Intracranial Aneurysms: Meta-Analysis Endovascular Treatment of Very Small

BACKGROUND AND PURPOSE: Outcomes of endovascular treatment of very small intracranial aneurysms are still not well-characterized. Recently, several series assessing coil embolization of tiny aneurysms have presented new promising results. Thus, we performed a systematic review and meta-analysis of studies evaluating endovascular treatment of very small intracranial aneurysms. MATERIALS AND METHODS: We conducted a computerized search of Scopus, Medline, and the Web of Science for studies on endovascular treatment of very small ( (cid:2) 3 mm in diameter) intracranial aneurysms published between January 1996 and May 2015. Using a random-effects model, we evaluated clinical and angiographic outcomes. RESULTS: Twenty-two studies with 1105 tiny aneurysms (844 ruptured and 261 unruptured) endovascularly treated were included. Post-operative and long-term complete occlusion was achieved in 85% (95% CI, 78%–90%) and 91% (95% CI, 87%–94%) of aneurysms, respectively. The recanalization rate was 6% (95% CI, 4%–11%) and retreatment occurred in 7% (95% CI, 5%–9%) of cases. Seventy-nine percent (95% CI, 64%–89%) of patients had good neurologic outcome at long-term follow-up. Intraprocedural rupture occurred in 7% (95% CI, 5%–9%) of the coiling procedures, while thromboembolic complications occurred in 4% (95% CI, 3%–6%). CONCLUSIONS: Coilembolizationofverysmallintracranialaneurysmscanbeperformedsafelyandeffectively.Inthecaseofunruptured aneurysms,procedure-relatedcomplicationsarenotnegligible.Patientsandprovidersshouldconsidersuchriskswhenengagedinashared

E ndovascular treatment is now the standard of care for most intracranial aneurysms (IAs). 1,2Despite endovascular treatment being safe with low morbidity and mortality rates, in certain groups of aneurysms, a detailed risk-benefit assessment of this treatment must be considered. 35][6] In the Barrow Ruptured Aneurysm Trial, very small aneurysm size was one of the main rea-sons for the high crossover rate from the coiling to the clipping group. 2 A previous meta-analysis that included only 7 studies observed a relatively high complication rate, especially in terms of periprocedural rupture risk. 7Since then, several technologic improvements, including better microcatheters and steerable soft microguidewires, compliant and easier-to-navigate balloons, and the availability of newer distal access catheters and very small endovascular coils have been developed, which may have enhanced our ability in coiling very small aneurysms.Since publication of the previous meta-analysis, many recent series have outlined results and complications in this specific subset of aneurysms.
In an attempt to examine the current safety and efficacy of endovascular treatment of tiny intracranial aneurysms, we performed an updated systematic review and meta-analysis of the literature addressing the endovascular treatment of very small intracranial aneurysms.We also compared results from studies included in a previously published meta-analysis from 2010 7 with more recently published studies to determine whether there have been improvements in outcomes with time.We hypothesized that recently published studies would demonstrate lower intraoperative rupture rates and higher rates of aneurysm occlusion.

Literature Search
Studies were identified by a search of Scopus, Medline, and the Web of Science for studies on the endovascular treatment of very small (Յ3 mm in diameter) intracranial aneurysms published between January 1996 and May 2015.The search was performed by using the following keywords: "cerebral aneurysm," "intracranial aneurysm," "coil," "small," "tiny," and "endovascular" in both AND and OR combinations.The eligibility assessment of the articles was initially performed independently by 2 authors (W.B. and V.N.Y.).In case of disagreement, a third author (G.L.) was consulted for a final decision.Reference lists of included articles were scanned as an additional means of identifying articles.

Study Selection
Consecutive case series studying the endovascular treatment of very small intracranial aneurysms (Յ3 mm in diameter) with Ն10 patients were included in this meta-analysis.Participants of any age with very small saccular IAs were considered.Studies addressing dissecting or "blister" IAs were excluded because these lesions have unique natural histories and pathologic characteristics.Series studying IAs in a specific location were also excluded to avoid selection bias.
Patients were divided into 3 groups: 1) unruptured coiled aneurysms, 2) ruptured coiled aneurysms, and 3) stent-assisted coiled aneurysms, ruptured and unruptured.Patients who received a stent without coiling were added to the stent-assisted group.In 1 article, 1 patient was treated with Onyx (Covidien, Irvine, California), thus excluding the article from our analysis.No patients were treated with flow-diverting stents.

Data Abstraction
Information was extracted through a prespecified data-extraction protocol.From each included study, we collected the following data: 1) immediate and long-term angiographic occlusion, 2) technical success, 3) recanalization rate, 4) retreatment rate, 5) morbidity and mortality related to the procedure, 6) procedurerelated rupture, 7) procedure-related thromboembolism, 8) long-term neurologic outcome, and 9) neurologic outcome at discharge.
The angiographic outcome data were divided into 2 groups: complete or near-complete angiographic occlusion and incomplete occlusion.Results reporting the angiographic findings by using the Raymond grading system (class 1, complete obliteration; class 2, neck remnant; class 3, aneurysm remnant) were translated to our methods as follows: Classes 1 and 2 were considered complete or near-complete angiographic occlusion, and class 3 was included in the incomplete-occlusion group.
The technical success was assessed in only prospective analysis or studies that reported failure of the endovascular treatment attempt.Hemorrhagic or ischemic events were not imputed to rupture or thromboembolism related to the procedure if they were not clearly described as a consequence of the procedure.With regard to the clinical outcome, only studies using the Glasgow Outcome Scale or the modified Rankin Scale were included.Good neurologic outcome was considered for grades 4 and 5 in the Glasgow Outcome Scale and for grades 0 and 1 in the mRS.Procedure-related morbidity and mortality were defined as morbidity or mortality related to thromboembolic or bleeding complications that occurred during the procedure.
We also compared results from studies included in a previously published meta-analysis from 2010 (time period 1) with more recently published studies (time period 2) to determine whether there were improvements in outcomes with time. 7For this comparison, we studied only outcomes of the pooled ruptured/unruptured group.

Statistical Analysis
All included studies were noncomparative.We estimated from each cohort the cumulative incidence (event rate) and 95% confidence interval for each outcome.Event rates for each intervention were pooled in a meta-analysis across studies by using the random-effects model. 8Anticipating heterogeneity among studies, we chose this model a priori because it incorporates withinstudy variance and between-study variance.Heterogeneity of the treatment effect across studies was evaluated by using the I 2 statistic. 9We considered an I 2 statistic of Ͼ50% to indicate substantial heterogeneity.We were unable to test for publication bias due to the noncomparative nature of the studies.The metaanalysis was performed by using the statistical software package Comprehensive Meta-Analysis, Version 3.0 (Englewood, New Jersey).

Characteristics of Included Studies
The literature search yielded 1046 articles, of which 22 met our inclusion criteria.A summary of the literature search process is All 22 studies included were single-center case series.Nine studies provided data on only ruptured IAs, 2 provided data on only unruptured IAs, and 11 included both ruptured and unruptured aneurysms.A total of 1105 intracranial aneurysms that received endovascular treatment (844 ruptured and 261 unruptured) were included in this meta-analysis.Stent-assisted coil embolization was performed in 86 cases.Table 1 provides a summary of included studies.
Comparison of Outcomes across Time.When we compared outcomes between studies included in the prior meta-analysis and those included in the current meta-analysis, angiographic outcomes were superior in studies included in the meta-analysis because initial complete/near-complete angiographic occlusion rates were 89.0%compared with 85.0% in recently published series (P ϭ .004). 7Long-term (Ն6-month) good clinical outcome rates were similar between studies (93.0% for time period 1 and 94.0% for time period 2).Procedure-related rupture rates were significantly lower in more recent studies (7.0% versus 3.0%, P ϭ .007).Retreatment rates were similar between time periods (5.0% for time period 1 and 4.0% for time period 2, P ϭ .93).These data are summarized in Table 3.

Study Heterogeneity
No I 2 statistics were higher than 50%, indicating substantial heterogeneity, for any outcomes in unruptured aneurysms or in stent-assisted coil-treated aneurysms.I 2 statistics were higher than 50%, indicating substantial heterogeneity, for the following outcomes in ruptured aneurysms: complete/near-complete initial occlusion and good neurologic outcome at discharge.I 2 statistics were higher than 50%, indicating substantial heterogeneity, for the following outcomes in all aneurysms: initiation complete/ near-complete occlusion, long-term good neurologic outcome, and recanalization.These data are summarized in Table 2.

DISCUSSION
Our updated meta-analysis of endovascular coiling of very small intracranial aneurysms demonstrated that coil embolization is effective, with angiographic occlusion rates of Ͼ90%.However, complications rates were not negligible at 6.5% for ruptured and 5% for unruptured aneurysms.Intraoperative rupture rates have decreased with time suggesting that technologic improvements may be enhancing patient outcomes.When we compared earlier experiences with more recent ones, there has been some improvement in complication rates as a result of newer technologic advances 5 and increased experience, which makes endovascular coil embolization a safe treatment technique in patients with ruptured very small aneurysms.Our results suggest that endovascular treatment of very small unruptured aneurysms should be pursued only in exceptional cases, given the very low risk of rupture of these aneurysms.
Our results showed high rates of complete or near-complete occlusion, both immediately following treatment and at followup.Technical success was achieved in 92% of patients, and recanalization rates were low.Only 7% of aneurysms were retreated with coiling or microsurgical clipping.In the prior meta-analysis including 7 studies, we reported slightly higher occlusion rates at short-and long-term, possibly due to factors related to publication and selection bias. 7Recently published series are reporting complete occlusion rates as high as 96% in the immediate postoperative setting and 94% in the long-term. 10Our report might be influenced by different treatment strategies adopted in the included studies.For example, Lu et al 11 suggested that complete coil occlusion of the aneurysm sac is not necessary to achieve good long-term outcomes.For challenging aneurysms with a very high rate of intraprocedural rupture, they suggested an initial partial occlusion, allowing progressive stable thrombus formation and complete occlusion in the long-term.
Intraprocedural rupture is a feared complication in the endovascular treatment of tiny aneurysms. 4,7,12Nguyen et al 4 found a relative risk of intraprocedural rupture of 5.2 (95% CI, 2.2-12.8)when comparing IAs of Յ3 mm versus Ͼ3 mm.However, there have been improvements in the procedural rupture rates with time.In a prior study, we reported an intraprocedural rupture rate of 10.7% of the ruptured IAs during the coiling procedure compared with 9% in the current study. 7The first large series addressing the endovascular treatment of tiny IAs reported an 11% rate of procedure-related rupture, 5 times higher compared with larger IAs. 4 However, recent reports have demonstrated considerably lower risk, with rates of approximately 3%-4%.Improvements in technologies with smaller and softer coils, in addition to increased operator experience, have likely contributed to this finding. 5New strategies and decisions by operators to avoid procedural rupture, such as partial initial occlusion of the sac as the treatment goal 11 and a lower threshold to convert to microsurgical clipping, 13,14 may be contributing to decreasing complication rates as well.
High rates of procedural rupture are expected in the treatment of tiny IAs.The smaller size of the aneurysm sac limits the movement of the microcatheter; thus, any unexpected movement during catheter positioning or coil deployment can result in rupture of the aneurysm sac. 10, 15 Cases of intraprocedural perforation are generally managed through reversal of anticoagulation followed by coiling of the aneurysm.The placement of a balloon at the side of the IA neck to stop hemorrhage has been advocated; however, the use of additional adjunctive devices during treatment of very small intracranial aneurysms has been associated with increased complication rates in some studies. 4,10,16lthough very helpful in packing coils and, in some cases, diverting flow from the aneurysmal sac, 17 stent placement has been associated with higher periprocedural complication rates in the endovascular treatment of tiny IAs, as demonstrated in our study. 4,18Lower rates of immediate angiographic occlusion were likely due to the less attenuated tamponade required for satisfactory results in stent-coiling procedures.Most interesting, recanalization rates were higher for aneurysms treated with stent-assisted coiling.The higher recanalization rate in the stent-assisted group might be related to antiplatelet therapy required with the use of intraluminal stents and the wider neck size of aneurysms requiring stent-assisted treatment.
In the management of very small unruptured intracranial aneurysms, it is very important to consider the natural history of these aneurysms.Sonobe et al, 19 reported an annual risk of rupture of 0.34% and 0.95% for single and multiple unruptured small aneurysms, respectively.Moreover, on the basis of the Population, Hypertension, Age, Size of Aneurysm, Earlier Subarachnoid Hemorrhage from Another Aneurysm, and Site of (PHASES) Aneurysm score, 20 patients with aneurysms of Ͻ7 mm in diameter were subject to 0%-7% risk of rupture within 5 years according to the patient's risk factors.Therefore, the very low probability of rupture in these aneurysms must be considered against the risk of procedural complications and the morbidity/mortality rates related to treatment.Given the low risk of rupture, very small unruptured aneurysms should not be treated except in selected cases. 21

Limitations
Our study has limitations.Fifteen of 22 included studies were retrospective; this feature might influence the outcomes pre-sented and increase the publication bias.A wide methodologic variability was found among studies, especially in assessing angiographic and clinical outcomes.Thus, for certain analyses, a limited number of studies were included.Moreover, lack of details about patients also hindered our analysis.The specific locations of IAs were not assessed in this meta-analysis due to lack of available data.Finally, the long-term angiographic and clinical outcomes were evaluated at different times in the studies analyzed.Publication bias is another limitation of this study.Nevertheless, our meta-analysis is currently the largest study assessing the endovascular treatment of very small intracranial aneurysms with Ͼ1000 included patients.

CONCLUSIONS
On the basis of our results, we conclude that the coiling of very small intracranial aneurysms can be performed safely and effectively with favorable long-term angiographic and neurologic outcomes.However, complications rates were not negligible at 6.5% for ruptured and 5% percent for unruptured aneurysms.These findings highlight the importance of individualized decisionmaking based on aneurysm natural history and treatment risks.

FIG 1 .
FIG 1. Flow diagram of the search strategy and study selection.

Table 2 : Meta-analysis outcomes
Note:-NA indicates not available.