Flow-diverter stents in intracranial aneurysm treatment: impact on covered cerebral artery branches

Objective: Flow diverter stents (FDSs) have attracted interest for intracranial aneurysm (IA) treatment; however, occlusion of side branches and related complications have been reported. This study aimed to investigate the effects of FDSs in IA management when different branches of intracranial arteries are covered. Materials and methods A cross-sectional study was conducted using PUBMED, Embase, Web of Science, and Cochrane databases to include randomized or nonrandomized comparative-designed studies from January 2000 to August 2022 which reported outcomes of occlusion/narrowing of branches after IA treatment using FDSs. The PRISMA guidelines were used for our report. A random-effects meta-analysis was conducted to pool the outcomes, which included incidence rates of occlusion/narrowing of FDS-covered branches, branch occlusion-related symptoms, obliteration of IAs, and ideal clinical outcomes (modified Rankin Scale score ≤2). Results: The authors identified 57 studies involving 3789 patients with IA managed by FDSs covering different branches. During the median imaging follow-up at 12 months, the IA obliteration rate was satisfactory (>70%) when covering the ophthalmic artery (OA), posterior communicating artery (PComA), anterior choroidal artery (AChoA) or anterior cerebral artery (ACA), but not the middle cerebral artery-M2 segment (MCA-M2; 69.5%; 95% CI: 60.8–77.5%) and posterior inferior cerebellar artery (PICA; 59.1%, 13/22). The overall ideal clinical outcome was observed in 97.4% of patients (95% CI: 95.5–98.9%). Higher rates of occlusion/narrowing of branches were identified when FDSs covered the ACA (66.6%; 95% CI: 45.1–85.3%), PComA (44.3%; 95% CI: 34.2–54.6%), or MCA-M2 (39.2%; 95% CI: 24.5–54.7%); the risks were lower when covering the OA (11.8%; 95% CI: 8.8–15.1%), PICA (6.8%; 95% CI: 1.5–14.5%), and AchoA (0.5%; 95% CI: 0.0–2.9%). The risk of branch occlusion-related complications was low (incidence rate <5%) for each of the six evaluated branches. Conclusions: Acceptable outcomes were identified following treatment of IAs when FDSs were placed across each of the six studied cerebral arteries. Treatment decisions regarding FDS placement across branch arteries should be made with the risk of complications from branch occlusion in mind.


Introduction
The rupture of intracranial aneurysms (IAs) results in poor outcomes for ~35% of patients with subarachnoid hemorrhage (SAH) [1] .Endovascular treatment has become a prime choice for IA treatment and the prevention of IA rupture.Flow diverter stents (FDSs), a customized tool, were initially designed for the management of large and wide-necked IA located in the internal carotid artery (ICA), and displayed immediate and long-term efficacy and safety [2] .As the technique matures, there have been several new attempts to apply FDS to treat IAs in the posterior circulation and other arteries beyond the circle of Willis [3] .However, when placed in a small-diameter artery, FDSs can induce stenosis or occlusion of the parent artery because of their higher metal mesh density compared to conventional stents.The induced changes in hemodynamics can reduce blood flow in FDScovered vessels, accelerate thrombus formation within the branches, and promote IA occlusion, increasing the risk of intracranial thrombosis, and becoming an important cause of severe complications and mortality.Reliable evidence regarding the risks of occlusion in different branches, related complications, and their potential mechanisms is needed to demonstrate the safety and efficacy of the procedure, and whether the indications for FDSs could be extended to IAs beyond the ICA in future clinical applications.Thus, this study aims to comprehensively summarize the existing evidence and assess the patency of intracranial arterial branches and related complications when they are covered by FDSs in the treatment of IAs.

Materials and methods
This study adhered to strengthening the reporting of cohort studies in surgery (STROCSS) criteria [4] (Supplemental Digital Content 1, http://links.lww.com/JS9/B225),PRISMA 2020 guidelines [5] and AMSTAR guidelines [6] and was registered with PROSPERO (CRD42022350339).Three authors (J.Y.L., F.C., N.B.Z.M.) independently performed record screening and data collection, while two authors (J.Y.L. and F.C.) conducted the risk-of-bias assessment.Subsequently, another author (Y.X.G.) reviewed the information for accuracy.Disagreements were discussed and resolved with J.X.Y. to reach a consensus.

Study selection
The studies included had a randomized or nonrandomized comparative design and had enrolled a minimum of 10 patients with IA.Studies needed to report data on occlusion/narrowing rates of branches and involve patients with IAs that had been treated with FDSs, regardless of the IA's location.We excluded reviews, case reports, animal studies, conference abstracts, and case series with fewer than 10 patients.

Data extraction
The characteristics of each included study were extracted according to a prespecified procedure.The outcomes of efficacy were identified as obliteration of IAs and ideal clinical outcomes (using the modified Rankin Scale [7] , less than 2, at the last clinical follow-up), and the outcomes of complications included occlusion/narrowing of FDS-covered branches and branch occlusionrelated symptoms.Patients experiencing a symptomatic acute stroke event were included as having a branch occlusion-related symptom, whether the condition was temporary or permanent.Most of these patients were evaluated through angiography when clinical symptoms appeared.For nonsymptomatic patients, branch occlusion/narrowing was identified in routine imaging follow-up.Details were described in Supplemental Methods (Supplemental Digital Content 2, http://links.lww.com/JS9/B226).

Risk-of-bias assessment
The risk-of-bias in each study was rated using the Quality in Prognosis Studies (QUIPS) tool [8] developed for prognosis studies.Six domains were considered to evaluate the validity and bias in prognostic factors: study participation, study attrition, prognostic factor measurement, confounding measurement and accounting, outcome measurement, and analysis and reporting (details in Appendix 2, Supplemental Digital Content 1, http:// links.lww.com/JS9/B225).We categorized the overall quality of the studies based on the risk-of-bias in each domain: studies with no domain at high risk were considered 'high quality,' those with one domain at high risk were 'moderate quality,' and those with more than one domain at high risk were deemed 'low quality.'

Statistical analysis
Statistical analyses were performed using the R meta package and software (R version 4.1.2;R Foundation for Statistical Computing).Random-effects meta-analyses of single proportions were conducted to obtain summary estimates of the incidence rates of outcomes for conservative results.The incidence rates of branch occlusion-related symptoms were estimated according to the risk of complications among the included patients or the risk caused by occluded/narrowed branches.Considering the presence of the rates of outcomes at 0% in several studies, the 95% CIs were calculated using the score method, and the variancestabilizing Freeman-Tukey transformation was applied to avoid excluding those studies from evidence synthesis.Cochran's Q test, tau (τ), and Higgins' I 2 statistics were used to evaluate heterogeneity.Subgroup analysis was conducted according to the published year of studies, location of the IA, follow-up period, location of medical centers, sample size, and covered branches [six branches of the ophthalmic artery (OA), posterior communicating artery (PComA), anterior choroidal artery (AChoA), anterior cerebral artery (ACA), middle cerebral artery-M2 segment (MCA-M2) and posterior inferior cerebellar artery (PICA)].We performed a sensitivity analysis in two ways: first, by excluding each study individually to assess the remaining studies, and second, by including only those studies with a sample size of greater than or equal to 20.Funnel plot asymmetry was tested using Egger's test to investigate the association between the sample size and outcomes when at least 10 studies were included.Statistical significance was set at a two-sided P-value of <0.05.When P < 0.05 in Egger's test, the trim-and-fill method was used to reduce publication bias.

HIGHLIGHTS
• Among the 57 studies, 3789 patients with IA were involved who were treated by FDSs.The overall obliteration rate was 77.5%, (95% CI: 74.0-80.9%)and 97.4% of patients (95% CI: 95.5-98.9%)had a modified Rankin Score ≤ 2 at the last follow-up (median, 12 months).• The pooled rates of branch occlusion-related complications were less than 5% in the evaluated cerebral branches.• Acceptable safety and efficacy were identified in the treatment of IA when FDSs were placed across branches.

Results
A total of 57 studies  , encompassing 4647 records, involved 3789 patients with IA who were treated using FDSs. These tudies employed a nonrandomized comparative design; the selection process is shown in Figure S1 (Supplemental Digital Content 2, http://links.lww.com/JS9/B226).The median total sample size was 28 patients (interquartile range: 19.5-58.5), the median age of participants was 54 years (52-58), and the median percentage of men was 25% (16-36%).The median imaging follow-up period was 12 months (7.9-15.2).In 75.4% (43/57) of the studies, outcomes were assessed using digital subtraction angiography as the sole imaging modality.The detailed characteristics of each study are listed in Table 1 and S2. ) and statistical analysis and reporting (66.7%, 38/57).However, 93.0% (53/57) failed to measure important confounders and 45.6% (26/57) did not provide details on participants lost to follow-up.A summary of these assessments is provided in Table S1 (Supplemental Digital Content 2, http://links.lww.com/JS9/B226).There was no evidence of publication bias for any outcome except the overall occlusion/narrowing of covered branches.

Discussion
This study systematically included 57 studies involving 3789 patients with IA treated with FDSs covering intracranial branch arteries.Our synthesis of evidence demonstrated the safety and efficacy of FDSs in managing IAs; 97.4% of patients had ideal clinical outcomes, and 77.5% of IAs were occluded at the last follow-up.However, further research is warranted considering the limited sample size, especially for IAs located distal to the internal carotid artery.Despite higher rates of occlusion/narrowing identified in the covered branches of ACA, PComA, and MCA-M2 compared to those of the OA, PICA, and AChoA, the risk of related complications was acceptable.Our findings showed a relatively higher risk of occlusion/narrowing for ACA, PComA, and MCA-M2 branches covered by FDSs; however, related clinical complications were rarely observed when covering ACA or PComA branches.This suggests that compensatory cerebral blood flow arises from the contralateral ACA via the anterior communicating artery when the ACA is blocked, and from the posterior cerebral artery when the PComA is blocked.However, ischemic symptoms may occur if predominant arteries are covered and there is insufficient compensation from other branches [66] .The M2 segment of the MCA, due to its relatively distal location, narrower diameter, and thinner walls, struggles to receive sufficient reperfusion when covered by FDSs.A study [67] on MCA-M2 thrombectomy reported that 58.1% of patients with MCA-M2 occlusion had good outcomes and another study [68] showed a similar rate of successful MCA-M2 reperfusion, whereas our results showed relatively lower branch occlusion-related complications when the MCA-M2 was covered by FDSs.Because hemodynamics in FDScovered branches can decrease gradually rather than abruptly in an acute thrombotic event, allowing time for collateral circulation to establish.Infarction of the AChoA can lead to a devastating outcome, AChoA syndrome, manifesting as the triad of hemiparesis, hemianesthesia, and hemianopia [69] .The absence of AChoA infarction was identified as an independent predictor of good outcomes for acute ICA occlusion [70] .However, we found a low risk of AChoA occlusion/narrowing and an unsatisfactory outcome despite cases without considering cases of ICA in-stent occlusion.Furthermore, when FDSs were placed across the OA, only 13.0% were occluded or narrowed, and 2.2% of patients exhibited new clinical symptoms.However, approximately onethird of patients with their OA occluded/narrowed meanwhile experienced new complications.Considering that paraclinoid aneurysms can grow to a large size and have a lower rate of rupture compared to other aneurysms, the source of new symptoms-either from OA occlusion or mass effect caused by coils or IA thrombogenesis-is difficult to pinpoint [71] .Additionally, no new symptoms were observed in patients with a PICA covered by FDSs.We hypothesize that the contralateral PICA or ipsilateral anterior inferior cerebellar artery may compensate for blood flow; however, further investigations with a larger sample size are needed to verify this.
The use of antiplatelets for maintaining branch patency allows for continued protection against further thrombosis induced by FDSs and coils, while minimizing hemorrhagic complications [72] .However, a subgroup analysis could not be conducted due to the variability in antiplatelet regimens among the included studies.Several studies [9,10,12,14,20,22,27,31,35,38,45,52,56,58,60,62,63] have reported symptomatic thrombotic events of in-stent thrombosis or branch occlusion/narrowing caused by the inappropriate use of antiplatelets; however, several of the included studies did not elaborate on this, so we could not evaluate this association accurately.Additionally, branch occlusion may be observed in patients with in-stent thrombosis; however, these cases were not classified in our study.Consequently, the risks of branch occlusion and related complications may be underestimated.
In IAs management, covering branches during FDSs placement should be avoided to reduce hemodynamic perturbations beyond IAs.If this is not feasible, our results suggest that placing FDSs across the OA, PComA, AChoA, MCA-M2, or ACA is relatively safe.Before the procedure, it is crucial to evaluate the vascular anatomy and hemodynamics to ensure adequate blood supply from collateral circulation.Strict administration of anticoagulants and antiplatelets should be initiated during the perioperative period to prevent thrombotic complications.Considering the occlusion rate of both IA and MCA-M2, caution is advised when placing FDSs at the MCA-M2 bifurcation, especially since microsurgery remains a feasible and safe alternative [73] .There is insufficient evidence on the treatment of IAs using FDSs covering the PICA or other unspecified cerebral arteries, warranting further research to determine their feasibility, effectiveness, and safety.
The risk-of-bias in the included studies mainly arises from issues controlling for confounders and loss to follow-up.Given that FDS is a relatively new innovation, earlier studies employed varied approaches in treating patients with IA, including different perioperative antiplatelet regimens, selection of implanted materials (FDS only, overlapping FDSs, or FDS with coiling), and imaging modalities.This has led to a heterogeneity in reported outcomes.It should be noted that clinicians had avoided obstructing branches while deploying FDSs, due to the risk of thrombosis caused by the decrease of blood flow in FDS-covered vessels.This caution has limited the sample size and design of existing studies.To assess the robustness of our pooled results, we conducted an additional sensitivity analysis among the studies with a sample size ≥ 20 patients, showing consistent results (Tables S3-S6, Supplemental Digital Content 2, http://links.lww.com/JS9/B226).Randomized clinical trials (RCTs) were valuable but not included in our study because the existing trails of PARAT [74] or FIAT [75] did not report on the outcomes of interest.Besides, the randomization was conducted into treatment groups of FDS or coiling with/without stent assistant instead of jailing branches or not jailing branches, and the potential confounders still existed.Even though 12.3% (7/57) of the included studies lost more than 20% of patients in follow-up, there were still a large proportion of included studies (42.1%, 24/57) without loss of participants or reporting details.Another limitation is that the available data restricted further analysis of different types of FDSs and antithrombotic regimens, as well as the use of overlapping or telescopic FDSs, which would also increase the risk of thrombosis due to a higher metal coverage on both IAs and branches.Additionally, significant heterogeneity exists due to variations in surgical techniques and skills among different operators, thus, perspective multicenter cohorts with standardized procedures and adequate follow-up are needed.
Our evaluation comprehensively assessed the occlusion/narrowing of FDS-covered branches and related complications, highlighting substantial evidence gaps in the management of IAs beyond the ICA.Acceptable outcomes were noted when FDSs were placed across the OA, PComA, AChoA, ACA, and MCA-M2.However, neurosurgeons, neurologists, and interventional neuroradiologists should be aware of the risk of complications of branch occlusions when making clinical decisions and should conduct precise preoperative assessments for individualized treatment.The limitations of existing studies underscore the urgent need for multicenter pragmatic trials aimed at long-term, patient-centered outcomes concerning the treatment goals for IA occlusion and fewer complications.

Figure 1 .
Figure 1.Forest plot for obliteration rates of intracranial aneurysms varied by different flow diverter stents-covering branches.

Figure 2 .
Figure 2. Forest plot for the rate of ideal clinical outcomes varied by studied populations.

Figure 3 .
Figure 3. Forest plot for rates of branch occlusion/narrowing varied by different flow diverter stents-covering branches.

Figure 4 .
Figure 4. Forest plot for rates of branch occlusion-related symptoms (per patient) varied by different Flow diverter stents-covering branches.

Table 1
Summary of included studies and baseline characteristics.