Impact of the type of anesthesia on adverse events during transcarotid artery revascularization

Abstract

OBJECTIVES.

The use of local or regional anesthesia (LRA) is encouraged during transcarotid artery revascularization (TCAR) since the procedure is performed through a small incision. LRA permits neurologic evaluation during the procedure and may reduce periprocedural cardiac morbidity compared to general anesthesia (GA). There is limited and conflicting information regarding the preferred anesthesia to use during TCAR. We compared periprocedural clinical and technical complications, and intraprocedural performance metrics of TCAR performed under GA versus LRA.

METHODS.

Patient, lesion, physician, and procedural information was collected in a worldwide quality assurance program of consecutive TCAR procedures. A composite clinical adverse event rate (death, stroke, transient ischemic attack, myocardial infarction) and a composite technical adverse event rate (aborted procedure, conversion to carotid endarterectomy, bleeding, dissection, cranial-nerve injury, device failure) in the periprocedural period were computed. Four intraprocedural performance measures (flow-reversal time, fluoroscopy time, contrast volume, and skin-to-skin time) were recorded. Deidentified data were analyzed independently at the Center for Vascular Research, Univ of Maryland. Poisson regressions were used to assess the impact of anesthesia type on adverse event rates. Linear regressions were used to compare performance measures.

RESULTS.

A total of 27,043 TCARs were performed by 1,456 physicians between 2012 and 2021. A majority (83%) of patients received GA, and this proportion increased over time (R²=0.74, p<0.0001). Some physicians (33.4%) used LRA in some of their procedures; only 2.7% used LRA in all of their procedures. Clinical risk-factors were more common in the LRA group (p<0.0001), and anatomic risk-factors in the GA group (p<0.0001); these differences were adjusted for in subsequent analyses. LRA was more likely to be used by vascular surgeons and by physicians with higher prior transfemoral carotid stenting experience (p<0.0001). When comparing GA vs. LRA, clinical adverse events (1.49% [95% CI 1.3,1.8] vs 1.55% [1.2,2.0], p=0.78), technical adverse events (5.6% [5.2,6.2] vs 5.3% [4.5,6.3], p=0.47), and intraprocedural performance measures did not differ by type of anesthesia.

CONCLUSIONS.

Almost 2/3^rds of physicians performed TCAR exclusively under GA, and the overall proportion of procedures performed under GA increased over time. A larger fraction of patients with severe medical risk-factors received LRA vs. GA, while a larger fraction of patients with anatomic risk-factors received GA. Periprocedural clinical and technical adverse events did not differ by type of anesthesia. Intraprocedural performance metrics that drive procedural cost were similar between groups; potential differences in procedural cost driven by anesthetic choice require further study.

Table of Contents Summary

Anesthetic choice did not significantly impact peri-procedural adverse event rates or performance metrics in this multicenter observational study of 27,043 TCAR procedures. TCAR can be safely performed under either general or local/regional anesthesia with similar outcomes; cost considerations require further study.

Introduction

Transcarotid artery revascularization (TCAR) is a hybrid approach to treating carotid stenosis combining elements of percutaneous transfemoral carotid artery stenting (TFCAS) and open surgical carotid endarterectomy (CEA).¹ TCAR can be performed under local or regional anesthesia (LRA) since the procedure requires a small incision. In fact, LRA is encouraged since this permits neurologic evaluation of the patient during the procedure² and it is believed to reduce periprocedural cardiac morbidity compared to general anesthesia (GA).

There is limited information on outcomes of TCAR performed under GA relative to LRA. Some studies report increased cardiac complications, death, and hospital length of stay with GA^3,4 while others report no differences.⁵ An analysis of the Vascular Quality Initiative (VQI) database reported a trend towards reduction in procedural transient ischemic attacks (TIA) and myocardial infarction (MI) with LRA.⁶ An analysis of the National Surgical Quality Improvement Program (NSQIP) database found significant reductions in mortality, hospital length of stay, and duration of the procedure, as well as a numeric though nonsignificant reduction in stroke rates with LRA.⁷ More studies have compared outcomes between GA and LRA during CEA, but with equally conflicting results. CEA performed under LRA has been associated with greater periprocedural hemodynamic stability, reduced operative time, shorter hospital lengths of stay⁸ and with reduced stroke and death rates.^9,10 Another analysis concluded that GA was a risk factor for MI.¹¹ Conversely, the randomized General Anesthesia Versus Local Anesthesia (GALA) trial found no difference by anesthetic choice in the periprocedural composite stroke, TIA, death, or MI rate.¹² The MI rates were numerically twice as high with GA compared to LRA in this trial, though this comparison was underpowered. Randomized trials have reported low MI rates with TFCAS, and this has generally been attributed to the near universal use of LRA for the procedure.¹³ As a hybrid procedure, TCAR shares some risk-factors with TFCAS and CEA, though not all risk-factors associated with TFCAS or CEA apply to TCAR. We recently reported that sex and age do not affect TCAR outcomes, unlike TFCAS and possibly CEA.¹⁴ Conclusions about TCAR cannot therefore be based solely on outcomes reported with CEA or TFCAS.

The aim of this study was to test whether the composite clinical adverse event rate for TCAR performed under GA is higher than that for LRA. In secondary analyses, we compared the composite technical adverse event rates and intraprocedural performance metrics for the two types of anesthesia.

Methods

The ENROUTE Transcarotid Neuroprotection and Stent Systems (Silk Road Medical Inc., Sunnyvale, CA) received Food and Drug Administration-approval in 2015.¹⁵ The device-manufacturer developed a training and quality assurance (QA) program in partnership with the Society for Vascular Surgery.¹⁶ After training, physicians performed TCAR independently at their institutions. As part of the QA program, clinically trained specialists from Silk Road Medical collected information on every TCAR procedure performed with the ENROUTE device, independent of participation in the Vascular Quality Initiative or other FDA studies. Data was collected on physician specialty and prior TFCAS experience (<5, 5 – 24, and ≥25 lifetime cases). Patient demographics, medical history, and lesion characteristics were obtained from the medical record.¹⁷ Patients were deemed symptomatic if they had experienced a stroke, TIA, or amaurosis fugax within 180 days before the procedure. The specialists were typically on site during the procedure to record procedural details and intra-operative adverse events. Anatomic characteristics of the common carotid artery (length and depth), stent size, and balloon size followed instructions for use (IFU) criteria and were not recorded. Postoperative adverse events were reported by physicians. Stroke was defined by neurological symptoms and signs with or without imaging confirmation, and by reviewing the medical record. Myocardial infarction was defined as clinical symptoms or electrocardiogram (EKG) changes; both had to be in conjunction with elevated troponin. Asymptomatic MIs based solely on EKG changes were not recorded, and there was no distinction made between STEMI vs NSTEMI. The clinical specialists underwent annual training to ensure accurate and complete data-collection. For this analysis, anesthesia type was dichotomized into general anesthesia (GA) and a composite of local and regional anesthesia (LRA). Prior studies addressing this question have used similar definitions to categorize the type of anesthetic provided. We acknowledge that there can be nuanced differences in levels of sedation and analgesia that can be provided within the two categories of anesthesia. The cohort comprised TCAR procedures performed worldwide between November 7, 2012, and April 21, 2021. Procedures with missing data on age, sex, symptomatic status, type of anesthesia, prior physician TFCAS experience, or adverse events, were excluded. Data collected through the QA program was stored and maintained by Silk Road Medical. The collected data was de-identified and sent to the University of Maryland for independent analysis. Deidentified data were analyzed independently at the Center for Vascular Research, University of Maryland School of Medicine (UMSOM). The study was reviewed and approved by the UMSOM Institutional Review Board and granted waiver of consent.

The primary outcome was the composite clinical adverse event rate (death, stroke, TIA, or MI) within 24 hours of the procedure. A secondary outcome was the technical composite adverse event rate (aborted procedure, conversion to carotid endarterectomy, bleeding necessitating reoperation or blood transfusion, arterial dissection, cranial-nerve injury, or device failure). The other secondary outcome was procedural performance metrics including skin-to-skin time (minutes from skin incision to skin closure), flow-reversal time (minutes from activation to deactivation of flow-reversal device), fluoroscopy time (minutes), and contrast medium volume (milliliters).

Statistical Analysis

Demographics, clinical and anatomic risk-factors, and physician characteristics were summarized as frequencies and percentages; differences by anesthesia group were assessed using Pearson χ² and Fisher’s exact tests, as appropriate. The number and proportion of cases performed each year were plotted in bar graphs stratified by anesthesia type. Linear regression was used to quantify year-to-year changes in anesthetic choice. To illustrate anesthetic choice and its relationship with TCAR experience, two plots were created: first, a histogram of the number and proportion of physicians using GA exclusively, LRA exclusively, and a mixture of GA and LRA; second, a bubble plot of procedures using LRA as a percentage of the total number of TCAR procedures performed at various levels of TCAR experience, quantified by the serial case number (serial case number 1 denotes each physician’s first TCAR procedure, 2 denotes each physician’s second procedure, etc.). The relationship between anesthetic used and prior TFCAS experience was depicted in another histogram.

Differences in clinical and technical outcomes by anesthetic type were assessed using multivariable Poisson models; multivariable linear models were used for proficiency metrics. Generalized estimating equations of Liang and Zeiger with an exchangeable correlation structure were used to account for the serial autocorrelation of procedures performed by the same physician; the unit of repeated measures was the physician.¹⁸ Models were adjusted for age, sex, symptomatic status, prior TFCAS experience, clinical risk, and anatomic risk. Clinical and anatomic risk were considered binary variables. Clinical risk was present if the patient had a history of angina, congestive heart failure (CHF) New York Heart Association (NYHA) class III or IV, permanent contralateral cranial nerve injury, chronic obstructive pulmonary disease (COPD), uncontrollable diabetes, need for open heart or other major surgery at the time of TCAR procedure, chronic renal insufficiency (serum creatinine ≥2.5 mg/dL), abnormal stress test, and history of MI within 6 weeks prior to TCAR. Anatomic risk included bilateral stenosis requiring treatment, restenosis after CEA, contralateral occlusion, high target lesion (above second cervical vertebra), hostile neck anatomy (cervical radiation, neck surgery, or cervical spine immobility), laryngeal palsy or laryngectomy, tandem stenoses ≥70%, and coronary artery disease in ≥2 vessels coupled with angina. Interactions between the independent variable (anesthesia type), and three of the dependent variables (symptomatic status, clinical risk, and anatomic risk) for each of our outcomes were not statistically significant in any of the models. Therefore, results from the models are presented without interaction terms. Results from the multivariable models are presented as adjusted adverse event rates (events per 100 procedures) and adjusted means, accompanied by 95% Confidence Intervals (CI). Data-points for our secondary outcomes that were ≥10 times the grand median (median of all data for each metric) were considered outliers and therefore excluded. A two-tailed p value <0.05 was considered significant. Analyses were conducted using SAS v9.4 (SAS Institute Inc., Cary, NC).

Results

Of the 30,621 TCAR procedures performed during the study period, 3,578 (12%) were excluded; 129 for treatment on arteries other than the carotid, and others missing details of the procedure (n=71), risk-factors (n=584), and physician information (n=2,794, Figure 1). The analytic cohort comprised 27,043 procedures performed by 1,456 physicians at 702 sites. Most procedures (83%) were performed under GA; only 17% were performed using LRA (Table I). The GA group was younger (p<0.0001), and had more females (p=0.02), uncontrolled diabetes (p=0.0002), prior CEA (p<0.0001), and spinal immobility (p<0.0001). The LRA group had more class III-IV CHF (p=0.01), COPD (p<0.0001), recent MI (p=0.002), and need for major cardiac and non-cardiac surgery (p<0.0001, p=0.04). Collectively, clinical risk-factors were more common in the LRA group (p<0.0001), and anatomic risk-factors were more common in the GA group (p<0.0001). LRA was used more frequently by vascular surgeons (p<0.0001) than other specialties and by physicians with more prior TFCAS experience (p<0.0001, Supplemental Table I). Most of the differences between groups were numerically small.

Figure 1.

Flow diagram showing cohort creation.

Table I.

Demographic & Clinical Characteristics

Count (Column %)		General Anesthesia N=22442 (83%)	Local/Regional Anesthesia N=4601 (17%)	p-value
Demographics
Female Sex		7,975 (35.5)	1,549 (33.7)	0.02
Age Group	≤64 years	3,772 (16.8)	715 (15.5)	0.04
	65–69 years	3,632 (16.2)	676 (14.7)	0.01
	70–74 years	4,496 (20.0)	895 (19.5)	0.37
	75–79 years	5,073 (22.6)	1,113 (24.2)	0.02
	≥80 years	5,469 (24.4)	1,202 (26.1)	0.01
Clinical Characteristics
Neurologic symptoms ≤180 days before procedure		8,585 (38.3)	1,731 (37.6)	0.42
Any severe medical comorbidity		4,285 (19.1)	1,044 (22.7)	<0.0001
History of angina		395 (1.8)	82 (1.8)	0.92
Congestive heart failure NYHA-class III or IV		679 (3.0)	172 (3.7)	0.01
Contralateral cranial nerve injury		42 (0.2)	4 (0.1)	0.13
Chronic obstructive pulmonary disease		1,559 (6.9)	428 (9.3)	<0.0001
Uncontrollable diabetes		836 (3.7)	128 (2.8)	0.002
Needs open heart surgery		173 (0.8)	71 (1.5)	<0.0001
Needs major surgery		349 (1.6)	91 (2.0)	0.04
Chronic renal insufficiency (creatinine ≥2.5 mg/dL)		317 (1.4)	55 (1.2)	0.25
Recent myocardial infarction		111 (0.5)	40 (0.9)	0.002
Anatomic Risk Factors
Any Anatomic risk factor		14,452 (64.4)	2,747 (59.7)	<0.0001
Bilateral stenosis requiring treatment		2,281 (10.2)	425 (9.2)	0.06
Post-CEA restenosis		3,028 (13.5)	492 (10.7)	<0.0001
Contralateral occlusion		1,442 (6.4)	273 (5.9)	0.21
Lesion located above second cervical vertebra		6,565 (29.3)	1,359 (29.5)	0.70
Hostile neck anatomy		2,459 (11.0)	496 (10.8)	0.73
Laryngeal nerve palsy or laryngectomy		69 (0.3)	14 (0.3)	0.97
Spinal immobility		1,094 (4.9)	156 (3.4)	<0.0001
Tandem Stenosis >70%		338 (1.5)	60 (1.3)	0.30

The number and proportion of procedures under GA increased over time (R²=0.74, p<0.0001, Figure 2). Most physicians (63.9%) used GA in all their procedures, 33.4% used GA in some, and only 2.7% used LRA in all their procedures (Figure 3A). Among the 486 physicians who used both anesthesia types, physicians first used LRA at a median of their 4^th case (interquartile range: [2, 9]). The proportion of cases using LRA showed no consistent overall trend and LRA remained below 30% throughout (Figure 3B). Physicians with more prior TFCAS experience (≥25 cases) incorporated LRA at a higher rate than those with less experience (Figure 4).

Figure 2. Distribution of choice of anesthesia for transcarotid artery revascularization over time.

(A) Number of TCAR procedures performed each year using general vs. local or regional anesthesia.

(B) The proportion of TCAR procedures performed each year using general anesthesia increased overtime while the proportion using local or regional anesthesia decreased each year.

Figure 3. Physicians’ preferred anesthetic for transcarotid artery revascularization.

(A) Most physicians (63.9%) used general anesthesia in all their procedures, while only 2.7% used local or regional anesthesia exclusively; 33.4% used general anesthesia sometimes but not in every procedure.

(B) Proportion of procedures performed using local or regional anesthesia as physician gains TCAR experience. Serial case number 1 denotes each physician’s first TCAR procedure, 2 denotes each physician’s second procedure, etc.

Figure 4. Impact of prior transfemoral carotid artery stenting experience on anesthetic choice for transcarotid artery revascularization.

(A) Number of physicians and percent out of all 1,456 physicians.

(B) Proportion of physicians using general anesthesia vs local/regional anesthesia, within each level of TFCAS experience.

Unadjusted adverse event rates were low in both anesthesia groups (Supplemental Table II). After adjusting for age, sex, symptomatic status, composite clinical risk-factors, composite anatomic risk-factors, and physician TFCAS experience, clinical adverse events occurred at a rate of 1.5% (95% CI [1.3, 1.8]) in the GA group, and 1.5% (95% CI [1.2, 2.0]) in the LRA group (p=0.78, Table II). Technical adverse events occurred at 5.6% (95% CI [5.2, 6.2]) with GA and 5.3% (95% CI [4.5, 6.3]) with LRA (p=0.47). There were no significant differences in the four procedural metrics by anesthesia type either (Table II). As a sensitivity analysis, we also performed a propensity score matched analysis. The resulting comparison between 18,224 procedures under GA and 4,601 procedures under LRA showed no difference in composite clinical event rates (1.4 [1.2, 1.8] vs. 1.6 [1.2, 2.1], p=0.57) or in composite technical event rates (5.7 [5.0, 6.4] vs. 5.4 [4.5, 6.4], p=0.50).

Table II.

Adjusted patient outcomes and procedural metrics of patients undergoing transcarotid revascularization under general versus local/regional anesthesia Rates reported as number of events per 100 procedures.

All rates and means were adjusted for patient age, sex, symptomatic status, clinical and anatomic risk factors, and prior physician experience with transfemoral carotid artery stenting.

Outcome	General Anesthesia	Local/Regional Anesthesia	p-value
Patient adverse event rates	Adjusted Rate (95% Confidence Interval)
Composite Clinical	1.5 (1.3, 1.8)	1.5 (1.2, 2.0)	0.78
Composite Technical	5.6 (5.2, 6.2)	5.3 (4.5, 6.3)	0.47
Procedural Metrics	Adjusted Mean (95% Confidence Interval)
Contrast Volume (mL)	29.5 (28.9, 30.2)	30.2 (29.1, 31.3)	0.18
Fluoroscopy Time (min)	5.5 (5.4, 5.6)	5.5 (5.3, 5.7)	0.76
Flow Reversal Time (min)	12.9 (12.7, 13.2)	13.2 (12.8, 13.6)	0.17
Skin-to-Skin Time (min)	74.0 (72.9, 75.1)	74.0 (72.5, 75.5)	0.98

Discussion

In patients undergoing TCAR, periprocedural composite clinical and composite technical adverse events were not affected by the type of anesthesia used (GA versus LRA, 1.5% vs 1.5% clinical [p=0.78] and 5.6% vs 5.3% technical [p=0.47]). Procedural metrics (contrast volume, fluoroscopy time, flow-reversal time, and skin-to-skin time) also did not differ by the type of anesthesia used. A larger fraction of patients with severe medical risk-factors received LRA rather than GA, while a larger fraction of patients with anatomic risk-factors received GA. Almost 2/3^rds of physicians performed TCAR exclusively under GA, and the proportion of procedures performed under GA increased over time. The use of LRA fluctuated as physicians gained TCAR experience but remained consistently lower than GA (<30%). LRA was used more often by physicians with high prior TFCAS experience (≥25 procedures).

Several considerations go into the choice of anesthesia for TCAR. 1) Safety (periprocedural stroke, death, and MI), 2) Patient factors that may increase adverse events (symptomatic status or co-existing clinical or anatomic risk-factors), 3) Patient preferences, 4) physician or facility familiarity with anesthetic options, and finally 5) cost considerations.

Differences in stroke and death rates between GA and LRA have been studied largely in the context of CEA with inconclusive results. A systematic review and an analysis of the NSQIP database showed reduced rates of stroke and death with LRA,^7,11 though another systematic review and an analysis of the VQI database did not find such differences.¹⁹ Two randomized trials failed to identify differences in composite stroke, death, and MI rates.^20,21 Comparisons of MI rates between GA and LRA have also shown mixed results. Analyses from NSQIP found higher MI rates with GA²² and that GA predicted periprocedural MI¹¹ while another retrospective study found no difference in MI rates between GA and LRA.²³ A subsequent randomized trial reported higher intraoperative hypertensive events with LRA, and hypotensive events with GA.²⁰ While MI rates were twice as high with GA, they were not significantly different. Similar results were reported from an analysis of the VQI database.¹⁹ While the randomized GALA trial did not find a difference in MI rates, the study was underpowered and confounded by clinical risk-factors that were not accounted for.²¹ In most randomized trials and registry analyses, TFCAS has been associated with low MI rates.^13,24 This has been attributed to the use of LRA for most TFCAS procedures.⁴ In general though, carotid revascularization regardless of approach, has become increasingly safe with extremely low complication rates. In most instances, statistical differences observed in stroke, death, or MI between the two anesthetic approaches were likely a function of the large sample sizes offered by multicenter registries. The numeric differences, when identified in such studies, have been relatively small and might not be clinically important.

We found no difference in clinical or technical adverse events between TCAR performed under GA versus LRA, even after adjusting for age, sex, symptomatic status, clinical risk factors, anatomic risk factors, and prior experience of operators with TFCAS (Table II). Unadjusted event rates of individual components of the composite outcomes were also numerically similar between the two groups (Supplemental Table II). Non-significant findings such as those we found, do not prove that there is no association between exposure and outcome, since one cannot prove the null. However, our study reviewed the largest database of TCAR performed worldwide, and the number of outcomes make it unlikely that we missed a true effect, i.e., the probability of a type-1 error, a false negative finding, is small. Compared to previous analyses of data contributed to registries or trials by select voluntarily participating clinical centers, this is an analysis of all 27,043 consecutively performed TCAR procedures in the country (with 385 events) between November 7, 2012, and April 21, 2021. This analysis has adequate power to detect clinically important differences; 91.7% power to find a difference of 0.7 composite clinical adverse events/100 procedures and 83.4% power to identify a difference of 0.6 events/100 procedures.

General anesthesia reduces global cerebral metabolism, leading some to suggest that symptomatic patients may be better served with GA to prevent additional hypoperfusion injury during revascularization.²⁵ TCAR operators, however, do not seem to believe this; there was no difference in the proportion of symptomatic patients undergoing TCAR under GA versus LRA (38.3% vs 37.6%, p=0.42, Table I). One study has reported higher periprocedural stroke with TFCAS performed under GA in symptomatic patients.⁴ We also found no significant interaction between symptomatic status and anesthesia type for clinical or technical outcomes or for procedural performance metrics. Therefore, TCAR achieves similar outcomes with the two anesthetic approaches regardless of symptomatic status. This is consistent with observations that cerebrovascular flow-reversal in TCAR is well-tolerated, and hypoperfusion events are not increased in TCAR compared to CEA or TFCAS.

More patients with severe medical risk-factors received LRA compared to GA (22.7% vs 19.1%, p<0.0001, Table I). Physicians likely assumed that LRA reduces periprocedural cardiac and pulmonary complications, though we found no interaction between clinical risk-factors and outcomes by anesthetic choice. We did not systematically track pulmonary complications, and therefore cannot determine whether LRA is better for patients with chronic pulmonary disease. More patients with anatomic risk-factors received GA compared to LRA (64.4% vs 59.7%, p<0.0001, Table I). Physicians likely anticipated more difficult access in such patients, though we did not find an interaction between anatomic risk-factors, outcomes, and anesthesia. Therefore, physicians were able to negotiate these anatomic challenges under either mode of anesthesia. Of note, procedures were performed per device IFUs, and patients with contraindicated anatomy (e.g., short or deep common carotid arteries) or difficult lesions (e.g., concentric calcification) were excluded. Therefore, our findings are not applicable to procedures performed off-IFU.

Based on a presumed reduction in periprocedural MI with LRA, physicians have been encouraged to transition from GA to LRA after becoming comfortable with TCAR. Physicians have not followed this recommendation. The fraction of procedures performed under LRA each year has decreased over time (p<0.0001, Figure 2B). Approximately 2/3^rds of physicians never transitioned to LR, 1/3^rd incorporated LRA in some of cases, and only 3% used LRA in every procedure (Figure 3). This was not anticipated at the inception of this QA database. Information was therefore not collected on physician reasoning for their anesthetic choice, which may have been multifactorial. First, physicians may have been reluctant to change the anesthesia they were most familiar with. Physicians with more prior experience with TFCAS incorporated LRA at a higher rate than those with less experience (73.2% vs 69.3%, p<0.0001, Figure 4). Second, physicians may have perceived that patients were more comfortable with GA due to the proximity of instrumentation near their face. Future studies will need to collect information on physician and patient preferences to objectively address this question.

In addition to being predictors of morbidity, procedural metrics (contrast volume, fluoroscopy time, flow-reversal time, and skin-to-skin time) are important drivers of procedural cost, and our analyses found that these procedural metrics did not differ by type of anesthesia. General anesthesia is associated with additional expenses (longer use of the operating room, cost of anesthetics, and anesthesiologist time). While cost data were not collected, it is reasonable to infer that the preference for GA by most physicians has cost implications for TCAR.²⁶

Limitations

Adverse events were infrequent, preventing assessments of individual components of the composite outcomes. The analysis of technical adverse events was appropriately focused on periprocedural events. In addition, logistical challenges of obtaining longer-term events for every procedure performed throughout the world precluded collection of these data. Health Insurance Portability and Accountability Act limitations prevented recording of exact age, and patients <65 years old were grouped together, as were those >80 years. However, randomized trials indicate the pivotal age for age-related differences in outcomes is 70–80 years and these patients formed a separate group in our analysis.^27–30 Being a relatively new procedure, it is possible that our results on anesthetic choice may have been confounded by an increasing number of new physician entrants to the procedure (first 3 bubbles in Figure 3B) and warrants a re-analysis of anesthetic choice several years later. The potential impact of anesthetic modality on cognition in this older population was not evaluated. A meta-analysis of surgical procedures suggests that GA may increase post-operative cognitive dysfunction³¹. Lastly, perceptions of physicians and patients for an awake procedure, degree of patient discomfort, change from LRA to GA within a patient, or use of vasoactive drugs was not recorded or evaluated. A prospective study will be required to address the cost implications of this distribution of anesthesia use.

Conclusions

The type of anesthesia used during TCAR does not affect periprocedural composite clinical adverse events (stroke, death, and MI) or composite technical adverse events. The impact of symptomatic status, clinical risk-factors or anatomic risk-factors on outcomes is not affected by anesthesia type either. Physician familiarity with TFCAS using LRA increases the use of LRA for TCAR. Procedural metrics that drive cost are similar between the two approaches, though potential differences in total operating room time and anesthesiology costs require further study.

Article Highlights.

Type of Research:

Multi-center retrospective observational study

Key Findings:

Among 27,043 patients who underwent transcarotid artery revascularization, the type of anesthesia used did not affect risk of peri-procedural clinical or technical complications. Procedural metrics that drive cost were similar between the two approaches, though potential differences in total operating room time and anesthesiology costs require further study.

Take home Message:

TCAR is safe and can be performed under either general or local/regional anesthesia with similar outcomes.