Thromboembolic complications that develop during endovascular coil embolization to treat cerebral aneurysms are the most serious and dangerous ones because they can result in serious neurological sequelae6,8,9,24). Their incidence has been reported as approximately 6–7%3,25). There are several methods for dissolving thromboembolism, including intravenous (IV) heparin, IV aspirin, thrombolysis using urokinase or tissue plasminogen activator, and mechanical thrombectomy. These methods all have attendant risks, including accompanying hemorrhage, so the medications used should have easy applications, have a short half-life, confer low bleeding tendency, and so forth. Recently, glycoprotein (GP) IIb/IIIa inhibitors have been used to treat thromboembolism that develops during endovascular coil embolization for cerebral aneurysm because they may fulfill these criteria1).

There are 2 types of GP IIb/IIIa inhibitors: reversible ones (tirofiban and eptifibatide) and an irreversible one (abciximab). In treating thromboembolism, 2 infusion routes are available for GP IIb/IIIa inhibitors: intra-arterial (IA) and IV. Systemic IV administration of tirofiban was the first method introduced, and its efficacy has been proven in patients with myocardial infarction. The IA administration of tirofiban was introduced later, its advantages being the focal elevation of tirofiban concentration at the lesion, decreased total dose of tirofiban and low incidence of hemorrhage. However, Brinjikji et al.3) reported no differences in terms of clinical outcomes and complication rates between the 2 administration routes.

We combined the 2 routes of administration, IA and IV infusion, of the reversible GP IIb/IIIa inhibitor, tirofiban to study their advantages and disadvantages for elevating focal concentration and systemic concentration of the drug at the same time. Here, we present the effectiveness and safety of combining infusion routes of tirofiban for treating thromboembolism during endovascular coil embolization.

Clinical and radiographical characteristics are described in Table 2. The postprocedural mRS score on day 1 showed no change from preprocedural mRS scores, except for 1 case (in which score changed from 2 to 4). In this patient, stent-assisted endovascular coil embolization of a right anterior cerebral artery, A2–3 aneurysm was performed, but severe vasospasm and thrombus occurred in right A2. Therefore, it was not possible to avoid the patient’s cerebral infarction and neurological deterioration.

The most common origin of thrombus formation was coil protrusions, seen in 15 cases (75%). These cases comprised 12 cases (60%) of simple coil protrusion and 3 cases (15%) of stent-assisted coil protrusion. The stent was the cause in 4 cases (20%), 2 of which had double stents (2 cases), 1 of which had a small parent vessel, and 1 of which had severe vasospasm. Microcatheter displacement caused thrombus formation in 1 patient (5%). Embolism from carotid plaque was the cause of thrombus formation in 1 patient, occurring in the presence of simultaneous coil protrusion.

Thromboembolism developed during endovascular coil embolization in 20 out of 264 cases of aneurysm (7.57%). Of these 20 cases, 5 were TICI grade 0, 2 were TICI grade 1, 7 were TICI grade 2a, and 6 were TICI grade 2b. In the 5 cases that were TICI grade 0, 1 achieved TICI grade 2a after recanalization, and 4 cases improved to TICI grade 3 after recanalization (Fig. 1, Table 3). Two cases that were TICI grade 1 achieved TICI grade 3 after recanalization (Fig. 2), and 13 cases that were TICI grade 2a, and 2b achieved TICI grade 3 after recanalization. The patient with a simultaneous embolus from a carotid bulb plaque improved from TICI grade 1 to grade 2 after recanalization (Table 3). Delayed irrelevant intracerebral hemorrhage was noted in 1 patient (5%) two weeks after the procedure, and thromboemboli-related (or procedure-related) cerebral infarction developed in 5 patients (25%), of which only 1 patient was symptomatic.

Thromboembolic events are the most common complications during endovascular coil embolization of cerebral aneurysm and have proven challenging to neurointerventionists. Possible causes of thrombus formation are flow disturbances and the effects of the procedure itself. Flow disturbances can occur at the coil surface or the interface between coil and vessel wall; they can be caused by an irregular coil surface or coil protrusions into the vessel lumen, resulting in vessel narrowing or vasospasm. Procedure-related causes include intimal injury from microcatheter displacement or stent deployment, in-stent thrombus or migration of a preexisting intra-aneurysmal thrombus4,5). In our patients, the most common cause of thrombus formation was coil protrusions (simple coil protrusion: 60%, stent-assisted coil protrusion: 15%), findings that are consistent with previous reports4,5). Thromboembolic events also have been associated with wide-necked aneurysms20,24). In our study, the mean dome-to-neck ratio of the aneurysms was 1.56±0.89 (0.73–4.71), and 16 patients had a wide-necked aneurysm (<2.0 dome-to-neck ratio), among which thromboembolic events occurred in four patients.

Two classes of medications are used for treating thromboembolism during endovascular coil embolization: thrombolytic agents (a tissue plasminogen activator or urokinase) and GP IIb/IIIa inhibitors. Cronqvist et al.7) reported on the outcomes of thrombolysis using urokinase in 19 patients and combining thrombolysis with mechanical thrombus fragmentation in 9 patients. In their report, results as seen on angiography were superb: complete recanalization was achieved in 10 patients, and partial recanalization was achieved in 9 patients. Clinical outcomes were also positive: a good clinical outcome was obtained in 14 patients, of whom 9 showed complete recanalization and 5 showed partial recanalization. However, significant hemorrhagic complications occurred in 3 patients. Koebbe et al.16) previously reported on the outcome of thrombolysis using urokinase or tissue plasminogen activator in 5 patients, 2 of whom experienced significant subarachnoid hemorrhage. In these studies, although thrombolysis using urokinase or tissue plasminogen activators effectively led to vessel recanalization, the rates of significant hemorrhage were high.

Hyperacute thrombi that develop during endovascular coil embolization consist of platelet aggregates that are not yet stabilized by fibrin cross-linking. Theoretically, GP IIb/IIIa inhibitors reversibly antagonize fibrinogen binding to the GP IIb/IIIa receptors of platelets and inhibit platelet aggregation in order to prevent thrombi from forming and propagating. Therefore, many clinicians try to dissolve thrombi during endovascular coil embolization using GP IIb/IIIa inhibitors. Fiorella et al.10) reported on 13 patients with thromboembolism during endovascular embolization who were treated with a GP IIb/IIIa inhibitor (abciximab). Although tissue plasminogen activator was simultaneously administered to 5 of these patients, 7 patients achieved complete recanalization, 6 patients achieved partial recanalization, and no new hemorrhagic complications developed. In another report, 52 of 54 patients showed complete or partial dissolution of their thrombi, and no new hemorrhagic complications developed11). In a recent meta-analysis by Brinjikji et al.2,3), the GP IIb/IIIa inhibitors abciximab, tirofiban, and eptifibatide were shown to be superior to thrombolytics in terms of recanalization rate (72% vs. 50%), perioperative morbidity (11% vs. 29%) and long-term morbidity (16% vs. 35%).

Abciximab is a monoclonal antibody with a large molecular weight. It binds irreversibly to the GP IIb/IIIa receptor at the β-chain of the integrin. Tirofiban is a nonpeptide tyrosine derivative with a small molecular weight based on a disintegrin polypeptide. Tirofiban reversibly and competitively antagonizes fibrinogen and von Willebrand factor binding to the GP IIb/IIIa receptors of platelets, and thereby inhibits platelet aggregation. Correspondingly, tirofiban’s potency for dissolving thromboemboli is inferior to abciximab’s, but its half-life is significantly shorter (1.5–2.0 hours for tirofiban vs. ≤48 hours for abciximab)11,25). In addition, with tirofiban, bleeding time normalizes within 4 hours after discontinuation compared with 12 hours for abciximab11).

Many authors have reported that abciximab is safe to use as a rescue treatment for intraprocedural/periprocedural thrombi that develop during endovascular embolization10,18,19,21). However, in theory, the affinity of abciximab to GP IIb/IIIa receptors is higher than tirofiban’s, and the half-life of tirofiban is usually much shorter than that of abciximab. Thus, the incidence of perioperative or postoperative hemorrhage is likely to be higher in patients treated with abciximab. Some authors have expressed concerns that abciximab could result in significant hemorrhagic complications22,23,25); other authors preferred tirofiban for treating patients who needed decompressive surgery and extraventricular drainage after undergoing endovascular coil embolization of cerebral aneurysm4). In contrast, Jeong and Jin14) reported that both abciximab and tirofiban exhibited similar safety and vessel recanalization rates. Recently, the meta-analysis by Brinjikji et al.2,3) compared patients receiving abciximab with those receiving tirofiban (and eptifibatide). In their report, patients receiving tirofiban (and eptifibatide) had significantly higher recanalization rates (83% for tirofiban vs. 66% for abciximab; p=0.05), but there was no difference in clinical outcomes between the 2 groups (postoperative hemorrhage rate; 14% for tirofiban vs. 7% in abciximab; p=0.12)3).

Tirofiban has been used in patients with acute myocardial infarction, usually with 1 of 2 regimens. The first regimen is an IV loading dose of 0.4 μg/min/kg for 30 minutes followed by an IV maintenance dose of 0.1 μg/min/kg for over 48 hours. The second regimen is an IV loading dose of 25 μg/kg for 3–5 minutes followed by an IV maintenance dose of 0.15 μg/min/kg for more than 48 hours17). When tirofiban began to be used for treating thromboemboli during endovascular coiling, the first regimen was used: an IV loading dose of 0.4 μg/min/kg for 30 minutes followed by an IV maintenance dose of 0.1 μg/min/kg for more than 48 hours. This regimen proved effective for dissolving thromboemboli during endovascular coil embolization4).

Recently, another regimen was introduced for treating thromboemboli during endovascular coiling. Tirofiban was repeatedly given in a loading dose (~4–8 μg/kg or 300 μg) via the IA route of administration for several minutes until the thromboemboli were dissolved, as confirmed on repeated cerebral angiography. IV maintenance therapy was then applied in a few cases. It is reported that this regimen was possibly able to elevate the focal concentration of tirofiban near the thromboemboli and thereby reduce the total dosage of tirofiban5,13,15). Cho et al.5) reported 87.2% of recanalization rates and no postoperative hemorrhages with IA tirofiban doses of 0.25 to 1.25 mg and Jeon et al.13) reported 80% of recanalization rates and no postoperative hemorrhages with IA tirofiban doses of 0.3 to 1.2 mg.

Many authors believe that both IA and IV administration of GP IIb/IIIa inhibitors is effective for dissolving thrombi during endovascular coil embolization. However, some authors believe that IA administration is better than the IV route for 2 reasons. First, the IA route can elevate the focal concentration of the drug with its rapid-dissolving action. Second, this dissolving action can be achieved with a relatively low dose of the drug and leads to a low rate of hemorrhagic complications11). Fiorella reported that the time it took for GP IIb/IIIa inhibitors to dissolve thrombi was less than 5 minutes via the IA route, but more than 10 minutes for the IV route10). However, the meta-analysis by Brinjikji et al.3) compared patients receiving GP IIb/IIIa inhibitors by IA route with those by IV route. In their report, there was no difference in recanalizaton rates (77% for IA route vs. 70% for IV route; p=0.36) and postoperative hemorrhage rates (5% for IA route vs. 7% for IV route; p=0.66) between the 2 groups.

Although the elevation of the focal concentration of tirofiban may be help for dissolving thrombus, we believed that tirofiban will operate above some systemic concentration like other drugs and we doubted that IA administration of the drug could focally elevate the drug’s concentration, resulting in more effective thrombus dissolution. If this were true, we reasoned, then such a focal elevation of the drug’s concentration could be dangerous and cause hemorrhagic lesions, such as a ruptured aneurysm. In addition, the repeated IA loading methods depends somewhat on the clinician’s experience, specially in point of drug dosage because the drug dosage is not specific. We would like to choose more standardized methods specially in terms of dosage and administration routes.

Because of this reason, we combined the 2 existing administration routes. We performed simultaneous IA and IV loading with approximately 80% (5 μg/kg, respectively) of the standard recommended dosage (i.e., 12 μg/kg) of tirofiban used in myocardial infarction patients for 3–5 minutes. Subsequently, IV maintenance therapy continued with 80% (0.08 μg/min/kg) of the standard recommended dosage (i.e., 0.1 μg/min/kg) used in myocardial infarction patients, for a maximum of 24 hours. With this method, we hoped to achieve focal and systemic elevation of tirofiban, good efficacy and fewer hemorrhagic complications. In our studies, 19 patients (95%) had their artery recanalized to TICI grade 3 from grade 0–2, with no related hemorrhagic complications. Five patients (25%) developed a thromboemboli-related cerebral infarction, 1 of whom was symptomatic.

In our study, 5 patients with ruptured cerebral aneurysm required mechanical devices to assist in thrombectomy. In 4 patients (including the embolic case), a microcatheter and microwire was used, and in 1 patient, a Solitaire stent was used. Of 3 patients who underwent thrombectomy assisted by microcatheter and microwire, preprocedure TICI grade 0 improved to TICI grade 2a after recanalization in 1 patient (who experienced severe vasospasm) and to TICI grade 3 in the other 2 patients. The fourth patient assisted by microcatheter and microwire had a distal embolism from carotid plaque that was TICI grade 1, but after recanalization of the vessel, the vessel improved to TICI grade 2a. In the patient who underwent thrombectomy with a Solitaire stent, the stent was originally intended to protect against coil protrusion, but it had to be retrieved because of a growing thrombus in spite of IA/IV tirofiban infusion and IV maintenance therapy lasting approximately 20 minutes. This patient’s vessel was recanalized, improving from TICI grade 1 to 3. Correspondingly, in 3 out of 5 patients with preprocedure TICI grade 0, and 2 out of 3 patients with preprocedure TICI grade 1, mechanical devices were used and the outcomes were relatively positive. Therefore, mechanical device-assisted thrombectomy could be a good adjunct option for dissolving thrombi during endovascular coil embolization, especially in patients with lower TICI grades. We decided to use mechanical devices within 20 minutes following tirofiban infusion, after obtaining serial angiograms at approximately 5-minute intervals because we believed that thrombus-dissolving action of tirofiban could be noted within several minutes. Thus, during the approximately 20-minute period, if the thrombus showed no changes or enlarged, we decided to use the mechanical devices for thrombectomy.

The limitations of this study were its retrospective analysis design and its small patient population. In addition, we could not explain the exact mechanisms of simultaneous IA and IV administration of tirofiban that were advantageous. However, in line with our results, simultaneous IA and IV loading and maintenance therapy appears to be a viable option for dissolving thrombi during endovascular coil embolization.

Thromboembolic events during endovascular coil embolization of cerebral aneurysm have been a challenge to neurointerventionists. Many pharmacological agents have been introduced to treat thromboembolism, and the outcomes appear promising. We found that in patients who develop thromboembolism during endovascular coil embolization of ruptured or unruptured cerebral aneurysms, administering a simultaneous IA and IV loading dose of tirofiban, followed by IV maintenance therapy, is a promising rescue treatment with good recanalization rates and low rates of bleeding complications.