Article Text

Download PDFPDF

Original research
Infarct growth despite full reperfusion in endovascular therapy for acute ischemic stroke
  1. Diogo C Haussen1,
  2. Raul G Nogueira1,
  3. Mohamed Samy Elhammady2,
  4. Dileep R Yavagal2,
  5. Mohammad Ali Aziz-Sultan3,
  6. Jeremiah N Johnson2,
  7. Brandon G Gaynor2,
  8. Shyian Jen1,
  9. Seena Dehkharghani1,
  10. Eric C Peterson2
  1. 1Emory University School of Medicine/Marcus Stroke & Neuroscience Center—Grady Memorial Hospital, Miami, Florida, USA
  2. 2University of Miami Miller School of Medicine/Jackson Memorial Hospital, Miami, Florida, USA
  3. 3Harvard Medical School/Brigham and Women's Hospital, Boston, Massachusetts, USA
  1. Correspondence to Dr Diogo C Haussen, Emory University School of Medicine/Marcus Stroke & Neuroscience Center—Grady Memorial Hospital, 49 Jesse Hill Jr Drive SE, Room #393, Atlanta, GA 30303, USA; diogo.haussen{at}


Aim To explore the predictors of infarct core expansion despite full reperfusion after intra-arterial therapy (IAT).

Methods We retrospectively reviewed 604 consecutive patients who underwent IAT for anterior circulation large vessel occlusion acute ischemic stroke in two tertiary centers (2008–2013/2010–2013). Sixty patients selected by MRI or CT perfusion presenting within <24 h of onset with modified Thrombolysis In Cerebral Infarction (mTICI) grade 3 or 2c reperfusion were included. Significant infarct growth (SIG) was defined as infarct expansion >11.6 mL.

Results Mean age was 67.0±13.7 years, 56% were men. Mean National Institute of Health Stroke Scale (NIHSS) score was 16.2±6.1, time from onset to puncture was 6.8±3.1 h, and procedure length was 1.3±0.6 h. MRI was used for baseline core analysis in 43% of patients. Mean baseline infarct volume was 17.1±19.1 mL, absolute infarct growth was 30.6±74.5 mL, and final infarct volume was 47.7±77.7 mL. Overall, 35% of patients had SIG. Three of 21 patients (14%) treated with stent-retrievers had SIG compared with 14 of 39 (36%) with first-generation devices. Eight of 21 patients (38%) with intravenous tissue plasminogen activator (IV t-PA) had infarct growth compared with 25/39 (64%) without. 23% of patients with SIG had a modified Rankin Scale score ≤2 at 3 months compared with 48% of those without SIG. Multivariate logistic regression indicated that race affected infarct growth. Use of IV t-PA (p=0.03) and stent-retrievers (p=0.03) were independently and inversely correlated with SIG.

Conclusions Despite full reperfusion, infarct growth is relatively frequent and may explain poor clinical outcomes in this setting. Ethnicity was found to influence SIG. Use of IV t-PA and stent-retrievers were associated with less infarct core expansion.

  • Angiography
  • Intervention
  • Stroke

Statistics from

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.


The benefits of intra-arterial therapy (IAT) for patients presenting with acute ischemic stroke (AIS) from large vessel occlusion (LVO) have been demonstrated in the Prolyse in Acute Cerebral Thromboembolism Trial (PROACT) II. However, more than half of the patients had a poor clinical outcome.1 Despite the technological advances and accrued experience, newer thrombectomy trials have demonstrated similar results.2–4 Although stent-retrievers have been shown to lead to faster and better reperfusion compared with previous technology, 42–60% of patients treated with these devices fail to achieve a good outcome.5–7 Even in cases where a full recanalization of all vessels is achieved, 20–49% of individuals with LVO have been reported to have poor outcomes.8 ,9

Infarct growth inversely correlates with chances of good outcome,10 ,11 and may in part explain why patients fully reperfused may not clinically benefit from IAT. We aim to explore the predictors of infarct core expansion in a cohort of fully revascularized patients.


Patient selection

We retrospectively reviewed consecutive patients who underwent IAT for LVO from two tertiary centers from 2008–2013 and 2010–2013, respectively. Patients with modified Thrombolysis In Cerebral Infarction (mTICI) grade 3 (complete antegrade reperfusion with absence of visualized occlusion in all distal branches) were included.12 Individuals with near-complete perfusion without clearly visible thrombus but with delay in contrast run-off (TICI 2c) from one institution were included to enhance power.13 All included patients presented within 24 h of the time last known normal, and underwent an admission non-contrast CT (NCCT) immediately followed by either MRI or CT perfusion for estimation of infarct core. Inclusion required an anterior circulation LVO at the internal carotid artery terminus (ICA-T), middle cerebral artery (MCA) M1 or M2 segment, cervical ICA, or tandem cervical ICA plus ICA-T and/or M1 and/or M2. All patients had follow-up MRI or NCCT 24 h after intervention.

Patient demographics, admission vital signs, laboratory data, preadmission medications, procedural and imaging data were collected. The use of either a primary or rescue thrombectomy was noted. Solitaire Flow Restoration Device and Trevo Retriever were classified as stent-retrievers. Given the very small relative number of patients treated with newer thromboaspiration technology during the study period, thromboaspiration was merged with Merci and classified as first-generation thrombectomy devices. The protocol for both centers was to use conscious sedation or monitored anesthesia care. General anesthesia was reserved for patients who could not protect the airway or had cardiopulmonary insufficiency.

Neuroimaging protocol

In one institution, all NCCT examinations were performed with a 64-section multidetector, 5 mm slices, 120 kVp, 37 mAs. MRI was performed on two 1.5 Tesla imaging systems with a 25 mT/m maximum gradient strength and local signal reception with a dedicated head coil including axial T2-FLAIR (TR: 8000 ms; TE: 125 ms; TI: 2500 ms; slice thickness: 5 mm; interslice gap: 2 mm; matrix size: 162×256) and axial single-shot Echoplanar DWI (TR: 4300 ms; TE: 109 ms; slice thickness: 5 mm; interslice gap: 1.5 mm; matrix size: 150×150). MRI perfusion sequence was a single shot EPI to acquire whole brain images with 2 s temporal resolution. Images were acquired for 1.5 min while injecting contrast using a power injector. Perfusion data were computed pixel-by-pixel to create mean transit time (MTT) maps (full width half maximum) and time-to-peak maps. CT perfusion (CTP) was performed with a 128-slice scanner with collimator of 128×0.6 mm at 80 kVp and 200 mAs, with total coverage of 80 mm. The plane of imaging was parallel to the floor of the anterior cranial fossa starting just above the orbits. Thirty cycles were obtained with a total scan time of 51 s. CTP data were analyzed using a stand-alone version of CTP software developed by Philips Medical Systems (Cleveland, Ohio, USA), which automatically calculated MTT and cerebral blood volume (CBV) defects. The anterior cerebral artery was manually used for arterial input function.

In the other institution, all patients underwent an institutional stroke imaging protocol to include NCCT, CT angiography (CTA), and CTP. CT imaging was performed on a 40 mm 64-detector row clinical system. Helical NCCT (kV 120, auto-mA 100–350) was obtained from the foramen magnum through the vertex at 2.5 mm slice thickness. Two contiguous CTP slabs were obtained for 8 cm combined coverage of the supratentorial brain, obtained at eight 5 mm slices per slab. Cine mode acquisition (kV 80, auto-mA 100) permitting high temporal resolution (q 1 s sampling interval) dynamic bolus passage imaging was obtained following the administration of iodinated contrast. Perfusion defect was defined by a delay of more than 6 s for the maximum of the tissue residue function (Tmax >6 s). The infarct core lesion was defined by a cerebral blood flow (CBF) reduction to <30% of the corresponding contralateral territory using an automated software (RAPID; iSchemaView, Menlo Park, California, USA).

Image analysis

All images were transferred to a separate workstation for analysis using a third party DICOM viewer (Osirix 64-bit; Pixmeo, Geneva, Switzerland). The baseline infarct core volume (mL) was calculated based on MRI diffusion-weighted imaging (DWI) or on CTP. CTP core infarct volumes were estimated according to institutional protocols using either a custom in-house developed perfusion analysis tool (RAPID) or using a threshold of 2% blood volume on CBV.14

Final infarct volume on follow-up imaging was measured using MRI FLAIR or DWI or NCCT (MRI preferred) using a semi-automated lesion outline and segmentation process. Infarct volume on follow-up imaging minus DWI or CBV lesion volume on admission scans were calculated, representing absolute infarct growth. Variable window width and center level settings were used for optimal ischemic hypoattenuation detection with NCCT and DWI images. Significant infarct growth (SIG) was defined by a volume of infarct growth >11.6 mL, which was the mean infarct growth in patients in the Diffusion and Perfusion Imaging for Understanding Stroke Evolution (DEFUSE) study.10

Due to its previously described correlation with larger baseline infarct volumes,15 the presence and extent of leukoaraoisis (LA) in the non-involved hemisphere were assessed using the van Swieten Scale (VSS), which grades white matter lesions anterior and posterior to the central sulcus on a 3-point scale from 0 (no LA) to 2 (confluent white matter involvement from the ventricles to the gray matter) for a total score ranging from 0 to 4.16 To minimize classification bias we dichotomized the degree of LA (VSS 0–2 vs 3–4). The evaluation of collateral grade on catheter-based angiography was performed. All diagnostic runs were analyzed for collateral circulation.17 A 5-point scale was used, with 0=no collaterals visible to the ischemic site; 1=slow collaterals to the periphery of the ischemic site with persistence of some of the defect; 2=rapid collaterals to the periphery of the ischemic site with persistence of some of the defect and to only a portion of the ischemic territory; 3=collaterals with slow but complete angiographic blood flow of the ischemic bed by late venous phase; and 4=complete and rapid collateral blood flow to the vascular bed in the entire ischemic territory by retrograde perfusion. In cases with insufficient information on collateral status (eg, contralateral or posterior circulation not injected due to the time sensitivity of stroke), the variable was not graded.

Statistical analysis

Continuous variables are reported as mean±SD and categorical variables are reported as proportions. Between-group comparisons for continuous/ordinal variables were made with the Student t test, Mann–Whitney U test or ANOVA, as appropriate. Categorical variables were compared by χ2 or Fisher exact test as appropriate. Correlation coefficients were calculated with Spearman or Pearson, as appropriate. Significance was set at p<0.05. Multivariate logistic regression analysis for predictors of SIG was performed for variables at the 0.1 level of significance on univariate analysis using a variable selection method.


Of 604 consecutive patients with AIS treated with IAT during the study period across both sites, 60 individuals fit our inclusion criteria. The baseline demographic and clinical variables are shown in table 1. The mean age was 67.0±13.7 years and 34 (56%) were men. Mean initial National Institute of Health Stroke Scale (NIHSS) score was 16.2±6.1 and mean time from last seen normal to groin puncture was 6.8±3.1 h. The mean duration of IAT was 1.3±0.6 h. MRI was used for baseline core analysis in 26 patients (43%) while the remaining patients were computed through CTP. The mean baseline core infarct volume was 17.1±19.1 mL and the mean baseline perfusion deficit volume was 169.3±78.6 mL. MRI was used for final infarct volumetric analysis in 27 patients (45%) while the remaining patients were analyzed by NCCT. The mean interval between baseline and follow-up scans was 6.3±9.5 days. The mean absolute infarct growth was 30.6±74.5 mL and the mean final infarct volume was 47.7±77.7 mL. Twenty-one patients had infarct growth of >11 mL (35%). More than two-thirds of the patients had mTICI 3 reperfusion (68.3%) while the remaining patients had TICI 2c. There was no statistical correlation between absolute infarct growth and TICI 2c versus mTICI 3 reperfusion (Spearman 0.18; p=0.15). Overall, 40% of patients had a modified Rankin Scale score ≤2 at 3 months.

Table 1

Baseline demographic and clinical variables

The comparison between patients with and without SIG showed that race had an influence on infarct growth. Due to small subgroup sample size for the different ethnic groups, Bonferroni post hoc analysis was not performed. Underlying diabetes was more common in patients with infarct growth, while the use of intravenous tissue plasminogen activator (IV t-PA) was inversely related to stroke expansion on univariate analysis. Baseline core volumes and perfusion defects were not significantly different between patients with and without SIG (table 1). The time from last seen normal to groin puncture and the procedure length were also not statistically different between the two groups (table 2). The occlusion site was similar between groups, as well as angiographic collaterals. The use of a stent-retriever was associated with less chance of SIG. Although SIG had no influence on hemorrhagic transformation, the chances of a good outcome were lower in patients with SIG (table 3).

Table 2

Procedural variables

Table 3

Clinical and imaging outcome

Multivariate analysis indicated that race influenced the chances of SIG. The administration of IV t-PA and the use of stent-retrievers were independently and inversely associated with SIG (table 4). Since no statistical correlation between use of stent-retrievers and procedure length was observed on sensitivity analysis, procedure length was forced into the multivariate model and did not reach significance (OR 2.1, 95% CI 0.8 to 5.5; p=0.11).

Table 4

Multivariate analysis for predictors of significant infarct growth


The primary finding of our study was that SIG occurred in one-third of patients with LVO successfully treated with IAT despite complete revascularization. The use of stent-retrievers and pre-interventional treatment with IV t-PA were associated with smaller infarct growth volume. Moreover, SIG was influenced by race and associated with a worse clinical outcome. This may partially explain the subset of patients that fails to achieve a good outcome despite full recanalization with IAT.

As previously stated, the mismatch between successful recanalization and good outcome has been shown in several recent studies of acute stroke. The mechanisms behind this phenomenon are not clear, but several possibilities exist. Chief among them is the so-called ‘no-reflow phenomenon’. Extensively documented in the cardiology literature, the no-reflow phenomenon refers to lack of appropriate capillary reperfusion despite large vessel recanalization. This phenomenon has been shown to be present in up to 50% of patients undergoing acute percutaneous coronary intervention and is a strong predictor of poor outcome.18 The potential causes underlying this phenomenon are multifactorial and include distal embolization, reperfusion injury, and ischemia-related microcirculatory dysfunction.19 In the acute stroke literature this phenomenon is less well studied, but ischemic changes at the microvascular level appear to play a critical role. As early as 1968 Ames et al20 demonstrated occlusion of the microvasculature in a rabbit model of acute stroke as early as 5 min after the onset of ischemia followed by reperfusion. Interestingly, the occlusion appeared to be only for erythrocytes. Mori et al21 demonstrated capillary occlusion from adhesion of leukocytes shortly after MCA occlusion. Indeed, Okada and colleagues examined the microvasculature in an acute stroke model and found accumulation of activated platelets as early as 2 h after MCA occlusion.22 This raises the possibility that the distal cerebral tissue is viable at the time of recanalization, yet local microvascular changes prevent reperfusion despite more proximal recanalization. mTICI grade 3 reperfusion is defined by full tissue level restoration of antegrade capillary blush, and therefore should minimize this confusion.12 However, complex biochemical and molecular mechanisms that lead to breakdown of neurovascular unit integrity and may lead to necrosis or apoptosis explain further neuronal demise despite local reperfusion.23

In addition, it is possible that the recanalized vessels reoccluded following endovascular intervention. Reocclusion has been demonstrated after a variety of clinical interventions for acute stroke, including IV t-PA (34% rate of reocclusion),24 IA t-PA (17%),25 and mechanical clot disruption (18%). Not surprisingly, arterial reocclusion is associated with a poor outcome.26 A study which evaluated reocclusion following thrombectomy using first-generation device technology by performing repeat catheter angiography 24 h post-intervention found a reocclusion rate of 9%.27 Symptomatic reocclusion was observed in four of 107 patients (3.5%) undergoing IAT (two patients had stents placed and the other two had at least two passes of MERCI).26 It is unknown whether stent-retrievers or the newer thromboaspiration devices lead to less chance of reocclusion. Newer technology leads to faster and more complete reperfusion with fewer device passes, and may lead to less endothelial disruption, less frequent need for rescue therapy, less residual thrombus, and a reduced need for angioplasty and stenting.5–7 The use of stent-retrievers was associated with a lower chance of SIG in our study, and this can be potentially explained by less endothelial injury and denudation, larger chances of retrieving the thrombus en bloc, and/or creation of an immediate perfusing channel after device unsheathing compared with first-generation devices.

The use of IV t-PA has been documented to lead to less infarct growth in reperfused patients.10 ,28 It has been shown that IV t-PA increases tissue reperfusion compared with placebo; furthermore, large vessel recanalization does not necessarily achieve the same effect in attenuating infarct growth as tissue reperfusion in patients receiving IV t-PA.29 Washout of the thrombus fragments that migrated to the microcirculation could explain the beneficial effects on infarct growth that were observed in the current study in patients who received pre-intervention IV thrombolysis. The impact of race on infarct growth was unexpected and we are unsure of its significance, given the small sample of each subgroup. Different risk factor burden and profile have been reported among races in the Northern Manhattan Stroke Study. Moreover, the etiology of stroke is diverse between different ethnic groups, which this study failed to control.30

There are several limitations to the study, including its relatively small size and the limitations inherent to retrospective analyses. We did not record the time between completion of the baseline imaging study and reperfusion; this variable is very representative and could have played an important role in the expansion of the ischemic core. Moreover, we did not systematically obtain follow-up vascular imaging to determine patency. Patients with TICI 2c reperfusion were included and, although there was no statistical impact on infarct growth compared with patients with mTICI 3, this could have influenced the results. The fact that the method for core and final infarct volume was heterogeneous (MRI vs CTP and MRI vs NCCT, respectively) may have affected the accuracy of the volumetric measurements, therefore limiting the analyses. We used 11.6 mL as a threshold for infarct growth, which derives from the DEFUSE cohort.10 Since patients in the DEFUSE study were not required to have LVO at baseline and because revascularization was used instead of reperfusion, we decided to use the mean overall infarct growth instead of focusing on mean volumes derived from patients who were revascularized or had good outcomes. The use of procedural heparin and the utilization of general anesthesia were not recorded, and these variables could theoretically have an effect on infarct growth.


In patients with acute anterior circulation stroke treated with endovascular therapy, a significant proportion had a poor outcome explained by infarct growth despite full reperfusion. The use of IV t-PA and the utilization of stent-retriever devices (versus first-generation technology) were inversely associated with infarct core expansion.



  • Contributors DCH: Study conception, design of the work, acquisition of data, statistical analysis, interpretation of data, drafting of the manuscript. RGN, MSE, DRY, MAA-S, BGG, SJ, SD: Data acquisition, critical revision of manuscript. JNJ: Data collection, critical revision of manuscript. ECP: Study conception, data acquisition, interpretation of data, critical revision of manuscript. All authors gave approval of the final version to be published and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

  • Competing interests MAA-S is a proctor for ev3/Covidien Vascular Therapies (Mansfield, Massachusetts, USA) and Codman (Raynham, Massachusetts, USA). DRY is a consultant to Boston Scientific, Micrus, Abbott Vascular, and Coaxia, and has received travel support from Abbott Vascular. RGN has potential conflicts with Stryker Neurovacular (Trevo-2 Trial PI, DAWN Trial PI), Covidien (SWIFT and SWIFT-PRIME Steering Committee, STAR Trial Core Lab) and Penumbra (3-D Separator Trial Executive Committee).

  • Ethics approval This study was reviewed and approved by the local Institutional Review Boards.

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

  • Data sharing statement The unpublished data from this dataset are held by Jackson Memorial Hospital/University of Miami School of Medicine and ECP as well as Grady Memorial Hospital/Emory University and DCH. Requests for data sharing would be required to be discussed with them directly.