Background and purpose Many studies have suggested a relationship between the type of anesthesia provided during intra-arterial therapy for acute ischemic stroke and patient outcomes. Variability in blood pressure and hypotension have previously been identified as possible reasons for worse outcomes in acute stroke. Our aim was to investigate hemodynamic parameters and neurological outcomes of patients receiving either general anesthesia or conscious sedation for intra-arterial therapy of acute stroke.
Methods We performed a retrospective review of patients undergoing intra-arterial therapy from December 2008 to March 2015. Demographic data, baseline National Institutes of Health Stroke Scale score, preoperative physiological variables, procedural details, systolic blood pressure (SBP), diastolic blood pressure (DBP), mean arterial pressure (MAP), heart rate, and modified Rankin Scale scores were recorded.
Results 99 patients were included in the study, with 38 receiving general anesthesia and 61 receiving conscious sedation. Patients who received general anesthesia had a lower maximum SBP (p=0.02), minimum SBP (p<0.0001), minimum DBP (p<0.0001), and minimum MAP (p<0.0001). On multivariate analysis, general anesthesia was associated with lower minimum SBP (p=0.04), DBP (p=0.02), and MAP (p=0.007). Conscious sedation was associated with more favorable neurological outcomes (p=0.02). Patients with favorable neurological outcomes had a lower maximum variability in SBP (p=0.01) and MAP (p=0.03), as well as a higher minimum DBP (p=0.03).
Conclusions Patients with acute ischemic stroke undergoing intra-arterial therapy with general anesthesia had lower minimum SBP, DBP, and MAP, greater fluctuations in blood pressure, and less favorable outcomes. More studies are needed to examine the implications of variable and reduced blood pressures and neurological outcomes.
- Blood Pressure
Statistics from Altmetric.com
The treatment of patients with acute ischemic stroke requires timely recanalization to prevent or slow neurological damage.1 Over the last decade, intra-arterial recanalization for acute ischemic stroke has supplemented or served as an alternative to intravenous tissue plasminogen activator (IV-tPA). Many factors affect the outcome of the patient undergoing intra-arterial therapy for acute stroke, including the type of anesthesia provided during the procedure.
The choice of anesthesia for patients undergoing intra-arterial therapy has been debated.2–4 Currently, moderate conscious sedation (CS) and general anesthesia (GA) are the two most commonly used anesthesia techniques for patients undergoing mechanical thrombectomy for acute ischemic stroke.4 Several recent retrospective studies, meta-analyses, and reviews have shown a strong association between GA and worse neurological outcomes and death compared with CS.5–11 The physiological or other factors that lead to these worse outcomes with GA compared with sedation are not completely understood. Previous investigators have postulated that variations in procedural blood pressures may exist due to the type of anesthetic management used.8 Notably, large variations in blood pressure have previously been linked to worse outcomes in acute stroke.12 ,13 In this current study we evaluated patients undergoing intra-arterial treatment for acute ischemic stroke, comparing their hemodynamic characteristics and neurological outcomes under GA and CS.
Institutional Review Board approval was obtained for a retrospective review of the records of patients with acute ischemic stroke and the requirement for individual patient consent was waived. Inclusion criteria for this study were patients with acute ischemic stroke treated with intra-arterial therapy from December 2008 to March 2015. One patient who converted from CS to GA and one patient receiving mechanical recanalization for stroke occurring during other neuroendovascular procedures were excluded. Strokes from venous thrombosis were similarly excluded. A total of five patients were excluded, leaving a total of 99 study patients.
For each patient, demographic variables and comorbidities including hypertension, diabetes, atrial fibrillation, coronary artery disease, obesity, and smoking were recorded. Smoking was recorded as a comorbidity for patients who had a history of smoking but may have quit at the time of the procedure. Obesity was defined as body mass index (BMI) 30–40 kg/m2 and morbid obesity as BMI >40 kg/m2. When available, the baseline National Institutes of Health Stroke Scale (NIHSS) score was recorded, generally from the emergency department and/or consulting neurologist. The location of the stroke was collected from the preoperative CT scan or from the mechanical thrombectomy notes. A T occlusion was defined as a lesion involved the internal carotid artery (ICA) terminus extending into both the A1 and M1. An L occlusion was defined as an occlusion that involved the ICA terminus extending into the M1.
Preoperative data collected included minimum oxygen saturation, minimum systolic blood pressure (SBP), minimum diastolic blood pressure (DBP), and blood glucose level. The use and time to use of IV-tPA were recorded when applicable.
Choice of anesthesia and procedure characteristics
Patients were divided into two groups based on type of anesthesia received during the procedure (GA or CS/local anesthesia). The choice of type of anesthesia was made by consensus among the neurologist, anesthesiologist, and the neurointerventionalist on a case-by-case basis. The primary factor contributing to the use of GA was the patient's ability to lie still during the procedure. All patients who arrived in the interventional neuroradiology suite with a pre-existing endotracheal tube in place received a general anesthetic.
Procedures were performed in a dedicated biplane neuroangiography suite by three experienced neurointerventionalists. Under sterile conditions, the femoral artery was punctured and accessed. Selective angiography and intra-arterial therapy were provided. A variety of thrombectomy devices and stents were used in the cerebral vasculature and, at the neurointerventionalist's discretion, combined with intra-arterial tPA. Anticoagulation was maintained during the procedures with heparin boluses. Control angiograms were obtained to monitor progression of the therapy and at the end of the procedure to assess reperfusion status. Blood pressure was measured for most of the patients with an oscillometric non-invasive blood pressure cuff generally at 2–5 min intervals. In seven patients an arterial line was used for blood pressure measurements at 1–2 min intervals due to the cuff size being too small or providing inaccurate measurements.
We studied the following intraoperative hemodynamic parameters: maximum and minimum measurements for SBP, DBP, mean arterial pressure (MAP), and heart rate (HR). These parameters were recorded during the time interval of the neurointerventionalist's procedure, as identified from the anesthesia records. In addition, we calculated the maximum variability in SBP, DBP, MAP, and HR. Variability was calculated as the maximum value minus the minimum value for each parameter (eg, maximal minus minimal SBP during the intervention). Additionally, the minimum SBP and DBP were collected during the induction phase of GA or during the initiation of CS. The difference between the preoperative minimum SBP and DBP and the minimum SBP and DBP during induction or initiation was also recorded.
The type of anesthesia, anesthetic drugs, and use of vasopressors (ephedrine, phenylephrine, epinephrine) or antihypertensives (esmolol, labetalol, isoproterenol) during the endovascular procedure were collected. Vasoactive use prior to the procedure was not available. The time to groin puncture and duration of intervention were recorded when available. Complications during the procedure and post-treatment intracerebral hemorrhage were noted.
Neurological outcome was determined using the modified Rankin Scale (mRS), with mRS ≤2 defined as a favorable outcome and mRS ≥3 defined as an unfavorable outcome. The mRS score was determined from the patient's follow-up visit or telephone call approximately 90 days following the stroke or the last available follow-up if 90-day follow-up was not available.
Two subgroup analyses of neurological outcomes were performed. A subgroup analysis was done to compare neurological outcomes among patients with only anterior circulation strokes. Another subgroup analysis was done to compare neurological outcomes among GA patients with pre-existing endotracheal tubes versus those intubated in the interventional suite.
Data were entered into a Microsoft Excel spreadsheet (Microsoft Corporation, Redmond, Washington, USA). Statistical analysis was performed with JMP V.10 (SAS Institute, Cary, North Carolina, USA). Categorical results are stated as relative frequencies and continuous results are stated as median and IQR or mean and SD. Patient demographics, preoperative characteristics, hemodynamics, timing of therapy, neurological outcome, and complications were compared between the GA and CS groups using the Wilcoxon rank-sum test for continuous variables and the Pearson χ2 test for categorical variables. We also performed a multivariate standard least squares analysis with effect emphasis in order to determine if anesthesia type was independently associated with the abovementioned hemodynamic outcomes. We adjusted for baseline variables which had a p value <0.10 when comparing the CS and GA groups (ie, minimum preoperative SBP, minimum preoperative DBP, age, diabetes mellitus, morbid obesity, and location of vessel occlusion) as well as baseline NIHSS score. In addition, we compared blood pressure measurements between patients with favorable and unfavorable neurological outcomes using the Wilcoxon rank-sum test.
A total of 99 patients with acute ischemic stroke who underwent intra-arterial therapy were included in the study, with 61 (62%) patients undergoing CS and 38 (38%) patients undergoing GA. In cases prior to 2012, 62% of patients received GA compared with 23% in cases including and following 2012. The distribution of age, gender, comorbidities, and number of patients receiving IV-tPA prior to treatment was similar between the two groups. However, the difference in age (p=0.07), diabetes mellitus (p=0.06), and morbid obesity (p=0.06) trended towards statistical significance. Patients in the GA group were more likely to have posterior circulation occlusions than those in the CS group (34% vs 2%, p<0.0001). There was no difference in the number of patients with left-sided strokes between the two groups (p=0.21). The patient characteristics are summarized in table 1.
Baseline NIHSS score and preoperative physiological parameters are summarized in table 2. Data for all variables were not available in the records for each study patient. The number of patients included in the analysis for each variable is indicated. The baseline NIHSS score of patients managed with CS was not significantly different from the NIHSS score of those managed with GA (median 18 vs 16; p=0.67). Additionally, the pre-intervention minimum oxygen saturation (p=0.54), minimum DBP (p=0.55), and blood glucose levels (p=0.17) were not significantly different. The difference in minimum SBP between the two groups trended towards statistical significance (p=0.09).
The drugs used for anesthesia are summarized in table 3. The anesthetic drug used was not available in 19 patients in the CS group. In the group receiving GA, 14 patients had pre-existing endotracheal tubes placed in the emergency room or in previous procedures, and one patient arrived at the hospital intubated.
Hemodynamics during induction and initiation of anesthesia
The hemodynamic parameters during the induction phase of GA or initiation of CS are summarized in table 4. Data for 47 patients in the CS group and 22 in the GA group were available. Patients managed with GA had a lower minimum SBP than patients managed with CS (115±23 mm Hg vs 139±23 mm Hg; p=0.0003). The difference in minimum DBP between the two study groups was not significant (p=0.14). During this phase of anesthesia, 32% of patients in the GA group had a fall in SBP of >10 mm Hg compared with 11% in the CS group (p=0.03). Additionally, 32% of patients in the GA group had a fall in DBP of >10 mm Hg compared with 4% in the CS group (p=0.002).
Hemodynamics during intra-arterial procedure
The hemodynamic parameters during the intra-arterial procedure are summarized in table 4. In patients managed with GA, the maximum SBP (154±24 mm Hg) and minimum SBP (94±16 mm Hg) were lower than the maximum SBP (165±25 mm Hg; p=0.02) and minimum SBP (119±26 mm Hg; p<0.0001) recorded for patients in the CS group. Patients receiving GA recorded a lower minimum DBP (50±12 mm Hg vs 63±14 mm Hg; p<0.0001) and minimum MAP (62±11 mm Hg vs 79±14 mm Hg; p<0.0001) than those receiving CS. The study group differences in maximum DBP (GA: 91±20 mm Hg; CS: 98±17 mm Hg; p=0.07) and maximum MAP (GA: 105±18 mm Hg; CS: 111±17 mm Hg; p=0.08) trended towards statistical significance.
The variability in SBP (GA: 60±20 mm Hg; CS: 46±26 mm Hg; p=0.002) and MAP (GA: 43±19 mm Hg; CS: 32±16 mm Hg; p=0.003) was larger in study patients receiving GA than in those under CS. The variability of DBP (GA: 41±19 mm Hg; CS: 32±18 mm Hg; p=0.08) trended towards significance. There was no statistically significant difference in the maximum (p=0.80) and minimum HR (p=0.10) between the two groups, although the variability in HR trended towards significance (p=0.08).
Use of vasoactive drugs during intra-arterial procedure
Study patients under GA were more likely to receive vasopressors (33/37, 89%) than patients managed with CS (34/55, 62%; p=0.004). In contrast, more patients in the CS group than in the GA group received antihypertensives (30/55 (55%) vs 7/37 (19%); p=0.0006).
Timing of therapy
There was no difference in the time to IV-tPA administration between the CS and GA patient groups (122±43 min vs 114±34 min; p=0.67). Additionally, both the time of onset to the start of the procedure (256±119 min vs 282±126 min; p=0.23) and the procedure duration (82±40 min vs 87±30 min; p=0.22) were similar between the two groups.
On multivariate analysis adjusting for age, baseline NIHSS, occlusion location, morbid obesity, diabetes mellitus, and minimum SBP and DBP at baseline, the use of GA was associated with a lower minimum SBP (p=0.04), DBP (p=0.02), and MAP (p=0.007). There was a significant difference in MAP variability (p=0.008). These data are summarized in table 5.
Complications during procedure
In the CS group there were no cases of iatrogenic dissection or vessel perforation, no cases of groin hematoma requiring transfusion or intervention, and no cases of pseudoaneurysm. There were two device failures in the CS group (1 stent fracture and 1 burst balloon). In the GA group there was one case of vessel perforation resulting in extravasation of contrast and one case of epistaxis which was controlled with local pressure.
Neurological outcomes and postoperative complications
Data for all variables were not available in the records for each study patient and were not controlled for demographic variables or clot location. The number of patients with a favorable neurological outcome (mRS ≤2) at 90 days was higher in the CS group than in the GA group (22/57 (39%) vs 6/37 (16%); p=0.02). In addition, 82% (47/57) patients in the CS group were alive at 90 days post-stroke compared with 70% (26/37) of patients in the GA group, but this difference was not statistically significant (p=0.17). Post-treatment imaging was available for 97 of the 99 patients. There were 17 cases of intracerebral hemorrhage in the CS group and 13 cases in the GA group (28% vs 35%; p=0.48).
We compared hemodynamic parameters in patients with favorable and unfavorable neurological outcomes.
Hemodynamic parameters and neurological outcome
Hemodynamic parameters and neurological outcomes are summarized in table 6. Patients with a good neurological outcome had a lower maximum variability in SBP (42±20 mm Hg vs 55±29 mm Hg; p=0.01) and MAP (31±13 mm Hg vs 39±20 mm Hg; p=0.03). Patients with good neurological outcomes had higher minimum DBP (64±15 mm Hg vs 56±14 mm Hg; p=0.03). There was a trend towards a lower minimum SBP in patients with a poor neurological outcome (p=0.09) and a lower minimum MAP (p=0.06). All other hemodynamic characteristics were similar between the groups.
A subgroup analysis was done to compare neurological outcomes among patients with only anterior circulation strokes. The difference in favorable neurological outcome (mRS ≤2) at 90 days between the CS and GA groups with only anterior circulation strokes trended towards statistical significance (22/57 (39%) patients in the CS group vs 5/24 (21%) in the GA group; p=0.10). In addition, 82% (46/56) of patients in the CS group were alive at 90 days post-stroke compared with 80% (20/25) of patients in the GA group (p=0.82). There was no significant difference in the rate of intracerebral hemorrhage between the two groups with exclusion of posterior circulation strokes (p=0.52). Additionally, there was a significant difference in the number of patients with anterior circulation strokes with a favorable neurological outcome compared with the number of patients with posterior circulation strokes (p=0.04).
Another subgroup analysis was done to compare neurological outcomes among patients in the GA group with pre-existing endotracheal tubes versus those who were intubated in the interventional suite. There was no significant difference in the rate of intracerebral hemorrhage (p=0.51), 90-day survival (p=0.17), or rate of favorable neurological outcome (p=0.69). Additionally, when excluding this subgroup of patients with pre-existing endotracheal tubes, there was no significant difference between the CS and GA groups in the rate of intracerebral hemorrhage (p=0.34) or 90-day survival (p=0.66). The difference in favorable neurological outcome (mRS ≤2) at 90 days trended towards statistical significance (p=0.08).
In this study we found that GA is associated with greater variations in SBP, DBP, and MAP than CS. The maximum and minimum SBP, DBP, and MAP are also lower in patients managed with GA. The differences in minimum SBP, DBP, and MAP seen between the GA and CS groups persisted on our multivariate analysis adjusting for factors such as baseline NIHSS score, baseline SBP and DBP, age, occlusion location, presence of morbid obesity, and diabetes mellitus. GA was also associated with more unfavorable outcomes (mRS ≥3). We also demonstrated that patients with poor neurological outcomes had greater variations in SBP and MAP than those with good neurological outcomes and had lower DBP. These findings are important as they suggest a possible mechanism for the poorer outcomes seen in patients undergoing mechanical thrombectomy receiving GA compared with those receiving CS.
A number of recently published studies and meta-analyses have demonstrated superior outcomes with CS compared with GA for patients with acute ischemic stroke undergoing mechanical thrombectomy. A multicenter study of 980 patients treated with intra-arterial therapy from 2005 to 2009 showed that the use of GA was associated with a poorer neurological outcome at 90 days (OR 2.33, 95% CI 1.67 to 3.25).5 A single-center study of 96 patients found that mortality was more likely in patients who received GA than in those managed with local anesthesia (relative risk 2.3; 95% CI 1.1 to 3.7).8 A recent meta-analysis of nine studies with a total of 1956 patients demonstrated that patients receiving GA had higher odds of death (OR 2.59, 95% CI 1.87 to 3.58) and respiratory complications (OR 2.09; 95% CI 1.36 to 3.23) and lower odds of good functional outcome (OR 0.43; 95% CI 0.35 to 0.53).7 More recently, in a subgroup analysis of the MR CLEAN trial, van den Berg et al examined 348 patients and found that patients managed without GA had significantly higher rates of good clinical outcome (OR 2.1, 95% Cl 1.02 to 4.31); however, after adjusting for various factors, the statistical significance was lost (OR 1.9, 95% CI 0.89 to 4.24).11
A number of mechanisms have been proposed to explain the worse outcomes seen with GA. High rates of hypotension and hemodynamic lability, especially during the induction phase of GA, may overwhelm the brain's ability to autoregulate cerebral blood flow.3 Takahashi et al14 found that blood pressure decreased significantly after induction of GA and remained lower during the procedure. Whalin et al15 associated GA with greater hypotension during induction and greater variations in blood pressure. Davis et al8 showed that a SBP below 140 mm Hg was a predictor of poor neurological outcome. During the acute phase of ischemic stroke, blood pressure reductions could be associated with brain injury and worse neurological outcomes.16 Our study adds to the growing body of evidence that GA is associated with lower blood pressures, increased fluctuations, and worse neurological outcomes. Autoregulation of cerebral blood flow only occurs within a specific and limited range. In the setting of stroke, a decrease in blood pressure below this range and a corresponding fall in cerebral perfusion pressure may worsen the extent of injury to the ischemic penumbra by compromising collateral flow, leading to complete infarction.17–19 Our findings support the argument that this mechanism could explain the worse outcomes consistently observed in patients managed with GA in this study and many recent studies.
Limitations of the study
An obvious limitation of this study is the retrospective and non-randomized nature. Comorbidities, NIHSS score, and other variables were different between the groups; however, even on our adjusted analysis, these differences persisted. The small sample size means that there are few patients in each group, thus limiting our power to detect differences between groups. Fifteen patients were also already intubated upon arrival at the neurological suite, reducing the number of patients where anesthesia type was chosen. In addition, some data were missing from our database, further reducing the number of cases available for inclusion in the analysis of the various variables. The site of vascular occlusion may also have played a role in the bias of the hemodynamic data. Basilar artery and posterior circulation strokes often have a worse prognosis and baseline respiratory compromise; a statistically greater portion of these patients were in the GA group. This study did not focus on the outcomes after intervention, but rather on comparing the intraoperative hemodynamic parameters between the two types of anesthesia. We wanted to draw attention to the greater decrease and variations in blood pressure seen with GA and its possible role in explaining poorer outcomes among patients managed with GA in our study and many previous studies. Other mechanisms could also help to explain this association, including the inability to monitor for clinical changes when the patient is fully anesthetized, excessive cerebral vasodilation and alterations in vasomotor reactivity with anesthetic gases, greater hypocarbia (and therefore potential cerebral vasoconstriction) in patients undergoing GA, and deleterious consequences from mechanical ventilation in patients undergoing GA who do not always get extubated immediately after the intervention. Our results could not be controlled for these other factors.
Like many centers, our rate of improved neurological outcome has improved in recent years. Additionally, the percentage of cases performed at our institution under GA has decreased since 2012. While the decreasing number of GA cases may play a role in the better neurological outcomes, our study cannot control for outside factors such as the use of newer stent retrievers and aspiration catheters and the improvement in periprocedural care available over the last few years.
Our study found that patients managed with GA had lower SBP, DBP, and MAP as well as greater fluctuations in blood pressure than those managed with CS during intra-arterial therapy for acute ischemic stroke. We also found that patients receiving GA had a worse neurological outcome after treatment, and that patients with worse outcomes have greater variation in MAP and SBP with lower minimum DBP. Our results suggest a hemodynamic explanation for the relationship between GA and poorer outcomes after acute endovascular stroke therapy. Prospective studies are needed to examine the relationship between blood pressure and neurological outcomes to eventually guide our clinical decision-making.
Contributors MJ contributed to the acquisition of the data. MJ and WB contributed to the analysis and interpretation of the data and drafting of the manuscript. WB, AAR, and DFK contributed to the conception and design of the project. WB, AAR, JJP, and DFK contributed to the critical revising of the manuscript and final approval of the version to be published.
Competing interests DFK: consultancy (ev3,* Medtronic,* Codman*); grants/grants pending (ev3,* MicroVention,* Sequent,* Codman*); payment for lectures, including service on speakers bureaus (MicroVention*); royalties (UVA Patent Foundation*); payment for development of educational presentations (ev3*); travel/accommodations/meeting expenses unrelated to activities listed (MicroVention*). Asterisks indicate money paid to the institution.
Ethics approval Ethics approval was obtained from the Mayo Clinic Institutional Review Board.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Any unpublished data related to this manuscript will remain with the corresponding author. They cannot be made available without additional IRB approval.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.