Procedural Predictors of Outcome in Patients Undergoing Endovascular Therapy for Acute Ischemic Stroke
To identify factors impacting outcome in patients undergoing interventions for acute ischemic stroke (AIS).
Materials and Methods
This was a retrospective analysis of patients undergoing endovascular therapy for AIS secondary during a 30 month period. Outcome was based on modified Rankin score at 3- to 6-month follow-up. Recanalization was defined as Thrombolysis in myocardial infarction score 2 to 3. Collaterals were graded based on pial circulation from the anterior cerebral artery either from an ipsilateral injection in cases of middle cerebral artery (MCA) occlusion or contralateral injection for internal carotid artery terminus (ICA) occlusion as follows: no collaterals (grade 0), some collaterals with retrograde opacification of the distal MCA territory (grade 1), and good collaterals with filling of the proximal MCA (M2) branches or retrograde opacification up to the occlusion site (grade 2). Occlusion site was divided into group 1 (ICA), group 2 (MCA with or without contiguous M2 involvement), and group 3 (isolated M2 or M3 branch occlusion).
A total of 89 patients were studied. Median age and National Institutes of health stroke scale (NIHSS) score was 71 and 15 years, respectively. Favorable outcome was seen in 49.4% of patients and mortality in 25.8% of patients. Younger age (P = 0.006), lower baseline NIHSS score (P = 0.001), successful recanalization (P < 0.0001), collateral support (P = 0.0008), distal occlusion (P = 0.001), and shorter procedure duration (P = 0.01) were associated with a favorable outcome. Factors affecting successful recanalization included younger age (P = 0.01), lower baseline NIHSS score (P = 0.05), collateral support (P = 0.01), and shorter procedure duration (P = 0.03). An ICA terminus occlusion (P < 0.0001), lack of collaterals (P = 0.0003), and unsuccessful recanalization (P = 0.005) were significantly associated with mortality.
Angiographic findings and preprocedure variables can help prognosticate procedure outcomes in patients undergoing endovascular therapy for AIS.
KeywordsNeurointerventions Endovascular treatment Stroke therapy Stroke
Two main factors drive an increasing awareness and interest in the endovascular treatment of acute ischemic stroke (AIS). The first, propelled by the medical industry and its quest to capture the stroke “market,” has resulted in several devices becoming available during the span of just a few years [1, 2, 3, 4, 5, 6]. The second is the increased number of physicians with the endovascular skills to perform these procedures [7, 8]. In this mix of multiple stroke devices and a variety of physicians with different skill levels using these devices, it is important to be able to convey realistic information regarding procedure success rates and outcomes to the patients and their families. This is important for obtaining true informed consent before the intervention and for communicating results and expectations at its conclusion. Preprocedure patient selection by functional perfusion imaging has been studied and holds promise, but is not the subject of this series. Our objective was primarily to identify angiographic factors that can predict procedure outcomes, which in turn can be used to guide discussions with patients. We did this by examining our own data set of anterior circulation strokes and by reviewing the literature regarding endovascular stroke therapies.
Materials and Methods
Institutional Review Board approval for this retrospective analysis was obtained. Only patients with AIS symptoms related to the anterior circulation plus a documented corresponding vascular occlusion on computed tomography angiography (CTA) were included. Our CTA protocol involves injection of 70 cc of intravenous contrast followed by a 20 cc saline flush injected at a rate of 4 cc/s through a large-bore antecubital vein. Scan coverage is from the aortic arch to the cranial vertex with 1 mm thick images reconstructed at 1-mm intervals.
All patients underwent endovascular therapy using pharmacological thrombolysis (n = 41 [46.1%]), mechanical thrombectomy (n = 21 [23.6%]). or both (n = 27 [30.3%]). Pharmacological thrombolysis was performed using recombinant tissue plasminogen activator (rt-PA), and the dose ranged from 4 to 24 mg (median 8 [mean 8.4 (±6]). The NIHSS score at admission was recorded as were risk factors for stroke, such as diabetes, hypertension, hyperlipidemia, atrial fibrillation, and smoking. Procedural factors that were captured included time to procedure from symptom onset, procedure duration, location of the vascular occlusion, presence of collateral support, and degree of recanalization based on the thrombolysis in myocardial infarction (TIMI) scoring system.
Procedural outcome was based on TIMI scoring, and recanalization was defined as a TIMI score of 2 or 3. Clinical outcome was determined at 3- to 6-month modified Rankin score (mRS). A favorable outcome was defined as a mRS of 0 to 2. Procedure duration was defined from arterial puncture to groin closure.
The significance of simple bivariate associations was assessed using chi-square tests or logistic regression as appropriate. Multiple logistic regressions were used when several factors were assessed simultaneously. All data analysis was performed using JMP statistical software (SAS Institute, Cary, NC).
A total of 89 patients with anterior circulation AIS who had undergone endovascular therapy and had a clinical outcome available were studied. The mean age was 68.4(±6.9) years (median 71 years [interquartile range [IQR] 58.5 to 82.5]). The median NIHSS score was 15 (IQR 10 to 19). There were 47 (52.8%) women and 42 (47.2%) men. Overall, a favorable outcome (mRS 0 to 2) was seen in 44 of 89 (49.4%) patients with a mortality rate of 23% of 89% (25.8%). The median time to groin puncture from symptom onset was 5 hours (IQR 4.3 to 6.1) (mean 5.9(±3.4).
Factors associated with a favorable outcome (mRS 0–2)
Factors (mean ± SD)
Favorable outcome, n = 44 (49.4%)
Poor outcome, n = 45 (50.6%)
Age (P = 0.006)
NIHSS at admission (P = 0.006)
Procedure duration (min) (P = 0.0049)
Site of occlusion (% [P < 0.0012])
Group 1 (ICA) (n = 16)
Group 2 (MCA) (n = 51)
Group 3 (M2/M3) (n = 22)
Collateral support (% [P = 0.0005])
Grade 0 (n = 27)
Grade 1–2 (n = 62)
Recanalization (% [P < 0.0001])
Unsuccessful (TIMI score 0–1) (n = 36 [40.4%])
Successful (TIMI score 2–3) (n = 53 [59.6%])
Analysis based on age >80 or <80 years
Age ≥ 80 y (n = 29)
Age < 80 y (n = 60)
Favorable outcome (% [mRS 0–2])
Recanalization (% [TIMI score 2–3])
Collateral support (%)
Factors associated with recanalization
Age (mean ± SD)
64 ± 18
74 ± 12
NIHSS at admission (mean ± SD)
13.6 ± 5.6
16.4 ± 7
Collateral support (% [grade 1 or 2])
Procedure duration (min) (mean ± SD)
66 ± 35
89 ± 44
The vascular occlusion site and its association with the NIHSS score, recanalization, outcome, and mortality are depicted in Figure 1. The most common occlusion site was the MCA. The more proximal the occlusion, the greater the NIHSS score, the lower the rate of recanalization and favorable outcome, and the greater the mortality.
Collateral circulation, recanalization, mortality, and outcome
No recanalization (n = 36)
OR (95% CI) P
Successful recanalization (n = 53)
OR (95% CI) P
No collateral support (grade 0 [n = 16])
Collateral support (grade 1–2 [n = 20])
No collateral support (grade 0 [n = 11])
Collateral support (grade 1–2 [n = 42])
P = 0.04
P = 0.0221
Favorable outcome (%)
P = 0.11
P = 0.037
Overall mortality was 25.8% (23 of 89). Mean age in the mortality group was 74 (SD ± 12) years versus a mean age of 66 (SD ± 18) years in the group without mortality (P = 0.04). The NIHSS was 18 ± 6 in patients with mortality versus 14 ± 6 in those without mortality (P = 0.01). Mortality was eight of 53 (15%) in successfully recanalized patients and 15 of 21 (42%) in nonrecanalized patients (odds ratio [OR] = 0.24 [95% CI (0.09 to 0.68); P = 0.005). An ICA terminus occlusion (Fig. 1) and lack of collateral support (Fig. 2) were other factors associated with a greater mortality.
During past years, endovascular stroke treatment has gained awareness not only amongst physicians but the general public as well. An increasing number of hospitals are aspiring to become primary stroke centers and offer the complete gamut of therapies, including endovascular interventions. Physicians without a traditional neurointerventional background, but with the prerequisite skills for endovascular stroke therapy, are becoming a valuable source in supplementing the “stroke call.” The purpose of this study was to identify factors that may impact procedure outcomes. Such information is useful when obtaining informed consent before the intervention and for setting expectations regarding prognosis at its conclusion.
Our data demonstrated an overall favorable outcome (mRS 0 to 2) of 49.4% compared with 27.7% for the MERCI trial , 36% for the multi-MERCI trial , and 25% for the PENUMBRA pivotal trial . Our overall mortality of 25.8% is slightly less than that in the reported literature for endovascular stroke therapy [4, 5, 6]. Factors affecting outcome were divided into preprocedural and intraprocedural and are discussed as such in the following text.
Younger age and lower NIHSS score were the most significant preprocedure predictors of a favorable outcome, whereas older age and greater baseline NIHSS score were associated with greater mortality, a finding supported by other studies [9, 10, 11]. In our analysis, a favorable outcome was seen in 58% of the patients in the <80 years group versus 31% in those ≥80 years (P = 0.01). Likewise, our recanalization rate was 67% in patients <80 years and 45% in those ≥80 years (P = 0.05). However, we did not find any difference in the extent of collateral support between the two groups. A trend toward greater mortality was seen in patients ≥80 years in our study (P = 0.07). Kim et al. concluded that despite greater mortality and lower good outcome, intra-arterial therapy can result in a nondisabling outcome in a quarter of the patients >80 years . Appropriate patient selection using imaging can perhaps aid in identifying these patients who are otherwise at greater risk of complications from endovascular therapy.
Factors influencing clinical outcomes, e.g., cerebrovascular and myocardial interventions, in elderly patients (>80 years) can be broadly divided into those that directly affect procedural success and those that are generally related to the ageing process [12, 13]. Several factors have been postulated to increase the risk of poor outcome and increased mortality in the elderly. These could be diseases, such as cerebral amyloid angiopathy fragile vasculature, and impaired rt-PA clearance [14, 15, 16]. It is possible that elderly patients have a greater likelihood of atherosclerotic narrowing with a superimposed thrombus versus younger patients, who are more likely to have embolic occlusion of a less-diseased vessel. Occlusions of vessels more likely to have atherosclerosis, such as cervical ICA, have lower rates of recanalization. We know from experience that increased vascular tortuosity and complex aortic arch anatomy in older patients makes endovascular interventions more challenging.
Our baseline median NIHSS score of 14.7 is slightly lower than that reported in other device trials [4, 5, 6]. This could be partly due to the exclusion of basilar strokes (high NIHSS score) and inclusion of smaller M2 and M3 branch vessel inclusions (lower NIHSS score). Regardless, baseline NIHSS score significantly impacted clinical outcome, a mean of 12 ± 5 in patients with favorable outcome versus 17 ± 6 in those with a poor outcome (P = 0.001). NIHSS score also strongly correlated with the site of vascular occlusion: The more proximal the occlusion, the greater the NIHSS score. This correlation has been previously documented [4, 5, 6, 17, 18]. Proximal occlusions, such as the ICA or MCA, not only affect a larger part of the brain but are also less supported by collateral circulation compared with more distal occlusions. The majority of patients in our cohort had a major vessel occlusion, either MCA (57%) or intracranial ICA (18%). NIHSS severity correlates with infarct volume and the hemisphere affected . Generally, patients with an NIHSS score ≥10 have been shown to have a vascular occlusion on angiography, whereas those with an NIHSS score <10 are more likely to have a normal angiogram [18, 20]. The Interventional management of stroke (IMS) I and IMS II trials treated patients with NIHSS score ≥10 [21, 22], whereas the trials for the Merci Retrieval System (Concentric Medical, Mountain View, CA) and the Penumbra devices (Penumbra Inc., Alameda, CA) used a score ≥8 to include patients for treatment [4, 6].
We achieved successful recanalization (TIMI score 2 or 3) in 59.5% of patients. Successful recanalization was a significant angiographic predictor of favorable outcome (P < 0.001), whereas unsuccessful recanalization was a significant predictor of mortality (P = 0.005). Factors associated with successful recanalization included younger age, lower NIHSS score, presence of collaterals, and shorter procedure duration. Recanalization as a predictor of favorable outcome is strongly supported by the literature. All endovascular stroke trials, especially the device trials, show it to be the most significant procedural predictor of good outcome [4, 5, 23]. The role of recanalization in good clinical outcome is intuitive. Successfully re-establishing blood flow can reverse potential ischemia in a viable brain. Although our overall recanalization rates are similar to or even less than those of other studies, the percentage of recanalized patients with a favorable outcome in our study was much greater than that reported by other device trials. We had a recanalization rate of 59.5%, which compares favorably with the 48% reported for the MERCI trial , 69.5% for the multi-Merci trial [5, 24], and 66% for the PROACT-II trial . However, almost 72% of the recanalized patients in our study achieved a mRS ≤ 2 versus 46% in the MERCI trial  and 49% in the multi-Merci trial . The highest reported recanalization rate of 82% is that for the Penumbra pivotal trial, which also had the lowest percentage of a favorable outcome in the recanalized group, only 29% . One possible reason for this discrepancy can relate to preprocedure patient selection. There is evidence that supports the role of imaging, whether with computed tomography or magnetic resonance, in identifying infarct core versus reversible ischemia [25, 26, 27, 28, 29], a discussion that is beyond the scope of the current article.
Different grading systems have been employed to assess the collateral circulation in AIS [24, 30, 31]. After reviewing these different methodologies, we adopted a simpler three-tier method for evaluating the collateral circulation one we believed could be easily and practically used by a stroke interventionalist. The most significant angiographic predictor of a favorable outcome after successful recanalization was presence of collateral blood supply, either grade 1 or 2 (P = 0.0008). Lack of collateral support was associated with greater mortality (P = 0.0003). The majority of patients in our study had collateral support, either grade 1 or 2. The association of collateral flow with perfusion parameters  and its impact on outcomes is supported by the literature . Our results are in concordance with those of Christoforidis et al.  in that thrombolytic treatment appears to have a greater clinical impact in those patients with good pial collateral formation. The investigators found both infarct volume and discharge mRS to be significantly lower in patients with better pial collateral than in those with poor pial collaterals, regardless of whether they had complete or partial recanalization . They also demonstrated a clear benefit to intra-arterial thrombolysis, even in patients with unfavorable pial collaterals, because it decreased infarct size . Liebeskind published a comprehensive review on the role of collateral circulation in AIS  that not only covers the different pathways of collateral circulation, both large vessel and pial collaterals, but also offers elegant correlation with physiologic imaging and clinical outcomes. Understanding these concepts and their clinical impact can help plan an intervention as well as offer prognostic information at its conclusion. For instance, aggressive interventions may be warranted in patients without any collateral support, such as ICA terminus occlusions, whereas additional risks may not be justified in patients with robust collaterals.
Site of Occlusion
A large percentage of the M2 or M3 occlusions (n = 22) in our cohort had a favorable outcome (82%], which can skew the results toward a greater percentage of overall favorable outcomes. However, Table 1 lists outcomes by site of occlusion showing a significantly lower rate of good outcomes for ICA and M1 occlusions. The fact that proximal occlusions cause greater morbidity and mortality is well known [4, 5, 6, 35]. Figure 1 illustrates the breakdown of recanalization, outcome, and mortality by site of vascular occlusion. As expected, the worst outcomes and highest mortality are seen in patients with an ICA occlusion. The impact of occlusion site on mortality is even more evident in nonrecanalized patients in whom mortality for an ICA occlusion was almost twice that of the MCA group. ICA terminus occlusions typically carry a greater clot burden and result in abrupt and complete cessation of blood flow to the affected hemisphere by blocking all potential collateral pathways . Thus, patients with a proximal occlusion and lack of collateral support, especially if not recanalized, constitute the highest-risk group in terms of morbidity and mortality .
Another important predictor of favorable outcome in our study was procedure duration. Shorter procedure duration was significantly associated with a favorable outcome (Table 1) as well as successful recanalization (Table 3). In contrast, longer procedure duration was associated with greater mortality. The length of the procedure can be impacted by different factors, such as vessel tortuosity, extent of clot burden, and mechanical versus pharmacological thrombolysis. Regardless of the contributing factors, procedure duration remains important in stroke interventions and may help define an end point for when a procedure should be terminated. The further out the patient from onset of a vascular occlusion and ischemic symptoms, the greater the likelihood that the collateral support sustaining an ischemic penumbra may be overcome, leading to irreversible damage. Perhaps an overlooked factor in longer procedures is that of operator fatigue. Although no real supporting data could be found, it is possible that prolonged procedures without success may lead to unwarranted risks on the part of the operator. In addition, even if recanalization is achieved at the end of a long procedure, it may be futile at best or detrimental at worst .
We had an overall mortality rate of 25.8%. Greater age and NIHSS score, proximal occlusion, lack of collateral circulation, unsuccessful recanalization, and longer procedure duration predicted greater mortality. Our mortality rate and associated risk factors are similar to other studies [4, 5, 6]. However, lack of collateral support as a predictor of mortality was not studied in the previous device trials. Because perfusion imaging can reflect the state of the collateral circulation , the ongoing MR and Recanalization of Stroke Clots Using Embolectomy (MR RESCUE) trial, with its inclusion of preprocedure imaging evaluation, will perhaps offer further indirect insight into the role of collateral circulation in endovascular treatment of AIS.
Limitations of the present study include the retrospective and single-center design. Our data demonstrated an overall favorable outcome (mRS 0 to 2) of 49.4%, which is greater than that reported in the recent device trials, i.e., 28% for MERCI, 36% for multi-MERCI, and 25% for PENUMBRA. This perhaps reflects an inherent selection bias at our institution with perfusion-based patient selection versus time-of-onset-based selection in these trials. We did not account for this selection bias, which is a limitation of the study. The treatment strategies were also somewhat heterogeneous in terms of intra-arterial thrombolytics, mechanical thrombectomy, or both. Because of the small sample size in each treatment group, we did not perform subgroup analysis based on type of endovascular treatment. Last, we excluded posterior circulation ischemic strokes, which carry a high mortality rate regardless of treatment . This is a limitation because the results cannot be generalized to posterior circulation strokes.
In summary, multiple preprocedural and intraprocedural factors can be used to predict a favorable outcome or mortality in a patient undergoing endovascular therapy for AIS. This knowledge is important when obtaining an informed consent and conveying expectations regarding procedure outcomes.
Conflict of interest
The authors declare that they have no conflict of interest.
- 2.Miteff F, Faulder KC, Goh AC, Steinfort BS, Sue C, Harrington TJ (2011) Mechanical thrombectomy with a self-expanding retrievable intracranial stent (Solitaire AB): experience in 26 patients with acute cerebral artery occlusion. AJNR Am J Neuroradiol 32(6):1078–81Google Scholar
- 11.Brinjikji W, Rabinstein AA, Kallmes DF, Cloft HJ (2011) Patient outcomes with endovascular embolectomy therapy for acute ischemic stroke: a study of the National inpatient sample: 2006 to 2008. StrokeGoogle Scholar
- 19.Fink JN, Selim MH, Kumar S, Silver B, Linfante I, Caplan LR et al (2002) Is the association of National institutes of health stroke scale scores and acute magnetic resonance imaging stroke volume equal for patients with right- and left-hemisphere ischemic stroke? Stroke 33(4):954–958PubMedCrossRefGoogle Scholar
- 23.Nogueira RG, Liebeskind DS, Sung G, Duckwiler G, Smith WS (2009) Predictors of good clinical outcomes, mortality, and successful revascularization in patients with acute ischemic stroke undergoing thrombectomy: pooled analysis of the mechanical embolus removal in cerebral ischemia (MERCI) and multi MERCI trials. Stroke 40(12):3777–3783PubMedCrossRefGoogle Scholar
- 24.del Zoppo GJ, Higashida RT, Furlan AJ, Pessin MS, Rowley HA, Gent M (1998) PROACT: a phase II randomized trial of recombinant pro-urokinase by direct arterial delivery in acute middle cerebral artery stroke. PROACT Investigators. Prolyse in acute cerebral thromboembolism. Stroke 29(1):4–11PubMedCrossRefGoogle Scholar
- 30.Qureshi AI (2002) New grading system for angiographic evaluation of arterial occlusions and recanalization response to intra-arterial thrombolysis in acute ischemic stroke. Neurosurgery 50(6):1405–1414 (discussion 1414–1415)Google Scholar