Preconditioning Effect on Cerebral Vasospasm in Patients With Aneurysmal Subarachnoid Hemorrhage
Kim, Young Woo MD*; Zipfel, Gregory J. MD‡; Ogilvy, Christopher S. MD§; Pricola, Katie L. MD§; Welch, Babu G. MD¶; Shakir, Nabeel BS¶; Patel, Bhuvic BS‡; Reavey-Cantwell, John F. MD‖; Kelman, Craig R. MD‖; Albuquerque, Felipe C. MD#; Kalani, M. Yashar S. MD, PhD#; Hoh, Brian L. MD**
*Department of Neurosurgery, Bucheon St. Mary's Hospital, Catholic University of Korea, Bucheon, Republic of Korea;
‡Department of Neurosurgery, Washington University School of Medicine, St. Louis, Missouri;
§Neurovascular Service, Massachusetts General Hospital, Boston, Massachusetts;
¶Department of Radiology, UT Southwestern Medical Center, Dallas, Texas;
‖Department of Neurosurgery, Virginia Commonwealth University, Richmond, Virginia;
#Division of Neurological Surgery, Barrow Neurological Institute, St. Joseph's Hospital and Medical Center, Phoenix, Arizona;
**Department of Neurosurgery, University of Florida, Gainesville, Florida
Correspondence: Brian L. Hoh, MD, Department of Neurosurgery, University of Florida, PO Box 100265, Gainesville, FL 32610. E-mail: firstname.lastname@example.org
Received April 26, 2013
Accepted December 20, 2013
BACKGROUND: Recent experimental evidence indicates that endogenous mechanisms against cerebral vasospasm can be induced via preconditioning.
OBJECTIVE: To determine whether these vascular protective mechanisms are also present in vivo in humans with aneurysmal subarachnoid hemorrhage.
METHODS: A multicenter retrospective cohort of patients with aneurysmal subarachnoid hemorrhage was examined for ischemic preconditioning stimulus: preexisting steno-occlusive cerebrovascular disease (CVD) and/or previous cerebral infarct. Generalized estimating equation models were performed to determine the effect of the preconditioning stimulus on the primary end points of radiographic vasospasm, symptomatic vasospasm, and vasospasm-related delayed cerebral infarction and the secondary end point of discharge modified Rankin Scale score.
RESULTS: Of 1043 patients, 321 (31%) had preexisting CVD and 437 (42%) had radiographic vasospasm. Patients with preexisting CVD were less likely to develop radiographic vasospasm (odds ratio = 0.67; 95% confidence interval = 0.489-0.930; P = .02) but had no differences in other end points. In terms of the secondary end point, patients with preexisting CVD did not differ significantly from patients without preexisting CVD in mortality or unfavorable outcome in multivariate analyses, although patients with preexisting CVD were marginally more likely to die (P = .06).
CONCLUSION: This retrospective case-control study suggests that endogenous protective mechanisms against cerebral vasospasm—a preconditioning effect—may exist in humans, although these results could be the effect of atherosclerosis or some combination of preconditioning and atherosclerosis. Additional studies investigating the potential of preconditioning in aneurysmal subarachnoid hemorrhage are warranted.
ABBREVIATIONS: aSAH, aneurysmal subarachnoid hemorrhage
CI, confidence interval
CVD, cerebrovascular disease
mRS, modified Rankin Scale
OR, odds ratio
VCI, vasospasm-related delayed cerebral infarction
Preconditioning refers to the phenomenon by which exposure to a mild injurious stimulus renders an organ resistant to a subsequent more severe insult. Since its first description in the myocardium,1 there has been growing interest in preconditioning and its protective effects in the brain.2-6 Numerous experimental studies have shown that a variety of preconditioning stimuli, including hypoxia, ischemia, inhalation anesthetics, and numerous drugs (some of which are already approved by the Food and Drug Administration for other conditions), induce powerful tolerance against subsequent focal and global cerebral ischemic insults.4-6 There are also descriptions of a remote preconditioning effect, in which ischemic exposure elsewhere in the body affords strong protection against a cerebral ischemic insult.7-9 Evidence for preconditioning-induced protection in humans also exists; patients with prior transient ischemic attacks have been shown to suffer less severe ischemic strokes than those without previous transient ischemic attacks.10,11
In aneurysmal subarachnoid hemorrhage (aSAH), vasospasm-related delayed cerebral infarction (VCI) is one of the major contributors to long-term disability and death.12 Given its impact on patient outcome and the wide therapeutic window of opportunity between aSAH and vasospasm (typically 4-14 days), many have considered aSAH to be an ideal candidate for application of a preconditioning-based therapeutic strategy.6,13 Only recently, however, has experimental evidence to support this notion been provided. The first came when we demonstrated that a preconditioning stimulus (hypoxia) induces powerful protection against vasospasm and neurological deficits in an animal model of SAH and that this protection is causally linked to endothelial nitric oxide synthase-derived nitric oxide, a key player in vasospasm pathophysiology.13 The second came when another preconditioning stimulus (lipopolysaccharide) was also shown to reduce vasospasm and to improve neurological outcome in a separate animal model of SAH.14 Whether a preconditioning stimulus reduces vasospasm and improves neurological outcome in patients with aSAH, however, has yet to be determined.
In this study, our objective was to examine the effect of potential preconditioning stimulus on cerebral vasospasm, VCI, and neurological outcome in patients with aSAH. We hypothesized that preexisting steno-occlusive cerebrovascular disease (CVD) and/or previous cerebral infarct would exert a preconditioning effect that would be associated with reduced vasospasm and vasospasm-induced neurological deficits in patients with aSAH.
MATERIALS AND METHODS
Institutional review board approval was obtained from each participating center. All data were obtained and maintained with the protection of patient privacy. We analyzed medical records and radiographic imaging for all consecutive aSAH patients treated between January 2006 and December 2011 at 6 centers (University of Florida, Massachusetts General Hospital, University of Texas Southwestern Medical Center, Virginia Commonwealth University, Washington University in St. Louis, and Barrow Neurological Institute). Inclusion criteria consisted of the following: age of at least 18 years; Fisher score of 3, 4, or 3/4; aSAH confirmed by computerized tomography (CT) scan; and radiographically documented ruptured cerebral aneurysm. Patients were excluded if they were ≤17 years of age, if they had nonaneurysmal SAH, if their aneurysms were untreated, or if their care was withdrawn within 48 hours of admission.
The following data were collected for each patient: age, sex, admission Hunt-Hess grade (the grade before external ventricular drain), neurological deficit on admission, aneurysm size and location, preexisting steno-occlusive vascular disease (extracranial and intracranial atherosclerotic disease, previous infarction on admission CT scan), treatment modality (surgical clipping vs endovascular coiling), presence of radiographic vasospasm on admission, statin use, endovascular treatment for vasospasm (balloon angioplasty with or without intra-arterial drug delivery vs intra-arterial drug delivery only), and vascular risk factors, including hypertension, diabetes mellitus, and dyslipidemia.
Preexisting CVD was defined as the presence of focal or diffuse narrowing of any cerebral artery or arteries on catheter angiography or CT angiography or the presence of previous infarction located in a vascular distribution on admission CT scan: intracranial and/or extracranial carotid/vertebral circulation narrowing (minimum stenosis >30%) and/or previous infarction on admission CT scan.
Radiographic vasospasm was defined as an arterial luminal narrowing of >30% compared with the normal parent vessel diameter on conventional cerebral angiography or CT angiography or abnormal transcranial Doppler measurements in the anterior circulation (Lindegaard ratio of ≥3, absolute mean flow velocity >200 cm/s, or an increase in mean flow velocity exceeding 50 cm/s in 24 hours).15 If there was clinical suspicion or transcranial Doppler evidence for vasospasm, CT angiography or conventional cerebral angiography was performed. This occurred in all but 10 patients. Symptomatic vasospasm was defined as documented new neurological deficits attributable to vasospasm in a corresponding arterial territory that is distinct from any deficit at baseline or on admission and that is not attributable to other causes, including hydrocephalus, hemorrhage, surgical complications, metabolic abnormalities, cognitive decline, hypoxia, seizures, or infection. VCI was defined as the development of a new hypodensity on CT scan or new area of restricted diffusion on magnetic resonance imaging attributable to vasospasm in a corresponding vascular territory that was not visible on the admission or immediate postoperative scan (new infarction attributable to vasospasm). Cerebral infarctions possibly related to complications of surgery, angiography, or endovascular treatment for vasospasm such as large-vessel occlusion, perforator vessel occlusion, or arterial rupture or dissection were not considered VCI.
The primary end points were radiographic vasospasm, symptomatic vasospasm, and VCI. The secondary end point was clinical outcome at hospital discharge as measured by the modified Rankin Scale (mRS). Outcomes were defined as unfavorable (moderate, severe disability, vegetative state, or death; mRS score 3-6) or favorable (good recovery or slight disability; mRS score 0-2).
The R statistical software package (version 2.15.0) was used to calculate means, standard deviations, and frequencies for all variables. The χ2, Fisher exact, analysis of variance, Kruskal-Wallis, Mann-Whitney, and t tests were performed as appropriate to compare the preexisting CVD and no preexisting CVD groups in terms of patient characteristics and outcomes. The SAS statistical software package (version 9.3, PROC GENMOD) was used to create generalized estimating equation models to estimate the effect of the preexisting condition on the primary and secondary end points when controlling for age, sex, Hunt-Hess grade, neurological deficit on admission (yes or no), aneurysm size and location (anterior or posterior), statin use, and treatment type (clipping or coiling). In these analyses, medical center was considered a repeated factor to account for the clustering of observations on centers. A binomial distribution was assumed for the 3 primary end points (radiographic vasospasm, symptomatic vasospasm, and VCI, all yes or no), and a normal distribution was assumed for the secondary end point (mRS at hospital discharge). In all models, we assumed an exchangeable working correlation structure. Values of P < .05 were considered significant.
A total of 1043 patients met the inclusion criteria. The mean age was 55.8 ± 13.6 years; 727 patients (70%) were female. On admission, 771 patients (74%) were Hunt-Hess grade 1 to 3, and 304 patients (29%) had neurological deficit on admission. There were 492 patients (47%) with a history of hypertension, 69 patients (7%) with diabetes mellitus, and 147 patients (14%) with dyslipidemia. The ruptured aneurysm was located in the anterior circulation in 868 patients (83%) and in the posterior circulation in the remaining 175 (17%). The 3 most common sites for an aneurysm were the anterior communicating artery (36%, 372 of 1043), posterior communicating artery (20%, 213 of 1043), and middle cerebral artery (14%, 147 of 1043). There were 506 patients (49%) who underwent surgical clipping and 537 patients (52%) with endovascular coil embolization. There were 289 patients (28%) who underwent endovascular treatment for vasospasm (balloon angioplasty with or without intra-arterial drug delivery in 86 patients and intra-arterial drug delivery only in 203). There were 321 of 1043 patients (31%) with preexisting CVD: 138 patients (30%) had intracranial vascular disease, 186 patients (52%) had extracranial carotid or vertebral vascular disease, and 80 patients (22%) had previous cerebral infarction on admission CT scan. The distribution of demographic, clinical, and radiological characteristics stratified by the presence of preexisting vascular disease is shown in Table 1.
Effect of Preexisting Vascular Disease on Primary End Points (Radiographic, Symptomatic Vasospasm, and VCI)
Of 1043 patients, 437 patients (42%) had radiographic vasospasm and 606 patients (58%) did not. Of the 437 patients with radiographic vasospasm, 242 patients (23% of all patients; 55% of patients with radiographic vasospasm) had symptomatic vasospasm and 165 patients (16% of all patients; 38% of patients with radiographic vasospasm) had VCI. In univariate analysis, patients with preexisting CVD were less likely to develop radiographic vasospasm (P < .001), but there was no difference in symptomatic vasospasm or VCI (Table 2). Logistic regression analyses controlling for patient age, sex, Hunt-Hess grade, presence of neurological deficit on admission, aneurysm location, statin use, and treatment type are shown in Table 3. Patients with preexisting CVD were less likely to develop radiographic vasospasm (odds ratio [OR] = 0.67; 95% confidence interval [CI] = 0.489-0.930; P = .02), but there was no difference in symptomatic vasospasm or VCI. Age, treatment modality, presence of neurological deficit on admission, and statin use were significantly associated with radiographic vasospasm; the probability of radiographic vasospasm decreased significantly with age (OR multiplies by 0.97 for each additional year of age; 95% CI = 0.966-0.981; P < .001). Patients whose aneurysms were clipped were significantly more likely to experience radiographic vasospasm than patients whose aneurysms were coiled (OR = 1.6; 95% CI = 1.22-2.20; P < .001); patients with neurological deficit on admission were significantly more likely to experience radiographic vasospasm (OR = 1.3; 95% CI = 1.02-1.78; P = .04); and patients who received statin were significantly less likely to experience radiographic vasospasm (OR = 0.63; 95% CI = 0.491-0.803; P < .001).
Age, sex, aneurysm location, and statin use were significantly associated with symptomatic vasospasm. The probability of symptomatic vasospasm decreased significantly with age (OR multiplies by 0.97 for each additional year of age; 95% CI = 0.960-0.990; P < .001). Women were significantly more likely to experience symptomatic vasospasm than men (OR = 1.3; 95% CI = 1.17-1.52; P < .001). Patients with anterior circulation aneurysms were significantly more likely to experience symptomatic vasospasm than patients with posterior circulation aneurysms (OR = 1.4; 95% CI = 1.12-1.80; P = .003). Patients who received statin were significantly less likely to experience symptomatic vasospasm (OR = 0.56; 95% CI = 0.460-0.683; P < .001).
Age (P = .001), sex (P = .02), treatment type (P < .001), and statin use (P < .001) were significantly associated with the occurrence of VCI. The probability of VCI decreased significantly with age (OR multiplies by 0.97 for each additional year of age; 95% CI = 0.957-0.990). Women were significantly more likely to experience VCI than men (OR = 1.4; 95% CI = 1.06-1.80). Patients whose aneurysms were clipped were significantly more likely to have VCI than those whose aneurysms were coiled (OR = 1.6; 95% CI = 1.24-1.96). Patients who received statin were significantly less likely to experience VCI (OR = 0.66; 95% CI = 0.564-0.778; P < .001).
Effect of Preexisting Vascular Disease on the Secondary End Point (mRS at Discharge)
Of the 1043 patients in this study, 156 patients (15%) died (mRS score, 6) and 628 patients (60%) had unfavorable outcome. Of 628 patients with unfavorable outcome, 222 patients (222 of 321, 69%) had CVD and 406 patients (406 of 722, 56%) had no preexisting CVD. In univariate analyses, patients with preexisting CVD had higher rates of mortality (P = .006) and unfavorable outcome (P < .001; Table 2). In multivariate analyses, however, patients with preexisting CVD did not significantly differ from patients without preexisting CVD in mortality or unfavorable outcome. Multivariate analysis of patients with complete data using the binary categorization of mRS found that unfavorable outcome was associated with increasing age, worse admission Hunt-Hess grade (>3), larger size, treatment modality (clipping), and neurological deficit on admission (Table 4).
Additionally, 96 of 628 patients (15%) had unfavorable outcomes attributed to nonneurological causes. To evaluate the effects of the covariates for patients with poor outcome (mRS score > 2) for neurological vs nonneurological conditions, we also performed 2 similar analyses, 1 analysis including only patients with good outcome and those with poor outcome for nonneurological causes and another analysis including only patients with good outcomes and those with poor outcome for neurological causes. In multivariate analysis including patients with mRS scores >2 for nonneurological causes, increasing age (P < .001), clipping (P < .001), absence of statin (P = .007), and larger aneurysm (P < .001) were significantly associated with a worse mRS score. In multivariate analysis of the patients with mRS score >2 for neurological causes, increasing age (P < .001), clipping (P = .03), presence of neurological deficit on admission (P < .001), Hunt-Hess grade >3 (P < .001), and larger aneurysm (P < .001) were significantly associated with a worse mRS score.
Propensity Score-Matched Analysis (Based on the Probability of Having Preexisting CVD)
To minimize confounding effects of age and comorbidities, we also performed a propensity score-matched analysis. We used logistic regression to generate a propensity score for each patient (the probability of having preexisting disease) and used caliper matching (tolerance = 0.1) with a target match ratio of 2:1 (no disease to disease) to select subjects from the original data set. The following variables were used as covariates in the regression to create propensity scores: age, sex, Hunt-Hess score, neurological deficit on admission, aneurysm size, aneurysm location (anterior or posterior), treatment type (clip or coil), hypertension, diabetes mellitus, dyslipidemia, and statin use. The matching process reduced the total sample size from 1043 to 603. There were 353 patients with no preexisting disease and 87 patients with preexisting disease who were dropped because suitable matches could not be found (data not shown). In an analysis of the matched subjects, results were qualitatively similar to those with the full data set. Patients with preexisting CVD were less likely to develop radiographic vasospasm (P < .02), and there was no difference in symptomatic vasospasm. Patients with preexisting CVD, however, were more likely to develop VCI (P = .04); although the original analysis showed a marginal association between preexisting CVD and VCI (P = .08), this association became significant in the matched analysis. In terms of mRS score at discharge, patients with preexisting CVD did not differ significantly from patients without preexisting CVD in mortality or unfavorable outcome (Table 5).
The primary finding from our study is that patients with preexisting CVD are less likely to develop radiographic vasospasm than patients without CVD. These protective effects, however, were not associated with reduced VCI or improved neurological outcome at discharge. This result suggests that endogenous mechanisms—preconditioning—in humans may exist against vasospasm. Although we believe these results are a preconditioning effect, it could be the effect of atherosclerosis or some combination of atherosclerosis and a preconditioning effect. Preexisting steno-occlusive CVD may affect the constricting or dilating capability of the intracranial vessels, which alters the response to aSAH.16 The vessels may be less vulnerable to vasospasm and subsequent delayed cerebral ischemia because atherosclerotic intracranial vessels are more rigid. However, when vasospasm occurs, the brain in a patient with preexisting CVD has a lower tolerance to vascular narrowing, and the presence of carotid stenosis/occlusion or intracranial stenosis/occlusion may cause ischemia or stroke.
Studies have shown that a transient exposure to a mild injurious stimulus like hypoxia or ischemia renders the brain resistant to a subsequent more severe insult, a concept known as preconditioning.4-6 However, many questions remain regarding the most favorable clinical setting to identify the effect of preconditioning, the adequate preconditioning stimulus, and whether a cerebral preconditioning response can be induced safely in humans, in contrast to laboratory animals.17,18 In many clinical settings, including acute ischemic stroke, clinical application of a preconditioning-based therapeutic strategy may not be easily achievable. However, in select populations in which patients are at high risk for ischemic complications such as cerebral aneurysm surgery,19 carotid endarterectomy,18 carotid artery stenting,20 and coronary artery bypass graft surgery,21 a preconditioning-based therapeutic strategy is eminently feasible and appears to hold great promise. aSAH is another potential candidate for such a strategy given that a significant portion of patients with aSAH experience vasospasm-induced VCI many days after the initial hemorrhagic event. Although no clinical studies have examined the protective role of preconditioning on cerebral vasospasm, recent evidence from animal studies indicates that preconditioning induces strong protection against vasospasm and neurological deficits after aSAH.13,14
Many physiological stimuli, including hypoxia, ischemia, and hypothermia, and a multitude of pharmacological agents have been shown to render the brain tolerant to a broad range of injurious stimuli such as focal ischemia, global ischemia, trauma, intracerebral hemorrhage, and, most recently, SAH.3,6,13,22 The adaptive responses induced by these preconditioning stimuli include a complex series of events involving molecular sensors and transducers, transcription factors, genes, and effectors that lead, after some time, to a “latent” cerebroprotective phenotype (for a review, see the article by Gidday3). For years, neurons were thought to be the principal cellular target responsible for this protective phenotype (neuronal preconditioning), but multiple lines of evidence now show that preconditioning-induced brain protection is also mediated via the glia (glial preconditioning)2 and the vasculature (vascular preconditioning). Evidence for the latter comes primarily from ischemia studies that show preconditioning increases postischemia blood flow,23,24 improves postischemia endothelium-dependent vasodilation,25 augments postischemia vascular patency,23 reduces postischemia blood-brain barrier breakdown and vasogenic edema,26-30 and decreases endothelial cell death.31 Given the fundamental role of vascular dysfunction in SAH outcome—most prominently cerebral vasospasm but also microvascular dysfunction,32 blood-brain barrier breakdown,33 and microvascular thrombosis34—and the high incidence of neuronal cell death after SAH,35,36 we hypothesized that the multifaceted protective effects of preconditioning may be beneficial in aSAH.
An interesting observation from our study is the disconnection between the protection against radiographic vasospasm vs the lack of protection against VCI and neurological outcome in patients with preexisting CVD. Whereas the former is consistent with past experimental studies that have shown that preconditioning produces strong vasospasm protection, the latter is in direct opposition to these studies that have also reported that preconditioning produces strong protection against aSAH-induced neurological deficits. There are a variety of potential reasons for this discrepancy. First, it is likely that the degree of preconditioning induced by chronic steno-occlusive arterial disease was more heterogeneous and likely milder than that caused by the acute, uniform, and strong hypoxic insult used in the animal study. Second, all patients in our study were administered nimodipine as prophylaxis against vasospasm-induced VCI (and some received other purported antivasospasm agents, including statins), whereas nimodipine or any other antivasospasm agent was not used in the animal studies. Third, rescue therapy with hemodynamic augmentation and/or endovascular intervention was routinely used in our patients, which may have diminished the protective effect of preconditioning on VCI and neurological outcome.
Our results could be affected by age or comorbidities, and our multivariable models for these outcomes all show this expected protective effect. However, when we adjusted for age in these models, any effect of preconditioning on the outcomes observed in these models was over and above the effect of age. Hence, we believe that the effects of preconditioning reported in this study are independent of age. To make our results clear, we also performed a propensity score-matched analysis to balance the groups on all of the covariates. Results of the propensity score-matched analysis were similar to those with the full data set and supported the effect of preconditioning.
Although we believe that the results of our study have value, our study has several limitations. The most important limitation is that our observations are not definitive proof of a preconditioning mechanism, and there is the possibility that there is another causal mechanism such as the effect of atherosclerosis causing stiffening of the vessels or some combination of an atherosclerotic and preconditioning effect. For the reasons we explain above, however, our findings cannot entirely be explained by an atherosclerotic effect, and we believe they can be attributed to a preconditioning effect. These findings form the basis for further investigation into the role of preconditioning in aSAH and vasospasm. Second, this is a retrospective analysis, and selection bias could not be eliminated. Additionally, there was no standardized management protocol for aSAH patients. Thus, patients received prophylaxis and treatment for seizure and vasospasm, but with different medications and varying dosages and durations. The decision of endovascular treatment for vasospasm was made at each center and was based on the opinions of the treating physicians. Given the retrospective design of this study, the exact mechanism or causality between preexisting vascular disease and development of vasospasm cannot be determined. There was also a lack of central clinical and radiological review that could have affected determination of radiographic vasospasm and VCI. Although we tried to restrict our definition of VCI on CT scan to lesions likely caused by vasospasm, we cannot completely eliminate the possibility that some of the ischemic lesions may have been due to other causes, including procedure-related complications. Finally, we did not include several other medical comorbidities such as pulmonary disease, renal disease, liver disease, and illicit substance abuse because these variables have a negative impact on outcome in patients with aSAH. Smoking also was not included among the factors analyzed and entered into the statistical analysis. However, there have been convincing data indicating a role for smoking in the incidence and severity of vasospasm after aSAH.
This retrospective case-control study suggests that endogenous protective mechanisms against cerebral vasospasm—a preconditioning effect—may exist in vivo in humans. We believe that establishing this relationship provides an important rationale for further investigation with larger prospective studies.
Dr Zipfel has received research grants from the National Institute of Health (P50 NS055977, 1R01 NS071011-01A1), Washington University Center for Investigation of Membrane Excitability Diseases, American Heart Association, Neurosurgery Research and Education Foundation, Brain Aneurysm Foundation, and Washington University Institute of Clinical and Translational Sciences. Dr Welch serves as a consultant to Stryker Neurovascular. Dr Hoh has received research grants from the National Institutes of Health. Dr Kalani has received research support from the American Medical Association. All other authors have no personal financial disclosures in this article.
We would like to thank Dan Neal, MS, for statistical support.
1. Murry CE, Jennings RB, Reimer KA. Preconditioning with ischemia: a delay of lethal cell injury in ischemic myocardium. Circulation. 1986;74(5):1124–1136.
2. Trendelenburg G, Dirnagl U. Neuroprotective role of astrocytes in cerebral ischemia: focus on ischemic preconditioning. Glia. 2005;50(4):307–320.
3. Gidday JM. Cerebral preconditioning and ischaemic tolerance. Nat Rev Neurosci. 2006;7(6):437–448.
4. Iadecola C, Anrather J. Stroke research at a crossroad: asking the brain for directions. Nat Neurosci. 2011;14(11):1363–1368.
5. Mergenthaler P, Dirnagl U. Protective conditioning of the brain: expressway or roadblock? J Physiol. 2011;589(pt 17):4147–4155.
6. Dirnagl U, Becker K, Meisel A. Preconditioning and tolerance against cerebral ischaemia: from experimental strategies to clinical use. Lancet Neurol. 2009;8(4):398–412.
7. Hahn CD, Manlhiot C, Schmidt MR, Nielsen TT, Redington AN. Remote ischemic per-conditioning: a novel therapy for acute stroke? Stroke. 2011;42(10):2960–2962.
8. Tropak MB, Shi H, Li J, Dai X, Redington AN, Askalan R. Potent neuroprotection induced by remote preconditioning in a rat model of neonatal cerebral hypoxic-ischemic injury. J Thorac Cardiovasc Surg. 2011;142(1):233–235.
9. Jensen HA, Loukogeorgakis S, Yannopoulos F, et al.. Remote ischemic preconditioning protects the brain against injury after hypothermic circulatory arrest. Circulation. 2011;123(7):714–721.
10. Sitzer M, Foerch C, Neumann-Haefelin T, et al.. Transient ischaemic attack preceding anterior circulation infarction is independently associated with favourable outcome. J Neurol Neurosurg Psychiatry. 2004;75(4):659–660.
11. Wegener S, Gottschalk B, Jovanovic V, et al.. Transient ischemic attacks before ischemic stroke: preconditioning the human brain? A multicenter magnetic resonance imaging study. Stroke. 2004;35(3):616–621.
12. Hop JW, Rinkel GJ, Algra A, van Gijn J. Case-fatality rates and functional outcome after subarachnoid hemorrhage: a systematic review. Stroke. 1997;28(3):660–664.
13. Vellimana AK, Milner E, Azad TD, et al.. Endothelial nitric oxide synthase mediates endogenous protection against subarachnoid hemorrhage-induced cerebral vasospasm. Stroke. 2011;42(3):776–782.
14. Smithason S, Moore SK, Provencio JJ. Low-dose lipopolysaccharide injection prior to subarachnoid hemorrhage modulates delayed deterioration associated with vasospasm in subarachnoid hemorrhage. Acta Neurochir Suppl. 2013;115:253–258.
15. Ryttlefors M, Enblad P, Ronne-Engstrom E, Persson L, Ilodigwe D, Macdonald RL. Patient age and vasospasm after subarachnoid hemorrhage. Neurosurgery. 2010;67(4):911–917.
16. Inagawa T. Cerebral vasospasm in elderly patients with ruptured intracranial aneurysms. Surg Neurol. 1991;36(2):91–98.
17. Koch S, Katsnelson M, Dong C, Perez-Pinzon M. Remote ischemic limb preconditioning after subarachnoid hemorrhage: a phase Ib study of safety and feasibility. Stroke. 2011;42(5):1387–1391.
18. Walsh SR, Nouraei SA, Tang TY, Sadat U, Carpenter RH, Gaunt ME. Remote ischemic preconditioning for cerebral and cardiac protection during carotid endarterectomy: results from a pilot randomized clinical trial. Vasc Endovascular Surg. 2010;44(6):434–439.
19. Chan MT, Boet R, Ng SC, Poon WS, Gin T. Effect of ischemic preconditioning on brain tissue gases and pH during temporary cerebral artery occlusion. Acta Neurochir Suppl. 2005;95:93–96.
20. Faries PL, DeRubertis B, Trocciola S, Karwowski J, Kent KC, Chaer RA. Ischemic preconditioning during the use of the PercuSurge occlusion balloon for carotid angioplasty and stenting. Vascular. 2008;16(1):1–9.
21. Hausenloy DJ, Candilio L, Laing C, et al.. Effect of Remote Ischemic Preconditioning on Clinical Outcomes in Patients Undergoing Coronary Artery Bypass Graft Surgery (ERICCA): rationale and study design of a multi-centre randomized double-blinded controlled clinical trial. Clin Res Cardiol. 2012;101(5):339–348.
22. Steiger HJ, Hänggi D. Ischaemic preconditioning of the brain, mechanisms and applications. Acta Neurochir (Wien). 2007;149(1):1–10.
23. Dawson DA, Furuya K, Gotoh J, Nakao Y, Hallenbeck JM. Cerebrovascular hemodynamics and ischemic tolerance: lipopolysaccharide-induced resistance to focal cerebral ischemia is not due to changes in severity of the initial ischemic insult, but is associated with preservation of microvascular perfusion. J Cereb Blood Flow Metab. 1999;19(6):616–623.
24. Zhao L, Nowak TS Jr. CBF changes associated with focal ischemic preconditioning in the spontaneously hypertensive rat. J Cereb Blood Flow Metab. 2006;26(9):1128–1140.
25. Bastide M, Gelé P, Pétrault O, et al.. Delayed cerebrovascular protective effect of lipopolysaccharide in parallel to brain ischemic tolerance. J Cereb Blood Flow Metab. 2003;23(4):399–405.
26. Masada T, Hua Y, Xi G, Ennis SR, Keep RF. Attenuation of ischemic brain edema and cerebrovascular injury after ischemic preconditioning in the rat. J Cereb Blood Flow Metab. 2001;21(1):22–33.
27. Lenzsér G, Kis B, Bari F, Busija DW. Diazoxide preconditioning attenuates global cerebral ischemia-induced blood-brain barrier permeability. Brain Res. 2005;1051(1-2):72–80.
28. Stowe AM, Altay T, Freie AB, Gidday JM. Repetitive hypoxia extends endogenous neurovascular protection for stroke. Ann Neurol. 2011;69(6):975–985.
29. Wacker BK, Park TS, Gidday JM. Hypoxic preconditioning-induced cerebral ischemic tolerance: role of microvascular sphingosine kinase 2. Stroke. 2009;40(10):3342–3348.
30. Wacker BK, Freie AB, Perfater JL, Gidday JM. Junctional protein regulation by sphingosine kinase 2 contributes to blood-brain barrier protection in hypoxic preconditioning-induced cerebral ischemic tolerance. J Cereb Blood Flow Metab. 2012;32(6):1014–1023.
31. Zhang Y, Park TS, Gidday JM. Hypoxic preconditioning protects human brain endothelium from ischemic apoptosis by Akt-dependent survivin activation. Am J Physiol Heart Circ Physiol. 2007;292(6):H2573–H2581.
32. Janjua N, Mayer SA. Cerebral vasospasm after subarachnoid hemorrhage. Curr Opin Crit Care. 2003;9(2):113–119.
33. Ayer RE, Zhang JH. Oxidative stress in subarachnoid haemorrhage: significance in acute brain injury and vasospasm. Acta Neurochir Suppl. 2008;104:33–41.
34. Wang Z, Chen G, Zhu WW, Bian JY, Shen XO, Zhou D. Influence of simvastatin on microthrombosis in the brain after subarachnoid hemorrhage in rats: a preliminary study. Ann Clin Lab Sci. 2010;40(1):32–42.
35. He Z, Ostrowski RP, Sun X, et al.. CHOP silencing reduces acute brain injury in the rat model of subarachnoid hemorrhage. Stroke. 2012;43(2):484–490.
36. Hasegawa Y, Suzuki H, Sozen T, Altay O, Zhang JH. Apoptotic mechanisms for neuronal cells in early brain injury after subarachnoid hemorrhage. Acta Neurochir Suppl. 2011;110(pt 1):43–48.
This is an interesting study that attempts to provide a clinical correlate to the well-recognized laboratory finding of preconditioning or protection from severe ischemic injury by exposing animals to earlier moderate ischemia. The authors propose that their finding of reduced radiographic spasm is reflective of preconditioning. However, in this large retrospective cohort, we remain unconvinced of a preconditioning effect because such an effect should be reflected by a reduction in vasospasm-related cerebral infarction and outcome rather than merely by radiographic spasm. This may be due to the extremely liberal use of stenosis >30% as a definition for cerebrovascular disease. It is well established that hemodynamic effects do not typically occur until stenosis exceeds 80%. This is further compounded in the brain because of an effective circle of Willis in the vast majority of patients that allows them to compensate for hemodynamic reductions. It is also unclear whether there is a temporal constraint on the preconditioning effect such that a prior stroke 3 years ago has lost its ability to precondition compared with a stroke 3 weeks ago.
We would suggest that future efforts consider much more constricted selection criteria for cerebrovascular disease, in terms of both degree of stenosis and age of previous stroke, to see whether preconditioning has a bearing, particularly on vasospasm-related cerebral infarction, because eventual outcome may be reflective of initial Hunt-Hess grade and other non-spasm-related factors.
Although this is a well-written article and clearly a herculean effort collecting a large and difficult data set, we remain not quite convinced of the inferences.
Grant C Sorkin
Adnan H. Siddiqui
Buffalo, New York
The authors present a multicenter retrospective cohort study on the effects of central preconditioning on the incidence of vasospasm and an outcome analysis. The size of the cohort is impressive, and the results are intriguing. The study is limited by the usually problems of a retrospective study of sample size and selection bias; however, the authors appropriately acknowledge this. The critical result was that patients with preexisting cerebral vascular disease were less likely to develop radiographic vasospasm than patients without cerebral vascular disease. Unfortunately, this effect was not associated with a reduced incidence of infarction or improved neurological outcome at discharge. This may be due to the rather arbitrary definition of ≥30% stenosis for cerebral vascular disease.
Nevertheless, the significance of this finding should not be discounted. Coupled with the animal literature that the authors briefly review, we agree with the authors that prospective preconditioning-based strategies for the treatment of subarachnoid hemorrhage or even in elective surgeries with potential ischemic complications such as carotid endarterectomy and extracranial-intracranial bypass ought to be investigated further.
Ryan P. Morton
Louis J. Kim
The authors present a very interesting article on the role of preconditioning in the development of vasospasm and delayed cerebral ischemia in aneurysmal subarachnoid patients. Using preclinical data, they posit that previous ischemia preconditions the cerebral vasculature and brain parenchyma against further ischemic insults precipitated by large-vessel vasospasm in the aftermath of cerebral aneurysm rupture. They retrospectively review the charts of aSAH patients at several institutions, paying particular attention to patients with high Fisher grade, the presence of previous cerebrovascular disease, and the development of radiographic vasospasm, delayed ischemic neurological deficits, and permanent neurological injury resulting from vasospasm. As a secondary end point, they reviewed favorable (0-2) vs unfavorable (3-6) modified Rankin Scale scores in their patient cohort. Although the authors did not find that previous cerebrovascular disease was associated with less clinically significant vasospasm or delayed cerebral infarction, there was less radiographic vasospasm in these patients. The limitations of retrospective reviews are clear from the data, even in a study such as this carried out rigorously by the authors. The cohort with previous cerebrovascular disease was made up of much older and (most likely) sicker patients as indicated by a higher proportion of patients with a high Hunt-Hess grade in this group. Although these patients may have had true preconditioning of the brain parenchyma from previous ischemia, this may have been masked by their overall decreased physiological reserve from other comorbidities. The role of previous ischemia in preconditioning subarachnoid patients against delayed ischemic insults is a worthy research effort that requires further laboratory studies to establish the proof of principle and purported mechanisms before it can be sufficiently applied at the bedside. Development of large prospective registries will also improve the robustness of patient cohorts who can be subjected to this type of analysis. We congratulate the authors of their efforts to further elucidate preconditioning as a neuroprotective mechanism.
Kristopher T. Kimmell
G. Edward Vates
Rochester, New York
The authors present a multicenter, retrospective review of patients with aneurysmal subarachnoid hemorrhage to evaluate the effect of ischemic preconditioning on the rate of radiographic and symptomatic vasospasm in this patient population. The study is a significant undertaking, with >1000 patient records reviewed from 6 treatment centers. On the basis of their findings, the authors argue that patients with a history of preexisting cerebrovascular disease, defined by previous infarct or major intracranial vessel stenosis >30%, have a lower rate of observed radiographic vasospasm during the 2 weeks after acute aneurysm rupture. However, this relationship does not hold true for symptomatic vasospasm, vasospasm-related infarction, or outcome based on modified Rankin Scale score at the time of discharge. Although a number of animal studies have suggested that ischemic preconditioning may be an important phenomenon in potentially mitigating cerebral injury after subarachnoid hemorrhage, the results of this study are difficult to interpret. It is logical to assume that patients with radiographic vasospasm would be more likely to have vasospasm-induced infarction and poorer outcomes compared with those who do not. In fact, it is interesting to note that other variables found to be significant predictors of radiographic vasospasm (age, sex, treatment, statin use) were also significant predictors of symptomatic spasm and infarction. Why then is preexisting cerebrovascular disease a predictor of only radiographic spasm, not symptomatic ischemia or infarction? This suggests that preexisting cerebrovascular disease within the confines of this study is not truly a predictor or that potential confounders are not being accounted for. Perhaps the largest limitation of the study, which the authors allude to, is the inability to rule out atherosclerotic disease as a confounder.
This study brings up a number of important questions regarding future study of ischemic preconditioning: How should we define preexisting cerebrovascular disease in the context of ischemic preconditioning? Does vessel stenosis in the absence of a history of ischemic symptoms weigh as heavily as a previous infarct? Is radiographic spasm more common on the side ipsilateral to previous ischemic symptoms or vessel stenosis after subarachnoid hemorrhage involving cisterns on both sides? These are important questions that need to be considered going forward. The authors should be congratulated for their efforts in trying to further elucidate this complicated yet therapeutically promising phenomenon.
1. What factor increases the incidence of symptomatic vasospasm?
A. Advanced age
B. Male gender
C. Ruptured anterior circulation aneurysm
D. Statin use
E. Aneurysm treatment by surgical clipping rather than coiling
2. When compared to patients without pre-existing cardiovascular disease (CVD), patients with pre-existing CVD are more likely to have which of the following beneficial effects after aneurysmal subarachnoid hemorrhage?
A. Reduced radiographic vasospasm
B. Reduced symptomatic vasospasm
C. Reduced vasospasm-related delayed cerebral infarction (VCI)
D. Reduced mortality
E. Improved modified Rankin scores (mRS) at discharge
3. Which factor is associated with poorer neurological outcomes after aneurysmal subarachnoid hemorrhage?
A. Young age
B. Coil embolization of the aneurysm
C. Absence of neurological deficit on admission
D. Larger aneurysm size
E. Treatment with a statin
Aneurysm; Preconditioning; Subarachnoid Hemorrhage; Vasospasm
Copyright © by the Congress of Neurological Surgeons
Highlight selected keywords in the article text.