Safety and efficacy of tirofiban in the management of stroke: A systematic review and meta-analysis of randomized controlled trials

Background: About 30 % of stroke patients have experienced unsuccessful reperfusion following endovascular therapy. Mechanical thrombectomy instruments may contribute to this by stimulating platelet aggregation. Tirofiban is a selective and rapidly activated antagonist of the platelets nonpeptide glycoprotein IIb/IIIa receptors that can reversibly suppress platelet aggregation. But, data from the medical literature are conflicting regarding its safety and efficacy for stroke patients. Hence, this study was designed to assess the safety and efficacy of tirofiban in stroke patients. Methods: Five major databases (PubMed, Scopus, Web of Science, Embase, and Cochrane library) were searched till December 2022. The Cochrane tool was used for risk of bias assessment, and the RevMan 5.4 was utilized for data analysis. Results: Seven RCTs with 2088 stroke patients were included. Tirofiban significantly increased the number of patients with mRS 0 score after 90 days than control; RR = 1.39, 95 %CI [1.15, 1.69]; p = 0.0006. Additionally, it reduced the NIHSS score after seven days; MD = (cid:0) 0.60, 95 %CI [ (cid:0) 1.14, (cid:0) 0.06]; p = 0.03. However, tirofiban increased the incidence of intracranial haemorrhage (ICH); RR = 1.22, 95 %CI [1.03, 1.44]; p = 0.02. Other assessed outcomes showed insignificant results. Conclusions: Tirofiban was associated with a higher mRS 0 score after three months and a lower NIHSS score after seven days. However, it is associated with higher ICH. Multicentric trials are required to provide more convincing proof of its utility.


Introduction
Acute ischemic stroke (AIS) represents the leading global cause of death and disability [1]. Currently, it is well established that individuals with AIS benefit from endovascular therapy (EVT) and intravenous thrombolysis (IVT) [2]. However, about 30 % of patients have experienced unsuccessful reperfusion following EVT [3]. Using IVT with alteplase before EVT is advised for such cases. But some patients may still experience re-occlusion [4]. Mechanical thrombectomy instruments may contribute to failed reperfusion by damaging the vascular endothelium and exposing the subendothelial matrix, which can initiate platelet activation and adhesion followed by aggregation and perhaps lead to re-occlusion and thromboembolic consequences [5,6].
Glycoprotein IIb/IIIa inhibitors are a group of highly specific platelet antagonists. They successfully and reversibly disrupt the fibrin binding to receptors preventing platelets aggregation [7]. Experimental data from glycoprotein IIb/IIIa inhibitors-treated stroke animals showed that infarct volume might be reduced even after delayed treatment [8].
Tirofiban is an inhibitor of the platelets nonpeptide glycoprotein IIb/ IIIa. It has excellent selectivity, rapid activation, deactivation, and a short half-life that has the ability to prevent platelet aggregation in a reversible manner. Using tirofiban has decreased the likelihood of thrombotic events through percutaneous coronary intervention [9]. Because of its benefits for the management of acute coronary syndromes, several studies have examined tirofiban effectiveness as an adjunct treatment for individuals undergoing endovascular therapy for stroke [10][11][12]. Tirofiban was linked to neurological function improvement 90 days following AIS [13]. Intravenous tirofiban significantly enhanced the clinical outcomes in progressive ischemic stroke patients [14][15][16]. According to many reports, individuals with AIS receiving endovascular therapy may benefit from continuous intravenous tirofiban therapy in terms of functional outcomes and mortality reduction with no bleeding risk [17]. Moreover, it outperformed the dual anti-platelet regimens [16]. Tirofiban has been reported as a safe agent for AIS; [18,19] however, it has been linked to an increased incidence of fatal ICH [20]. So, it is crucial to investigate the safety of tirofiban besides its efficacy for stroke patients.
Medical literature still shows controversy regarding the tirofiban effect for stroke. The available evidence is derived from observational, nonrandomized, or retrospective studies with few randomized controlled trials (RCTs). Hence, we conduct this systematic review of the RCTs evaluating tirofiban in stroke patients aiming to provide reliable evidence regarding its safety and efficacy.

Methods
This study was carried out in concordance with the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines [21] and the Cochrane handbook of systematic reviews [22].

Searching databases and data collection
Data were retrieved from five databases (PubMed, Web of Science, Scopus, EMBASE, and Cochrane) till December 2022. We used the following keywords: tirofiban, aggrastat, stroke, cerebrovascular accident, cerebrovascular apoplexy, and brain vascular accident.

Eligibility criteria
Any study that achieved the following criteria was included: participants: stroke patients undergoing regular treatment or endovascular thrombectomy; Intervention: tirofiban; Comparator: placebo of regular treatment; Outcomes: efficacy and safety outcomes; Study design: RCTs.

Studies screening
Studies collected from the databases search were sent to an Excel workbook using EndNote X8.0.1 version. The collected studies were screened using a three-step screening based on the eligibility criteria. First, the studies' titles and abstracts were screened. Second, the relevant trials' full text was screened for eligibility. Finally, manual screening of references from the included trials was performed. Two different authors performed each step. Another author solved any disparities.

Data extraction
Following study screening, the applicable studies' data were extracted. This included data of two major categories: (1) general characteristics of the involved trials and patients (2) data on the planned outcomes. This process was performed by two independent authors. Any discrepancies were solved by a third one.
The summary of the included studies comprised; the site of study conduction, study registration number, their inclusion criteria, intervention details, primary outcomes, and conclusion. In addition, the baseline characteristics of the included participant were extracted. This included groups, sample size, age, sex, stroke aetiology, and risk factors.
The efficacy outcomes involved modified Rankin scale (mRS) scores  after 90 days and National Institutes of Health Stroke Scale (NIHSS) score changes after 24 h and seven days. The safety outcomes comprised the frequency of symptomatic intracranial haemorrhage (sICH), any ICH, and deaths.

Quality assessment
Cochrane's risk of bias tool was utilized to assess the RCTs quality [22]. The tool appraises seven important domains 1) the randomization process; 2) the allocation of patients to each group; 3) participants and personnel blinding; 4) blinding of the outcome assessors; 5) attrition bias; 6) reporting bias; 7) other bias. Two separate authors evaluated the studies' quality. A third person was brought in to settle any disagreements. Also, we assessed the methodological quality using the JADAD scale. The certainty level of evidence was assessed using GRADE.

Data Analysis
Data analysis was done by Review Manager Software (RevMan 5.4.1). The mean difference (MD) and 95 % confidence interval (CI) were used to evaluate the continuous data. The risk ratio (RR) and 95 % CI were used for dichotomous data. Homogeneous outcomes were studied using the fixed-effect model. In contrast, heterogeneous outcomes were performed under the random-effect model. Chi-square tests NR: not reported, NIHSS: national institutes of health stroke scale, AIS-ND: acute ischemic stroke with neurological deterioration, IVT: intravenous thrombolysis. and the I 2 index [22] were used to measure the heterogeneity. Heterogeneous outcomes were identified when p˂ 0.1 with I 2 > 50 %. For heterogeneous outcomes, Cochrane's leave-one-out technique was used to solve it [22]. The significant outcomes were identified as p˂ 0.05. The publication bias was assessed by visual inspection and the p-value of Egger's test.

Summary of results of database search and study selection
After duplicate removal, 1873 articles were handled from various databases searched and screened for eligibility. Only 32 studies underwent full-text screening. According to the eligibility criteria, seven trials were included in our systematic review [14,15,[23][24][25][26][27], while only six of them [14,15,[24][25][26][27] were eligible for meta-analysis. Fig. 1 shows the PRISMA flow diagram of the database search and study selection.

General characteristics of included participants
We included 2088 patients from seven RCTs (of them, 1057 were treated with tirofiban while 1031 received the control of regular treatment regimen or endovascular thromboembolectomy). The involved patients showed male predominance (61 %) with a mean age of 65.6 years old. Tables 1 and 2 show the summary of studies and participants' characteristics.

Quality assessment results
The final quality rating was at a low risk of bias. Regarding selection bias, all studies had low bias except for three trials [14,23,27] that did not report enough data regarding the randomization process. The same was for allocation concealment. Two trials [14,27] were at unclear risk. As for the blinding, only two trials [23,25] were at low risk, while the others were at high risk or unclear. All RCTs were at a low risk of attrition bias, while four trials [15,[25][26][27] were at a low risk of reporting bias. Fig. 2 summarizes the results of the quality assessment process. Using JADAD, only two studies were scored as low quality, while others were high quality. Table S1 3.4. Efficacy outcomes
Six trials reported the NIHSS change after seven days [14,15,[24][25][26][27]. Tirofiban was associated with a significant decrease after seven days compared with the control groups; MD= − 0.90, 95 %CI [− 1.63, − 0.17]; p = 0.02. The data showed heterogeneity; p = 0.02, I 2 = 64 %. Fig. 6 By visual inspection and the p-value of Egger's test, the results showed no publication bias among all subgroups (p > 0.05). Fig. S7 and 8 Using GRADE, the certainty level of the NIHSS change after 24 h was very low, while the NIHSS change after seven days was important. Table S2 3.5. Safety outcomes
The results showed no publication bias by visual inspection and the p-value of Egger's test (p = 0.18). Fig. S9 Using GRADE, the certainty level of any ICH outcome was high. Table S2 3. 5

.2. Symptomatic intracranial haemorrhage (sICH)
Three trials reported this outcome [15,25,26]. No difference was noticed between tirofiban and the control groups; RR= 1.5, 95 %CI [0.97, 2.32]; p = 0.07. The data were homogenous; p = 0.78, I 2 = 0 %. Fig. 8 We could not assess the publication bias as the included studies were three in this outcome. Using GRADE, the certainty level of the sICH outcome was moderate. Table S2 3. 5
The results of NIHSS after 24 h and seven days were homogeneous by excluding Han et al; [15]. Regarding death outcome, the results become homogeneous by excluding Shuai et al [25]. (p = 0.63, I 2 = 0 %). The pooled effect estimate revealed that the control group had considerably more deaths than the tirofiban group.; 18 vs 4 patients in each group respectively; RR= 0.23, 95 %CI[0.08, 0.66]; p = 0.007. Fig. S15

Quantitative synthesis
Bai et al [23]. conducted a double-blinded multicentric RCT on 66 stroke patients with mild to severe motor damage. Half of them received tirofiban at 0.1 g/kg body weight with slaine over 30 min or more, while the control group received only IV saline infusion for 21 days in addition to the rehabilitation course of treatment for both groups. Their findings imply that tirofiban, along with rehabilitation therapy, has a high safety profile and encourages additional motor function regeneration in stroke patients whose motor function has been severely compromised. They revealed that tirofiban significantly improved motor function compared with the control p˂ 0.05. Tirofiban also significantly improved patients with severe motor damage; p˂ ˂ 0.01. Only one patient from each group exhibited a serious adverse event (SAE). However, they reported that these SAEs were not linked to their medical management. In the tirofiban group, SAE was represented by cholecystitis triggered by cholelith, while the in the placebo group, it was the hemorrhagic

Discussion
This meta-analysis of 2088 patients from seven RCTs shows that tirofiban tends to decrease the mRS score after 90 days. It significantly increased the number of patients with the mRS 0 scores. Additionally, it reduced the NIHSS score after seven days. However, there were insignificant results regarding NIHSS after 24 h, combined mRS score, and mRS scores from 1 to 6. Notably, tirofiban showed higher any ICH and did not significantly differ from the control regarding sICH, although it was slightly higher than the control group. As for mortality, there were insignificant results regarding the mortality rates, although it was slightly lower with tirofiban. After sensitivity analysis, tirofiban showed a significantly lower incidence of deaths within 90 days.
Early IV infusion of tirofiban following alteplase injection for at least 24 h notably lowered the rates of re-occlusion, NIHSS after seven days, and mRS scores after 90 days [28]. In a recent meta-analysis, Liu et al [29]. estimated the influence of early administration of tirofiban in acute ischemic stroke patients following the IVT plus or minus mechanical EVT. Tirofiban exhibited excellent functional outcomes in terms of mRS 0-1. It also enhanced mRS 0-2. However, results differed according to the management plan, either IVT or IVT bridging followed by the EVT. Tirofiban following IVT alone was associated with significantly better functional outcomes. However, there was an insignificant difference in the subgroup of IVT bridging followed by EVT.
Another meta-analysis of 2028 stroke patients receiving EVT reported insignificant results regarding the favorable functional outcomes (defined as mRS of 0-1, 0-2, or 0-3 according to the included studies). It also did not increase the recanalization rates compared with the control [30]. Additionally, in a meta-analysis of 3251 patients, tirofiban increased the number of patients with mRS 0-2 score than the control; p = 0.002. There was an increase in recanalization in the tirofiban group compared to the placebo group [31]. However, a meta-analysis by Gong et al [18]. revealed no improvement in mRS 0-2 and recanalization rates with the tirofiban group.
Glycoprotein IIb/IIIa comprises a group of anti-platelet agents. They have been broadly utilized in patients with high-risk myocardial infarction receiving IVT or percutaneous coronary intervention to stop platelet aggregation and subsequent thrombus formation [32]. However, the risk of hemorrhagic transformation and the disease's diverse pathophysiology prevented immediate utilization of glycoprotein IIb/IIIa inhibitors in stroke despite the success demonstrated in cardiovascular treatment [33]. According to the 2018 guidelines, IV thrombolysis should not be used concurrently with anti-platelet medications [34]. This is evidenced by outcomes of adding anti-platelet drugs such as aspirin, eptifibatide, or abciximab, in addition to the lack of convincing evidence that early anti-platelet therapy improves outcomes. So, assessment of the safety of such agents is critically important.
It has been reported that the combination of tirofiban and alteplase (thrombolytic agent) significantly decreased the mortality probability with no risk of developing sICH in acute ischemic cerebral infarction patients [24]. Also, Huo et al [35]. reported no difference between the two cohorts regarding the incidence of any ICH, sICH, or distal embolization. But it reduced the long-term mortality rates following EVT bridged by IVT. On the other hand, Kellert et al [20]. displayed a higher fatal ICH in the tirofiban group, with a higher risk for those with a stroke of the anterior circulation. Moreover, they reported that age, in addition to tirofiban treatment, was identified as an independent predictor for developing fatal ICH and poor outcomes.
Our meta-analysis with pooled effect estimate solved this controversy by pooling data from high-evidence RCTs rather than observational, non-randomized studies. We report that tirofiban did not raise the risk of sICH or death within 90 days. However, it was associated with higher radiological ICH than the controls (207/952 vs 175/934; p = 0.02). These findings were in concordance with the findings of the meta-analysis of Gong et al [18]., which reported that tirofiban did not raise the incidence of sICH or deaths. But it significantly raised the fatal ICH, especially with the intra-arterial route. Our findings differed from that of meta-analyses by Liu et al [29]. and Tang et al [31]. Both meta-analyses found that tirofiban reduced mortality with no increase of sICH or any ICH. Fu et al [30]. revealed similar findings without reporting the incidence of any ICH. Zhou et al [36]. divided the included studies into a group in which tirofiban was used alone (monotherapy) and another group in which tirofiban was combined with IVT. They revealed that neither group was associated with higher ICH, sICH, and mortality risks.
Notably, all the previously published meta-analyses pooled data from RCTs, retrospective, non-randomized trials, or observational studies. We included only results from RCTs which represents the highest evidence. All of the included studies were conducted in China except for Siebler et al [24]., which was conducted in Germany. This makes it harder to generalize our findings for stroke patients worldwide. Some outcomes showed heterogeneity between the included trials' findings. But, this was solved with sensitivity analysis reported by the Cochrane handbook for systematic review and meta-analysis [22]. The heterogeneity could be attributed to the different general and clinical characteristics of the involved patients, such as differences in the age groups between certain studies, the aetiology of the stroke, the times of intervention, or the dosage of tirofiban.

Conclusion
Tirofiban significantly increased the number of patients with an mRS 0 score and reduced the NIHSS score after seven days. However, the tirofiban group did not differ significantly from the controls for the NIHSS after 24 h, the mean total mRS score, and the mRS scores from 1 to 6. Tirofiban did not significantly increase the incidence of sICH or mortality. However, it increased the incidence of any ICH.

Funding
None.

Declaration of Competing Interest
The authors declare no conflict of interest.