Treatment of giant intracranial aneurysms using the Pipeline flow-diverting stent: Long-term results from the International Retrospective Study of the Pipeline Embolization Device (IntrePED) study

Abstract

Background

Traditional endovascular treatments of giant intracranial aneurysms are associated with high rates of complications and retreatment. Our objective was to examine the safety and long-term efficacy of the Pipeline Embolization Device for treatment of these aneurysms.

Methods

This retrospective study using the IntrePED database included all patients with giant intracranial aneurysms treated with the Pipeline device between July 2008 and February 2013. Efficacy outcomes were stratified by using the Raymond-Roy Occlusion Classification. Predefined safety outcomes included spontaneous rupture of the target aneurysm; ipsilateral intracranial hemorrhage; ischemic stroke; parent artery stenosis; and sustained cranial neuropathy.

Results

Sixty-six embolizations were performed to treat 63 giant intracranial aneurysms (including 2 ruptured): 49 (77.8%) in the anterior and 14 (22.2%) in the posterior circulation. The median follow-up was 22.4 (0.1–60.5) months. Class I angiographic occlusion was achieved in 72.0% (36/50). The neurological morbidity/mortality rate was 23.8% (15/63), with higher rates in the posterior circulation than in the anterior circulation (22.4% vs. 28.6%). Among seven deaths, five had neurological causes. The procedure-related neurological morbidity and mortality rates were 22.7% (15/66) and 7.6% (5/66), respectively. The spontaneous rupture rate was 4.5% (3/66). Two spontaneous ruptures (1 death), 4/4 postprocedural intracranial hemorrhages, and 6/9 ischemic events occurred within 30 days. In-stent stenosis and new-onset cranial neuropathy were not observed during the angiographic follow-up period.

Conclusions

Although procedure-related neurological morbidity/mortality rates were not insignificant, this study confirms the feasibility and long-term efficacy of the Pipeline Embolization Device to treat giant intracranial aneurysms.

Abbreviations

GIA = giant intracranial aneurysm; PED = Pipeline Embolization Device;

Introduction

Giant intracranial aneurysms (GIAs), with a diameter of ≥25 mm, are rare and occur at sites of high-velocity flow and flow redirection.^1–3 Although most unruptured intracranial aneurysms are small and carry a lower annual rupture risk, the natural history of GIAs shows a higher risk of rupture.^4–6 GIAs carry greater morbidity because of compression of neural structures and thromboembolic events. In patients with untreated symptomatic GIAs, mortality rates exceed 60% within two years of diagnosis, and 80% of patients are incapacitated or die within 5 years.^7–9

Both traditional open microsurgery and endovascular intervention carry a higher risk and complication profile and rates of retreatment, including considerable perioperative morbidity/mortality,^{5, 8, 10–18} in GIAs compared with treatment of smaller aneurysms because of recanalization and coil compaction.^18–21 Morbidity/mortality rates up to 35% have been reported in modern open surgical series, with aneurysm rupture, older age, mass effect, and posterior circulation location predicting poor surgical outcomes.^22–26

Endoluminal flow diverter therapy offers high rates of complete aneurysm occlusion even in cases of large and giant aneurysms.^27–29 The Pipeline Embolization Device (PED, Covidien, Irvine, CA) was the first flow-diverting stent to be approved by the U.S. Food and Drug Administration (FDA), with approval for the treatment of large and giant wide-necked aneurysms of the internal carotid artery (ICA) in 2011. The IntrePED (International Retrospective Study of Pipeline Embolization Device) registry was established to capture consecutive patients treated at high-volume centers using common data elements. In this study, we examined the safety and long-term effectiveness of the PED for treatment of GIAs and report the long-term results recorded in the IntrePED registry. We hypothesized that flow diversion using the PED for the treatment of GIAs is safe and effective, with a complication profile on par with PED treatment of small intracranial aneurysms and more favorable than traditional open surgical treatment of GIAs.

Methods

Study criteria

For this investigation, the IntrePED registry was queried to identify patients treated for GIAs with the PED between July 2008 and February 2013 at 17 experienced centers in 6 countries (after the date of regulatory approval in that region or country). The IntrePED retrospective study had no defined antiplatelet protocol; thus, each center followed its own treatment algorithm. Patients were excluded if they had not undergone “meaningful” follow-up, defined as lack imaging and clinical evaluation to inform the presence of treatment failure or the ability to assess neurologic morbidity. Angiographic outcomes and major complications from this cohort have been published previously ³⁰ ; the current investigation extends the previous work as a comprehensive analysis of clinical, procedural, and radiographic information including long-term follow-up, all of which have not been previously published. Local ethics committees at each site approved the study.

Data collection

Patient demographic information and aneurysm characteristics, procedural details, and follow-up information were collected. Angiographic and clinical follow-up were defined as the time from the first treatment session to the last angiogram available and the last post-treatment clinical neurologic examination, respectively. Although all centers used a common study protocol, procedural details and periprocedural patient management varied across centers. The primary outcomes for efficacy of PED treatment were stratified according to the Raymond-Roy Occlusion Classification; successful outcome was defined by complete obliteration (Class I) reported by the individual operators.

Predefined safety outcomes included: (1) spontaneous rupture of the target aneurysm; (2) ipsilateral intracranial hemorrhage; (3) ischemic stroke; (4) parent artery stenosis; and (5) sustained cranial neuropathy. Procedure-related morbidity was defined as any of the predefined safety outcomes with development of a new neurologic deficit after a treatment session, with major adverse events being death or any clinical deficit persisting for 7 days and minor adverse events defined as clinical deficits that resolved within 7 days of onset with no clinical sequelae.

Statistical analysis

Data from all centers were collected and managed by the first author. Descriptive statistics were calculated. Frequencies were calculated for categorical and ordinal variables and means (±standard deviations) or median for continuous variables.

Results

Patient population and aneurysm characteristics

A total of 63 patients underwent 66 endovascular embolization procedures using the PED for GIA treatment (Table 1). The previous publication of the IntrePED study ³⁰ reported 66 patients, but after detailed review, we have corrected this information. The patient population consisted of 43 (68.3%) female and 20 male patients, with a median age of 61 years. Twenty-six patients (41.3%) presented with a baseline cranial nerve deficit; two (3.2%) presented with ruptured aneurysm. Most patients were able to function independently (pretreatment mRS 0–2); 8/45 (17.8%) were neurologically intact at baseline with no symptoms from the GIA.

Table 1.

Demographic and aneurysm characteristics for 63 patients with GIAs.

Variable	Value a
Female	43 (68.3%)
Median age (years)	61 (range 15–86 years)
Baseline cranial nerve deficit	26 (41.3%)
Ruptured on presentation	2 (3.2%)
Presentation mRS score 0-2	56/58 (98.2%) b
Mean aneurysm size (mm)	Dome 29.0 ± 5.5 Neck 11.5 ± 8.1
Saccular aneurysms	44 (69.8%)
Anterior circulation aneurysms	49 (77.8%) Mean size 29.2 ± 5.6 mm
Posterior circulation aneurysms	14 (22.2%) Mean size 28.4 ± 6.2 mm
Previously treated?	4 anterior circulation (2 stent-assisted coiling, 1 primary coiling, 1 Silk flow diverter) 0 posterior circulation

^a Value reported as no. of patients (%) unless otherwise indicated.

^b Six patients with missing data.

mRS, modified Rankin Scale.

Overall, the mean aneurysm size was 29.0 ± 5.5 mm [range 25–55 mm], with a neck width of 11.5 ± 8.1 mm (Table 2). Saccular aneurysms comprised 69.8% (44/63) of the GIAs, with the remainder being fusiform morphology. Forty-nine aneurysms (73.0%), with an average size of 29.2 ± 5.6 mm, were located in the anterior circulation, and 14 (22.2%), with an average size of 28.4 ± 6.2 mm, were located in the posterior circulation. Four (6.3%) GIAs (all involving the ICA) had been previously treated, 2 with stent-assisted coiling, 1 with primary coiling, and 1 with a Silk flow diverter.

Table 2.

Procedure characteristics for the 66 embolization procedures.

Variable	Value a
Median procedural time in min (mean, range)	114 (130.4, 54–330)
Mean number of PEDs used (range)	2.4 (1–11)
Multiple PEDs used	37 (56.1%)
PED alone for treatment	54 (81.8%)
Coils adjunctively used	12 (18.2%)
Saccular aneurysms	44 (66.7%)
Ruptured on presentation	2 (3.0%)

^a Value reported as no. of patients (%) unless otherwise indicated.

PED, Pipeline Embolization Device.

Procedure characteristics

A total of 66 embolization procedures were performed to treat 63 GIAs. Three patients with GIAs of the ICA with persistent filling Raymond-Roy Class 3 underwent retreatment an average of 8.8 months after initial treatment. The median procedural time for aneurysm embolization was 114 min (mean, 130.4 min; range, 54–330 min) (Table 2). The mean number of PEDs used for target aneurysm embolization was 2.4 (range, 1–11). Multiple PEDs were used in 37 cases (56.1%). PEDs alone were used for treatment of 54 aneurysms (81.8%), and coils were used adjunctively in 12 embolizations (18.2%).

Angiographic outcomes

The final available angiogram for patients in our series was obtained a mean of 9.1 ± 6.5 months after the initial treatment (Figure 1A); 13 (20.6%) patients were lost to long-term follow-up. Among surviving patients with follow-up imaging, Class I angiographic occlusion was obtained in 72.0% (36/50) of the treated GIAs (Figure 1B), including 70.7% (29/41) of anterior circulation GIAs and 77.8% (7/9) of posterior circulation GIAs. Saccular GIAs demonstrated a 76.5% (26/34) rate of Class I occlusion. All of the patients who required subsequent treatment with PED to address persistent aneurysm filling had angiographic occlusion of the target lesion.

Figure 1.

Flow diagrams demonstrate angiographic outcomes for patients in this series. (A) Retreatments; (B) Raymond-Roy Class I outcomes of patients with imaging available.

Clinical outcomes, neurological morbidity/mortality

Seventeen patients (27.0%) experienced morbidity and/or mortality (Table 3). The overall neurological morbidity/mortality rate was 23.8% (15/63). Among the subset of patients with anterior circulation GIAs, the neurological morbidity and mortality rate was 22.4% (11/49), compared with 28.6% (4/14) for those with posterior circulation GIAs. The median clinical follow-up time for the cohort was 22.4 months (mean 22.4 months; range 0.1–60.5 months). Among patients with follow-up mRS scores, 39/48 (81.3%) were functioning independently (including 33/38 patients with anterior circulation GIA and 6/10 patients with posterior circulation GIA). Seven patients died (4/49 patients with anterior circulation aneurysms and 3/14 patients with posterior circulation aneurysms); however, two of the deaths had non-neurologic causes: a gastrointestinal bleed and pneumonia. Including the 7 individuals who died, 10 patients had worsened neurological status at last clinical follow-up; 17 individuals demonstrated clinical improvement based on mRS scores.

Table 3.

Clinical outcomes, morbidity, mortality in patients with GIA treated with PED.

Variable	Value a
Mean follow-up time in months (range)	22.4 (0.1–60.5)
Mean radiographic follow-up time in months (median, range)	9.1 ± 6.5 months (6.3, 0.2–27.1 months)
Overall outcomes	n = 63
mRS score 0-2 Anterior circulation Posterior circulation	39/48 (81.3%) 33/38 (86.8%) 6/10 (60.0%)
Morbidity/mortality	17 (27.0%)
Neurological morbidity/mortality Anterior circulation Posterior circulation	15 (23.8%) 11/49 (22.4%) 4/14 (28.6%)
Death a Anterior circulation Posterior circulation	7 (11.1%) 4/49 (8.2%) b 3/14 (21.4%) c
Procedure-related neurological outcomes	n = 66
Morbidity	15 (22.7%)
Mortality	5 (7.6%)
Anterior circulation morbidity/mortality	11/52 (21.1%) 2 aneurysm rupture after treatment 3 ipsilateral intracranial hemorrhage 6 ischemic stroke
Posterior circulation morbidity/mortality	4/14 (28.6%) 1 aneurysm rupture after treatment 1 ipsilateral intracranial hemorrhage 2 ischemic stroke

Data reported as no. of patients (%) unless otherwise indicated.

^a Two had non-neurological causes.

^b 54-year-old man with gastrointestinal bleed two months after treatment of unruptured ICA GIA.

^c 86-year-old woman developed pneumonia six months after treatment of ruptured basilar GIA.

mRS, modified Rankin Scale.

The procedure-related neurological morbidity and mortality rates were 22.7% (15/66) and 7.6% (5/66), respectively (Table 3). All of the events qualified as major adverse events, with most resulting in permanent deficit. Among embolization procedures for anterior circulation GIAs, the procedure-related neurological morbidity/mortality rate was 21.1% (11/52): 2 spontaneous ruptures of the aneurysm after treatment, 3 ipsilateral intracranial hemorrhages, and 6 ischemic strokes after treatment. Embolization procedures for posterior circulation GIAs carried a morbidity/mortality rate of 28.6% (4/14): 1 spontaneous rupture of the treated GIA, 1 ipsilateral intracranial hemorrhage, and 2 ischemic strokes. Morbidity was most likely to occur within 30 days of the index procedure (11/15). Baseline and treatment characteristics of the 15 major complications can be found in Table 4. When embolizations for treatment of ruptured, dissecting, or fusiform aneurysms were excluded, the procedure-related morbidity/mortality rates reduced to 15.2% and 5.1% in the anterior and posterior circulation, respectively.

Table 4.

Demographic and treatment characteristics for patients who experienced a major complication.

Variable	Value (n = 15)
Aneurysm dome size (mean, mm)	31
Aneurysm neck size (mean, mm)	15
Anterior circulation	11 (73)
Multiple PEDs used	9 (60)
PED with or without adjuncts
PED alone for treatment	10 (66)
PED with coils	3 (20)
PED alone – prior treatment with stent/coils with recurrence	1 (6)
PED alone – prior treatment with coils with recurrence	1 (6)
Age (mean, years)	57
Female	8 (53)
Saccular aneurysms	9 (66)
Unruptured	15 (100)

Values reported as n (%) unless otherwise indicated.

The spontaneous rupture rate was 4.5% (3/66), with two of the three cases occurring within 30 days, resulting in one patient death. Two of the cases involved saccular GIAs of the ICA in which the target aneurysms were treated with a single PED without adjunctive coiling. The remaining case involved a 30.5-mm fusiform aneurysm of the vertebral artery that was treated with placement of 5 PEDs with no coiling. In two of the cases, subarachnoid hemorrhage (SAH) resulted; in the other, a cavernous-carotid fistula developed. The ipsilateral intracranial hemorrhage rate was 6.1% (4/66), and ischemic complications occurred after 13.6% of treatment sessions (9/66). All of the postprocedural intracranial hemorrhages and most (6/8) ischemic events occurred within 30 days of the index procedure. There were no cases in which in-stent stenosis occurred during the angiographic follow-up period, nor were there any instances of new-onset cranial neuropathy.

Illustrative case

A 68-year-old woman presented with visual acuity deterioration over several weeks, which led to the discovery of a GIA of the ophthalmic segment of the left ICA (Figure 2A–C). The patient underwent endovascular embolization of the aneurysm with one PED used without adjunctive coiling (Figure 2D). Stasis of contrast was noted within the aneurysm upon completion of the procedure (Figure 2E). Five days post-procedurally, the patient was found unresponsive at home. Her admission Glasgow Coma Scale score was 4T; a head CT scan showed diffuse, Fisher Grade 3 subarachnoid blood (Figure 2F–G). Platelets were administered, and an external ventricular drain was placed. The patient’s neurologic examination results improved gradually, and the drain was weaned. Retreatment involved placement of 2 additional PEDs. Two months after her bleeding event, she had recovered significantly, with a resolving mild right-sided hemiparesis. Follow-up angiography 10 months later demonstrated complete aneurysm occlusion (Figure 2H).

Figure 2.

Case illustration demonstrates a 68-year-old woman with a GIA of the left ICA on MRI (A). (B, C) Angiography demonstrated a GIA of the ophthalmic segment of the left ICA. (D) The patient underwent endovascular embolization of the aneurysm with one PED used without adjunctive coiling. (E) Stasis of contrast was noted within the aneurysm upon completion of the procedure. Five days post procedurally, the patient was found unresponsive at home. Her admission Glasgow Coma Scale score was 4T. (F, G) A CT scan of the head showed diffuse, Fisher Grade 3 subarachnoid blood. (H) Follow-up angiography 10 months later, 8 months after recovery demonstrated complete aneurysm occlusion.

Discussion

In this large, multicenter study of flow-diversion therapy for the treatment of GIAs, the use of PED demonstrated an efficacy (Raymond-Roy Class I) of 72.0%. The associated procedural morbidity and mortality rates were 22.7% and 7.6%, respectively. The spontaneous rupture rate was 4.5%; ipsilateral intracranial hemorrhage occurred after 6.1% of cases and ischemic complications were seen after 13.6% of treatment sessions. Taken in context with modern surgical series and large case series involving endovascular treatment of GIAs, our data suggest that use of the PED for treatment of GIAs is efficacious, with acceptable rates of morbidity/mortality.

The use of open surgery is widely reported,^{22–24, 26} but with the continued innovation and improvement in endovascular techniques, the treatment paradigm has shifted to new approaches using endovascular treatment and flow diversion. In 2009, Fiorella and colleagues ³¹ described using telescoping PEDs for successful endoluminal reconstruction of a giant basilar trunk aneurysm in a pediatric patient. Lylyk et al. ³² demonstrated aneurysm occlusion in 12/14 large and giant aneurysms treated with the PED at 6 months. McAuliffe and colleagues ³³ and Saatci et al. ³⁴ demonstrated 100% and 90% occlusion rates based on 6-month follow-up arteriography in their individual series of 7 and 27 giant aneurysms, respectively. The largest experience with the use of the PED for treatment of large and giant aneurysms was reported in the initial IntrePED study. ³⁰ The Pipeline for Uncoilable or Failed Aneurysms (PUFS) study, published in 2013, which was a multicenter, prospective, single-arm clinical trial involving aneurysms within the anterior circulation, ³⁵ demonstrated an occlusion rate of 73.6% at 6 months and 86.8% at 1 year, with a 5.6% rate of major stroke or neurologic death.

A meta-analysis of outcomes among 1451 patients with 1654 intracranial aneurysms of various sizes treated with flow diverters demonstrated a complete occlusion rate for GIAs of 76% at 6 months ²⁸ ; however, patients with GIAs were more likely to experience SAH from delayed aneurysm rupture and ischemic strokes after treatment than patients who underwent embolization of smaller aneurysms. The current study represents a large multicenter effort to further define the risk profile of treating GIAs with PED; previous literature on outcomes after flow-diverter treatment of GIAs is limited to single-center case series, leading to selection and publication bias, making generalization of data difficult.

The complication rate experienced in our cohort was higher than that of patients in the PUFS trial. The estimated procedure-related rate of experiencing a serious adverse event was 19.6% in the PUFS cohort. In addition, procedure-related morbidity/mortality is higher with PED treatment of GIAs when compared with smaller aneurysms, ³⁰ which is especially noteworthy considering GIAs represented only 20.4% of aneurysms in the PUFS trial. ³⁵ However, we included 14 posterior circulation GIAs in our cohort. The worse outcomes observed in the current series with the treatment of posterior circulation GIAs are consistent with the literature reported to date. Ertl et al. ³⁶ reported on a series of patients with giant vertebrobasilar aneurysms who underwent treatment with flow diverters. Although there were no primary periprocedural complications, 4 of 6 patients died during follow-up, and the remainder had a worsened outcomes.

Fujii et al. ³⁷ reported on 84 Japanese patients with 90 giant ICA aneurysms treated with PED. After exclusions, 60 of 77 aneurysms (77.9%) demonstrated complete occlusion after treatment. On multivariate analysis, the authors identified age >70 years, wide aneurysm neck, and non-use of coiling as adjunct risk factors for incomplete occlusion. The occlusion rates of Fujii et al. ³⁷ and the current study are lower than the 81.9–93.3% occlusion rates reported in historical large-scale reports.^{35, 38, 39} Bae et al. ⁴⁰ recently identified excessive use of balloon angioplasty and incorporated vessel as independent risk factors for a suboptimal outcome after flow diversion in very large or giant aneurysms.

Our results are on par with others who have used various flow-diverting technologies for the treatment of large and giant aneurysms. Peschillo et al. ⁴¹ reported a one-year occlusion rate of 72.7% (61.5% in combination with coil embolization vs. 88.9% with flow diversion only; not solely PED). Interestingly, despite a small number of posterior circulation aneurysms treated, we achieved a nearly 80% complete occlusion rate, a finding that requires further, large-scale study. Leonardi et al. ⁴² demonstrated aneurysm thrombosis at 1 year in 14/18 patients with large and giant aneurysms treated with the Silk device. Zhou and colleagues ⁴³ reported their experience using the Tubridge flow diverter. Compared with our study, the authors used overlapping flow diverters much less frequently (17.9%) and performed adjunctive coil embolization more often (64.2%). There were no procedure-related morbidity or mortality events among those with angiographic follow-up, and 72.0% of aneurysms showed complete occlusion. ⁴⁴ In 2019, the Surpass Intracranial Aneurysm Embolization System Pivotal Trial to Treat Large or Giant Wide Neck Aneurysms also showed promising results with the Surpass flow diverter, with an occlusion rate of 75.5% in 12 months (Raymond-Roy Classes 1 and 2) and a 12-month major stroke or death rate of 8.3%. ⁴⁵

Although the rate of complete occlusion is lower in this study at a mean follow-up of 2 years, we believe this is likely representative of “real-world” results—and, importantly, that long-term follow-up of these complex lesions is warranted.

The role of newer technologies, including newer generations of flow diverters and intrasaccular devices, remains to be seen depending on how they compare with flow diverters and classical endovascular and open vascular treatment modalities in tackling this challenging pathology. Since the completion of this study, the strategy employed for flow diversion has evolved, including more frequent use of coiling as an adjunct and limited use of multiple flow-diverting devices. The high complication rate in the current series is reflective of the early experience with PED and the associated learning curve.

Limitations

Several limitations must be considered when interpreting the results of this study. This is a retrospective study and thus subject to inherent limitations. Although we have described a large series of a rare pathology, a detailed statistical analysis determining specific risk factors for complications and incomplete occlusion was not performed because the low number of patients means the study is underpowered to detect meaningful differences. We acknowledge there are factors that may contribute to these outcomes, such as adjunctive coiling, single vs. multiple PED use, and aneurysm size, but the current report does not have the sample size to correctly and responsibly analyze and report this data. Most significantly, there was a variety of management with respect to antiplatelet medication regimen length and procedural characteristics among institutions. Importantly, all study adverse events collected were prespecified in the study protocol and were adjudicated by the Adverse Events Review Committee to maintain consistency, but no imaging protocol was required and the investigators followed their own institutions’ standard procedures. There was no specification regarding the minimum duration of follow-up, and follow-up timing was per standard of care for the treating physician and institution. Finally, the number of posterior circulation aneurysms included in this registry was small. Despite these limitations, this large series of GIAs treated with flow diversion demonstrates the viability of this option for treatment of these complicated lesions.

Conclusion

This study represents a large analysis of the use of flow-diversion therapy for treatment of GIAs, with a focus on long-term efficacy and complications incurred with use of the PED. Although the rates of procedure-related neurological morbidity/mortality are not insignificant, our study confirms that the PED represents a feasible means of obliterating these difficult-to-treat aneurysms with a reasonable morbidity/mortality profile. Our findings should be considered when deciding the best therapeutic option for patients harboring such intracranial pathology. Future large-scale, multicenter prospective series with standardization of treatment regimens and follow-up to further refine management are merited.