Controversies in the management of aneurysmal subarachnoid hemorrhage*

Neeraj S. Naval, MD; Robert D. Stevens, MD; Marek A. Mirski, MD, PhD; Anish Bhardwaj, MD, FCCM

Background: The care of patients with aneurysmal subarach- Data Source: Search of MEDLINE and Cochrane databases and noid hemorrhage has evolved significantly with the advent of new manual review of article bibliographies. diagnostic and therapeutic modalities. Although it is believed that Data Synthesis and Conclusions: Many aspects of care in these advances have contributed to improved outcomes, consid- patients with aneurysmal subarachnoid hemorrhage remain erable uncertainty persists regarding key areas of management. highly controversial and warrant further resolution with hypoth- Objective: To review selected controversies in the manage- esis-driven clinical or translational research. It is anticipated that ment of aneurysmal subarachnoid hemorrhage, with a special the rigorous evaluation and implementation of such data will emphasis on endovascular vs. surgical techniques for securing provide a basis for improvements in short- and long-term out- aneurysms, the diagnosis and therapy of cerebral vasospasm, comes. (Crit Care Med 2006; 34:511–524) neuroprotection, antithrombotic and anticonvulsant agents, cere- KEY WORDS: aneurysm; subarachnoid; hemorrhage; vasospasm; bral salt wasting, and myocardial dysfunction, and to suggest ischemia venues for further clinical investigation.

he rupture of an intracranial sepsis, and thromboembolism. As many and management of cerebral vasospasm, aneurysm may be associated of these complications are life-threaten- neuroprotective strategies, use of anti- with an array of severe distur- ing but reversible, it is widely believed thrombotic agents (thrombolytic agents, bances in intracranial and sys- that patients with aSAH can benefit from heparin, and platelet inhibitors), prophy- Ttemic physiology that represent a unique laxis of seizures, and the approach to ce- management in an intensive care setting. challenge to the clinician. Surgical man- Recent years have seen a considerable rebral salt wasting and to cardiac dys- agement has traditionally emphasized the expansion in the use of image-guided en- function after aSAH. For each prevention of renewed intracranial bleed- dovascular therapies for aSAH, including controversy, a critical evaluation of the ing by clipping or wrapping the respon- coiling of aneurysms and balloon angio- available evidence is coupled with recom- sible aneurysm. Medical management is plasty or intraarterial drug delivery for mendations for further clinical investiga- based on the detection and treatment of cerebral vasospasm (1–4). These ad- tion. The review is selective, focusing on cerebral and extracerebral complications vances have occurred in a general setting the principal debates at the expense of of aneurysmal subarachnoid hemorrhage of increasing knowledge of aSAH epide- other equally important but arguably less (aSAH). Cerebral complications of aSAH miology, pathophysiology, diagnosis, and controversial issues (e.g., post-aSAH hy- include recurrent intracranial hemor- prevention and of significant refinements drocephalus and ventricular drainage). rhage, vasospasm, cerebral infarction, hy- in microsurgical technique and in medi- drocephalus, cerebral edema, and intra- cal therapy. There is a widespread percep- cranial hypertension; extracerebral Controversy 1: Surgical vs. tion that this broader understanding and Endovascular Aneurysm Repair complications include respiratory failure, expertise is yielding benefits in the form derangements of water and electrolyte of improved outcome after aSAH. Indeed, Endovascular coiling emerged as an homeostasis, myocardial dysfunction, a progressive increase in aSAH survival alternative to surgery in patients with over the past three decades has been re- intracranial aneurysms who were deemed ported in several studies (5–7). However, poor surgical candidates due to signifi- *See also p. 571. evidence of a direct relationship between cant neurologic injury, the presence of From the Division of Neurosciences Critical Care, aSAH outcomes and a specific strategy or severe medical co-morbidities, or difficult Departments of Neurology, Anesthesiology/Critical Care Medicine, and Neurological Surgery, Johns Hop- intervention is limited (2, 3, 8, 9). In surgical access to the aneurysm (1). More kins University School of Medicine, Baltimore, MD. several key areas of management, sup- recent work has sought to extend the Supported, in part, by U.S. Public Health Service porting data are lacking or equivocal in indications of endovascular coiling to National Institutes of Health NS NS046379 and by an nature, generating uncertainty and con- other patient categories. In a small ran- Established Investigator Grant (A. Bhardwaj) from the troversy among clinicians. American Heart Association. domized trial of 109 patients with aSAH, The authors have no financial interests to disclose. This review focuses on controversies 3- and 12-month clinical and neuropsy- Copyright © 2006 by the Society of Critical Care that are central to the acute management chological outcomes were the same be- Medicine and Lippincott Williams & Wilkins of aSAH. These include surgical vs. endo- tween the surgical group and the endo- DOI: 10.1097/01.CCM.0000198331.45998.85 vascular aneurysm repair, the diagnosis vascular group (10). This was followed by

Crit Care Med 2006 Vol. 34, No. 2 511 the International Subarachnoid Aneu- after the first year for the endovascular spasm was either not significantly differ- rysm Trial (ISAT), a multiple-center, ran- group and 3,107 patient years of fol- ent (19–21) or slightly higher (22, 23) domized study of endovascular coiling vs. low-up for the neurosurgical group, with among surgically treated patients. An surgical clipping conducted in 2,143 pa- a mean follow-up of 4 yrs, risk of recur- early nonrandomized study of 156 pa- tients with aSAH who were deemed suit- rent aSAH was higher in patients ran- tients had suggested a higher rate of ce- able for either therapy. Posterior circula- domized to coiling (seven patients) com- rebral infarction in patients receiving en- tion aneurysms accounted for only 58 of pared with clipping (two patients), but dovascular vs. surgical therapy; however, 2,143 patients (as many of these patients mortality related to recurrent aSAH was the proportion of patients with poor ini- were not enrolled because coiling was equal in both groups. A higher risk for tial neurologic presentation was higher considered the preferred modality of seizures and poor cognitive outcomes in the endovascular group (24). treatment). At 1 yr, endovascular coiling was seen in the surgical group, and cu- Comment. The available data indicate was associated with dependency or death mulative 7-yr mortality curves showed that in patients with good neurologic in 23.5% of patients compared with more deaths in the surgical group com- grade aSAH who undergo treatment for 30.9% in the surgical group, a relative pared with coiling. The increased risk of aneurysms in the anterior circulation, risk reduction of 22.6% (p Ͻ .001). Of re-bleeding in the coiling group did not 1-yr outcomes are clearly superior after concern, however, nonprocedural re- seem to reverse the early benefit seen endovascular coiling when compared bleeding within 1 yr was higher in pa- with this modality. Notwithstanding its with surgical clipping. Follow-up of pa- tients randomized to endovascular treat- limitations, ISAT represents level I evi- tients receiving endovascular treatment ment (40 recurrent aSAHs, with 22 dence that in this patient population, en- in some studies suggests that despite deaths) compared with patients allocated dovascular coiling is associated with bet- good clinical results, a significant propor- to neurosurgical treatment (33 patients ter 1-yr outcomes, with trends toward tion of patients may require repeat endo- with aSAHs, 30-day mortality in 21 pa- superior long-term outcomes, when vascular or surgical therapy for residual tients) (2, 3). compared with surgical clipping. or recurrent aneurysmal lesions. How- Although widely regarded as a land- Apart from the recently published in- ever, long-term (Ͼ1 yr) clinical trends in mark trial, the ISAT has been criticized formation stemming from the ISAT trial, patients enrolled in ISAT do seem to sug- with regard to biases in patient selection, randomized trials comparing long-term gest that endovascular coiling is likely to low rates of randomization of eligible pa- outcomes after coiling vs. clipping of rup- retain its advantage over clipping as a tients, the definition of clinical equipoise, tured aneurysms are not available. How- superior procedure, and further fol- expertise of the neurosurgeons and inter- ever, in an analysis of patients who un- low-up of these patients is likely to add ventionists, the failure to use an opera- derwent coiling of unruptured aneurysms insight to this debate. Preliminary results tive microscope, the higher than ex- and were followed for a median of 22.3 from various studies have failed to con- pected morbidity in the surgically treated months, annual re-bleeding rates were sistently show an association between group, the absence of angiographic data 0.8% in the first year, 0.6% in the second treatment modality and the rate of de- after the initial treatment, the lack of year, and 2.4% in the third year after layed cerebral ischemia. For posterior long-term (Ͼ1 yr) follow-up, and the ap- embolization, with no re-bleeding in sub- circulation aneurysms and for large, an- propriateness of the reported outcome as- sequent years (16). In a study of 29 pa- atomically complex or wide-necked aneu- sessment scale (11, 12). An expert panel tients with giant aneurysms treated with rysms, further studies are needed to de- suggested that 1-yr outcome as reported endovascular coiling and followed up for fine the best therapeutic approach. in the ISAT is not an appropriate end a median of 50 months, long-term clini- point for the comparison of therapies, cal outcomes were good in 79% of pa- Controversy 2: Diagnosis of given that endovascular coiling is be- tients; however, the stability of the coil Delayed Cerebral Ischemia lieved to carry a risk of aneurysmal re- over time was poor, requiring repeat coil- bleeding that extends beyond 1 yr. In ing, surgery, or parent-vessel occlusion A critical concern in the management defense of ISAT, its authors observed that in more than half of the aneurysms ini- of patients with aSAH is the prediction outcomes of surgically treated patients in tially treated with coils (17). Finally, and accurate diagnosis of delayed cere- ISAT were comparable with those in a Friedman et al. (18) reported clinical bral ischemia because timely institution prospective, multiple-center North Amer- (mean, 19.1 months) and angiographic of therapeutic interventions may prevent ican trial (13). They acknowledge that the (mean, 11.6 months) data in 83 patients the occurrence of tissue infarction. rate of randomization of eligible patients with aSAHs treated with endovascular Spasm of large arteries in the circle of was low (22.4%); however, they point out coils. Neurologic outcome was good in Willis is a significant determinant of ce- that this rate was comparable with other 77% of patients; however, 26% had a rebral ischemia after aSAH and carries a large randomized trials of vascular ther- “dog-ear” remnant, 35% had a residual 15–20% risk of stroke or death (25). apy such as the North American Symp- neck, and 3% had residual aneurysm fill- Four-vessel cerebral digital subtraction tomatic Carotid Endarterectomy Trial ing. Two or more coiling procedures were angiography is the gold standard for di- (14) or the Asymptomatic Carotid Athero- required in 34% of patients. agnosing vasospasm, but given the inher- sclerosis Study (15), studies that have An important question concerns the ent risks and allocation of time and re- significantly influenced the treatment of influence of surgical or endovascular sources required for this procedure, patients with carotid artery stenosis. treatment on the prevalence of cerebral alternative and complementary diagnos- A recent report from the ISAT trial ischemic complications. In several non- tic tools have been proposed (26). These group provides further crucial informa- randomized comparisons of surgery and include transcranial Doppler (TCD), com- tion regarding long-term outcomes (3). endovascular therapy of ruptured aneu- puted tomographic angiography (CTA), Based on 3,258 patient years of follow-up rysms, the rate of symptomatic vaso- magnetic resonance imaging, radionu-

512 Crit Care Med 2006 Vol. 34, No. 2 clide imaging, cerebral microdialysis, and comparing transcranial color-coded tion, whereas perfusion-weighted imag- electroencephalography. sonography and conventional TCD in the ing reveals asymmetries in regional per- TCD is an ultrasound-based monitor prediction of angiographic spasm, sensi- fusion. In a study of 14 patients with that uses the principle that the velocity of tivity and specificity of transcranial color- aSAH, diffusion-weighted imaging abnor- blood flow in an artery is proportional to coded sonography for MCA and internal malities were noted in all patients with the ratio of flow to the luminal surface carotid artery vasospasm were higher vasospasm suggested by TCD, whereas no area of that vessel (27). TCD is a nonin- than those for TCD (34). The validity of such abnormalities were observed in pa- vasive, low-risk procedure that can be TCD in diagnosing angiographic vaso- tients without abnormal TCD findings readily performed at the bedside and spasm was summarized in a recent sys- (40). A study of aSAH patients using per- lends itself to repeated (e.g., daily) obser- tematic review (35). For the MCA, sensi- fusion-weighted imaging revealed areas vations, enabling trend analysis. How- tivity of TCD was 67% and specificity was of hypoperfusion that correlated well ever, there is debate about the correlation 99%, with a positive predictive value of with delayed ischemic neurologic deficits between increased TCD flow velocities 97% and negative predictive value of (DIND) and were larger than the areas of and a) angiographic vasospasm and b) 78%. The accuracy of TCD was consider- diffusion-weighted imaging abnormality clinically significant or “symptomatic” ably less for detecting spasm in vessels performed at the same time. Of note, vasospasm. Although mean middle cere- other than the MCA. whereas all 15 patients with DIND bral artery (MCA) cerebral blood flow Interest has developed in CTA as a showed alterations in perfusion-weighted (CBF) velocities of Ͼ200 cm/sec accu- diagnostic modality for detecting cerebral imaging, TCD evidence of vasospasm was rately predict angiographic vasospasm, vasospasm. CTA may be combined with noted in only seven of these patients (41). velocities in the 120–200 cm/sec range perfusion computed tomography (CT), al- Radionuclide-based studies of CBF have a far lower predictive value (28). lowing characterization of both vascular and metabolism include single-photon Moreover, it has been observed that TCD anatomy and associated cerebral perfu- emission computed tomography (SPECT) is not as reliable in estimating distal MCA sion abnormalities. Using CTA, a positive and positron emission tomography. In a vasospasm compared with the more prox- predictive value of 100% for angiographic series of 129 patients with aSAH, Rajen- imal portions of the MCA (29). Studies vasospasm was reported in a recent small dran et al. (42) found that 89 had SPECT showing a significant moment-to-mo- series (36). Another group reported that evidence of hypoperfusion, and this cor- ment variability during continuous mea- CTA was highly accurate in detecting se- related with TCD evidence of vasospasm surement of CBF velocities have height- vere cerebral vasospasm in proximal ar- in only 64% of cases. In another study, ened concern about the accuracy of this terial locations but was less accurate for Jabre et al. (43) observed that the sensi- technique (30). detecting mild and moderate spasm in tivity of SPECT for symptomatic vaso- To address these limitations, several distal locations (37). A third group noted spasm was inferior to TCD; however, refinements have been proposed. Investi- an excellent concordance between the se- SPECT was more specific. Studies using gators have suggested that incremental verity of vasospasm as determined by CTA positron emission tomography in pa- increases in CBF velocities over time are and conventional angiography in both tients with aSAH suggest this technique more helpful than velocities taken in iso- proximal and distal arterial segments can help differentiate neurologic deficits lation. One study indicated that an in- (38). Of note, none of these studies in- due to reversible ischemia or to irrevers- crease of Ͼ50 cm/sec in 24 hrs was cluded Ͼ20 patients. ible infarction (44); however, the accu- closely associated with angiographic and Although the phenomenon of cerebral racy of positron emission tomography clinical vasospasm (31). To overcome the vasospasm has historically elicited much (45) in the diagnosis of vasospasm is un- numerical dependence of flow velocities scientific interest, recent investigations known. in the measurement of CBF, Lindegaard have turned to the related question of Monitoring of neurochemical markers et al. (32) developed a “hemispheric in- brain ischemia and infarction in patients of ischemia with cerebral microdialysis dex” that normalizes flow velocity in the with aSAH. The predictors of CT-defined has been proposed as a technique for de- MCA to that in the ipsilateral extracranial cerebral infarction after aSAH were ana- tecting vasospasm and delayed cerebral internal carotid artery (an index of Ͼ3is lyzed in a recent study. Multivariate anal- ischemia. In a study of 97 patients with strongly predictive of angiographic vaso- ysis revealed that both TCD and angio- aSAH, neurochemical changes indicative spasm). Torbey et al. (33), observing that graphic vasospasm were independently of ischemia were observed before the on- angiographic vasospasm occurs at lower associated with cerebral infarction, how- set of symptoms in 83% of patients with TCD velocities in older patients, sug- ever there was agreement between the DIND (46). In another report, an isch- gested that the accuracy of TCD can be two tests only in 73% of cases, and the emic pattern of cerebral metabolites pre- enhanced by using an age-adjusted no- validity of the two tests alone or in com- ceded the occurrence of DIND by a mean mogram based on a quadratic relation- bination was disappointing (combined interval of 11 hrs (47). In a comparison ship between age and CBF velocities. sensitivity, 0.72; specificity, 0.68; positive with TCD and angiography, microdialysis Some investigators have proposed that predictive value, 0.67; negative predictive was found to have the highest specificity the diagnostic capability of TCD may be value, 0.72) (39). These results are con- and likelihood ratio, but lower sensitivity, enhanced by transcranial color-coded sistent with data suggesting that the as a diagnostic tool for DIND (48). In a sonography. Transcranial color-coded pathophysiologic model of cerebral isch- study of 13 patients combining cerebral sonography is an ultrasound-based neu- emia after aSAH may need to consider microdialysis and positron emission to- roimaging technique that allows real- factors other than large artery spasm. mography, it was noted that transient time visualization of intracranial vascular Magnetic resonance diffusion- reductions in regional CBF correlated structures in addition to measurement of weighted imaging accurately identifies with elevations in extracellular glutamate CBF velocities. In a prospective study brain tissue that is at high risk of infarc- and glycerol, whereas the lactate/pyru-

Crit Care Med 2006 Vol. 34, No. 2 513 vate ratio was sensitive only after longer might expand the therapeutic window for transfusion and albumin was associated periods of hypoperfusion (49). Although treating ischemia in patients with aSAH; with a significantly lower rate of symp- these results are encouraging, several in- however, their introduction into clinical tomatic vasospasm and death when com- herent limitations of cerebral microdialy- practice needs further confirmation in pared with a strategy of blood pressure sis have been recognized, including the clinical trials. control with diuretics. Lennihan et al. difficulty of extrapolating from measure- (62) randomly assigned 82 aSAH patients ments made in a very restricted volume Controversy 3: Management of on the day after aneurysm clipping to of tissue, the development of reactive gli- Delayed Cerebral Ischemia receive albumin fluid boluses titrated to osis around the catheter tip decreasing normal or high cardiac filling pressures the accuracy of measurements, the inter- The principal options for treating de- (central venous and pulmonary artery di- subject variability in basal neurochemical layed cerebral ischemia are hemody- astolic pressures). They discovered that values, and the tissue trauma after probe namic augmentation and endovascular higher filling pressures were not associ- implantation (50). The use of this tech- therapy. Alternative therapies include in- ated with any significant change in CBF nique as a routine diagnostic method in traaortic balloon counterpulsation (as measured by xenon CT) or blood vol- patients with aSAH was not supported in (IABC), therapeutic hypothermia, and ume, nor were there any differences in a recent systematic review (51). barbiturate coma. the rate of symptomatic vasospasm, cere- Several groups have assessed the role The concept of hemodynamic aug- bral infarction, or 3-month Glasgow Out- of electroencephalography in the diagno- mentation—also referred to as hyperten- come Scale. Finally, in a study of 32 aSAH sis of cerebral ischemia after aSAH. In a sion, hypervolemia, hemodilution or tri- patients randomized to hypertensive/ study of 151 patients, Rivierez et al. (52) ple-H therapy—in patients with aSAH hypervolemic vs. normotensive/normo- noted that that focal areas of slowing evolved out of two important and (and volemic management protocols, Egge et correlated with angiographic vasospasm related) observations. The first, suggested al. (63) reported no difference in vaso- in 96% of cases. Vespa et al. (53), using in studies from the 1950s and 1960s, was spasm rates, CBF as measured by SPECT, continuous electroencephalography, the strong association between cardiovas- or 1-yr Glasgow Outcome Scale, whereas found that relative alpha variability was cular variables, such as intravascular vol- a higher rate of complications (hemor- decreased in 19 patients with angio- ume status, cardiac output, and blood rhage, coagulopathy, congestive heart graphically proven vasospasm; in ten of pressure in the days and weeks ensuing failure) was noted in the hypertensive/ these patients, the relative alpha variabil- aSAH, and clinical outcomes, such as hypervolemic group. When the results of ity change preceded clinical symptoms by symptomatic vasospasm and long-term these three trials were pooled in a sys- nearly 3 days. Finally, Claassen et al. (54), neurologic function. The second observa- tematic review, no significant effect of in comparative-analysis quantitative elec- tion was that cerebral vasospasm is char- prophylactic triple-H therapy on the rate troencephalographic variables in 34 pa- acterized by a shift in cerebrovascular of symptomatic vasospasm, DIND, or tients with Hunt and Hess grade IV or V resistance away from the penetrating ar- death was noted (64). A Cochrane meta- aSAH, determined that a decrease in al- terioles to the major branches of the cir- analysis reached a similar conclusion (9). pha to delta ratio (ADR) was strongly cle of Willis and their proximal branches, Triple-H therapy has many inherent associated with delayed cerebral ischemia vessels that are incapable of effective au- limitations, notably a) the association of defined by clinical or CT criteria. In this toregulation. As a result, CBF becomes hemodynamic augmentation with severe report, a 50% decrement in alpha to delta passively dependent on systemic blood complications such as congestive heart ratio has a sensitivity of 89% and a spec- pressure, greatly increasing the risk of failure, noncardiogenic pulmonary ificity of 84% for delayed cerebral isch- cerebral ischemia. Using the Poiseuille edema, myocardial ischemia, intracranial emia. relationship, it can be predicted that ef- hemorrhage, global cerebral edema, and Comment. The prediction and accu- forts to increase systemic blood pressure death; b) failure to reverse neurologic rate diagnosis of cerebral ischemia is a or to decrease blood viscosity will ame- deterioration in certain patients; and c) cardinal goal in the critical care of pa- liorate cerebral perfusion and reverse contraindications to its use such as the tients with aSAH. Using serial TCD and ischemia (55). presence of significant preexisting or ac- transcranial color-coded sonography, in- In an early case series of patients with quired cardiopulmonary dysfunction. Al- cremental changes in flow velocities and DIND, Kosnik and Hunt noted significant ternative approaches to treating vaso- calculation of the hemispheric index may improvements in neurologic function af- spasm have emerged, including a) a provide valuable information with regard ter initiating therapy with phenylephrine strategy targeting cardiac output rather to underlying angiographic vasospasm, in and colloid fluid expansion (56). This was than mean arterial pressure as a physio- particular, vasospasm involving the MCA. followed by several other uncontrolled logic end point, b) the endovascular ther- The diagnostic capability of CTA is un- studies further substantiating the clinical apies, including balloon angioplasty and clear and needs investigation in a pro- efficacy of hemodynamic augmentation intraarterial vasodilator administration, spective study. Recent studies have (57–60). Randomized studies evaluating and c) the use of IABC. shifted away from the characterization of the efficacy of hemodynamic augmenta- To augment cardiac output, the ino- vessel lumen diameter to the detection of tion therapy in aSAH are limited in num- tropic agent dobutamine was adminis- changes in CBF and of cerebral ischemia. ber. In 30 aSAH patients who were being tered in combination with hypervolemic Perfusion CT, diffusion- and perfusion- studied before they underwent aneurysm preload enhancement to 23 patients with weighted magnetic resonance imaging, clipping, Rosenwasser et al. (61) noted vasospasm whose neurologic examina- radionuclide-based perfusion studies, ce- that a strategy of blood pressure control tion failed to improve after preload en- rebral microdialysis, and electroencepha- with vasodilators in addition to volume hancement alone. The authors noted a lography hold promise as techniques that expansion with packed red blood cell 52% increase in cardiac index, and clini-

514 Crit Care Med 2006 Vol. 34, No. 2 cal reversal of ischemic symptoms was Although the efficacy of TBA in treat- ple-H therapy has a favorable effect on evident in 18 of the 23 patients (65). In a ing vasospasm is well documented, inter- neurologic outcomes or survival and is more recent study, xenon CT was used to est has developed on the preventive use of safe when compared with a strategy of evaluate CBF in 16 patients with symp- this technique in high-risk patients. normovolemia and normotension. Fur- tomatic vasospasm who underwent vol- Based on results from an animal model ther study is needed to better understand ume expansion combined with either showing that TBA performed on the day the effects of increased cardiac output, as mean arterial pressure augmentation of the hemorrhage prevented the devel- opposed to hypertensive therapy, on CBF with phenylephrine or cardiac output opment of angiographic spasm on day 7 and on the reversal of symptomatic vaso- augmentation with dobutamine. The in- after aSAH, Muizelaar et al. (74) evalu- spasm. At this time, TBA and intraarterial crease in mean CBF was similar in both ated prophylactic TBA in 13 patients with vasodilators administration are reason- groups (66). The mechanism whereby in- Fisher grade III aSAH who had a high able options in treating vasospasm refrac- creases in cardiac output without probability of developing vasospasm. Of tory to medical management. However, changes in mean arterial pressure affect these patients, none developed DIND or the relative efficacy and harm of TBA vs. CBF is unclear, and this phenomenon more than mild TCD-defined vasospasm. medical management needs further sub- may be unique to the setting of vaso- At 3 months posttreatment, eight pa- stantiation. This might take the form of a spasm, as a relationship between cardiac tients had a good recovery, two were randomized trial comparing immediate output and CBF was not observed in pa- moderately disabled, and three had died angioplasty vs. triple-H therapy in pa- tients with traumatic brain injury (67). It (one because of a vessel rupture during tients who have developed symptomatic has been postulated that the wider pulse TBA). These authors are conducting a vasospasm. In patients with symptomatic pressure and enhanced pulsatile flow as- randomized trial (Balloon Prophylaxis of vasospasm in whom triple-H therapy and sociated with the administration of ino- Aneurysmal Vasospasm trial; see http:// endovascular options have either failed or tropic agents may ameliorate flow www.strokecenter.org/trials for details) are contraindicated, consideration should through collateral vessels and through to determine whether the efficacy of pro- be given to IABC or to neuroprotective the microvasculature (68). phylactic TBA is sufficient to justify the interventions such therapeutic hypother- Recent years have seen a significant risks and to clarify which vessels need to mia and pentobarbital coma (see below). development in the use of endovascular be dilated prophylactically In recent years, several groups have therapies for patients with, or at risk of, Controversy 4: Cerebral reported on the use of IABC in patients cerebral vasospasm (69). Endovascular Protection with delayed cerebral ischemia (75–79). treatments include transluminal balloon IABC has been associated with reversal of The development of cerebral ischemic angioplasty (TBA) and the intraarterial neurologic deficits and significant in- complications in a significant proportion delivery of vasodilating compounds. creases in CBF in patients who did not of patients after aSAH has prompted great These techniques are most commonly respond to conventional triple-H therapy interest in the possibility of preventing or used in patients with symptomatic vaso- (79) or who had cardiopulmonary dys- limiting irreversible brain injury. The spasm that has been resistant to triple-H function contraindicating it (75, 76). Of risk of a cerebral ischemic event is par- therapy. TBA is highly effective in reliev- note, despite the invasive nature of this ticularly high during specific events as- ing focal spasm involving proximal seg- technique, the prevalence of severe ad- sociated with aSAH, namely a) the initial ments of the circle of Willis; however, verse effects was remarkably low, and a aneurysmal rupture (risk of global cere- when vasospasm is diffuse or more distal, recent series evaluated the prophylactic bral ischemia secondary to increased in- the selective intraarterial infusion of va- placement use of IABC in six patients who tracranial pressure, hypotension, and hy- sodilators may be helpful. Clinical im- were considered at high risk for cerebral poxemia), b) the procedure undertaken provement after angioplasty is well doc- vasospasm (77). As with pharmacologic for securing the aneurysm (risk of stroke umented and is generally durable, cardiac output augmentation, the benefi- in relation to surgical clipping or endo- whereas the response to intraarterial va- cial effects of IABC are not well under- vascular coiling), and c) delayed cerebral sodilators is transitory. Complications of stood and might reflect augmentation of ischemia and vasospasm. Cerebral pro- TBA include trauma to the arterial wall diastolic perfusion through the carotid tective strategies that have been studied leading to dissection, rupture, and and vertebral arteries and heightened in these settings include a) pharmaco- thrombosis, with consequent cerebral in- pulsatility, which has been associated logic agents with cytoprotective or vaso- farction or hemorrhage; in addition, with improved microvascular blood flow dilatory properties, b) therapeutic hypo- reperfusion of cerebral tissue in which (78). thermia, and c) hemodynamic infarction has already occurred may in- Comment. Hypovolemia and hypoten- augmentation or endovascular therapies, duce edema and hemorrhage. The most sion after aSAH are strongly linked to which are discussed above. Pharmaco- commonly administered cerebral intraar- adverse outcome and should be avoided logic cerebral protectants that have been terial vasodilator is papaverine. However, in all patients. The existing data do not tested in clinical studies of aSAH include papaverine is neurotoxic and has been support the prophylactic use of triple-H calcium channel antagonists, tirilazad linked to seizures, coma, blindness, and therapy in patients with aSAH who do not mesylate, glucocorticoids, magnesium, irreversible cortical injury (70). Vera- have clinical evidence of vasospasm. In endothelin receptor antagonists, and hy- pamil (71), nimodipine (72), and nicardi- patients with symptomatic vasospasm, droxymethylglutaryl coenzyme A reduc- pine (73) have been used as alternative hemodynamic augmentation may reverse tase inhibitors. intraarterial cerebral vasodilators; how- neurologic deterioration; however, an ad- A meta-analysis of trials using calcium ever, data on their efficacy and safety is equately powered randomized trial is channel antagonists after aSAH was re- largely anecdotal. needed to test the hypothesis that tri- ported by the Cochrane group (8). Of 11

Crit Care Med 2006 Vol. 34, No. 2 515 studies (2,804 patients), eight involved rate of symptomatic vasospasm that was with vasospasm but were a more general nimodipine (1,574 patients), two involved significantly lower in women who re- marker of neuronal damage (91). nicardipine (954 patients), and one in- ceived tirilazad, 15 mg/kg, without any Initially developed as cholesterol- volved AT877 (276 patients). When com- effect on mortality (81). In the other trial, lowering agents, the hydroxymethylglu- pared with placebo, calcium channel an- the rate of symptomatic vasospasm was taryl coenzyme A reductase inhibitors or tagonists were associated with a not affected by tirilazad therapy; however, statins are capable of modulating endo- significantly reduced risk of poor out- mortality was significantly lower in thelial function by reducing vascular in- come, with a relative risk (RR) of 0.82 treated patients with Hunt and Hess flammation, inhibiting vascular smooth (95% confidence interval [CI], 0.72– grades IV and V (82). When results of the muscle cell proliferation, decreasing 0.93), an absolute risk reduction of 5.1%, four trials were combined in a meta- platelet aggregation, and promoting NO- and numbers needed to treat of 20. The analysis, it was concluded that overall mediated cerebral vasodilation (92). In a results were most robust for oral nimo- mortality was decreased in patients who murine model of aSAH, pretreatment dipine (RR of poor outcome, 0.70; 95% received tirilazad, with an effect that with simvastatin was associated with a CI, 0.58–0.84). The RR of death in pa- seemed most pronounced in patients reduction in vasospasm and with in- tients treated with calcium antagonists with poor neurologic grade (83). creased expression of NO synthetase (93). was 0.94 (95% CI, 0.80–1.10), that of Animal studies have shown that glu- A retrospective review of 60 aSAH pa- ischemic neurologic deficits was 0.67 cocorticoids may effectively prevent de- tients suggested that patients who were (95% CI, 0.59–0.76), and that of CT or layed cerebral ischemia after aSAH (84– taking statins, when compared with a MR documented cerebral infarction was 86). The potential benefit of control group who were not, had a sig- 0.80 (95% CI, 0.71–0.89). Of note, the glucocorticoids in humans was suggested nificantly lower rate of DIND and cerebral prevalence of angiographic vasospasm in a nonrandomized comparative study of infarctions of any type, but no impact on was not influenced by treatment alloca- 21 patients by Chyatte et al. in which mortality or global outcome (modified tion, suggesting that the benefit of nimo- patients receiving high-dose methylpred- Rankin scale) was detected (94). Another dipine was linked to its cytoprotective nisolone had a significantly lower rate of retrospective study showed an increased rather than its cerebral vasodilatory prop- DIND (87). However, these results have risk of vasospasm associated with use of erties. The effects of nimodipine on aSAH not been reproduced in any other clinical statins (95). The effects of statins have outcome are fairly consistent across pub- trial. been further explored in two recent ran- lished trials. However, it is plausible that The neuroprotective effects of magne- domized, placebo-controlled trials. In the the results were confounded because ni- sium have been reported in experimental first, therapy with simvastatin (19 pa- modipine-treated patients may have re- models of traumatic brain injury, cere- tients) initiated within 48 hrs of aSAH ceived greater amounts of intravenous bral ischemia, and aSAH. Magnesium sul- was linked to reduced serum levels of fluids to counteract the effect of the drug fate therapy is both safe and effective in brain injury biomarkers and a decrease in on systemic blood pressure, surrepti- preventing neurologic complications in the prevalence of DIND confirmed by tiously exposing them to a more aggres- obstetric patients with preeclampsia. A TCD or angiography when compared with sive regimen of triple-H therapy. randomized trial of 40 patients demon- placebo (20 patients) (96). In the second, Tirilazad mesylate, a non–glucocorti- strated that high-dose intravenous mag- treatment with pravastatin started within coid 21-aminosteroid that inhibits lipid nesium sulfate is safe in patients with 72 hrs of aSAH ameliorated cerebral va- peroxidation, has been evaluated in four aSAH (88). In a more recent randomized, sospasm, improved cerebral autoregula- randomized, placebo-controlled trials of placebo-controlled trial of 283 patients, tion, and reduced vasospasm-related patients with aSAH. The first trial, con- an infusion of magnesium started within DIND by 83% and mortality by 75% (97). ducted in Europe and Australasia, dem- 4 days of aSAH and continued up to 14 Additional agents that have generated in- onstrated an improvement in symptom- days postaneurysm occlusion was associ- terest as cerebral protectants after aSAH atic (but not angiographic) vasospasm, ated with a 34% RR reduction for the include erythropoietin (98), nitric oxide 3-month survival, and Glasgow Outcome primary outcome of delayed cerebral donors (99), and potassium channel acti- Scale in patients receiving 6 mg/kg tiril- ischemia, however, this result did not vators (100); to our knowledge, there are azad for 10 days (80). These results were achieve statistical significance (89). Of no published clinical trials of these com- not reproduced in a second study under- note, significantly more patients treated pounds in patients with aSAH. taken in North America in which tirilazad with magnesium had excellent outcome The neuroprotective effect of hypo- administration had no significant effect (defined as a Rankin score of 0) at 3 thermia is supported by a large body of on symptomatic vasospasm, 3-month months (RR, 0.91; 95% CI, 0.84–0.98). experimental and clinical evidence. This mortality, or Glasgow Outcome Scale Following experimental evidence im- effect may result from decreased excita- (13). The divergence in these results was plicating the vasoconstrictor endothelin tory amino acid release and free-radical believed to reflect differences in manage- peptides and their receptors in the patho- production, reduced intracellular cal- ment protocols, including the widespread physiology of cerebral vasospasm, a ran- cium accumulation, stabilization of the use of anticonvulsants (e.g., phenytoin) domized, placebo-controlled trial of an blood–brain barrier, and decreased cere- in the American trial, which may have endothelin receptor antagonist, TAK-044, bral edema (101). Hypothermia also has diminished the bioavailability of tirilazad. was conducted in 412 patients with aSAH serious deleterious consequences, includ- Further analysis of these two studies in- (90). A nonsignificant decrease in the rate ing cardiovascular depression, immune dicated that any detected improvement in of delayed cerebral ischemia was noted in suppression, coagulopathy, and electro- outcome seemed limited to men. This led the treated group. A more recent obser- lyte abnormality. The use of moderate to two other studies conducted in women vational study suggested that increased hypothermia (28–32°C) to protect the only. The first of these demonstrated a CSF levels of endothelin did not correlate brain during cerebral aneurysm surgery

516 Crit Care Med 2006 Vol. 34, No. 2 was first described in the 1950s; however, that clot formation, both in the intravas- tients, p Ͻ .001), without any increase in interest in this technique declined in cular and extravascular compartments, is the rate of intracranial bleeding (113). light of very poor clinical outcomes. More a key determinant of cerebral injury after There was a potential bias because patients recently, several studies have evaluated aSAH. The presence of a thick clot in the in the placebo group were more severely mild hypothermia (38–35°C) as a protec- subarachnoid space is consistently re- affected (worse Hunt and Hess grade). Of tive strategy during aneurysm surgery. In ported as one of the strongest predictors note, another larger randomized trial of an early randomized trial, 114 patients of vasospasm after aSAH (108). The poor 170 patients suggested that enoxaparin had with ruptured and unruptured intracra- neuroanatomic correlation between ra- no effect on the outcome of aSAH and nial aneurysms who received mild intra- diologically defined vasospasm and cere- seemed to increase the risk of intracranial operative hypothermia had a lower rate of bral infarction suggests alternate or com- hemorrhage (114). neurologic deterioration at 24 and 72 hrs plementary hypotheses of DIND, among Antiplatelet Agents. Platelet aggrega- after surgery, a greater frequency of dis- which intravascular thrombosis is a pos- tion and release of thromboxanes have charge to home, and a greater rate of tulated mechanism. been demonstrated after aSAH, and these good long-term outcomes (102). How- Thorombolytics. Experimental and changes are more pronounced in patients ever, a subsequent multiple-center study clinical evidence indicate that a sub- who develop DIND (115). These data, in of 1,001 patients with aSAH who had arachnoid clot may be removed with the combination with the identification of World Federation of Neurologic Surgeons help of thrombolytic agents. Cisternal ir- microembolic signals by TCD in patients scores of I, II, or III failed to detect any rigation with recombinant tissue plas- with aSAH (116), have prompted several beneficial effect of intraoperative cooling minogen activator was found to be safe small clinical trials examining the thera- on 3-month neurologic outcomes or on (109), and intrathecal urokinase infusion peutic potential of antiplatelet agents. any perioperative outcomes (103). Mov- has been associated with a decreased rate Meta-analysis of these results indicated ing beyond the perioperative setting, of vasospasm and permanent neurologic that in patients treated with aspirin or small-scale investigations indicate that deficits (110). A systematic review of nine antiplatelet agents, the RR of DIND was mild hypothermia may be helpful in cases studies (only one of which was random- significantly reduced (RR, 0.65; 95% CI, in which vasospasm is refractory to con- ized) noted that cisternal thrombolysis 0.47–0.89), without a concomitant in- ventional treatment (104, 105). The ef- was associated with an absolute risk re- crease in the risk of intracranial hemor- fects of hypothermia on the relationship duction of 14.4% for DIND (p Ͻ .001), rhage. There was also a trend toward a between CBF and cerebral metabolic ox- 9.5% for poor Glasgow Outcome Scale reduced risk of poor outcome in patients ygen demand in patients with aSAH are scores (p Ͻ .01), and 4.5% for death (p Ͻ receiving antiplatelet agents (RR, 0.87; poorly understood. Hypothermia was as- .05) (111). Treatment effects did not sig- 95% CI, 0.65–1.17) (117). sociated with evidence of brain ischemia nificantly differ among the studies on the Comment. Preliminary data indicate in one study (106), but another study did basis of the type of thrombolytic agent that placement of thrombolytic agents in not confirm this finding (107). used (recombinant tissue plasminogen the subarachnoid space may improve out- Comment. Although considerable re- activator vs. urokinase) or the method of come after aSAH, but adequately pow- search has focused on the possibility of administration (intraoperative vs. postop- ered, randomized, controlled studies are neuroprotection in patients with aSAH, erative). Studies that enrolled only pa- needed to substantiate these findings. only oral nimodipine has been clearly as- tients at high risk for vasospasm seemed The two randomized trials assessing the sociated with improved outcome. Meta- to demonstrate greater treatment effects association between low-molecular- analysis suggests a survival benefit with (109). In an intriguing recent trial, weight heparin and aSAH outcome are high-dose tirilazad mesylate in patients Hamada et al. (110) randomized 110 pa- contradictory. Regarding antiplatelet with poor neurologic grade. Preliminary tients to endovascular coiling with or therapy, although a beneficial effect is results with magnesium and statins need without infusion of urokinase into the suggested, the evidence does not support substantiation in larger randomized pla- subarachnoid space via a microcatheter the routine use of these agents in aSAH at cebo-controlled trials. Finally, no benefit inserted in the lumbar space. The rate of the present time. Randomized trials with was seen for intraoperative mild hypo- symptomatic vasospasm was 8.8% in the greater power would be needed to clarify thermia in aSAH patients with good neu- urokinase-treated group vs. 30.2% in the the role of these therapies. rologic grade; however, further studies untreated group (p ϭ .012). are needed to explore other methodolo- Anticoagulants. Heparin, an anti- Controversy 6: Seizure gies and applications of therapeutic hypo- thrombin III agonist, has biological prop- Prophylaxis thermia, for example, more prolonged erties that include anticoagulation, modu- cooling protocols and its effects on pa- lation of inflammation, neuroprotection, Aneurysmal SAH has been linked to an tients with poor neurologic grade or re- and antiproliferative effects. In an animal increased risk of seizures, yet the role of fractory vasospasm. model of aSAH, administration of heparin seizure prophylaxis is controversial. In a has been associated with smooth muscle retrospective analysis of 217 surgically Controversy 5: Antithrombotic relaxation, increased CBF, and a reduction treated patients with aSAH over a 2-yr Agents in proliferative angiopathy (112). A recent period, 20% experienced one or more sei- randomized, placebo-controlled trial that zures, with more than half of these oc- Although the conventional view of an- enrolled 120 patients indicated that low- curring in the perioperative period; of eurysmal rupture emphasizes the patho- molecular-weight heparin significantly re- note, the occurrence of perioperative sei- physiologic role of hemorrhage and the duced the rate of vasospasm-related cere- zures did not predict epilepsy in this need to prevent recurrent intracranial bral infarction (3.5% of enoxaparin-treated study (118). Another investigation re- bleeding, data have accrued to suggest patients and 28.3% of placebo-treated pa- vealed that 24 of 121 aSAH patients with

Crit Care Med 2006 Vol. 34, No. 2 517 Table 1. Summary of controversies in the management of aneurysmal subarachnoid hemorrhage (SAH)

Controversy Best Available Evidence and Recommendations Future Studies

Surgical vs. endovascular aneurysm Level I evidence in favor of endovascular Comparative long-term follow-up of endovascular vs. exclusion management (grade A) surgical patients. New randomized trials to test endovascular vs. surgical therapy in patient subsets that were not represented in the ISAT trial. Diagnosis of vasospasm Transcranial Doppler (TCD) Level III evidence for high PPV and specificity of Prospective study comparing predictive value of CTA TCD for MCA (grade C) and TCD with angiography in detecting vasospasm. Computerized tomography angiography Level III comparing CTA to angiography and TCD Observational study to compare CTA/CT perfusion, (CTA) (grade C) perfusion/diffusion weighted MRI, and conventional angiography in detecting clinically significant vasospasm. MRI, PET, SPECT Level V evidence (grade C) Cerebral microdialysis Level V evidence Prospective clinical studies to understand the correlation between neurochemical abnormalities and clinical events in SAH. Outcome-based trials to assess the effect of microdialysis-guided management. Treatment of vasospasm Vasospasm prophylaxis Level II evidence showing no effect of triple-H Prospective randomized study comparing outcomes in prophylaxis on DIND (grade B) patients in vasospasm receiving either triple-H therapy or undergoing immediate endovascular intervention. Prospective studies to identify high- risk patients who may be candidates for prophylactic management of vasospasm. Vasospasm treatment Level III evidence supporting use of hemodynamic Adequately powered prospective, randomized trial augmentation as treatment of vasospasm (grade evaluating hemodynamic augmentation vs. C) conventional hemodynamic goals in symptomatic vasospasm. Prospective, randomized trials to compare a strategy of cardiac output augmentation vs. arterial pressure augmentation. Hemodynamic end points Level IV evidence for use of cardiac output goals over arterial pressure goals for hemodynamic augmentation (grade C) Intraaortic balloon counterpulsation (IABC) Level IV evidence for use of IABC Neuroprotection Caϩ channel blockers Level I evidence favoring use of nimodipine and Search for alternative neuroprotectants based on against use of nicardipine and AT877 (grade A) animal studies for possible synergy with or superiority to nimodipine. Tirilazad mesylate Level II evidence in favor of tirilazad for high grade SAH (grade A) Glucocorticoids, magnesium, Level II evidence for statins, magnesium (grade B) Randomized, placebo-controlled study of these agents endothelin receptor antagonists, and in patients with aneurysmal SAH; larger hydroxymethylglutaryl coenzyme A randomized trials evaluating statins and magnesium reductase inhibitors in SAH. Level V evidence favoring use of other agents (grade C) Hypothermia Level I evidence showing no benefit of Prospective, randomized studies to define the role of intraoperative hypothermia (grade A); use in hypothermia in patients at high risk to develop other settings unclear vasospasm (prophylaxis) or in patients with clinical vasospasm. Thrombolytics Level IIa evidence for use of thrombolytics (grade Prospective, randomized trials with sufficient B) statistical power to detect the efficacy of antiplatelet agents, anticoagulants, and intraoperative thrombolytics with appropriate end points (recurrent hemorrhage, DIND, mortality, and morbidity/functional outcome). Anticoagulation Antiplatelet agents Level IIb evidence in favor of use of antiplatelet agents (grade B) Seizure prophylaxis Level V evidence in favor of prophylaxis Randomized, placebo-controlled studies of anticonvulsants in aneurysmal SAH with risk stratification based on grade of SAH, location of aneurysm, and surgical intervention (craniotomy). Level III evidence against prophylaxis (grade C) Cardiac sequelae of SAH Myocardial dysfunction Level III evidence in favor of adrenergic receptor Prospective, observational studies to identify blockade pathophysiology, diagnosis, management, and outcome Level IV evidence in favor of inotropic support or IABC of neurocardiogenic injury and to distinguish from other forms of myocardial dysfunction.

ISAT, International Subarachnoid Aneurysm Trial; MRI, magnetic resonance imaging; PET, positron emission tomography; SPECT, single-photon emission computed tomography; PPV, positive predictive value; MCA, middle cerebral artery; CT, computerized tomography; triple-H, hypertension/ hypervolemia/hemodilution therapy; DIND, delayed ischemic neurologic deficits. aIntervention decreased DIND and improved neurologic outcomes but not mortality rates; bintervention decreased DIND but did not improve neurologic outcome.

518 Crit Care Med 2006 Vol. 34, No. 2 Table 2. Levels of evidence and strength of recommendation

Level of evidence Level I Data from randomized trials with low false-positive (␣) and low false-negative (␤) errors Level II Data from randomized trials with high false-positive (␣) or high false-negative (␤) errors Level III Data from nonrandomized concurrent cohort studies Level IV Data from nonrandomized cohort studies using historical controls Level V Data from anecdotal case series Strength of recommendation Grade A Supported by level I evidence Grade B Supported by level II evidence Grade C Supported by levels III–V evidence

Adapted from Cook et al. (157) and Broderick JP et al (158). clipped aneurysms had at least one sei- tion based on focal parenchymal pathol- (BNP) (130–133). Others have observed a zure and that ten patients had two or ogy, aneurysm location, aSAH grade, age, link between increased levels of BNP and more seizures after hospital discharge and history of hypertension. the development of cerebral vasospasm (119). Seizure risk has been linked to the (134, 135). Recent work suggests that thickness of the aSAH clot (120), to loca- Controversy 7: Cerebral Salt cardiac dysfunction (136) and triple-H tion of the aneurysm on the MCA (121), Wasting Syndrome therapy (137) are factors that might stim- to the presence of a subdural hematoma, ulate BNP release after aSAH. When these and to cerebral infarction (122). In one Hyponatremia occurs in up to 30% of results are considered collectively, it re- study, variables associated with the devel- patients after aSAH and has been associated mains unclear whether the relationship opment of epilepsy were, in order of im- with several disorders, most notably cere- between increased BNP levels and the de- portance, a history of hypertension, cere- bral salt wasting (CSW) syndrome and the velopment of CSW or vasospasm is caus- bral infarction, and duration of impaired syndrome of inappropriate antidiuretic ative or merely circumstantial. consciousness after the seizure (123). hormone (SIADH) secretion (126). CSW in- Therapeutic options for CSW are lim- Although empirical prophylaxis of sei- volves renal salt loss leading to a negative ited. It has been argued that increasing zures after aSAH might seem reasonable, sodium balance, hyponatremia, and intra- salt intake during CSW only further en- the benefit of preventive anticonvulsant vascular volume depletion, whereas SIADH hances sodium excretion. The mineralo- administration has yet to be demon- is characterized by an inability to appropri- corticoid fludrocortisone, which acts di- strated in a prospective randomized trial. ately excrete free water, resulting in a euv- rectly on the kidney tubules to enhance A cohort study of 123 aSAH patients fol- olemic or hypervolemic state. Because sodium resorption, has been shown to lowed up at 4 to 7 yrs found no evidence therapeutic interventions for CSW and SI- prevent intravascular volume depletion for the effectiveness of prophylactic anti- ADH are radically opposed, clinical differ- in patients with aSAH (138). In a related convulsants (119). A recent retrospective entiation between the two entities is essen- study, Hasan et al. (139) found that 0.2 evaluation of 527 patients with aSAH in- tial. It has been suggested that mg of fludrocortisone given intrave- dicated a strong association between phe- hypouricemia and an increased fractional nously or orally twice a day in 46 patients nytoin exposure in the setting and func- excretion of uric acid are more consistent with aSAH significantly reduced the fre- tional and cognitive disability (124). In a with a diagnosis of CSW (127). However, quency of a negative sodium balance and study of 101 patients with aSAH who had physiologic indicators of intravascular vol- led to smaller decreases in plasma vol- unexplained coma or neurologic deterio- ume are generally viewed as the most reli- ume and, consequently, less risk of cere- ration, eight (all of whom were receiving able way to distinguish between these con- bral ischemia. More recently, Moro et al. prophylactic anticonvulsants) were found ditions (128). (140) demonstrated that hydrocortisone to be in nonconvulsive status epilepticus. The pathogenesis of CSW is poorly un- attenuated excessive natriuresis in the Despite successful termination of sei- derstood (129). The main pathologic pro- setting of aSAH with development of hy- zures in five of eight patients, all eight cesses that have been linked with this ponatremia in 43% of treated patients vs. patients eventually died after a period of derangement are a) decreased sympa- 0% in the untreated group. prolonged coma (125). thetic input to the kidney leading to a Comment. Although the pathogenesis Comment. The available evidence sug- deficient regulation of proximal tubule of CSW needs clarification, its conse- gests that prophylaxis of seizures may be sodium resorption and to an inadequate quences of hyponatremia and intravascu- useful in aSAH patients with stroke or rise in renin and aldosterone in response lar volume depletion can be deleterious other distinct focal pathology. Indiscrim- to hypovolemia, and b) increased levels of in the setting of aSAH. Preliminary data inate administration of anticonvulsants circulating natriuretic peptides. The na- support the careful use of salt-conserving to patients with aSAH has been linked triuretic peptides are produced in the medications such as fludrocortisone or with unfavorable functional and cognitive heart, brain, and endothelium and pro- hydrocortisone in patients with aSAH. outcomes. A randomized, placebo-con- mote vasodilation, sodium excretion, and trolled trial is warranted to assess the diuresis. Several groups have demon- Controversy 8: Management of effect of anticonvulsant prophylaxis on strated an association between aSAH, hy- Cardiac Dysfunction After aSAH the prevalence of early and late seizures ponatremia, volume depletion, and in- in patients with aSAH. Such a trial would creased blood natriuretic peptide levels, A spectrum of cardiac sequelae have ideally incorporate predefined stratifica- in particular B-type natriuretic peptide been described in patients after aSAH,

Crit Care Med 2006 Vol. 34, No. 2 519 incidence of congestive heart failure with the use of inotropic agents such as (148). Another group found that the sen- dobutamine (65, 147, 156). Finally, pa- any aspects of sitivity of cardiac troponin I in the detec- tients presenting with severe neurocar- care in patients tion of echocardiographically demon- diomyopathy or cardiogenic shock and strated left ventricular dysfunction was vasospasm may benefit from temporary with aneurys- significantly higher than CK-MB (100% support with IABC (108, 110). M compared with 29%) (149). In a third Comment. Cardiac dysfunction com- mal subarachnoid hemor- study, mild elevations in cardiac troponin plicating aSAH is common and has been I(Ͻ2.8 ng/mL) associated with depressed linked to poor neurologic outcomes. rhage remain highly contro- left ventricular function in the absence of These complications may contraindicate, versial and warrant further significant electrocardiographic changes or be exacerbated by, brain-targeted ther- were proposed as characteristics that apies such as hemodynamic augmenta- resolution with hypothesis- might aid in the differentiation of neuro- tion. Studies are needed to elucidate risk genic stunned myocardium from myocar- factors, pathogeneses, natural history, di- driven clinical or transla- dial infarction (150). Recently, there has agnostic markers, management strate- been increasing recognition of the impor- gies, and prognostic impact of these com- tional research. tance of BNP as a diagnostic and prog- plications. nostic marker in patients with cardiac dysfunction. Increased serum levels of CONCLUSIONS including electrocardiographic changes, BNP are highly predictive of the short- reversible cardiomyopathy (stunned myo- and long-term risk of cardiac death This review highlights some of the cardium), and release of cardiac-specific across the entire spectrum of acute cor- controversies that exist in the clinical markers (141, 142). Aneurysmal SAH has onary syndromes and in patients with de- management of patients with aSAH, a also been linked to hypoxemic respiratory compensated congestive heart failure summary of which is presented in Table failure, which may take the form of con- (151). A correlation between increased 1, along with a classification of levels of gestive heart failure, neurogenic pulmo- levels of BNP and myocardial necrosis, evidence (157, 158) in Table 2. nary edema, or acute lung injury (143). pulmonary edema, and both systolic and It is anticipated that these areas of Extracerebral organ dysfunction in aSAH diastolic left ventricular dysfunction after uncertainty will generate hypothesis- has been associated with a greater risk of aSAH was demonstrated in a recent study driven experimental and translational re- poor neurologic outcome and death (136). However, there is debate as to the search, prospective observational studies, (144). Although there is general agree- source—cardiac or cerebral—of BNP in and clinical trials, with the goal of im- ment regarding the importance of cardiac patients with aSAH (152). A recent study proving outcomes in these complex and complications after aSAH, many aspects evaluated serum and CSF levels of atrial challenging patients. of their pathophysiology, diagnosis, and natriuretic peptide, BNP, and troponin T treatment remain undefined. in patients with aSAH. Of note, increased ACKNOWLEDGMENT A widely held view postulates a se- levels of atrial natriuretic peptide and quence of events involving acute cerebral BNP were detected in the serum but not We thank Tzipora Sofare, MA, for her injury, derangements in autonomic func- CSF, leading to the conclusion that these editorial assistance in preparing this ar- tion, release of endogenous cat- peptides were exclusively of cardiac ori- ticle. echolamines, and activation of adrenergic gin (153). This contradicted previous re- receptors, culminating in target organ sults in which a selective increase of BNP, REFERENCES damage (145). The severity of neurologic but not atrial natriuretic peptide, sug- compromise as measured by the Hunt gested that the hormone was secreted in 1. Guglielmi G, Vinuela F, Duckwiler G, et al: and Hess grade is highly predictive of the brain (134). Endovascular treatment of posterior circu- myocardial necrosis after aSAH, support- The therapeutic approach to patients lation aneurysms by electrothrombosis us- ing the hypothesis that cardiac injury af- with cardiac dysfunction acquired in the ing electrically detachable coils. J Neuro- ter aSAH is a neurally mediated process setting of aSAH is a matter of further surg 1992; 77:515–524 2. Molyneux A, Kerr R, Stratton I, et al: Inter- (146). Significantly, impaired left ventric- controversy. In light of the postulated national Subarachnoid Aneurysm Trial ular function and low cardiac output el- role of catecholamines in the develop- (ISAT) of neurosurgical clipping versus en- evate the risk of cerebral ischemia and ment of this disorder, early randomized dovascular coiling in 2143 patients with have been identified as independent pre- trials evaluated the effect of pharmaco- ruptured intracranial aneurysms: A ran- dictors of symptomatic vasospasm (147). logic ␣ and ␤ adrenergic receptor block- domized trial. International Subarachnoid Regarding diagnosis, several investi- ade (e.g., phentolamine and propanolol) Aneurysm Trial (ISAT) Collaborative Group. gators have evaluated the significance of in patients with aSAH, noting a favorable Lancet 2002; 360:1267–1274 cardiac-specific markers such as cardiac effect on the rate of myocardial necrosis, 3. Molyneux A, Kerr R, Ly-Mee Y, et al: Inter- troponin I in patients after aSAH. In a survival, and neurologic outcome (154, national subarachnoid aneurysm trial 7-day observation of 39 patients with 155). These data are at variance with (ISAT) of neurosurgical clipping versus en- dovascular coiling in 2143 patients with aSAH, measurements of cardiac troponin more recent studies that suggest that ruptured intracranial aneurysms: A ran- I revealed a higher incidence of myocar- neurogenically induced ventricular dys- domised comparison of effects on survival, dial injury (21%) than predicted by function and its associated low cardiac dependency, seizures, rebleeding, sub- CK-MB (13%), and elevations of cardiac output/pulmonary edema may exacerbate groups, and aneurysm occlusion. Lancet troponin I were associated with a higher cerebral ischemia and should be treated 2005; 366:809–817

520 Crit Care Med 2006 Vol. 34, No. 2 4. Barreau X, Pastore M, Piotin M, et al: En- Guglielmi detachable coil treatment of rup- rebral artery vasospasm following subarach- dovascular treatment of cerebral vasospasm tured saccular cerebral aneurysms: Retro- noid haemorrhage. Br J Neurosurg 1993; following S.A.H. Acta Neurochir Suppl spective review of a 10-year single-center 7:291–298 2001; 77:177–180 experience. AJNR Am J Neuroradiol 2003; 32. Lindegaard KF, Nornes H, Bakke SJ, et al: 5. Hop JW, Rinkel GJ, Algra A, et al: Quality of 24:526–533 Cerebral vasospasm after subarachnoid life in patients and partners after aneurys- 19. Charpentier C, Audibert G, Guillemin F, et haemorrhage investigated by means of mal subarachnoid hemorrhage. Stroke al: Multivariate analysis of predictors of ce- transcranial Doppler ultrasound. Acta Neu- 1998; 29:798–804 rebral vasospasm occurrence after aneurys- rochir Suppl (Wien) 1988; 42:81–84 6. Ostbye T, Levy AR, Mayo NE: Hospitaliza- mal subarachnoid hemorrhage. Stroke 33. Torbey MT, Hauser TK, Bhardwaj A, et al: tion and case-fatality rates for subarachnoid 1999; 30:1402–1408 Effect of age on cerebral blood flow velocity hemorrhage in Canada from 1982 through 20. Dehdashti AR, Mermillod B, Rufenacht DA, and incidence of vasospasm after aneurys- 1991: The Canadian Collaborative Study et al: Does treatment modality of intracra- mal subarachnoid hemorrhage. Stroke Group of Stroke Hospitalizations. Stroke nial ruptured aneurysms influence the in- 2001; 32:2005–2011 1997; 28:793–798 cidence of cerebral vasospasm and clinical 34. Proust F, Callonec F, Clavier E, et al: Use- 7. Cesarini KG, Hardemark HG, Persson L: outcome? Cerebrovasc Dis 2004; 17:53–60 fulness of transcranial color-coded sonogra- Improved survival after aneurysmal sub- 21. Goddard AJ, Raju PP, Gholkar A: Does the phy in the diagnosis of cerebral vasospasm. arachnoid hemorrhage: Review of case method of treatment of acutely ruptured Stroke 1999; 30:1091–1098 management during a 12-year period. intracranial aneurysms influence the inci- 35. Lysakowski C, Walder B, Costanza MC, et al: J Neurosurg 1999; 90:664–672 dence and duration of cerebral vasospasm Transcranial Doppler versus angiography in 8. Rinkel GJ, Feigin VL, Algra A, et al: Calcium and clinical outcome? J Neurol Neurosurg patients with vasospasm due to a ruptured antagonists for aneurysmal subarachnoid Psychiatry 2004; 75:868–872 cerebral aneurysm: A systematic review. haemorrhage. Cochrane Database Syst Rev 22. Groden C, Kremer C, Regelsberger J, et al: Stroke 2001; 32:2292–2298 2005; 25:CD000277 Comparison of operative and endovascular 36. Takagi R, Hayashi H, Kobayashi H, et al: 9. Rinkel G, Feigin V, Algra A, et al: Circula- treatment of anterior circulation aneu- Three-dimensional CT angiography of in- tory volume expansion therapy for aneurys- rysms in patients in poor grades. Neurora- tracranial vasospasm following subarach- mal subarachnoid haemorrhage. Cochrane diology 2001; 43:778–783 noid haemorrhage. Neuroradiology 1998; Database Syst Rev 2004; 4:CD000483 23. Rabinstein AA, Pichelmann MA, Friedman 40:631–635 10. Koivisto T, Vanninen R, Hurskainen H, et JA, et al: Symptomatic vasospasm and out- 37. Anderson GB, Ashforth R, Steinke DE, et al: al: Outcomes of early endovascular versus comes following aneurysmal subarachnoid CT angiography for the detection of cere- surgical treatment of ruptured cerebral an- hemorrhage: A comparison between surgi- bral vasospasm in patients with acute sub- eurysms: A prospective randomized study. cal repair and endovascular coil occlusion. arachnoid hemorrhage. AJNR Am J Neuro- Stroke 2000; 31:2369–2377 J Neurosurg 2003; 98:319–325 radiol 2000; 21:1011–1015 11. Raabe A, Schmiedek P, Seifert V, et al: Ger- 24. Gruber A, Ungersbock K, Reinprecht A, et 38. Otawara Y, Ogasawara K, Ogawa A, et al: man Society of Neurosurgery Section on al: Evaluation of cerebral vasospasm after Evaluation of vasospasm after subarachnoid Vascular Neurosurgery: Position statement early surgical and endovascular treatment hemorrhage by use of multislice computed on the International Subarachnoid Hemor- of ruptured intracranial aneurysms. Neuro- tomographic angiography. Neurosurgery rhage Trial (ISAT). Zentralbl Neurochir surgery 1998; 42:258–267 2002; 51:939–942 2003; 64:99–103 25. Bleck TP: Rebleeding and vasospasm after 39. Rabinstein AA, Friedman JA, Weigand SD, 12. Sade B, Mohr G: Critical appraisal of the SAH: New strategies for improving out- et al: Predictors of cerebral infarction in International Subarachnoid Aneurysm Trial come. J Crit Illn 1997; 12:572–582 aneurysmal subarachnoid hemorrhage. (ISAT). Neurol India 2004; 52:32–35 26. Cloft HJ, Joseph GJ, Dion JE: Risk of cere- Stroke 2004; 35:1862–1866 13. Haley EC Jr, Kassell NF, Apperson-Hansen bral angiography in patients with subarach- 40. Condette-Auliac S, Bracard S, Anxionnat R, C, et al: A randomized, double-blind, vehi- noid hemorrhage, cerebral aneurysm, and et al: Vasospasm after subarachnoid hemor- cle-controlled trial of tirilazad mesylate in arteriovenous malformation: A meta- rhage: Interest in diffusion-weighted MR patients with aneurysmal subarachnoid analysis. Stroke 1999; 30:317–320 imaging. Stroke 2001; 32:1818–1824 hemorrhage: A cooperative study in North 27. Aaslid R, Huber P, Nornes H: Evaluation of 41. Hertel F, Walter C, Bettag M, et al: Perfu- America. J Neurosurg 1997; 86:467–474 cerebrovascular spasm with transcranial sion-weighted magnetic resonance imaging 14. Beneficial effect of carotid endarterectomy Doppler ultrasound. J Neurosurg 1984; 60: in patients with Vasospasm: A useful new in symptomatic patients with high-grade 37–41 tool in the management of patients with carotid stenosis: North American Symp- 28. Vora YY, Suarez-Almazor M, Steinke DE, et subarachnoid hemorrhage. Neurosurgery tomatic Carotid Endarterectomy Trial Col- al: Role of transcranial Doppler monitoring 2005; 56:28–35 laborators. N Engl J Med 1991; 325: in the diagnosis of cerebral vasospasm after 42. Rajendran JG, Lewis DH, Newell DW, et al: 445–453 subarachnoid hemorrhage. Neurosurgery Brain SPECT used to evaluate vasospasm 15. Endarterectomy for asymptomatic carotid 1999; 44:1237–1247 after subarachnoid hemorrhage: Correla- artery stenosis: Executive Committee for 29. Okada Y, Shima T, Nishida M, et al: Com- tion with angiography and transcranial the Asymptomatic Carotid Atherosclerosis parison of transcranial Doppler investiga- Doppler. Clin Nucl Med 2001; 26:125–130 Study. JAMA 1995; 273:1421–1428 tion of aneurysmal vasospasm with digital 43. Jabre A, Babikian V, Powsner RA, et al: Role 16. Byrne JV, Sohn MJ, Molyneux AJ, et al: subtraction angiographic and clinical find- of single photon emission computed to- Five-year experience in using coil emboli- ings. Neurosurgery 1999; 45:443–449 mography and transcranial Doppler ultra- zation for ruptured intracranial aneurysms: 30. Venkatesh B, Shen Q, Lipman J: Continu- sonography in clinical vasospasm. J Clin Outcomes and incidence of late rebleeding. ous measurement of cerebral blood flow Neurosci 2002; 9:400–403 J Neurosurg 1999; 90:656–663 velocity using transcranial Doppler reveals 44. Powers WJ, Grubb RL Jr, Baker RP, et al: 17. Sluzewski M, Menovsky T, van Rooij WJ, et significant moment-to-moment variability Regional cerebral blood flow and metabo- al: Coiling of very large or giant cerebral of data in healthy volunteers and in patients lism in reversible ischemia due to vaso- aneurysms: Long-term clinical and serial with subarachnoid hemorrhage. Crit Care spasm: Determination by positron emission angiographic results. AJNR Am J Neurora- Med 2002; 30:563–569 tomography. J Neurosurg 1985; 62: diol 2003; 24:257–262 31. Grosset DG, Straiton J, McDonald I, et al: 539–546 18. Friedman JA, Nichols DA, Meyer FB, et al: Angiographic and Doppler diagnosis of ce- 45. Carpenter DA, Grubb RL Jr, Tempel LW, et

Crit Care Med 2006 Vol. 34, No. 2 521 al: Cerebral oxygen metabolism after aneu- and induced arterial hypertension. Neuro- liminary results. AJNR Am J Neuroradiol rysmal subarachnoid hemorrhage. J Cereb surgery 1990; 11:337–343 2004; 25:1067–1076 Blood Flow Metab 1991; 11:837–844 60. Origitano TC, Wascher TM, Reichman OH, 73. Badjatia N, Topcuoglu MA, Pryor JC, et al: 46. Sarrafzadeh AS, Sakowitz OW, Kiening KL, et al: Sustained increased cerebral blood Preliminary experience with intra-arterial et al: Bedside microdialysis: A tool to mon- flow with prophylactic hypertensive hyper- nicardipine as a treatment for cerebral va- itor cerebral metabolism in subarachnoid volemic hemodilution (“triple-H” therapy) sospasm. AJNR Am J Neuroradiol 2004; 25: hemorrhage patients? Crit Care Med 2002; after subarachnoid hemorrhage. Neurosur- 819–826 30:1062–1070 gery 1990; 27:729–739 74. Muizelaar JP, Zwienenberg M, Rudisill NA, 47. Skjoth-Rasmussen J, Schulz M, Kristensen 61. Rosenwasser RH, Delgado TE, Buchheit et al: The prophylactic use of transluminal SR, et al: Delayed neurological deficits de- WA, et al: Control of hypertension and pro- balloon angioplasty in patients with Fisher tected by an ischemic pattern in the extra- phylaxis against vasospasm in cases of sub- Grade 3 subarachnoid hemorrhage: A pilot cellular cerebral metabolites in patients arachnoid hemorrhage: A preliminary re- study. J Neurosurg 1999; 91:51–58 with aneurysmal subarachnoid hemor- port. Neurosurgery 1983; 12:658–660 75. Apostolides PJ, Greene KA, Zabramski JM, rhage. J Neurosurg 2004; 100:8–15 62. Lennihan L, Mayer SA, Fink ME, et al: Ef- et al: Intra-aortic balloon pump counterpul- 48. Unterberg AW, Sakowitz OW, Sarrafzadeh fect of hypervolemic therapy on cerebral sation in the management of concomitant AS, et al: Role of bedside microdialysis in blood flow after subarachnoid hemorrhage: cerebral vasospasm and cardiac failure after the diagnosis of cerebral vasospasm follow- A randomized controlled trial. Stroke 2000; subarachnoid hemorrhage: Technical case ing aneurysmal subarachnoid hemorrhage. 31:383–391 report. Neurosurgery 1996; 38:1056–1059 J Neurosurg 2001; 94:740–749 63. Egge A, Waterloo K, Sjoholm H, et al: Pro- 76. Nussbaum ES, Sebring LA, Ganz WF, et al: 49. Sarrafzadeh AS, Haux D, Ludemann L, et al: phylactic hyperdynamic postoperative fluid Intra-aortic balloon counterpulsation aug- Cerebral ischemia in aneurysmal subarach- therapy after aneurysmal subarachnoid ments cerebral blood flow in the patient noid hemorrhage: A correlative microdialy- hemorrhage: A clinical, prospective, ran- with cerebral vasospasm: A xenon-enhanced sis-PET study. Stroke 2004; 35:638–643 domized, controlled study. Neurosurgery computed tomography study. Neurosurgery 50. Bhardwaj A: In vivo regional neurochemis- 2001; 49:593–605 1998; 42:206–213 try in stroke: Clinical applications, limita- 64. Treggiari MM, Walder B, Suter PM, et al: 77. Spann RG, Lang DA, Birch AA, et al: Intra- tions, and future directions. Stroke 2004; Systematic review of the prevention of de- aortic balloon counterpulsation: Augmenta- 35:e74–e76 layed ischemic neurological deficits with tion of cerebral blood flow after aneurysmal 51. Peerdeman SM, van Tulder MW, Vandertop hypertension, hypervolemia, and hemodilu- subarachnoid haemorrhage. Acta Neuro- WP: Cerebral microdialysis as a monitoring tion therapy following subarachnoid hem- chir (Wien) 2001; 143:115–123 method in subarachnoid hemorrhage pa- orrhage. J Neurosurg 2003; 98:978–984 78. Montessuit M, Chevalley C, King J, et al: tients, and correlation with clinical events: 65. Levy ML, Rabb CH, Zelman V, et al: Cardiac The use of intra-aortic counterpulsation A systematic review. J Neurol 2003; 250: performance enhancement from dobut- balloon for the treatment of cerebral vaso- 797–805 amine in patients refractory to hypervole- spasm and edema. Surgery 2000; 127: 52. Rivierez M, Landau-Ferey J, Grob R, et al: mic therapy for cerebral vasospasm. J Neu- 230–233 Value of electroencephalogram in predic- rosurg 1993; 79:494–499 79. Rosen CL, Sekhar LN, Duong DH: Use of tion and diagnosis of vasospasm after intra- 66. Kim D, Joseph M, Ziadi S, et al: Increases in intra-aortic balloon pump counterpulsation cranial aneurysm rupture. Acta Neurochir cardiac output can reverse flow deficits for refractory symptomatic vasospasm. Acta (Wien) 1991; 110:17–23 from vasospasm independent of blood pres- Neurochir (Wien) 2000; 142:25–32 53. Vespa PM, Nuwer MR, Juhasz C, et al: Early sure: A study using xenon computed tomo- 80. Kassell NF, Haley EC Jr, Apperson-Hansen detection of vasospasm after acute sub- graphic measurement of cerebral blood C, et al: Randomized, double-blind, vehicle- arachnoid hemorrhage using continuous flow. Neurosurgery 2003; 53:1044–1051 controlled trial of tirilazad mesylate in pa- EEG ICU monitoring. Electroencephalogr 67. Bouma GJ, Muizelaar JP: Relationship be- tients with aneurysmal subarachnoid hem- Clin Neurophysiol 1997; 103:607–615 tween cardiac output and cerebral blood orrhage: A cooperative study in Europe, 54. Claassen J, Hirsch LJ, Kreiter KT, et al: flow in patients with intact and with im- Australia, and New Zealand. J Neurosurg Quantitative continuous EEG for detecting paired autoregulation. J Neurosurg 1990; 1996; 84:221–228 delayed cerebral ischemia in patients with 73:368–374 81. Lanzino G, Kassell NF, Dorsch NW, et al: poor-grade subarachnoid hemorrhage. Clin 68. Milnor WR: Pulsatile blood flow. N Engl Double-blind, randomized, vehicle-con- Neurophysiol 2004; 115:2699–2710 J Med 1972; 287:27–34 trolled study of high-dose tirilazad mesylate 55. Sen J, Belli A, Albon H, et al: Triple-H 69. Morgan MK, Jonker B, Finfer S, et al: Ag- in women with aneurysmal subarachnoid therapy in the management of aneurysmal gressive management of aneurysmal sub- hemorrhage: Part I. A cooperative study in subarachnoid haemorrhage. Lancet Neurol arachnoid haemorrhage based on a papav- Europe, Australia, New Zealand, and South 2003; 2:614–621 erine angioplasty protocol. J Clin Neurosci Africa. J Neurosurg 1999; 90:1011–1017 56. Kosnik EJ, Hunt WE: Postoperative hyper- 2000; 7:305–308 82. Lanzino G, Kassell NF: Double-blind, ran- tension in the management of patients with 70. Smith WS, Dowd CF, Johnston SC, et al: domized, vehicle-controlled study of high- intracranial arterial aneurysms. J Neuro- Neurotoxicity of intra-arterial papaverine dose tirilazad mesylate in women with an- surg 1976; 45:148–154 preserved with chlorobutanol used for the eurysmal subarachnoid hemorrhage: Part 57. Giannotta SL, McGillicuddy JE, Kindt GW: treatment of cerebral vasospasm after aneu- II. A cooperative study in North America. Diagnosis and treatment of postoperative rysmal subarachnoid hemorrhage. Stroke J Neurosurg 1999; 90:1018–1024 cerebral vasospasm. Surg Neurol 1977; 2004; 35:2518–2522 83. Dorsch NW, Kassell NF, Sinkula MS: Meta- 8:286–290 71. Feng L, Fitzsimmons BF, Young WL, et al: analysis of trials of tirilazad mesylate in 58. Awad IA, Carter LP, Spetzler RF, et al: Clin- Intraarterially administered verapamil as aneurysmal SAH. Acta Neurochir Suppl ical vasospasm after subarachnoid hemor- adjunct therapy for cerebral vasospasm: 2001; 77:233–235 rhage: Response to hypervolemic hemodilu- Safety and 2-year experience. AJNR Am J 84. Chen D, Nishizawa S, Yokota N, et al: High- tion and arterial hypertension. Stroke 1987; Neuroradiol 2002; 23:1284–1290 dose methylprednisolone prevents vaso- 18:365–372 72. Biondi A, Ricciardi GK, Puybasset L, et al: spasm after subarachnoid hemorrhage 59. Kassell NF, Peerless SJ, Durward QJ, et al: Intra-arterial nimodipine for the treatment through inhibition of protein kinase C ac- Treatment of ischemic deficits from vaso- of symptomatic cerebral vasospasm after tivation. Neurol Res 2002; 24:215–222 spasm with intravascular volume expansion aneurysmal subarachnoid hemorrhage: pre- 85. Shibata S, Suzuki S, Ohkuma H, et al: Ef-

522 Crit Care Med 2006 Vol. 34, No. 2 fects of intracisternal methylprednisolone bits subjected to subarachnoid hemorrhage. 111. Amin-Hanjani S, Ogilvy CS, Barker FG: on lipid peroxidation in experimental sub- Proc Natl Acad SciUSA2002; 99: Does intracisternal thrombolysis prevent arachnoid haemorrhage. Acta Neurochir 5627–5631 vasospasm after aneurysmal subarachnoid (Wien) 1999; 141:529–532 99. Wolf EW, Banerjee A, Soble-Smith J, et al: hemorrhage? A meta-analysis. Neurosur- 86. Chyatte D: Prevention of chronic cerebral Reversal of cerebral vasospasm using an in- gery 2004; 54:326–334 vasospasm in dogs with ibuprofen and high- trathecally administered nitric oxide donor. 112. Chimowitz MI, Pessin MS: Is there a role for dose methylprednisolone. Stroke 1989; 20: J Neurosurg 1998; 89:279–288 heparin in the management of complica- 1021–1026 100. Kwan AL, Lin CL, Yanamoto H, et al: Sys- tions of subarachnoid hemorrhage? Stroke 87. Chyatte D, Fode NC, Nichols DA, et al: Pre- temic administration of the potassium 1987; 18:1169–1172 liminary report: Effects of high dose meth- channel activator cromakalim attenuates 113. Wurm G, Tomancok B, Nussbaumer K, et ylprednisolone on delayed cerebral isch- cerebral vasospasm after experimental sub- al: Reduction of ischemic sequelae follow- emia in patients at high risk for vasospasm arachnoid hemorrhage. Neurosurgery ing spontaneous subarachnoid hemor- after aneurysmal subarachnoid hemor- 1998; 42:347–350 rhage: A double-blind, randomized compar- rhage. Neurosurgery 1987; 21:157–160 101. Thome C, Schubert GA, Schilling L: Hypo- ison of enoxaparin versus placebo. Clin 88. Veyna RS, Seyfried D, Burke DG, et al: Mag- thermia as a neuroprotective strategy in Neurol Neurosurg 2004; 106:97–103 nesium sulfate therapy after aneurysmal subarachnoid hemorrhage: A pathophysio- 114. Siironen J, Juvela S, Varis J, et al: No effect subarachnoid hemorrhage. J Neurosurg logical review focusing on the acute phase. of enoxaparin on outcome of aneurysmal 2002; 96:510–514 Neurol Res 2005; 27:229–237 subarachnoid hemorrhage: A randomized, 89. Van den Bergh: Magnesium sulfate in aneu- 102. Hindman BJ, Todd MM, Gelb AW, et al: Mild double-blind, placebo-controlled clinical rysmal subarachnoid hemorrhage: A ran- hypothermia as a protective therapy during trial. J Neurosurg 2003; 99:953–959 domized controlled trial. MASH Study intracranial aneurysm surgery: A random- 115. Ohkuma H, Suzuki S, Kimura M, et al: Role Group (abstract). Stroke 2005; 36:434(A85) ized prospective pilot trial. Neurosurgery of platelet function in symptomatic cerebral 90. Shaw MD, Vermeulen M, Murray GD, et al: 1999; 44:23–32 vasospasm following aneurysmal subarach- Efficacy and safety of the endothelin, recep- 103. Todd MM, Hindman BJ, Clarke WR, et al: noid hemorrhage. Stroke 1991; 22:854–859 tor antagonist TAK-044 in treating sub- Mild intraoperative hypothermia during 116. Romano JG, Forteza AM, Concha M, et al: arachnoid hemorrhage: A report by the surgery for intracranial aneurysm: Intraop- Detection of microemboli by transcranial Steering Committee on behalf of the UK/ erative Hypothermia for Aneurysm Surgery Doppler ultrasonography in aneurysmal Netherlands/Eire TAK-044 Subarachnoid Trial (IHAST) Investigators. N Engl J Med subarachnoid hemorrhage. Neurosurgery Haemorrhage Study Group. J Neurosurg 2005; 352:135–145 2002; 50:1026–1030 2000; 93:992–997 104. Nagao S, Irie K, Kawai N, et al: The use of 117. Dorhout Mees SM, Rinkel GJ, Hop JW, et al: 91. Mascia L, Fedorko L, Stewart DJ, et al: Tem- mild hypothermia for patients with severe Antiplatelet therapy in aneurysmal sub- poral relationship between endothelin-1 vasospasm: A preliminary report. J Clin arachnoid hemorrhage: A systematic re- concentrations and cerebral vasospasm in Neurosci 2003; 10:208–212 view. Stroke 2003; 34:2285–2289 patients with aneurysmal subarachnoid 105. Gasser S, Khan N, Yonekawa Y, et al: Long- 118. Lin CL, Dumont AS, Lieu AS, et al: Char- hemorrhage. Stroke 2001; 32:1185–1190 term hypothermia in patients with severe acterization of perioperative seizures and 92. Davignon J: Beneficial cardiovascular pleio- brain edema after poor-grade subarachnoid epilepsy following aneurysmal subarach- tropic effects of statins. Circulation 2004; hemorrhage: Feasibility and intensive care noid hemorrhage. J Neurosurg 2003; 99: 109(23 Suppl 1):III39–III43 complications. J Neurosurg Anesthesiol 978–985 93. McGirt MJ, Lynch JR, Parra A, et al: Sim- 2003; 15:240–248 119. Ogden JA, Utley T, Mee EW: Neurological vastatin increases endothelial nitric oxide 106. Sato K, Sato K, Yoshimoto T: Systemic and and psychosocial outcome 4 to 7 years after synthase and ameliorates cerebral vaso- cerebral haemodynamics during craniot- subarachnoid hemorrhage. Neurosurgery spasm resulting from subarachnoid hemor- omy under mild hypothermia in patients 1997; 41:25–34 rhage. Stroke 2002; 33:2950–2956 with acute subarachnoid haemorrhage. 120. Rhoney DH, Tipps LB, Murry KR, et al: 94. Parra A, Kreiter KT, Williams S, et al: Effect Acta Neurochir (Wien) 2000; 142: Anticonvulsant prophylaxis and timing of of prior statin use on functional outcome 1013–1019 seizures after aneurysmal subarachnoid and delayed vasospasm after acute aneurys- 107. Kawamura S, Suzuki A, Hadeishi H, et al: hemorrhage. Neurology 2000; 55:258–265 mal subarachnoid hemorrhage: A matched Cerebral blood flow and oxygen metabolism 121. Butzkueven H, Evans AH, Pitman A, et al: controlled cohort study. Neurosurgery during mild hypothermia in patients with Seizures independently predict poor out- 2005; 56:476–484 subarachnoid haemorrhage. Acta Neuro- come after subarachnoid hemorrhage. Neu- 95. Singhal AB, Topcuoglu MA, Dorer DJ, et al: chir (Wien) 2000; 142:1117–1121 rology 2000; 55:1315–1320 SSRI and statin use increases the risk for 108. Fisher CM, Kistler JP, Davis JM: Relation of 122. Claassen J, Peery S, Kreiter KT, et al: Pre- vasospasm after subarachnoid hemorrhage. cerebral vasospasm to subarachnoid hem- dictors and clinical impact of epilepsy after Neurology 2005; 64:1008–1013 orrhage visualized by computerized tomo- subarachnoid hemorrhage. Neurology 96. Lynch JR, Wang H, McGirt MJ, et al: Sim- graphic scanning. Neurosurgery 1980; 2003; 60:208–214 vastatin reduces vasospasm after aneurys- 6:1–9 123. Ohman J: Hypertension as a risk factor for mal subarachnoid hemorrhage: Results of a 109. Sasaki T, Kodama N, Kawakami M, et al: epilepsy after aneurysmal subarachnoid pilot randomized clinical trial. Stroke 2005; Urokinase cisternal irrigation therapy for hemorrhage and surgery. Neurosurgery 36:2024–2026 prevention of symptomatic vasospasm after 1990; 27:578–581 97. Tseng MY, Czosnyka M, Richards H, et al: aneurysmal subarachnoid hemorrhage: A 124. Naidech AM, Kreiter KT, Janjua N, et al: Effects of acute treatment with pravastatin study of urokinase concentration and the Phenytoin exposure is associated with func- on cerebral vasospasm, autoregulation, and fibrinolytic system. Stroke 2000; 31: tional and cognitive disability after sub- delayed ischemic deficits after aneurysmal 1256–1262 arachnoid hemorrhage. Stroke 2005; 36: subarachnoid hemorrhage: A phase II ran- 110. Hamada J, Kai Y, Morioka M, et al: Effect on 583–587 domized placebo-controlled trial. Stroke cerebral vasospasm of coil embolization fol- 125. Dennis LJ, Claassen J, Hirsch LJ, et al: Non- 2005; 36:1627–1632 lowed by microcatheter intrathecal uroki- convulsive status epilepticus after sub- 98. Grasso G, Buemi M, Alafaci C, et al: Bene- nase infusion into the cisterna magna: A arachnoid hemorrhage. Neurosurgery ficial effects of systemic administration of prospective randomized study. Stroke 2003; 2002; 51:1136–1143 recombinant human erythropoietin in rab- 34:2549–2554 126. Ogawasara K, Kinouchi H, Nagamine Y, et

Crit Care Med 2006 Vol. 34, No. 2 523 al: Differential diagnosis of hyponatremia 138. Wijdicks EFM, Vermeulen M, van Brum- orrhage. J Am Coll Cardiol 2000; 36: following subarachnoid hemorrhage. No melen P, et al: The effect of fludrocortisone 1328–1335 Shinkei Geka 1998; 26:501–505 acetate on plasma volume and natriuresis in 149. Deibert E, Barzilai B, Braverman AC, et al: 127. Maesaka JK, Gupta S, Fishbane S: Cerebral patients with aneurysmal subarachnoid Clinical significance of elevated troponin I salt-wasting syndrome: Does it exist? hemorrhage. Clin Neurol Neurosurg 1988; levels in patients with nontraumatic sub- Nephron 1999; 82:100–9 90:209–214 arachnoid hemorrhage. J Neurosurg 2003; 128. Palmer BF: Hyponatremia in patients with 139. Hasan D, Lindsay KW, Wijdicks EFM, et al: 98:741–746 central nervous system disease: SIADH ver- Effect of fludrocortisone acetate in patients 150. Bulsara KR, McGirt MJ, Liao L, et al: Use of sus CSW. Trends Endocrinol Metab 2003; with subarachnoid hemorrhage. Stroke the peak troponin value to differentiate 14:182–187 1989; 20:1156–1161 myocardial infarction from reversible neu- 129. Harrigan MR: Cerebral salt wasting syn- 140. Moro N, Katayama Y, Kojima J, et al: Pro- rogenic left ventricular dysfunction associ- drome: A review. Neurosurgery 1996; 38: phylactic management of excessive natri- ated with aneurysmal subarachnoid hemor- 152–160 uresis with hydrocortisone for efficient hy- rhage. J Neurosurg 2003; 98:524–528 130. Diringer M, Ladenson PW, Stern BJ, et al: pervolemic therapy after subarachnoid 151. Cowie MR, Jourdain P, Maisel A, et al: Clin- Plasma atrial natriuretic factor and sub- hemorrhage. Stroke 2003; 34:2807–2811 arachnoid hemorrhage. Stroke 1988; 19: 141. Macmillan CS, Grant IS, Andrews PJ: Pul- ical applications of B-type natriuretic pep- 1119–1124 monary and cardiac sequelae of subarach- tide (BNP) testing. Eur Heart J 2003; 24: 131. Berendes E, Walter M, Cullen P, et al: Se- noid haemorrhage: Time for active manage- 1710–1718 cretion of brain natriuretic peptide in pa- ment? Intensive Care Med 2002; 28: 152. Clerico A, Emdin M: Diagnostic accuracy tients with aneurysmal subarachnoid haem- 1012–1023 and prognostic relevance of the measure- orrhage. Lancet 1997; 349:245–249 142. Sakr YL, Ghosn I, Vincent JL: Cardiac man- ment of cardiac natriuretic peptides: A re- 132. Isotani E, Suzuki R, Tomita K, et al: Alter- ifestations after subarachnoid hemorrhage: view. Clin Chem 2004; 50:33–50 ations in plasma concentrations of natri- A systematic review of the literature. Prog 153. Espiner EA, Leikis R, Ferch RD, et al: The uretic peptides and antidiuretic hormone Cardiovasc Dis 2002; 45:67–80 neuro-cardio-endocrine response to acute after subarachnoid hemorrhage. Stroke 143. Mayer SA, Fink ME, Homma S, et al: Car- subarachnoid haemorrhage. Clin Endocri- 1994; 25:2198–2203 diac injury associated with neurogenic pul- nol 2002; 56:629–635 133. Tomida M, Muraki M, Uemura K, et al: monary edema following subarachnoid 154. Neil-Dwyer G, Walter P, Cruickshank JM, et Plasma concentrations of brain natriuretic hemorrhage. Neurology 1994; 44:815–820 al: Effect of propranolol and phentolamine peptide in patients with subarachnoid hem- 144. Gruber A, Reinprecht A, Illievich UM, et al: on myocardial necrosis after subarachnoid orrhage. Stroke 1998; 29:1584–1587 Extracerebral organ dysfunction and neuro- haemorrhage. BMJ 1978; 2:990–992 134. Sviri GE, Feinsod M, Soustiel JF: Brain na- logic outcome after aneurysmal subarach- 155. Walter P, Neil-Dwyer G, Cruickshank JM: triuretic peptide and cerebral vasospasm in noid hemorrhage. Crit Care Med 1999; 27: Beneficial effects of adrenergic blockade in subarachnoid hemorrhage: Clinical and 505–514 patients with subarachnoid haemorrhage. TCD correlations. Stroke 2000; 31:118–122 145. Masuda T, Sato K, Yamamoto S, et al: Sym- BMJ 1982; 284:1661–1664 135. McGirt MJ, Blessing R, Nimjee SM, et al: pathetic nervous activity and myocardial 156. Parr MJ, Finfer SR, Morgan MK: Revers- Correlation of serum brain natriuretic pep- damage immediately after subarachnoid ible cardiogenic shock complicating sub- tide with hyponatremia and delayed isch- hemorrhage in a unique animal model. arachnoid haemorrhage. BMJ 1996; 313: emic neurological deficits after subarach- Stroke 2002; 33:1671–1676 681–683 noid hemorrhage. Neurosurgery 2004; 54: 146. Tung P, Kopelnik A, Banki N, et al: Predic- 1369–1373 tors of neurocardiogenic injury after sub- 157. Cook DJ, Guyatt GH, Laupacis A, et al: 136. Tung PP, Olmsted E, Kopelnik A, et al: arachnoid hemorrhage. Stroke 2004; 35: Rules of evidence and clinical recommen- Plasma B-type natriuretic peptide levels are 548–551 dations on the use of antithrombotic associated with early cardiac dysfunction 147. Mayer SA, Lin J, Homma S, et al: Myocar- agents. Chest 1992; 102:305S–311S after subarachnoid hemorrhage. Stroke dial injury and left ventricular performance 158. Broderick JP, Adams HP Jr, Barsan W, et al: 2005; 36:1567–1569 after subarachnoid hemorrhage. Stroke Guidelines for the management for sponta- 137. Inoha S, Inamura T, Nakamizo A, et al: 1994; 30:780–786 neous intracerebral hemorrhage: State- Fluid loading in rats increases serum brain 148. Parekh N, Venkatesh B, Cross D, et al: Car- ment for healthcare professionals from a natriuretic peptide concentration. Neurol diac troponin I predicts myocardial dys- special writing group of the Stroke Council, Res 2001; 23:93–95 function in aneurysmal subarachnoid hem- AHA. Stroke 1999; 30:905–915

524 Crit Care Med 2006 Vol. 34, No. 2