Reversible Cerebral Vasoconstriction Syndrome and Intracranial Hemorrhage

• acute stroke
• angiography
• cerebral vasoconstriction
• headaches
• intracerebral hemorrhage

Reversible cerebral vasoconstriction syndrome (RCVS) is a fairly recently described constellation of clinical and radiological features: sudden, severe (“thunderclap”) headache; transient, multifocal, segmental vasoconstriction of cerebral arteries lasting several weeks to months; and focal neurological symptoms, sometimes with stroke. The features of what is now called RCVS have been repeatedly recognized, especially since an influential early report by Drs Call and Fleming et al¹ published in 1988. A similar syndrome has attracted many different names, depending on whether patients present to specialists in headache, obstetrics, or neurology. These include thunderclap headache with vasospasm, migrainous vasospasm (or “crash” migraine), migrainous angiitis, drug-induced angiitis, postpartum angiopathy, and benign angiopathy of the central nervous system. The label RCVS is an attempt to unify these various descriptions of a distinct clinico-radiological pattern. Nevertheless, there remains uncertainty about the status of RCVS as a true “disease,” which can generally be defined, with progressively less certainty, according to (1) a defined cause; (2) the underlying mechanism (pathophysiology); or (3) a pattern of symptoms and investigation findings (syndrome). As knowledge evolves, some syndromes become more clearly defined by cause or mechanisms, whereas others may not endure. RCVS is a syndrome still in the process of being defined; our understanding and knowledge are increasing as neuroimaging methods improve and more cases are discovered. However, it seems likely that there are many causes (or triggers) for RCVS, and probably more than one underlying pathophysiological mechanism causing abnormal vascular tone. Multiple etiologies for RCVS are suggested by the many and varied associations, including sympathomimetic, serotonergic, or other drugs, pregnancy, and the puerperium,² although no obvious cause can be found in approximately one-half of cases.

It has increasingly been reported that intracranial hemorrhage, especially cortical subarachnoid bleeding, can accompany RCVS.^3–7 In this issue of Stroke, Ducros et al⁸ from the Lariboisiere hospital report on a series of 89 consecutive patients with RCVS diagnosed in their service (with specialist expertise in both headache and stroke), of whom 30 (34%) had at least 1 type of intracranial hemorrhage, compared to only 5 (6%) with cerebral infarction (which tended to occur later than hemorrhage). This large series (for a fairly rare condition) allowed the authors not only to describe the patterns of bleeding but also to investigate the risk factors for hemorrhage in RCVS. The authors found that cortical subarachnoid hemorrhage was the most common pattern, affecting 27 of the 40 cases of hemorrhage; women and those with migraine were at highest risk. Although recruitment in such a specialist referral center is not free of bias, and the diagnostic work-up was not standardized, the findings do suggest that intracranial hemorrhage in RCVS is much more common than previously thought and further challenges the concept of RCVS as a benign self-limiting syndrome.²

What causes the high rate of hemorrhagic complications in RCVS? The authors suggest that the intracranial bleeding is a result of the vasospasm, rather than the trigger. However, the causal direction of the relationship is difficult to prove. This is important because the most commonly understood reason for vasospasm or vasoconstriction in stroke is as a result of subarachnoid hemorrhage (usually, but not always, aneurysmal).⁹ A clinical presentation indistinguishable from RCVS has been documented in the context of aneurysmal subarachnoid hemorrhage.¹⁰ One way to strengthen the argument that vasospasm is a cause of bleeding is to show that the bleeding complications occur after the documentation of vasospasm. However, in the present study, vasospasm was concomitant with intracerebral hemorrhage in 15, and after hemorrhage in 14 patients; the mean time of diagnosis of intracranial hemorrhage was 2.2 days from headache onset and the mean time to diagnosis of vasoconstriction was 6.6 days. The authors argue that the pattern of subarachnoid blood in their cohort (unilateral or bilateral focal cortical) is unlikely to cause multifocal vasospasm. Although this may be true, it also remains unclear how multifocal vasospasm could cause usually rather focal subarachnoid hemorrhage. Speculated mechanisms include reperfusion injury attributable to vasodilatation after severe vasoconstriction or as a consequence of hypertension and disturbed autoregulation (as may occur in another poorly understood syndrome that overlaps with RCVS, posterior reversible encephalopathy syndrome).

So, on the basis of current understanding, what is the value of a diagnosis of RCVS to practicing stroke physicians, and what should be the impact of the increasingly recognized association with intracranial hemorrhage? A diagnosis of RCVS could help clinicians if it could usefully guide treatment or refine prognosis. There is currently no treatment shown to improve outcome in RCVS. It may well be that therapy needs to be tailored to the underlying causes or mechanisms, although current approaches focus on the shared common factor of vasospasm, which logically might reduce the main complications of infarction and hemorrhage. Because of the similarities with postsubarachnoid hemorrhage-related vasospasm, calcium channel blockers are used in some centers;¹¹ high-dose glucocorticoids are sometimes recommended¹² because of radiological similarities with vasculitic disorders and experimental evidence that glucocorticoids reduce vasoconstriction. Other reports emphasize the value of supportive care and careful observation.¹³ We will only know whether any treatments are helpful by conducting randomized controlled trials, which will require clearer diagnostic criteria to allow appropriate patient selection. Challenges here are the heterogeneity and relative rarity of RCVS, and the requirement to demonstrate that vasoconstriction has resolved after 1 or 2 months, at which time the most useful window for intervention may have passed. This can only be resolved by reliably predicting a diagnosis of RCVS without waiting for reversibility to be proven, which currently is not possible.

Finally, with regard to prognosis, RCVS has been considered a useful label on account of the generally benign clinical outcome, in parallel with resolution of vasospasm. The present study has clearly shown that the prognosis is often not benign because cerebral hemorrhage and infarction may occur, with a significant minority unable to work because of persistent deficits at 6 months. Patients therefore cannot be reassured that RCVS is always benign and will wish to be offered effective treatment. However, before treatment for RCVS can be confidently recommended, trials targeting those at highest risk for poor outcome with rational therapeutic approaches are needed; however, such studies must await a clearer understanding of the underlying mechanisms and prognostic indicators.

Therefore, the true value of RCVS as a diagnostic entity in the clinic remains uncertain, because it cannot yet reliably guide treatment or prognosis. The present study, nevertheless, is an important contribution to the field. It has increased our understanding of the association with intracranial hemorrhage and the range of clinical outcomes, and it has begun to unravel the clinical and radiological predictors of hemorrhage. However, the article raises many important questions and should stimulate further work, ideally collaboratively at multiple centers to increase study power and generalizability, on this fascinating and enigmatic clinical syndrome.