Aneurysmal subarachnoid hemorrhage (SAH), a type of hemorrhagic stroke due to rupture of an intracranial aneurysm, affects approximately 30,000 Americans each year and has a mortality rate of nearly 45%. At least 15% of people with SAH die before reaching the hospital. Of those who survive, a substantial proportion is left with significant disability. Prompt diagnosis, treatment, and anticipation of complications may improve outcome. Case fatality rates have been declining and functional outcomes have been improving. These changes in the natural history of SAH may be attributable to early aneurysm repair and aggressive management of medical complications. This chapter reviews major clinical management points and discusses the relevant literature.
Emergency Setting
In the emergency setting, once the diagnosis of SAH has been established, initial goals are to stabilize the patient’s airway, breathing, and circulation. Early referral to a large-volume center with experienced vascular neurosurgeons, neuroendovascular specialists, and dedicated neurointensivists should be considered. Four studies have demonstrated that hospital volume of SAH patients and procedural experience correlate with improved mortality.
SAH-Related Complications
Rebleeding
Aneurysmal rebleeding is one of the most serious initial threats to the patient. The incidence may be as high as 30%, with the greatest risk (roughly 4%) during the first 24 hours. Temporizing medical measures are used to reduce the risk of rebleeding until the culprit aneurysm is excluded from the circulation through surgical or endovascular means.
Medical Measures
Bed rest does not alter the incidence of rebleeding, but it has become a standard practice. Blood pressure control is widely recommended to reduce the risk of aneurysmal rebleeding. The benefit of blood pressure reduction must be weighed against the risk of precipitating cerebral ischemia. Although there are no prospective studies that demonstrate the efficacy of antihypertensive therapy, retrospective data suggest an association between hypertension and aneurysmal rebleeding. Ohkuma et al. found a statistically significant increase in the incidence of prehospitalization rebleeding in patients whose systolic blood pressure was greater than 160 mm Hg. Interpretation of these data is confounded by variable times at which rebleeding was observed and variations in antihypertensive therapies. Because rebleeding may be related to aneurysm expansion, which is largely dictated by changes in transmural pressure, surges in blood pressure may be more important than absolute levels of blood pressure. Therefore, it is reasonable to treat extreme hypertension and to minimize blood pressure lability with a short-acting, intravenous agent that has a predictable dose–response relationship. Premorbid baseline blood pressure should be taken into consideration for setting blood pressure goals, and hypotension should be avoided.
Antifibrinolytics
Antifibrinolytic agents such as tranexamic acid and epsilon-aminocaproic acid have been well studied. Ten prospective randomized studies (1904 participants) have been performed ( Table 63-1 ) and were included in a 2013 Cochrane Review. In sum, death and poor outcome (death, vegetative state, or severe disability) were not influenced by treatment. It appears that, although antifibrinolytic medications reduce the risk of rebleeding, their benefit is offset by an increased risk of cerebral infarction. In several of these studies, patients received antifibrinolytic therapy for weeks, well after the risk for rebleeding had declined and the risk of delayed cerebral ischemia (DCI) increased. Two recent case-control studies have shown that an early and short course of epsilon-aminocaproic acid before exclusion of the aneurysm from the circulation might reduce the risk of rebleeding without significantly increasing complications. In one study, patients treated with epsilon-aminocaproic acid had an eightfold increase in deep venous thrombosis without an increase in pulmonary embolism. This nonrandomized study was not adequately powered to determine the effect of antifibrinolytic therapy on overall patient outcome. A Dutch multicenter, randomized, open-label study, Ultra-Early Tranexamic Acid after Subarachnoid Hemorrhage (ULTRA), began enrolling patients in 2013. It is designed to study the effect on functional outcome (blinded endpoint) of early administration of tranexamic acid in patients with moderate to high-grade aneurysmal SAH for a duration of up to 24 hours. The results of this study might inform whether an early and short course of antifibrinolytic therapy is clinically warranted. The Neurocritical Care Society Consensus Guidelines advise that an early and short course of antifibrinolytic therapy should be considered for patients who are at high risk of rebleeding (such as those with high clinical grade) and in whom definitive aneurysm treatment will be delayed. However, antifibrinolytic agents should be avoided in patients who are at high risk of thromboembolic complications and in whom the risk of rebleeding is reduced. Patients treated with antifibrinolytic therapy should be monitored for systemic and cerebral thrombotic complications.
| Study, Year ∗ | Number of Subjects (Intervention/No Intervention) | Study Design | Intervention | Control | Outcomes |
|---|---|---|---|---|---|
| Girvin, 1973 | 66 (39/27) | Episilon-aminocaproic acid | Standard treatment | No effect on rebleeding, ischemia, or mortality | |
| van Rossum, 1977 | 51 (26/25) | DB, P | Tranexamic acid | Placebo | No effect on rebleeding or mortality |
| Chandra, 1978 | 39 (20/19) | DB, P | Tranexamic acid | Placebo | No effect on rebleeding or mortality |
| Maurice, 1978 | 79 (38/41) | Tranexamic acid | Standard treatment | No effect on rebleeding or mortality | |
| Kaste, 1979 | 64 (32/32) | DB, P | Tranexamic acid | Placebo | No effect on rebleeding or mortality |
| Fodstad, 1981 | 59 (30/29) | Tranexamic acid | Standard treatment | No effect on rebleeding, cerebral ischemia, or mortality | |
| Vermeulen, 1984 | 479 (241/238) | DB, P | Tranexamic acid | Placebo | Decreased rebleeding, increased cerebral ischemia; no effect on outcome or mortality |
| Tsementzis, 1990 | 100 (50/50) | DB, P | Tranexamic acid | Placebo | Increased cerebral ischemia; no effect on rebleeding, outcome, or mortality |
| Roos, 2000 | 452 (229/223) | DB, P | Tranexamic acid | Placebo | Decreased rebleeding; no effect on ischemia, outcome, or mortality |
| Hillman, 2002 | 505 (254/251) | Tranexamic acid | Standard treatment | Decreased rebleeding; no effect on cerebral ischemia, outcome, or mortality | |
| ULTRA | 940 (470/470) | Tranexamic acid | Standard treatment | Primary endpoint: functional outcome; secondary endpoints: case fatality, rebleeding rate, complication rates |
Surgical and Endovascular Measures
There are two primary methods for excluding aneurysms from the circulation: (1) surgical, in which a craniotomy is performed and a clip is placed across the aneurysm neck and (2) endovascular, in which detachable coils are placed into the aneurysm by means of catheter-based techniques. On occasion, an endovascular technique known as flow diversion may be used, in which a stent is placed into the parent vessel across the neck of the aneurysm, allowing blood flow to bypass the aneurysm. The International Subarachnoid Aneurysm Trial (ISAT) is the only large prospective trial comparing the two primary methods. In this trial, 2143 of 9559 patients were deemed good candidates for either therapy and were randomized to surgical or endovascular aneurysm treatment. In the short term, endovascular therapy was associated with less disability (15.6% vs. 21.6%) but lower rates of complete aneurysm obliteration (58% vs. 81%) and higher recurrent SAH rates (2.9% per year vs. 0.9% per year). At 1 year, there was no difference in mortality. Long-term follow-up of these patients found low rates of rebleeding from the treated aneurysm in both groups (10 in the coiling group, 3 in the clipped group), which was insignificant by intention-to-treat analysis. At 5 years, the risk of death was significantly lower in the endovascular arm compared with the surgical arm (11% vs. 14%), but in patients who survived there was no difference in good outcome (modified Rankin Scale [mRS] >2).
Although endovascular therapy has proven to be effective in the short term, aneurysm re-canalization remains a significant limitation. In a retrospective analysis, aneurysm recurrence was found in 33.6% of coiled aneurysms within 1 month and up to 2 years at 0.5 to 24 (mean ± SD) after treatment. Another retrospective review suggested that the use of a high-porosity stent (to retain coils within the aneurysm) was associated with a higher rate of complete aneurysm obliteration but also with increased morbidity and mortality, likely due to the need for dual-antiplatelet therapy. Therefore the use of stents should be avoided if safer alternatives exist.
Whether to clip or to coil an aneurysm is a complex decision that depends on patient factors (age, comorbidities), aneurysm factors (size, shape, location), and availability of local resources and expertise. On the basis of several single-institution retrospective case series and nonrandomized prospective studies, there is evidence that patients with middle cerebral artery (MCA) aneurysms and large (> 50 mL) hematomas might benefit from microsurgical clipping, whereas patients who are older, are seen during the vasospasm period, have poor clinical grade, or have a basilar apex aneurysm might be considered for endovascular therapy. Ideally, experienced neurosurgeons and interventional neuroradiologists collaboratively make the decision.
Timing
In recent years, there has been a trend toward early aneurysm treatment. Multiple retrospective and prospective studies have established an association between a longer interval to treatment and increased risk of pretreatment hemorrhage. The International Cooperative study on the Timing of Aneurysm Surgery explored early versus late surgical intervention based on the neurosurgeons’ intention to treat. Patients whose surgery was planned for within the first 3 days had an overall mortality rate equal to the patients whose surgery was planned for between days 11 and 32. However, patients in the early surgical group had a significantly better clinical recovery than those whose surgery was delayed ( P < .01). The patients with the highest mortality were those whose surgery was planned for days 7 to 10 after ictus, a time when risk of vasospasm and delayed cerebral injury is greatest. On the basis of this study, early surgery/endovascular therapy is recommended.
Hydrocephalus
Acute hydrocephalus (enlargement of the ventricles) occurs in 15% to 30% of SAH patients. The presence of hydrocephalus correlates with worse radiographic and clinical grades and with an unfavorable prognosis. The symptoms associated with hydrocephalus range from no symptoms to signs of intracranial hypertension, such as impairment of upward gaze, sixth nerve palsy, and headache. Hydrocephalus may be “noncommunicating” because of obstruction (by blood) within the ventricular system or “communicating” because of obstruction of cerebrospinal fluid (CSF) reabsorption into the venous system.
If severe, hydrocephalus may impair the level of consciousness and should be treated immediately with CSF diversion. Ventriculostomy is the most common method of treatment; however, in a select group of patients with communicating hydrocephalus, who are not at risk for central or tonsillar herniation, lumbar drainage may be reasonable. Two small, single-institution studies suggested that in appropriately selected patients, lumbar CSF drainage is associated with a reduction in “clinical vasospasm” (i.e., neurological deficits not attributable to other structural or metabolic causes). The EARLYDRAIN (Outcome After Early Lumbar CSF-drainage in Aneurysmal SAH) trial is an ongoing two-arm randomized controlled study comparing the effect of early continuous lumbar CSF drainage and standard neurointensive care with standard neurointensive care alone on functional outcome (disability at 6 months). CSF drainage usually leads to an improvement in symptoms.
Hydrocephalus and the need for CSF diversion are typically temporary. In some patients, hydrocephalus does not resolve, and ongoing CSF diversion with a permanent indwelling shunt is necessary. In a single-center, prospective, randomized controlled trial, extending the duration of weaning external ventricular drainage for more than 24 hours did not affect the need for permanent shunting and was associated with increased length of both intensive care unit (ICU) and hospital stay. There is no role for routine fenestration of the lamina terminalis to decrease the rate of permanent shunting. Data regarding treatment of hydrocephalus in SAH are largely retrospective; optimal management of patients with mild symptoms is unknown.
Seizures
The evidence regarding the incidence, prophylaxis, and treatment of seizures is mostly retrospective. The reported incidence of seizures after SAH varies from 8% to 35%. In one retrospective cohort study, most seizures after SAH occurred before hospitalization, and the incidence of in-hospital seizures was 4.1%. These seizures occurred despite prophylaxis with an antiepileptic drug (AED) and occurred at least 1 week after aneurysmal rupture. Risk factors associated with the development of seizures include ruptured MCA aneurysm, intracerebral hemorrhage, thicker cisternal clot, rebleeding, ischemic infarct, and a history of hypertension. Two studies demonstrated no difference in outcome between patients who had seizures and those who did not. However, a third study found that seizures at the time of hemorrhage were associated with poor outcome.
The incidence of generalized convulsive status epilepticus (GCSE) is 0.2%, but the incidence of nonconvulsive status epilepticus (NSE) is much higher. A prospective study found that 31% of stuporous or comatose SAH patients had NSE when monitored with continuous electroencephalography (cEEG). The mean onset of NSE was 18 days after hemorrhage. GCSE and NSE are associated with worse outcome. Therefore it is reasonable to use periodic or cEEG to assess unconscious patients and those who have a change in neurologic examination for seizures.
The benefit of prophylactic AEDs has not been definitively established. It is reasonable to use AEDs before aneurysm treatment because of the risk of seizure-related rebleeding (due to a surge in blood pressure). However, there is no evidence to support the long-term use of AEDs in patients without a history of seizure. In fact, cumulative phenytoin exposure is associated with a worse cognitive outcome at 3 months.
Delayed Cerebral Ischemia
DCI (also referred to as delayed ischemic neurologic deficits [DINDs]) is defined as neurologic deterioration presumed to be ischemic that lasts for more than an hour and that cannot be attributed to another cause. DCI accounts for most morbidity and mortality from SAH; therefore its detection and treatment are the major foci of intensive care. Although historically attributed exclusively to cerebral vasospasm (narrowing of the large caliber arteries at the base of the brain), DCI likely has protean causes, including a local inflammatory and hypercoagulable state that result in formation of microthrombi and microembolism. Consequently, consensus statements recommend that the term vasospasm be reserved to describe only radiologic findings of vessel narrowing and not clinical deterioration. DCI may present with agitation followed by an indolent decrease in level of consciousness or focal neurologic deficits that vary depending on the affected arterial distribution. Vasospasm and DCI usually begin at day 3 after bleed, peak at days 6 to 8, and resolve over 2 to 4 weeks. Thickness of cisternal clot has been associated with the development of vasospasm. Almost one third of patients who survive the initial SAH have DCI, and approximately half of these patients die.
The diagnosis of and the decision to treat DINDs due to vasospasm are made with an observed clinical deterioration along with the radiographic finding of vasospasm. For SAH patients who have impaired consciousness and in whom subtle clinical deteriorations are not easily seen, bedside monitoring modalities such as cEEG, transcranial Doppler ultrasound (TCD), and invasive physiologic monitors might serve as surrogates for clinical changes.
Detection : Monitoring for ischemia includes clinical, radiologic, and physiologic assessments. Although by definition, DCI is detected by serial neurologic examination, not all ischemic insults are clinically apparent, especially in comatose patients.
Radiologic monitoring includes methods to assess for cerebral vasospasm and methods that assess cerebral blood flow (CBF) (perfusion). The gold standard for vasospasm detection is invasive digital subtraction angiography (DSA). Risks associated with DSA include hematoma, infection, peripheral thromboembolic events, and stroke. The rate of neurologic complications in SAH patients is 1.8%. Noninvasive angiography with computed tomography (CT) or magnetic resonance imaging (MRI) is less sensitive for detecting vasospasm. CT angiography (CTA) has a sensitivity of 86% to 91.6% and is better suited to detect vasospasm of proximal arterial segments. CTA has a high negative predictive value (95% to 99%); therefore it may be used as a screening tool to limit the use of DSA. Magnetic resonance angiography (MRA) has a sensitivity for vasospasm detection of 45.6% compared with conventional angiography.
TCD detects increased cerebral blood flow velocity (CBFV) associated with vasospasm. This noninvasive study may be performed daily at the bedside and is less expensive than many other monitoring tests. TCD is most useful in detecting evidence of vasospasm in the middle cerebral and basilar arteries. Compared with DSA, TCD has a relatively high specificity for, but poor sensitivity (42% to 67%) for, vasospasm detection. Several conditions other than cerebral vasospasm increase CBFV, such as increased blood pressure and hyperemia. The Lindegaard ratio (hemispheric index), the ratio between the blood flow velocities in the MCA and the ipsilateral extracranial internal carotid artery, may be used to distinguish increased CBFV due to vasospasm from other causes. Lindegaard ratios between 3 and 6 correlate with mild and moderate vasospasm, whereas indices greater than 6 suggest severe vasospasm. Importantly, elevated TCD velocities do not correlate with the development of DIND. No study has shown that TCD monitoring affects outcome after SAH.
Although imaging blood flow may be a more direct way to assess for ischemia than imaging of blood vessels (for vasospasm), it has been less well studied. Methods for blood flow imaging include CT perfusion (CTP), xenon CT (Xe-CT), MR perfusion, and single-photon emission CT. In the ICU, CT-based imaging studies (CTP, Xe-CT) are typically more practical because they involve less time than MRI and nuclear imaging studies and may be done at the bedside with a portable CT scanner. Although Xe-CT is a well-established tool that provides quantitative blood flow information, xenon gas is no longer approved by the U.S. Food and Drug Administration for this use. Therefore Xe-CT blood flow imaging is currently not in clinical use. The literature regarding the utility of CTP consists of multiple small studies, and there are considerably fewer studies on the utility of MR perfusion. Neither CT nor MR perfusion imaging is widely used for detection of DCI, and further study of these modalities is required.
Physiologic monitoring for DCI includes invasive techniques, such as regional brain tissue oxygen monitoring, regional CBF monitoring, and regional biochemical monitoring, and noninvasive techniques, such as quantitative cEEG and near-infrared spectroscopy. Although there is much enthusiasm for regional brain tissue oxygen monitoring and cerebral microdialysis, there is little supportive evidence in this setting. Although traditionally used to detect seizures, cEEG is also becoming a tool for detection of ischemia. Ischemia produces characteristic changes on EEG, namely loss of fast-frequency waves followed by an increase in slow-frequency waves and ultimately suppression of brain waves. Various software analytic tools are available that provide some measure of the relative proportion of fast waves to slow waves; therefore, they might allow cEEG to serve as a continuous, noninvasive ischemia detector. The optimal EEG parameters for DCI detection and the effect on outcome of therapy based on cEEG ischemia monitoring are unknown. The literature consists exclusively of small prospective and retrospective single-center observational studies that used varying definitions of DCI and that included patients with varying severities of SAH. These studies suggest that it might be possible to detect ischemia by cEEG 1 to 3 days before changes in clinical examination. As with other physiologic monitors, quantitative cEEG is not widely used and its utility requires further study.
Prevention and Treatment : Table 63-2 summarizes the randomized trials that have been performed on therapies to treat vasospasm and DCI. Therapy consists of medical and endovascular measures.
| Study, Year ∗ | Number of Subjects (Intervention/No Intervention) | Study Design | Intervention | Control | Outcomes |
|---|---|---|---|---|---|
| Hemodynamic Augmentation | |||||
| Lennihan, 2000 | 82 (41/41) | Hypervolemic therapy | Normovolemic therapy | No difference in symptomatic vasospasm | |
| Egge, 2001 | 32 (16/16) | Hypervolemic hypertensive hemodilution therapy | Normovolemic therapy | No difference in DIND or TCD vasospasm | |
| HIMALAIA | 240 (120/120) | Induced hypertension | Standard therapy without induced hypertension | Primary endpoint: functional outcome Secondary endpoints: adverse effects, CBF measured by perfusion CT | |
| Magnesium Therapy | |||||
| van den Bergh, 2005 | 283 (139/144) | DB, P | Magnesium IV | Placebo | Decreased incidence of DINDs; improved clinical outcome at 3 months |
| Veyna, 2002 | 40 (20/20) | Magnesium IV | Standard therapy | Trend toward improved clinical outcome | |
| Wong, 2006 | 60 (?/?) | DB | Magnesium IV | Saline | Trend toward decrease in symptomatic vasospasm; decrease TCD vasospasm timeframe; no difference in clinical outcome |
| Schmid-Eisaeser, 2006 | 104 (53/51) | Magnesium IV | Nimodipine IV | Incidence of vasospasm and clinical outcome comparable | |
| Muroi, 2008 | 58 (31/27) | P | Magnesium IV | Placebo | No difference in DINDs; improved clinical outcome at 3 months |
| IMASH, 2010 | 328 (169/159) | DB, P | Magnesium IV | Saline | No difference in clinical outcome at 6 months |
| MASH-2, 2012 | 1203 (606/597) | DB, P | Magnesium IV | Saline | No difference in clinical outcome at 3 months |
| Calcium Channel Blockers | |||||
| Allen, 1983 | 116 (56/60) | DB, P | Nimodipine PO | Placebo | Decreased incidence of DINDs |
| Philippon, 1986 | 70 (?/?) | DB, P | Nimodipine PO | Placebo | No difference in vasospasm; decreased incidence of DINDs; improved mortality |
| Neil-Dwyer, 1987 | 75 (?/?) | DB, P | Nimodipine PO | Placebo | Improved clinical outcome at 3 months |
| Petruk, 1988 | 154 (72/82) | DB, P | Nimodipine PO | Placebo | Decreased incidence of DINDs; improved clinical outcome at 3 months |
| Pickard, 1989 | 554 (278/276) | DB, P | Nimodipine PO | Placebo | Decreased incidence of DINDs; improved clinical outcome at 3 months |
| Haley, 1993 | 906 (449/457) | DB, P | Nicardipine IV | Placebo | Decreased incidence of vasospasm; no difference in clinical outcome |
| Haley, 1994 | 365 (184/181) | DB | High-dose nicardipine IV | Low-dose nicardipine IV | Incidence of vasospasm and clinical outcome comparable |
| Statin Therapy | |||||
| Lynch, 2005 | 39 (19/20) | DB, P | Simvastatin | Placebo | Decreased incidence of vasospasm |
| Tseng, 2005 | 80 (40/40) | DB, P | Pravastatin | Placebo | Decreased incidence of vasospasm and DINDs; improved mortality |
| Tseng, 2006 | 80 (40/40) | DB, P | Pravastatin | Placebo | Improved clinical outcome at 6 months |
| Chou, 2008 | 39 (19/20) | DB, P | Simvastatin | Placebo | No difference in vasospasm or DINDs; trend toward decreased mortality |
| STASH, 2014 | 812 (391/421) | DB, P | Simvastatin | Placebo | No difference in short- and long-term outcome |
| Wong | 240 (120/120) | DB | Simvastatin 80 mg | Simvastatin 40 mg | Presence of DIND at 1 month |
| Other Approach | |||||
| Bulters, 2013 | 71 (35/36) | Intra-aortic balloon pump | Hypervolemic therapy | No difference in clinical outcome, mean cardiac output, or CBF | |
Related posts:
What Lessons Have Intensivists Learned During the Evidence-Based Medicine Era?
What Is the Role of Invasive Hemodynamic Monitoring in Critical Care?
Should Fever Be Treated?
Are Computerized Algorithms Useful in Managing the Critically Ill Patient?
Inhaled Vasodilators in ARDS: Do They Make a Difference?
How Can We Monitor the Microcirculation in Sepsis? Does It Improve Outcome?
Full access? Get Clinical Tree
