You are here

  • Home
  • Archive
  • Volume 14, Issue 7
  • Predictors of poor outcome despite successful endovascular treatment for ischemic stroke: results from the MR CLEAN Registry

Article Text

Article menu
Ischemic stroke
Original research
Predictors of poor outcome despite successful endovascular treatment for ischemic stroke: results from the MR CLEAN Registry
Free
  1. Rob A van de Graaf1,2,
  2. Noor Samuels1,2,3,
  3. Vicky Chalos1,2,3,
  4. Geert J Lycklama a Nijeholt4,
  5. Heleen van Beusekom5,
  6. Albert J Yoo6,
  7. Wim H van Zwam7,
  8. Charles B L M Majoie8,
  9. Yvo B W E M Roos9,
  10. Pieter Jan van Doormaal1,
  11. Wagih Ben Hassen10,
  12. Aad van der Lugt1,
  13. Diederik W J Dippel2,
  14. Hester F Lingsma3,
  15. Adriaan C G M van Es11,
  16. Bob Roozenbeek1,2
  17. MR CLEAN Registry investigators
    1. 1 Radiology and Nuclear Medicine, Erasmus MC University Medical Center, Rotterdam, Netherlands
    2. 2 Neurology, Erasmus MC University Medical Center, Rotterdam, Netherlands
    3. 3 Public Health, Erasmus MC University Medical Center, Rotterdam, The Netherlands
    4. 4 Radiology, HMC Westeinde, Den Haag, Zuid-Holland, The Netherlands
    5. 5 Cardiology, Erasmus MC University Medical Center, Rotterdam, The Netherlands
    6. 6 Neurointervention, Texas Stroke Institute, Plano, Texas, USA
    7. 7 Radiology and Nuclear Medicine, Maastricht University Medical Center, Maastricht, The Netherlands
    8. 8 Radiology and Nuclear Medicine, Amsterdam UMC Locatie AMC, Amsterdam, The Netherlands
    9. 9 Neurology, Amsterdam UMC Locatie AMC, Amsterdam, The Netherlands
    10. 10 Interventional Neuroradiology, Saint Anne Hospital Centre, Paris, Île-de-France, France
    11. 11 Radiology and Nuclear Medicine, Leiden Universitair Medisch Centrum, Leiden, The Netherlands
    1. Correspondence to Rob A van de Graaf, Radiology & Nuclear Medicine and Neurology, Erasmus Medical Center, 3015 GD Rotterdam, Zuid-Holland, The Netherlands; r.a.vandegraaf{at}erasmusmc.nl

    Abstract

    Background Approximately one-third of patients with ischemic stroke treated with endovascular treatment do not recover to functional independence despite rapid and successful recanalization. We aimed to quantify the importance of predictors of poor functional outcome despite successful reperfusion.

    Methods We analyzed patients from the MR CLEAN Registry between March 2014 and November 2017 with successful reperfusion (extended Thrombolysis In Cerebral Infarction ≥2B). First, predictors were selected based on expert opinion and were clustered according to acquisition over time (ie, baseline patient factors, imaging factors, treatment factors, and postprocedural factors). Second, several models were constructed to predict 90-day functional outcome (modified Rankin Scale (mRS)). The relative importance of individual predictors in the most extensive model was expressed by the proportion of unique added χ2 to the model of that individual predictor.

    Results Of 3180 patients, 1913 (60%) had successful reperfusion. Of these 1913 patients, 1046 (55%) were functionally dependent at 90 days (mRS >2). The most important predictors for mRS were baseline patient factors (ie, pre-stroke mRS, added χ2 0.16; National Institutes of Health Stroke Scale score at baseline, added χ2 0.12; age, added χ2 0.10), and postprocedural factors (ie, symptomatic intracranial hemorrhage (sICH), added χ2 0.12; pneumonia, added χ2 0.09). The probability of functional independence for a typical stroke patient with sICH was 54% (95% CI 36% to 72%) lower compared with no sICH, and 21% (95% CI 4% to 38%) for pneumonia compared with no pneumonia.

    Conclusion Baseline patient factors and postprocedural adverse events are important predictors of poor functional outcome in successfully reperfused patients with ischemic stroke. This implies that prevention of postprocedural adverse events has the greatest potential to further improve outcomes in these patients.

    • stroke
    • thrombectomy
    • intervention

    Data availability statement

    No data are available. Data cannot be made available, as no patient approval has been obtained for sharing coded data. However, R syntax and output files of the analyses will be made available on request.

    https://doi.org/10.1136/neurintsurg-2021-017726

    Statistics from Altmetric.com

      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.

      Introduction

      Approximately 50% of patients with ischemic stroke caused by a proximal large vessel occlusion in the anterior circulation do not recover to functional independence, even when successful reperfusion is achieved by endovascular treatment (EVT).1 Factors such as age, National Institutes of Health Stroke Scale (NIHSS) score at baseline, and Alberta Stroke Program Early CT Score (ASPECTS) are associated with poor outcome after successful reperfusion.2–5 A better understanding of the key determinants of poor recovery despite successful reperfusion after EVT could guide researchers and physicians in the development of new treatments to further improve outcomes. Therefore, we aimed to quantify the importance of predictors of poor functional outcome despite successful reperfusion.

      Methods

      Study design

      We used data from the MR CLEAN Registry, which is a national, prospective, open, multicenter, observational monitoring study for stroke intervention centers that perform EVT in the Netherlands. The complete methods and description of variables of the MR CLEAN Registry have been described elsewhere.6 For the present study, we selected patients who were registered between March 2014 and November 2017 and complied with the following criteria: age ≥18 years; presence of a proximal intracranial occlusion in the anterior circulation confirmed on CT angiography (intracranial carotid artery (ICA/ICA-T), middle cerebral artery (M1/M2), anterior cerebral artery (A1/A2)); groin puncture within 6.5 hours after symptom onset; treatment in a center that participated in the MR CLEAN trial; and successful postinterventional macrovascular reperfusion status (extended Thrombolysis In Cerebral Infarction (eTICI) ≥2B) assessed by an independent core laboratory. The current observational study was guided by the STROBE statement.7

      Measures and outcomes

      We constructed multivariable ordinal regression models to predict functional outcome measured with the modified Rankin Scale (mRS) at 90 days. We selected candidate predictors based on expert opinion and availability.2–5 8 In the selection, priority was given to causal and modifiable factors. The predictors were clustered in four groups according to the time of acquisition: baseline patient factors, imaging factors, treatment factors, and postprocedural factors (ie, adverse events). We successively added each group of predictors to a basic model only including baseline patient factors. This resulted in four multivariable ordinal regression models of increasing extensiveness. The most extensive model was used to quantify the relative importance of the individual predictors. Finally, we evaluated the overall explained variance of the most extensive model.

      Subsequently, we repeated these analyses (1) for the subgroup with excellent reperfusion (eTICI ≥2C) and (2) using a modified NIHSS score at 24–48 hours as the outcome. A modification of the NIHSS score at 24–48 hours was necessary to also include patients who died within 48 hours, by assigning them the maximum NIHSS score of 42. This early modified NIHSS score may be a better representation of direct stroke-related factors associated with outcome after EVT as opposed to the mRS and less inflicted by patients who died early.9

      Statistical methods

      Any mRS score assessed within 30 days of symptom onset was considered invalid and treated as missing. For the purpose of unbiased estimation of associations of outcome with baseline characteristics, we replaced missing outcome and predictor values by values derived from multiple imputation by chained equations with five imputations.10 11 After constructing ordinal logistic or linear regression models as appropriate, quantification measures were derived. Nagelkerke’s pseudo-R2 for the mRS (ordinal outcome) and R2 for the modified NIHSS at 24–48 hours (continuous outcome) were applied to quantify the explained variance in outcome by the models. This derived (pseudo-)R2 reflects the explained variance in outcome of the models by the included predictors, ideally aiming to achieve a highest possible score of 1 representing complete variance explanation. Subsequently, the strength of relationship between an individual predictor in the model and the outcome was quantified by the proportion of unique added value in that particular model using the Wald χ2 test, with penalization for df. Further explanation on this approach is provided in online supplemental material 1. For the most important modifiable predictors associated with functional outcome, we calculated the absolute difference in predicted probability for good functional outcome (mRS ≤2) for a typical stroke patient.

      To account for non-linearity of the associations between continuous parameters and outcome, the variables age and systolic blood pressure were handled using restricted cubic splines with three knots based on prior knowledge.8 12 The modified NIHSS at 24–48 hours was log transformed, after adding one point to all NIHSS scores, to best satisfy the linear model (normal distribution of residuals and homoscedasticity).9

      Confidence intervals for individual predictor importance were calculated using bootstrapping with 10 000 iterations. All statistical analyses were performed with R version 3.5.0 (R foundation for Statistical Computing, Vienna, Austria).

      Results

      Study population

      In total, 3180 patients were analyzed. Successful reperfusion was achieved in 1913/3180 (60%) and excellent reperfusion in 1218/3180 (38%) patients (figure 1). The characteristics of patients with successful and excellent reperfusion, together with clustering of predictors according to the four predefined groups, are shown in table 1 (and per reperfusion grade in online supplemental table).

      View this table:
      Table 1

      Cohort characteristics and predictor clustering

      Figure 1
      Figure 1

      Flowchart.

      Within 90 days following EVT, 900/1913 (51%) patients in whom successful reperfusion was achieved remained functionally dependent (mRS >2) or died. This was similar for patients in whom excellent reperfusion was achieved, with 554/1218 (49%) being dependent or died at 90 days. Of the 1913 patients with successful reperfusion, 78 (4%) died within 48 hours and, of the 1218 patients in whom excellent reperfusion was achieved, 46 (4%) died within 48 hours. The median modified NIHSS score at 24–48 hours was 8 (IQR 3, 15) for successful reperfusion and 7 (IQR 3, 14) for excellent reperfusion.

      Variable importance

      The basic model including patient factors already explained the largest proportion in variance (figure 2). Successive addition of the other grouped factors based on clustering on acquisition over time (ie, imaging factors, treatment factors, and postprocedural factors) increased the explained variance by the models, but these were relatively less important than the contribution of patient factors. The four most extensive models, including all independent variables, explained between 42% and 47% of the variation in outcome prediction among patients with reperfusion after EVT. In patients with successful reperfusion (eTICI ≥2B), the five most important individual predictors of functional outcome at 90 days were pre-stroke mRS (added χ2: 0.16), NIHSS at baseline (added χ2: 0.12), symptomatic intracranial hemorrhage (sICH) (added χ2: 0.12), age (added χ2: 0.10), and pneumonia (added χ2: 0.09; figure 3A). The five individual predictors with the highest added χ2 were similar in patients with excellent reperfusion (eTICI ≥2C), although the order of importance and the quantity of added χ2 differed: pre-stroke mRS (added χ2: 0.19), pneumonia (added χ2: 0.12), sICH (added χ2: 0.11), NIHSS at baseline (added χ2: 0.10), and age (added χ2: 0.09; figure 3C). The most important predictors of the modified NIHSS at 24–48 hours as outcome were NIHSS on admission (added χ2: 0.26), sICH that occurred within 24 hours (added χ2: 0.07), collaterals (added χ2: 0.06), duration of the procedure (added χ2: 0.03), and ASPECTS on admission (added χ2: 0.02; figure 3B). In patients with excellent reperfusion, the order of importance of added χ2 of the four most important predictors was similar to those with successful reperfusion, only the fifth most important predictor differed: NIHSS at baseline (added χ2: 0.28), sICH that occurred within 24 hours (added χ2: 0.08), collaterals (added χ2: 0.05), duration of the procedure (added χ2: 0.02), and glucose (added χ2: 0.02; figure 3D).

      Figure 2
      Figure 2

      Performance of models with increasing extensiveness in patients with successful reperfusion defined as eTICI ≥2B (A, B) and excellent reperfusion defined as eTICI ≥2C (C, D), predicting mRS at 90 days (A, C), and modified NIHSS at 24–48 hours (B, D). aDeath within 48 hours was assigned the maximum score of 42. bFor the model with the modified NIHSS at 24–48 hours as outcome, postprocedural factors in the model were restricted to occurrence within 24 hours, which was only recorded for symptomatic intracranial hemorrhage (resulting in exclusion of new ischemic stroke and pneumonia for these analyses). eTICI, extended Thrombolysis In Cerebral Infarction; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale.

      Figure 3
      Figure 3

      Individual penalized variable importance within the most extensive models in patients with successful reperfusion defined as eTICI ≥2B (A, B) and excellent reperfusion defined as eTICI ≥2C (C, D), predicting mRS at 90 days (A, C) and modified NIHSS at 24–48 hours (B, D). Confidence intervals for individual predictor importance were calculated using bootstrapping with 10 000 iterations. aDeath within 48 hours was assigned the maximum score of 42. bHandled using three restricted cubic spines. ASPECTS, Alberta Stroke Program Early CT Score; ER, emergency room; eTICI, extended Thrombolysis In Cerebral Infarction; EVT, endovascular treatment; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale.

      Probability of good functional outcome (mRS ≤2) for a typical stroke patient with sICH was 54% (95% CI 36% to 72%) lower compared with a patient without sICH, and 21% (95% CI 4% to 38%) for pneumonia compared with no pneumonia.

      Discussion

      In this study, in which we evaluated the importance of predictors according to their acquisition over time, we found that baseline patient factors and postprocedural adverse events are the most important predictors of poor functional outcome in patients with ischemic stroke with successful reperfusion after EVT. It is conceivable that prevention of postprocedural adverse events (ie, sICH, pneumonia) has the greatest potential to further improve outcomes.

      Strategies currently investigated that could be of benefit in the prevention of postprocedural adverse events are (1) direct EVT without preceding intravenous alteplase to reduce sICH (MR CLEAN-NO IV (ISRCTN80619088), DIRECT MT (NCT03469206), SKIP (UMIN000021488), SWIFT-DIRECT (NCT03192332), DIRECT-SAFE (NCT03494920), DEVT (ChiCTR-IOR-17013568)), (2) strict blood pressure control to reduce the risk of intracranial hemorrhage (BP TARGET (NCT03160677)), and (3) pharmacological strategies reducing complications like pneumonia (PRECIOUS (ISRCTN82217627)).

      Comparing the most important predictors for the group with successful reperfusion and those with excellent reperfusion, this resulted only in minor differences, assuming that predictor importance seems relatively constant with regard to the level of reperfusion. The results of this study should not be used to determine for which reperfusion grade one should strive as this was not our aim. The observation that pre-stroke mRS was a strong predictor in explaining outcomes after 90 days confirms the hypothesis that most patients with poor outcomes at 90 days already have poor outcomes at baseline and vice versa. Yet, as no perfect prediction was observed, other factors contribute to the prediction of outcomes at 90 days. We observed that the time from stroke onset to admission to the emergency room of the intervention hospital and time from admission to the emergency room of the intervention hospital to groin puncture were relatively less important than the duration of the procedure. It is possible that the importance of preinterventional time intervals was negated by the achievement of successful reperfusion as our analyses were inherent to this selection criterion. The relative importance of duration of the procedure could reflect the difficulty of the procedure caused, for example, by agitation of the patient, tortuosity of the vessels, or performance of multiple attempts, which is associated with poor functional outcomes.13

      Six earlier studies also found that the non-modifiable patient factor baseline NIHSS score was a very important predictor of poor functional outcome despite reperfusion.3–5 14–16 In five of these studies, age was found to be an important predictor of poor functional outcome.3–5 15 16 Two studies found that EVT without prior IV alteplase administration was associated with a poor outcome.4 14 Lower diffusion weighted imaging (DWI) ASPECTS on admission was found to be related to a poor outcome in two studies.5 16 Factors such as collateral status, blood glucose, occlusion location, diabetes mellitus, neutrophil-to-lymphocyte ratio, delayed EVT, mTICI 2B (vs mTICI 3), procedural complications, and higher number of passes (≥3) were mentioned in only one study to be associated with poor functional outcome.4 5 14 15 Our results confirm the suggestion of an opinion review that unfavorable non-modifiable patient factors (reflecting limited ‘brain reserve’) are the most notable factors explaining why patients do not recover despite reperfusion.17 Remarkably, none of these studies evaluated the influence of postprocedural factors such as sICH, pneumonia, and new ischemic stroke (ie, imaging of new brain infarction with corresponding clinical neurologic deficit within 90 days) which, based on our results, are very important in explaining why some patients with successful reperfusion recover well and others do not. It should be kept in mind that the identified factor is not necessarily causal in explaining the detrimental outcomes. For example, it is possible that the occurrence of pneumonia is associated with other conditions like heart failure and sepsis followed by hemodynamic instability and hypoperfusion requiring ICU admission, which actually explains why these patients do worse.

      The variance in outcome explained by the models varied between 42% and 47% so a substantial part of the variability in outcome after successful reperfusion is still unexplained. Therefore, we advocate incentivizing the identification of new predictors as well as optimizing the determination of current predictors. Considering the identification of new predictors, additional information on quantification of perfusion at the microvascular level—preferably at an early stage—could be a useful new approach to improve outcome prediction. Current visual scoring techniques are unable to evaluate vessels <90 μm in diameter,18 yet it is believed that microvascular dysfunction (vasculature <90 μm in diameter) following reperfusion could contribute to poor functional outcomes despite macrovascular reperfusion.19

      Regarding optimization of predictor determination, our current models could be optimized even more by improving both preinterventional and postinterventional quantification of brain tissue status with more advanced neuroimaging techniques such as MRI or CT perfusion instead of CT.

      As no baseline MRI-DWI or CT perfusion data (eg, information on preinterventional perfusion status such as cerebral blood flow and core volume) were available in this observational registry, our analyses were limited to ASPECTS evaluation, which is probably a less accurate measure of the infarct core. Also, in-depth information on, for example, the periprocedural device technique used (eg, use of balloon protection, assisted aspiration), periprocedural blood pressure course, and malignant brain edema could be of additional value to further improve outcome explanation. Furthermore, addition of the follow-up infarct volume might have improved the model’s performance further.20–22 However, this assessment was not available in our dataset. Besides, as we only documented the occurrence of pneumonia, not the occurrence of other infections, this could have limited our study. Nevertheless, pneumonia accounts for at least half of all stroke-related infections and is by far the strongest prognostic factor among the stroke-associated infections.23 24 Another limitation is the possibility of information bias as factors were selected for the model based on prior knowledge. Furthermore, it should be considered that the chosen outcome of the modified NIHSS (including death) should be interpreted with caution as it is not known whether assigning the maximum NIHSS score of 42 to patients who died before 24–48 hours NIHSS assessment is the most optimal strategy for the evaluation of early stroke-related outcome after EVT. However, this outcome is a strong mediator of the mRS at 90 days and might be seen as a more essential evaluation of neurological deficit and directly stroke-related outcome measure, which is less inflicted by early death.9 Finally, although we did not detect a large influence of the included modifiable factors—that is, use of antithrombotic medication (eg, antiplatelets, heparin) or anesthesia type—on poor functional outcome despite reperfusion, it should be kept in mind that these treatments were not assigned systematically and confounding by indication may have occurred. To improve outcomes further, we suggest evaluating the effect of modifiable factors in randomized studies as well as incentivizing the identification of additional modifiable predictors. As these treatments are modifiable, this warrants a further randomized study.

      Conclusions

      Both patient and postprocedural factors are important predictors of outcome in successfully reperfused patients with ischemic stroke. This implies that prevention of postprocedural adverse events has the greatest potential to further improve clinical outcomes in these patients.

      Data availability statement

      No data are available. Data cannot be made available, as no patient approval has been obtained for sharing coded data. However, R syntax and output files of the analyses will be made available on request.

      Ethics statements

      Patient consent for publication

      Ethics approval

      The central medical ethics committee of the Erasmus Medical Centre Rotterdam, the Netherlands, evaluated the study protocol and granted permission to carry out the study as a registry.

      Acknowledgments

      We thank the MR CLEAN (Multicenter Randomized Controlled Trial of Endovascular Treatment for Acute Ischemic Stroke in the Netherlands) Registry investigators.

      References

      1. ↵
        2. Goyal M ,
        3. Menon BK ,
        4. van Zwam WH , et al
        . Endovascular thrombectomy after large-vessel ischaemic stroke: a meta-analysis of individual patient data from five randomised trials. Lancet 2016;387:1723–31.doi:10.1016/S0140-6736(16)00163-X pmid:http://www.ncbi.nlm.nih.gov/pubmed/26898852
        OpenUrlCrossRefPubMed
      2. ↵
        2. Hussein HM ,
        3. Georgiadis AL ,
        4. Vazquez G , et al
        . Occurrence and predictors of futile recanalization following endovascular treatment among patients with acute ischemic stroke: a multicenter study. AJNR Am J Neuroradiol 2010;31:454–8.doi:10.3174/ajnr.A2006 pmid:http://www.ncbi.nlm.nih.gov/pubmed/20075087
        OpenUrlAbstract/FREE Full Text
      3. ↵
        2. Lee S-H ,
        3. Kim BJ ,
        4. Han M-K , et al
        . Futile reperfusion and predicted therapeutic benefits after successful endovascular treatment according to initial stroke severity. BMC Neurol 2019;19:11.doi:10.1186/s12883-019-1237-2 pmid:http://www.ncbi.nlm.nih.gov/pubmed/30646858
        OpenUrlPubMed
      4. ↵
        2. Linfante I ,
        3. Starosciak AK ,
        4. Walker GR , et al
        . Predictors of poor outcome despite recanalization: a multiple regression analysis of the NASA registry. J Neurointerv Surg 2016;8:224–9.doi:10.1136/neurintsurg-2014-011525 pmid:http://www.ncbi.nlm.nih.gov/pubmed/25564538
        OpenUrlAbstract/FREE Full Text
      5. ↵
        2. Tajima Y ,
        3. Hayasaka M ,
        4. Ebihara K
        . Predictors of poor outcome after succesful mechanical thrombectomy in patients with acute anterior circulation stroke. J Clin Interv Radiol 2017;1:139–43.doi:10.1055/s-0037-1612600
        OpenUrl
      6. ↵
        2. Jansen IGH ,
        3. Mulder MJHL ,
        4. Goldhoorn R-JB , et al
        . Endovascular treatment for acute ischaemic stroke in routine clinical practice: prospective, observational cohort study (MR CLEAN registry). BMJ 2018;360:k949.doi:10.1136/bmj.k949 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29523557
        OpenUrlAbstract/FREE Full Text
      7. ↵
        2. von Elm E ,
        3. Altman DG ,
        4. Egger M , et al
        . The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet 2007;370:1453–7.doi:10.1016/S0140-6736(07)61602-X pmid:http://www.ncbi.nlm.nih.gov/pubmed/18064739
        OpenUrlCrossRefPubMedWeb of Science
      8. ↵
        2. Venema E ,
        3. Mulder MJHL ,
        4. Roozenbeek B , et al
        . Selection of patients for intra-arterial treatment for acute ischaemic stroke: development and validation of a clinical decision tool in two randomised trials. BMJ 2017;357:j1710.doi:10.1136/bmj.j1710 pmid:http://www.ncbi.nlm.nih.gov/pubmed/28468840
        OpenUrlAbstract/FREE Full Text
      9. ↵
        2. Chalos V ,
        3. van der Ende NAM ,
        4. Lingsma HF , et al
        . National Institutes of Health Stroke Scale: an alternative primary outcome measure for trials of acute treatment for ischemic stroke. Stroke 2020;51:282–90.doi:10.1161/STROKEAHA.119.026791 pmid:http://www.ncbi.nlm.nih.gov/pubmed/31795895
        OpenUrlCrossRefPubMed
      10. ↵
        2. Donders ART ,
        3. van der Heijden GJMG ,
        4. Stijnen T , et al
        . Review: a gentle introduction to imputation of missing values. J Clin Epidemiol 2006;59:1087–91.doi:10.1016/j.jclinepi.2006.01.014 pmid:http://www.ncbi.nlm.nih.gov/pubmed/16980149
        OpenUrlCrossRefPubMedWeb of Science
      11. ↵
        2. Schafer JL ,
        3. Olsen MK
        . Multiple imputation for multivariate missing-data problems: a data analyst's perspective. Multivariate Behav Res 1998;33:545–71.doi:10.1207/s15327906mbr3304_5 pmid:http://www.ncbi.nlm.nih.gov/pubmed/26753828
        OpenUrlCrossRefPubMedWeb of Science
      12. ↵
        2. Mulder MJHL ,
        3. Ergezen S ,
        4. Lingsma HF , et al
        . Baseline blood pressure effect on the benefit and safety of intra-arterial treatment in MR CLEAN (Multicenter Randomized Clinical Trial of Endovascular Treatment of Acute Ischemic Stroke in the Netherlands). Stroke 2017;48:1869–76.doi:10.1161/STROKEAHA.116.016225 pmid:http://www.ncbi.nlm.nih.gov/pubmed/28432266
        OpenUrlAbstract/FREE Full Text
      13. ↵
        2. Zaidat OO ,
        3. Castonguay AC ,
        4. Linfante I , et al
        . First pass effect: a new measure for stroke thrombectomy devices. Stroke 2018;49:660–6.doi:10.1161/STROKEAHA.117.020315 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29459390
        OpenUrlAbstract/FREE Full Text
      14. ↵
        2. Wang H ,
        3. Zhang M ,
        4. Hao Y , et al
        . Early prediction of poor outcome despite successful recanalization after endovascular treatment for anterior large vessel occlusion stroke. World Neurosurg 2018;115:e312–21.doi:10.1016/j.wneu.2018.04.042 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29673825
        OpenUrlPubMed
      15. ↵
        2. Shi Z-S ,
        3. Liebeskind DS ,
        4. Xiang B , et al
        . Predictors of functional dependence despite successful revascularization in large-vessel occlusion strokes. Stroke 2014;45:1977–84.doi:10.1161/STROKEAHA.114.005603 pmid:http://www.ncbi.nlm.nih.gov/pubmed/24876082
        OpenUrlAbstract/FREE Full Text
      16. ↵
        2. van Horn N ,
        3. Kniep H ,
        4. Leischner H , et al
        . Predictors of poor clinical outcome despite complete reperfusion in acute ischemic stroke patients. J Neurointerv Surg 2021;13:14–18.doi:10.1136/neurintsurg-2020-015889 pmid:http://www.ncbi.nlm.nih.gov/pubmed/32414889
        OpenUrlAbstract/FREE Full Text
      17. ↵
        2. Rabinstein AA ,
        3. Albers GW ,
        4. Brinjikji W , et al
        . Factors that may contribute to poor outcome despite good reperfusion after acute endovascular stroke therapy. Int J Stroke 2019;14:23–31.doi:10.1177/1747493018799979 pmid:http://www.ncbi.nlm.nih.gov/pubmed/30188259
        OpenUrlPubMed
      18. ↵
        2. Gkanatsios NA ,
        3. Huda W ,
        4. Peters KR
        . Effect of radiographic techniques (kVp and mAs) on image quality and patient doses in digital subtraction angiography. Med Phys 2002;29:1643–50.doi:10.1118/1.1493213 pmid:http://www.ncbi.nlm.nih.gov/pubmed/12201409
        OpenUrlCrossRefPubMed
      19. ↵
        2. Dalkara T ,
        3. Arsava EM
        . Can restoring incomplete microcirculatory reperfusion improve stroke outcome after thrombolysis? J Cereb Blood Flow Metab 2012;32:2091–9.doi:10.1038/jcbfm.2012.139 pmid:http://www.ncbi.nlm.nih.gov/pubmed/23047270
        OpenUrlCrossRefPubMed
      20. ↵
        2. Boers AMM ,
        3. Jansen IGH ,
        4. Beenen LFM , et al
        . Association of follow-up infarct volume with functional outcome in acute ischemic stroke: a pooled analysis of seven randomized trials. J Neurointerv Surg 2018;10:1137–42.doi:10.1136/neurintsurg-2017-013724 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29627794
        OpenUrlAbstract/FREE Full Text
      21. ↵
        2. Compagne KCJ ,
        3. Boers AMM ,
        4. Marquering HA , et al
        . Follow-up infarct volume as a mediator of endovascular treatment effect on functional outcome in ischaemic stroke. Eur Radiol 2019;29:736–44.doi:10.1007/s00330-018-5578-9 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29987421
        OpenUrlPubMed
      22. ↵
        2. Ernst M ,
        3. Boers AMM ,
        4. Aigner A , et al
        . Association of computed tomography ischemic lesion location with functional outcome in acute large vessel occlusion ischemic stroke. Stroke 2017;48:2426–33.doi:10.1161/STROKEAHA.117.017513 pmid:http://www.ncbi.nlm.nih.gov/pubmed/28765288
        OpenUrlAbstract/FREE Full Text
      23. ↵
        2. Vermeij FH ,
        3. Scholte op Reimer WJM ,
        4. de Man P , et al
        . Stroke-associated infection is an independent risk factor for poor outcome after acute ischemic stroke: data from the Netherlands Stroke Survey. Cerebrovasc Dis 2009;27:465–71.doi:10.1159/000210093 pmid:http://www.ncbi.nlm.nih.gov/pubmed/19329851
        OpenUrlCrossRefPubMed
      24. ↵
        2. Westendorp WF ,
        3. Nederkoorn PJ ,
        4. Vermeij J-D , et al
        . Post-stroke infection: a systematic review and meta-analysis. BMC Neurol 2011;11:110.doi:10.1186/1471-2377-11-110 pmid:http://www.ncbi.nlm.nih.gov/pubmed/21933425
        OpenUrlCrossRefPubMed

      Supplementary materials

      • Supplementary Data

        This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

      Footnotes

      • RAvdG and NS contributed equally.

      • Collaborators MR CLEAN Registry investigators: Robert J van Oostenbrugge; Jelis Boiten; Jan Albert Vos; Ivo GH Jansen; Maxim JHL Mulder; Robert-Jan B Goldhoorn; Kars CJ Compagne; Manon Kappelhof; Josje Brouwer; Sanne J den Hartog; Wouter H Hinsenveld; Bart J Emmer; Jonathan M Coutinho; Wouter J Schonewille; Marieke JH Wermer; Marianne AA van Walderveen; Julie Staals; Jeannette Hofmeijer; Jasper M Martens; Sebastiaan F de Bruijn; Lukas C van Dijk; H Bart van der Worp; Rob H Lo; Ewoud J van Dijk; Hieronymus D Boogaarts; J de Vries; Paul LM de Kort; Julia van Tuijl; Jo P Peluso; Puck Fransen; Jan SP van den Berg; Boudewijn AAM van Hasselt; Leo AM Aerden; René J Dallinga; Maarten Uyttenboogaart; Omid Eschgi; Reinoud PH Bokkers; Tobien HCML Schreuder; Roel JJ Heijboer; Koos Keizer; Lonneke SF Yo; Heleen M den Hertog; Emiel JC Sturm; Paul JAM Brouwers; Marieke ES Sprengers; Sjoerd FM Jenniskens; René van den Berg; Ludo FM Beenen; Alida A Postma; Stefan D Roosendaal; Bas FW van der Kallen; Ido R van den Wijngaard; Joost Bot; Anton Meijer; Elyas Ghariq; Marc P van Proosdij; G Menno Krietemeijer; Rob Lo; Dick Gerrits; Wouter Dinkelaar; Auke PA Appelman; Bas Hammer; Sjoert Pegge; Anouk van der Hoorn; Saman Vinke; H Zwenneke Flach; Naziha el Ghannouti; Martin Sterrenberg; Corina Puppels; Wilma Pellikaan; Rita Sprengers; Marjan Elfrink; Michelle Simons; Marjolein Vossers; Joke de Meris; Tamara Vermeulen; Annet Geerlings; Gina van Vemde; Tiny Simons; Cathelijn van Rijswijk; Gert Messchendorp; Nynke Nicolaij; Hester Bongenaar; Karin Bodde; Sandra Kleijn; Jasmijn Lodico; Hanneke Droste; Maureen Wollaert; Sabrina Verheesen; D Jeurrissen; Erna Bos; Yvonne Drabbe; Michelle Sandiman; Nicoline Aaldering; Berber Zweedijk; Mostafa Khalilzada; Jocova Vervoort; Eva Ponjee; Sharon Romviel; Karin Kanselaar; Denn Barning; Esmee Venema; Ralph R Geuskens; Tim van Straaten; Saliha Ergezen; Roger RM Harmsma; Daan Muijres; Anouk de Jong; Olvert A Berkhemer; Anna MM Boers; J Huguet; PFC Groot; Marieke A Mens; Katinka R van Kranendonk; Kilian M Treurniet; Manon L Tolhuisen; Heitor Alves; Annick J Weterings; Eleonora LF Kirkels; Eva JHF Voogd; Lieve M Schupp; Sabine Collette; Adrien ED Groot; Natalie E LeCouffe; Praneeta R Konduri; Haryadi Prasetya; Nerea Arrarte-Terreros; Lucas A Ramos.

      • Contributors RAvdG, NS, and BR had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: RAvdG, NS, VC, HFL, BR. Acquisition, analysis, or interpretation of data: RAvdG, NS, VC, GJLaN, HvB, AJY, WHvZ, CBLM, YBWR, PJVD, WBH, AvdL, DWJD; HFL, ACGvE, BR. Drafting of the manuscript: RAvdG, NS, VC, BR. Critical revision of the manuscript for important intellectual content: GJLaN, HvB, AJY, WHvZ, CBLM, YBWR, PJVDL, WBH, AvdL, DWJD; HFL, ACGvE. Statistical analysis: RAvdG, NS, VC, HFL, BR. Study supervision: BR, DWJD, AvdL.

      • Funding The authors received no funding for this study. The MR CLEAN Registry is partially funded by unrestricted grants from Toegepast Wetenschappelijk Instituut voor Neuromodulatie, Twente University (TWIN), Erasmus MC University Medical Center, Maastricht University Medical Center, and Amsterdam UMC.

      • Disclaimer The funding sources had no role in study design, patient enrolment, data collection, analysis, writing of the manuscript, approval of the manuscript, and decision to submit the manuscript for publication.

      • Competing interests DWJD reports funding from the Dutch Heart Foundation, Brain Foundation Netherlands, The Netherlands Organisation for Health Research and Development, Health Holland Top Sector Life Sciences & Health, and unrestricted grants from Penumbra, Stryker European Operations BV, Medtronic, Thrombolytic Science, and Cerenovus for research, all paid to the institution. AvdL reports funding from the Dutch Heart Foundation, Dutch Brain Foundation, Stryker, Angiocare BV, Medtronic/Covidien/EV3, MEDAC GmbH/LAMEPRO, Penumbra, and Top Medical Concentric, all paid to the institution. CBLMM reports funding from CVON/Dutch Heart Foundation, Stryker, Health Evaluation Netherlands, all paid to the institution, and is a shareholder of Nico.lab, a company that focuses on the use of artificial intelligence for medical imaging analysis. YBWR reports funding from CVON/Dutch Heart Foundation, Stryker, Health Evaluation Netherlands, all paid to the institution, and is a shareholder of Nico.lab, a company that focuses on the use of artificial intelligence for medical imaging analysis.

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

      • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.