Improving mechanical thrombectomy time metrics in the angiography suite: Stroke cart, parallel workflows, and conscious sedation

Abstract

Purpose

Earlier reperfusion of large-vessel occlusion (LVO) stroke improves functional outcomes. We hypothesize that use of a stroke cart in the angiography suite, containing all commonly used procedural equipment in a mechanical thrombectomy, combined with parallel staff workflows, and use of conscious sedation when possible, improve mechanical thrombectomy time metrics.

Methods

We identified 47 consecutive LVO patients who underwent mechanical thrombectomy at our center, retrospectively and prospectively from implementation of these three workflow changes (19 pre- and 28 post-). For each patient, last known normal, NIHSS, angiography suite in-room time, type of anesthesia, groin puncture time, on-clot time, recanalization time, LVO location, number of passes, device(s) used, mTICI score, and outcome (mRS) were recorded. Between-group comparisons of time metrics and multivariate regression were performed.

Results

Stroke cart, parallel workflows, and primary use of conscious sedation decreased in-room time to groin puncture (−21.3 min, p < 0.0001), in-room to on-clot time (−24.1 min, p = 0.001), and in-room to reperfusion time (−29.5 min, p = 0.01). In a multivariate analysis, endotracheal intubation and general anesthesia were found to significantly increase in-room to on-clot time (p = 0.01), in-room to reperfusion time (p = 0.01), and groin puncture to on-clot time (p = 0.05). The number of patients achieving a good outcome (mRS 0−2), however, did not significantly differ between the two groups (9/18 (47%) vs 14/28 (50%), p = 0.60).

Conclusions

Use of a stroke cart, parallel workflows by neurointerventionalists, technologists, and nursing staff, and use of conscious sedation may be useful to other institutions in efforts to improve procedural times.

Introduction

Mechanical thrombectomy is now the standard of care for patients presenting with large-vessel occlusion (LVO) acute ischemic stroke (AIS).^1–5 Timely reperfusion is critical as earlier reperfusion improves functional outcomes,^6–8 with every 30-minute delay reducing the absolute rate of a good outcome by 11%. ⁹ The rapid, efficient, and team-oriented approaches used in the clinical trials have prompted many institutions to analyze their mechanical thrombectomy workflows. Several authors have highlighted the importance of process improvement to maximize clinical benefit of mechanical thrombectomy.^10–12 A growing number of strategies within the angiography suite to hasten groin puncture and decrease the time to which a device makes first contact with a thrombus have emerged: use of conscious sedation (CS) instead of general anesthesia (GA), as in the majority of mechanical thrombectomy cases in the 2015 clinical trials,^1–5 preparation of devices and room prior to patient arrival, such as use of a pre-arranged stroke tray in the angiography suite, or brisk recanalization ischemic stroke kit (BRISK),^10,11,13 standardization of thrombectomy techniques,^14,15 parallel workflows by neurointerventional staff, ¹¹ and the use of additional personnel and/or cross-training of x-ray and computed tomography (CT) technicians for after-hours cases. ¹³

In an effort to decrease time to groin puncture, improve procedural efficiency, and facilitate early reperfusion, we implemented three simple concurrent changes to our pre-procedural workflow within the angiography suite. One is the use of a stroke cart for use in LVO interventions. The stroke cart, modeled after the American Heart Association Advanced Cardiac Life Support (ACLS) crash cart, is a self-contained mobile unit that contains two copies of virtually all the equipment required to perform endovascular treatment of LVO (Figure 1). The stroke cart is located inside the angiography suite during a mechanical thrombectomy procedure, near the patient and operators. The second workflow change was the implementation of parallel workflows for the neurointerventionalists, technologists, and nursing staff for angiography suite and equipment preparation. The third workflow change was a shift away from use of GA toward CS whenever possible. We hypothesize that use of stroke cart, parallel workflows, and use of CS (when possible) improves efficiency in endovascular treatment of LVO, leads to earlier groin puncture and reperfusion times, and may improve patient outcomes.

Figure 1.

Stroke cart. Contents include all devices commonly used for mechanical thrombectomy: short and long vascular sheaths measuring 5 to 10 F, a variety of microcatheters, guidewires and microwires, exchange wires, Tuohy-Borsts, one-way stopcocks, 1 to 60 ml syringes, 7 to 9 F balloon and 5 to 8 F non-balloon-guided catheters, inner catheters, stent-retrievers of varying sizes, aspiration catheters, aspiration tubing, and an aspiration system.

Materials and methods

After institutional review board approval, we identified 50 consecutive patients with LVOAIS who underwent mechanical thrombectomy at our institution. The inclusion criteria for this study were: AIS symptoms with National Institutes of Health Stroke Scale (NIHSS) > 6, noncontrast CT Alberta Stroke Program Early CT Score (ASPECT) > 5, CT angiography (CTA) confirmation of an LVO, followed by mechanical thrombectomy starting within six hours from last seen normal. One case in each group was excluded because of spontaneous recanalization seen on angiography. In addition, we excluded one case in the post-workflow change implementation group because of a biplane angiography equipment failure/shutdown during the procedure. This resulted in a final sample size of 47 patients. We retrospectively collected data on 19 consecutive patients who underwent mechanical thrombectomy prior to implementation of the workflow changes (stroke cart, use of parallel workflows, and shift toward use of CS) in April 2015, and prospectively collected the same data for 28 consecutive patients subsequent to implementation of the changes.

The neurointerventional section at our institution houses three separate but adjacent biplane angiography suites that share storage cabinets located outside the biplane rooms. Before the implementation of the stroke cart, the catheters, wires, and devices required for a mechanical thrombectomy procedure were located on different carts in different rooms. As a result, during preparation for a mechanical thrombectomy, and, occasionally, during the mechanical thrombectomy procedure itself, staff would shuttle from cart to cart, and room to room, to locate and retrieve the necessary equipment. The goal of the stroke cart is improve workflow and efficiency in the angiography suite by housing only a limited number of individual units of essentially all the necessary devices for a mechanical thrombectomy procedure, including: sheaths, microcatheters, guidewires, Tuohy-Borsts, stopcocks, 2–60 ml syringes, guide catheters, inner catheters, stent-retrievers, aspiration catheters, aspiration tubing, and an aspiration system (Figure 1, Table 1). The stroke cart (Datel, Holland, MI) measures 78 × 36 × 24 inches (height, width, depth), including caster wheels. Concurrent with the introduction of the stroke cart in April 2015, we also implemented use of parallel workflows for personnel in the angiography suite for faster room and equipment preparation.

Table 1.

Stroke cart contents.

Sheaths (one each)	Inner/Diagnostic catheters	Microwires (one each)
5 F 11 cm	5 F Berenstein II/UCSF II	Synchro2 standard
6 F 11 cm	5 F Terumo GlideCath	Synchro2 soft
6 F 23 cm	5 F Simmons 2	Transcend EX Platinum Wire
7 F 23 cm		Fathom 016
8 F 11 cm	Stent retrievers (two each)
8 F 23 cm	Trevo XP (varying sizes)	Guidewires (one each)
9 F 11 cm	Solitaire 2 (varying sizes)	Bentson
9 F 23 cm	MindFrame	Terumo Glide
10 F 11 cm		Terumo torque device
	Aspiration supplies (one each)	Storq (exchange length)
Guide catheters (one each)	Penumbra 088 Max STR	Terumo Glide (exchange length)
9 F Merci Balloon Guide	Penumbra 088 Max MP
8 F Flowgate Balloon Guide	Penumbra 070 STR	Additional supplies
AXS Infinity	Penumbra 070 MP	60 cc syringe (×2)
6 F MPD	Penumbra ACE68	1 cc yellow Medallion syringe (×2)
6F MPD DA	Penumbra ACE64	Tuohy Borst (×3)
6F MPD XB	Penumbra 5MaxACE	One-way stopcock (×4)
6F MPD XB DA	Penumbra 5Max	Three-way stopcock (×2)
6F STR	Penumbra 4Max	K50 extension tubing
6F STR DA	Penumbra 3Max	Small blue bowl (×1)
6F STR XB	Aspiration pump	30 cc syringe (×1)
6F STR XB DA	Tubing set
66 MPC
6F MPC DA	Microcatheters (one each)
6F MPC XB	Penumbra Velocity
6F MDC XB DA	Prowler Plus Straight
	Prowler Plus 45

For each mechanical thrombectomy, the treatment team in the angiography suite consists, at a minimum, of one or two interventional neuroradiologist(s) (attending and/or fellow), one technologist, one radiology nurse, one emergency department (ED) nurse, and two anesthesiologists (attending and resident/fellow). In the parallel workflow approach, preparation of different aspects of the procedure occurs simultaneously instead of in series. The patient is transferred to the angiography table by the anesthesiologist, technologist, and nursing staff. During sterile preparation of the groin and draping by the technologist, a neurointerventionalist (attending and/or fellow) prepares devices for the procedure on the sterile back table. Once the nurse has obtained heparinized saline bags, the sterile technologist or neurointerventionalist assists nursing staff with connection and clearing of heparinized saline drip tubing and connection to pressurized saline bags. Finally, if more than one neurointerventionalist is present (attending or fellow), one is tasked with obtaining groin access, while additional devices continue to be prepared by the other neurointerventionalist. The sterile technologist may also assist with device preparation, especially if a second neurointerventionalist is not present. During daytime business hours, any additional available neurointerventional staff (neurointerventionalists, technologists, nurses) also may assist with the above tasks. This standardized parallel workflow approach was used in each patient in the post-workflow change group.

Lastly, a policy to avoid GA and use CS whenever possible was also implemented in April 2015. In the angiography suite, the type of anesthesia is at the discretion of the attending neurointerventionalist. CS or GA was administered by Department of Anesthesia staff in all patients. Some patients treated under GA in both groups were intubated in the ED for airway protection and some had to be intubated in the angiography suite (e.g., because of an inability to follow commands). For this reason, we also identified a subgroup of patients intubated in the angiography suite (ETTIR).

For each patient, we recorded patient age, time of stroke onset, presentation NIHSS score, in-room arrival time, LVO location, groin puncture time, on-clot time (defined as the time the catheter was positioned at the clot), recanalization time, post-thrombectomy modified Thrombolysis in Cerebral Infarction (mTICI) score, number of passes required to achieve final mTICI score, thrombectomy device(s) used, and type of anesthesia (CS vs GA), whether an endotracheal tube placement in angiography suite (ETTIR) versus in the ED, and placement of a radial arterial line versus arterial pressure transduction from the femoral sheath. These data were readily available in all cases from the patient electronic medical record, including time stamps on angiographic images, angiography technologist log books, procedure reports, and anesthesia records. We calculated in-room time to groin puncture (IRGP), in-room time to on-clot time (IROCT), in-room time to recanalization time (IRRT), groin puncture time to on-clot time (GPOCT), and groin puncture to recanalization time (GPRT). We also determined the final outcome for each patient, defining a good outcome as 90-day modified Rankin Scale (mRS) score of 0–2. If 90-day follow-up was not available, we defined a good outcome as either 30-day mRS 0–2, or 24-hour NIHSS score decrease of greater than eight points.

An a priori power analysis was performed to determine the sample size needed for this study. Based on a Type I error rate of 0.05, a Type II error rate of 0.20, a moderate effect size of 25% reduction in time metrics (based on IRGP means and standard deviation obtained in a preliminary analysis), and three independent variables, the sample size was calculated to be 47 patients. One-tailed Student's t-tests were used to determine if continuous variables significantly differed between the two groups. Two-sample test of proportions was used to determine if binary variables were significantly different between groups. For variables with more than two categories, chi-square tests were used to determine significant differences between the two groups. To determine if any of the workflow changes were independent predictors of treatment time metrics, multiple regressions for each outcome variable (IRGP, IROCT, IRRT, GPOCT, and GPRT) were performed. This resulted in five different models, each using predictor variables relevant to the outcome variable of interest. IRGP was modeled using predictor variables: stroke cart+parallel workflow, ETTIR, arterial line placement by anesthesia team versus femoral sheath transduction, and anesthesia type (monitored anesthesia care (MAC)/CS vs GA). IROCT was modeled using predictor variables: stroke cart+parallel workflow, ETTIR, arterial line placement by anesthesia team versus femoral sheath transduction, anesthesia type (MAC/CS vs GA) and LVO location. IRRT was modeled using predictor variables: stroke cart+parallel workflow, ETTIR, arterial line placement by anesthesia team versus femoral sheath transduction, anesthesia type (MAC/CS vs GA) and number of passes before recanalization. GPOCT was modeled using predictor variables: stroke cart+parallel workflow, ETTIR, arterial line placement by anesthesia team versus femoral sheath transduction, anesthesia type (MAC/CS vs GA) and LVO location. GPRT was modeled using predictor variables: stroke cart+parallel workflow, ETTIR, arterial line placement by anesthesia team versus femoral sheath transduction, anesthesia type (MAC/CS vs GA), and number of passes before recanalization. LVO location was not included in the IRRT and GPRT to avoid an over-stratified analysis for our sample size. For categorical variables showing significance in the full model, linear contrasts were used to determine significant differences between levels of the categorical variable. An alpha of 0.05 was considered significant. All statistical analyses were conducted using STATA (version 13.1, Stata Corp, College Station, TX, USA). No correction was made for multiple corrections as the comparisons were planned before statistical analysis.

Results

A total of 47 consecutive patients with LVOAIS who underwent mechanical thrombectomy were included in this study, 19 prior to implementation of workflow changes (stroke cart, parallel workflows, and shift toward CS), and 28 after. Characteristics of patients are summarized in Table 2. Comparison between pre- and post-workflow change groups revealed no significant differences in patient age, gender, time from last seen normal to in-room time, presenting NIHSS score, LVO location, number of passes, or mTICI 2b or 3 reperfusion outcomes (Table 2). Significant differences were present, however, in the types of thrombectomy devices used (exclusively stent-retrievers in the pre-workflow change group vs combination of stent-retrievers and Solumbra in the post-workflow change group), and type of anesthesia (Table 2).

Table 2.

Characteristics of mechanical thrombectomy patients.

	Pre-workflow changes	Post-workflow changes	p value
Age (mean, SD)	67.2 years ( ± 12.2)	73.3 years ( ± 13.5)	0.1205
Male gender (%)	10 (52.6%)	13 (46.4%)	0.6765
LVO location (%)	ICA: T–3	ICA: T–6	0.4861
	MCA: M1–9	MCA: M1–13
	M2–3	M2–4
	Basilar: 3	Basilar: 3
	Tandem: 0	Tandem: 2
Devices	Stent-retriever: 18	Stent-retriever: 11	0.0060 b
	ADAPT: 0	ADAPT: 1
	Solumbra: 0	Solumbra: 16
Average number of passes	2.4	2.2	0.6680
TICI score 2B or 3	15/19 (79%)	22/28 (79%)	0.9753
CS	1/19 (5%)	18/28 (64%)	0.0001 a
ETT placement in IR suite c	13/19 (68%)	9/28 (32%)	0.0181 a
LSN to in-room time (minutes, average)	224	246	0.31
Presenting NIHSS score (mean, SD)	18.3 ( ± 7.4)	19.4 ( ± 5.4)	0.32

LVO: large-vessel occlusion; ICA: internal carotid artery; MCA: middle cerebral artery; ADAPT: A Direct Aspiration First Pass Technique; TICI: Thrombolysis in Cerebral Infarction; CS: conscious sedation; ETT: endotracheal tube; IR: interventional radiology; LSN: last seen normal; NIHSS: National Institutes of Health Stroke Scale.

^aP value indicates a significant difference in proportions between groups.

^bP value indicates a significant difference between groups for proportion of cases using stent-retrievers and Solumbra combined.

^cSome patients in both groups who were treated under general anesthesia had ETT placement in the emergency department.

Improvements in several time metrics were seen after implementation of the workflow changes (Figures 2 and 3), with statistically significant decreases in IRGP, 43.0 ± 8.0 min versus 21.7 ± 4.4 min (−21.3 min, p < 0.0001), IROCT, 84.8 ± 11.7 min versus 60.7 ± 8.0 min, (−24.1 min, p = 0.001), and IRRT, 123.0 ± 14.6 versus 93.5 ± 16.7 min (−29.5 min, p = 0.01). GPOCT (p = 0.30) and GPRT (p = 0.16) were not significantly different between groups.

Figure 2.

Effect of workflow changes (stroke cart/parallel workflow/primary use of conscious sedation) on mechanical thrombectomy time metrics. Statistically significant improvements were seen in IRGP, 43.0 ± 8.0 min versus 21.7 ± 4.4 min (−21.3 min, p < 0.0001), IROCT, 84.8 ± 11.7 min versus 60.7 ± 8.0 min, (−24.1 min, p = 0.001), and IRRT, 123.0 ± 14.6 versus 93.5 ± 16.7 min (−29.5 min, p = 0.01), were seen. GPOCT (p = 0.30) and GPRT (p = 0.16) were not significantly different between groups. IRGP: in-room time to groin puncture; IROCT: in-room time to on-clot time; IRRT: in-room time to recanalization time; GPOCT: groin puncture time to on-clot time; GPRT: groin puncture time to recanalization time.

Figure 3.

Effect of workflow changes (stroke cart/parallel workflow/primary use of conscious sedation) on mechanical thrombectomy time metrics. Stacked bar plot demonstrating significant effect of workflow changes on IRGP time. No significant effect was seen on GPOCT or OCTRT. IRGP: In-room time to groin puncture; IROCT: in-room time to on-clot time; IRRT: in-room time to recanalization time; GPOCT: groin puncture time to on-clot time; GPRT: groin puncture time to recanalization time; OCTRT: on-clot time to recanalization time.

Multivariate regression analysis of time metrics for pre- and post-workflow changes was performed and included stroke cart presence and parallel workflows (as a combined variable), ETTIR, arterial line placement in IR suite, type of anesthesia, LVO location, and number of passes, as described in the Methods section. Stroke cart utilization and parallel processing was found to be an independent predictor of improvements in IRGP (coefficient, −15.81, 95% confidence interval (CI) −30.22 to −1.40, p = 0.033), IROCT (coefficient, −32.08, 95% CI −54.65 to −9.50, p = 0.007), and IRRT (coefficient: −37.50, 95% CI −62.10 to −12.89, p = 0.004). In addition, trends to decreasing GPOCT and GPRT were noted with stroke cart and parallel workflow (coefficient: −16.27, 95% CI −35.55 to 3.01, p = 0.095 and coefficient: −19.82, 95% CI −42.76 to 3.12, p = 0.087, respectively) although these did not reach statistical significance (Table 3). ETTIR and the type of anesthesia used also had an effect on time metrics. ETTIR was found to an independent predictor of increases in IROCT and IRRT (coefficient: 44.01, 95% CI 10.20–77.83, p = 0.013 and coefficient: −50.22, 95% CI 13.65−86.78, p = 0.009, respectively). Similarly, GA was an independent predictor of increased GPOCT (coefficient: 36.97, 95% CI 0.69–73.26, p = 0.046) and showed near significance for increasing IROCT (coefficient: 42.00, 95% CI −0.49 to 84.5, p = 0.052). While controlling for all other factors, LVO of the basilar artery was found to have significantly higher IROCT and GPOCT compared to other LVO locations (p < 0.05) (Table 3). Lastly, increasing number of passes was a predictor of GPRT and IRRT (p < 0.05) (Table 3). Arterial line placement in IR was not found to be an independent predictor of any of the time metrics.

Table 3.

Multivariate regression results.

Variable	Coefficient	p value	95% CI
In-room time to groin puncture time
Stroke cart+parallel workflow	−15.81	0.033*	−30.22 to −1.40
ETTIR	19.12	0.080	−2.47 to 40.71
Arterial line by anesthesiologist	5.44	0.389	−7.32 to 18.19
General anesthesia	5.03	0.706	−22.10 to 32.15
In-room to on-clot time
Stroke cart+parallel workflow	−32.08	0.007*	−54.65 to −9.50
ETTIR	44.01	0.013*	10.20 to 77.83
Arterial line by anesthesiologist	3.88	0.693	−16.10 to 23.86
Anesthesia	42.00	0.052	−0.49 to 84.5
Basilar > L ICA	9.28	0.722
Basilar > L ICA T	62.21	0.009*
Basilar > L M1	42.79	0.029*
Basilar > L M2	65.44	0.004*
Basilar > L tandem	30.67	0.294
Basilar > R ICA T	48.80	0.017*
Basilar > R M1	49.21	0.029*
Basilar > R M2	45.23	0.165
Basilar > R tandem	22.70	0.456
In-room time to recanalization time
Stroke cart+parallel workflow	−37.50	0.004*	−62.10 to −12.89
ETTIR	50.22	0.009*	13.65 to 86.78
Arterial line by anesthesiologist	4.12	0.724	−19.76 to 28.00
General anesthesia	42.04	0.071	−3.93 to 88.00
Pass 1 vs pass 2	20.58	0.166	−9.17 to 50.34
Pass 1 vs pass 3	56.42	0.004*	20.33 to 92.51
Pass 1 vs pass 4	41.27	0.008*	11.69 to 70.85
Pass 1 vs pass 5	64.55	0.008*	18.88 to 110.21
Pass 1 vs pass 7	164.26	p < 0.001*	103.52 to 225.00
Groin puncture time to on-clot time
Stroke cart+parallel workflow	−16.27	0.095	−35.55 to 3.01
ETTIR	24.89	0.088	−3.99 to 53.77
Arterial line by anesthesiologist	−1.56	0.852	−18.62 to 15.50
General anesthesia	36.97	0.046*	0.69 to 73.26
Basilar > L ICA	11.42	0.6092
Basilar > L ICA T	49.27	0.0144*
Basilar > L M1	32.58	0.0487*
Basilar > L M2	42.81	0.0223*
Basilar > L tandem	10.61	0.6680
Basilar > R ICA T	35.17	0.0406*
Basilar > R M1	42.25	0.0285*
Basilar > R M2	23.68	0.3895
Basilar > R tandem	4.21	0.8707
Groin puncture time to recanalization time
Stroke cart+parallel workflow	−19.82	0.087	−42.76 to 3.12
ETTIR	29.96	0.082	−4.13 to 64.06
Arterial line by anesthesiologist	−2.55	0.815	−24.82 to 19.72
General anesthesia	35.19	0.103	−7.67 to 78.05
Pass 1 vs pass 2	15.067	0.273	−12.67 to 42.81
Pass 1 vs pass 3	49.56	0.006*	15.91 to 83.20
Pass 1 vs pass 4	61.80	p < 0.001*	34.22 to 89.37
Pass 1 vs pass 5	69.86	0.002*	27.29 to 112.44
Pass 1 vs pass 7	152.15	p < 0.001*	95.51 to 208.78

ETTIR: patients intubated in the angiography suite; L: left; R: right; ICA: internal carotid artery; CI: confidence interval. *p value < 0.05.

The number of patients achieving a good outcome (defined as 90-day or 30-day mRS of 0−2, or 24-hour NIHSS score decrease of greater than eight points) did not differ between the two workflow groups, with a good outcome seen in 9/18 (47%) patients pre-workflow changes, and 14/28 (50%) of patients post-workflow changes (p = 0.60).

Discussion

The results of this study indicate that a few simple workflow changes in the angiography suite, namely use of a stroke cart, parallel workflows, and use of CS whenever possible, can have a significant impact on mechanical thrombectomy in-room time metrics. An analysis of Endovascular Treatment for Small Core and Proximal Occlusion Ischemic Stroke (ESCAPE) trial data showed that use of GA and endovascular techniques offer major opportunities for improvement in workflow, ¹⁶ and a similar study of SWIFT-PRIME trial data showed that attention to workflows with recording and tracking times and setting aggressive target time goals coupled with provider feedback could contribute to improved procedural efficiency. ¹⁷ An emphasis on process improvement to maximize clinical benefit of mechanical thrombectomy^10–12 has produced a growing number of strategies to hasten groin puncture within the angiography suite and decrease the time to which a device makes first contact with a thrombus. In addition, the importance of rapid initiation of a mechanical thrombectomy procedure is highlighted by the development of subspecialty society “ideal” targets, such as the 60-minute door-to-groin puncture time advocated for by the Society of NeuroInterventional Surgery. ¹⁸

One strategy is the use of CS instead of GA. This is supported by the use of CS in the majority of mechanical thrombectomy cases in the 2015 clinical trials.^1–5 Most of the time delay in starting a mechanical thrombectomy procedure when using GA is the placement of the ETT and induction. Since 5/18 patients in the pre-workflow change cohort had ETT placed and induction in the ED, and only one patient in the pre-workflow change group did not receive GA, the variable GA in our analysis may not be a reliable predictor of time metrics in our study. As a result, we performed our analyses using ETT placement in the angio suite (ETTIR) as a distinct variable. ETTIR at our center resulted in a statistically significant prolongation of IROCT and IRRT time metrics and a trend toward delayed IRGP (Table 3). Furthermore, ETTIR does not always occur prior to groin puncture. Some patients started the procedure under CS and subsequently required ETTIR after groin puncture and conversion to GA. This may explain why the effect of ETTIR on IRGP approached but did not reach statistical significance.

Controversy still surrounds the type of anesthesia a patient should receive during mechanical thrombectomy procedures. In a study by Whalin et al., small decreases (≥10 mmHg from baseline) in mean arterial pressure during procedures, regardless of the type of sedation used, were associated with increased odds ratio (OR) for poorer outcome of 2.22 and that every 10-mm Hg drop in mean arterial pressure below 100 mm Hg had an OR of 1.28. ¹⁹ Retrospective studies have supported the use of CS/MAC in avoiding hypotension-induced expansion of the core infarct^20,21 and achieving better outcomes.^21–23 A subgroup analysis of the Multicenter Randomized Clinical Trial of Endovascular Treatment for Acute Ischemic Stroke in the Netherlands (MR CLEAN) trial showed that of the 216 patients, 79 patients had thrombectomy performed under GA and 137 under CS/MAC. While there was no significant difference in time to revascularization between the two groups, the MAC group was an average of 28 min faster to treatment initiation. Further analysis adjusting for age, NIHSS, time from onset, previous stroke, diabetes, atrial fibrillation, and terminal internal carotid artery occlusion showed that the beneficial treatment effect of thrombectomy was lost on the intubation group.^5,24,25 In contrast, however, the results of the Intubation for Endovascular Stroke TreAtment (SIESTA) trial, which randomized 150 mechanical thrombectomy patients at several German centers to receive GA or CS, showed improvements in time metrics with CS but no difference in patient outcomes at 24 h and a higher rate of functional independence (mRS 0–2) after three months in the GA group compared to the CS group (37% vs 18%, respectively, p = 0.01). ²⁶ Shortcomings of the SIESTA trial, however, include the exclusion of one-third of thrombectomy patients from randomization, the GA protocol used, relatively long procedure times, and relatively low rates of good clinical outcomes in both arms. Performing mechanical thrombectomy with CS may be justified as a primary approach unless there are concerns about airway protection, hemodynamic stability, or inability of the patient to cooperate, jeopardizing patient safety during the procedure.

In this study, utilization of a stroke cart also contributed to improvements in time metrics. A stroke cart, containing all commonly utilized procedural equipment (Figure 1, Table 1), from groin puncture to recanalization, facilitates efficient parallel workflows whereby team members can readily obtain and open equipment, or prepare the next procedural step without leaving the angiography suite. The characteristics of an ideal stroke cart include: enough storage space for stocking all essential items, clear visibility and accessibility of contents, lightweight construction, swivel caster wheels, and being narrow enough to be easily transported from one room to another.

Another strategy is the use of a pre-arranged stroke tray in the angiography suite, or BRISK, containing most of the items necessary for a angiography procedure, including a Chloraprep, sterile drapes, cups, marker, micropuncture access kit, syringes, needles, scalpel, tubing, sterile gauze, and towels.^10,11,13,14 Some high-volume centers prepare BRISK trays every morning in anticipation of a mechanical thrombectomy patient later in the day or evening. Maintaining such a kit on stand-by, however, may violate operating room sterility protocols at some centers, especially if the angiography suite is to be used for other procedures. Time metrics have also been shown to improve after standardization of thrombectomy techniques.^14,15 Parallel workflows have been demonstrated in computer science to greatly improve the speed of processing by spreading tasks across multiple processors in parallel computing environments. ²⁷ Similarly, neurointerventionalists, technologists, nursing, and anesthesia staff performing tasks in parallel can shorten procedure times. ¹¹ Additional time savings may be achieved by avoiding placement of a radial arterial line and using the groin sheath for transducing pressures, avoiding a urinary Foley catheter, and omitting shaving of the groin region. ²⁸ In addition, enlisting the assistance of additional personnel and/or cross-training of x-ray and CT technicians, ¹³ especially for after-hours cases, may increase efficiency, although there is a limit to how many people can be in an angiography suite at the same time before it can become detrimental.

Despite significant decreases in time metrics observed with implementation of a stroke cart, parallel workflow, and use of CS, this did not translate into a statistically significant improvement in patient outcomes in this study, as in the SIESTA and Anesthesia During Stroke (AnStroke) trials.^26,29 Owing to the small sample size, however, our study was not powered to detect differences in patient outcomes between groups. Additional limitations of this study include data obtained from a single center, and the use of a retrospective comparison group, which may have introduced performance and chronology biases. In addition, the use of a combined variable of stroke cart and parallel workflow in this study does not permit conclusions to be drawn regarding the effect utility of these two variables independent of one another. Validation of these results in a larger prospective study would be ideal; however, randomizing patients to undergo procedures with less efficient workflows would be unethical. The patients included this study were treated before and after several clinical trials supporting the use of emergent mechanical thrombectomy in LVOAIS were published.^1–5 In addition, concurrent with implementation of the three workflow changes, thrombectomy technique at our institution shifted from stent-retriever plus balloon-guided catheter to increased use of stent-retriever, aspiration catheter, and balloon guide catheter. The number of LVO passes and proportion of cases achieving mTICI 2B or 3 reperfusion did not differ significantly between groups, however, so this device/technique shift is unlikely to have significantly affected time metrics or outcomes. Although we attempted to control for as many possible confounding variables as possible, there may be other logistical or procedural modifications or “learning curves” that may have occurred during the study period that may affect the efficiency of thrombectomy procedures.

In summary, use of a stroke cart containing all commonly used procedural equipment in a mechanical thrombectomy, placed in the angiography suite, combined with parallel workflows by neurointerventionalists, technologists, and nursing staff, and first-line use of CS, improves time metrics in acute LVOAIS thrombectomy procedures.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.