Analysis of Complications and Recurrences of Aneurysm Coiling with Special Emphasis on the Stent-Assisted Technique

BACKGROUND AND PURPOSE: Stent-assisted coiling has expanded the treatment of intracranial aneurysms. With the use of continuously compiled data, we reviewed the role and drawbacks of stent-assisted coiling. MATERIALS AND METHODS: We compiled data from consecutive patients from 2003–2012 who underwent coiling, with or without stent assistance. Clinical and angiographic results were analyzed retrospectively. RESULTS: Of 1815 saccular aneurysms in 1505 patients, 323 (17.8%) were treated with stents (299 procedures) and 1492 (82.2%) without stents (1400 procedures). Procedure-related complications occurred in 9.4% with stents versus 5.6% without stents (P = .016, relative risk 1.5; 95% CI, 1.1–2.7). Ischemic complications were more frequent in the stent group than in the no-stent group (7.0% versus 3.5%; P = .005; relative risk, 1.7; 95% CI 1.2–2.5), as were hemorrhagic complications (2.3% versus 1.9%, P = .64). Procedure-induced mortality occurred in 2.7% (8/299) with stents versus 1.1% (15/1400) without stents (P = .029; relative risk, 2.0; 95% CI, 1.1–3.5). Logistic regression analysis identified wide-neck aneurysms as the most significant independent predictor of complications. A total of 64.1% (207/323) of aneurysms treated with stents and 70.3% (1049/1492) treated without stents have been followed, disclosing angiographic recurrence in 15.5% (32/207) versus 35.5% (372/1049), respectively (P < .0001). Logistic regression analysis showed that the presence of a stent was the most important factor for the reduction of angiographic recurrence (P < .0001; relative risk, 2.3; 95% CI, 1.6–3.3). CONCLUSIONS: The stent-assisted coiling technique is associated with a significant decrease in recurrences but a significant increase in complications. The treatment of wide-neck aneurysms remains hazardous.

T he stent-assisted coiling technique has broadened the indication for coil embolization, and numerous reports have depicted the value of stents in the treatment of cerebral aneurysms. [1][2][3][4][5][6] We present herein the clinical and angiographic results of a consecutive series of 1815 aneurysms treated over a 9-year period. The aims of this retrospective study were to place the role of stentassisted coiling into perspective and to determine the factors associated with procedural complications.

Data Collection
From a prospectively gathered data base of all patients with intracranial saccular aneurysms (with no prior endovascular treat-ment) who were treated with coils in our institution between January 2003 (when we initiated the use of self-expandable stents) and March 2012, 1505 patients were identified and constituted our study population. The data base included data on the patient (age, sex), the endovascular procedure technique (balloon-assisted, stent-assisted, stand-alone coiling), aneurysm status (ruptured or not), procedure-related morbidity and mortality, immediate and follow-up angiographic results, and the rate of (re-)hemorrhage. All data were reviewed and statistically analyzed. Institutional review board acceptance was obtained for this retrospective study, and the need for informed consent was waived.

Endovascular Procedures
Coiling was performed under general anesthesia and full anticoagulation with heparin in all cases. In all patients with no history of subarachnoid hemorrhage within the previous 4 weeks, 250 mg aspirin was given intravenously. Heparin was discontinued after embolization in most patients. Whenever stent placement was anticipated, patients were given dual antiplatelet therapy before surgery (75-150 mg clopidogrel, 250 mg aspirin daily initiated 15 days before the procedure and continued for 6 months). There were no strict exclusion criteria, but a massive subarachnoid hemorrhage potentially requiring ventricular drainage was considered to be a contraindication to stent placement.

Follow-Up Protocol
Our standard follow-up protocol consisted of fixed angiographic follow-up, the first follow-up being performed at 1-6 months after endovascular treatment (depending of the quality of the initial angiographic occlusion and the presentation). The second angiographic follow-up was performed 12 months after the first follow-up, and the third follow-up was performed 24 months after the second follow-up. In the case of angiographic recurrence and/or associated aneurysms left untreated, the follow-up was continued on a yearly basis. A single reader prospectively evaluated all angiograms. The degree of aneurysm occlusion was classified according to the Montreal simplified classification. 7 Aneurysm recurrence was dichotomized as absent or present. At follow-up, an aneurysm was considered recurrent if a previously totally occluded aneurysm (class 1) had a partial recurrence of the neck (class 2) and/or the sac (class 3). An aneurysm was considered remnant regrowth if a subtotally occluded aneurysm (class 2 or 3) was found to have an increasing neck remnant or residual aneurysm. Aneurysm dimensions were determined on 3D images derived from rotational angiography. Indication for retreatment (class 3 angiographic results) was decided on a case-to-case basis, depending on the patient age and medical history. Aneurysm retreatments were noted.

Statistical Analysis
Data are presented as means for continuous variables and frequencies for categoric variables. Statistical analysis of categoric variables was carried out by use of 2 and Wilcoxon tests. Analysis of variance followed by Bonferroni post hoc testing was used to assess differences between the stent-assisted and no-stent groups.
A multivariable logistic regression analysis was carried out to determine predictors of procedural complications and aneurysm recanalization. P values of Յ.05 were considered statistically significant. Statistical analysis was carried out with R software version 2.15 (http://www.rproject.org/).

Baseline Demographics and Procedures
Among 1505 consecutive patients, 1815 intracranial aneurysms were treated by coil embolization in 1699 procedures. A total of 323 aneurysms were treated with the assistance of self-expandable stents in 299 procedures; 1492 aneurysms were treated without stent assistance in 1400 procedures. The baseline demographics of all of the aneurysms according to stent use are shown in Table 1.

Immediate and Follow-Up Angiographic Results
Immediately, there were more complete obliterations in the nostent group and more residual aneurysms in the stent-assisted group, but this trend was reversed at last follow-up (Table 2).

Risk Factors for Angiographic Recurrence
In the univariate analysis, the risk factors for angiographic recurrence were ruptured status at presentation, larger aneurysm size, no stent, wider neck, younger age, and lower coil packing attenuation ( Table 4). The identified reductive factor for angiographic recurrence was multiple aneurysms. In the logistic regression analysis, the independent variables were ruptured status at presentation, absence of stent, larger aneurysm size, lower packing attenuation, younger age, and wider neck (Table 4).
In the univariate analysis, the risk factors for aneurysm retreatment were larger aneurysm, lower packing attenuation, and incomplete initial aneurysm occlusion. In the logistic regression analysis, the independent variables favoring retreatment were larger aneurysm size, absence of stent, hemorrhagic presentation, lower packing attenuation, and initial incomplete occlusion (Table 5).

Procedure-Related Complications
The procedure-related complications were counted for any intracranial hemorrhages (including a wire perforation or an aneurysm rupture) and any ischemic events (that resulted in patient morbidity or mortality). Hydrocephalus after the treatment of unruptured aneurysms was also counted as a complication. In the no-stent group, procedure-related complications occurred in 5.6% of procedures, including complication-related deaths in 1.1% ( Table 6). The 49 ischemic complications resulted in 5 deaths, and the 27 perforations resulted in 10 deaths. There were 3 hydrocephali after treatment of unruptured aneurysms with hydrogel-coated coils.
In the stent-assisted group, complications occurred in 9.4% of procedures and led to 8 deaths (2.7%). The 21 ischemic events led to 5 deaths, and the 7 hemorrhagic events led to 3 deaths. Overall, there were significantly more complications in the stent-assisted group than in the no-stent group (9.4% versus 5.6%; P ϭ .016).
In the univariate analysis, the risk factors for procedure-related complications were wider neck, stent-assisted coiling, MCA location, and larger aneurysm. In logistic regression analysis, the independent variable was a wider neck. Stent-assisted coiling and aneurysm size were related to neck width (Table 7).

(Re-)Bleeding After Coiling
Only 21 of 1815 aneurysms (1.2%) bled after endovascular treatment ( Table 8). The rates of (re-)bleeding were 0.3% in the stentassisted group and 1.4% in the no-stent group. In the stent-assisted group, 2 patients bled after treatment of unruptured aneurysms (1 from the aneurysm 12 months after the treatment; the second was readmitted 3 weeks after the treatment with a remote intraparenchymal hematoma). No rebleeding was seen from previously ruptured aneurysms. In the no-stent group, 1 unruptured aneurysm bled 12 months after coiling (modified Rankin Scale 4). Nineteen rebleedings occurred after coil embolization of 729 ruptured aneurysms (2.6%) and led to 13 deaths.

DISCUSSION
Stent-assisted coiling creates a mechanical scaffold to prevent coil protrusion into the parent vessels. Thus, the indication for this technique had mostly been motivated by aneurysm morphology (large neck). Accordingly, our stent-assisted group included aneurysms with wider necks. Ruptured aneurysms were underrepresented to avoid antiplatelet therapy in the setting of subarachnoid hemorrhage. In the stent-assisted group, internal carotid and MCA aneurysms were overrepresented, whereas posterior circulation and pericallosal aneurysms were scarce, explained by a lower incidence of wide-neck aneurysms in these locations. Multiple aneurysms, which were more often treated with coiling alone, tended to have a smaller size, a narrower neck, and fewer ruptured aneurysms compared with the group of single aneurysms.

Immediate and Follow-Up Angiographic Results
Immediate angiographic complete occlusions were obtained less frequently in the stent-assisted than in the no-stent group (51.1% versus 61.5%). This is because larger aneurysms were more frequent in the     stent-assisted group and because dual antiplatelet therapy affected the immediate intra-aneurysmal thrombosis. Catheter kickback out of the stent also affected tight packing. Conversely, at follow-up, complete occlusions increased to 73.4% in the stent-assisted group, whereas these diminished to 54.0% in the no-stent group. For stentassisted coiling, numerous articles have reported a broad range (13.2-94.4%) of immediate complete occlusion. 1,2,[8][9][10][11][12][13] However, similar to the present series, most mid-to-long-term follow-up series have reported augmented rates of angiographic complete occlusion at follow-up (range, 54 -81%). [8][9][10][11][12][13][14] Angiographic Recurrences Recurrences were statistically less likely in the stent-assisted group (15.5% versus 35.5%; P Ͻ .0001). Low rates of recurrence for stent-assisted coiling have also been reported in other recent studies (range, 0 -15.2%). 8,10,11,[13][14][15] In line with our findings, rup-tured aneurysms, larger size, lower packing attenuation, and wide necks are well-established risk factors for recurrence. 7,[16][17][18][19][20] We also identified younger age to be a risk factor for recurrence. The effect of age is not yet clear, but younger age was one of the predicting factors of late retreatment in the International Subarachnoid Aneurysm Trial. 21 Multiple aneurysms reduced the likelihood of recurrence, but these aneurysms were generally smaller, had narrower necks, and had fewer ruptured lesions.
In our study, the use of a balloon-assisted technique had no influence on recurrence. Conversely, Shapiro et al 22 found both initial and follow-up aneurysm occlusion rates to be higher in balloon-assisted cases. In our study, an absence of a stent was identified as one of the most relevant factors for recurrence. Our results help to confirm the evidence that stent-assisted coiling augments treatment durability and contributes to progressive occlusion. 10,23,24 This durability can be explained by the combination of biologic, geometric, and hemodynamic mechanisms. [25][26][27]

Procedural Complications
Our no-stent results show occurrences of complications (5.6%) and mortality (1.1%) similar to various other series without stents. [28][29][30][31] Henkes et al 28 reported procedural morbidity of 5.0% and mortality of 1.5%. van Rooij et al 30 reported procedural complications with a morbidity rate of 3.2% and a mortality rate of 2.6%. A recent meta-analysis for unruptured aneurysms (mainly of studies without the use of stents) found a morbidity rate of 4.8% and a mortality rate of 1.2%. 29 Our complication and mortality  rates in the stent-assisted group were 9.4% and 2.7%, respectively. These are similar to other series of stent-assisted coiling that have reported 2.9 -11% morbidity rate and 0 -4.8% mortality rate. 1,3,5,8,32,33 Overall, our complication rate in the stent-assisted group was higher than in the stand-alone coiling series (9.4% versus 5.6%; P ϭ .016). The use of a stent was linked to a significantly higher morbidity in our series, but not as an independent factor because stents were generally used in aneurysms with wider necks. Three cases (0.2%) of hydrocephalus occurred after the treatment of unruptured aneurysms with hydrogel-coated coils, and there were no cases in the smaller stent-assisted group. The HELPS trial also reported that there was no significant difference in the occurrence of hydrocephalus between the hydrogel-coated and bare-platinum coil groups. 34 A large aneurysm size was identified as a risk factor for complications but was not identified as an independent value, unlike neck width, which also appeared to be an independent factor. In the CLARITY study, aneurysms Ͼ10 mm had a higher risk of thromboembolic events, with higher morbidity-mortality rates compared with smaller lesions. 35 Similar results have also been reported in the ATENA study. 33 Moreover, the amount of procedural aneurysmal ruptures was significantly higher for smaller aneurysms. 33,36 In CLARITY, the morbidity-mortality secondary to thromboembolic events was higher in the group of aneurysms for which the neck was Ͼ4 mm. 35 In the same study, a neck Ͼ4 mm was also identified as an independent risk factor of intraoperative rupture. 35 As in our study, age was not identified as a risk factor for complications in the series from van Rooij et al. 30 However, Sedat et al 14 reported that thromboembolic events were more frequent among elderly patients, whereas in the CLARITY study, procedural ruptures were more likely to occur in patients Ͻ65 years of age. 35 As did van Rooij et al, 30 we did not find any specific locations to be linked with complications. However, we found that MCA location was a risk factor in the univariate analysis. Similarly, the CLARITY study reported that thromboembolic events were more likely to occur in MCA aneurysms than in aneurysms in other locations, such as intraoperative ruptures. 35 The treatment of several aneurysms during the same procedure was not found to result in more complications, as was previously found in the CLARITY study. 35 However, only 6.4% of our procedures were multiple aneurysm treatments. The balloon-assisted technique was not identified as a risk factor for complications, unlike in some previous reports. 28,30,37 However, the ATENA and CLARITY studies showed that the balloon-assisted technique was as safe as conventional coiling. 33,35 A ruptured aneurysm was not found to be a significant risk factor for complications in the current study. However, this is contrary to various other studies. For example, Ng et al 18 reported that intraprocedural ruptures occurred more frequently for ruptured aneurysms than for unruptured aneurysms. Ishibashi et al 38 reported that hyper intensities were seen more frequently on diffusion-weighted MRI after the treatment of ruptured than unruptured aneurysms. Also, Ross and Dhillon 39 reported that the risks of vessel or aneurysm rupture or thromboembolic stroke were greater during the treatment of ruptured aneurysms. Similarly, Park et al 36 reported that pro-cedural morbidity and mortality rates were higher for ruptured than unruptured aneurysms. We cannot postulate as to why our results are different from these studies.

(Re-)Bleeding
The (re-)bleeding rates were low for both stented and nonstented aneurysms, which is in line with previous reports. [40][41][42][43] None of the ruptured aneurysms that were treated with stent-assisted coiling rebled. However, no definitive conclusions can be drawn because of the small number of ruptured aneurysms that were treated with stents.

Study Limitations
Our study has the inherent limitations of a retrospective study. The duration of follow-up was shorter for stented aneurysms because most of the stents were implanted during the last 6 years. Moreover, the rate of aneurysms followed by angiography was lower.

CONCLUSIONS
The stent-assisted coiling technique was associated with a significant decrease in recurrences but a significant increase in complications. The treatment of wide-neck aneurysms remains more hazardous.