Article Text

Download PDFPDF

Original research
Factors associated with rapid neurological improvement 24 h following intra-arterial thrombolytic treatment for acute ischemic stroke
  1. Gregorgy Christoforidis1,
  2. Yousef M Mohammad2,
  3. Mohammed Khadir1,
  4. Ming Yang3,
  5. Andrew P Slivka3
  1. 1Department of Radiology, University of Chicago, Chicago, Illinois, USA
  2. 2Department of Neurology, Rush University, Chicago, Illinois, USA
  3. 3Department of Neurology, Ohio State University, Columbus, Ohio, USA
  1. Correspondence to Dr G Christoforidis, Department of Radiology, University of Chicago, 5841 S Maryland Ave, MC2026 Chicago, IL 60637, USA; gchristoforidis{at}


Background and Purpose Significant 24 h improvement is the strongest indicator of functional recovery following thrombolytic treatment for acute ischemic stroke. This study sought to analyze factors contributing to rapid neurological improvement (RNI) following intra-arterial thrombolytic treatment (IATT).

Methods Angiograms and clinical information derived from consecutive patients receiving treatment initiated within 6 h of stroke onset were retrospectively reviewed. RNI was defined as at least 50% 24 h improvement on the National Institutes of Health Stroke Scale score. Logistic regression analysis identified factors associated with RNI. Variables tested included: age, gender, serum glucose, platelet count, pial collateral formation, presenting National Institutes of Health Stroke Scale score, time to treatment, extent of reperfusion, site and location of occlusion, treatment agent and systolic blood pressure.

Results Greater than 50% reperfusion of the involved territory, time to treatment within 270 min and good pial collateral formation (large penumbra zone) significantly predicted RNI. RNI occurred in 31% of the 112 patients studied. RNI occurred in 21/26 (80.8%) patients exhibiting all three favorable variables whereas patients with only one favorable variable had a 6.5% chance of RNI. 94% of patients displaying RNI had a modified Rankin Scale score of 2 or less at 3 months compared with 28.6% without RNI.

Conclusions RNI following IATT for stroke is more likely when at least two of the following are present: good reperfusion, good pial collateral formation and treatment within 4.5 h of symptom onset, and is strongly predictive of 3 month outcomes. Important to clinical management, IATT may need to be reconsidered in patients with poor pial collateral formation if time to treatment exceeds 4.5 h.

Statistics from


Intra-arterial thrombolytic treatment (IATT) has the potential to rescue brain within an ischemic penumbra, which is characterized by delayed tissue perfusion surrounding already infarcted brain tissue. Not all patients benefit from this reconstitution of flow to the ischemic penumbra. Furthermore, the degree and time course of improvement varies from patient to patient. Although gradual clinical improvement tends to occur over time following ischemic stroke, significant improvement within the first 24 h of treatment is a desired outcome and the strongest indicator for subsequent functional recovery. As an outcome measure, significant clinical improvement within the first 24 h of treatment may require smaller enrollments for trials.1 Identifying factors that are predictive of early improvement is critical to understanding and replicating this improvement. This study sought to determine factors which contribute to rapid neurological improvement (RNI) following IATT for acute ischemic stroke.


This study was approved by the local institutional review committee. Data were collected from the retrospective review of consecutive patients who underwent IATT at a single institution. Patients seen within 6 h of symptoms of a carotid or vertebrobasilar stroke who were considered thrombolytic candidates using clinical, laboratory, CT and angiographic criteria derived from the Prolyse in Acute Cerebral Thromboembolism (PROACT) study underwent screening cerebral angiography.2 ,3 Unlike the PROACT study, this study also included patients with ischemic stroke involving territories other than the middle cerebral artery (MCA) and patients over the age of 85 years. Only patients demonstrating intracranial thromboembolic occlusion involving an artery at or distal to the internal carotid bifurcation, at or distal to the basilar artery or posterior inferior cerebellar artery on angiography were included. Patients receiving treatment for basilar artery thrombosis were treated up to 12 h following symptom onset if, based on clinical history, these patients were thought to have an National Institutes of Health Stroke Scale score (NIHSSS) <4 and deteriorated to an NIHSSS above 4 points within 6 h of presentation. Time to treatment for patients with vetebrobasilar occlusion was recorded from the time last known to be at baseline. Following FDA approval for tissue plasminogen activator (tPA) in the setting of acute ischemic stroke, all patients seen between 0 and 3 h were offered intravenous tPA as an alternative to IATT. Patients who received intravenous tPA or bridging intravenous/intra-arterial tPA therapy were not included in this study. Although microcatheter manipulation was left to the discretion of the operator, the microcatheter was embedded within the thrombus during infusion. Patients who underwent embolectomy were not included in this study. Treatment endpoints were considered to be: (1) angiographic signs of hemorrhage such as extravasation of contrast reagent; (2) complete reperfusion with complete recanalization; and (3) maximal thrombolytic dose was reached (100 mg tPA, 1 000 000 units urokinase (UK), 9 mg pro-urokinase (proUK)). If the patient did not recanalize by 6 h, continued infusion of thrombolytic was left at the discretion of the physicians involved. Patients receiving tPA as a rule were not to be anticoagulated following treatment. Presentation NIHSSS, NIHSSS 24–36 h following ictus (24hNIHSSS), NIHSSS at the time of hospital discharge (dcNIHSSS), modified Rankin Scale score (mRS) at 3 months, presentation laboratory values (platelet count, glucose level), systolic blood pressure, history of diabetes mellitus, time to treatment, age and sex were all recorded. Change in NIHSSS (ΔNIHSSS) from the time of ictus was calculated at 24–36 h (Δ24hNIHSSS) and at the time of hospital discharge (ΔdcNIHSSS). NIHSSS was determined by a stroke neurologist (AS, YM). A patient was considered to display RNI when the ratio of Δ24hNIHSSS to presenting NIHSSS was at least 50%.

All angiograms were reviewed for occlusion site, pial collateral formation and reperfusion by an interventional neuroradiologist (GAC) who was blinded to all clinical information during this review. Both pial collateral formation and reperfusion have been shown to be associated with improved outcomes following thrombolytic treatment.4–7 Pial collaterals were graded on a five point scale based on anatomic extent, as defined previously.4 Intraobserver agreement for this scoring method for pial collateral formation was identical in 87% of patients and differed by 1 scoring point in the rest of the patients (κ=0.81, SE=0.065). Grade 1 and grade 2 pial collaterals are considered to represent good pial collaterals and are equivalent to grades 3 and 4 described by the collateral flow grading scheme defined by Higashida and Furlan.7 Only collaterals which reconstituted the proximal portion of the segment adjacent to the occluded vessel (ie, if there was M1 segment occlusion with reconstitution to the proximal M2 vessel segments) were considered good. Internal carotid bifurcation occlusions with patent anterior communicating artery were scored similar to M1 occlusions. Reperfusion was assessed as the revascularized percentage of the affected vascular territory by taking the ratio of the number of reperfused branches of the affected territory to the number of branches occluded prior to recanalization. In this study, complete reperfusion was considered to be complete recanalization with complete reperfusion.8 Reperfusion was then scored on the basis of the modified thrombolysis in myocardial ischemia score (Mori score).9 Patients with no reperfusion were assigned a score of 0, 0–50% reperfusion a score of 1, 50–100% reperfusion a score of 2 and 100% reperfusion a score of 3. Reperfusion scores of 2 and 3 were considered to represent good or favorable reperfusion. Occlusions were considered proximal if the site of occlusion was at the internal carotid artery, basilar artery or at a proximal cerebral segment (m1, a1 or p1 without posterior communicating artery contribution). Occlusions were considered distal if the occlusion was at or more distal to the m2, a2 or p2 cerebral arterial segments or involved cerebellar branches.

Presentation clinical and angiographic information as well as outcome measures were analyzed for significant differences between patients with and without RNI using pairwise comparisons. P values were calculated with Pearson's χ2 test for ordinal and nominal data. Because continuous data were not normally distributed, a two sample Wilcoxon rank sums test was used to determine statistical significance. NIHSSS was treated as non-parametric continuous data. Bonferroni's correction for statistical significance was applied for multiple pairwise comparisons. To identify factors associated with RNI, multivariate logistic regression analysis was carried out for RNI using variables previously shown to influence outcomes or hemorrhage rates in acute ischemic stroke.1 ,5 ,10–14 Factors entered into the model included: age >65 years, gender, platelet count <200 K/μl, glucose <150 mg/dl, presenting systolic blood pressure <150, presenting NIHSSS, time to treatment less than or greater than 270 min, good reperfusion, location of occlusion (internal carotid artery, MCA, anterior cerebral artery, vertebrobasilar), site of occlusion (distal versus proximal), treatment agent (UK, proUK, tPA) and good pial collateral formation. All factors with p<0.10 were entered into the final model as predictors for clinical outcome using backward selection. Factors found to be associated with RNI were further analyzed to estimate potential additive effects using contingency analysis. RNI was compared with Δ24hNIHSSS, ΔdcNIHSSS, (24hNIHSSS – ΔdcNIHSSS) and mRS score at 3 months using a two sample Wilcoxon rank sums test and contingency analysis.


A total of 119 patients received IATT between May 1995 and June 2007. Arteriograms were available in 112. Median treatment doses were: 750 000 units (quartile range 627 500–1 000 000 units) in 22 patients receiving UK, 9 mg in two patients receiving proUK and 40 mg (quartile range 20–65 mg) in 87 patients receiving alteplase. Thrombolytics were stopped in 45 patients due to complete reperfusion, in eight patients due to near complete reperfusion with residual thrombus in vessels too distal (eg, M4 or M5) to attempt further thrombolysis, in 33 patients due to time, in 22 patients due to maximum dose and in four patients with angiographic evidence of hemorrhage (all four completely occluded at the site of the hemorrhage). Median per cent revascularization was 75% (quartile range of 0–100%). Mori reperfusion scores were score 3 in 45 patients, score 2 in 15 patients, score 1 in 15 patients and score 0 in 30 patients. Presenting data are displayed in table 1. Logistic regression analysis for RNI is presented in table 2.

Table 1

Presenting clinical information and angiographic data dichotomized by presence of rapid neurological improvement

Table 2

Nominal logistic regression analysis for rapid neurological improvement

The rates of RNI in patients with greater than 50% reperfusion of the involved territory, good pial collateral formation and time to treatment within 270 min were 51.7%, 41.2% and 51.0%, respectively, which compare favorably with the overall rate of RNI of 31.2%. Figures 1 and 2 display the rate of RNI relative to the predictors derived from regression analysis and their additive effects. RNI occurred in 21/26 (80.8%) patients when all three variables were present. When only one of the three variables was present, only two of 31 patients (6.5%) had RNI. Twelve of 40 patients (30%) with only two variables present exhibited RNI. Nine of these 12 patients had a combination of good pial collateral formation and favorable reperfusion. This combination occurred in 21 patients among whom 43% exhibited RNI. Thus among the 35 patients exhibiting RNI, 30 (86%) had at least good pial collaterals and favorable reperfusion present. Among patients who displayed RNI, time to treatment was less than 270 min in 71% and within 180 min in 29%.

Figure 1

Rate of rapid neurological improvement by pial collateral score, reperfusion score and time to treatment (Pearson χ2).

Figure 2

Additive effects of factors associated with rapid neurological improvement (Pearson χ2).

In addition, 94.1% of patients who exhibited RNI had a 3 month mRS score of 2 or less compared with 28.6% of patients who did not exhibit RNI (p<0.0001, Pearson). Table 3 compares RNI to clinical outcome measures. It is noteworthy that two patients with RNI had a mRS score >2. One died of unrelated causes and the other had a prolonged recovery related to hypersomnolence associated with thalamic infarction but eventually went on to a mRS score of 0.

Table 3

Comparison of clinical outcome measures to rapid neurological improvement


The purpose of this study was to identify factors which lead to rapid clinical improvement following acute ischemic stroke. The current study indicates that immediate clinical improvement, defined as RNI, following IATT in the setting of acute ischemic stroke due to embolic occlusion, is influenced by the presence of good pial collaterals to the ischemic territory, favorable reperfusion and time to treatment within 270 min (table 2). Furthermore, the likelihood of immediate clinical improvement increases when more factors are present (figure 2). Only 6.5% of patients with only one favorable variable exhibited RNI. Based on ORs derived from logistic regression analysis, patients with a combination of good pial collateral formation and favorable reperfusion are more likely to demonstrate RNI. Indeed, the results collected show that 86% of patients with RNI had these two variables.

Using data collected from the National Institute of Neurological Disorders and Stroke (NINDS) trial, Broderick et al demonstrated that change in NIHSSS in the first 24 h post ictus (ie, 24hNIHSSS ≤2 and Δ24hNIHSSS ≥15) was the “most powerful measure of the effectiveness of tPA” during the first 3 months following ischemic stroke.1 A 24hNIHSSS ≤2 occurred in 24% of the treated population derived from part II data from the NINDS tPA trial and in 12% of the population in the current study. The lower rate of 24hNIHSSS ≤2 in the current study is attributable to differences in selection criteria, such as later treatment times. Because a lower rate would be expected to result in lower predictive power, a less stringent criterion would be more appropriate to define significant early improvement in the population reviewed in this study. A per cent improvement in Δ24hNIHSSS rather than an absolute cut-off would be able to modulate for improvement when comparing patients with lower and higher presenting NIHSS scores. With this in mind, a 50% improvement in Δ24hNIHSSS is a reasonable measure from which to identify factors which lead to rapid and substantial clinical improvement following acute ischemic stroke.

The data collected here point to a strong relationship between good pial collateral formation and RNI. Conversely, no patient with a pial collateral score of 4 or 5 exhibited RNI. This is substantiated by studies indicating that good leptomeningeal collateral formation and reperfusion help limit infarct size and improve functional outcome.14–16 Theoretically, residual blood flow, maintained by leptomeningeal collaterals during MCA occlusion, is able to delay apoptosis within endangered tissue in the ischemic pneumbra.15 Patients with a large ischemic penumbra, as indicated by the pial collateral score in this study, benefit the most from reperfusion.

Reperfusion takes place when the occluded vessel recanalizes and restores antegrade blood flow.17 The PROACT II trial established that IATT within 6 h of stroke onset was associated with better outcomes.2 Data presented here indicate that reperfusion of more than 50% of the involved territory following IATT was associated with higher rates of RNI (table 2). The Mechanical Embolus Removal in Cerebral Ischemia (MERCI) 1 study reinforced the fact that significant recovery occurs in patients who recanalized.18 This is further corroborated by a meta-analysis demonstrating the positive impact of recanalization on ischemic stroke outcome, however, despite the potential for reperfusion injury.19

Earlier time to treatment was also found to be independently associated with rapid clinical improvement. In this study, among patients presenting for treatment within 4.5 h, 51.2% displayed RNI compared with 15.9% of those presenting after 4.5 h. Table 1 illustrates that patients with shorter time to treatment (198 min vs 299 min) were more likely to exhibit RNI (p=0.0002). Furthermore, good pial collateral formation and reperfusion additively influence the rate of RNI associated with earlier treatment times, as shown in figure 2. Patients treated with intravenous tPA for acute ischemic stroke within 4.5 h of onset have been shown to have better clinical outcomes at 3 months.10 ,20 The NINDS sponsored tPA for acute ischemic stroke trial found a statistically significant improvement in NIHSSS at 24 h when patients were treated with intravenous tPA within 90 min of ictus but not between 90 and 180 min.10 In this trial, 47% of patients had an improvement of ≥4 points on the NIHSS at 24 h when treated with intravenous tPA within 3 h of symptom onset compared with 39% in the placebo group.10 A re-examination of data derived from the NINDS study found that an NIHSSS of ≤2 at 24 h and a change in NIHSS ≥15 at 24 h were the best measures of intravenous tPA activity.1 The European Cooperative Acute Stroke Study (ECASS) demonstrated a 24 h improvement by a median difference of 2 points on the NIHSSS in their target population compared with no change in the placebo group.21 Finally, Suarez et al found 43% of patients improved ≥4 on the NIHSS at 24 h when treated with intra-arterial tPA within 6 h of symptom onset.22 Based on the results from these studies as well as on data presented here, significant clinical improvement is strongly dependant on early reperfusion treatment.

The current study has several limitations that deserve mention. First, assessment of the extent of pial collateral flow was done retrospectively which has the potential to introduce bias. Arteriograms were reviewed at a separate sitting with the reviewer blinded to the patient's information. Some patients were excluded because angiograms were no longer available for review or angiograms did not provide an adequate evaluation of pial collateral formation. The study was limited to single center experience in patients receiving intra-arterial thrombolytics. RNI outcomes do not necessarily apply to regions with differing demographics and patient care or patients undergoing mechanical thrombectomy or intravenous thrombolysis. As experience with these procedures increased with time, fewer patients with CT hypodensity and pial collateral formation grades 4 and 5 were treated because it had become clear that those patients would not benefit from treatment.4 ,13 As a result, the patient population included fewer patients with poor collateral formation. It would not be reasonable to assume that this would significantly alter conclusions derived from the study. Both anterior and posterior circulation occlusions were included in this study, as was in done the Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke (ATLANTIS), NINDS and ECASS III trials. This was felt to be reasonable because this is true of most clinical practices, and clinical outcomes between the two circulations have been shown to be similar.23 Differences in potential leptomeningeal collateral formation between the hemispheric and posterior fossa circulations are largely unknown but should not influence conclusions since they are taken into consideration. Indeed, experimental evidence suggests that leptomeningeal collateral formation between territories is similar.24 ,25 Assessment of leptomeningeal collateral formation as a dichotomized variable is designed to give a best estimate. Thus differences in leptomeningeal collateral assessment between territories are expected to be small with little impact on this study's conclusions. Patients with posterior circulation strokes in this study were too few to perform adequate subgroup comparison analysis. To the extent possible, RNI occurred in 30.1% (31 of 103) of patients with anterior circulation strokes and in 44.4% (five of nine) of patients with posterior circulation stroke. The difference was not significant and can be accounted for by favorable pial collateral formation (89% vs 70%) and rate of reperfusion over 50% (66.7% vs 52.4%) in posterior circulation infarctions with similar time to treatment <270 min (44.4% vs 43.7%). Finally, a permissive inclusion criterion was used for intra-arterial treatment of basilar artery occlusions (NIHSSS on presentation <4 that increased to above 4 within 6 h). This is not uncommon in many practices because the time of onset of basilar artery occlusion is often vague (ie, may begin with nausea, mild imbalance or diplopia) but suddenly progresses, possibly due to conversion of partial to complete occlusion. It was therefore felt to be appropriate.


Patients who are more likely to display RNI (a 24h improvement in NIHSSS by at least 50%) tend to have at least two of the following presenting variables: good pial collaterals (large ischemic penumbra zone), good reperfusion and rapid time to treatment (within 4.5 h) compared with patients who do not display RNI. Overall, RNI occurred at a rate of 31% of the population studied here and a rate of 81% when all three favorable variables were present. If only a single favorable variable was present RNI occurred in only 6.5% of cases. Important to clinical management, if a patient had poor pial collateral formation and time to treatment exceeded 270 min, the chances for RNI were negligible, regardless of reperfusion status. The combination of good pial collateral formation and reperfusion appeared to represent the strongest predictors for RNI. In 94% of cases, patients displaying RNI had an mRS score of 2 or less at 3 months whereas patients who did not experience RNI were less likely to demonstrate a substantial clinical benefit by the time of hospital discharge and at 3 months. The study adds to the body of literature indicating that early collateral augmentation and reliable reperfusion should lead to beneficial outcomes.


The authors wish to acknowledge Peggy Notestine, BS, and Hoda Jradi, MS, for their help.



  • Competing interests None.

  • Ethics approval The study was approved by the local institutional review committee.

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

  • Data sharing statement Data sharing will be considered on request after publication. Please contact the first author of this publication.

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.