Article Text

Download PDFPDF

Original research
Onset to reperfusion time as a determinant of outcomes across a wide range of ASPECTS in endovascular thrombectomy: pooled analysis of the SWIFT, SWIFT PRIME, and STAR studies
  1. Joon-Tae Kim1,
  2. Mayank Goyal2,
  3. Elad I Levy3,
  4. David Liebeskind4,
  5. Reza Jahan5,
  6. Vitor M Pereira6,
  7. Jan Gralla7,
  8. Alain Bonafe8,
  9. Jeffrey L Saver4
  1. 1Department of Neurology, Chonnam National University Medical School, Chonnam National University Hospital, Gwangju, Korea
  2. 2Departments of Radiology and Clinical Neurosciences, University of Calgary, Calgary, Alberta, Canada
  3. 3Department of Neurosurgery, University at Buffalo, State University of New York, Buffalo, New York, USA
  4. 4Department of Neurology, UCLA, Los Angeles, California, USA
  5. 5Department of Interventional Neuroradiology, UCLA Medical Center, Los Angeles, California, USA
  6. 6Medical Imaging, Toronto Western Hospital, Toronto, Ontario, Canada
  7. 7Department for Diagnostic and Interventional Neuroradiology, Inselspital, University of Bern, Bern, Switzerland
  8. 8Department of Neuroradiology, Hopital Gui De Chauliac, Montpellier, France
  1. Correspondence to Dr Joon-Tae Kim, Department of Neurology, Chonnam National University Medical School, Chonnam National University Hospital, Gwangju 501-757, Korea; alldelight2{at}


Background The time–benefit relationship of endovascular thrombectomy (EVT) according to the size of the core infarct has been incompletely explored in prior studies. We investigated whether established infarct core size on baseline imaging modifies the relationship between onset-to-reperfusion time (OTR) and functional outcomes in patients with acute ischemic stroke treated with EVT.

Methods We analyzed a database containing individual patient data pooled from three prospective Solitaire stent retriever studies. The inclusion criteria were treatment with a Solitaire device and achievement of substantial reperfusion (modified Thrombolysis in Cerebral Infarction 2b–3). Main analyses were performed in patients with baseline Alberta Stroke Program Early CT Scores (ASPECTSs) of 7–10.

Results Among the 305 patients (mean age 67±13 years, 58% women), the proportions of patients in different categories of pretreatment infarct extent were: small (ASPECTS 9–10) 52.0%, moderate (ASPECTS 7–8) 37.1%, and large (ASPECTS 0–6) 7.6%. The mean OTR was 297±95 min. At 3 months, 60.1% of the patients achieved a good outcome. For OTRs of 2–8 hours, the rates of good outcomes at all time points were higher with higher baseline ASPECTS but declined with similar steepness. Both baseline ASPECTS (OR 1.23 (95% CI 1.04 to 1.45)) and OTR (every 30 min delay, OR 0.80 (95% CI 0.73 to 0.88)) were independently associated with a good 3-month outcome. No interaction between OTR and baseline ASPECTS was observed.

Conclusions Although patients with higher baseline ASPECTS are more likely to have good clinical outcomes at all OTR intervals after 2 hours, this benefit consistently declines with time, even in patients with a small infarct core, reinforcing the need to treat all patients as quickly as possible.

  • endovascular thrombectomy
  • solitaire stent
  • onset to reperfusion
  • acute ischemic stroke
View Full Text

Statistics from


Recent randomized trials of endovascular thrombectomy (EVT) have demonstrated its efficacy over conventional therapy in treating acute ischemic stroke with large artery occlusion.1–5 After the failure of previous trials of intra-arterial therapy,6–8 recent trials were performed with newer devices, stricter imaging selection criteria, and more efficient workflows. The resulting higher reperfusion rates, enrollment confined to informative patients, and reduced time to reperfusion enhanced treatment benefits.9–11 In addition, these trials with an early time window of within 8 hours of stroke onset demonstrated that a shorter time from onset to reperfusion (OTR) is associated with better outcomes after endovascular therapy.12–14

An aspect of the time–benefit relationship of EVT that has been incompletely explored in prior early time window studies is the extent to which the pace of decline of the benefit with a longer onset to treatment time can be modified by the core volume at the time of presentation. Among EVT patients, baseline infarct volume is highly correlated with final infarct volume and is a major determinant of functional outcomes.15 Patients with small baseline infarcts may respond to EVT in a ‘time-independent’ manner, experiencing improved outcomes regardless of how long after onset that reperfusion occurs.16–18 As baseline core size reflects the pace of infarct progression through imaging time, patients with larger cores at a particular time after onset may experience their infarcts more quickly, thus resulting in a worse outcome compared with patients with smaller cores.19 Although baseline core volume might be able to identify patients with better prognosis after EVT, this association does not imply that the time to treatment may not be important in patients with a small infarct volume.

Therefore, we undertook this study to determine whether the established core size on baseline imaging modifies the relationship between the OTR and functional outcomes in patients undergoing EVT within an early time window.


We analyzed a database of pooled individual patient-level data within an early window obtained from two randomized controlled trials (the SWIFT and SWIFT PRIME) and a prospective monitored registry (the STAR) (Triple-S registry).2 9 20 Each trial had a dedicated core laboratory, which was described as part of the original publications in the three trials/registry studies. The detailed methods and primary results of these studies have been reported.20–22 Briefly, patients were eligible for these studies if they had acute ischemic stroke with moderate to severe neurological deficits, harbored confirmed occlusions of cerebral arteries of the proximal anterior circulation, and were treatable with mechanical thrombectomy within 8 hours from symptom onset. The local ethics committee at every site approved the study protocol and all patients or their legal representatives provided written informed consent.

In this analysis we included only patients treated with the Solitaire retriever device (Medtronic, Irvine, California, USA) who achieved substantial reperfusion, which was defined as modified Thrombolysis in Cerebral Infarction (mTICI) scores of 2b or 3 as determined by each core laboratory assessment. The time to reperfusion was defined in terms of the following workflow metrics: (1) the OTR; (2) the time from onset to qualifying imaging (OTI); and (3) the time from qualifying imaging to reperfusion (ITR). The OTR was defined as the time from when a patient was last known to be well until visualization of successful reperfusion as defined above in all treatable vessels. To display unadjusted data, OTR intervals were categorized in 60 min intervals as follows: from the initial OTR time to 120 min; 120–179 min, 180–239 min, 240–299 min, 300–359 min, 360–419 min, 420–479 min, and 480 min or more. The baseline Alberta Stroke Program Early CT Scores (ASPECTSs) were categorized into the following groups: small core size (ASPECTS 9–10), moderate core size (ASPECTS 7–8), and large core size (ASPECTS ≤6); Triple-S registries usually exclude patients with baseline ASPECTS of 0–4. Three-month outcomes were assessed by the full range of the modified Rankin Scale (mRS) and dichotomized into functional outcomes as follows: good and poor outcomes (functional independence vs dependence; mRS score 0–2 vs 3–6) or excellent and poor outcomes (freedom from disability vs disability; mRS score 0–1 vs 2–6).

Statistical analysis

The percentage, mean (SD), or median (IQR) is reported depending on the variable characteristics. Categorical variables were analyzed using Pearson’s χ2 test for multigroup comparisons, and Fisher’s exact test was used for two-group comparisons. Continuous variables were analyzed using analysis of variance or the Kruskal–Wallis test as appropriate.

Baseline characteristics were compared among the three baseline ASPECTS categories. However, the main analyses were performed in patients with small and moderate core sizes due to the small number of patients with a large core size (n=33). We investigated the association between the OTR (also the ITR and OTI) and functional outcomes in overall patients and in each of the baseline ASPECTS categories. Multivariable logistic regression analyses were performed to explore the relationships between the OTR (the ITR or OTI in some analyses) and dichotomized functional outcomes, including good and poor outcomes (mRS 0–2) or excellent and poor outcomes (mRS 0–1) at 3 months. Adjusted variables were predefined as follows: age, baseline National Institutes of Health Stroke Scale, atrial fibrillation, target occlusion location, the OTR (the ITR or OTI based on the analysis), and baseline ASPECTS. The interaction between the OTR and baseline ASPECTS (or ASPECTS categories) regarding functional outcomes was investigated, and ORs and 95% CIs were calculated using logistic regression. For graphical display, predicted rates of functional outcomes were obtained from logistic regression models based on the ASPECTS category and OTR (ITR and OTI), which were adjusted for baseline characteristics as cited above. These analyses were performed in all patients, including those with large core sizes. Two-sided p values <0.05 were considered significant. The statistical analyses were performed in R version 3.2 (R Foundation for Statistical Computing, Vienna, Austria).


A total of 305 patients with substantial reperfusion after Solitaire stent retriever therapy (mean age 67±13 years; 42.0% men) were screened for this study. The median pretreatment infarct extent ASPECTS was 9 (IQR 8–10). ASPECTSs were missing for three patients. The proportions of patients in the different categories of pretreatment core extent were as follows: small core (ASPECTS 9–10) 52.0%, moderate core (ASPECTS 7–8) 37.1%, and large core (ASPECTS 0–6) 7.6%. The mean OTR of the entire population was 297±95 min. At 3 months, an excellent outcome (mRS 0–1) was achieved in 43.9% of the patients and a good outcome (mRS 0–2) was achieved in 60.1% of the patients. The general characteristics of the patients according to the baseline ASPECTS categories are presented in table 1. A non-significant trend towards a longer OTR was noted in patients with a large core (p=0.21). However, the median ITR was significantly longer in patients with a small core (148 min) (p=0.01) than in patients with moderate and large cores (120.5 min and 145.5 min, respectively). In contrast, the median OTI was shorter in the ASPECTS 9–10 category than in the ASPECTS 7–8 and ≤6 categories.

Table 1

Characteristics according to baseline ASPECTS categories of patients with substantial reperfusion (mTICI 2b–3) after endovascular therapy

The crude rates of outcomes among the baseline ASPECTS categories are shown in table 2. Among the patients in the three ASPECTS categories, patients with a small core (68.6%) most frequently had good outcomes at 3 months (moderate and large cores; 56.3% and 39.4%, respectively, p=0.001). Similarly, the unadjusted rate of an excellent outcome at 3 months was highest in patients with a small core (p=0.002). Mortality within 3 months occurred more frequently in patients with a large core than in patients with moderate and small cores (p=0.007). For comparison, general characteristics and crude outcome among the baseline ASPECTS categories in patients without substantial reperfusion are shown in online supplementary tables 1 and 2.

Table 2

Unadjusted rates of outcomes according to the baseline ASPECTS categories

In the adjusted binary logistic regression analysis, in the population with small and moderate core size (ASPECTS 7–10), every 30 min delay in the OTR was independently associated with reduced odds of an excellent outcome (OR 0.74, 95% CI 0.66 to 0.83) and a good outcome (OR 0.80, 95% CI 0.73 to 0.88) at 90 days (table 3). Similarly, every 30 min delay in the OTI and ITR resulted in a lower likelihood of good and excellent outcomes at 3 months. No interaction was observed between the OTR and baseline ASPECTS (p=0.51 for individual ASPECTSs, p=0.38 for ASPECTS categories 9–10/7–8).

Table 3

Associations between the OTR interval and functional outcomes in patients with substantial reperfusion (mTICI 2b–3) and baseline ASPECTSs of 7–10

The time–benefit curves for efficacy according to the individual ASPECTSs and the ASPECTS categories are shown in figure 1 and online supplementary figures 1 and 2, respectively. For an OTR of 2–8 hours, the adjusted rates of both good and excellent outcomes at all time points were higher with higher ASPECTSs but gradually declined with similar steepness (figure 1). Similarly, among the ASPECTS categories 9–10 and 7–8, the treatment effect of the OTR did not differ in magnitude and exhibited a similar direction (p interaction >0.1). Similar curves were obtained for the OTI and ITR (see online supplementary figures 1 and 2).

Figure 1

Estimated probabilities of a good outcome (A) and an excellent outcome (B) at 90 days by the time to reperfusion according to the baseline ASPECTS scores (ASPECTS 7–10). mRS, modified Rankin Scale.


In this analysis of more than 300 patients treated with the Solitaire stent retriever who achieved substantial reperfusion (mTICI 2b–3) as recorded in the pooled database of the STAR, SWIFT, and SWIFT PRIME studies, the relationships between the OTR and functional outcomes were robust even in patients with a small core (baseline ASPECTS 9–10) and were not modified by the baseline ASPECTS. With an OTR between 2 and 8 hours, the time–benefit curves show that the benefit gradually and consistently decreased across all baseline individual ASPECTSs and the ASPECTS categories.

Our study provides important information regarding the probability of greater functional outcome decreases with increasing OTRs in patients with small to moderate infarct cores in the early treatment window. A substantially different probability of a good outcome may exist between an earlier OTR and a later OTR, even in patients with the least severe core size (ASPECTS 10). Among patients with small and moderate cores, every 30 min delay in the OTR reduced the likelihood of a good outcome at 90 days by 18% and 26%, respectively. However, because the number of patients with a large core was too low (n=33), the effect estimates of the OTR on outcomes were underpowered (OR 0.99). The estimated probability of a good outcome at 90 days declined similarly and steadily from an OTR of 120 to 480 min among patients in the ASPECTS categories. These results suggest that the magnitude of the likelihood of functional independence at 3 months with an OTR within the early time window appears to be similar among patients with small to moderate infarct cores. These findings may be explained by the inclusion of study populations treated within early time windows and with small-to-moderate core sizes. In the early time window, small infarct cores could grow without early reperfusion; therefore, functional outcomes should not be simply assumed based on baseline infarct core size in the early time window. Our results also provide evidence supporting recent guidelines in which advanced imaging, such as perfusion imaging or MRI, to select patients for EVT in <6 hours is not recommended. Some findings from the current study may conflict with the results of recent studies showing that improved functional outcomes after endovascular therapy were not time-dependent in patients with target mismatch or high ASPECTSs.18 23 24 Ribo et al also found that the OTR was strongly associated with outcomes in patients with low ASPECTSs but not in patients with high ASPECTSs.23 However, these studies differed from our study in either diffusion and/or perfusion imaging (target mismatch)-based EVT, delayed time intervals beyond 6 hours of symptom onset,18 24 or population characteristics analyzed.

The HERMES meta-analysis showed that when EVT could be performed within 7.3 hours, functional outcomes were better the sooner after symptom onset that endovascular reperfusion was achieved.14 Distinct from this meta-analysis, which underemphasized that baseline core infarct significantly affects outcome after EVT, our study provides information regarding the clinical implications of both reperfusion time and baseline infarct core on the outcome of EVT.

Additionally, our study showed that, compared with patients with a large core, patients with a small core had the shortest OTI (small core: median 102 min, moderate core: median 150 min, and large core: median 162 min) but the most delayed ITR (small core: median 148 min, moderate core: median 121 min, and large core: median 146 min). Recently, a new concept of interval times in stroke treatment (ie, the OTI and ITR) was suggested.25 Considering the decreasing rates of good outcomes at the later OTR intervals in patients with higher ASPECTSs, a focus on shortening the ITR may increase the likelihood of good outcomes in these patients.25 Our results showed that a delayed ITR was less likely to result in a better functional outcome at 90 days, which emphasizes the importance of reducing the ITR in the early window treatment. Therefore, these results reinforce the importance of timeliness in the workflow, especially the time to reperfusion, and provide important reassurance regarding the beneficial effects of a shorter OTR on functional outcomes.

Our study had several limitations. First, this study was a post hoc analysis of a pooled dataset in limited time window studies only. This study is also limited by the relatively small sample sizes in other published datasets. Second, as the current study included only patients who were treatable within 8 hours from symptom onset, interpretation should be limited to an early time window. Third, as expected, as the numbers of patients in the lowest ASPECTS categories (no patients with ASPECTSs of 0–2 and only 10 patients with ASPECTSs of 3–4) and patients who were treated at the end of the OTR intervals were low, the effect estimates for these areas should be interpreted with caution. Further studies with larger populations, especially more patients with low ASPECTSs, are warranted to confirm our results. Third, we focused on a specialized subgroup representative of patients who achieved substantial angiographic reperfusion. The benefits of time to treatment according to the ASPECTS may be different for all patients treated with endovascular therapy, including those with non-substantial reperfusion. Although we provided data available for patients without substantial reperfusion (mTICI 0–2a) after EVT, the results were limited because of the small number of samples and no reperfusion times in these subgroups. In addition, given the denominator fallacy,26 the association of the OTR with a good outcome may be more substantial with higher versus lower ASPECTSs among all patients undergoing endovascular therapy. However, our results support current efforts aimed to improve the workflows and stringent imaging selection criteria of recent trials with respect to mechanical thrombectomy. Additionally, these results should be interpreted with caution because controls were not included in the analysis.


In the context of EVT cohorts with substantial reperfusion within an early time window, we found that the relationship between the OTR and functional outcomes could be robust and is not modified by baseline imaging. The likelihood of a good outcome consistently declines with similar steepness across all baseline ASPECTS categories. Our study highlights the critical importance of reducing the OTR to achieve better functional outcomes, even in patients with a small baseline core size, and supports the improvement of workflows for EVT within an early window.


We thank to Medtronic to statistical and overall support for this study.


View Abstract


  • Contributors Study concept and design: J-TK, JLS. Acquisition of data: MG, EIL, DL, RJ, VMP, JG, AB, JLS. Analysis and interpretation of data: J-TK, JLS. Drafting of the manuscript: J-TK, JLS. Critical revision of the manuscript for important intellectual content: J-TK, MG, EIL, DL, RJ, VMP, JG, AB, JLS. Statistical analysis: J-TK, JLS.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests MG: served as a consultant to Medtronic for design and conduct of the SWIFT PRIME trial; has also received honoraria from Medtronic/Covidien for speaking and teaching engagements; and was also one of the Principal Investigators for the ESCAPE trial. The ESCAPE trial was partially funded by Covidien through an unrestricted research grant to the University of Calgary. EIL: serves as a scientific consultant to Medtronic; has shareholder/ownership interest with Intratech Medical and Blockade Medical; has received fees from Abbott for carotid training; and has served as an expert witness for Renders Medical/Legal opinion. DL: a consultant to Stryker (modest) and Covidien (modest) and employed by the University of California, which holds a patent on retriever devices for stroke. RJ is a consultant for Medtronic. VMP: is a consultant for Medtronic (PI for SWIFT PRIME trial), Stryker (SC for DAWN trial [DWI or CTP Assessment With Clinical Mismatch in the Triage of Wake-Up and Late Presenting Strokes Undergoing Neurointervention]), Penumbra (SC for PROMISE study), and BALT (proctorship of products unrelated to ischemic stroke) and receives research grant from Philips. JG: is a global PI of STAR and SWIFT DIRECT (Medtronic), Consultancy; CEC member of the PROMISE study (Penumbra), Consultancy; and receives SNSF grants for magnetic resonance imaging in stroke. AB: has been a consultant for Covidien and has a licensing agreement with GE. JLS is an employee of the University of California. The University of California Regents receive funding for JLS’s services as a scientific consultant regarding trial design and conduct to Medtronic/Covidien, Stryker, Neuravia, BrainsGate, Pfizer, Squibb, Boehringer Ingelheim (prevention only), ZZ Biotech, and St Jude Medical. JLS has served as an unpaid site investigator in multicenter trials run by Lundbeck for which the University of California Regents received payments on the basis of clinical trial contracts for the number of subjects enrolled. JLS serves as an unpaid consultant to Genentech advising on the design and conduct of the PRISMS trial; neither the University of California nor JLS received any payments for this voluntary service. The University of California has patent rights in retrieval devices for stroke.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Patient consent for publication Not required.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.