Periprocedural Morbidity and Mortality Associated with Endovascular Treatment of Intracranial Aneurysms

Procedural Complications for Ruptured Aneurysms

Procedure-related complications for initial coil embolizations are summarized in Table 3. We recorded no technical failures. Twenty-seven (22.9%) procedural complications occurred with 118 procedures: 10 had no clinical consequences, one was associated with transient neurologic changes, seven resulted in persistent neurologic morbidity, and nine resulted in death. The procedure-related neurologic morbidity rate on discharge was 5.9%, and the mortality rate was 7.6% (13.5% combined).

Intraprocedural Aneurysmal Rupture

Intraprocedural aneurysmal rupture occurred in nine (7.6%) of 118 ruptured aneurysms due to protrusion of a tip of the coil (n = 6) or microcatheter (n = 3) outside the limit of the aneurysmal sac during the procedure. (Seven perforations were accompanied by angiographically visible extravasation of contrast agent.) Eight perforated aneurysms were small (3–8 mm), and one was large (12 mm). All ruptures occurred during the early stage of the procedure before detachment of the halfway coil. Coil-induced ruptures were caused by the first coil in three small (4–8 mm) aneurysms, the third coil in two small (6–8 mm) aneurysms, and the 11th coil in one large aneurysm (12 mm). Microcatheter-induced ruptures occurred in three small (4–5 mm) aneurysms during initial catheter placement in one case and during deployment of the second coil in two cases. All nine intraprocedural ruptures occurred during embolization of acutely ruptured aneurysms: four in the anterior communicating artery, one in the anterior choroidal artery, one in the middle cerebral artery bifurcation, one in the basilar tip, one in the basilar trunk, and one in the vertebrobasilar confluence. Preoperative clinical grades of the nine lesions were Hunt and Hess grade I in one, grade II in two, grade III in three, and grade IV in three.

All intraprocedural ruptures were managed with a systemic protamine injection for heparin reversal and further coil embolization. In three cases, one of the interventionalists (M.H.) inserted ventricular drainage catheters in the angiography suite immediately after aneurysmal perforation, and in five cases, a ventricular catheter had been placed before coil placement.

All three deaths were associated with intraprocedural ruptures that occurred before detachment of the first coil and with initial Hunt-Hess grades of III (n = 1) or IV (n = 2). Six intraprocedural ruptures had no clinical consequence, and three resulted in death. In 118 procedures, the morbidity rate from intraprocedural rupture of an already ruptured aneurysm was 0%, and the mortality rate was 2.5%. Overall morbidity and mortality rates at discharge from intraprocedural aneurysm rupture in 210 cases were 0% and 1.4%, respectively.

Periprocedural Thromboembolism

Thromboembolic complications occurred in 13 (11.0%) of 118 procedures; one resulted in no neurologic consequence, one resulted in transient neurologic deficit (expressive aphasia), seven resulted in persistent neurologic abnormalities on discharge, and four resulted in death. The morbidity rate on discharge due to thromboembolic complications was 5.9%, and the mortality rate was 3.4%. Seven thromboembolic complications were evident during the procedure, and six were documented on postembolization follow-up radiographic studies. Eight events were managed with local or systemic administration of fibrinolytic or antiplatelet agents. Fibrinolytic therapy consisted of intravenous tissue-type plasminogen activator (tPA) 10 mg along with 10 mg given intra-arterially at the occlusion site. When a distal multifocal occlusion was observed, patients were treated with intravenous abciximab or eptifibatide at the appropriate dose according to their weight.

Postembolization Rebleeding

One (0.8%) patient had fatal rebleeding the 6th day after partial embolization of a large acutely ruptured basilar-tip aneurysm. At the conclusion of the procedure, a few loops of a coil were projecting outside the limit of aneurysmal sac, and the lesion demonstrated residual filling of its base.

Coil Migration

Coil migration during treatment occurred in two (1.7%) of 118 aneurysms. Both cases were associated with coil embolization of small posterior communicating artery aneurysms with unfavorable fundus-to-neck ratios. Migrated coil could be retrieved successfully in one case with no clinical consequence. Another case experienced a fatal stroke after unsuccessful coil retrieval.

Parent-Vessel Injury

Parent-vessel injury occurred in two (1.7%) of 118 procedures, with no clinical consequence. Each involved extracranial dissection of a vertebral artery. The dissections occurred during catheter or wire manipulations. Both cases were treated with intravenous heparin for 24–48 hours after the procedure.

Nonprocedural Complications Attributable to SAH

Of 118 patients with SAH, one (0.9%) had early rebleeding just before coil embolization. Seven (5.9%) had symptomatic vasospasm: Five were treated with intra-arterial papaverine infusion with balloon angioplasty; one, with intra-arterial papaverine infusion only; and one, with triple-H therapy. Seven (5.9%) had hydrocephalus requiring shunt placement. Fourteen (11.9%) deaths occurred after SAH and were not related to direct complications of endovascular treatment. Twelve deaths were attributable to the initial effect of SAH; one death was due to early rebleeding in a patient not treated with coils and one death was due to vasospasm resistant to aggressive treatment.

Procedural Complications for Unruptured Aneurysms

Six (8.3%) procedural complications occurred after 72 procedures for unruptured aneurysms. Thromboembolic complications occurred in five (6.9%), and cranial nerve palsy (thought to be the result of the compressive effect of a coiled aneurysm) occurred in one (1.4%). We noted two technical failures: one due to coil prolapse into the parent artery and one due to intraprocedural distal thromboembolism. Two complications (thromboembolism) had no clinical consequence, two (one thromboembolism and one cranial nerve palsy) were associated with neurologic morbidity after the procedure, one (thromboembolism) resulted in persistent neurologic morbidity on discharge, and one (thromboembolism) resulted in death. Procedure-related neurologic morbidity and mortality rates at the time of discharge were 1.4% and 1.4%, respectively.

Procedural Complications for Re-embolization

Fourteen aneurysms were re-embolized once, and two were re-embolized three times. Additional procedures were necessary because of an inability to deposit coils during the initial attempt (one procedure), incomplete obliteration after initial embolization (14 procedures), or recurrent (or residual) aneurysms at follow-up angiography (five procedures). One technical failure occurred when a fourth embolization for a small residual of a previously ruptured, large basilar-tip aneurysm was aborted because of coil prolapse into the parent vessel. Of 20 re-embolization procedures, four (20%) had complications. All four complications were thromboembolic: two had no clinical consequence and two resulted in persistent neurologic morbidity on discharge. The neurologic morbidity rate related to re-embolization was 10%, and the mortality rate was 0%.

Intraprocedural Aneurysmal Rupture

Rates of aneurysmal rupture during coil embolization are 1.9–16% for ruptured aneurysms and 0–1.3% for unruptured aneurysms (2–4, 6, 12–14). Many authors have found that iatrogenic rupture almost invariably involves previously ruptured lesions (2, 3, 13–15). Kuether et al (2) reported the result of 93 embolization procedures in 77 aneurysms with a 2.1% overall incidence of aneurysm perforation (6.5% in 31 ruptured aneurysms and 0% in 46 unruptured aneurysms). Valavanis et al (15) reported 128 endovascular cases, noting perforation only during treatment of previously ruptured lesions (4.2% in 69 ruptured aneurysms). Ng et al (3) also noted intraprocedural rupture more commonly during treatment of ruptured than treatment of unruptured ones (16% vs. 1.3%).

Small aneurysms are also associated with an increased incidence of intraprocedural rupture (6, 12, 13, 15). Ricolfi et al (12) observed four procedure-related perforations in small aneurysms only (4 mm or smaller in diameter). Vinuela et al (6) reported that nine of 11 perforated aneurysms measured 4–10 mm in diameter. Valavanis et al (15) reported that two of three perforated aneurysms measured 3–4 mm in diameter, and in the series by Sluzewski et al (13), six of the seven perforated aneurysms measured 2–10 mm.

In our series, intraprocedural rupture occurred in nine (4.2%) cases of ruptured aneurysms. The rate of aneurysm perforation was 7.6% in 118 ruptured aneurysms and 0% in 72 unruptured aneurysms. All ruptures occurred before or during the early phase of coil embolization. Anterior communicating artery aneurysms were most commonly perforated (four of nine perforations were interior communicating aneurysms), and all but one intraprocedural rupture occurred in small (3.5–8 mm) aneurysms. Of the nine ruptures, six had good outcomes with no clinical consequence, and three resulted in death. This seems to be consistent with other publications (2, 4, 16–21). Fatal outcomes were associated with intraprocedural ruptures that occurred before detachment of the first coil and poorer initial Hunt-Hess grades (grade III for one and grade IV for two). Morbidity and mortality rates attributed to intraprocedural aneurysm rupture were 0% and 1.5% respectively among a total of 210 procedures. Once an aneurysm ruptured, 33% of the patients died, while 66% did well.

Thromboembolism

Endovascular coil embolization of intracranial aneurysms is associated with a risk of thromboembolic and ischemic complications. Despite routine use of heparin in conjunction with this procedure, complications that related to thromboembolic events occurred in 2.5–28% of patients treated (6, 14, 18, 22). In our study, thromboembolic complications were observed in 22 (10.4%) of 210 procedures; 13 (11%) of 118 ruptured aneurysms, five (6.9%) of 72 unruptured aneurysms, and four (20%) of 20 re-treated aneurysms. Thromboembolic complications accounted for 59.5% of all complications. All thromboembolic complications observed occurred during (n = 12) or within 24 hours (n = 10) of the procedures.

Periprocedural thromboembolism can be treated with selective intra-arterial administration of thrombolytic agents (urokinase or tPA). Cronqvist et al (23) reported the results of superselective intra-arterial fibrinolytic therapy in 19 patients to treat thromboembolic complications during endovascular aneurysm treatment. Complete recanalization was achieved in 10, and partial recanalization, in nine. Fourteen patients had a good clinical recovery, and one died of a large intracerebral hematoma. Aneurysm rupture with SAH occurred in two patients. We previously reported five cases in which intra-arterial thrombolytic agents were used to treat acute vascular occlusion associated with endovascular coil placement of intracranial aneurysms (24). Mechanical clot disruption was performed before intra-arterial thrombolytics were delivered. Angiographic recanalization was observed in four patients, with clinical improvement in two. Two patients experienced hemorrhage, which caused severe morbidity or death.

In view of the modest results and limitations of thrombolytic therapy, several authors (25–28) have proposed the use of abciximab—a chimeric monoclonal antibody fragment against the platelet glycoprotein IIb/IIIa receptor complex—to manage acute thromboembolism during endovascular coil placement in intracranial aneurysms. This drug, or others like it, may be more effective and logical for intraprocedural thromboembolism because most patients have already received full heparinization, making it unlikely that a fibrin clot is present. Platelet clumps, however, are known to quickly form on intravascular foreign bodies, even with full heparin anticoagulation. Lempert et al (27) reported a case of successful abciximab treatment for an acute thrombosis during coil embolization of a basilar apex aneurysm. They observed complete thrombus dissolution after 30 minutes of systemic drug infusion. Ng et al (28) also reported successful thromboembolic treatment with systemic abciximab during endovascular coil placement of a ruptured intracranial aneurysm after superselective infusion of a fibrinolytic agent was unsuccessful. Alexander et al (25) described a similar experience with intravenous abciximab for the treatment of intraprocedural arterial thrombus. Duncan and Fourie (26) described their experience with catheter-directed intra-arterial abciximab in five cases of acute intracranial arterial occlusion during neurointerventional procedures. Rapid thrombus dissolution was observed in four cases, and partial dissolution, in one. Two patients had no neurologic deficit, one had a pre-existing neurologic deficit, one had mild residual thumb paresis, and one had a severe neurologic deficit. No abciximab-related complication occurred. The authors concluded that the intra-arterial administration of abciximab could result in rapid clot lysis.

Of 22 thromboembolic complications in our series, eight were treated by using superselective intra-arterial or intravenous abciximab, and one was treated by using local tPA. Of the eight cases treated with abciximab, we observed complete recanalization in six and partial recanalization in two. One case treated with tPA showed partial recanalization. Two cases with complete recanalization had no neurologic consequence, two had transient neurologic deficit, and two had persistent neurologic deficit. Two of three partially recanalized cases had persistent neurologic deficit, and one resulted in death. Overall morbidity and mortality rates from thromboembolic complications were 4.8% (10 of 210) and 2.4% (five of 210), respectively.

Thromboembolic complication was the major cause of morbidity and mortality in our series. All morbidity from our 210 embolization procedures was associated with thromboembolic complications, which also accounted for one-half of all procedure-related deaths. Once a thromboembolic event occurred, morbidity was 45% and mortality was 23%. With successful experiences in the use of IIb/IIIa inhibitor for these events, acute platelet aggregation is implicated as an important constituent (22, 25, 27, 29).

In view of these findings, our present approach (since November 2002) to decrease the incidence of thromboembolic complications includes the prophylactic use of intravenous eptifibatide (Integrilin; CGL Therapeutics, San Francisco, CA) at the conclusion of the procedure. The effects of eptifibatide, a synthetic heptapeptide with high specificity and affinity for glycoprotein IIb/IIIa receptors, rapidly reverse (within 2–4 hours) after its discontinuation (30–32). This effect may be beneficial in the event of serious bleeding (including aneurysmal rebleeding), in the routine removal of a femoral arterial sheath, and in consideration of early ambulation and discharge after a successful procedure. Preventive use of eptifibatide during coil embolization of an aneurysm poses a substantial risk if iatrogenic rupture should necessitate emergent neurosurgical intervention to evacuate a hematoma. To avoid this quandary, we currently use eptifibatide only at the completion of embolization to prevent or reduce the incidence of postoperative thromboembolic complications.

Nonprocedural Complications Attributable to SAH

Hydrocephalus.—The incidence of SAH-associated hydrocephalus is reported to range from 6% to 67% (33–39). In the International Cooperative Study on the Timing of Aneurysm Surgery (40), 8.0% of patients underwent CSF shunt surgery. In terms of the effect of the type of procedure performed on the incidence of shunt surgery, a number of reports indicate no difference between endovascular and exovascular therapy. In a series of 242 patients with aneurysmal SAH treated with early surgery or early endovascular treatment, Gruber et al (41) reported incidences of 23.2% in the shunt-dependent hydrocephalus in the surgical group and 17.7% in the endovascular group. Sethi et al (42) retrospectively studied 100 matched patients with aneurysmal SAH to determine the association between treatment method and the development of chronic hydrocephalus. They also concluded that the technical procedure does not significantly affect the incidence of shunted hydrocephalus (18% for surgical clip placement and 16% for coil embolization). In our study, the incidence of shunt-dependent hydrocephalus was 5.9% in patients with aneurysmal SAH.

Vasospasm.—Delayed ischemic neurologic deficit (symptomatic vasospasm) is a major cause of morbidity and mortality in patients with aneurysmal SAH. Symptomatic vasospasm is reported in 22–40% of patients with SAH, resulting in 34% morbidity and 30% mortality rates (31, 40, 43–46). Early aneurysmal surgery followed by proper postoperative management, including preventive use of intravenous nimodipine, is reported to reduce the incidence of symptomatic vasospasm to 10% or less (47–49). Although some have suggested that the cisternal removal of blood may reduce the incidence of cerebral vasospasm, only partial physical removal of subarachnoid blood is possible during early aneurysmal surgery (50–53). The International Cooperative Study on the Timing of Aneurysm Surgery (40, 54) did not demonstrate a major influence of early surgery on morbidity and mortality due to chronic cerebral vasospasm. The difficulty in removing tenacious subarachnoid blood clots, the effects of blood breakdown products, or the surgical-mechanical manipulation of vessels may explain this failure of early surgery to substantially alter the incidence of vasospasm (40, 46, 55, 56). Murayama et al (57) reported a 23% incidence of symptomatic vasospasm after endovascular coil occlusion of acutely ruptured; this rate compares favorably with that found in conventional surgical series. Gruber et al (58), however, noted an increased incidence of vasospasm-related infarctions in patients treated endovascularly (37.7% vs. 21.6% with surgery). However, when patients with Fisher grade 4 and Hunt and Hess grade V lesions were excluded, the difference between the treatment groups was no longer significant.

Other authors (57, 59, 60) have not found an increased risk of vasospasm with endovascular therapy. Yalamanchili et al (60) retrospectively compared 19 patients treated by surgical clip placement with 18 patients treated by endovascular coil embolization after aneurysmal SAH, excluding patients with Hunt and Hess grade IV or V SAH. Symptomatic vasospasm was more frequent and more severe in the surgical group (74% vs. 22%). Charpentier et al (59) performed a multivariate analysis to determine the role of treatment technique on cerebral vasospasm after aneurysmal SAH. Symptomatic vasospasm occurred in 25 (22.2%) of 145 surgical patients compared with 22 (17.2%) of 99 endovascular patients. They concluded that the type of treatment was not associated with an increased risk of cerebral vasospasm. Rabinstein et al (61) studied 415 consecutive patients with aneurysmal SAH. Symptomatic vasospasm occurred in 39% treated with surgical clip placement and 30% treated with endovascular coil occlusion. In a univariate analysis, the incidence of vasospasm did not differ between the groups.

In our study, the incidence of symptomatic vasospasm among 118 patients with ruptured aneurysms was 5.9% (n = 7). Excluding 35 patients with Hunt and Hess grades of IV and V, in whom it is difficult to accurately detect a vasospastic neurologic deterioration, the incidence was 4.8%. Our rate tended to be lower than those previously reported, even when we did not make the exclusion just described.