Abstract
Objectives
To report demographics and characteristics of reversible cerebral vasoconstriction syndrome (RCVS) in the Korean cohort.
Methods
We prospectively recruited patients with definite (imaging-proven) RCVS and probable (imaging-negative) RCVS who visited Samsung Medical Center between June 2012 and September 2016. Clinical manifestations, neuroimaging, treatment, and clinical outcomes were evaluated in all patients. Characteristics of RCVS without typical causes (“idiopathic RCVS”) were compared with those of RCVS with identifiable causes (“secondary RCVS”). International Classification of Headache Disorders (ICHD)-3 beta criteria for 6.7.3 RCVS and 6.7.3.1 probable RCVS were tested.
Results
A total of 138 patients (104 definite and 34 probable RCVS) were included in this study. Patients with definite RCVS were predominantly female (85.6%) and middle-aged (mean, 50.7 [range, 23–82] years). Probable RCVS was associated with less female predominance (70.6%, p = 0.049), more typical manifestations (p < 0.001), and none of neurological complications. One-hundred and one (97.1%) patients with definite RCVS had headache, but the typical “recurrent and/or triggered” thunderclap headache was reported in only 83 (82.2%). In most patients with definite RCVS (84.6%), RCVS was idiopathic, while only 16 (15.4%) had secondary causes. Compared to those with secondary RCVS, patients with idiopathic RCVS were older (52.8 ± 11.42 vs. 39.1 ± 9.55 years, p < 0.001). Patients with secondary RCVS had more complications than those with idiopathic RCVS (40.5% vs. 12.5%, p = 0.018). Among idiopathic RCVS patients, 33 (37.5%) reported a preceding event or a change in lifestyle, environment, health, or medication within one month before onset.
Conclusion
In our cohort, RCVS was benign and idiopathic in most patients, and occurred frequently in middle-aged women. Manifestations of RCVS were more diverse than previously recognized, and forms without any headache existed. Different genetic, social, and environmental factors should be taken into account to unveil the spectrum and pathophysiology of RCVS.
Keywords
Introduction
Reversible cerebral vasoconstriction syndrome (RCVS) comprises a group of conditions that show reversible multifocal narrowing of the cerebral arteries and acute onset severe headaches (1). With advances in knowledge of the disease, RCVS has been increasingly diagnosed in patients with thunderclap headache (1–3). Typical causes of RCVS are postpartum status and use of vasoactive substances, while it has no identifiable cause in a substantial number of patients (1–3).
To date, only one prospective series of RCVS cases in an Asian population has been reported (4). Compared to Western patients, Asian patients are mostly middle-aged women and have fewer neurological (especially hemorrhagic) complications (4–7). However, it is unclear whether the differences are attributable to genetic, social, or medical factors, or different modes of recruitment. To elucidate the differences, a new, large series of RCVS cases should be reported from another Asian population. The aim of this study was to present characteristics of RCVS in Korean patients and test the current diagnostic criteria. We also investigated factors related to idiopathic RCVS, that is, without identifiable causes.
Methods
We prospectively collected patients with RCVS at Samsung Medical Center between June 2012 and September 2016. Patients were recruited from an outpatient headache clinic, emergency unit, acute stroke unit, and in-patient consultations. All first visit patients in the outpatient headache clinic and acute stroke unit were collected in a prospective registry during the study period. From the emergency unit, patients with acute onset severe headache were referred to a neurologist after excluding aneurysmal subarachnoid hemorrhage (SAH). Our protocol for differential diagnosis of thunderclap headache was described previously (8). Definite RCVS was diagnosed by characteristic angiographic findings and reversibility: 1) Invasive or noninvasive angiographic evidence of multifocal (>2) segmental vasoconstriction in multiple (≥2) intracranial arteries; 2) no evidence of aneurysmal subarachnoid hemorrhage; and 3) reversibility of angiographic abnormalities within 3–6 months after onset (modified from Calabrese et al. (1)). Probable RCVS was diagnosed based on the third beta edition of the International Classification of Headache Disorders (ICHD-3 beta) in patients with normal brain imaging (9).
Under the Korean national health insurance system, there is little prerequisite to visit an outpatient clinic and no prerequisite to visit the emergency room in a tertiary hospital. Patients can visit our hospital at their discretion. As a result, patients can primarily visit our clinic or ER immediately after they had a first headache attack, while some patients presented to us for a second opinion when they had already completed a diagnostic procedure and even treatments at other hospitals. In this study, we reviewed all the outside images and interviewed patients about clinical features at the acute phase of the disease.
Clinical and radiological evaluation
Headache characteristics were collected with a structured questionnaire. When patients first presented to us, we assessed the headache features as thunderclap (onset-to-peak of <1 min) or not, single or recurrent attacks, and whether they were remitted (self-limiting) or persistent at the time of the evaluation. Typical trigger factors were defined as a Valsalva maneuver, defecation, urination, sexual activity, swimming, exercise, showering, shampooing, bending, coughing, and an emotional surge. A cause of RCVS was identified if there was a precipitating condition with a clear temporal association with RCVS. Conditions which have been reported in literature were accepted as a cause of RCVS, for example postpartum status or use of serotonergic or adrenergic medication (1,10–13). RCVS without the above-mentioned conditions was classified as idiopathic. BP surge was determined only in admitted patients.
All patients were asked if there had been any change in lifestyle, environment, medications, or emotions within one month prior to onset. Any transient neurological symptoms that were reported by patients were recorded. Neurological complications were determined when patients showed seizures, ischemic stroke, intracranial hemorrhage, or posterior reversible encephalopathy syndrome (PRES) confirmed by magnetic resonance imaging as defined below (1,10,14,15).
Magnetic resonance angiography (MRA) was the preferred method for diagnosis of RCVS in a majority of patients (n = 127); the remaining 11 patients underwent computed tomographic angiography (CTA). A total of 129 patients underwent brain magnetic resonance imaging (MRI). Among the nine patients (all with definite RCVS) who did not have an MRI scan, the reasons were cost (n = 7), contraindication to MRI (n = 1; tissue expander in the breast), and an alleged diagnosis by CTA and treatment at another hospital (n = 1). Ischemic stroke, SAH, and PRES were diagnosed with diffusion weighted-imaging, gradient-echo, and fluid-attenuated inversion recovery imaging, respectively.
Follow-up of patients with RCVS
A follow-up visit was scheduled at one, three, and six months after RCVS was diagnosed. Nimodipine was used as standard treatment in 100 (72.5%) patients, at a dose of 60–90 mg per day. Any recurrence of thunderclap headache or development of new neurological deficits after treatment was recorded at every visit. The reversibility of intracranial vessels was determined by follow-up MRA or CTA at 3–6 months after onset.
Statistical analyses
Characteristics were compared between definite and probable RCVS. Other analyses were performed only in patients with definite RCVS. Statistical analyses were performed using SPSS version 18.0 software (SPSS Inc., Chicago, IL, USA). Variables were presented as proportion, mean ± SD or median (interquartile ranges). The chi-square and Fisher’s exact tests were used for categorical variables to assess the characteristics of each group. Two-tailed p-values <0.05 were considered significant.
Results
We collected 138 patients with RCVS (104 with definite RCVS and 34 with probable RCVS) from the outpatient headache clinic (n = 80), emergency unit (n = 50), acute stroke unit (n = 2), and inpatient consultation (n = 6).
Demographics and characteristics of RCVS
Demographics and characteristics of definite and probable RCVS are summarized in Table 1. In patients with definite RCVS, the mean age was 50.7 years (range, 23–82 years), with a female preponderance (85.6%). In most patients (84.6%), the cause of RCVS was not identified (idiopathic RCVS), while only 16 (15.4%) had secondary causes. Thunderclap headache was the presenting symptom in most patients (89.1%). However, 14 (13.5%) did not have thunderclap headache; 11 (10.6%) presented with severe headache of gradual onset, and three (2.9%) did not have any headache. Among 11 patients with gradual onset of headache, nine clearly remembered the time from onset to peak intensity, with the median of 5 min (range 2–120). The remaining two patients reported a mild preceding headache for the day before the onset, and awakened from sleep with severe headache. Six (4.3%) patients reported transient neurological deficits such as left arm hypesthesia, diplopia with ptosis, and transient visual field defect, all of which occurred during or before the headache onset. Neurological complications were present in 16 (15.4%) patients, including seizure (2.9%), ischemic stroke (6.3%), cortical SAH (7.4%), and PRES (5.3%). Typical imaging findings are shown in Figure 1.
Typical neuroimaging in reversible cerebral vasoconstriction syndrome and associated complications. (a–c) Serial MR angiogram at (a) seven days, (b) 11 days, and (c) one month after onset. (d–f) Examples of neurological complications in RCVS: (d) Cortical subarachnoid hemorrhage seen in the gradient echo sequence image; (e) vasogenic edema suggestive of posterior reversible encephalopathy syndrome in the T2-weighed image; and (f) ischemic infarction in the territory of the left anterior cerebral artery visualized in a diffusion-weighted image. Demographics and characteristics of reversible cerebral vasoconstriction syndrome. Numbers are presented as mean ± SD, median (interquartile range), or number (%). RCVS: reversible cerebral vasoconstriction syndrome; BP: blood pressure; SAH: subarachnoid hemorrhage; PRES: posterior reversible encephalopathy syndrome Others: vertebral artery dissection, carotid artery bypass Evaluated in 101 patients with definite RCVS who had headache as presenting symptoms and all patients with probable RCVS Analyzed in 83 patients admitted in our hospital during the acute phase Some patients underwent their first scan at other hospitals before visiting our hospital.
Compared to definite RCVS, probable RCVS was associated with less female predominance (70.6%, p = 0.049) and more current smoking (14.7%, p = 0.041; Table 1). Headache characteristics were more typical in probable RCVS, because the diagnostic criteria for probable RCVS are stricter than that of definite RCVS. However, no patients with probable RCVS showed focal deficits or neurological complications. BP surge was present in 47.1% among admitted patients. Age of onset, associated medical conditions, premorbid migraine, and median time from onset to neuroimaging were not different between patients with definite and probable RCVS. When classified into weeks, patients with probable RCVS had their neuroimaging at a wider range of timing, with a trend of more delay than patients with definite RCVS (p for trend = 0.048).
Diagnosis of RCVS in patients with or without headache
ICHD-3 beta criteria testing in patients with headache and definite RCVS.
Only 101 patients with definite RCVS presenting with headache were included in this analysis
ICHD-3 beta criteria testing in patients with probable RCVS.
Three patients with definite RCVS did not have any headache. Serial neuroimaging findings showed reversible stenosis in multifocal intracranial arteries in these patients (Figure 2). A 36-year-old female (case 1) was initially admitted with multiple infarctions in the right medulla and cerebellum, which is attributed to dissection of the right vertebral artery. Intravenous thrombolysis was performed. The patient reported increased dizziness and right-sided ptosis immediately after thrombolysis. Brain MRI was followed up at three days after onset and revealed new infarctions and multiple stenosis in intracranial arteries other than the right vertebral artery (Figure 2(a)), which were much improved after three months (Figure 2(b)). Another patient (case 2) was a 48-year-old female who developed sudden-onset spontaneous vertigo that recurred every day. MRI taken at another hospital showed multiple intracranial arterial stenosis (Figure 2(c)). A follow-up CTA at three months after onset was normalized (Figure 2(d)). Case 3 was a 37-year-old woman. She had a seizure three hours after delivery. Initial CTA and MRA showed progressive vasoconstrictions in multiple intracranial arteries (Figure 2(e); MRA only), which was improved at three months after onset (Figure 2(f)).
Neuroimaging findings of RCVS without headache. (a–b) A 36-year-old female who had dissection of the right vertebral artery developed multiple mild stenosis in multifocal cerebral arteries (arrows in (a)). Follow-up MR angiogram showed normalization (arrows in (b)). (c–d) A 48-year-old female developed sudden-onset spontaneous recurrent vertigo. MR angiogram showed multiple severe stenosis in intracranial arteries (arrows in (c)), which were improved in the follow-up CT angiogram (arrows in (d)). (e–f) A 37-year-old female had a seizure on the day of delivery. Her MR angiogram showed progressive vasoconstrictions in multiple cranial arteries (arrows in (e)), which were normalized at three months after onset (arrows in (f)).
Conditions precipitating RCVS
Causes of definite RCVS.
Only patients with definite RCVS were included in this analysis
Vertebral artery dissection that clearly preceded RCVS
Preceding conditions associated with reversible cerebral vasoconstriction syndrome.
Only patients with definite RCVS were included in this analysis
Factors related to idiopathic RCVS
Characteristics of idiopathic vs. secondary RCVS.
Only patients with definite RCVS were included in this analysis
Determined in 95 patients (80 with idiopathic RCVS and 15 with secondary RCVS) who underwent brain magnetic resonance imaging
Assessed in 101 patients (87 with idiopathic RCVS and 14 with secondary RCVS) presenting with headache
Analyzed in 68 patients admitted in our hospital during the acute phase
RCVS: reversible cerebral vasoconstriction syndrome; BP: blood pressure; SAH: subarachnoid hemorrhage; PRES: posterior reversible encephalopathy syndrome
Treatment of RCVS
Nimodipine was used as first-line treatment in 100 (72.5%) patients. Of these, recurrence of thunderclap headache was investigated with regard to the time after nimodipine therapy in 90 patients. The other 10 patients could not recall the date of recurrence after starting nimodipine therapy. A total of 62 (68.9%) patients achieved complete remission of their thunderclap headache immediately after taking nimodipine. In 11 (12.2%) patients, headache remitted within one week during nimodipine therapy. The remaining 17 patients had remission within four weeks. No neurological complications occurred after starting nimodipine therapy.
Discussion
In our analysis of a new, large series of Korean patients with RCVS, 1) most patients were middle-aged women and the rate of neurological complications was low, 2) probable RCVS was diagnosed in one-fourth of all RCVS patients, 3) not all patients with definite RCVS had typical manifestations, and 4) idiopathic RCVS was more common than reported in Western cohorts. Various, potentially associated preceding conditions were found in about half of the idiopathic RCVS cases.
Most patients in our series were middle-aged women. Women are more affected by RCVS than men, with varying proportions in different RCVS cohorts (4–7). Among these, a Taiwanese cohort showed the highest female predominance (4). In addition, similar to the Taiwanese cohort, most of our patients were middle-aged (4), whereas Western patients were younger (5,6). The demographic features of Asian (i.e. Korean and Taiwanese) patients might be attributed to the rate of secondary causes. Singhal et al. showed that patients with drug-related or postpartum RCVS are younger than those with idiopathic RCVS (6). Male patients are also younger and more likely to have vasoactive drug use as a cause of RCVS (5,6). In contrast, both the Taiwanese and our studies identified a secondary cause in only one out of 10 patients, which might affect the demographic features. Although not reported as a subgroup, female patients with idiopathic RCVS might also be middle-aged in Western cohorts when postpartum patients, who are usually aged below 40, are excluded (5–7,15). In addition, genetic factors might be associated with the demographics in Asian patients (15).
Probable RCVS was diagnosed in one-fourth of all RCVS patients. While probable RCVS was still associated with female predominance, males were more frequent in probable than definite RCVS. This can be attributed to females having more precipitating conditions such as delivery or antidepressant medications. However, the result remained the same when analyzed only in patients with idiopathic RCVS. Then, females might have more of a genetic tendency to develop vasoconstrictions within major divisions of cerebral arteries. The positive association between smoking and probable RCVS may simply reflect increased male prevalence. However, it cannot be excluded that smoking can trigger RCVS-like phenomena only in small vessels, which is invisible in noninvasive imaging. Interestingly, the timing of neuroimaging in patients with probable RCVS was not earlier, rather it was delayed compared to patients with definite RCVS. This indicates that very early imaging was not the only explanation for negative imaging findings in probable RCVS. In fact, vasoconstrictions in middle-sized arteries might not be sufficient to sensitively diagnose RCVS. As we recently reported, blood-brain barrier disruption can be present in RCVS patients with normal angiograms, which may indicate a possible involvement of small arterioles and capillaries (8).
As noted earlier, our study showed a high rate of idiopathic RCVS. Idiopathic RCVS was more benign in our study, showing fewer neurological complications and more self-limited headaches, compared to secondary RCVS. The rate of idiopathic RCVS was highest in the Taiwanese cohort, while it was much lower in Western cohorts (4–7). Vasoactive drug use and a postpartum state have been reported in more than half of the Western patients (5,6,16). Social and medical factors may be responsible for this disparity. First, illicit drug use is a rare factor in RCVS in Korea and Taiwan (17), while it is a major cause of RCVS in studies of Western patients (5–7). Second, the national healthcare system of Korea and Taiwan allows easy access to university hospitals with a relatively low cost for almost all residents (18). Consequently, mild and benign cases might be more easily identified (6).
We identified minor preceding conditions in one-third of patients with idiopathic RCVS (19). A various change in the lifestyle, environment, health, or medication preceded the onset of RCVS. Some of those changes could induce sympathetic hyperactivities, while others seemed not related to a sympathetic response. Altered sympathetic response to physiologic events, abnormal vascular response to increased sympathetic tone, or both might be present in susceptible patients. Further research is needed to identify additive or interactive roles of genetic and environmental factors in the pathophysiology of RCVS.
We found a wide variability in the manifestation of RCVS. About 10% of patients did not have thunderclap headache, and three did not have headache at all. There are increasing reports of RCVS without thunderclap headache (3,20). One-third of patients were diagnosed after a single thunderclap headache without recurrence. Headaches were triggered by typical precipitants only in two-thirds of patients, while the remainder reported a spontaneous headache or neurological event. In addition, we had a patient aged 82, which might be the oldest age reported in the literature. She showed recurrent thunderclap headaches, which were triggered by bending and shampooing and reversible vasoconstrictions. Furthermore, infarct locations in our patients were not restricted to the watershed area, which was previously considered typical for RCVS-associated infarction (11). The presentation of RCVS might be more diverse than previously recognized.
ICHD-3 beta classification was applicable for all patients with definite RCVS presenting with headache (9). However, its sensitivity and specificity has not been tested yet. As Wolff and Ducros discussed, any headache that led to angiography can be sufficient for the diagnosis based on ICHD-3 beta (3,9). Our results suggest that not all patients have typical headaches. Also, the ICHD-3 beta classification requires evaluations at different time points. For patients who visited the hospital immediately after their first headache, headache characteristics such as recurrence or triggering factors might not be properly assessed at the date of the first visit. We are now conducting a validation study to test ICHD-3 beta criteria for RCVS and probable RCVS serially at the first and follow-up visits.
In our study, the rate of neurological complications was lower than that in Western cohorts (4–7). It is difficult to estimate the rate of hemorrhagic complication unless all the patients with intracranial hemorrhages undergo a screening for RCVS. However, the mode of patient recruitment is also responsible for estimating complication rates. In retrospective studies (15,21), most subjects were in-patients who were evaluated by a stroke specialist, leading to the possibility of more inclusion of complicated cases. In our study, patients were screened equally from headache clinics, emergency departments, and stroke units, showing a broader spectrum of the disease. In addition, genetic factors should be considered. When our data are combined with that of the Taiwan cohort, Asian patients appear to have more idiopathic but benign RCVS. However, it is not clear whether Asian patients have a greater risk of mild RCVS, even in the absence of a causative condition, or whether they are more resistant to the complications of RCVS. Asians have greater risk of intracranial atherosclerosis, dissection of intracranial arteries, and moyamoya disease. Polymorphism of the RNF213 gene predisposes to Moyamoya disease and intracranial atherosclerosis, and is frequently found in East Asians (22–26). Similarly, a certain genetic factor may modulate disease susceptibilities and complication rates of RCVS in different ethnicities. Further studies are warranted to determine genetic factors that are protective for neurologic complications or responsible for the susceptibility for RCVS.
Four out of five patients showed an excellent response to nimodipine treatment in our study. The response rate to nimodipine was also similar in a Taiwanese cohort (27), while Western studies reported different rates of response to nimodipine (5–7). Genetic susceptibility might affect the treatment outcome. Conversely, Asian patients with RCVS may have a more benign disease course, regardless of treatment. Future randomized controlled studies using different ethnic populations should be performed to determine the effect of nimodipine treatment in different populations and clinical settings.
There are several strengths of this study. First, we prospectively included a large number of patients with RCVS. This is the first study of its kind from Korea, and the second from the Asian population. Second, the mode of recruitment was comprehensive. All patients from the headache clinic and stroke unit were prospectively collected by the registry. In the emergency unit, all patients with thunderclap headache were referred to a neurologist after excluding aneurysmal SAH. Third, we reported preceding conditions that have not been recognized as conventional causes of RCVS. The limitation of this study was the single institution-based recruitment with a single ethnicity (i.e. Koreans); therefore, external validation is warranted before generalizing the findings of this study. In addition, we did not routinely repeat the angiogram for patients with normal angiograms, because many patients with probable RCVS showed a typical self-limiting course. Patients can be falsely classified into probable RCVS because the angiogram can be normal during the first few days. Fourth, our claims on “RCVS without headache” have not been validated in other studies. Whether we can refer all kinds of reversible vasoconstriction to RCVS should be determined in future. Also, one case (case 2) was diagnosed by using different imaging modalities with suboptimal quality. Finally, neurological complications were determined by MRI. Although there was no difference in characteristics in patients who did not undergo an MRI scan, it is possible that patients with neurological complications would be more likely to have brain MRIs because the Korean national health insurance reimburses the cost of MRI only when neurological deficits other than headache are present. Thus, the rate of neurological complications might be over-estimated in our study.
Conclusions
In conclusion, our patients with RCVS show distinctive demographics and characteristics; more cases are benign and idiopathic, and it occurs most frequently in middle-aged women. Manifestations of RCVS are more diverse than previously recognized. This study’s results may broaden our understanding of the characteristics, clinical courses, and preceding conditions of RCVS. Different genetic, social, and environmental factors should be taken into account to unveil the spectrum and pathophysiology of RCVS.
Footnotes
Clinical implications
In our large cohort of Korean patients, RCVS was benign and idiopathic in most cases, and occurred most frequently in middle-aged women.
Probable RCVS is associated with more typical presentations, less female predominance, and increased rate of current smoking. Timing of neuroimaging was rather delayed in patients with probable RCVS.
ICHD-3 beta criteria for RCVS are applicable in all patients with definite RCVS whose presenting symptoms were headache. However, about 20% of patients did not have a characteristic headache as suggested in the ICHD-3 criterion C2 (recurrent and/or triggered thunderclap headache), leaving the specificity of diagnostic criteria in doubt.
While typical causes of RCVS such as illicit drug use were not frequent in our patients, various preceding conditions are found in more than one-third of idiopathic RCVS, which might be potentially associated with the development of RCVS.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.