Centripetal Propagation of Vasoconstriction at the Time of Headache Resolution in Patients with Reversible Cerebral Vasoconstriction Syndrome

In this retrospective cohort study, the authors evaluated 16 patients diagnosed with reversible cerebral vasoconstriction syndrome who underwent MR imaging, including MRA, within 72 hours of RCVS onset (initial MRA) and within 48 hours of thunderclap headache remission. In 14 of the 16 patients (87.5%), centripetal propagation of vasoconstriction occurred from the initial MRA to remission of thunderclap headache, with typical segmental vasoconstriction of major vessels (M1, P1, A1). The authors conclude that there is evidence of centripetal propagation of vasoconstriction on MRA performed at the time of remission of the thunderclap headache, and this time point may represent a useful opportunity to diagnose RCVS with greater confidence. BACKGROUND AND PURPOSE: Reversible cerebral vasoconstriction syndrome is characterized by thunderclap headache and diffuse segmental vasoconstriction that resolves spontaneously within 3 months. Previous reports have proposed that vasoconstriction first involves small distal arteries and then progresses toward major vessels at the time of thunderclap headache remission. The purpose of this study was to confirm centripetal propagation of vasoconstriction on MRA at the time of thunderclap headache remission compared with MRA at the time of reversible cerebral vasoconstriction syndrome onset. MATERIALS AND METHODS: Of the 39 patients diagnosed with reversible cerebral vasoconstriction syndrome at our hospital during the study period, participants comprised the 16 patients who underwent MR imaging, including MRA, within 72 hours of reversible cerebral vasoconstriction syndrome onset (initial MRA) and within 48 hours of thunderclap headache remission. RESULTS: In 14 of the 16 patients (87.5%), centripetal propagation of vasoconstriction occurred from the initial MRA to remission of thunderclap headache, with typical segmental vasoconstriction of major vessels. These mainly involved the M1 portion of the MCA (10 cases), P1 portion of the posterior cerebral artery (10 cases), and A1 portion of the anterior cerebral artery (5 cases). CONCLUSIONS: This study found evidence of centripetal propagation of vasoconstriction on MRA obtained at the time of thunderclap headache remission, compared with MRA obtained at the time of reversible cerebral vasoconstriction syndrome onset. If clinicians remain unsure of the diagnosis during early-stage reversible cerebral vasoconstriction syndrome, this time point represents the best opportunity to diagnose reversible cerebral vasoconstriction syndrome with confidence.

R eversible cerebral vasoconstriction syndrome (RCVS) is characterized by thunderclap headache (TCH) and diffuse segmental constriction of cerebral arteries that resolves spontaneously within 3 months. 1,2 RCVS can be spontaneous or evoked by various factors such as postpartum status, vasoactive drugs, catecholamine-secreting tumors, immunosuppressants, or blood products. 3,4 RCVS is poorly understood and has recently been estimated to be more common than previously thought. 2 Unpredictable and transient failure of the regulation of cerebral arterial tone through sympathetic overactivity is thought to play a role in the development of RCVS. 1,5 TCH is usually the first symptom of RCVS, and typically recurs for 1-2 weeks after onset. [3][4][5] The final TCH occurs at a mean of 7-8 days after onset, and all noteworthy headaches are generally gone within 3 weeks after onset. 4,6,7 Conventional DSA was crucial to the diagnosis of RCVS in early case reports and typically showed diffuse segmental vasoconstrictions involving major vessels, like a "string of beads" or "string of sausages." 4 More recently, MRA has been widely used as a noninferior tool to evaluate vasoconstriction in patients with RCVS. 3,4,6 Findings of typical segmental vasoconstriction of major vessels are not obtained in about one-third of patients with RCVS if the initial MRA is performed within 10 days of headache onset. 4,6,8 Ducros et al 4,9 postulated that the arterial abnormalities first involve small distal arteries and then progress toward major vessels.
This study retrospectively investigated whether centripetal propagation of vasoconstriction on MRA at the time of TCH re-mission was evident compared with MRA at the time of RCVS onset.

Patient Population
Patients diagnosed with RCVS in Tokai University Hachioji Hospital according to the following 5 criteria for RCVS were included in our data base: 1) acute and severe headache (often TCH) with or without focal deficits or seizures; 2) a uniphasic course without any new symptoms Ͼ1 month after clinical onset; 3) segmental, multifocal vasoconstriction of cerebral arteries as shown by DSA or indirect angiography (eg, MRA or 3D CTA); 4) no evidence of aneurysmal SAH; and 5) complete or substantial normalization of arteries on follow-up DSA or indirect angiography performed within 12 weeks after clinical onset. 2,4 CSF analysis was not routinely performed on patients in this study.
Subjects were selected from among the 39 patients entered into the RCVS data base at Tokai University Hachioji Hospital between October 2010 and April 2015. Ten of these patients who did not undergo initial MRA/MR imaging within 72 hours of RCVS onset were excluded because they were admitted in the subacute phase (Ն3 days after onset). In addition, 13 patients with RCVS who did not undergo MRA/MR imaging within 48 hours of TCH remission were excluded because they did not visit an outpatient clinic at the time of TCH remission (subacute phase) due to their comfortable medical condition. The remaining 16 patients with RCVS who underwent MR imaging within 72 hours of RCVS onset, within 48 hours of remission of TCH, and 3 months after RCVS onset were included in this retrospective study.
Clinical and imaging records were evaluated retrospectively. Study approval was obtained from the Institutional Review Board for Clinical Research at Tokai University School of Medicine.

Imaging Protocol
We differentiated primary angiitis of the central nervous system from RCVS on the basis of an insidious onset, low incidence of TCH, mostly irreversible angiographic irregularities, and inflammatory reactions evident in the CSF. 2

Definition of Variables
"TCH" was defined as a severe pain peaking within seconds. The presence of TCH was diagnosed by a thorough interview of the patient. "TCH remission" was defined as the time at which the last TCH improved. "Hypertensive emergency" was defined as sys-tolic blood pressure of Ͼ180 mm Hg or diastolic blood pressure of Ͼ120 mm Hg.
"Cortical SAH" was defined as a nontraumatic, nonaneurysmal SAH characterized by clots distributed over the cerebral convexity, sparing the basal cistern. 11 In patients with cortical SAH, conventional DSA or 3D CTA or both were routinely performed to rule out other causes of SAH, such as aneurysm or vessel malformation. "Posterior reversible encephalopathy syndrome" was defined as a characteristic MR imaging pattern of vasogenic edema with hemispheric boundary zones of hyperintensities on FLAIR, with increased ADC values affecting the cortex and subcortical and deep white matter to varying degrees. 2,12 Localization of vasoconstriction of the cerebral arteries was evaluated on MRA. "Centripetal propagation of vasoconstriction" was defined as a vasoconstriction that progressed from the distal arteries at the time of TCH (ie, on MRA performed within 72 hours of RCVS onset) to the major cerebral arteries of the circle of Willis (defined as the ICA, the A1 portion of the anterior cerebral artery, and/or the P1 portion of the posterior cerebral artery) and the M1 portion of the MCA, basilar artery, and vertebral artery on MRA performed within 48 hours of TCH remission. A "hyperintense vessel sign" was defined as a focal or tubular hyperintensity in the subarachnoid space on FLAIR 13 that normalized within 12 weeks of clinical onset.
MR imaging findings were interpreted by at least 2 senior stroke neurosurgeons (M.S. and S.O., with 32 and 27 years of experience, respectively). In the case of disagreement between raters, diagnosis was obtained by consensus decision. Outcomes were assessed at 3 months by using the mRS.

Treatment Protocol
Vasoactive medications were stopped in all patients. Symptomatic analgesic treatment was used in all patients without a standard protocol. Oral administration of lomerizine hydrochloride was recommended for prevention of cerebral vasoconstriction. Steroids were avoided. For patients with severe TCH, a low dose of propofol was infused intravenously at 30 -50 mg/h. For patients who experienced a hypertensive emergency, nicardipine was used with intravenous infusion of the dose adapted to normalize blood pressure levels.

Clinical Features and Associated Lesions
The On-line Table summarizes the clinical characteristics of the 16 patients with RCVS who underwent MR imaging within 72 hours of RCVS onset and within 48 hours of the remission of TCH. Ten of these 16 patients had a history of migraine before the onset of RCVS, and 3 of these 10 patients experienced an aura. All 10 patients who had a history of migraine stated that their headaches were completely different and more severe than their usual attacks. The 16 patients with RCVS included 2 postpartum women. Six of the 16 patients were using a vasoactive substance such as a selective serotonin reuptake inhibitor or triptans. Overuse of triptan was identified in only 1 case. Two of the 16 patients experienced bathing-related RCVS. Four showed a "blood pressure surge" defined as a hypertensive emergency during the peak of the TCH, and no patients displayed permanently elevated blood pressure. The 4 patients with a hypertensive emergency showed vasoconstriction of the distal MCA and posterior cerebral artery on initial MRA.
All patients reported multiple episodes of TCH. Because the clinical records were insufficient, we were unable to investigate the exact number of events of TCH retrospectively. TCH was the only symptom of RCVS in 13 patients (81.2%).
A mean delay of 0.69 days (range, 0 -2 days) was seen between RCVS onset and initial MR imaging. In 9 of the 16 patients (56.2%), initial MR imaging was performed on day 0 (ie, within 24 hours of onset). In the remaining 7 patients, initial MR imaging was performed 1-2 days after onset.
Stroke as a complication of RCVS occurred in 6 of the 16 patients (37.5%). Evidence of cortical SAH was found on MR imaging performed on days 0 -2 in 5 patients. Conventional angiography, 3D CTA, MR imaging, and MRA revealed no other causes of SAH. No patients had intracerebral hemorrhage or microbleeds on T2*-weighted imaging. One patient developed MCA/posterior cerebral artery watershed infarction on day 0. Evidence of posterior reversible encephalopathy syndrome was found on MR imaging performed on day 2 in 1 patient. Findings of brain lesions on follow-up MR imaging are summarized in the On-line Table. Hyperintense vessel signs on FLAIR imaging with the initial MR imaging were found in 5 patients. Of these 5 patients, 4 displayed evidence of some kind of stroke lesion on initial MR imaging as a complication of RCVS (On-line Table). Among these 5 patients who showed findings of vessel hyperintensity on initial MR imaging, the findings remained on MR imaging at the time of TCH remission in 2 patients.

Centripetal Propagation of Vasoconstriction
The Table presents

Clinical Features of the Chronic Stage
By definition, all 16 patients with RCVS displayed confirmed multifocal vasoconstriction that resolved within 12 weeks, as assessed by MRA. At the 3-month follow-up, the mRS scores of all 16 patients with RCVS were zero, and all patients were able to return to their prior professional activities.

Clinical Course and Mechanisms of TCH of RCVS
In patients with RCVS, TCH is the presenting symptom and usually recurs during ϳ1 week. In general, TCH occurs before any detectable vasoconstriction and vasoconstriction outlasts TCH. 4 Whereas TCH, cortical SAH, intracerebral hemorrhage, and posterior reversible encephalopathy syndrome represent early manifestations that lead to a suspicion of RCVS, ischemic stroke can occur as late as 1-2 weeks after clinical onset, sometimes when the headache has resolved or after the patient has been discharged. 3,4,6,7,9 Dysregulation of cerebral vascular tone is known to result

Mechanism of Centripetal Propagation of Vasoconstriction in RCVS
Ducros 2 suggested that vasoconstriction of the second and first segments of large cerebral arteries might represent a reaction to distal blood-flow abnormalities, increasing during the ensuing 1 or 2 weeks. However, the precise mechanisms underlying centripetal propagation of vasoconstriction in RCVS remain unclear. Centripetal progression of vasoconstriction from the manifestation of recurrent TCH to the remission of headaches has not been precisely delineated by using neuroradiologic imaging modalities such as MRA. Chen et al 6 showed that vasoconstriction is maximal at TCH remission. However, they did not mention the centripetal propagation of vasoconstriction at TCH remission. Their initial MRA was performed at a mean of 10.4 Ϯ 7.0 days (median, 9 days; range, 1-30 days) after TCH onset. We consider that the lower ratio of patients who underwent initial MRA/MR imaging within 72 hours of RCVS onset would have largely prevented identification of the centripetal propagation of vasoconstriction at TCH remission. Another review by Chen et al 14 described that only a small number of their patients showed cen-tripetal propagation of vasoconstriction, citing their initial article. 6 On the other hand, the purpose of this study was to prove centripetal propagation of vasoconstriction on MRA at the time of TCH remission, compared with MRA at the time of RCVS onset. The present investigation showed centripetal propagation of vasoconstriction from RCVS onset to the remission of TCH on MRA in 14 of the 16 patients with RCVS. Although the pathophysiologic mechanisms behind the centripetal propagation of vasoconstriction are unknown, this phenomenon is an important neuroradiologic finding for diagnosis.

Neuroradiologic Diagnosis of RCVS
RCVS should be considered for patients in the acute phase of cortical SAH or cryptogenic stroke with recurrent TCH. 4 However, in the acute phase of RCVS, clinicians are likely to overlook a diagnosis of RCVS in patients who have no neuroradiologic findings of parenchymal lesions such as posterior reversible encephalopathy syndrome, intracerebral hemorrhage, or SAH on MR imaging because clinicians should diagnose RCVS only from a finding of vasoconstriction on MRA. The diagnosis of RCVS is reportedly missed in one-third of patients who undergo initial MRA within 10 days of headache onset. 8 The diagnosis of RCVS in the acute phase by using MR imaging remains difficult because the vasoconstricted vessels are located peripherally at this time.
Patients with RCVS are predominantly women 50 years of age or younger. MRA of healthy young individuals usually depicts the MCA, posterior cerebral artery, and anterior cerebral artery branches distal to the circle of Willis and the first branch points of major vessels relatively well. A diagnosis of RCVS should thus be considered in young women reporting TCH if the depiction of peripheral vessels on MRA is poor. Furthermore, at the time of TCH remission, if a clinician confirms centripetal propagation of vasoconstriction with typical segmental vasoconstriction of the circle of Willis vessels, RCVS can be diagnosed with confidence.
Recently, a hyperintense vessel sign on FLAIR images has been reported as one of the neuroradiologic findings of vasoconstriction of a distal artery in the acute phase of RCVS. 8,15 The hyperintense vessel sign could simply represent slow flow in collateral vessels distal to a proximal narrowing. Chen and Wang 8 reported the presence of a hyperintense vessel sign on FLAIR images in 22% of patients with RCVS, and this sign in patients with RCVS was associated with more severe vasoconstriction that may progress to ischemic stroke or posterior reversible encephalopathy syndrome. In our study, the incidence of a hyperintense vessel sign on initial MR imaging was also low (5 of 16 patients, 31%). This sign tended to be more frequent in patients with RCVS who experienced stroke. We considered that a hyperintense vessel sign might be an indicator of the severity of RCVS, rather than a diagnostic tool.

Limitations
TOF-MRA is a widely available, noninvasive technique that requires no radiation exposure and no administration of contrast material. However, the limitations of TOF-MRA include a lower spatial resolution than DSA and CTA 16 and the presence of flowinduced artifacts. These limitations may diminish the ability to depict vasoconstriction in small vessels.
Chen et al 6 reported that vasoconstriction of M1 and P2 is important for the development of posterior reversible encephalopathy syndrome and ischemic stroke in patients with RCVS. Although the present study included only 1 patient with posterior reversible encephalopathy syndrome and 1 patient with cerebral infarction, a high proportion of patients displayed vasoconstriction of M1 or P2. This study thus found no relationship between the location of vasoconstricted vessels and the type of stroke. We believe that a more detailed examination of the relationships between the locations of constricted vessels and the type of stroke is necessary in patients with RCVS.
This was a retrospective study of a small group of patients, and prospective studies with a greater number of cases are necessary.

CONCLUSIONS
Compared with MRA at the time of RCVS onset, this study found evidence of centripetal propagation of vasoconstriction on MRA performed at the time of remission of TCH. This time point therefore may represent a useful opportunity to diagnose RCVS with greater confidence if the clinician remains unsure of the diagnosis during the early stages of RCVS.