Reversible cerebral vasoconstriction syndrome

INTRODUCTION — Reversible cerebral vasoconstriction syndrome (RCVS) represents a group of conditions that show reversible multifocal narrowing of the cerebral arteries with clinical manifestations that typically include thunderclap headache and sometimes include neurologic deficits related to brain edema, stroke, or seizure. The clinical outcome is usually benign, although major strokes can result in severe disability or death in a minority.

This topic will review RCVS. Other conditions associated with thunderclap headache are discussed separately. (See "Overview of thunderclap headache" and "Primary cough headache" and "Exercise (exertional) headache" and "Primary headache associated with sexual activity".)

TERMINOLOGY — RCVS has been reported using variable terminology, including the following:

●Migrainous vasospasm or migraine angiitis [1,2]

●Call-Fleming syndrome (or Call syndrome) [3,4]

●Thunderclap headache-associated vasospasm [5-7]

●Drug-induced cerebral arteritis [8]

●Postpartum cerebral angiopathy [9]

●Benign angiopathy of the central nervous system [10]

●Central nervous system pseudovasculitis [11]

These conditions are characterized by clinical manifestations that typically include thunderclap headache and, less commonly, focal neurologic deficits related to brain edema, stroke, or seizure and angiographic reversible multifocal narrowing of the cerebral arteries. It is now apparent that patients with reversible cerebral arterial narrowing have nearly identical clinical, laboratory, imaging, and prognostic features regardless of the associated condition [12-14]. The descriptive term "reversible cerebral vasoconstriction syndrome" has been proposed to facilitate the recognition and management of this group of disorders [15]. The adoption of the broad term RCVS, along with its main clinical and imaging features, has encouraged relatively large retrospective and prospective studies that have helped characterize the syndrome [15-23].

PATHOPHYSIOLOGY — The pathophysiology of the abrupt-onset headache and of the prolonged but reversible vasoconstriction is not known. Reversible angiographic narrowing suggests an abnormality in the control of cerebrovascular tone [24]. It remains unclear whether the angiographic abnormalities trigger the headaches or result from severe headache, but there certainly is a close relationship [25,26]. The anatomic basis to explain both the vasoconstriction and headaches is the innervation of cerebral blood vessels with sensory afferents from the trigeminal nerve (V1) and dorsal root of C2. Cerebral vasoconstriction, when severe or progressive, may result in ischemic stroke and in some cases brain hemorrhages that probably reflect postischemic reperfusion injury due to the dynamic and reversible nature of the arterial narrowing. Some patients develop convexal subarachnoid hemorrhages, presumably from the rupture of small surface arteries undergoing dynamic vasoconstriction-vasodilation.

The pathophysiologies of RCVS and reversible posterior leukoencephalopathy syndrome (RPLS) may overlap as both entities can present with reversible brain lesions, including transient brain edema in patients with RCVS and transient cerebral angiographic abnormalities in patients with RPLS [27-34]. (See "Reversible posterior leukoencephalopathy syndrome".)

EPIDEMIOLOGY — The true incidence of RCVS is unknown. In a study using administrative claims data from 2016 in the United States, the incidence of RCVS hospitalizations was approximately three per one million adults [35]. Clinical experience suggests the true incidence of RCVS is likely much higher [36]. RCVS is being reported with increasing frequency, presumably due to greater awareness of the syndrome, higher detection rates due to the widespread use of relatively noninvasive imaging tests such as computed tomography angiography (CTA) and magnetic resonance angiography (MRA), and perhaps the escalating use of illicit drugs and vasoconstrictive medications [37].

In adults, RCVS predominantly affects females, with female to male ratios ranging from 2:1 to 10:1, depending on the case series [35,36,38,39]. In contrast, a 2017 review of pediatric RCVS found that most cases affected males (11 of 13) [40].

The mean age of affected individuals across published studies is 42 to 48 years, with an age range of 4 months to 65 years [16,22,36,39-42]. RCVS occurs worldwide in individuals of all races.

RISK FACTORS AND ASSOCIATED CONDITIONS — RCVS has been associated with a variety of conditions including pregnancy [8,43], migraine [1,2,44], use of vasoconstrictive drugs [8,36,45,46] and other medications [29,47], neurosurgical procedures [48], hypercalcemia [49], unruptured saccular aneurysms [5,50], cervical artery dissection [50,51], cerebral venous thrombosis [52,53], and others [41,54-57].

The individual risk factors, triggers, and conditions associated with RCVS (table 1) appear unrelated (ie, without a common pathophysiological theme) and may simply reflect the biases of investigators in attributing risk. Indeed, the variable nosology previously used by different physician groups (eg, stroke neurologists, headache specialists, obstetricians, internists, and rheumatologists) to report this clinical-angiographic syndrome reflects uncertainties concerning the pathogenesis and clinical approach. Authors have implicated the listed conditions, including commonly used medications such as serotonergic antidepressants, based on their known vasoconstrictive effects or the temporal relationship with the onset headaches [4]. However, epidemiologic evidence to support a causal relationship is lacking. Some authors have speculated that the vasoconstriction is related to transient vasculitis, but there is no evidence to support a role for inflammation. Cerebrospinal fluid examination and extensive serological tests are normal, and pathological studies of the brain and temporal arteries have shown no abnormality [58].

CLINICAL PRESENTATION AND COURSE

●Thunderclap headaches – The clinical presentation of RCVS is usually dramatic, with sudden, excruciating headaches that reach peak intensity within seconds, meeting the definition for "thunderclap headache" [59,60]. The thunderclap headaches tend to recur over a span of days to weeks.

The onset headaches are usually diffuse or located in the occipital region or vertex. They are often accompanied with nausea and photosensitivity. The character of these headaches is usually different from the patient's prior migraine headaches, if any. Most patients experience moderate pain relief within a few minutes to hours, only to be followed by sudden, severe exacerbations that can recur for days. In one study, patients reported an average of four recurrences [16].

Less than 10 percent of patients with RCVS present with subacute or less severe headaches; the absence of headache at onset is exceptional [16,22,61,62].

●Triggering factors – Many patients have triggering factors, such as orgasm, physical exertion, acute stressful or emotional situations, Valsalva maneuvers (eg, straining, coughing, sneezing), bathing, and swimming [22,63].

●Blood pressure – The initial blood pressure can be elevated with RCVS due to severe headache pain, the disease itself, or the associated condition (eg, eclampsia, cocaine exposure).

●Neurologic involvement – Headache remains the only symptom in many patients with RCVS; others develop focal deficits from underlying ischemic stroke, intracerebral hemorrhage, or reversible cerebral edema [15,16,22]. In published series, the frequency of focal neurologic deficits ranged from 9 to 63 percent, being higher in inpatient case series. In one report of 139 patients with RCVS, a majority (81 percent) eventually developed brain lesions including ischemic infarction (39 percent), brain edema (38 percent), convexity subarachnoid hemorrhage (33 percent), and lobar hemorrhage (20 percent) [22]. Generalized tonic-clonic seizures are reported in 0 to 21 percent of patients at the time of presentation; however, recurrent seizures are rare.

Hemiplegia, tremor, hyperreflexia, ataxia, and aphasia can develop. Visual deficits, including scotomas, blurring, hemianopia, and cortical blindness, are common, and these patients typically have concomitant reversible posterior leukoencephalopathy syndrome [62]. Many patients show features of Balint syndrome, which is made up of the triad of simultanagnosia (the inability to integrate a visual scene despite adequate acuity to resolve individual elements), optic ataxia (the inability to reach accurately under visual guidance), and ocular apraxia (the inability to direct gaze accurately to a new target, frequently leading to difficulty reading) [22,64].

●Neuroimaging – Brain imaging is often normal early in the course of RCVS. Typical abnormalities include vasogenic edema and/or fluid-attenuated inversion recovery (FLAIR) sulcal hyperintensities (dot sign) on magnetic resonance imaging (MRI). Infarcts, if present, are usually symmetric and distributed along border zones of arterial territories. Intraparenchymal hemorrhage and/or nonaneurysmal convexity subarachnoid hemorrhage may be present in some cases of RCVS. Multifocal segmental cerebral artery vasoconstriction on cerebral angiography is the hallmark of RCVS. These findings are discussed in detail below. (See 'Brain imaging' below and 'Neurovascular imaging' below.)

●Time course – The resolution of the different components of RCVS, including thunderclap headaches, focal deficits, and angiographic narrowing, usually occurs over days to weeks but does not always follow the same time course. (See 'Clinical course and prognosis' below.)

EVALUATION

Urgent evaluation — Nearly all patients with RCVS present with one or more thunderclap headaches. Thunderclap headache must be evaluated and treated as a medical emergency, beginning with an evaluation for potentially serious secondary causes such as a ruptured brain aneurysm, brain hemorrhage, cervical artery dissection, and other conditions listed in the table (table 2). Urgent brain and cerebral vascular imaging with a cranial computed tomography (CT) or brain magnetic resonance imaging (MRI) and head and neck CT angiography (CTA) or magnetic resonance angiography (MRA) is appropriate. If initial imaging is normal, lumbar puncture with measurement of opening pressure and cerebrospinal fluid examination for cell counts, glucose and protein levels, and xanthochromia should be pursued to exclude subarachnoid hemorrhage and infectious causes of thunderclap headache.

The past medical history should inquire specifically about associated conditions and possible triggering factors for RCVS, such as those listed in the table (table 1) and discussed above. (See 'Risk factors and associated conditions' above and 'Clinical presentation and course' above.)

The systemic examination of patients with RCVS is usually unrevealing, although the initial blood pressure can be elevated due to either severe headache pain, the disease itself, or an associated condition (eg, eclampsia, cocaine exposure).

Brain imaging — Between 30 and 70 percent of patients with RCVS have no abnormality on initial neuroimaging studies with cranial CT or MRI, despite having (eventually) widespread cerebral vasoconstriction [16,22,26,36,65,66]. However, approximately 75 percent of admitted patients eventually develop parenchymal lesions (image 1 and image 2). The most frequent lesions are ischemic stroke and cortical surface (convexity) nonaneurysmal subarachnoid hemorrhage, followed by reversible vasogenic brain edema and parenchymal hemorrhage [16,22,36]. Any combination of lesions can be present. CT and MRI remain normal in approximately 25 percent of cases reported from in-hospital settings; this number is much higher in emergency department case series.

Infarcts are often bilateral and symmetrical, located in arterial border-zone (ie, watershed) regions of the cerebral hemispheres or in the cortical-subcortical junction. Larger infarcts are often wedge shaped. Perfusion-weighted MRI may show areas of hypoperfusion in border-zone regions. Cortical surface (convexity) subarachnoid hemorrhage is typically minor, restricted to a few sulcal spaces [27,67,68]. Several studies have shown that RCVS is the most frequent cause of cortical surface (convexity) subarachnoid hemorrhage (image 3 and image 1) in individuals below age 60 years [69-71].

Single as well as multiple lobar hemorrhages can occur, and brain hemorrhages can develop a few days after onset, which again suggests a mechanistic role for reperfusion injury. Subdural hemorrhage has been reported [38]. Fluid-attenuated inversion recovery (FLAIR) MRI often shows indirect signs of RCVS in the form of dot (ie, the dot sign) or linear hyperintensities within sulcal spaces, which are believed to represent slow flow within dilated surface vessels [72,73]. Intravascular hyperdensities can be hard to distinguish from convexity subarachnoid hemorrhage (image 1).

Neurovascular imaging — Abnormal cerebral angiography is the primary diagnostic feature of RCVS. Cerebral angiographic abnormalities are dynamic, progressing over time from distal to proximal vessels, and appear as intermittent areas of focal narrowing ("sausage on a string" appearance) of the circle of Willis arteries and their branches. Smooth, tapered narrowing followed by abnormal dilated segments of second- and third-order branches of the cerebral arteries (image 1) is the most characteristic abnormality.

●Study selection – CTA or MRA are preferred initial modalities to document the segmental cerebral arterial narrowing and vasodilatation (image 1 and image 4). Digital subtraction catheter angiography (DSA) is typically reserved for cases when CTA or MRA are nondiagnostic or in atypical cases to exclude other entities in the differential diagnosis. DSA is more sensitive than either CTA or MRA for identifying abnormalities in smaller vessels (ie, lumen diameter <2 mm) (image 5) but is invasive and carries a higher risk than the noninvasive vascular studies. The diagnosis of RCVS can be made with accuracy based on history and the results of initial brain and noninvasive vascular imaging alone [36,62].

Transcranial Doppler ultrasound has been used for diagnosis; however, normal results do not exclude this diagnosis [9]. This noninvasive bedside tool has utility in monitoring the progression of vasoconstriction [17].

●Timing of imaging – Initial CTA or MRA can be normal because the condition starts distally in small vessels that are not well visualized; one study found that 21 percent had normal findings on initial MRA and 9 percent had normal findings on both MRA and transcranial Doppler ultrasonography [16]. Vasoconstriction may not become evident for more than one week after clinical onset, peaks at two to three weeks, and is typically reversible within three months (image 4). For patients with a high degree of clinical suspicion for RCVS, a follow-up CTA or MRA should be done after three to five days.

Vascular imaging may reveal concomitant cervicocephalic arterial dissection or unruptured aneurysms [5,51,74,75]. In some patients, the extracranial internal carotid or vertebral artery can be affected by RCVS. Systemic arteries are rarely involved [76,77].

Other tests — Serum and urine toxicology screens should be routinely performed to investigate for exposure to vasoconstrictive drugs such as cannabinoids and cocaine. Laboratory evaluation should also include urine vanillylmandelic acid and 5-hydroxyindoleacetic levels to evaluate for vasoactive tumors (eg, pheochromocytoma, carcinoid) that are associated with RCVS, and a serum calcium level to exclude hypercalcemia as a cause of RCVS, if there is clinical suspicion for these conditions based on symptoms or signs. Serum magnesium should be obtained if there is local preference to treat vasoconstriction with intravenous magnesium.

When there is uncertainty about the cause of cerebral arteriopathy, we obtain complete blood count, electrolytes, liver and renal function tests, and tests for inflammation (eg, erythrocyte sedimentation rate, rheumatoid factor, and antinuclear cytoplasmic antibodies), all of which are typically normal in patients with RCVS. However, these tests are not necessary if the diagnosis of RCVS is highly likely, based upon the presence of recurrent thunderclap headaches (table 3) or RCVS2 score (table 4 and table 5). (See 'Diagnosis' below and 'Angiographic differential' below.)

●Lumbar puncture – Although lumbar puncture is required in patients presenting with thunderclap headache to exclude secondary causes such as a ruptured cerebral aneurysm or meningitis, it could be avoided in patients with multiple thunderclap headaches, since three or more recurrent thunderclap headaches are diagnostic for RCVS [36,62].

In patients with a single thunderclap headache, lumbar puncture may be needed to exclude secondary causes unless there is clear evidence for RCVS on CTA or MRA with multifocal segmental narrowing of the cerebral arteries [36,62].

Patients with RCVS typically have normal cerebrospinal fluid findings (ie, protein level <60 mg/dL, ≤5 white blood cells per mm³). In one series, with cerebrospinal fluid examination performed in over 100 patients with RCVS, results were normal in approximately 85 percent [22]. Minor abnormalities can result from ischemic or hemorrhagic strokes. The classic lumbar puncture findings of aneurysmal subarachnoid hemorrhage (ie, an elevated opening pressure, an elevated red blood cell count that does not diminish from tube 1 to tube 4, and xanthochromia) are absent in RCVS.

●Biopsy – There is generally no role for brain biopsy or temporal artery biopsy unless the diagnosis remains unclear despite a thorough evaluation and there is at least moderate suspicion for cerebral vasculitis.

DIAGNOSIS — The diagnosis of RCVS is based upon the characteristic clinical, brain imaging, and angiographic features. Key components of the diagnosis (table 3 and table 6 and table 4 and table 5) are single or recurrent thunderclap headaches, absence of aneurysmal subarachnoid hemorrhage, and typical brain imaging findings (image 2) (eg, normal, or variable presence of vasogenic edema, fluid-attenuated inversion recovery sulcal hyperintensities [dot sign] on magnetic resonance imaging, symmetric border-zone infarcts, intraparenchymal hemorrhage, and/or nonaneurysmal convexity subarachnoid hemorrhage), combined with multifocal segmental cerebral artery vasoconstriction on angiography, which usually develops within a week of symptom onset [15,36,62].

The presence of multiple thunderclap headaches recurring over a few days has nearly 100 percent sensitivity and specificity for the diagnosis of RCVS [36,62]. The sensitivity and specificity of variables useful to diagnose RCVS and to distinguish it from primary angiitis of the central nervous system (a historic mimic of RCVS) is shown in the table (table 6). In patients with a newly detected cerebral arteriopathy, the RCVS₂ score (table 4 and table 5) has excellent sensitivity and specificity for diagnosing RCVS and distinguishing it from a variety of other cerebral arteriopathies.

DIFFERENTIAL DIAGNOSIS — To the experienced clinician, RCVS is an instantly recognizable entity based on certain features (see 'Clinical presentation and course' above). Most patients report severe thunderclap headaches and characteristic brain imaging features and vascular abnormalities that resolve over a few weeks. The syndrome is known to occur in certain clinical settings (table 1). Individually, however, the clinical and imaging features carry a wide range of differential diagnoses. In the past, patients with RCVS have been misinterpreted as having primary angiitis of the central nervous system (PACNS) or aneurysmal subarachnoid hemorrhage due to overlapping features such as headache, stroke, and cerebral angiographic narrowing.

Headache differential

●Aneurysmal subarachnoid hemorrhage is a major consideration in the differential of RCVS because of the presence of thunderclap headaches, subarachnoid blood, and cerebral artery narrowing [19,67,68]. However, the recurrent nature of thunderclap headaches related to RCVS, the superficial location and small quantity of subarachnoid blood, and the widespread, symmetric vasoconstriction distinguish RCVS from aneurysmal bleeds [19]. (See "Aneurysmal subarachnoid hemorrhage: Clinical manifestations and diagnosis".)

●Other causes of thunderclap headache should be considered. The presence of recurrent thunderclap headaches over the span of a few days is pathognomonic for RCVS [36,62]. Nevertheless, isolated thunderclap headache can signify a variety of ominous conditions, including cerebral artery dissection, cerebral venous sinus thrombosis, ischemic stroke, intracranial infection, spontaneous intracranial hypotension, reversible posterior leukoencephalopathy syndrome, pituitary apoplexy, and colloid cyst of the third ventricle (table 2 and table 7). These conditions are differentiated with appropriate evaluation and imaging. The clinical features, imaging characteristics, and cerebrospinal fluid findings of the more common causes of thunderclap headache are summarized in the table (table 8). (See "Overview of thunderclap headache" and "Overview of thunderclap headache", section on 'Diagnostic evaluation' and "Cerebral and cervical artery dissection: Clinical features and diagnosis" and "Cerebral venous thrombosis: Etiology, clinical features, and diagnosis".)

If secondary causes of thunderclap headache are excluded, the differential diagnosis narrows to include RCVS, primary thunderclap headache, and associated primary headaches (ie, primary cough headache, primary exercise headache, and primary headache associated with sexual activity). These conditions are closely related [26]. The segmental angiographic abnormalities that accompany RCVS may be absent in the early stages of the condition; hence, the patient initially may be misdiagnosed as having a primary thunderclap headache. In such cases, a follow-up cranial computed tomography angiography (CTA) or magnetic resonance angiography (MRA) after approximately one week should be performed to investigate for RCVS. In one study, 39 percent of patients presenting with thunderclap headache and normal brain magnetic resonance imaging (MRI) findings proved to have vasoconstriction on MRA, and those with and without vasoconstriction had similar clinical features, suggesting that RCVS and primary thunderclap headache belong to the same spectrum of disorders [26]. (See "Overview of thunderclap headache", section on 'Primary TCH' and "Primary cough headache" and "Exercise (exertional) headache" and "Primary headache associated with sexual activity".)

●Migraine is another consideration in the differential diagnosis of RCVS, and the misdiagnosis of migraine can lead to inappropriate treatment with antimigraine agents such as triptans, which can exacerbate vasoconstriction and stroke risk [4,78]. Although there may be some overlap, RCVS appears distinct from migraine because, unlike migraine, RCVS rarely recurs, the sudden-onset headaches of RCVS are quite different from migraine headaches, the brain and vascular imaging abnormalities are inconsistent with migraine, and the angiographic abnormalities of RCVS persist for several weeks. (See "Pathophysiology, clinical manifestations, and diagnosis of migraine in adults" and "Pathophysiology, clinical features, and diagnosis of migraine in children".)

Angiographic differential

●Intracranial arteriopathies – The angiographic abnormalities of RCVS can raise concern for intracranial atherosclerosis, infectious arteritis, vasculitis, moyamoya disease, fibromuscular dysplasia, and other cerebral arteriopathies. (See "Intracranial large artery atherosclerosis: Epidemiology, clinical manifestations, and diagnosis" and "Moyamoya disease and moyamoya syndrome: Etiology, clinical features, and diagnosis" and "Clinical manifestations and diagnosis of fibromuscular dysplasia".)

A number of features on initial evaluation help to distinguish RCVS from other intracranial arteriopathies that affect large and medium-sized vessels. In a retrospective study of consecutive patients with RCVS (n = 30) or non-RCVS arteriopathy (n = 80), recurrent or single thunderclap headache, vasoconstrictive trigger, female sex, and convexity subarachnoid hemorrhage were predictors of RCVS; luminal irregularities of the intracranial carotid artery was a negative predictor, being more frequently present in non-RCVS (mainly moyamoya disease) compared with RCVS (58 versus 20 percent) [62].

These features were incorporated into the RCVS₂ score (table 4):

•Recurrent or single thunderclap headache: present 5, absent 0

•Carotid artery (intracranial segment) narrowing: affected -2, not affected 0

•Vasoconstrictive trigger: present 3, absent 0

•Sex: female 1, male 0

•Subarachnoid hemorrhage: present 1, absent 0

The RCVS₂ score can now be used for diagnosis shortly after admission, even before angiographic reversal occurs, with high accuracy [62]. In the derivation cohort, RCVS₂ scores ≥5 had a high specificity and sensitivity (99 and 90 percent, respectively) for diagnosing RCVS, while scores ≤2 had a high specificity and sensitivity (100 and 85) percent for excluding RCVS; intermediate scores of 3 to 4 had a lower specificity and sensitivity (86 and 10 percent) for diagnosing RCVS (table 5) [62]. Performance was similar in the validation cohort of 156 patients with RCVS and 47 with PACNS. Among patients in the derivation and validation cohorts with RCVS₂ scores of 3 or 4, clinical features of recurrent thunderclap headaches, vasoactive triggers, and normal brain imaging or the presence of convexity subarachnoid hemorrhage correctly identified 25 of 37 patients with RCVS.

●Primary angiitis of the central nervous system – Historically, it was considered difficult to exclude PACNS from RCVS because features such as headache, focal deficits, stroke, seizures, and angiographic irregularities are common to both conditions. (See "Primary angiitis of the central nervous system in adults".)

While there is overlap, the nature of the headaches and imaging abnormalities are quite different [15,16,22,36,61,79,80]. Patients with PACNS usually have an insidious progressive clinical course with chronic headaches and rarely have thunderclap headache that is typical of RCVS. The characteristic vasoconstriction of RCVS usually manifests as smooth, tapered narrowing followed by abnormal dilated segments of second- and third-order branches of the cerebral arteries. This angiographic appearance distinguishes RCVS from PACNS, where the arterial narrowing is much more irregular. Brain imaging in RCVS can be normal or show watershed infarcts or lobar hemorrhages, whereas PACNS is usually associated with accumulating T2-hyperintense brain lesions, leptomeningeal enhancement, and scattered deep infarcts [36].

In a retrospective report that compared 159 patients with RCVS and 47 patients with PACNS, several features had 98 to 100 percent specificity for the diagnosis of RCVS and a similarly high positive predictive value (ie, the likelihood that a patient with a positive finding has the disease) (table 6) [36]. These were (1) recurrent thunderclap headache or (2) single thunderclap headache combined with normal neuroimaging, border zone infarcts, or vasogenic edema, or (3) no thunderclap headache but abnormal angiography and no brain lesions on neuroimaging. Note that the absence of brain lesions virtually rules out PACNS.

These criteria were independently validated in a study comparing cohorts of 173 patients with RCVS and 110 patients with PACNS [81]. They can be used for bedside diagnosis at the time of admission, even without cerebral angiography or documentation of vasoconstriction reversal on follow-up imaging.

There have been rare reports of cases with severe and prolonged vasoconstriction associated with irreversible angiographic changes, making the distinction between vasculitis and vasoconstriction extremely difficult [82]. In challenging cases, 3D contrast-enhanced vessel wall imaging may help to identify arterial wall enhancement that has been associated with cases of cerebral vasculitis but not in cases of RCVS (image 6) [83]. However, the utility of this test remains to be confirmed [84].

MANAGEMENT

Supportive care — There is no proven or established therapy for RCVS. While most patients fully recover with time, up to one-third can develop transient symptoms in the initial few days, and rare cases can develop a progressive clinical course [18]. Therefore, it is reasonable to admit patients for observation, pain control, and supportive care for the first few days after symptom onset.

●Blood pressure – Patients with severe angiographic abnormalities are often admitted to the intensive care unit for neurologic monitoring and blood pressure management. The goals of blood pressure control are controversial. While there is no consensus, we generally allow a broad range of systolic blood pressure, from 90 to 180 mmHg. We treat hypotension (systolic <90 mmHg) with intravenous fluids, although the threshold of 90 mmHg may be too low if vasoconstriction is severe. High blood pressure (systolic >180 mmHg) can be treated with labetalol or nicardipine. Theoretically, pharmacologically induced hypertension can induce further cerebral vasoconstriction or result in brain hemorrhage and, in the setting of cerebral vasoconstriction, even mild hypotension can trigger ischemic stroke [85].

●Pain – The pain of RCVS-associated headache is extreme and frequently warrants the use of opioid analgesics in addition to nonsteroidal anti-inflammatory drugs (NSAIDs). In our experience, oral treatment with hydromorphone or acetaminophen-codeine is usually sufficient to manage pain. Thunderclap headaches typically decrease in intensity and frequency over a span of days to weeks. Triptans and the ergot derivatives are contraindicated because of their vasoconstrictive actions [4,78].

●Seizures – Acute seizures warrant treatment with antiseizure medications, though seizures are usually present only upon presentation and do not recur. Therefore, long-term seizure prophylaxis is probably unnecessary. No seizure prophylaxis is needed for patients who do not have a seizure. (See "Evaluation and management of the first seizure in adults", section on 'Acute symptomatic seizures' and "Evaluation and management of the first seizure in adults", section on 'When to start antiseizure medication therapy'.)

●Avoid empiric glucocorticoids – We suggest not using empiric glucocorticoid therapy for possible primary angiitis of the central nervous system (PACNS) when RCVS is suspected. However, glucocorticoids are often administered to minimize the risk of delaying treatment in patients who may actually have PACNS, a condition that shares certain features with RCVS (see 'Differential diagnosis' above) and is believed to be progressive and potentially fatal without prompt immunosuppressive therapy. Unfortunately, many patients remain on glucocorticoids for prolonged durations and incur the risk of serious steroid-related adverse effects.

There are several reasons to avoid glucocorticoid therapy:

•Distinguishing RCVS and PACNS in the acute setting is generally straightforward (table 6 and table 4 and table 5). (See 'Angiographic differential' above.)

•There is little evidence that a therapeutic delay of a few days would increase the risk for worse outcome in PACNS; even with diagnostically challenging cases, the diagnosis usually becomes apparent after a brief period of observation.

•Glucocorticoids are associated with worse outcome in RCVS [22,86].

Bedside efforts should focus on distinguishing RCVS from PACNS on the basis of the initial clinical and imaging features and reserve empiric glucocorticoid therapy for the rare patient with a rapidly worsening clinical course while the diagnosis remains uncertain.

Vasoconstriction — Because clinical and angiographic resolution occur spontaneously without any medical intervention in approximately 90 percent of patients with RCVS, we generally do not use any agent to treat vasoconstriction.

In the absence of controlled trials, management of vasoconstriction is guided by observational data and expert opinion. Empiric therapy is not justified for patients who present with thunderclap headache but have not yet undergone vascular imaging. Even when cerebral vasoconstriction has been documented, specific treatment remains undefined. While the literature is replete with various treatment approaches associated with good outcome, these reports probably reflect publication bias.

●Pharmacologic treatment – Calcium channel blockers such as nimodipine and verapamil [87] and brief courses of magnesium sulfate [27,88], serotonin antagonists, and dantrolene [89] have been administered in an effort to relieve the vasoconstriction. Data from two prospective case series suggest that nimodipine does not affect the time course of cerebral vasoconstriction [16,17]. However, nimodipine might relieve the number and intensity of headaches and has documented effects on the smaller vasculature not easily imaged by angiography. Calcium channel blockers can be discontinued after resolution of symptoms or angiographic abnormalities if they are used.

●Intra-arterial vasodilation – We reserve intra-arterial measures for patients exhibiting clear signs of clinical progression, particularly since over 90 to 95 percent of RCVS patients have a benign, self-limited syndrome despite the presence of severe angiographic vasoconstriction and ischemic or hemorrhagic brain lesions. Unfortunately, no known clinical or imaging features reliably predict disease progression.

Balloon angioplasty and direct intra-arterial administration of nicardipine, papaverine, milrinone, and nimodipine have been used with variable success [90-92]. In patients with RCVS, intra-arterial infusion of vasodilators into a single constricted artery can promptly reverse vasoconstriction in that artery and often in multiple brain arteries, including the contralateral arteries. A similar but milder response has rarely been observed in RCVS mimics such as PACNS and intracranial atherosclerosis. On this basis, the demonstration of arterial dilatation using intra-arterial vasodilator infusions has been proposed as a "diagnostic test" for RCVS [93]. However, intra-arterial interventions carry a risk for reperfusion injury.

Prevention and counseling — In the acute setting, it is logical to avoid further exposure to any potential precipitating factors, such as marijuana, cocaine, exercise stimulants, amphetamines and triptans, serotonergic antidepressants, or other vasoconstrictive medications that can worsen the clinical course. Patients should avoid physical exertion, sexual activity, the Valsalva maneuver, and other known triggers of recurrent headaches for a few weeks. Laxatives and stool softeners should be used to avoid constipation (which can trigger the Valsalva maneuver), especially in patients receiving opioids for head pain.

The risk of recurrent RCVS is low; hence, reexposure to the potential precipitating factor (eg, antidepressants) can be considered if clinically necessary and after other therapeutic options are exhausted.

Usual secondary stroke preventive medications, such as antiplatelet agents, anticoagulants, and cholesterol-lowering agents, are probably not indicated.

There are no known genetic implications of RCVS.

CLINICAL COURSE AND PROGNOSIS — The resolution of the different components of RCVS, including headaches, focal deficits, and angiographic narrowing, does not always follow the same time course. The thunderclap headaches typically resolve over days to weeks. Similarly, most patients show resolution of visual and other focal neurologic signs and symptoms within days to weeks. Less than 15 to 20 percent are left with residual deficits from stroke, and in most cases the deficits are relatively minor or moderate (ie, 90 to 95 percent have a modified Rankin scale score (table 9) of 0 to 2 at discharge) [94].

Progressive cerebral arterial vasoconstriction culminating in massive strokes, brain edema, severe morbidity, or death occurs in less than five percent of cases, and these fulminant cases have been more commonly reported in postpartum patients [58,95-97]. Predictors of poor outcome in retrospective studies include baseline clinical factors (advanced age, comorbid medical conditions), severe imaging findings (infarction, intracerebral hemorrhage, cerebral edema), and exposure to glucocorticoid therapy [39,86].

Some patients go on to have intractable chronic migraine-like headaches or depression [94].

The time course of vasoconstriction is variable, but most patients show resolution within three months. Note that "reversible" in the term RCVS refers to the dynamic and reversible nature of vasoconstriction; clinical deficits from brain damage might persist, and the vasoconstriction (particularly if severe and prolonged) may not fully reverse in rare cases.

Recurrence of an episode of RCVS after resolution of the initial symptomatic period is uncommon, approximately 5 to 6 percent in two studies [98,99], and usually manifests as an isolated thunderclap headache without complications such as stroke [98,100].

SUMMARY AND RECOMMENDATIONS

●Definition and epidemiology – Reversible cerebral vasoconstriction syndrome (RCVS) represents a group of conditions (including Call-Fleming syndrome, benign angiopathy of the central nervous system, and postpartum angiopathy) characterized by reversible narrowing and dilatation of the cerebral arteries. The mean age of onset of RCVS is approximately 42 years. In adults, RCVS affects females more often than males. (See 'Terminology' above and 'Epidemiology' above.)

●Clinical presentation and risk factors – The clinical presentation of RCVS is usually dramatic with sudden, severe thunderclap headaches that often recur over a span of days to weeks. Some patients develop seizures or focal neurologic deficits. (See 'Clinical presentation and course' above.)

Many patients have triggering factors, such as orgasm, physical exertion, acute stressful or emotional situations, Valsalva maneuvers, bathing, or swimming. A variety of conditions have been associated with RCVS including exposure to vasoconstrictive drugs and medications, other headache disorders, and recent pregnancy (table 1). (See 'Risk factors and associated conditions' above.)

●Neuroimaging features – Cerebral angiographic abnormalities of RCVS are dynamic and progress proximally, resulting in a "sausage on a string" appearance of the circle of Willis arteries and their branches. These abnormalities resolve spontaneously over a few weeks. (See 'Neurovascular imaging' above.)

Magnetic resonance imaging (MRI) of the brain is normal in over 50 percent of patients with RCVS. In the ensuing days, many patients may develop complications such as ischemic stroke, convexity (nonaneurysmal) subarachnoid hemorrhage, lobar hemorrhage, and reversible brain edema, alone or in combination. (See 'Brain imaging' above.)

●Diagnostic evaluation – Patients who present with thunderclap headache must be evaluated as a medical emergency, beginning with cranial computed tomography (CT) or brain MRI and head and neck CT angiography (CTA) or magnetic resonance angiography (MRA). If imaging is normal, lumbar puncture and cerebrospinal fluid analysis is appropriate to exclude other causes such as subarachnoid hemorrhage. (See 'Urgent evaluation' above.)

The diagnosis of RCVS is based upon the characteristic clinical, brain imaging, and angiographic features, as summarized in the tables (table 3 and table 6 and table 4 and table 5). (See 'Diagnosis' above.)

●Differential diagnosis – The individual clinical and imaging features if RCVS carry a wide range of differential diagnoses, particularly aneurysmal subarachnoid hemorrhage, other conditions associated with thunderclap headache, and intracranial arteriopathies including intracranial atherosclerosis, primary angiitis of the central nervous system, moyamoya disease, and fibromuscular dysplasia. (See 'Differential diagnosis' above.)

●Management – There is no proven therapy for RCVS. Supportive care is directed toward managing blood pressure, severe headaches, and other complications such as seizures. We generally do not use calcium channel blockers or other agents to treat vasoconstriction, as evidence for this strategy is lacking. Intra-arterial vasodilator therapy has been attempted in fulminant cases with variable success. (See 'Management' above.)

●Prognosis – The clinical outcome is benign in 90 to 95 percent of patients. Rare patients develop severe irreversible deficits or death from progressive strokes or cerebral edema. Recurrence of an episode of RCVS is rare. (See 'Clinical course and prognosis' above.)