Background and purpose Reversible cerebral vasoconstriction syndromes (RCVS) represent a heterogeneous group of cerebrovascular disease characterized by acute presentations and transient segmental narrowing of the distal intracranial arteries. A series of patients with RCVS were studied to better understand the clinical and imaging characteristics of this rare pathology.
Methods A retrospective study was performed on patients that met inclusion criteria for a diagnosis of RCVS. Pertinent clinical and laboratory data, initial and follow-up imaging, treatment and outcomes were studied.
Results 11 patients (10 women, mean age 42 years) diagnosed with RCVS presented with acute onset of severe headache, neurological symptoms and subarachnoid hemorrhage (SAH). Cross sectional imaging (CT/MRI) identified presentations of cortical SAH (n=9) and/or acute infarcts (n=3). Initial cerebral angiography (digital subtraction angiography n=10 or MR angiography n=1) confirmed diffuse vasoconstriction involving the intracranial vasculature. Rheumatological panel (n=9) and CSF analysis (n=8) were not supportive of vasculitis in any patient. In nearly all cases, reversal of vasoconstriction was noted on follow-up cerebral angiography with early resolution in less than 3 months.
Conclusions RCVS classically presents with ‘thunderclap’ headaches and neurological symptoms but cortical SAH is not an uncommon presentation with a unique and focal distribution overlying the cerebral sulci. Although the initial clinical and angiographic appearance of RCVS may be confused for vasospasm related to aneurysmal SAH or primary angiitis of the CNS, its clinical, laboratory and imaging features assist in diagnosis.
Statistics from Altmetric.com
In 1988, neurologists Gregory Call and Marie Fleming described clinical and angiographic findings in four patients with reversible vasospasm of the cerebral vasculature in the setting of thunderclap headache and normal or near normal CSF analysis. They referred to this entity as reversible cerebral segmental vasoconstriction.1 Since then, some authors have referred to this constellation of clinical and imaging findings as Call–Fleming syndrome. Similar findings in a case report were reported as early as 1978 by Snyder and McClelland.2 Over the years, there have been numerous reports of various types of reversible vasoconstriction syndromes within the literature under a multitude of nosologies including, but not limited to, the following: benign isolated arteritis of the CNS,3 benign angiopathy of the CNS,4–6 migrainous vasospasm,7 postpartum cerebral angiopathy,8 thunderclap headache with vasospasm9–12 and drug induced vasospasm.13–18
The different terminology often varied based on the associated clinical features and/or the primary service responsible for the patient. The majority of the existing literature and definition of this entity were published in neurology or rheumatology journals.19 Unfortunately, reversible cerebral vasoconstriction syndromes (RCVS) is not well represented in the neuroradiology literature despite its diagnosis being supported by cross sectional imaging as well as initial and short term follow-up angiographic studies.
RCVS has classically been described as occurring without subarachnoid hemorrhage (SAH) in the setting of normal or near normal CSF analysis.1–4 9 10 17 18 Although there have been reports of non-aneurysmal SAH associated with RCVS,20–24 there have been few series demonstrating SAH as an important feature of RCVS.5 25–27 We report a case series of 11 RCVS patients, the majority presenting with a characteristic pattern of focal, cortical subarachnoid blood along the high convexity sulci. We discuss the clinical and neuroimaging findings of RCVS required for diagnosis and specific features distinguishing it from SAH related vasospasm and primary angiitis of the central nervous system (PACNS), diseases with which RCVS is often confused.
Institutional review board approval and waiver for informed consent was obtained for this HIPAA compliant retrospective study.
Using the computerized radiology information system, a patient search was performed to identify RCVS patients using the following terms: headache, vasculitis, vasculopathy, SAH, reversible cerebral vasoconstriction or Call Fleming syndrome.
Subsequently, individual patient medical and imaging records were reviewed for the following RCVS inclusion criteria, as per Calabrese et al19: (1) acute severe headache with or without neurological signs and symptoms; (2) benign CSF analysis with protein level <80 mg/dl, leucocytes <10/mm3 and normal glucose; (3) no evidence of underlying ruptured intracranial aneurysm/vascular malformation or aneurysmal SAH; and (4) digital subtraction angiography (DSA), MR angiography (MRA) or CT angiography (CTA) evidence of arteriopathy limited to the CNS demonstrating multifocal areas of segmental arterial narrowing which reverses on short term follow-up evaluation (<6 months in the current study).
Pertinent clinical data were recorded as the following: demographics, presenting signs and symptoms, medications, social history, laboratory studies (including rheumatologic serology, CSF analysis, microbiology and virology), biopsy results, clinical course, treatment, complications and follow-up. Short-term outcome was recorded on the basis of the last available medical record.
Two experienced neuroradiologists independently confirmed the initial head CT/CTA, MRI/MRA and DSA findings as well as follow-up brain and cerebrovascular imaging. Cross sectional imaging (CT/MRI) was assessed for the presence of intracranial hemorrhage, SAH (Fischer grade) and acute/subacute infarcts.
Angiographic studies were assessed for evidence of aneurysms, vascular malformations and areas of segmental vascular narrowing. Arterial narrowing or irregularity was referred to as arteriopathy and subjectively graded as mild, moderate or severe. Its distribution was categorized as focal (1 or 2 vessel distributions) or diffuse (>2 vessel distributions). When there was disagreement between the neuroradiologists regarding presence or absence of segmental narrowing, distribution or severity, the studies were jointly reviewed for consensus.
Using RCVS inclusion criteria,19 17 potential patients were identified from the radiology information system database but six patients were excluded because either initial or follow-up angiography was not performed or available for review to confirm short term reversibility (<6 months) of vasoconstriction. Pertinent clinical and laboratory data are listed in table 1.
The patient population with a mean age of 42±11 years was predominantly female (n=10) and Caucasian (n=9). Patient No 9 was 10 days postpartum and patient No 11 was 35 weeks' pregnant at presentation. Five patients were being treated with antidepressants at the time of symptom onset and three patients were taking a selective serotonin reuptake inhibitor (SSRI).
All 11 patients presented with an acute onset of severe headaches over several days prompting medical treatment. Associated neurological symptoms were noted in eight patients (73%), including photophobia, blurred vision, visual field cuts, extremity weakness, seizures, dizziness, confusion and/or cognitive impairment.
CSF analysis in eight patients was near normal except for the presence of red blood cells (n=5) indicating SAH and mild protein elevation (n=5). However, the mean protein level for this population remained normal, measuring 48 mg/dl (range 22–66 mg/dl), with no values exceeding 80 mg/dl, as per our inclusion criteria. None of the patients exhibited leukocytosis indicative of inflammation. Three patients underwent brain biopsies (patients 1, 6 and 10) demonstrating non-specific changes in the media of the arterial wall but no angioinvasive inflammation of the leptomeninges or cortex to support a diagnosis of PACNS. Furthermore, an extensive rheumatology panel in nine patients was negative except for elevated erythrocyte sedimentation rate and C reactive protein in patient 1.
All patients underwent an initial non-contrast head CT and MRI on admission. Imaging results, initial and follow-up angiographic evaluations, and treatment are summarized in table 2.
Cortical SAH was identified in nine patients (81%) by CT/MRI (figure 1A). All SAHs were classified as Fischer grade 2 (blood layering <1 mm in thickness) with a characteristic and subtle imaging pattern: focal and superficial, overlying a few high convexity sulci. In fact, CSF analysis failed to detect SAH in one patient (patient 4). Acute infarcts were noted in three patients (27%) in distal cortical and/or watershed distributions.
Initial angiographic assessment included DSA (n=10), MRA (n=7) and CTA (n=2) but RCVS findings were best visualized with DSA. All DSA studies supported initial findings on cross sectional CTA/MRA. All patients demonstrated arteriopathy, manifested as segmental areas of arterial narrowing or vasoconstriction. These changes were diffuse in distribution, involving greater than two vessel distributions in all patients. Arteriopathy invariably involved the second and third order branch vessels of the intracranial circulation, but occasionally in severe cases, the proximal cisternal segments (A1, M1, P1 segments, basilar artery) were also affected, notably in the posterior circulation. Vasoconstriction typically appeared as a ‘beaded pattern’ on angiography with alternating areas of segmental arterial narrowing. Furthermore, arteriopathy was subjectively graded as severe (n=5), moderate (n=2) and mild (n=4). The diffuse extent of arteriopathy was disproportionate and often distant to the small and focal amount of subarachnoid blood.
Two illustrative RCVS cases are provided as examples (figures 1 and 2) to depict presentations of SAH and acute infarcts as well as vasoconstriction of distal versus proximal cisternal intracranial segments, respectively.
The average number of days between documented SAH and arteriopathy was 2.9 days (range 0–11) (table 2). Moreover, vasoconstriction was verified within 24 h of identifying SAH in five patients.
Due to the retrospective nature of this study, both angiographic and clinical follow-up were variable. Follow-up angiographic evaluation comprised DSA (n=10) and/or MRA (n=2) for a mean follow-up time of 52 days (range 10–172). Complete resolution of arteriopathy was observed in eight patients on follow-up. The remaining three patients exhibited marked improvement in arteriopathy with no new areas of segmental arterial narrowing. Reversibility or resolution of vasoconstriction was identified in nearly all patients (n=10) within 90 days (3 months).
Clinical evaluation consisted of a mean follow-up time of 15.6 months (range 1.4–80.2) with no recurrent presentations. All patients reported complete resolution of the presenting ‘thunderclap’ headaches but six patients (54%) complained of residual, less severe headaches. The majority of associated neurological deficits (eight patients) either resolved or markedly improved on follow-up without significant visual, motor, sensory or cognitive deficits, except for mild residual extremity weakness in one patient (figure 2). No new or progressive neurological deficits developed following initial hospitalization and treatment.
Since the diagnosis of RCVS was delayed or initially confused for PACNS, variable treatments were offered to this population, with anti-inflammatory or immunomodulating drugs administered to four patients. Medical treatment consisted of one or more of the following: calcium channel blockers (verapamil or nimodipine, n=6), steroids (n=4), cyclophosphamide (n=1), thorazine (n=1), oxycontin (n=1) and nortryptyline (n=1). However, no patient required long term steroids or anti-inflammatory medication (>3 months). A single patient was observed without prescribing any medication.
RCVS represent a group of clinical syndromes characterized by severe ‘thunderclap’ headache with or without neurological deficits, benign CSF analysis and multifocal segmental areas of cerebral arterial narrowing or vasoconstriction that usually resolve within weeks.19
An increasing number of patients present to the emergency room with headaches of unknown etiology and normal CSF analysis. With the paucity of neuroradiology and neurosurgical literature pertaining to RCVS, recognition of its clinical and neuroimaging features is important for neurointerventionalists who may be consulted on these imaging findings, particularly in the context of SAH. As CTA rapidly replaces DSA in the initial screening of patients presenting with SAH, the threshold for DSA should remain low for non-diagnostic cases as it remains superior in the evaluation of small aneurysms/vascular malformations and the distal intracranial arteries predominantly affected by RCVS.
The early literature describes RCVS as classically occurring in the absence of SAH with normal or near normal CSF analysis,1–4 9 10 other than isolated case reports.20–23 Subsequently, Hajj-Ali et al5 and Spitzer et al25 reported series in which 13% and 33% of RCVS cases presented with cortical SAH. More recently, Ducros et al26 27 described a 22% incidence of cortical SAH in the largest prospective series of RCVS patients to date, with a 34% incidence of intracranial hemorrhage in their follow-up series of 89 patients. Statistically significant risk factors predisposing RCVS patients to hemorrhage included an older population, female gender and migrainers. In our series, we noted an unusually high incidence of cortical SAH, occurring in 81% of patients. In fact, RCVS presenting with non-aneurysmal SAH may be an under recognized phenomenon as its presence on cross sectional imaging or CSF analysis can be subtle or even missed if not performed in the acute SAH setting. Moreover, patients with RCVS have been shown to delay their presentations (∼7 days) from the onset of their thunderclap headaches.26 Conversely, our case series probably overestimates the presence of SAH due to a selection bias of identifying and diagnosing RCVS patients from a database search yielding predominantly cerebral angiograms, invasive studies always initiated in a setting of SAH at our institution. Furthermore, RCVS cases without SAH may have been misdiagnosed or undiagnosed when routine cross sectional or even non-invasive angiographic studies were performed without a prospective awareness of this relatively rare pathology. In our retrospective study, at least six patients were excluded due to inadequate angiographic follow-up to confirm a diagnosis of RCVS.
RCVS associated SAH has a characteristic appearance: focal and superficial, overlying the cortical convexity along the cerebral sulci. It should be emphasized that cortical SAH may be a subtle finding on imaging and can be easily missed. It may manifest as only one or two hyperdense sulci on CT or linear areas of high signal on MRI FLAIR sequences. Even CSF analysis may fail to detect SAH as in a single case in our series (table 1; patients No 4). Moustafa et al24 reported on two cases of Call–Fleming syndrome presenting with SAH and characterized a similar imaging pattern.
Since recurrent thunderclap headaches can occur over 1 week, delaying presentation in some patients,26 27 recognition of RCVS as a cause of non-aneurysmal SAH may be confused for vasospasm as a result of aneurysmal or even traumatic SAH. Interestingly, the mean interval between documented SAH and vasoconstriction was <2.9 days in our study with four of nine patients presenting with vasoconstriction on the day of documented SAH. Unfortunately, due to the retrospective nature of our study, we were unable to account for the number of days from the onset of our patients' headaches to their documentation of SAH (approximately 4–5 days based on prior large prospective series).26 27 However, the imaging findings on CT/MRI suggest that SAH occurred relatively recently, potentially even exacerbating headaches and prompting their presentation. This is highly unusual for vasospasm that rarely occurs prior to 4–7 days after aneurysmal SAH, with patients presenting very early (0–2 days) after onset of a severe headache.
Based on our observations and the existing literature,4 5 9–12 19 26 RCVS associated SAH may be distinguished from SAH related vasospasm based on the following imaging criteria: (1) absence of ruptured aneurysm or a vascular malformation on angiography; (2) diffuse distribution and disproportionate extent of vasoconstriction relative to the focal cortical subarachnoid blood; (3) beaded appearance of alternating areas of segmental vasoconstriction preferentially involving the distal second and third order intracranial branch vessels; and (4) temporal relationship of segmental vasoconstriction occurring earlier (<4 days) and/or persisting longer (weeks) than SAH related vasospasm In contrast, SAH related vasospasm generally causes smooth, long segment narrowing of the proximal cisternal segments (A1, M1, P1 segments) correlating with the location and amount of subarachnoid blood, peaking between 4 and 14 days after aneurysm rupture. Although the cisternal segments (A1, M1, P1 segments) are rarely affected in RCVS, in severe cases, we noted unusual proximal involvement in a few patients, particularly of the proximal posterior circulation, with segmental vasoconstriction affecting the anterior inferior cerebellar, superior cerebellar and posterior cerebral arteries. This may represent a temporal progression of distal to proximal vessel vasoconstriction, as previously suggested.26
While the angiographic appearance of RCVS may be differentiated from SAH related vasospasm, it is non-specific and may be indistinguishable from vasculitis. RCVS is often mistaken for PACNS and thus patients may be subjected to unnecessary brain biopsy and/or long term treatment with high dose steroids and/or cytotoxic agents that may have harmful side effects. Failure to differentiate these disorders occurs because the diagnosis of PACNS is often made solely on the basis of initial angiography without documenting reversibility or resolution of vasoconstriction seen in RCVS on early follow-up angiography. Given the frequent misdiagnosis as well as heterogeneity of these disorders, Calabrese et al19 have suggested repeating angiography within 12 weeks to document reversibility. Progressive clinical deterioration and/or incomplete resolution of angiographic abnormalities would support an alternative diagnosis and warrants further investigation for PACNS. Interestingly, the angiographic findings of PACNS have limited diagnostic value, with a reported specificity ranging from 26% to 60%.28 29 Other diseases may also demonstrate an overlapping initial angiographic appearance with both RCVS and PACNS, including intracranial atherosclerosis, CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy), emboli, infection and intravascular neoplasm.28 29
In addition to early follow-up angiography, the distinction between RCVS and PACNS is possible with diligent clinical assessment and CSF analysis. Acute severe ‘thunderclap’ headaches are a characteristic presentation of RCVS as opposed to PACNS patients that present with a history of dull headaches. Moreover, PACNS usually has an insidious course of progressive neurologic deterioration whereas RCVS occurs acutely and has a self-limited course.19 26 30 31 CSF analysis may be helpful and is always normal or near normal in patients with RCVS, with occasional mild protein or white blood cell elevation.5 In contrast, patients with PACNS typically demonstrate abnormal CSF values with elevated protein levels greater than 100 mg/dl and leukocytosis in excess of 5–10 cells/mm3.28 30 31 In a single study, CSF analysis yielded both a sensitivity and negative predictive value of 100% in patients with biopsy proven PACNS.28
The natural history of RCVS has not been well studied until recently; in part because it is often misdiagnosed and treated as a vasculitis. RCVS has been reported as a relatively benign, self-limited disease with few sequelae. The most common complaint at follow-up was a low grade residual headache in 54% of our patients. While typically a benign disorder, the most important factors impacting outcome are related to complications of ischemic stroke and intraparenchymal hemorrhage.5 6 11 15 17 18 30 31 MRI may demonstrate focal areas of infarction, frequently in a watershed distribution, but also may be normal in patients with RCVS.11 15 17 18 In contrast, MRI is nearly always abnormal in patients with CNS vasculitis demonstrating numerous areas of cortical and subcortical ischemia in multiple vascular territories.31–34 In our series, 27% of patients had imaging evidence of acute infarction in a watershed or distal cortical distribution with one patient exhibiting a symptomatic neurological deficit on short term follow-up. Diffuse severe vasoconstriction of the distal intracranial arteries may limit cerebral blood flow enough for ischemia to develop in the distal cortical and watershed regions. Similar results have been reported in a case series of 16 patients with RCVS in which 43% of patients had ischemic stroke and 44% of subjects complained of minor headache at short term follow-up.5 However, in the larger prospective series by Ducros et al, only 9% of patients exhibited an acute infarct on MRI with an additional 9% developing reversible posterior leukoencephalopathy.30
The pathophysiology of RCVS remains for conjecture. Normal histology at brain biopsy in three of our patients supports the argument that the segmental arterial narrowing in RCVS is not due to an arteritis but literally due to physiologic vasoconstriction secondary to an inciting event or substance.6 In contrast, RCVS may represent a subclinical and self-limited form of vasculitis. Low grade or transient inflammation in the context of anti-inflammatory treatment that many patients mistakenly receive may not be easily recognized.
Although considered idiopathic, RCVS has been associated with a multitude of factors, including vasoactive substances, medications, postpartum state and headache disorders. The implicated substances have included alcohol, amphetamines, cannabis, cocaine, methylendioxymethamphetamine (Ecstasy), bromocriptine, ergotamine, pseudophedrine, sympathomimetics, SSRIs, interferon, nicotine patches and triptans.13–19 26 27 The association with sympathomimetic drugs and female predominance of the disease has lead to the hypotheses of increased activation of CNS sympathetic fibers and estrogen modulation of sympathetic tone as pathophysiologic causes for reversible vasoconstriction.16 Support for the role of serotonin induced vasospasm in RCVS includes reports of vasoconstriction following the use of serotonergic medications, reports of migraine related stroke in patients with serotonin syndrome and animal studies demonstrating serotonergic–sympathomimetic synergism precipitating cerebral vasoconstriction.13–17 35 36 Of note, in our case series, 45% of subjects were on antidepressant therapy and 27% were being treated with SSRIs.
Unfortunately, treatment of our patients was not standardized, due to the retrospective nature of this report, other than discontinuing all vasoactive substances. A few patients were initially treated with high dose steroids and cytotoxic agents due to a misdiagnosis of PACNS. These agents were later discontinued and all subjects showed no further progression of symptoms and resolution of angiographic abnormalities on follow-up imaging. In other patients correctly diagnosed with RCVS, the treatment approach included calcium channel blockers, short term glucocorticoids and observation, all of which have been reported to be effective.1 2 5–7 9 10 15 18 26
Since RCVS may result in cerebral ischemia or infarction, the use of verapamil or nimodipine as first line therapy has been suggested, particularly if a patient has focal neurological symptoms.19 Although endovascular treatment with intra-arterial calcium channel blockers or phosphodiesterase inhibitors has been described in a few case reports,37 38 the diagnosis of RCVS is often confused or delayed from the initial diagnostic cerebral angiogram and oral/parenteral medications are typically prescribed. However, if a prospective diagnosis of RCVS is suggested or there is an unusually severe presentation, intra-arterial vasodilators may be indicated and effective, with few contraindications or adverse reactions. As opposed to SAH related vasospasm, intracranial angioplasty is rarely indicated in RCVS, but may be considered in severe ischemic presentations involving the proximal cisternal segments.
RCVS is an important cause of cerebral arteriopathy or vasoconstriction. It has classically been described as occurring in the absence of SAH. However, as demonstrated in our series, cortical SAH may be a more common presentation of RCVS than previously recognized, exhibiting a characteristic pattern of focal and superficial SAH along the high convexity sulci. RCVS should be included in the differential diagnosis of non-aneurysmal SAH.
Several specific clinical, laboratory and neuroimaging findings can assist in differentiating RCVS from SAH related vasospasm or PACNS; diseases with which RCVS is often confused. Acute relapsing thunderclap headaches with or without neurological deficits, presence of cortical SAH over the convexity, early onset of diffuse vasoconstriction involving the distal intracranial vessels (out of proportion to the amount of SAH), near normal CSF analysis, short term reversibility of vasoconstriction (<3 months) and benign clinical outcomes are typical of this entity.
Reversible cerebral vasoconstriction syndromes (RCVS) represent a group of clinical syndromes characterized by severe ‘thunderclap’ headache with or without neurological deficits, benign CSF analysis and multifocal segmental areas of cerebral arterial narrowing or vasoconstriction that usually resolve within weeks. Although there have been reports of non-aneurysmal SAH associated with RCVS, there have been few series demonstrating SAH as an important feature of RCVS. We report a case series of 11 RCVS patients, the majority presenting with a characteristic pattern of focal, cortical subarachnoid blood along the high convexity sulci. Cortical SAH may be a more common presentation of RCVS than previously recognized and RCVS should be included in the differential diagnosis of non-aneurysmal SAH.
SA and TJR contributed equally to this work.
Competing interests None.
Ethics approval This study was conducted with the approval of the Institutional Review Board, University of Michigan Health System.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.