.^1,3,5-7 Although the precise prevalence and incidence of RCVS are unknown, it has been suggested that RCVS is not particularly rare based on the ease of recruitment into prospective studies at tertiary care referral centers.^1,3,7 A US population–based study using administrative claims data found the incidence of RCVS requiring hospitalization was 3 per million per year and the incidence in women was 4 times higher than men.⁸ Among 2 Asian cohorts, there was a slightly stronger female predominance,^9,10 and the age of presentation tends to be 10 years younger for men than for women.^3,7 There is uncertainty as to whether differences between cohorts reflects genetic difference, disparate rates of exposure to vasoactive triggers, differences in inclusion and exclusion criteria, or recruitment biases.¹

Diagnosis

RCVS diagnosis is made by recognizing typical clinical symptoms and course of illness, finding vasoconstriction on cerebral angiography, and excluding other causes. Detailed history taking is crucial to identify TCH and possible triggers. Diagnostic criteria for RCVS proposed by Calabrese et al in 2007² have been slightly modified (Box 2) and incorporated into the International Classification of Headache Disorders, 3rd edition.¹¹

Clinical Features

Headache

The most common clinical presentation of RCVS is TCH, defined as abrupt onset of a severe headache reaching peak intensity in less than 1 minute. Differential diagnosis for TCH is broad, including aneurysmal SAH, cerebral venous thrombosis, and pituitary apoplexy, which should all be ruled out.

TCH was the presenting symptom in 85% to 99% of RCVS in several reported case series at tertiary referral centers.^3,7,9,10 Among people presenting to the emergency department with TCH, 8.8% were eventually diagnosed with RCVS in a single-center study.¹² The headache is often bilateral and posterior but can be unilateral and diffuse, and may be associated with photophobia, phonophobia, and nausea and vomiting. The severe headache is usually short-lived, lasting minutes to hours. Recurrent TCH over a 2-to 4-week period is common in RCVS, occurring in 82% to 94% of people with RCVS in US and French cohorts, respectively.^3,7 In the presence of cerebral vasoconstriction and absence of aneurysmal SAH, recurrent TCH is highly specific for RCVS with a reported specificity of 99%.¹³ In a Korean series of 138 people with RCVS, however, 10% presented with headache of more gradual onset and only 76% had recurrent TCH.⁹ In an Italian cohort of 102 white individuals with RCVS, TCH was present in only 67%, with more than one-third presenting with headache of more gradual onset.¹⁴ Associated neck pain should prompt evaluation for cervical artery dissection, which has been described in association with RCVS.^15,16 A moderate persistent headache in between attacks has been reported. Absence of headache at onset is very rare.^1,13

A history of migraine is common in persons with RCVS. Upon careful questioning, most describe the headache associated with RCVS to be different than their usual migraines.³ Vasoactive medications such as triptans and ergots can trigger or aggravate RCVS and people diagnosed with RCVS should be advised to avoid use of vasoactive medications.

Hemorrhagic Complications

Hemorrhagic complications, most frequently convexity SAH, tend to occur early and most commonly in the first week of illness and are often mild.^2,17-20 Convexity SAH may be asymptomatic and unilateral or bilateral.^1,18,19,21 Intraparenchymal hemorrhage and, more rarely, subdural hemorrhage may also occur. The frequency of hemorrhage is reported to be 34% to 43% of patients.^17-19 Those with hemorrhagic RCVS have a higher rate of persistent focal deficits, cerebral infarctions, and posterior reversible encephalopathy syndrome (PRES) compared with those without hemorrhagic complications.^17,18

Angiographic findings appear similar in both hemorrhagic and nonhemorrhagic RCVS. Hemorrhagic RCVS tends to occur in older and more commonly female persons. In a French cohort, a history of migraine was found to be an independent risk factor for hemorrhagic complications,¹⁸ although this has not been corroborated in other cohorts.^17,19 A US study did not find any association of age, sex, migraine, smoking status, peripartum state, acute or chronic arterial hypertension, or use of vasoactive medications with hemorrhagic complications.¹⁷

Ischemic Complications

Cerebral infarctions tend to occur later in the course of illness compared with hemorrhagic complications, usually 1 to 2 weeks after the initial TCH and often after recurrent TCHs have resolved.^3,10,17,19 Focal neurologic deficits were present in 8% to 43% of people with RCVS across 3 large case series,^3,7,10 and ischemic stroke was diagnosed, up to 4 weeks after onset, in 16% of people with RCVS in a US population–based study.⁸ The variability and relatively high rates of neurologic complications may be due to recruitment bias in case series, with more severe cases being more likely to have presented to specialized centers. Transient focal neurologic deficits may reflect transient vasoconstriction, rapid recovery from infarction, or migraine-like aura phenomenon.²¹ Cerebral infarctions are associated with proximal vasoconstriction^10,17 and are predominantly in border zone distributions of the cerebral hemispheres between the posterior cerebral arteries and carotid territories, although cerebellar infarctions are also reported.¹ Ischemic infarction on imaging is associated with clinical worsening.¹⁹

Posterior Reversible Encephalopathy Syndrome and Seizures

PRES is an acute or subacute clinicoradiographic syndrome characterized by subcortical brain edema and acute neurologic symptoms, including seizures, encephalopathy, headache, and visual disturbances.²² There is overlap between PRES and RCVS suggesting a common pathophysiology including endothelial dysfunction causing disturbances in vascular tone and capillary leakage.²³ PRES shares risk factors with RCVS, such as pregnancy and the peripartum state, the use of immunosuppressants, and its propensity to affect young and middle-aged women.²³ Overlapping clinical features include acute headaches, seizures, and focal neurologic deficits. Common radiographic findings include brain lesion predilection for arterial border-zone cerebral vasoconstriction on angiography.²⁴ A significant minority of people with RCVS develop clinical and MRI findings consistent with PRES at rates reported to be 6% to 39% .^9,17,19,25

Acute symptomatic seizures have been reported in 5% to 17% of patients with RCVS and may be focal or generalized.^7,8,14 Development of epilepsy has not been reported, although those who develop intracranial hemorrhage or brain infarction are potentially at risk.

Radiographic Features

Brain parenchyma imaging is often normal, especially on first presentation. Depending on methods of study recruitment, brain lesions are noted on the initial imaging of 12% to 80% of RCVS.^1,14,25,26 In a cohort study, 55% of people with RCVS had normal imaging on presentation but 81% had developed brain lesions on repeat imaging.⁷ There may be more than 1 lesion and multiple lesion types, including nonaneurysmal convexity SAH, intraparenchymal hemorrhage, and cerebral infarction.

Noninvasive or invasive cerebral angiography shows segmental narrowing and dilation of 1 or more cerebral arteries. Vasoconstriction can affect the anterior and posterior circulations and can be bilateral and diffuse.¹ There is no currently validated method of diagnosing cerebral vasoconstriction of RCVS. Vasoconstriction may be delayed up to a week and may be difficult to detect in small distal arteries, regardless of the imaging modality used.^1,3,27 Digital subtraction angiography (DSA) is considered the gold standard; the sensitivity of both CT and MR angiography (CTA and MRA) for detecting RCVS-related vasoconstriction compared with DSA is approximately 80%.²⁷ CTA and MRA have the advantage of also screening for parenchymal abnormalities. These include convexity SAH, parenchymal hemorrhage, infarction, and vasogenic edema suggestive of PRES. MRI fluid-attenuated inversion recovery imaging (FLAIR) may show hyperintense vessels, thought to be the result of slow flow through constricted vessels.²⁷ Although still experimental, MRI with contrast enhanced vessel wall imaging (VWI) may help distinguish inflammatory and noninflammatory causes of cerebral vasoconstriction. Transcranial Doppler (TCD) studies are abnormal in 69% to 81% of RCVS cases, although the sensitivity for detecting vasospasm is extrapolated from aneurysmal SAH, and TCD is likely insensitive for detecting distal vasoconstriction.²⁷

Laboratory Evaluation

Serum blood counts, metabolic markers, and inflammatory markers are usually normal in RCVS. Rheumatologic serologies such as rheumatoid factor, antinuclear and antineutrophil cytoplasmic antibodies are usually negative. Urine and serum toxicology should be done to screen for drug use that may have precipitated RCVS. The results of cerebrospinal fluid (CSF) studies are generally normal or nonspecific. CSF may show a mild pleocytosis and an excess of red cells, especially in the setting of concomitant SAH, and elevated CSF protein up to 100 mg/dL has been seen.¹ These abnormalities should normalize within a few days on repeat lumbar puncture.

Differential Diagnosis

Correctly arriving at early diagnosis is imperative because RCVS presentations can overlap with diverse syndromes with disparate treatments. TCH differential diagnosis is broad and beyond the scope of this review. Anyone presenting with new onset TCH should be evaluated for aneurysmal SAH. Unlike RCVS, the distribution of aneurysmal SAH is typically cisternal and deep compared with the superficial convexity and sulcal distribution of the convexity SAH in RCVS.¹ By definition, diagnosis of RCVS requires exclusion of aneurysmal SAH (Box 2).²

Primary angiitis of the central nervous system (PACNS) is often considered in the differential diagnosis because of its presentation with headaches and multifocal cerebral vasoconstriction. Comparative studies of RCVS and PACNS have been completed in large US²⁸ and French²⁵ cohorts. Individuals with RCVS are more likely to be younger, female persons who have an identified trigger, single or recurrent TCH, a history of migraine, normal CSF studies, and convexity SAH or normal brain parenchymal imaging at presentation. In contrast, those with PACNS are more likely to have nonthunderclap and insidious-onset headache, absent headache, and focal neurologic signs at initial presentation, and may have other systemic symptoms such as fever, weight loss, rashes, cognitive deterioration, and personality changes. People with PACNS often have abnormal initial brain parenchymal imaging that may show leptomeningeal vascular enhancement, as well as a CSF profile with elevated white blood cells and elevated protein. For those who present with or develop parenchymal brain lesions, the lesions are often deep infarctions of the brain stem or deep gray structures in PACNS, whereas brain lesions in RCVS tend to be in the hemispheric vascular border zones.^25,28

Rocha and colleagues recently proposed a diagnostic scoring algorithm (Table 1) as a clinical tool to differentiate RCVS from other large and medium vessel intracranial arteriopathies such as PACNS in the acute setting.¹³ The score was derived from a retrospective analysis of people age 18 to 55 years with RCVS (n=30) or other intracranial arteriopathy (n=80). All had abnormal cerebrovascular imaging at presentation. The score was then validated in a previously published cohort²⁸ comparing RCVS (n=159) with PACNS (n=23). A score of at least 5 diagnosed RCVS with 90% sensitivity and 99% specificity, whereas a 2 or less excluded RCVS with 85% sensitivity and 100% specificity for nonRCVS arteriopathy. The score has not been validated for people less than age 18 years or more than age 55.

Precipitants and Triggers

At least half of RCVS cases are referable to a specific trigger or precipitating factor. Triggers include the Valsalva maneuver, sexual activity, physical exertion, bathing, laughing, and stressful situations.¹ Well-established precipitating factors are listed in Table 2.^2,29 In some series, the most common vasoactive substance identified was cannabis,^3,20,25 and the presence of cannabinoids has been associated with multifocal cerebral vasoconstriction in clinical and preclinical studies.³⁰ Individuals with RCVS caused by cannabis use are more likely to be male persons of younger age than noncannabis users with RCVS.²⁰

Pathophysiology

The pathophysiology of RCVS is not yet fully elucidated. Several features of the clinical and angiographic course support the theory of triggers or precipitating factors activating a central mechanism of endothelial dysfunction and vasoconstriction that begins with distal arterioles and progresses centripetally to more proximal arteries over time.^1,31 Observations supporting this theory include the peripheral distribution of parenchymal lesions, SAH at the cortical surfaces, and the delayed evolution of parenchymal and vascular imaging abnormalities. Several authors have suggested that early constriction and dilation of small arterioles causes TCH by triggering pain pathways involving the trigeminal system and C2 nerve root, whereas vasoconstriction initially remains below the resolution of MRA, CTA, or catheter angiography.^1,17 The severity of distal arteriolar vasoconstriction has been associated with greater hemorrhagic burden. As vasoconstriction progresses to more proximal vessels, ischemia becomes more likely, explaining the temporal delay of ischemic lesions. The mechanism by which the cascade of vasoconstriction is activated is unknown.

Management

There are no randomized controlled trials assessing management of RCVS, and guidance is limited to observational data and expert opinion. As described, case series underscore the significant rates of delayed sequelae (eg, intracranial hemorrhage, cerebral infarction, seizures, and PRES), and every person with RCVS should be monitored for complications.^3,7,17,19,26 Whether all require inpatient observation and for what duration is not addressed in the literature. Supportive care includes rest and avoidance of triggering activities and precipitating factors, including discontinuation of identified vasoactive medications when possible. Whether and when it is safe to resume vasoactive medications are questions that require further research. There is a recurrence rate of 5%,^26,32 and the risks and benefits of treatments should be weighed on case-by-case basis. There are no data on optimal blood pressure management in patients with RCVS, and some experts have recommended adhering to guidelines for acute stroke with care taken to avoid hypotension in the setting of cerebral vasoconstriction.¹

Calcium channel blockers (CCBs) such as nimodipine, verapamil, or nicardipine are commonly used in reported case series via oral, intravenous, or intra-arterial routes.^{3,7,9,10,14,19} Definitive data on efficacy, however, are sparse. A small case series of recurrent primary TCH (n=11), published before RCVS diagnostic criteria were developed, reported that nimodipine was effective in aborting TCH.³³ Although CCBs are often used for headache management, they have not been shown to modify angiographic vasoconstriction or the risk of neurologic complications.^3,7,18 A retrospective study of 59 people with RCVS from 2 institutions found 19 of 20 patients who experienced clinical worsening after the RCVS diagnosis were taking CCBs at the time of worsening.¹⁹ Intra-arterial vasodilator therapy is usually not recommended, although has sometimes been used in severe refractory cases. Other vasodilator strategies (eg, magnesium sulfate) have been used in anecdotal reports.^1,26

Glucocorticoids should be avoided in the management of RCVS. In a large single-center study (n=162), glucocorticoid use was found to be strong independent predictor of clinical, angiographic, and radiologic worsening and predicted poor clinical outcome.³⁴ The authors postulated that glucocorticoids may amplify the effects of endogenous vasoconstrictors, and thereby worsen outcomes.

Prognosis

Long-term outcomes are favorable in the vast majority of cases and most individuals can return to their prior level of functioning without neurologic deficits.^7,19,31 A population–based US study of over 1,000 people with RCVS, however, found 14% required readmission and 37% who were readmitted were discharged to a rehabilitation facility rather than home.⁶ The most common reasons for readmission were acute ischemic or hemorrhagic stroke, continued or recurrent symptoms of RCVS, infections, and headaches. In another study, more than a quarter of those with hemorrhage were unable to return to their previous work at 6 months.¹⁸

RCVS can recur long after recovery and the return of TCHs should prompt repeat evaluation. In a prospective cohort study from Taiwan, 5% (n=168) with mean follow-up 37.5 months had RCVS recurrence at a rate of 1.71 per 100 person-years.³² Recurrence happened as early as 6 months and as late as 7 years after the initial episode. Only sexual activity as a trigger was found to be significantly associated with risk of recurrence. A prospective study in a French cohort with a median follow up of 9.8 years also found a 5% recurrence but at a lower rate of 0.65 per 100 person-years. Risk factors predicting recurrence included a presentation of exercise-induced TCH and history of migraine.²⁶ Remarkably, more patients in the French cohort had an identified precipitating factor compared with the Taiwanese cohort (43% vs 1.5% respectively) and counseling to avoid vasoactive triggers may have contributed to the lower rate of RCVS recurrence.