Procedural factors associated with successful recanalization in patients with acute ischemic stroke treated with endovascular thrombectomy—a nationwide register-based observational study

Abstract

Introduction

Several studies have addressed technical aspects of endovascular thrombectomy (EVT), but it is not well known how procedural factors contribute to technical success in routine healthcare. The aim was to explore factors associated with technically successful EVT on nationwide scale.

Methods

We did an observational register-based study assessing factors associated with technical success off anterior circulation EVT in Sweden. The main outcome was successful recanalization defined as modified treatment in cerebral ischemia score 2b-3. The association between baseline and treatment variables and successful recanalization were explored using Chi-square(d) test and univariable logistic regression. Multivariable logistic regression was used to define predictors of successful recanalization.

Results

The study included 3211 patients treated during 2015 to 2020. Successful recanalization was achieved in 83.1% (2667) with a gradual improvement in technical outcome over the period. After adjustment for age and occlusion location, thet use of general anesthesia, balloon guide catheter (BGC) and an operator with an overall success rate of >85% were independent predictors of successful recanalization. An overall operator success rate of <80% or 80–85%, and an annual center volume lower than 50 were predicitors of recanalization failure.

Conclusion

This study illustrates factors associated with procedural success in endovascular thrombectomy on a nationwide scale including the use of general anesthesia, BGC, annual center volumes >50 cases per year and the overall success rate of the individual operator. It highlights the potential benefit of systematic performance measurements, benchmarking, and continuous training to bring all centers and operators to the highest level of performance.

Introduction

Since the publication of the five groundbreaking trials of endovascular thrombectomy (EVT) for acute ischemic stroke (AIS) in 2015, ¹ EVT has become the standard of care for AIS caused by large artery occlusions. From the original proof of principle, additional studies have contributed additional knowledge of the benefit of EVT, including EVT in the extended time window of 8 to 24 h,^2,3 direct aspiration technique, ⁴ and lately for patients with large ischemic core regions at presentation. ⁵

In order to achieve the maximal impact of EVT it is desireable to improve the technical success rate, as well as to extend the treatment to groups typically not included in the randomized controlled trials (RCTs), for example older patients and patients with higher co-morbidity burden, as well as to patients with more challenging vascular anatomies and distal occlusion sites.

The RCTs are typically performed in academic high-volume centers by highly dedicated operators. When introducing EVT into clinical routine care, additional challenges emerge such as less experienced operators performing EVT during nonoffice hours in centers with relatively low patient volumes without access to senior colleauges. The transition of EVT into clinical routine is best assessed by intention-to-treat registries, ensuring that all patients are included in the analysis. One such example is the MR CLEAN registry that showed that the transition from the MR CLEAN RCT into clinical routine care was successful. ⁶ No Swedish centers were involved in the RCTs and therefore did not have the same experience from previous treatment within trials, but nevertheless we have previously shown that EVT for anterior circulation strokes caused by Large Vessel Occlusions (LVO) in clinical routine care has similar results as in the trials, both within 6 h of onset ⁷ and in the extended time-window. ⁸

Technical success in EVT is generally assessed according to the modified treatment in cerebral ischemia (mTICI) score. Good recanalization was in several studies shown to be one of the most important factors for a favorable functional outcome ⁹ and poor revascularisation was associated with a higher risk of symptomatic intracranial bleeding in a recently published study. ¹⁰ Several studies have addressed technical aspects of EVT with increased technical success rates achieved by refining the procedural technique,^11,12 and the recent Swift-Direct study reported recanalization rates above 93% in a multicenter study. ¹³ But these reports are all based on highly selected patients treated in internationally renowned centers, and although it has been shown that similar results can be obtained when introducing EVT in clinical routine care on a national level ⁷ , as well as in countries with less developed stroke care, ¹⁴ but it is still not well known to what extent different procedural factors contribute to technical success in routine healthcare, especially factors that are not easily controllable in a trial such as the time of the day or the day of the week. ¹⁵

The aim of this study was to explore factors associated with technically successful endovascular thrombectomy for AIS on a nationwide scale.

Methods

Study design

We performed an observational register-based study assessing factors associated with technical success for EVT in the treatment of acute ischemic anterior circulation stroke caused by large vessel occlusions.

Data sources

The study was based on the Swedish Stroke Register—Riksstroke, Sweden's national quality register for stroke care since 1994, and the Swedish EndoVAscular treatment for acute Stroke register—EVAS, that started in 2012. The EVAS register did not capture complete data until 2015–2016, but since then the national coverage has been >95% compared to ICD-10 diagnoses in the National Patient Register, which is considered as the ground truth. Riksstroke has had consistently high coverage (>90%) over the last decades. Data from the two registers were combined, creating a database of all EVT's performed between 2015 and 2020, covering the whole chain of stroke care.

Setting

Both registers are hospital-based, and data are entered in Riksstroke by dedicated hospital personnel, while data in EVAS are entered in separate modules by nurses, neurologists, and interventionalists, respectively. EVAS data are entered in all 6 comprehensive stroke centers, while Riksstroke data are entered in all 72 centers handling patients with acute stroke.

Participants

We selected all patients with AIS caused by a large vessel occlusion of the anterior circulation that were treated or attempted to be treated with mechanical thrombectomy. Isolated anterior cerebral artery occlusions were very rare (<1%) and consequently excluded. We also excluded patients without a recorded mTICI score post-EVT. The study period was from 1 January 2015 to 31 December 2020.

Variables

Since the research question was strictly technical, the only demographic factors reported were age, sex, and prestroke independency. The latter was defined as living in own home without home care service and being independent in dressing, toileting, and mobility.

Site of the vascular occlusion was defined on digital subtraction angiography (DSA) by the proximal end of the thrombus, and was categorized into carotid type I, carotid type T, MCA-M1 (middle cerebral artery M1-segment of the MCA) or MCA-M2 and beyond.

Tandem occlusion was defined as an extracranial internal carotid complete occlusion or significant stenosis, combined with an ipsilateral intracranial occlusion of any of the types mentioned above.

The strategy of the first attempt was defined as stent retriever, combined or not combined with a distal aspiration catheter, contact aspiration alone or any other strategy (including intra-arterial thrombolysis, angioplasty, stenting or cases where the exact strategy was not registered). Use of a balloon guide catheter (BGC) was noted.

Attempted EVT was defined as an anterior circulation LVO in combination with arterial puncture, but without any thrombectomy maneuver performed, either due to resolvement of the clot or the clot not being accessed.

Successful recanalization was defined as mTICI 2b–3 on post-EVT DSA.

Office hours were defined by a procedure start between 07.30 am and 16.30 pm on Mondays to Fridays.

Individual operator performance was calculated as the percentage of successful treatments (mTICI 2b–3) of all treatments registered for that specific operator. Interventionist ID was first introduced in mid-2016, treatments performed before that are therefore anonymous (12.7%). Operators were divided into three groups based on their total overall mean performance: <80% success rate, 80–85% success rate, and >85% success rate. Total volume of treatments per interventionalist was defined as the number of EVTs performed since 2016 when the ID was introduced.

Yearly center volume was defined as the yearly total number of EVT procedures at a center.

Statistical methods

SPSS v 28 was used. Baseline data were presented as medians or simple proportions. The association between baseline and treatment variables and successful recanalization were explored using x² test and univariate logistic regression. A p-value <0.05 was considered statistically significant. A multivariable logistic regression was used to define independent predictors of successful recanalization.

Results

Descriptive data

The study included 3211 patients treated endovascularly for AIS in the anterior circulation during 2015–2020. Baseline patient charactaristics and procedural parameters are shown in Table 1. Median age was 74 years and 50.5% were female. M1 occlusions accounted for 53.6% and median National Institute och Health Stroke Scale (NIHSS) score was 16. Over the period there was a gradual increase in overall EVT procedures, the proportion of M2 occlusions increased, and a gradual increase was seen in the use of general anesthesia (GA) (Figure 1).

Table 1.

Baseline characteristics in 3211 patients with anterior circulation AIS treated with EVT in 2015–2020.

Variable	Frequency % (n) N = 3211
Demographics
Median age (IQR)	74 (66–82)
Female sex	50.4% (1618)
Prestroke independency	82.8% (2660)
Stroke characteristics
Occlusion location on DSA
Carotid I	6.9% (220)
Carotid T	12.4% (399)
MCA M1	53.6% (1720)
MCA M2 and beyond	26.4% (849)
Tandem lesion	9.0% (289)
Median NIHSS (IQR)	16 (11–20)
Treatment characteristics
IVT treatment	48.0% (1541)
Type of anesthesia
General anesthesia	35.6% (1142)
Conscious sedation	64.2% (2060)
BGC use	47.7% (1531)
Groin to recanalization time (IQR)	48 (28–81)
Strategy for first attempt (n = 3091)
Stent retriever ± aspiration catheter	60.2% (1933)
Aspiration catheter only	34.1% (1094)
Not specified	2% (64)
Only attempted EVT	3.7% (120)
Intra-operative complication	12.5% (402)
Process data
Number of centers	6
Number of cases per year
2015	261
2016	357
2017	470
2018	634
2019	761
2020	728
Number of procedures by operator success rate categories
<80% success rate (n = 7 operators)	24.5% (786)
80–85% success rate (n = 6 operators)	13.2% (424)
85% success rate (n = 14 operators)	49.6% (1594)
Time of day
Office hours	40.1% (1287)
On-call hours	59.4% (1907)

Missing values were <1%, except for operator performance that was first introduced in mid-2016 and displayed 12.7% missing.

AIS: acute ischemic stroke; BGC: Balloon Guide Catheter; DSA: digital subtraction angiography; EVT: endovascular thrombectomy; IQR: interquartile range; IVT: intravenous thrombolysis; M1: M1-segment of the MCA; M2: M2-segment of the MCA; MCA: middle cerebral artery; NIHSS: National Institute och Health Stroke Scale.

Figure 1.

(A) Increase in EVT procedures during the studied period separated by vessel segment. (B) Proportion of counscious sedation (CS, brown/dark color) and general anesthesia (GA, green/light color) for each study year. (C) Proportion of recanalization according to the mTICI scale during the study period. The mTICI grade 2c was introduced in 2017. (D) Annual success rate from 2017 to 2020 for all operators with the dotted blue line incicating the mean success rate for each year. (E) Annual success rate from 2017 to 2020 for the five operators with the highest and lowest overall success rate (Top5 and Low 5, green/regular and red/dotted lines, respectively) with the dashed green and dashed/dotted red lines indicating the mean of each group. (F) Delta-NIHSS score grouped for operators with an overall success rate <80% (red), 80–85% (orange), and >85% (green) over the entire period. Boxplots of IQR with the median indicated by black lines and dotted colored lines for each group. For the operators with an average success rates <80% the median NIHSS-reduction (improvement) was 4 points (IQR 0–9), compared to 6 points (IQR 0–11) for operators with an average overall success rate of 80–85%, and 6 points (IQR 2–12) for operators with an average overall success rate of >85%. EVT: endovascular thrombectomy; IQR: Interquartile range; mTICI: modified treatment in cerebral ischemia; NIHSS: National Institute och Health Stroke Scale.

Associations between successful recanalization and patient/treatment characteristics

Successful recanalization (mTICI 2b-3) was achieved in 83.1% (2667) and there was a gradual improvement over the observation period as seen in Figure 1. During the observed period, the mTICI scale was altered by the introduction of the 2c grade in 2017. The chance of successful recanalization decreased with patient age (p = 0.001) and there was no difference in recanalization grade between sexes. Successful recanalization was achived more often in M1 occlusions (85.7%) than in M2 and beyond (79.9%, p < 0.001) and Carotid I occlusions (80.5%, p = 0.041). Successful recanalization in Carotid T occlusions was 84.2%, not significantly different compared to M1 occlusions. Intraoperative complications (p < 0.001) and longer procedure times (p < 0.001) were also associated with unsuccessful recanalization.

GA increased the chance of successful recanalization to 87%, compared to 80.9% with contious sedation (p < 0.001). The use of BGC increased the chance of successful recanalization from 80.8% to 85.6% (p < 0.001).

Regarding the initial technical strategy, the univariate analyses show a significant benefit for contact aspiration, but in the multivariable analysis the use of stentretriever as initial technical strategy was instead independently associated with increased success rate (Tables 2 and 3). Combined strategies were introduced during this period and cannot be systematically analyzed based on the registry data.

Table 2.

Association of successful recanalization (mTICI 2b–3) to patient, occlusion, and treatment characteristics in univariate analyses of 3211 patients treated with EVT in 2015–2020.

Variable N = 3211	Successful recanalization % (n)	p-value	OR and 95% CI of successful recanalization (univariate)	p-value
Patient factors
Age	–		0.988 (0.980–0.995)	<0.001
Sex		0.103
Female	82.0% (1327)		0.858 (0.713–1.032)	0.103
Male (ref.)	84.2% (1340)		1
Occlusion characteristics
Location		0.003
Carotid I	80.5% (177)		0.687 (0.480–0.984)	0.041
Carotid T	84.2% (336)		0.890 (0.659–1.203)	0.448
MCA M1 (ref.)	85.7% (1474)		1
MCA M2 and beyond	79.9% (678)		0.662 (0.534–0.821)	<0.001
Tandem lesion		0.864
Yes	82.7% (239)		0.973 (0.706–1.339)	0.864
No (ref.)	83.1% (2428)		1
Treatment characteristics
IVT		0.130
Yes	84.1% (1296)		1.154 (0.959–1.388)	0.130
No (ref.)	82.1% (1371)		1
Type of anesthesia		<0.001
General	87% (993)		1.571 (1.281–1.927)	<0.001
Conscious sedation (ref.)	80.9% (1667)		1
BGC use		<0.001
Yes	85.6% (1310)		1.411 (1.170–1.701)	<0.001
No	80.8% (1357)		1
First attempt strategy		0.007
Stent retriever (ref.)	84.4% (1632)		1
Aspiration catheter	88.0% (963)		1.139 (0.873–1.488)	0.007
Intraoperative complication		<0.001
Yes	73.9% (297)		0.518 (0.403–0.666)	<0.001
No	84.5% (1803)		1
Median groin to recanalization time	_		0.986 (0.984–0.988)	<0.001
Operator experience
Operator success rate		<0.001
<80%	73.8% (580)		0.334 (0.267–0.418)	<0.001
80–85%	81.8% (347)		0.534 (0.398–0.717)	<0.001
>85% (ref.)	89.4% (1425)		1
Operator EVT volume		0.370
<50	83.6% (46)		0.960 (0.463–1.991)	0.913
50–99	88.2% (149)		1.400 (0.858–2.283)	0.178
100–200	83.0% (1018)		0.915 (0.742–1.128)	0.406
>200 (ref.)	84.2% (1123)		1
Center experience
Mean yearly volume		<0.001
<50	83% (240)		0.711 (496–1.021)	0.065
50–100	87.4% (271)		1.009 (0.686–1.484)	0.963
100–200	79.8% (1316)		0.572 (0.457–0.716)	<0.001
>200 (ref.)	87.3% (840)		1
Time of day		0.198
Office hours (ref.)	84.1% (1082)		1
On-call hours	82.3% (1570)		0.883 (0.730–1.067)	0.198

BGC: Balloon Guide Catheter; CI: confidenc interval; EVT: endovascular thrombectomy; IVT: intravenous thrombolysis; mTICI: modified treatment in cerebral ischemia; M1: M1-segment of the MCA; M2: M2-segment of the MCA; MCA: middle cerebral artery; NIHSS: National Institute och Health Stroke Scale; OR: odds ratio.

Table 3.

Factors associated with successful recanalization on a nationwide scale, in a multivariable logistic regression analysis with backwards conditional approach.

Variable	OR	95% CI
		Lower	Upper	p-value
Age	0.985	0.975	0.996	0.007
Occlusion location
MCA M1 (ref)	1
Carotid I	0.842	0.507	1.399	0.508
Carotid T	1.249	0.809	1.929	0.315
MCA M2/M3	0.756	0.560	1.021	0.068
General (conscious sedation ref.)	1.567	1.077	2.280	0.019
BGC use (no ref.)	1.545	0.057	2.258	0.025
Intraoperative complication (no ref.)	0.338	0.225	0.507	<0.001
Mean individual operator success rate
<80%	0.310	0.226	0.424	<0.001
80–85%	0.457	0.299	0.698	<0.001
>85% (ref.)	1
Strategy for first attempt
Aspiration catheter (stent retriever ref.)	0.653	0.441	0.967	0.034
Mean yearly center volume
<50	0.303	0.164	0.561	<0.001
50–100	0.593	0.287	1.225	0.158
100–200	0.470	0.263	0.840	0.011
>200 (ref.)	1
On-call hours (office hours ref.)	0.805	0.616	1.053	0.113

The only variables excluded from the model were sex and total operator volume.

BGC: Balloon Guide Catheter; CI: confidenc interval; M1: M1-segment of the MCA; M2: M2-segment of the MCA; M3: M3-segment of the MCA; MCA: middle cerebral artery; OR: odds ratio.

For the initial technical strategy, there was no significant difference in successful recanalization between stent retriever (with or without distal aspiration catheters) and contact aspiration as the initial strategy in univariable analysis but in the multivariable analysis the use of stent-retriever as initial technical strategy was independently associated with increased success rate (Tables 2 and 3). Combined strategies were introduced during this period and cannot be systematically analyzed in detailindependantly based on the registry data.

EVT performed in centers with less than 50 EVT procedures per year was associated with unsuccessful recanalization in the multivariable analysis (p = 0.003, Table 3).

Operators were divided in three groups based on their overall success rate in the entire material. Operators with <10 procedures at any particular year were not included in the analysis for that year. The overall performance was the strongest predictor of procedural success or failure with an odds ratio of successful recanalization of 0.334 (0.267–0.418) for operators with an average overall success rate <80%, and 0.534 (0.398–0.717) for operators with an average overall success rate of 80–85%, compared to operators with an average overall success rate of >85%. There was no association between total operator volume and success rate.

The annual success rate of the five individual operators with the highest overall success rates was consistently in the order of 20% better than the five individual operators with the lowest success rates, and there was no indication of that gap decreasing over time (Figure 1E-F).

There was also a significant (p < 0.001) difference in positive delta-NIHSS score, that is, the difference in NIHSS score on the day after the procedure compared with the initial NIHSS score. For the operators with an average success rate <80% the median NIHSS score reduction (improvement) was 4 points (interquartile range (IQR) 0–9), compared with 6 points (IQR 0–11) for operators with an average overall success rate of 80–85%, and 6 points (IQR 2–12) for operators with an average overall success rate of >85%.

There was also a difference in the proportion of patients with a worsening of the NIHSS scores, which occurred in 16.8% of the patients treated by operators in the <80% group, 15.8% in patients operated by the 80–85% group, and 12% in patients treated by operators with >85% success rate (p = 0.003). Numerically, the complication rate was higher for operators with lower mean success rate (8.5% for operators with <80% success rate; 4.9% for operators with 80–85% success rate; and 6.8% for the operators with >85% success rate), but the difference was not statistically significant, p = 0.094.

Independent predictors of successful recanalization

Factors that differed significantly between patients with successful recanalization and patients with failed recanalization were included in a multivariable analysis to assess which factors that were independently associated with successful recanalization, as shown in Table 3.

After adjustment for age and occlusion location, the technical and procedural factors that were independently associated with successful recanalization were the use of GA, BGC, and stent retriever or combined technique as first approach.

Factors with a decreased odds of recanalization were patient age, procedure-related complications, procedures performed by operators with <85% overall success rate, and procedures performed at centers with <50 cases/year.

Discussion

This study is based on two merged Swedish quality registries and cover >90% of all EVTs for AIS performed in Sweden by the 6 comprehensive stroke centers serving the Swedish population during this period. All patients undergoing endovascular treatment (defined as a performed arterial puncture) for AIS caused by an LVO are registered in the EVAS registry and all patients receiving in-hospital care for acute stroke are registered in Riksstroke.

In this study we consider mTICI 2b–3 as technical success and there was a gradual improvement over the observation period as seen in Figure 1. Since mTICI grade 2c was introduced and gradually adopted during the study period, we have chosen not to analyze success rate according to mTICI 2c.

The majority of cases were done using stent retrievers as the intial strategy with or without distal aspiration catheters. Combined techniques were gradually introduced during this period and therefore the use of different combined strategies was not possible to analyze separately. Also there is a substantial variety of combined techniques, ¹⁶ currently not registered in the EVAS registry. The multivariable analysis suggests that the use of stent retrievers as first strategy is independently associated with successful recanalization, but the univariable analysis does not show a significant difference and the difference is therefore likely to be small.

Key findings

General anesthesia increases the technical success rate from 81% (sedation) to 87%

The preferred type of anesthesia during EVT has been debated since the start, and and in the early years GA was found to be associated with higher in-hospital mortality, primarily due to higher incidence of postprocedural pneumonia. ¹⁷ However, with the increasing number of EVT procedures, the anesthesia teams are improving their skills of rapid induction of EVT patients to minimize the door-to-puncture time and avoid drops in blood preasure and other potentially negative consequences of GA. The benefits for sedation could therefore not be found in later meta-analyses of RCTs comparing GA to sedation. ¹⁸ In two recent meta-analyses of six RCTs published between 2016 and 2022, GA was associated with a higher technical success rate as well as a more favorable functional outcome at 3 months, without any difference in mortality.^19,20 This development is most certainly a result of improved skills by the anesthesiology teams, however, another possible explanation is the trend to perform more challenging EVT procedures including EVTs for more distal occlusions and in older patients. Our results also show that the technical success rate is lower for M2 occlusions and decreasing with increasing patient age.

The success rate is similar for tandem-lesions and for isolated intracranial occlusions

The presence of a proximal occlusion or severe stenosis requiring treatment in the internal carotid artery is considered an additional challenge and such patients have generally not been included in the EVT RCTs. In our national data, we see that the success rate is similar with our without a tandem lesion. However, the prevalence of tandem occlusions (9%) is lower than expected, and may therefore be incomplete in the EVAS register. This result is somewhat counter-intuitive but in line with previous real-world data from four US centers ²¹ and suggests that most operators today master techniques to overcome the additional technical challenge of the tandem lesion.

The use of Ballon Guide Catheters increases the technical success rate from 81% to 86%

Every technical aspect of a procedure may not be tested individually in RCTs, such as the choices of catheters or guidewires. The benefit of using BGCs to increase the success rate and reduce distal emboli has been shown in vitro models and experimental in vivo models.^22,23 It has also been confirmed in clinical studies ²⁴ and meta-analyses.^25,26 Our results confirm these previous results on the benefit of BGCs.

There is no significant drop in success rate for low volume operators

During the early roll-out of EVT following the successful RCTs, a discernible learning curve was evident. ²⁷ Such a steep improvement curve was, however, only shown in the very early phase of EVT implementation and initial success raterates around 55% to 60% has since improved significantly. But even so, a learning curve for technical success has previously been shown ²⁷ and is associated with procedure times as well as success rate.^28,29 Simulator training has been shown to improve operator performance and may provide a useful tool for EVT training. ³⁰

In our study we did not discover any significant difference in technical success rate for operators that perform <50 EVT procedures/year. This suggest that the training of new operators before performing EVT without supervision at the Swedish centers appear to be sufficient.

Center volume of <50 cases per year is an independent predictor of lower success rate, and our results thereby support the consensus statement, recommending EVT to be performed only if the annual center volume exceeds 50. ³¹ However, our data do not support any additional benefit of concentrating EVT procedures to centers with an even higher annual case load.

The overall success rate of the operator is the most discriminating successfactor

Even though successful recanalization is the most important factor for outcome in AIS caused by LVO in the anterior circulation, there are few studies that have investigated the contribution of interoperator variability on the technical success rate. The presence of a learning curve is clearly shown for EVT^27,28,29 as well as for other neurointerventional^32,33,34 and surgical procedures. ³⁵ It has also recently been shown that the outcome following EVT differs significantly between centers, ³⁶ but there are few studies on the interoperator variability after the initial learning phase.

Our study shows that there is a large interoperator variability in achieving a successful recanalization. The magnitude of this interoperator variability is considerable; if a patient meets an operator with <80% overall success rate, the chance of achieving a succssfull recanalization is 74%, to be compared to 89% if the patient instead meets an operator with >85% overall success rate. In our analysis this is the single largest factor determining successful recanalization.

Several factors could contribute to this variability between operators. Since a low total number of cases (<50/operator annually in the entire material) is not associated with a lower success rate, it is unlikely that the operator experience is explaining the variability. Looking at the annual average success rate for the Top 5 and Low 5 operators (Figure 1E), it is also evident that these operators have a stable success rate over the years without any catch-up effect for the low-performing operators. Since this variability is not explained by a low number of cases for the operator and since the individual performance is largely similar per operator during the observation period, our finding does not appear to represent a learning curve.

Another explanation would be different adherence to the mTICI scale, since it has been shown that operators generally exaggerate their procedural results in self-assessments compared to core-lab evaluations.^37,38 But in our material we can show that the differences are corresponding to clinical improvement (or deterioration) as measured by the delta-NIHSS score (Figure 1F), suggesting that differences in adherence to the mTICI scale may not alone explain the variability.

Another possible explanation would be a selection bias so that senior operators manage the most challenging and difficult cases leading to a higher risk of an unfavorable technical result, but since the majority of EVT cases are performed during on-call hours it is unlikely that any such selection bias is explaining this result.

Instead it seems that individual differences, for example in the choice of strategy or material, bail-out strategies or different individual thresholds for ending a procedure before obtaining successful recanalization, may be more important, in addition to an unavoidable human factor.

If this result can be confirmed by other groups it will likely have major impact in the way we evaluate the performance at an individual level and benchmark procedural techniques to achieve the full potential of the EVT procedures on nation-wide scale.

Limitations

This study has several limitations. The mTICI is self-reported by the operator or someone else at the performing center and over- and underreporting of mTICI may represent a systematic error. Previous validation studies have shown that under- and overreporting are more common in mTICI 2c and 3, and less common between mTICI 2a and 2b.

Aspiration as first attempt is registered in EVAS, but for the early years it was not possible to distinguish between stent retriver only or combined techinques, therefore, the two are grouped.

The results are not related to the proportion of patients with AIS treated by EVT at each center, therefore case selection by centers and/or operators may affect the results.

The EVAS register currently does not register EVT strategy as an independent variable, instead the strategy for the first attempt has been constructed by several other variables, and it is consequently not possible currently to analyze different EVT techniques beyond the choice of the initial strategy.

Conclusions

This study investigates procedural factors associated with procedural success in endovascular thrombectomy on a nationwide scale and show that the use of GA, BGC, an annual center volumes >50 cases per year, and the overall success rate of the individual operator are the most important successfactors for EVT. This work highlights the potential benefit of systemic performance measurement, benchmarking, and continuous training to bring all centers and operators to the highest level of performance.

Abbreviations

AIS

Acute ischemic stroke

BGC

Balloon guide catheter

DSA

Digital subtraction angiography

EVAS

the Swedish EndoVAscular treatment for acute Stroke register

EVT

Endovascular thrombectomy

GA

General anesthesia

IQR

Interquartile range

LAO

Large artery occlusion

mTICI

modified treatment in cerebral ischemia score

M1

The most proximal segment of the middle cerebral artery

M2

The first segments distal of the M1-segment of the middle cerebral artery

MCA

Middle cerebral artery

NIHSS

National Institute och Health Stroke Scale

RCT

Randomized controlled trial