Efficacy and safety of flow diverters in posterior circulation aneurysms and comparison with their efficacy in anterior circulation aneurysms: A systematic review and meta-analysis

Abstract

Purpose

We aimed to assess the efficacy and safety of flow-diverter stents (FDs) in the management of posterior circulation cerebral aneurysms and compare FD efficacy between anterior and posterior circulation aneurysms.

Methods

We searched the PubMed, Scopus, Cochrane, and Web of Science databases for relevant studies through March 2020. Studies assessing FDs for posterior circulation aneurysms that included ≥20 treated aneurysms were included. Moreover, the studies compared FD efficacy between anterior and posterior circulation aneurysms were included. Data regarding angiographic aneurysmal occlusion, procedural complications, mortality, and morbidity were extracted and pooled in a random-effects meta-analysis model.

Results

Fourteen studies with a total of 659 patients and 676 posterior circulation aneurysms were included. The pooled rate of aneurysmal occlusion at long-term angiographic follow-up was 78% [95% confidence interval (CI), 71–85]. The pooled rates of intraparenchymal hemorrhage, ischemia, and procedure-related mortality and neurological morbidity were 2%, 8%, 7%, and 6%, respectively. Complete occlusion occurred in 82.4% of the posterior circulation aneurysm subgroup and 77.5% of the anterior circulation aneurysm subgroup. The difference was not significant (relative risk 1.01; 95% CI, 0.86–1.19; p = 0.91). Regression analysis showed that elderly patients and females had higher morbidity.

Conclusion

Posterior circulation aneurysms can be effectively treated with FDs with comparable occlusion rates to those in anterior circulation aneurysms. However, periprocedural complications are not negligible.

Introduction

Endovascular therapy is the treatment of choice for ruptured and unruptured intracranial aneurysms, as it is superior to surgical clipping.¹,² However, large and wide-neck aneurysms are generally not appropriate for traditional coiling due to lower complete occlusion rates, which is associated with risk of post-treatment rupture.³,⁴ Recently, flow-diverter stents (FDs) have been employed as an alternative approach in treating intracranial aneurysms, mainly fusiform, wide-neck, and large aneurysms.⁵,⁶

FDs are inserted within the parent artery rather than the aneurysmal sac.⁷ Unlike regular stents used in stent-assisted coiling, FDs are low porosity with a greater coverage area and thus cover the aneurysmal ostium, preventing intra-aneurysmal blood flow and lowering the rate of recanalization.⁸ The first FD to be approved by the FDA was the Pipeline embolization device (PED) (Covidien, Mansfield, MA, USA) in 2011⁹. Other FDs include the Silk stent (Balt Extrusion, Montmorency, France), Surpass (Stryker, Neurovascular, Fremont, CA, USA), flow redirection endoluminal device (FRED) Microvention, Tustin, CA, USA), and p64 Flow Modulation Device (phenox, Bochum, Germany). Reported rates of aneurysmal occlusion and complications for each device have varied.⁷,⁹,¹⁰

FDs are commonly used for anterior circulation aneurysms.¹¹ They have been recently introduced to treat posterior circulation aneurysms and are considered preferable in dissecting and fusiform aneurysms due to their inherent endoluminal reconstructive nature.¹² However, FD usage for posterior circulation aneurysms is still considered off-label.¹³ Moreover, their safety has not been well-defined and many studies have shown that posterior circulation aneurysms are associated with a higher rate of ischemic complications due to occlusion of perforators.¹⁴

We performed this systematic literature review to assess the efficacy and safety of FDs in the management of posterior circulation cerebral aneurysms and compare their efficacy between anterior and posterior circulation aneurysms.

Methods

All steps of this systematic review were performed according to the Cochrane handbook of systematic reviews and meta-analysis. We also followed the preferred reporting items for systematic reviews and meta-analyses (PRISMA) statement guidelines while drafting our manuscript.¹⁵,¹⁶

Search strategy

We searched the PubMed, Scopus, Web of Science, and Cochrane databases through March 2020 to retrieve all published studies reporting original data on the role of FDs in posterior circulation aneurysms. The search strategy was formulated using the following terms: “posterior circulation aneurysm,” “basilar artery aneurysm,” “posterior cerebral artery aneurysm,” “vertebral aneurysm,” “posterior inferior cerebellar aneurysm,” “flow diverter”, “flow diversion”, “pipeline”, “PED”, “SILK”, “p64”, “FRED,” and “SURPASS” in both “AND” and “OR” combinations. All references from the retrieved studies were further examined to identify additional relevant studies.

Study eligibility

Three reviewers (M.A., A.R., and A.K.A) independently evaluated the studies for inclusion eligibility. Any discrepancies regarding study inclusion were resolved by mutual agreement. Studies meeting the following criteria were included:

1. Population: studies including at least 20 posterior circulation aneurysms, regardless of size and shape, treated with FDs

2. Study design: both prospective and retrospective

3. Outcome: studies reporting at least one of the following outcomes: aneurysmal occlusion rate, procedure-related neurological morbidity, mortality, and complications.

We excluded review articles, case reports, experimental or animal studies, non-English language studies, and studies reporting data from overlapping sample populations.

Study selection

Titles and abstracts of retrieved records were screened for eligibility, followed by full-text screening. If all authors stated that a study did not meet inclusion criteria, the study was excluded. Any disagreements were resolved by discussion. Prospective and retrospective studies that evaluated FDs in the treatment of posterior circulation cerebral aneurysms were selected. In cases of studies with overlapping patient populations, we selected the series with the largest number of subjects or the most detailed data. If data from a single-center study was included in another multicenter one, we excluded the single-center study.

Data extraction

Three reviewers (M A, AM, and M A H T) independently extracted the following data from articles using data abstraction forms: general study description (including the study name, year, title, journal, study design, centers, country, study duration, median angiographic follow up in months, and type of FD), baseline characteristics (number of patients, aneurysms, and procedures; prior treatment; aneurysmal rupture; patient age and gender; and aneurysmal size, location, and morphology), and outcomes: aneurysmal occlusion (based on O’Kelly Marotta grading scale (OKM)¹⁷) the average number of FD devices, retreatment of aneurysms, postoperative aneurysmal rupture, hemorrhagic complications, ischemic complications, parent artery occlusion, perforator artery occlusion, mortality, morbidity, early and delayed, transient and permanent complications. In the studies where data was available, we compared anterior and posterior circulation aneurysm occlusion rates.¹²,^18–21 Any discrepancies were resolved through discussion.

Endpoint/outcomes

The primary objectives of the study were to define the efficacy (aneurysmal occlusion, retreatment, and adjuvant coiling) and the safety (intraparenchymal hemorrhage and ischemia, postoperative aneurysmal rupture and perforator artery occlusion, mortality and morbidity rates of FD across the posterior circulation. The secondary objective was to compare FD efficacy between anterior and posterior circulation aneurysms.

Risk of bias assessment

Three reviewers (M A, AM, and M A H T) separately graded study quality using the Newcastle-Ottawa Scale,²² which assesses the quality of non-randomized studies in meta-analyses. A star rating of 0–9 was allocated to each study based on three domains: S, selection (0–4); C, comparability (0–2); and O, outcome (0–3). Good quality was defined as 3 or 4 stars (★) in the selection domain AND 1 or 2 stars in the comparability domain AND 2 or 3 stars in the outcome domain. Fair quality was defined as 2 stars in the selection domain AND 1 or 2 stars in the comparability domain AND 2 or 3 stars in the outcome/exposure domain. Poor quality was defined as 0 or 1 star in the selection domain OR 0 stars in the comparability domain OR 0 or 1 star in the outcome/exposure domain.²²,²³

Statistical analysis

Statistical analyses were performed using STATA 16.0 and RevMan 5.3 software. The event rate with 95% confidence interval (CI) were calculated for each outcome. The Cochran Q test was used to compare the effect estimates among the included studies. Inconsistency was assessed by testing the chi-square distribution of the Cochran Q values with p < 0.1 indicating significant statistical heterogeneity between studies. Heterogeneity across studies was also investigated using the I² statistic, where studies with I² values <50% were considered to have an acceptable level of statistical heterogeneity. In the case of significant heterogeneity, we used a random-effects model meta-analysis. Meta-regression analyses were used to identify potential moderator variables (age, gender, aneurysmal size, aneurysmal location, aneurysmal morphology, mass effect, number of FDs used, and median angiographic follow-up duration) of both aneurysmal occlusion and neurological morbidity. P < 0.05 was considered significant.

Results

Literature search results

Our search retrieved 1799 studies. Of these, 311 were duplicates. After screening titles and abstracts, 1377 studies were excluded. We assessed the full text of 111 studies and excluded 97 based on criteria. Finally, 14 studies^{12,18–21,24–32} were included for meta-analysis. Figure 1 shows a flow diagram of the study selection process.

Figure 1.

Flow diagram of the study selection process.

Characteristics of the included studies

Of the 14 included studies, 12 were retrospective^{12,19–21,24–30,32} and 2 were prospective.¹⁸,³¹ Eight studies were multicenter and six were single-center case series. Thirteen studies^{12,18–21,24–29,31,32} reported aneurysm occlusion rates and all reported procedure-related neurological complication rates.

Our meta-analysis included a total of 659 patients with 676 posterior circulation aneurysms treated with FDs. Mean patient age was 55.5 ± 4.3 years and 51.4% were male. Mean aneurysm size was 10.7 ± 1.9 mm. The characteristics of the included studies are shown in Table 1.

Table 1.

Characteristics of the included studies.

Study name	Year	Study design	Centers	Country	Study duration	Median Angiographic follow up (mo)	Type of FD	Patient number	Procedure number
Zhang	2019	Retrospective	Single-center	China	2015–2016	4.9	PED	30	32
Wallace	2019	Retrospective	Multicenter	USA	–	14	PED	35	36
Bender	2019	Retrospective	Single centre	USA	2011–2017	11.8	PED	55	59
Griessenauer	2018	Retrospective	Multicenter	USA, Europe	2009–2016	11	PED	129	129
Zammar	2018	Retrospective	Multicenter	USA	2011–2015	9	PED	21	22
Liang	2018	Retrospective	Single-center	China	2015–2016	5.5	PED	35	37
Lopes	2017	Retrospective	Multicenter	USA, Irland, Korea, Turkey, Argentina	2008–2013	21.1	PED	91	NA
Taschner	2017	Retrospective	Multicenter	USA, UK, Germany, Japan, Netherlands, Argentina, Hungary, France	–	11.3	Surpass	52	53
Bhogal	2017	Retrospective	Single-center	Germany	2009–2016	25.2	PED/P64	56	NA
Wakhloo	2015	Prospective	Multicenter	Europe, South America, and Japan		6	Surpass FD	27	NA
Fischer	2014	Retrospective	Single-center	Germany	2009–2013	27.4	PED	38	38
Phillips	2012	Prospective	Multicenter	Australia	2009–2011	21	PED	32	32
Briganti	2012	Retrospective	Multicenter	Italy	2009-2010	3	SILK/PED	37	NA
De Barros Faria	2011	Retrospective	Single-center	Argentina	2007–2010	6	PED	21	21

Quality of evidence assessment

As shown in Table 2, the quality of evidence was moderate in all studies according to the Newcastle-Ottawa Quality Assessment Scale. The average score was 5.7.

Table 2.

Quality assessment of the included studies according to the modified Newcastle-Ottawa Quality Assessment Scale.

Study Name	Selection				Comparability		Outcome			Total
	1	2	3	4	a	b	1	2	3
Retrospective design (score 0–8)
Zhang 2019	*	*			*	*	*		*	6
Wallace 2019	*	*				*	*		*	5
Bender 2019	*	*				*	*		*	5
Griessenauer 2018	*	*				*	*		*	5
Zammar 2018	*	*			*	*	*		*	6
Liang 2018	*	*			*	*	*		*	6
Lopes 2017	*	*				*	*		*	5
Taschner 2017	*	*				*	*		*	5
Bhogal 2017	*	*				*	*	*	*	6
Fischer 2014	*	*			*	*	*	*	*	7
Briganti 2012	*	*			*	*	*		*	6
De Barros Faria 2011	*	*				*	*		*	5
Prospective design (score 0–8)
Wakhloo 2015	*	*	*	*	*	*	*		*	7
Phillips 2012	*	*	*	*		*	*		*	6

FD efficacy

Aneurysmal occlusion

Mean angiographic follow up was 10.9 months (range, 3–27.4). The overall complete aneurysm occlusion rate was 78% (95% CI, 71–85) at the last angiographic follow-up (Figure 2).

Figure 2.

Aneurysmal occlusion rates.

Subgroup meta-analysis of aneurysmal occlusion rates with aneurysms stratified by FD device used showed that a higher occlusion rate was obtained with the PED compared to other devices (Figure 3).

Figure 3.

Subgroup analysis of aneurysmal occlusion rates with aneurysms stratified by flow-diverter stent device used.

A meta-regression analysis was performed to evaluate the potential impact of age, gender, aneurysmal size, location and morphology, mass effect, number of FDs used, and median angiographic follow-up duration on the complete aneurysm occlusion rate. Meta-regression showed that the less mass effect and vertebral aneurysmal location of aneurysm favor good angiographic outcome (p-value = 0.043, 0.032 respectively). None of the other factors were significant predictors of occlusion rate (all p > 0.05; Table 3).

Table 3.

Meta-regression analysis results.

	Regression coefficient	Standard error	P-value	[95% CI]
Aneurysmal occlusion
Age	−0.0133	0.00911	0.144	−0.0312	0.00456
Gender	0.0892	0.19705	0.651	−0.297	0.4754
Aneurysmal size	−0.0044	0.03045	0.886	−0.0641	0.0553
Aneurysmal location	0.278	0.024	0.032	0.024	0.531
Aneurysmal morphology	–0.048	0.092	0.605	–0.228	0.133
Mass effect	–0.009	0.004	0.043	–0.017	–.0000
Number of flow-diverter stents used	−0.0349	0.02973	0.241	−0.0932	0.02339
Median Angiographic follow up duration	−0.0014	0.00549	0.799	−0.0122	0.00936
Neurological morbidity
Age	0.00474	0.00209	0.023	0.00064	0.00884
Gender	−0.1338	0.05227	0.01	−0.2363	−0.0314
Aneurysmal size	−0.0073	0.00669	0.273	−0.0204	0.00578
Aneurysmal location	–0.112	0.034	0.008	–0.227	–0.033
Aneurysmal morphology	–0.130	0.050	0.001	–0.178	–0.046
Mass effect	0.002	0.001	0.083	–0.000	0.005
Number of flow-diverter stents used	0.01118	0.00909	0.218	−0.0066	0.02899
Median Angiographic follow up duration	0.00223	0.0015	0.137	−0.0007	0.00516

Retreatment and adjunctive coiling

Five studies²⁰,²⁴,²⁶,²⁹,³⁰ reported data on the need for another session of treatment. The overall retreatment rate was 3% (95% CI, 0–5; Figure 4).

Figure 4.

Retreatment rates.

Moreover, eight studies¹²,^24–26,^29–32 reported data regarding the need for adjunctive coiling in addition to FD deployment. Overall, 22% (95% CI, 17–27) required adjunctive coiling (Figure 5).

Figure 5.

Adjunctive coiling rates.

Comparison of efficacy in anterior and posterior circulation aneurysms

Five studies including a total of 712 patients compared angiographic outcomes between anterior (n = 570) and posterior circulation aneurysms (n = 142).¹²,^18–21 The complete occlusion rate was 82.4% (117/142) in the posterior circulation group and 77.5% (442/570) in the anterior circulation group. Our analysis showed that the difference was not significant (relative risk (RR) 1.01; 95% CI, 0.86–1.19; p = 0.91; Figure 6).

Figure 6.

Comparison of angiographic outcomes between anterior and posterior circulation aneurysms.

The pooled studies were heterogeneous (p = 0.004; I² = 74%). The heterogeneity was best resolved by a leave-one-out sensitivity analysis that removed Liang (2018). However, the resulting effect estimates showed no significant angiographic occlusion rate difference between the anterior and posterior circulation aneurysms (RR 0.92; 95% CI, 0.82–1.04; p = 0.18; Figure 7).

Figure 7.

Comparison of angiographic outcomes results after leave-one-out sensitivity analysis.

FD safety

In studies reporting data, the pooled rates of intraparenchymal hemorrhage and ischemia were 2% (95% CI, 1–4) and 8% (95% CI, 5–11), respectively (Figures 8 and 9).

Figure 8.

Intraparenchymal hemorrhage rates.

Figure 9.

Ischemia rates.

In terms of the ischemia rates, the pooled studies were heterogeneous (p = 0.02; I² = 52%). The heterogeneity was best resolved by a leave-out-one sensitivity analysis that removed Griessenauer (2018), which resulted in a decrease in overall effect estimate from 8.0% to 6.5%.

The pooled rates of postoperative aneurysmal rupture and perforator artery occlusion were 1% and 5%, respectively (Figures 10 and 11).

Figure 10.

Postoperative aneurysmal rupture rates.

Figure 11.

Perforator artery occlusion rates.

Thirteen studies reported data regarding procedure-related mortality and procedure-related neurological morbidity. Although the pooled studies were heterogeneous, leave-one-out sensitivity analysis showed that the estimated pooled effect size was not driven by one single study. Pooled mortality was 7% (95% CI, 4–8) (Figure 12). Pooled neurological morbidity was 6% (95% CI, 4–8; Figure 13). Meta-regression analysis showed that both age and gender were significant predictors of neurological morbidity (p = 0.02 and 0.01, respectively). Also, the vertebral location and saccular morphology show lower morbidity rate (p-value = 0.008, 0.001 respectively) Table 3.

Figure 12.

Procedure-related mortality.

Figure 13.

Procedure-related neurological morbidity.

Thirteen studies reported data regarding procedure-related mortality and procedure-related neurological morbidity. Although the pooled studies were heterogeneous, a leave-one-out sensitivity analysis showed that the estimated pooled effect size was not driven by one study. Pooled mortality was 7% (95% CI, 4–8) (Figure 12). Pooled neurological morbidity was 6% (95% CI, 4–8; Figure 13). Meta-regression analysis showed that both age and gender were significant predictors of neurological morbidity (p = 0.02 and 0.01, respectively. The less mass effect and vertebral aneurysmal location of aneurysm favor good angiographic outcome (p-value = 0.043, 0.032 respectively), Table 3. Regarding the periprocedural/early events, their rate was 23% (95% CI, 12.4% to 33.8%), while the delayed events rate was 6.8% (95% CI, 1.7% to 11.9%). The rate of transient complications was 15% (95% CI, 7.9% to 22.4%), while the rate of permanent complications was 11.7% (95% CI, 6.3 to 17.2%) (online supplementary Appendix 1).

Discussion

The FD is an emerging treatment option for cerebral aneurysms and currently has a more established role in the anterior circulation. Studies that have investigated FDs in posterior circulation aneurysms have focused on dissecting and fusiform aneurysms and have found acceptable levels of efficacy and safety.

We conducted this systematic review and meta-analysis to investigate FD safety and efficacy in the management of posterior circulation aneurysms. Fourteen studies comprised of 659 patients with 676 aneurysms were included, with a mean angiographic follow up of 10.9 months. The pooled complete aneurysm occlusion rate was 78% and the occlusion rate obtained with the PED device was higher than that obtained with the others. Adjunctive coiling was required in 22%, while another treatment session was needed in 3%.

Complete occlusion occurred in 82.4% (117/142) of patients in the posterior circulation group compared to 77.5% (442/570) of patients in the anterior circulation group. However, the difference was not significant (RR 1.01; p = 0.91). Postoperative aneurysmal rupture and perforator artery occlusion occurred in 1% and 5%, respectively. Intraparenchymal hemorrhage and ischemia occurred in 2% and 8%, respectively. Procedure-related mortality and neurologic morbidity occurred in 7% and 6%, respectively.

In 2016, Wang et al. performed a meta-analysis to investigate FDs in posterior circulation aneurysms.³³ They included 14 studies with a total of 220 patients with 225 aneurysms and reported a pooled aneurysmal occlusion rate of 84% (95% CI, 60–94), ischemia rate of 11% (95% CI, 7–17), intraparenchymal hemorrhage rate of 4% (95% CI, 1–8), postoperative subarachnoid hemorrhage rate of 3% (95% CI, 1–6), and mortality of 15% (95% CI, 1–21). Another meta-analysis performed by Kiyofuji et al. investigated FDs in non-saccular posterior circulation aneurysms and included 13 studies with a total of 129 patients with 131 aneurysms.³⁴ They reported a pooled long-term aneurysmal occlusion rate of 52% (95% CI, 29–76), average number of FDs per treated aneurysm of 4.33, retreatment rate of 5% (95% CI, 0–14), ischemia rate of 23% (95% CI, 10–39), morbidity of 26% (95% CI, 12–42), and mortality of 26% (95% CI, 12–42). A recent meta-analysis by Liang et al. that investigated the PED in posterior circulation aneurysms included 12 studies with 358 patients with 365 aneurysms.³⁵ They reported a pooled long-term aneurysmal occlusion rate of 82% (95% CI, 73–90) and pooled procedure-related complication rate of 18% (95% CI, 14–22). Despite differences in inclusion criteria, the aneurysmal occlusion rates in these studies are in line with our findings. Moreover, we compared the occlusion rates between anterior and posterior circulation aneurysms and found them to be comparable without a significant difference.

To the best of our knowledge, no previous large observational study or meta-analysis has conducted such a comparison. Our finding supports extending the indications for FDs to include posterior circulation aneurysms and approving them for on-label use in this role. We found that neither patient nor aneurysmal characteristics were significant predictors of aneurysmal occlusion rate. Although a previous meta-analysis by Liang et al.³⁵ found that patient age is associated with complete obliteration rates for posterior circulation aneurysms treated with FDs and increasing age predicted incomplete obliteration, we found that older patients had lower complete occlusion rates; however, our finding was not statistically significant.

On the other hand, our regression analysis showed that age and female gender were significant predictors of morbidity. A study by Brinkinji et al. noted that increasing age is associated with higher neurological morbidity and mortality after FD embolization of intracranial aneurysms.³⁶ However, the overall complication rate of FDs in such patients was acceptably low. Accordingly, old age alone should not be considered an exclusion criterion for treating posterior circulation aneurysms with FDs.

Regarding the association between the type of FD and complete occlusion rate, we found better results in studies that used only the PED (83.2% complete occlusion rate) compared to the Surpass stent (68.4% complete occlusion rate) and other types. A previous meta-analysis by Arrese et al.³⁷ investigated the efficacy of FDs for intracranial aneurysms and reported that studies that used only the PED had significantly higher occlusion rates than those that used the Silk stent (88% vs. 68%), in agreement with our finding.

Earlier studies have reported high complication rates related to use of FDs in the posterior circulation,^38–41 which might be attributed to the presence of numerous perforating arteries that supply the brainstem.¹³ For basilar artery aneurysms, use of endovascular devices including FDs has been associated with low efficacy and high morbidity and mortality.¹⁷,⁴² In our meta-analysis, 8.3% of the patients overall developed post-procedural ischemia, but after leave-one-out sensitivity analysis, the overall effect estimate decreased to 6.5%. The higher thromboembolic rate reported in the other studies may be related to the placement of multiple overlapping FDs, which adds more impedance to perforator artery filling due to increased surface coverage area.²⁹

Mortality and morbidity are important factors in evaluating treatment outcomes. We found low rates of both (approximately 6%), which are considerably lower than rates reported in previous studies and possibly attributed to our exclusion of small case series. Such studies reporting any adverse events may be published more frequently, thus maximizing the overall complication rate.

Our meta-analysis has several strengths. It is the largest meta-analysis to investigate FDs in posterior circulation aneurysms to date, as we included 14 large studies (comprised of at least 20 posterior circulation aneurysms) with 659 patients and 676 aneurysms. We also performed subgroup and regression analyses and found that age and gender were significant predictors of neurological morbidity. Moreover, this is also the first meta-analysis to compare FD efficacy in anterior and posterior circulation aneurysms.

The major limitation of our meta-analysis was the heterogeneity observed among the included studies, even though we included large studies. In addition, the lack of any specific outcomes associated with patient and aneurysmal characteristics precluded evaluating predictors in the statistical analysis. Finally, most of the included studies were retrospective in nature.

Flow diverters are an effective treatment option for posterior circulation aneurysms. High complete occlusion rates comparable to those in the anterior circulation can be achieved. However, procedure-related complications are not negligible. Older age and female gender are associated with higher morbidity.