Type of Anesthesia and Differences in Clinical Outcome After Intra-Arterial Treatment for Ischemic Stroke
Background and Purpose—Intra-arterial treatment (IAT) in patients with acute ischemic stroke (AIS) can be performed with or without general anesthesia (GA). Previous studies suggested that IAT without the use of GA (non-GA) is associated with better clinical outcome. Nevertheless, no consensus exists about the anesthetic management during IAT of AIS patients. This study investigates the association between type of anesthesia and clinical outcome in a large cohort of patients with AIS treated with IAT.
Methods—All consecutive patients with AIS of the anterior circulation who received IAT between 2002 and 2013 in 16 Dutch hospitals were included in the study. Primary outcome was functional outcome on the modified Rankin Scale at discharge. Difference in primary outcome between GA and non-GA was estimated using multiple ordinal regression analysis, adjusting for age, stroke severity, occlusion of the internal carotid artery terminus, previous stroke, atrial fibrillation, and diabetes mellitus.
Results—Three hundred forty-eight patients were included in the analysis; 70 patients received GA and 278 patients did not receive GA. Non-GA was significantly associated with good clinical outcome (odds ratio 2.1, 95% confidence interval 1.02–4.31). After adjusting for prespecified prognostic factors, the point estimate remained similar; statistical significance, however, was lost (odds ratio 1.9, 95% confidence interval 0.89–4.24).
Conclusions—Our study suggests that patients with AIS of the anterior circulation undergoing IAT without GA have a higher probability of good clinical outcome compared with patients treated with general anesthesia.
Intra-arterial treatment (IAT) has been proven effective and safe for patients with acute ischemic stroke (AIS).1–3 Numerous studies have evaluated the effect of different thrombolytic agents and devices.4,5 However, less is known about the effect of anesthesia during IAT. During intervention, patients receive either general anesthesia (GA) or no GA (non-GA), referring to local anesthesia at groin puncture site with or without conscious sedation (CS). Recent retrospective studies suggest that non-GA is as feasible as GA and that GA may be associated with a lower rate of successful recanalization and worse clinical outcome.6–10 Several factors could contribute to these findings. Induction and recovery phases in GA are stressful and could lead to cardiac arrhythmias and cardiac ischemia. Furthermore, inhaled and intravenous anesthetic agents are known to alter blood carbon dioxide (CO2) and can cause blood pressure shifts that could lead to changes in cerebral autoregulation with decreased cerebral perfusion.11
Currently, no consensus exists about the optimal anesthetic management of AIS patients during IAT. Previous studies had several methodological limitations that prevent to draw definite conclusions.12 Most important was the imbalance in stroke severity at baseline in most studies, resulting in more severe strokes in the GA group as compared with the non-GA group. Furthermore, the majority of studies had small numbers of patients. In the absence of definite evidence, current practice is largely based on local protocols and preferences of the neurointerventionalists.13 Possible advantages of GA are (1) immobilization of the patient to prevent wire-induced vessel injury and to facilitate navigation with a quicker recanalization; (2) adequate ventilation and airway protection; and (3) limiting patient discomfort. On the other hand, a non-GA approach (1) may reduce time to treatment initiation; (2) allow neurological assessments during and after the procedure, (3) does not induce blood pressure lowering, and (4) does not require intubation. Nonetheless, when using a non-GA approach, there is a chance of a need to convert acutely to GA accompanied by emergency intubation, which is associated with a higher rate of aspiration pneumonia and poor outcome.14
In this retrospective study among 16 Dutch hospitals, we aimed to evaluate the relation between anesthetic management during IAT and clinical outcome. In most intervention centers in the Netherlands, a standard strategy regarding anesthetic management for acute stroke interventions is applied, thereby limiting bias through patient selection by baseline stroke severity in this study. We hypothesized that a non-GA approach during IAT in patients with AIS of the anterior circulation is associated with a better clinical outcome compared with GA based on a potentially shorter time from onset to treatment initiation, avoidance of potentially harmful blood pressure changes, and quicker recovery without the use of GA.
We conducted a retrospective cohort study in patients from the pretrial cohort of the Multicenter Randomized Clinical Trial of Endovascular Treatment for Acute Ischemic Stroke in The Netherlands (MR CLEAN), which consists of all consecutive patients with AIS treated with IAT in 16 stroke centers in The Netherlands. Information concerning procedures and treated patients was gathered to assess pretrial experience in centers that were committed to participate in the MR CLEAN trial.15 The registry started in October 2002 and continued until a center started participation in the trial. The institutional review board from the coordinating institution approved registration and use of the data. We only included patients with an anterior circulation stroke in our analysis. Patients were treated intra-arterially with a thrombolytic agent, a dedicated clot retriever or a retrievable stent. The method of IAT was left to the discretion of the treating neurointerventionalists.
All centers kept a prospective registry of patients who received IAT. Data collection itself was largely retrospective. Demographic variables, premorbid stroke risk factors, National Institutes of Health Stroke Scale (NIHSS) score at baseline, use of intravenous tissue-type plasminogen activator, timing of baseline and treatment procedures, treatment type (intra-arterial thrombolytics, mechanical treatment, or both), and type of anesthesia (GA or non-GA) were obtained from medical charts and intervention reports by trained medical researchers. When necessary and possible, NIHSS at baseline was reconstructed from clinical data with a modified algorithm.16 When missing, IAT time points were reconstructed using angiogram times: for start of IAT, time of first scan minus 5 minutes; for end of IAT, time of last scan plus 5 minutes.
Modified Rankin Scale (mRS) for functional outcome at discharge was assessed by a certified neurologist or neurology fellow.17 Good clinical outcome was defined as mRS of ≤2. Grade of recanalization was assessed with the modified Thrombolysis in Cerebral Infarction score (mTICI).18 Recanalization was defined as mTICI score 2b or 3 on Digital Subtraction Angiography imaging at the end of the procedure. Three experienced observers from a center that was not involved in the treatment assessed all Digital Subtraction Angiography runs. Observers were blinded for baseline data of the patient and for intervention center. All periprocedural and postprocedural complications, including conversion from local to GA, were recorded from intervention and imaging reports and patient records. Symptomatic intracranial hemorrhage (SICH) was defined as parenchymal hemorrhage at any site in the brain on the CT-scan, being compatible with documented neurological deterioration. Asymptomatic intracranial hemorrhage was defined as parenchymal hemorrhaged at any site of the brain found on follow-up CT-scan without neurological deterioration.
Analyses were based on the intention to treat principle. Conversions from non-GA to GA were therefore counted in the non-GA arm of the study. Descriptive statistics was expressed as means with standard deviation or medians with interquartile range (IQR). Groups (non-GA versus GA) were compared by the chi-square test for categorical variables and the Student-t test or, in case of a non-normal distribution, the Mann–Whitney U test for continuous variables. Univariable logistic analysis was performed to determine an association between type of anesthesia and good clinical outcome. Multivariable logistic regression was performed to adjust for predefined prognostic variables: age, stroke severity (NIHSS) at baseline, occlusion of the internal carotid artery terminus, history of previous stroke, atrial fibrillation, and diabetes mellitus. Additionally, we performed multivariable ordinal logistic regression analysis to assess the adjusted common odds ratio for a shift in direction of a better outcome on the mRS, adjusted for the aforementioned variables. Statistical analyses were performed using SPSS version 22.0.
We identified 369 patients with an anterior circulation stroke and available information on anesthetic management during IAT and functional outcome at discharge. Of these 369 patients, we excluded 21 patients for multiple reasons, for example, patients already under GA for other procedures, lack of information on timing of procedures, or cross over to no IAT (see online-only Data Supplement for patient flow-chart). Three hundred forty-eight patients were used for the analysis; 278 patients were treated without GA and 70 patients with GA. Information on the use of CS and specific agents were not available in most of the cases. Patients received non-GA based on standard strategy in 274 cases. In 4 cases, procedure was started without GA, despite the local standard strategy indicating GA. The majority of patients (N=63) received GA as initial treatment modality, based on the local standard strategy. Seven patients received GA because of agitation, respiratory insufficiency, or decreased level of consciousness before start of the treatment, whereas they would normally be treated without GA.
Ten patients (10/278 [4%]) in the non-GA group converted to GA during treatment. In 9 patients, reason for conversion was agitation and patient movement. One patient had respiratory insufficiency during treatment initiation. These converted cases were included in the non-GA group based on the intention to treat principle.
Patients treated under GA were significantly younger (57 years versus 62 years) and less often had atrial fibrillation (9/70 [29%] versus 40/278 [16%]). Furthermore, patients in the GA group had a longer time from onset of symptoms to start of IAT of 00:20 hours (median 04:01; interquartile range 01:53 hours versus 03:40; interquartile range 01:41) and were more frequently treated with mechanical thrombectomy only (32/70 [46%] versus 61/278 [22%]). The distribution of baseline stroke severity (NIHSS), pretreatment with intravenous tissue-type plasminogen activator, and occlusion site was similar in both groups (Table 1).
A total of 82 (82/348 [24%]) patients were functionally independent (mRS 0–2) at discharge. Good clinical outcome was seen in 26% (72/278) of patients in the non-GA group and in 14% (10/70) of patients in the GA group. A higher mortality rate was seen in the GA group (15/70 [21%]) compared with the non-GA group (46/278 [17%]); however, this difference was not statistically significant (Table 2). The distribution of the mRS in both treatment groups is presented in Figure.
In unadjusted logistic regression analysis, non-GA was significantly associated with good clinical outcome (odds ratio 2.1, 95% confidence interval 1.02–4.31). After adjusting for prespecified prognostic factors, the point estimate remained positive and, however, did not reach statistically significance (odds ratio 1.9, 95% confidence interval 0.89–4.24). The additional multivariable ordinal regression analysis showed a shift in distribution on the mRS in favor of the non-GA group (adjusted common odds ratio 1.6, 95% confidence interval 0.98–2.54). This also was not statistically significant.
Vessel perforation was seen in 4 patients (4/278 [1%]) treated without GA and did not occur in patients treated under GA. Two of these 4 patients had an accompanying SICH with an outcome of respectively 4 and 5 on the mRS at discharge. From one patient, neither SICH nor asymptomatic intracranial hemorrhage was reported and had an mRS of 3 at discharge, and one patient had an asymptomatic intracranial hemorrhage with mRS 4 at discharge. Dissection of the internal carotid artery during treatment was seen in both groups (non-GA: 12/278 [4%] versus GA: 2/70 [3%]), as well as device-related complications (non-GA: 6/278 [2%] versus GA: 3/70 [4%]). These included failure to deploy the retrievable stent, a broken guidewire, a broken stent, and a part of device unable to retrieve.
Postprocedural complications are summarized in Table 2. There was no difference in occurrence of SICH or asymptomatic intracranial hemorrhage between the 2 treatment groups. Progression of ischemic stroke and seizures was seen more often in the GA group. Pneumonia and other infections were more frequent in the non-GA group. However, these differences were not statistically significant.
mTICI scores were not available for 13 patients in the non-GA group. Of the available scores, full recanalization (mTICI 2b/3) was reached in 113/265 (43%) of patients in the non-GA group versus 34/70 (49%) in the non-GA group. All scores on the mTICI are summarized in Table 2.
Our study suggests that patients with anterior circulation AIS treated with IAT, who did not receive GA, have a higher probability of good clinical outcome compared with patients who received GA. Furthermore, we observed that IAT was initiated sooner after symptom onset in patients treated without GA as compared with GA. We did not find major differences with regard to safety parameters between the 2 treatment modalities.
Our findings are consistent with earlier findings in both terms of clinical and safety outcomes between the 2 treatment types. However, previous studies reported an imbalance in baseline NIHSS in favor of non-GA-treated patients, which could have influenced outcome. In contrast, our study had equal scores on baseline NIHSS. Hence, difference in baseline stroke severity is not the reason for improved clinical outcome after non-GA patients in our cohort.
How can we explain improved outcome in patients treated without the use of GA? First of all, it is known that inhaled or intravenous anesthetic agents can alter blood CO2 levels and blood pressure shifts, which can lead to changes in cerebral autoregulation and consequently in decrease of cerebral bloodflow, leading to extension of ischemic injury. Use of propofol and induction dosages of fentanyl predicted postinduction hypotension in a study of Reich and colleagues.19 Furthermore, some anesthetic gases might act as a vasodilatator, resulting in the reverse Robin Hood syndrome, with steal from blood flow of the affected vascular territories toward unaffected territories, further compromising flow in the ischemic area.20 There are data that support these findings in AIS patients treated with IAT. Davis et al found that lower blood pressures were associated with worse outcomes in patients undergoing CS or GA, and the mean systolic blood pressure in patients undergoing CS was 135 mm Hg compared with 104 mm Hg in patients with GA.21 Additionally, in a retrospective study of 126 patients with a middle cerebral artery stroke treated with IAT, Jumaa et al showed that final infarct volume was significantly larger in intubated patients versus nonintubated patients (mean infarct volume [cm3] 147 versus 80.2, P=0.002).7 In our study, we were unable to collect adequate information on type of anesthetic agents, blood pressure, CO2, and cerebral bloodflow during treatment nor final infarct volumes to confirm these data.
Another reason often suggested for the difference in outcome could be a higher rate of aspiration and pneumonia in intubated patients and contribution of pneumonia to poor outcome.9 However, we found a lower rate of pneumonia in the GA group, and therefore, this phenomenon cannot explain the differences in clinical outcome in our study. Conversely, lack of airway protection by the absence of intubation could lead to higher rates of pulmonary aspiration in non-GA patients. Patients with AIS of large cerebral artery may have a degree of dysphagia and are unlikely to have been fasted before intervention. The urgent need for conversion to GA may occur, accompanied by a higher risk of aspiration. Most previous studies did not examine the rate of conversion from non-GA to GA. In our study, only 10 patients in the non-GA group were converted to GA. No effect on clinical outcome was seen in these patients. The small number of converted patients in our study demonstrates that in current medical practice, the risk of conversion to GA is relatively small, thereby not clearly influencing clinical outcome in the non-GA group.
The most important factor leading to poor outcome could be that GA may lead to treatment delay resulting in a prolonged onset to recanalization time and therefore reduce the chance of good clinical outcome. However, 2 previous studies that investigated this perception found no difference in time to treatment between GA and non-GA and between intubated and nonintubated state, respectively.6,7 In our cohort, IAT in patients treated under GA was started 20 minutes later than in patients treated without GA. Because time from stroke onset to treatment is an important factor for outcome after acute stroke treatment, this may account for difference in clinical outcome. To our knowledge, this is the first study to demonstrate a difference in time to treatment between GA and non-GA. Future studies need to confirm this and should use specific time points to provide insight into the point at which most time is lost.
The main reason for neurointerventionalists to use GA is to minimize patient movement. Awake patients could be agitated during treatment, resulting in head movements that affect Digital Subtraction Angiography images. As a result, longer times to recanalization may occur. Major concern, of course, is increasing risk of procedural complications, such as vessel perforation or dissection and subsequent intracranial hemorrhage. In our group of non-GA patients, rate of vessel perforation was low and SICH was seen as often as in patients treated under GA. Other studies showed similar safety result, indicating that a non-GA approach seems to be a safe choice.
As we know from previous studies, higher recanalization leads to better clinical outcome.22 In a meta-analysis from Brinjikji et al, which included all available studies on anesthesia and IAT of AIS, a significant difference was found in recanalization grades in favor of non-GA.10 In our study, full recanalization was reached in similar percentages of patients in both treatment groups. So, we can conclude that higher recanalization may not account for better outcomes in the non-GA group in our cohort of patients.
The effect of anesthesia on clinical outcome in AIS patients remains a black box, containing several factors that could influence outcome. Our study did not answer the question which individual parameters are responsible for worse clinical outcome in patients treated under GA. Faster initiation of treatment from stroke onset could be one of the major factors in this study.
Our study does have several limitations. One of the major limitations is the retrospective and nonrandomized nature. Choice of anesthesia was based on standard local strategy or preference of the neurointerventionalist. The latter could have led to selection bias or confounding by indication or center, although a standard strategy regarding anesthetic management for acute stroke interventions is applied in most centers. Also the majority of centers and operators preferred not to use GA; therefore, group sizes were unequal. Optimal method would include randomization between GA and non-GA. Currently, the ANSTROKE (Sedation Versus General Anesthesia for Endovascular Therapy in Acute Stroke—Impact on Neurological Outcome) trial is randomizing AIS patients between GA and sedation only.23 Furthermore, mRS scores were only available at discharge. It is preferable to assess the effect of anesthesia on clinical outcome over a longer period of time.
Overall, the results of our study are in line with previous studies and show that patients who do not receive GA have a higher probability of good clinical outcome and do not have higher complication rates than patients who undergo GA. Local anesthesia, with the possible use of CS, during IAT for AIS seems a good strategy if possible.
The MR CLEAN pretrial study group.
Participating centers with local investigators in order of enrollment (N):
Department of Neurology and Radiology, Sint Antonius Hospital, Nieuwegein, the Netherlands (136), Wouter Schonewille, MD, PhD, Jan Albert Vos, MD, PhD; Department of Neurology and Radiology, Medical Center Haaglanden, the Hague, the Netherlands (108), Jelis Boiten, MD, PhD, Geert Lycklama à Nijeholt, MD, PhD; Department of Neurology and Radiology, HAGA Hospital, the Hague, the Netherlands (44) Sebastiaan de Bruijn, MD, PhD, Lukas van Dijk, MD; Department of Neurology and Radiology, Maastricht University Medical Center and Cardiovascular Research Institute Maastricht, the Netherlands (34), Robert van Oostenbrugge, MD, PhD, Wim van Zwam, MD, PhD; Department of Neurology and Radiology, Erasmus MC University Medical Center Rotterdam, the Netherlands (34), Diederik Dippel, MD, PhD, Aad van der Lugt, MD, PhD; Department of Neurology and Radiology, University Medical Center Utrecht, the Netherlands (30), Jaap Kappelle, MD, PhD, Rob Lo, MD; Department of Neurology and Radiology, Academic Medical Center Amsterdam, the Netherlands (26), Yvo Roos, MD, PhD, Charles Majoie, MD, PhD; Department of Neurology and Radiology, Sint Elisabeth Hospital, Tilburg, the Netherlands (24), Paul de Kort, MD, PhD, Willem Jan van Rooij, MD, PhD; Department of Neurology and Radiology, Rijnstate Hospital, Arnhem, the Netherlands (23), Jeannette Hofmeijer, MD, PhD, Jacques van Oostayen, MD, PhD; Department of Neurology and Neurosurgery, Radboud University Medical Center, Nijmegen, the Netherlands (15) Ewoud van Dijk, MD, PhD, Joost de Vries, MD, PhD; Department of Neurology and Radiology, Atrium Medical Center, Heerlen, the Netherlands (13), Tobien Schreuder, MD, Roel Heijboer, MD; Department of Neurology and Radiology, University Medical Center Groningen, the Netherlands (10), Patrick Vroomen, MD, PhD, Omid Eshghi, MD; Department of Neurology and Radiology, Reinier de Graaf Gasthuis, Delft, the Netherlands (8), Leo Aerden, MD, PhD, René Dallinga, MD; Department of Neurology and Radiology, Isala Klinieken, Zwolle, the Netherlands (6) Jan van den Berg, MD, PhD, Boudewijn van Hasselt, MD; Department of Neurology and Radiology, Medical Spectrum Twente, Enschede, the Netherlands (2), Heleen den Hertog, MD, PhD, Alexander Tielbeek, MD, PhD; Department of Neurology and Radiology, Leiden University Medical Center, the Netherlands (1), Marieke Wermer, MD, PhD, Marianne van Walderveen, MD, PhD.
Dr Majoie’s institution received fees for his role as a consultant for Stryker (speakers bureau/lecture fees). Dr Boiten has received honoraria for his role as a consultant for Boehringer Ingelheim. The other authors report no conflicts.
* A full list of the MR CLEAN pretrial study group is given in the Appendix.
Presented in part at the World Stroke Conference, Istanbul, October 2014.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.115.008699/-/DC1.
- Received January 29, 2015.
- Revision received March 5, 2015.
- Accepted March 9, 2015.
- © 2015 American Heart Association, Inc.
- del Zoppo GJ,
- Higashida RT,
- Furlan AJ,
- Pessin MS,
- Rowley HA,
- Gent M.
- Lee M,
- Hong KS,
- Saver JL.
- Gralla J,
- Brekenfeld C,
- Mordasini P,
- Schroth G.
- Abou-Chebl A,
- Lin R,
- Hussain MS,
- Jovin TG,
- Levy EI,
- Liebeskind DS,
- et al
- Jumaa MA,
- Zhang F,
- Ruiz-Ares G,
- Gelzinis T,
- Malik AM,
- Aleu A,
- et al
- Nichols C,
- Carrozzella J,
- Yeatts S,
- Tomsick T,
- Broderick J,
- Khatri P.
- Hassan AE,
- Chaudhry SA,
- Zacharatos H,
- Khatri R,
- Akbar U,
- Suri MF,
- et al
- Brinjikji W,
- Murad MH,
- Rabinstein AA,
- Cloft HJ,
- Lanzino G,
- Kallmes DF.
- Talke PO,
- Sharma D,
- Heyer EJ,
- Bergese SD,
- Blackham KA,
- Stevens RD.
- Fransen PS,
- Beumer D,
- Berkhemer OA,
- van den Berg LA,
- Lingsma H,
- van der Lugt A,
- et al
- Williams LS,
- Yilmaz EY,
- Lopez-Yunez AM.
- van Swieten JC,
- Koudstaal PJ,
- Visser MC,
- Schouten HJ,
- van Gijn J.
- Alexandrov AV,
- Sharma VK,
- Lao AY,
- Tsivgoulis G,
- Malkoff MD,
- Alexandrov AW.
- Rha JH,
- Saver JL.
- 23.↵Sahlgrenska University Hospital, Sweden. Sedation Versus General Anesthesia for Endovascular Therapy in Acute Stroke-Impact on Neurological Outcome (ANSTROKE). ClinicalTrials.gov. https://clinicaltrials.gov/ct2/show/NCT01872884. Accessed January 19, 2015.