Introduction

Aneurysmal subarachnoid hemorrhage (aSAH), a deadly form of hemorrhagic stroke caused by the rupture of a brain aneurysm, affects 30,000 people annually in the United States [1, 2]. Of the 85% of patients who survive the initial hemorrhage, one third develop further brain injury or “delayed cerebral ischemia” (DCI) [3]. DCI is defined as any neurological deterioration presumed to be related to ischemia that lasts for more than an hour and cannot be attributed to other causes [4]. DCI occurs insidiously within the first 2 weeks after subarachnoid hemorrhage (SAH) and accounts for the majority of morbidity and mortality related to aSAH [5–7].

Following early observations that electrical, chemical, and mechanical stimuli could cause cortical arterial constriction [8–10], Robertson [11] in 1949, speculated that arterial spasm was a possible cause of ischemia after aSAH. Two years later, a small study of patients with aSAH demonstrated angiographic evidence of narrowing of the large arteries at the base of the brain, and the connection between vasospasm and ischemia was established [12]. Subsequent studies reported that the onset of vasospasm occurs at least 3 days after aneurysm rupture and coincides with clinical deterioration [13–15]. Vessel narrowing was found to peak at 5 to 14 days and gradually resolved within 2 to 4 weeks [13–15]. In the ensuing years, arterial narrowing became accepted as the mechanism for DCI, and the term “vasospasm” was used interchangeably to describe angiographic vessel narrowing, elevated cerebral blood flow velocities measured by transcranial Doppler ultrasonography (“TCD-vasospasm”), and clinical deterioration (“clinical vasospasm”). However, there remained an imperfect correlation between angiographic vasospasm, TCD-vasospasm, and clinical deterioration, and it has become apparent that arterial vasospasm, elevated cerebral blood flow velocities, and clinical vasospasm are not synonymous [11, 16–18]. Vasospasm is associated with clinical signs in roughly half of the cases, and half of the symptomatic cases culminate in cerebral infarction [3, 19, 20]. Recent consensus statements recommend that the term “vasospasm” be reserved to describe radiological findings only [4, 21]. Indeed, mechanisms other than arterial narrowing might contribute to DCI [22–25]. Elsewhere, we have reviewed both the evidence against vasospasm as the sole cause of cerebral ischemia and alternate mechanisms of DCI [26]. Unsurprisingly, management of patients with aSAH has lagged behind evolving conceptual changes in the pathogenesis of delayed cerebral ischemia. Despite growing recognition that vasospasm is not the sole cause of DCI, current management remains focused on detection and treatment of vasospasm.

In this article, we review the current state of knowledge regarding the treatment options for cerebral vasospasm, defined as arterial narrowing. In the first section, we discuss established therapies that are either evidence-based or supported by expert guidelines. In the second section, we discuss experimental therapies that have theoretical rationales, but have not yet been established as effective.

Although there are no known methods to prevent vasospasm, numerous treatment strategies have been studied. As these strategies are discussed as follows, it is important to bear in mind that assessment of efficacy is difficult due to significant limitations in the extant literature. One major problem is inconsistent use and definitions of the term “vasospasm.” Anther major problem is that vasospasm may be an inappropriate endpoint for therapeutic studies, as radiological infarction correlates better with functional outcome. An expert panel recently suggested that brain infarction and functional outcome be used as primary endpoints in future therapeutic studies [4, 21], and not vasospasm or clinical deterioration.

Established Therapies

Calcium Channel Antagonists

Multiple clinical studies, including 5 double-blind, placebo-controlled trials, have demonstrated that the dihydropyridine calcium channel antagonist, nimodipine, improves clinical outcome in patients with aSAH [27–31]. However, the mechanism underlying this clinical benefit is unclear, as nimodipine does not influence the incidence of vasospasm [32]. Because nimodipine improves clinical outcome, is cost effective, and safe, oral administration (60 mg every 4 h for 21 days) is recommended as the standard of care [33]. Conversely, 2 randomized controlled trials of intravenous nicardipine, another dihydropyridine calcium channel antagonist, demonstrated a reduction in “symptomatic vasospasm,” but had no impact on 3-month clinical outcomes [34, 35]. Intravenous nicardipine is therefore not recommended for routine use [35].

Hemodynamic Augmentation

The cornerstone of medical management of cerebral vasospasm is hemodynamic augmentation, but the strategy has changed with time. Originally, patients were treated with a combination of hypervolemia, hemodilution, and induced hypertension (“triple-H therapy”) to augment cerebral blood flow by increasing cerebral perfusion pressure and lowering blood viscosity [36–38]. However, as the various components of triple-H therapy were studied, the focus of therapy has shifted to euvolemic-induced hypertension.

Although early retrospective studies of hypervolemia as part of a triple-H therapy suggested benefit, subsequent physiological and prospective observational studies suggest harm [39, 40]. Two randomized trials compared prophylactic hypervolemia to normovolemia [41, 42]. Neither found a significant improvement in cerebral blood flow (CBF), incidence of neurological decline, or functional outcome. Patients randomized to hypervolemia had an excess of bleeding, congestive heart failure, and infections. Hypervolemia was also associated with higher cost. Taken together, these studies suggest that hypervolemia does not improve outcome and is associated with increased cardiopulmonary complications. Therefore, hypervolemia is recommended neither for prophylaxis nor for treatment of vasospasm.

Hemodilution has also fallen out of favor as a treatment strategy for vasospasm, however, controversy exists regarding the “optimal” hemoglobin in SAH patients. Although hemodilution increases CBF through improved rheology, it also reduces oxygen-carrying content and results in no net increase in cerebral oxygen delivery [43]. Furthermore, hemodilution is achieved by volume loading, which itself is associated with complications. Whether transfusion of red blood cells (the opposite of hemodilution) improves outcome is unknown. Blood transfusion increases arterial oxygen content and augments cerebral oxygen delivery. However, whether this translates to increased tissue utilization of oxygen is unclear. Furthermore, it has become apparent in the general critical care population that transfusion is associated with mortality, independent of disease severity [44, 45]. Observational studies in the SAH population suggest that while anemia is associated with increased incidence of cerebral infarction and worse functional outcome [46, 47], transfusion of red blood cells is associated with increased risk of poor outcome, extracerebral complications, and increased intensive care unit and total hospital length of stay [48, 49]. Whether the risks of transfusion outweigh the risks of anemia is unclear. A pilot study suggested that randomizing SAH patients to a restrictive versus a liberal transfusion strategy is feasible [50]. A definitive randomized study of the effects of transfusion strategies on outcome has not been performed. It is likely that the “optimal” hemoglobin level varies from patient to patient depending on individual physiology and that no single hemoglobin threshold is optimal for all patients. Presently, neither hemodilution nor red blood cell transfusion is recommended for routine treatment of vasospasm or DCI [21].

Induced hypertension remains the sole recommended component of triple-H therapy. This strategy does not treat or reverse vasospasm, but aims to ameliorate ischemia, a potential consequence of vasospasm. Induced hypertension increases CBF and brain tissue oxygenation, and reverses neurological deficits believed to be from vasospasm [51–53]. These effects appear to be independent of volume status. Several case series suggest that inotropic agents improve CBF and reverse neurological deficits in patients who do not respond to vasopressors [54, 55]. There is scant data to guide choice of vasopressor or inotropic agent and the choice must be made based on empiricism and theoretical advantages that are based on pharmacology and individual patient physiology.

Endovascular Therapies

Endovascular therapies include percutaneous transluminal angioplasty (PTA), and intra-arterial infusion of vasodilators. Although studies have defined the feasibility and safety profiles of these therapies, none has definitively demonstrated an impact on outcome [56–60].

PTA results in durable dilation of vasospastic vessels [61], augments CBF [61, 62], and may reverse neurological deficits [56]; however, its impact on long-term outcome is unclear. Prophylactic PTA (in the absence of arterial narrowing) reduces the incidence of DCI, but is associated with significant complications, resulting in no net benefit [63]. Complications of PTA include vessel occlusion, dissection, and rupture, dislodging of aneurysm clips, and thrombus formation [64–67]. PTA is typically limited to proximal segments of large cerebral vessels, which are easier to access and presumably less prone to rupture due to thicker muscular walls.

Catheter-based delivery of intra-arterial (IA) vasodilators is a commonly used rescue strategy. Papaverine is the most widely studied IA vasodilator. The administration of papaverine is associated with short-lived (roughly 3 h) effects on both vessel caliber and CBF, and multiple side effects, including brainstem depression, elevated intracranial pressure, seizures, hypotension, and occasionally paradoxical worsening of vasospasm [68, 69]. Intracranial hypertension occurs in ≤42% of cases and may be associated with death and poor outcome [70–73]. Therefore, ICP must be monitored during therapy [71]. Due to the short half-life of papaverine, multiple serial injections are frequently required. However, repeated injections have been associated with worse clinical outcomes in several case series [74–76]. Advancements in microcatheter technology have made the super selective catheterization of third and fourth order cerebral vessels possible, allowing for slow infusion of vasodilators, perhaps reducing side effects [70, 71, 77, 78]. Case series report the use of other IA vasodilators, including verapamil, nicardipine, nimodipine, and milrinone [35, 79–83]. Although these agents appear to be safer than papaverine and provide a more durable response, adequate trials that establish safety and efficacy are lacking.

Endovascular therapies are currently recommended as optional rescue strategies for patients that do not adequately respond to medical therapy [21, 33]. The optimal methods, timing, frequency, and combination of endovascular therapies remain unknown.

Experimental Therapies

Statins

Statins have a favorable safety profile and pleiotropic effects that make them an attractive theoretical treatment for vasospasm and DCI. Levels of several inflammatory cytokines, as well as markers of the central nervous system and endothelial cell injury are increased after SAH, and have been correlated with occurrence of vasospasm [84, 85]. Statins downregulate inflammation and upregulate endothelial nitric oxide synthase, which improve cerebral vasomotor reactivity and CBF [86, 87].

Multiple retrospective studies, and 6 small, prospective, randomized controlled studies (4 of which were published in peer-reviewed journals, 2 of which were published as abstracts) investigated the use of statins in aSAH [88–93]. Study results were inconsistent. A recent meta-analysis of the prospective studies demonstrated a small reduction in DCI and a small reduction in mortality, which was lost when patients from the non-peer-reviewed studies were included [94]. Another meta-analysis that included the 4 high-quality trials reported no benefit with regard to TCD-vasospasm, functional outcome, or mortality [95]. A phase III randomized controlled study, Simvastatin in Aneurysmal Subarachnoid Hemorrhage (STASH), is underway. Currently it is recommended that statins be continued in patients who had been on therapy prior to hospitalization, and that therapy may be considered in statin-naïve patients, pending the outcome of ongoing trials [21].

Magnesium Sulfate

Multiple experimental and clinical observations have justified therapeutic studies of magnesium in patients with aSAH. Magnesium inhibits glutamate release and causes cerebral vasodilation through noncompetitive antagonism of voltage-gated calcium channels [96, 97]. Numerous experimental stroke models have demonstrated that magnesium is neuroprotective. In experimental SAH models, magnesium sulfate reverses cerebral vasospasm and infarct volume [98–100]. Hypomagnesemia occurs in >50% of patients with aSAH and is associated with the occurrence of DCI and poor outcome [101].

Six phase II studies have examined the effects of various magnesium regimens on a variety of endpoints [102–107]. In general, these studies suggest that magnesium therapy is safe. A meta-analysis that included 476 patients suggested that magnesium was associated with a reduced risk of poor outcome (death, vegetative state, or dependency; odds ratio, 0.54 [confidence interval, 0.36-0.81]) and no difference in mortality [108]. A single phase III study, The Intravenous Magnesium Sulfate for Aneurysmal Subararchnoid Hemorrhage (IMASH) trial [109] compared magnesium to a placebo in 327 patients. No difference between groups was detected with respect to 6-month outcomes. A second phase III study, Magnesium in Aneurysmal Subarachnoid Hemorrhage (MASH-II) is ongoing.

Other therapies that have been studied include antiplatelet agents [110–112] endothelin antagonists [113], intracisternal fibrinolysis [114, 115], nitric oxide, and nitric oxide donor compounds [116, 117], anticoagulants [118, 119] corticosteroids [120], antioxidants [121, 122], free radical scavengers [123], and serine protease inhibitors [124]. To date, none have proven efficacious in high quality human studies.

Conclusions

The only therapy currently established for the treatment of cerebral vasospasm after SAH is euvolemic-induced hypertension. Although nimodipine improves outcome and is standard of care, it does not prevent or treat vasospasm per se. Rescue therapy with balloon angioplasty and intra-arterial vasodilator therapy is widely used and supported by expert opinion, but its impact outcome is unclear. As the pathophysiology of aSAH is further elucidated and alternate mechanisms of DCI are uncovered, less emphasis may be placed on the treatment of vasospasm. Future therapeutic trials should use consistent terminology and choose surrogate outcome measures, such as cerebral infarction, and perhaps, in the future, chemical biomarkers that correlate more closely with long-term clinical outcome.