Background To examine the effect of stenting on progressive occlusion of small and incompletely occluded unruptured intracranial aneurysms (UIAs) ≤10 mm in size using a propensity score matched case controlled analysis.
Methods 715 small UIAs consecutively treated by coiling between 2008 and 2010 were eligible for study. Time of flight MR angiography and/or catheter angiography were used to estimate extent of occlusion after coiling. Complete occlusion at 6 months post embolization of a sac filled with contrast immediately after coiling constituted progressive occlusion. A propensity score matched analysis was conducted, based on the probability of stent deployment.
Results 206 (28.8%) small UIAs showed residual sac filling directly after coiling. Of these, 182 (88.3%) displayed progressive occlusion at 6 months. Aneurysm size (p<0.01), neck size (p<0.01), and embolization attempt (p<0.01) differed significantly for stented and non-stented lesions, but the incidence of progressive occlusion did not differ (p=0.78) between the groups. After 1:1 propensity score matching, however, the rate of complete occlusion in stented subjects (97.5%) surpassed that of the non-stented counterparts (OR=9.75, p=0.01).
Conclusions Small UIAs with residual sac filling after coiling showed a complete occlusion rate of 88.3% at 6 months post embolization. Stent deployment seems to promote complete occlusion in such lesions.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Clinical management of small unruptured intracranial aneurysms (UIAs) ≤10 mm in size is open to debate. Although the International Study of Unruptured Intracranial Aneurysms (ISUIA)1 demonstrated low rupture risk in small aneurysms, rates up to 87.9% are quoted for small ruptures.2 Consequently, configuration, location, prior history of subarachnoid hemorrhage, familial history, and procedural risks should also be considered in related therapeutic decisions.
With advances in endovascular technologies, coil embolization of UIAs is on the rise; however, rates of recanalization (20.8%) and retreatment (10.3%) are disturbing.3 Risk factors for recanalization include rupture, larger size, younger patient age, wide necked lesions, and incomplete initial occlusion.4 ,5 Thus complete occlusion (ie, absence of contrast filling after coiling) to prevent recanalization is the goal of all neurointerventionalists. Although the immediate post interventional rate of angiographic incomplete occlusion remains at 8.8%,3 despite every effort, inordinate attempts to achieve complete occlusion may culminate in serious procedural complications. Clinical outcomes (total progressive occlusion vs recanalization) of the aneurysms with residual sac filling, especially small UIAs, are therefore of particular interest.
Self-expanding stents enabling higher packing density, flow diversion, and endothelization of the aneurysm neck are gaining in popularity.6 ,7 Chalouhi et al7 reported a 12% rate of recanalization and a 6.4% retreatment rate with stent assisted coiling. Unfortunately, previous studies addressing the impact of protective stenting on recanalization after coil embolization have largely involved heterogeneous samplings (completely and incompletely occluded aneurysms). Hence it remains to be determined whether complete occlusion (via progressive thrombosis) in small UIAs showing residual sac filling directly after coiling may be enhanced by protective stenting. Taking into account confounding factors for the probability of stent deployment, such as configuration and size of the aneurysm, arterial branching, anatomic location, and packing density, we conducted a case controlled study, using a propensity score analysis, to determine the effect of stenting on progressive occlusion in small UIAs with residual sac filling.
A total of 898 patients with 1035 intracranial aneurysms underwent coil embolization at a single institution between January 2008 and December 2010. Patients with non-saccular (ie, fusiform, blood blister-like, pseudo, or dissecting aneurysms), large (>10 mm), or ruptured aneurysms, or those lacking 6 month follow-up imaging were excluded, leaving 715 small UIAs eligible for study (92.7% followed for 6 months). Ultimately, the cohort was limited to 206 small UIAs with contrast filled sacs directly after coil embolization, as stipulated by the Raymond Scale.8 ,9
Patient variables extracted from the medical records included gender, age, comorbidities (hypertension, diabetes mellitus, hyperlipidemia, and coronary artery disease), smoking history, and antiplatelet medications. Angiographic variables such as diameter of the aneurysm, neck size, location (anterior vs posterior circulation),10 depth to neck ratio, nature of affected artery (non-branching vs branching), packing density, embolization attempt (first vs second due to recanalization), and type of coil used (hydrogel modified (HydroSoft coils; MicroVention, Aliso Viejo, California, USA) vs polyglycolic acid/lactide copolymer coated (Matrix coils; Boston Scientific, Fremont, California USA)) were recorded. Based on the length of the bioactive coil inserted (relative to total), aneurysms were grouped as bare (≤50%) or bioactive (>50%). This study was approved by our institutional review board.
Most of the endovascular procedures were conducted under general anesthesia. Dual agent antiplatelet therapy (loading doses of clopidogrel and aspirin (ie, 300 mg each), given 1 day prior, and another 75 mg of clopidogrel and 100 mg of aspirin, given the morning of the procedure) was administered if stent protection (Enterprise stent; Codman, Raynham, Massachusetts, USA; Neuroform stent, Boston Scientific, Natick, Massachusetts, USA) was anticipated.11 In patients were poor responders to clopidogrel, signaled by VerifyNow P2Y12 assay, cilostazol was added. If stent placement was unlikely, clopidogrel alone served as premedication, adding aspirin in poor responders.
A bolus of heparin (3000 IU), given on femoral arterial sheath placement, was bolstered thereafter by hourly doses (1000 IU), and activated clotting time was monitored hourly. In patients with a stent, continued dual antiplatelet therapy was advised for at least 3 months post embolization, followed by single agent maintenance for at least 1 year. Otherwise, a maintenance antiplatelet regimen was not routinely prescribed, except in instances of prior antiplatelet agent use, coil protrusion, or procedural thromboembolism. An Integris V (Philips Medical Systems, Best, The Netherlands) scanner was engaged for each procedure.
Angiographic outcome and follow-up monitoring
Immediate post embolization angiographic results were categorized into three groups based on the Raymond classification.9 Among them, only aneurysms with residual contrast filled sacs were included in this investigation, checking for progressive occlusion.12
In all patients, time of flight MR angiography (TOF-MRA), with three-dimensional reconstruction and source image, was recommended at 6 months post embolization. Conventional angiography was advised if post-treatment MRA was not feasible or if recanalization was suspected by MRA (see supplementary figure). Anatomic outcomes on follow-up were interpreted as complete or incomplete occlusion. Progressive occlusion was equated with complete saccular occlusion (no flow or contrast within the aneurysm on MRA or angiography) in any lesion previously filled with contrast. Incomplete occlusion was defined as slight filling at the aneurysm neck or prominent filling of the sac.13 Radiologic assessments were conducted by two independent blinded neuroradiologists.
Categorical variables are shown as numbers and percentages, with continuous data expressed as mean±SD. The χ2, Fisher's exact, or Student's t test was used to estimate categorical and continuous variables. A propensity score matched analysis was done (based on the probability of stent deployment), using multivariate logistic regression analysis of relevant variables (gender, age, location, aneurysm size, neck size, arterial branching status, packing density, and use of HydroSoft coil) to calculate individual propensity scores. A greedy matching technique was applied for one to one match. The Kaplan–Meier product limit method and generalized Wilcoxon test were used to estimate cumulative survival without progressive thrombosis by stent deployment status. Statistical significance was set at p<0.05. Separate programs were enlisted for score matched analysis (SAS Release V.9.1.3; SAS Institute, Cary, North Carolina, USA) and other calculations (SPSS V.19; SPSS Inc, Chicago, Illinois, USA).
A total of 715 small UIAs qualified for the study (497 (69.5%) in women; mean age 58.4±10.0 years). Angiographic outcomes of complete occlusion or residual neck were achieved in 509 (71.2%) of these. Arteries affected were as follows: internal carotid (n=289, 40.4%), middle cerebral (n=117, 16.4%), anterior cerebral and anterior communicating (n=176, 24.6%), and posterior cerebral and posterior communicating (n=133, 18.6%). Directly after coil embolization, 206 UIAs (28.8%) showed residual sac filling. Multivariate analysis linked female gender (p=0.02), advanced age (p=0.01), smaller lesion size (p<0.01), non-branching vessels (p<0.01), and wide necked aneurysms (depth to neck ratio ≤1; p=0.01) with contrast retention immediately post embolization (table 1).
Propensity score adjusted outcomes
Of the 206 aneurysms with residual sac filling, 182 (88.3%) underwent progressive complete occlusion at the 6 month follow-up imaging. To determine the impact of stenting on progressive occlusion in this context, aneurysms with contrast filled sacs were divided into stented (n=72) and non-stented (n=134) groups. For the non-stented group, balloon remodeling and double microcatheter techniques were used in 42 (31.3%) and 29 (21.6%) cases, respectively. Aneurysm size (p<0.01), neck size (p<0.01), and embolization attempt (p<0.01) differed significantly between the groups whereas the incidence of progressive occlusion did not differ (p=0.78). Procedure related complications were noted in 10 patients (symptomatic cerebral infarction, n=7; thrombus resolved by tirofiban, n=3). The difference in complications rates between the stented and non-stented groups did not reach statistical significance (n=3, 4.2% in the stent group vs n=8, 6.0% in the non-stent group, p=0.58). Major recanalization was found in 11 patients (5.3%). Among them, retreatment was performed in seven patients. There was no delayed infarction or aneurysm bleeding during the follow-up period.
After 1:1 propensity score matching of subjects, respective group demographics were similar (female gender: 80% vs 80%, p=1.00; age: 61.6±10.8 vs 60.5±10.0 years, p=0.65; hypertension: 45% vs 55%, p=0.37; diabetes mellitus: 12.5% vs 7.5%, p=0.71; hyperlipidemia: 17.5% vs 25%, p=0.41; coronary artery disease: 10% vs 7.5%, p=1.00; and smoking: 2.5% vs 12.5%, p=0.20). Angiographic variables are shown in table 2. In small UIAs with residual sac filling immediately after coil embolization, the rate of progressive occlusion achieved through stent assisted coil embolization was significantly better than that of simple coiling (stented, 97.5%; non-stented, 80%; OR=9.75 (95% CI 1.16 to 82.11); p=0.01) (figure 1).
Kaplan–Meier estimates of cumulative survival 36 months after coil embolization without recanalization are shown in figure 2, before and after propensity score matched analysis (before: 84.8% (non-stented) vs 81.4% (stented); after: 73.5% (non-stented) vs 94.4% (stented)). Using the generalized Wilcoxon test, cumulative survival without recanalization by stent deployment status also reached significance after propensity score matched analysis (p=0.016).
Incompletely occluded endovascularly treated saccular aneurysms are subject to recanalization and recurrent hemorrhage over time;14 ,15 hence the aim of every neurointerventionalist is complete occlusion Nevertheless, a recent systemic review has cited an 8.8% rate of incomplete occlusion after coil embolization.3 Stents have been used to increase packing density during coiling and alter hemodynamics (ie, flow diversion and endothelization of aneurysm neck).6 ,7 Lawson et al6 have reported that stent associated flow remodeling contributes to progressive occlusion of incompletely coiled aneurysms. In contrast with our study, minimal residual contrast at lesion base (Raymond class II) constituted incomplete occlusion in their study. They insisted that stent deployment (OR=18.54), smaller lesion size (OR=0.84), and higher packing density (OR=1.09) impacted progressive occlusion. However, observed differences in age (stented, 57.0±13.6 years; non-stented, 63.4±11.4 years), aneurysm size (stented, 9.1±4.2 mm; non-stented 11.2±4.9 mm), and neck size (stented, 5.4±2.0 mm; non-stented, 5.1±1.8 mm) were problematic when assessing the effect of protective stenting on progressive occlusion. Furthermore, confounding factors, such as location, initial presentation (unruptured vs ruptured), and packing density, contributed to progressive thrombosis in incompletely occluded aneurysms. To address such inherent biases, we conducted a propensity score matched analysis. Our results showed that stent assisted coiling enhanced the rate of progressive occlusion in small UIAs with residual sac filling, as opposed to simple coiling.
The significance of contrast filling directly after coil embolization has been investigated. Mascitelli et al16 grouped subjects by contrast found within coiled masses (class IIIa) and contrast lining the walls (class IIIb) of incompletely occluded aneurysms (Raymond–Roy class III). Compared with class IIIa lesions, class IIIb aneurysms displayed a lower progressive occlusion rate, with higher rates of retreatment and subsequent rupture. However, their cohort included a high proportion of ruptured aneurysms (54.87%), and the ratio of procedural assistance differed (stent alone, 61.8%; stent assisted coil embolization, 28.2%). Moreover, class IIIb aneurysms were larger, had wider necks, and were less densely packed by comparison. Hence the effect of protective stenting on progressive occlusion could not be accurately assessed. Hwang et al17 interpreted contrast stasis as a benign angiographic feature, unrelated to recanalization. However, proportionate differences in stent deployment between groups with (19.2%) and without (10.1%) contrast stasis limit the interpretation of the results.
In our cohort, aneurysm size (p<0.01), neck size (p<0.01), and instances of second embolization (p<0.01) differed significantly in stented (n=72) and non-stented (n=134) subjects (see table 2). Although the rate of progressive occlusion did not differ significantly by group (stented, 87.5%; non-stented 88.8%; p=0.78) in unmatched subjects, protective stenting proved beneficial in propensity score matched analysis of small UIAs incompletely occluded directly after coil embolization (OR=9.75; p=0.01). This benefit was conferred despite a significant disparity in antiplatelet maintenance therapy of stented (40/40, 100%) and non-stented (14/40, 35%) subsets (p<0.01).
Some critics refute the significance of progressive occlusion and its association with stent deployment under these circumstances in small sized UIAs (≤10 mm). From reviewing the literature, we acknowledge that UIAs >10 mm in size require interventions, such as clipping or coiling. However, treatment of small UIAs is still debatable in clinical practice, given the contradictory data on risk of rupture (low) and proportion of small ruptures (high).18 Thus procedural risk is especially important to avoid in the endovascular treatment of small UIAs, which sometimes defy all attempts at occlusion. Undue efforts to achieve complete occlusion predisposes to procedural complications. Knowing the clinical outcomes of incompletely occluded aneurysms (total progressive occlusion vs recanalization) is therefore of considerable importance. However, previous studies have largely focused on the risk factors of recanalization in heterogeneous samplings of completely and incompletely occluded aneurysms.
Angiographic outcome immediately after coil embolization may vary. Lawson et al6 demonstrated the effect of protective stents on progressive occlusion in initially incompletely occluded aneurysms. In their study, 44.8% (145/324) of aneurysms showed incomplete occlusion of Raymond classification 2 (neck remnant) and Raymond classification 3 (residual sac). Piotin et al19 also revealed that the rate of aneurysm recurrence was lower in stent assisted coil embolization than in simple coiling. More specifically, in the stent group, complete occlusion, neck remnant, and residual sac were noted in 46.3% (100/216), 19.0% (41/216), and 34.7% (75/216), respectively, compared with 63.5% (704/1109), 18.2% (202/1109), and 18.3% (203/1109) in the non-stent group. Ozretic et al20 showed that only 36.0% (102/283) of aneurysms had complete occlusion. In particular, patients who underwent stent assisted coil embolization showed complete occlusion in 20%, residual neck in 15%, and residual sac in 65%. Here we only included coiled aneurysms with residual sac filling according to the Raymond Scale,9 because we encountered a relatively higher number of residual neck (n=535, 63.6%) in coiled small aneurysms12 than in previous studies. Although the Raymond Scale has been widely used due to its simplicity, its application during industry sponsored registries revealed significant interobserver variability which may limit its application and efficacy as a tool for outcome measurement.21 ,22 Accordingly, we decided that coiled aneurysms with residual sac filling would be more appropriate to assess the effect of stenting on complete occlusion in our cohort.
As a result of this study, we determined an 88.3% rate of progressive complete occlusion 6 months after embolization in small UIAs initially showing residual sac filling. By deploying a stent during coil embolization, neurointerventionalists anticipate some flow remodeling effect, but saccular thrombosis is the only recognized means by which stent placement leads to occlusion.23 On the other hand, the effect of stenting on progressive occlusion over time has not been critically examined. Due to differences in factors impacting on progressive thrombosis, we similarly found no significant association between stenting and progressive occlusion in the 206 incompletely occluded small UIAs of our cohort. Although previous studies6 ,19 have demonstrated the effect of protective stents on progressive occlusion and aneurysm recurrence, we speculated that some confounding factors, such as initial presentation (unruptured vs subarachnoid hemorrhage) and aneurysm configuration (bifurcation vs side wall) could be of concern when assessing the effect of stenting on progressive occlusion. Therefore, we conducted a propensity score matched analysis based on the probability of stent deployment using multivariate logistic regression analysis of relevant variables, such as aneurysm location, size, arterial branching status, and packing density in unruptured small aneurysms. A propensity score matched analysis conveyed less error in estimates than simple regression analysis.24 Based on the latter, our data suggest that protective stenting promotes progressive occlusion in small UIAs with residual sac filling.
Limitations of the study include its non-randomized, retrospective, and observational approach, applying TOF-MRA for follow-up (as dictated by institutional protocol) without baseline MRA imaging shortly after coil embolization. However retrospective in nature, measured variables related to the probability of stent deployment assumed less proportional influence due to propensity score matching. Some studies25 ,26 have shown the feasibility of TOF-MRA to evaluate aneurysm occlusion using DSA as the reference. In our previous report,27 TOF-MRA (MIP and source image) has been shown to be helpful in estimating the degree of recanalization. Nevertheless, it is possible for stent artifacts to obscure minor recanalization and thus to influence recurrence rates in some cases. Thus the possibility needs to be taken into account when interpreting the results.
Progressive aneurysm occlusion was found in 88.3% of small UIAs with residual sac filling directly after coil embolization. Stenting may enhance progressive occlusion in this setting. A prospective multicenter study of longer duration is required to confirm our results.
Competing interests None declared
Contributors Study concept and design: JPJ and YDC. Data collection: JKR, JJP, and W-SC. Literature research: H-SK and JEK. Drafting the manuscript: JPJ. Revision of the manuscript: YDC and MHH. Approval of the final version of the manuscript: all authors.
Ethics approval Ethics approval was obtained from the Seoul National University Hospital institutional review board.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Additional data can be requested by email from the first author (firstname.lastname@example.org).