The impact of intraarterial, intravenous, and combined tirofiban on endovascular treatment for acute intracranial atherosclerotic occlusion

Background and purpose Adjunctive tirofiban administration in patients undergoing endovascular treatment (EVT) for acute large vessel occlusion (LVO) has been investigated in several studies. However, the findings are conflict. This study aimed to compare the effect of different administration pathways of tirofiban on patients undergoing EVT for acute LVO with intracranial atherosclerotic disease (ICAD). Methods Patients were selected from the ANGEL-ACT Registry (Endovascular Treatment Key Technique and Emergency Workflow Improvement of Acute Ischemic Stroke: A Prospective Multicenter Registry Study) and divided into four groups: intra-arterial (IA), intravenous (IV), and intra-arterial plus intravenous (IA+IV) and non-tirofiban. The primary outcome was 90-day ordinal modified Rankin Scale (mRS) score, and the secondary outcomes included the rates of mRS 0–1, 0–2, and 0–3 at 90-day, successful recanalization. The safety outcomes were symptomatic intracranial hemorrhage (sICH) and other safety endpoints. The multivariable logistic regression models adjusting for potential baseline confounders were performed to compare the outcomes. A propensity score matching (PSM) with a 1:1:1:1 ratio was conducted among four groups, and the outcomes were then compared in the post-matched population. Results A total of 502 patients were included, 80 of which were in the IA-tirofiban group, 73 in IV-tirofiban, 181 in (IA+IV)-tirofiban group, and 168 in the non-tirofiban group. The median (IQR) 90-day mRS score in the four groups of IA, IV, IA+IV, and non-tirofiban was, respectively 3(0–5) vs. 1(0–4) vs. 1(0–4) vs. 3(0–5). The adjusted common odds ratio (OR) for 90-day ordinal modified Rankin Scale distribution with IA-tirofiban vs. non-tirofiban was 0.77 (95% CI, 0.45–1.30, P = 0.330), with IV-tirofiban vs. non-tirofiban was 1.36 (95% CI, 0.78–2.36, P = 0.276), and with (IA+IV)-tirofiban vs. non-tirofiban was 1.03 (95% CI, 0.64–1.64, P = 0.912). The adjusted OR for mRS 0–1 and mRS 0–2 at 90-day with IA-tirofiban vs. non-tirofiban was, respectively 0.51 (95% CI, 0.27–0.98, P = 0.042) and 0.50 (95% CI, 0.26–0.94, P = 0.033). The other outcomes of each group were similar with non-tirofiban group, all P was >0.05. After PSM, the common odds ratio (OR) for 90-day ordinal modified Rankin Scale distribution with IA-tirofiban vs. non-tirofiban was 0.41 (95% CI, 0.18–0.94, P = 0.036), and the OR for mRS 0–1 and mRS 0–2 at 90-day with IA-tirofiban vs. non-tirofiban was, respectively 0.28 (95% CI, 0.11–0.74, P = 0.011) and 0.25 (95% CI, 0.09–0.67, P = 0.006). Conclusions Intra-arterial administration of tirofiban was associated with worse outcome than non-tirofiban, which suggested that intra-arterial tirofiban had a harmful effect on patients undergoing EVT for ICAD-LVO. Clinical trial registration http://www.clinicaltrials.gov, Unique identifier: NCT03370939.

Background and purpose: Adjunctive tirofiban administration in patients undergoing endovascular treatment (EVT) for acute large vessel occlusion (LVO) has been investigated in several studies.However, the findings are conflict.This study aimed to compare the e ect of di erent administration pathways of tirofiban on patients undergoing EVT for acute LVO with intracranial atherosclerotic disease (ICAD).
Methods: Patients were selected from the ANGEL-ACT Registry (Endovascular Treatment Key Technique and Emergency Workflow Improvement of Acute Ischemic Stroke: A Prospective Multicenter Registry Study) and divided into four groups: intra-arterial (IA), intravenous (IV), and intra-arterial plus intravenous (IA+IV) and non-tirofiban.The primary outcome was -day ordinal modified Rankin Scale (mRS) score, and the secondary outcomes included the rates of mRS -, -, and -at -day, successful recanalization.The safety outcomes were symptomatic intracranial hemorrhage (sICH) and other safety endpoints.The multivariable logistic regression models adjusting for potential baseline confounders were performed to compare the outcomes.A propensity score matching (PSM) with a : : : ratio was conducted among four groups, and the outcomes were then compared in the post-matched population.
Results: A total of patients were included, of which were in the IAtirofiban group, in IV-tirofiban, in (IA+IV)-tirofiban group, and in the non-tirofiban group.The median (IQR) -day mRS score in the four groups of IA, IV, IA+IV, and non-tirofiban was, respectively ( -) vs. ( -) vs. ( -) vs.

Introduction
The benefit of tirofiban administration in patients with AIS undergoing mechanical thrombectomy is still unknow.The recently published randomized RESCUE BT (The Endovascular Treatment With vs. Without Tirofiban for Patients with Large Vessel Occlusion Stroke) trial explored the safety and efficacy of intravenous tirofiban in patients with acute anterior circulation LVO, this study didn't find that tirofiban could improve significantly outcomes compared with placebo, but suggested intravenous tirofiban could benefit the patients with large artery atherosclerosis (LAA) in subgroup analysis (1).Furthermore, in the post-hoc analysis of RESCUE BT trial, it was found that intravenous tirofiban was an effective adjunctive medication for patients with ICAD-related LVO undergoing EVT (2).In addition, it was reported that patients treated with intra-arterial tirofiban undergoing EVT for AIS suffered from an increased risk of symptomatic and fatal intracerebral hemorrhage (3).A recent meta-analysis also reported that treatment with tirofiban in patients with AIS undergoing EVT was effective in improving prognosis, particularly in patients with large atherosclerotic stroke, and intravenous administration of tirofiban improved more significantly clinical prognosis of patients than arterial administration (4).In practice, the dose and administration pathways of tirofiban are decided by interventionists at discretion, but the dose and pathway of tirofiban in the RESCUE BT trial were restricted, so whether the result of the RESCUE BT trial can be generalized to a broader population and clinical setting is still needed to be explored.Therefore, in this study, we aimed to explore the efficacy and the safety of different administration pathways of tirofiban on patients undergoing EVT for acute ICAD-related LVO based on the data of the ANGEL-ACT (a prospective nationwide registry study).

Study population
The patients were selected from the ANGEL-ACT (endovascular treatment key technique and emergency workflow improvement of acute ischemic stroke) registry, which was a nationwide prospective cohort including 1,793 consecutive adult patients with AIS undergoing EVT for LVO at 111 hospitals from 26 provinces in China between November 2017 and March 2019.The inclusion/exclusion criteria, data collection and methods of the ANGEL-ACT registry were reported in previous article (5).The exclusion criteria of this analysis were as follows: (1) EVT records unavailable; (2) No underlying ICAD after reopening of the occluded vessel or underlying ICAD could not be evaluated according to our criteria; (3) 90-day mRS missing; (4) Only intravenous bolus injection of tirofiban.Finally, we included 502 eligible patients in this analysis.The study protocol of ANGEL-ACT registry was approved by the Ethics Committees of Beijing Tiantan Hospital, Capital Medical University, and all participating centers, written informed consent were provided by the patients or their legally authorized representatives before the study enrollment.
Finally, 502 patients were divided into four groups.One hundred and sixty-two patients who did not receive any tirofiban treatment were assigned to the control (non-tirofiban) group, 80 patients who only received intraarterial tirofiban were assigned to the IA-tirofiban group, 73 patients who only received intravenous tirofiban were assigned to the IV-tirofiban group, and 181 patients who received both intraarterial and intravenous tirofiban were assigned to the (IA+IV)-tirofiban group.

Intervention of tirofiban
The decision of tirofiban treatment and administration pathway was at the discretion of interventionists.In general, tirofiban was administrated when interventionists encountered the following conditions: (1) Emergency stenting or balloon angioplasty for severe residual stenosis or instant re-occlusion; (2) Severe atherosclerosis disease in occlusive site with a high risk of early re-occlusion; (3) Successful mechanical recanalization with three or more passes with a stent retriever for presumed endothelial damage.The regimen of administration was not mandatory.It was generally recommended that, if necessary, tirofiban should be applied after recanalization, with local arterial administration through the guiding catheter.A low-dose intra-arterial bolus (400-500 µg) followed by a continuous intravenous infusion (300-480 µg/h) for 24 h was proposed as a standard administration and at 4 h prior to the end of the infusion, dual antiplatelet agents (aspirin 100 mg and clopidogrel 75 mg) were taken orally if ICH was excluded by computed tomography or magnetic resonance imaging.

Definition of ICAD-LVO
The ICAD-LVO was defined as occlusion located in the intracranial artery and was caused by acute in situ thrombus secondary to underlying ICAD.Underlying ICAD was defined as fixed residual stenosis degree >50% or stenosis with distal blood flow impairment or evidence of repeated re-occlusion, can also be determined according to previous images indicating stenotic lesion at the occlusion, at the same time vasospasm, dissection, vasculitis, or Moyamoya disease were excluded (10,11), and was diagnosed by the imaging core laboratory.

Outcomes measurement
The primary outcome was the 90-day ordinal mRS with scores ranging from 0 (no symptoms) to 6 (death).The secondary outcomes included the rates of mRS 0-1, 0-2, and 0-3 at 90 days, changes in the NIHSS score at 24 h and 7 days from baseline as assessed, rates of successful recanalization at final angiogram defined as modified Thrombolysis in Cerebral Infarction (mTICI) of 2b to 3 and complete recanalization defined mTICI of 3, and the pass numbers of thrombectomy.The safety outcomes were any ICH and symptomatic ICH within postprocedural 24 h according to the Heidelberg Bleeding Classification (12), intraprocedural embolization in procedure, and death within 90 days.

Statistical analysis
Continuous variables were presented as the median [interquartile range (IQR)] and categorical variables as a number (percentage).Comparisons of the baseline, procedural characteristics and outcomes among four groups were performed using the Kruskal-Wallis test for continuous and Pearson χ 2test for categorical variables.In the univariable analysis, the P-value < 0.05 indicated that variables were not entirely equal in the four groups and served as confounder.we performed ordinal/binary logistic regression or generalized linear models by adjusting for confounders to calculate the common odds ratios (OR), OR or β-coefficients with 95% confidence intervals (CI), comparing the clinical outcomes of the three groups of tirofiban with non-tirofiban.To reduce selection bias, a propensity score matching (PSM) was performed among the four groups.All variables with P < 0.05 in the univariable analysis were included to generate the propensity score.We matched four groups using a greedy-matching algorithm without replacement at a 1:1:1:1 ratio, with a caliper width ≤ 0.2 of the standard deviation of the logit of the propensity scores.Finally, we explored whether the effects of different administration Pathways of tirofiban on the primary outcome differed in certain subgroups by testing the administration pathway of tirofiban by subgroup interaction effect using an ordinal logistic regression model in the following subgroups: gender (female vs. male), age (age<65 vs. age ≥ 65), baseline NIHSS (<15 vs. ≥15), baseline ASPECTS (≤9 vs. =10), occlusion location (posterior circulation vs. anterior circulation), tandem lesions (yes vs. no), onset-to-puncture time (<6 vs. ≥6 h), pretreatment with IVT (yes vs. no), rescue balloon/stenting angioplasty (yes vs. no).Significance level was set to P = 0.05 (2-sided).We used SAS software v.9.4 (SAS Institute, Cary, NC, USA) to conduct the statistical analyses.

Results
We calculated that in 261 patients receiving IA-tirofiban, the median bolus dose of IA-tirofiban was 5.7 µg/kg, the interquartile range was 4.3-7.6 µg/kg and in 254 patients receiving IV-tirofiban, median intravenous infusion speed was 0.07 µg/kg/min, the interquartile range was 0.06-0.08µg/kg/min.Finally, 168 patients were assigned to non-tirofiban group, 80 people were assigned to the IA-tirofiban group, 73 patients were assigned to the IVtirofiban group, and 181 patients were assigned to the (IA+IV)tirofiban group.

Subgroup analysis
As shown in Table 4, there were significant interaction effects between administration pathway of tirofiban and onset-to-puncture time (p for interaction = 0.028) and rescue balloon/stenting angioplasty (P for interaction = 0.008) on the 90-day mRS score.Intra-arterial tirofiban tended to be associated with worse outcome of 90-day mRS (adjusted common OR, 0.46; 95% CI, 0.21-1.00)inpatients with onsetto-puncture time more than or equal 6 h, whereas not (adjusted common OR, 1.01; 95% CI, 0.45-2.27) in patients with onsetto-puncture time <6 h; intravenous tirofiban was associated with better outcome of 90-day mRS (adjusted common OR, 3.64; 95% CI, 1.57-8.43) in patients without receiving rescue balloon/stenting angioplasty, whereas not (adjusted common OR, 0.55; 95% CI, 0.25-1.22) in patients with receiving rescue balloon/stenting angioplasty.However, no interaction effect was found in other subgroups (all P for interaction >0.05;Table 4).

Discussion
In our study, we focused on patients with ICAD related LVO undergoing EVT, and we didn't find IA-tirofiban, IV-tirofiban, and (IA+IV) could significantly improve the primary outcome  of the 90-day ordinal mRS.However, we found IA-tirofiban decreased the rates of mRS 0-1 and mRS 0-2 at 90 d in both the prematched and postmatched population.After PSM, we found that IA-tirofiban was associated with more severe disability of the 90-day ordinal mRS.Furthermore, in subgroup analysis we found IA-tirofiban tended to be harm to the primary outcome of the 90day ordinal mRS in patients with onset-to-puncture time more than or equal 6 h and IV-tirofiban could improve the primary outcome of the 90-day ordinal mRS in patients without rescue balloon/stenting angioplasty.
Our study failed to find IV-tirofiban could improve prognosis of patients with ICAD related LVO, which isn't consistent with the subgroup analysis of the RESCUE BT trial (2).The following reasons were considered: (1) Our study was observational, although a multivariate logistic regression analysis was conducted, selective bias still existed in intravenous tirofiban and nontirofiban groups.(2) Because effect size was small, our sample size couldn't detect difference.(3) The dose of tirofiban in our study was much lower than RESCUE BT.Infusion median speed of tirofiban in our study was 0.07 µg/kg/min, but in RESCUE BT was 0.15 µg/kg/min.In addition, we didn't find (IV+IA)-tirofiban was sufficient to improve prognostic outcomes.Even though in initial univariable analysis, the rate of complete recanalization was the highest in (IV+IA) group, (IV+IA)tirofiban neither associated with complete recanalization nor successful recanalization after adjusting potential confounders.This finding was supported by a recent pool-analysis (4), and was inconsistent with previous study (13,14).Given that our studies were non-randomized controlled, further research will be needed to verify.
Although the risk of symptomatic ICH was not significantly different between the three tirofiban groups compared with the non-tirofiban group, we found IA-tirofiban could decreased the rates of mRS 0-1 and mRS 0-2 at 90 d in patients with ICADrelated LVO undergoing EVT.We speculated that there were several reasons for this.First, intra-arterial injection increases the local drug concentration of tirofiban which might aggravate damage to the blood-brain barrier and lead to intracerebral hemorrhage due to the presence of ischemic brain tissue.This viewpoint is supported by two studies (3,15), Wu et al. reported intra-arterial tirofiban administration increases risk of major ICH after endovascular thrombectomy for acute ischemic stroke, and doses more than 6.7 µ g/kg were associated with symptomatic and fatal intracranial hemorrhage (3).In our study, median bolus dose of IA-tirofiban was 5.7 µg/kg.Secondly, administration of IA-tirofiban before EVT might increase the risk of thrombus migration toward distal blood vessel, because the clot in ICAD-LVO is fresh and intra-arterial injection allows tirofiban directly contacts with the thrombus and might promote thrombus evolution.It was reported that IV-tPA (Tissue-Type  Plasminogen Activator) administration before EVT for LVO was associated with distal embolization, which in turn might reduce the chance that recanalization was achieved (16), so we speculated that IA-tirofiban had the same effect.Thirdly, arterial injection of tirofiban alone is not a continuous dose which might have a poor preventive effect on late in situ thrombosis and reocclusion (17)(18)(19).However, some study showed contrary conclusion that intra-arterial tirofiban didn't increase risk of ICH, even improved the clinical outcome of patients undergoing EVT (20).Up to now, there has been no randomized controlled trial about intra-arterial tirofiban administration in patients undergoing EVT.All studies were based on real world observational study.The safety of intra-arterial tirofiban treatment in patients with ICAD related LVO requires further randomized controlled trial to be verified.Therefore, intra-arterial tirofiban should be administrated cautiously during EVT, if necessary, a low dose may be more feasible.
In subgroup analyses, we found intra-arterial tirofiban tended to be associated with worse outcome of 90-day mRS in patient with onset-to-puncture time more than or equal 360 min.Possible explanation for this finding was that infarct grew, and vascular bed was destroyed more widely with a longer onset-to-puncture time, so IA-tirofiban administration increased risk of ICH.We also found intravenous tirofiban was associated with better outcome of 90-day mRS in patients without receiving rescue balloon/stenting angioplasty.It was reported that re-occlusion after an initial recanalization with SR thrombectomy in ICAS-related LVO was very frequent (65%) (21), the main causes for re-occlusion are residual plaque and platelet activation leading to thrombosis (22).If patients with ICAS-related LVO don't receive rescue balloon/stenting angioplasty, we believe that early and long-lasting intravenous tirofiban is preventive for the re-occlusion.
Our study has several limitations.First, this was not a randomized control trial, patients didn't have equal chance to enter each group, and measured and unmeasured variables still acted on the effect size, although we conducted a logistic regression to adjust for confounders.Second, there was no unified mandatory regime for tirofiban, the use of tirofiban was finally at the discretion of the treating physician and local practice in the present study, and different dose and period of the procedure may lead to different endpoint events.Third, we did not analyze the status of the perfusion, collateral, social background, economic situation, and genes of patients which are important factors for a good prognosis.Fourth, in our study underlying ICAD was defined as fixed stenosis degree >70% or stenosis >50% with distal blood flow impairment or evidence of repeated re-occlusion, this definition might mistake residual thrombus after thrombectomy as an intracranial atherosclerotic stenosis lesion.Last, our study population was limited to the Chinese population, which confined the generalizability of our results.

Conclusion
Administration of IA-tirofiban had a harm effect on patients undergoing EVT for ICAD-related LVO, especially in patients with onset-to-puncture time more than or equal 6 h instead of increasing the rates of complete recanalization and successful recanalization compared with non-tirofiban; administration of intravenous tirofiban could improve the prognosis of patients undergoing EVT for ICAD-related LVO without receiving rescue balloon/stenting angioplasty.
TABLE Baseline characteristics of di erent tirofiban groups. 2 missing data, ASPECTS for anterior circulation stroke and pc-ASPECTS for posterior circulation stroke.
TABLE Common OR or OR of safety and e cacy outcome according to di erent regimen of tirofiban.IA, intra-arterial; IV, intravenous; mRS, modified Rankin Scale; IQR, inter quartile range; NIHSS, National Institutes of Health Stroke Scale; ICH, intracranial hemorrhage.Adjusting for confounders including Male, Hypertension, Hyperlipidemia, Atrial fibrillation, Smoking history, admission systolic blood pressure in admission, admission ASPECTS, Occlusion sites, onset-to-door time, door to-puncture time, onset-to-recanalization time, onset-to-puncture time, Prior use of antiplatelet agents, IAT, and Rescue balloon/stenting angioplasty.31missing data, the β-coefficients were calculated using a generalized linear model.TABLE Subgroup analysis regarding -day mRS of di erent tirofiban groups.
* 39 missing data, the β-coefficients were calculated using a generalized linear model.?? 13 missing data.?! 15 missing data.IA, intra-arterial; IV, intravenous; NIHSS, National Institutes of Health Stroke Scale; ASPECTS, Alberta Stroke Program Early CT Score; IVT, intravenous thrombolysis.*Adjusting for confounders including Male, Hypertension, Hyperlipidemia, Atrial fibrillation, Smoking history, admission systolic blood pressure in admission, admission ASPECTS, Occlusion sites, onset-to-door time, door to-puncture time, onset-to-recanalization time, onset-to-puncture time, Prior use of antiplatelet agents, IAT, and Rescue balloon/stenting angioplasty.