Background and purpose The Pipeline Embolization Device (PED) has been shown to effectively treat complex internal carotid artery aneurysms while maintaining patency of covered side branches. The purpose of this retrospective matched cohort study is to evaluate the effect of flow diversion on the patency of the ophthalmic artery when treating ophthalmic artery aneurysms.
Methods A retrospective review of our prospectively collected institutional database identified 19 ophthalmic artery aneurysms treated with a PED. These were matched according to aneurysm diameter in a 1:2 fashion to ophthalmic artery aneurysms treated via coil embolization, although it is important to note that there was a statistically significance difference in the neck diameter between the two groups (p=0.045). Clinical and angiographic outcomes were recorded and analyzed.
Results On follow-up angiography, decreased flow through the ophthalmic artery was observed in 26% of the PED cohort and 0% of the coil embolization cohort (p=0.003). No ophthalmologic complications were noted in either cohort. Complete occlusion at 12 months was more common following PED treatment than coil embolization (74% vs 47%; p=0.089), although lower than reported in previous trials. This may be due to inflow into the ophthalmic artery keeping the aneurysm patent. Retreatments were more common following coil embolization than PED (24% vs 11%), but this was not significant (p=0.304). Permanent morbidity rates were not significantly different between the PED (11%) and coil embolization (3%) cohorts (p=0.255).
Conclusions Our results suggest that ophthalmic artery aneurysms may be adequately and safely treated with either the PED or coil embolization. However, treatment with the PED carries a higher risk of impeding flow to the ophthalmic artery, although this did not result in clinical sequelae in the current study.
- Blood Flow
- Flow Diverter
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.
In 2011, the Pipeline Embolization Device (PED; eV3, Irvine, California, USA) received FDA approval for the treatment of large or giant wide-necked intracranial aneurysms arising from the internal carotid artery between the petrous and ophthalmic segments. Studies of the device have shown it to be very effective at occluding complex aneurysms, previously untreatable with coil embolization alone.1–3 Complete occlusion rates for ophthalmic or ophthalmic segment aneurysms following PED deployment have been reported to be between 73% and 80%.4–7 This compares to an occlusion rate of 63–86% following coil embolization.8–10 However, it is important to note that these occlusion rates are not derived from direct comparisons of PED and coil embolization, as the literature typically describes the use of PED in aneurysms that are not amenable to treatment with coil embolization alone.
A major concern with the treatment of ophthalmic aneurysms is the patency of the ophthalmic artery and associated ophthalmologic complications. The judicious use and careful placement of coils can lead to adequate treatment of the aneurysm without impinging on the ophthalmic artery. However, the only way to treat an ophthalmic aneurysm with a PED is to cover the origin of the ophthalmic artery. Prior studies have found that the ophthalmic artery can become sluggish or occluded in 4–32% of cases following placement of a PED across the origin of the otherwise normal artery.11–15 However, it remains unclear what happens to the ophthalmic artery when a PED is used to treat an aneurysm arising at the base of the ophthalmic artery.
The aims of this retrospective matched cohort study are to (1) compare the angiographic and clinical outcomes of patients with unruptured ophthalmic artery aneurysms who were treated with PED versus coil embolization, and (2) determine the rate and clinical significance of ophthalmic artery patency after treatment with PED versus coil embolization.
This HIPAA compliant study was performed under the auspices of our Institutional Review Board.
Retrospective evaluation of a prospectively collected patient log identified 19 consecutive patients with an unruptured ophthalmic artery aneurysm who were treated with a PED and 85 consecutive patients with an unruptured ophthalmic artery aneurysm who were treated by coil embolization. Only ophthalmic aneurysms arising from the dorsal internal carotid artery at or immediately adjacent to the origin of the ophthalmic artery were included in accordance with Day's definition of an ophthalmic artery aneurysm.16 In an effort to limit selection bias, aneurysms treated by coil embolization following the introduction of the PED at our institution were excluded. Medical records were reviewed by a single member of the research team. Radiographic imaging was reviewed by three neurointerventionalists and consensus was reached regarding treatment outcomes. Patient demographics, treatment details, complications, and follow-up details were recorded.
Immediate treatment outcomes were documented in all cases using digital subtraction angiography. Patients were scheduled for 6 and 12 month follow-up with catheter angiography, and yearly follow-up thereafter via CT angiography, MR angiography, or catheter angiography. Long-term follow-up imaging analysis was limited to those patients with at least 12 months of radiographic follow-up. All follow-up imaging was evaluated for evidence of recurrence, stability, or progressive occlusion of the aneurysm. In cases of recurrent or residual aneurysm, the remnant was stratified according to the Raymond classification.17 Patients treated with a PED who demonstrated residual filling at 6 months follow-up were initially asked to stop taking clopidogrel. If the aneurysm still did not thrombose, it was retreated with an additional PED at or after 12 months. Patients who were initially treated by coil embolization were considered for retreatment with additional coils if there was coil compaction or regrowth of the aneurysm resulting in exposure of a wall or dome of the aneurysm.
In all cases, the transit time through the ophthalmic artery was evaluated before treatment and on follow-up imaging.
The patients’ presentation and treatment outcomes were recorded. Major complications were considered to be any clinically significant infarction, hemorrhage, or dissection that could conceivably be ascribed to the procedure. Clinical status was scored using the modified Rankin Scale (mRS), with a score of 0–2 considered a good outcome.18
Data are presented as mean and range for continuous variables, and as frequency for categorical variables. Blinded to outcome, PED and coil patients were matched in a 1:2 fashion based on aneurysm size. Matched analysis was carried out with matched Student t test, Wilcoxon rank sum test, and McNamara test, as appropriate. Univariate conditional matched analysis was used to test covariates predictive of the following dependent variables: any treatment complications, follow-up aneurysm occlusion (>95%), and follow-up overall outcome (mRS 0–1 vs 2–6). Interaction and confounding was assessed through stratification and relevant expansion covariates. Factors predictive in univariate analysis (p<0.20) were entered into a multivariate conditional logistic regression analysis.19 p Values of ≤0.05 were considered statistically significant. Statistical analysis was carried out with Stata V.10.0 (College Station, Texas, USA).
Patient demographics and aneurysm characteristics
After matching 1:2 according to aneurysm diameter while blinded to outcome, there were 19 patients who underwent treatment with a PED and 38 patients who underwent coil embolization of their unruptured ophthalmic artery aneurysms. Of the patients treated by coil embolization, seven required a stent and three underwent balloon remodeling. There was no statistically significant difference in patient age or gender between the two cohorts (table 1). The majority of aneurysms in the PED cohort were found incidentally, while patients in the coil embolization cohort were more likely to have presented with symptoms, although this was not significant (p=0.244). There was no significant difference in the patients’ mRS at presentation (p=0.255).
No statistical difference was identified in the diameter (p=0.709) or volume (p=0.780) of the aneurysms. However, the neck widths of the aneurysms tended to be larger in the PED cohort (p=0.045).
Sixteen of the 19 patients in the PED cohort were treated with a single PED and three were initially treated with two PEDs.
The clinical and radiographic outcomes of the PED and coil embolization cohorts are summarized in table 2.
The incidence of major complications in the PED cohort (11%) was not statistically different from that of the coil embolization cohort (3%; p=0.255). The first complication in the PED cohort resulted from acute clot formation during the procedure, which embolized to an M3 pre-Rolandic branch. The clot resolved following intra-arterial administration of abciximab. The patient initially awoke with left-sided weakness and headache, although this significantly improved and his discharge mRS was the same as his mRS at presentation. The second complication in the PED cohort occurred approximately 4 months after treatment. The patient had completed 3 months of treatment with clopidogrel but had also decreased her aspirin dosage from 325 to 81 mg against medical advice. She presented with new-onset blurry vision of the ipsilateral eye and was found to have several distal focal infarcts in the ipsilateral anterior cerebral artery and middle cerebral artery distribution, although the ophthalmic artery and choroidal blush remained normal.
The only complication in the coil embolization cohort was also a stroke in the ipsilateral anterior vascular territory. During treatment the patient had a coil loop at the neck of the aneurysm without herniation into the parent artery. The patient was discharged on 325 mg aspirin daily but returned 2 weeks later with left upper extremity paresis, which improved significantly following administration of an intravenous heparin infusion. On imaging, he was found to have a small infarction with interval occlusion of the ipsilateral internal carotid artery. It is believed that clot formed on the coil ball with subsequent parent artery occlusion.
Univariate predictors of any post-treatment complication were baseline visual deficit (p=0.037), larger aneurysm diameter (p=0.031), larger aneurysm volume (p=0.033), and ophthalmic artery patency (p=0.149). However, a multivariate analysis identified the larger aneurysm diameter as the only independent predictor of a post-treatment complication (p=0.026). Treatment with a PED or coil embolization was not a predictor of complications (p=0.571).
There was no significant difference in the mRS at discharge (p=0.255) or follow-up (p=0.594) between the PED and coil embolization cohorts. After accounting for the patients’ mRS at presentation, patients of both cohorts were unlikely to have a worse mRS at follow-up (p=0.255). Multivariate analysis identified baseline mRS as the only predictor of mRS at follow-up (p=0.001). While the use of a PED versus coils was a univariate predictor of poor outcomes (p=0.007), the multivariate analysis demonstrated that the treatment type was not predictive of long-term clinical outcome (p=0.296).
At follow-up, 14 aneurysms (74%) in the PED cohort had completely thrombosed compared with 18 aneurysms (47%) in the coil embolization cohort (p=0.089). The proportions of aneurysms that demonstrated at least 95% occlusion in the PED (79%) and coil embolization (68%) cohorts were not significantly different (p=0.537).
Aneurysms in the coil embolization cohort (24%) were more likely to recur than aneurysms in the PED cohort (0%; p=0.022). However, there was no statistical difference in the proportion of patients in the PED (11%) or coil embolization (24%) cohorts who required retreatment (p=0.304). The relationship of the ophthalmic artery to the aneurysm was not predictive of the likelihood of retreatment. In the PED cohort, one retreatment was required in an aneurysm from which the ophthalmic artery arose from the base of the aneurysm and one retreatment was required for an aneurysm from which the ophthalmic artery arose from the neck or wall of the aneurysm (p=1.000; figure 1A). In the PED cohort, six retreatments were required for aneurysms from which the ophthalmic artery arose from the base of the aneurysm and three retreatments were required for aneurysms from which the ophthalmic artery arose from the neck or wall of the aneurysm (p=0.454; figure 1B).
On follow-up angiography (figure 2), five patients (26%) in the PED cohort demonstrated sluggish flow through the ophthalmic artery compared with none of the patients (0%) in the coil embolization cohort (p=0.003). Of the five patients with sluggish flow through their ophthalmic artery, all five demonstrated normal flow immediately after PED deployment but delayed flow on or after their initial follow-up examination. Interestingly, the relationship of the ophthalmic artery to the aneurysm neck was not a significant factor in the likelihood of the ophthalmic artery demonstrating sluggish flow at follow-up. Ophthalmic arteries in which the aneurysm arose at the takeoff of the ophthalmic artery from the internal carotid artery were just as likely to appear sluggish at follow-up (25%) as those in which the ophthalmic artery arose from the aneurysm neck or wall (27%; p=1.000). Similarly, the number of PEDs used during treatment was not a significant predictor of sluggish flow. All five patients with sluggish flow were treated with a single PED. None of the patients were symptomatic. Neither the P2Y12 Reactivity Units (PRU) nor the Aspirin Reactivity Units (ARU) were predictive of sluggish flow in the ophthalmic artery after treatment. The PRU was higher in the subgroup with sluggish flow (115.2±81.3) than in the subgroup with normal flow (83.2±71.5), but this was not significant (p=0.465). Similarly, the ARU was higher in the subgroup with sluggish flow (451.6±77.4) than in the subgroup with normal flow (419.1±74.5), but this was also not significant (p=0.446).
The recent development of flow diversion devices has provided the neurointerventionalist with a supplementary tool for the treatment of intracranial aneurysms. Pipeline embolization has been shown to be a safe and efficacious option for the treatment of large wide-necked intracranial aneurysms not amenable to coil embolization.1 ,3 However, more data are needed to determine if the PED has the same risk to benefit profile for aneurysms that could otherwise be treated via conventional methods.
Comparison of initial occlusion rates is misleading. An argument can be made that coil embolization provides the patient with more immediate protection against rupture. However, while none of the aneurysms demonstrated immediate or ≥95% occlusion immediately following PED deployment, the fluid dynamics within the aneurysm were altered. Computational models suggest that flow diversion may impart an immediate reduction in wall shear stress and the wall shear stress gradient, thereby reducing the risk of further aneurysm growth and/or rupture.20–22
Complete obliteration rates at follow-up were greater following PED deployment (71%) than coil embolization (47%; p=0.089). Prior studies comparing PED versus coil embolization of intracranial aneurysms report similar findings with a significant difference in complete occlusion rates at follow-up.23 ,24 While these prior studies reported a higher obliteration rate following PED, these studies included aneurysms in multiple locations. The lower rates of complete obliteration reported here may be related to the propensity for the ophthalmic artery to arise from the base or neck of the aneurysm. So, while the flow dynamics within the aneurysm are altered, these aneurysms may be less likely to thrombose over time due to the continued flow through the ophthalmic artery.
Aneurysms required retreatment more frequently following coil embolization (24%) than PED (11%), although the difference was not significant (p=0.304). A previous study evaluating various aneurysms found that the retreatment rate following coil embolization was significantly greater than aneurysms treated by PED.23 This finding is not altogether surprising. Coil compaction or underpacking at the neck of the aneurysm may allow continued flow within the aneurysm. In contrast, a properly deployed PED will completely cover the neck so that, once an aneurysm thromboses, it should not be able to recur. The difference here, though, is that even if a coiled aneurysm has a recurrence at the neck or base of the aneurysm, the dome and walls of the aneurysm remain protected by the coil construct. However, until an aneurysm treated with a PED thromboses, the dome and walls remain underprotected and the patient must rely on the altered flow mechanics within the aneurysm sac to limit the risk of rupture.
The benefit of a higher occlusion rate with PED must be weighed against the reported higher morbidity. A meta-analysis of 7172 patients found that 11.5% had an unfavorable outcome following PED treatment while only 3.1% had an unfavorable outcome following coil embolization with modern coils.25 The problem with comparing PED and coil embolization from reports in the literature is that there is undoubtedly a selection bias as aneurysms treated with a PED tend to be those that cannot be treated easily with coil embolization alone. After matching aneurysms according to diameter, this study found that patients who were treated with a PED were almost four times more likely to have a major complication than those treated with coil embolization (11% vs 3%), although this was not statistically significant (p=0.255). Results from the LARGE prospective randomized trial comparing flow diversion to coil embolization for the treatment of large or giant aneurysms should ultimately provide greater insight into the complication rates of the two groups.26
Following thrombosis of the aneurysm, the PED endothelializes to form a permanent seal at the neck of the aneurysm.27 This has led to some concern about the patency of normal side branches after being covered by the PED. The ophthalmic artery, in particular, may be more susceptible to this phenomenon. Prior studies have reported that 4–32% of ophthalmic arteries covered by a PED became sluggish or occluded over time.11–15 In these prior studies, the ophthalmic artery was not involved with the aneurysm. As such, it was unclear how the presence of a thrombosing aneurysm at the origin of the ophthalmic artery would alter the flow dynamics through the vessel. In the current study, 26% of ophthalmic arteries were found to be sluggish after PED treatment of an ophthalmic artery aneurysm. In blood vessels with good collateral flow, the pressure drop across the PED may allow the pressure head in the collateral branches to overcome the lower pressure in the parent artery. This theory is supported by the reported higher incidence of sluggish flow in ophthalmic arteries following PED deployment compared with the anterior choroidal artery (3–7%), which does not have as robust a collateral supply.28 ,29 Similarly, another study found that all of the A1 branches in systems with a normal anterior communicating artery demonstrated sluggish flow after being covered by a PED.15
Several factors may impart some limitations in the generalizability of the findings in this study to the general population. The coil embolization cohort tended to have a smaller neck, which results from a selection bias. However, all aneurysms were matched according to aneurysm dome size. Despite using a prospective patient log, the study is still retrospective in nature and subject to all of the biases therein. The quality of the data is therefore dependent on the clinical skill and reporting practices of multiple physicians.
The relationship of the ophthalmic artery to an ophthalmic artery aneurysm and the extensive collateral network to the ophthalmic artery produces a unique hemodynamic relationship between an aneurysm and its parent artery. Our results suggest that ophthalmic artery aneurysms may be adequately and safely treated with either the PED or coil embolization. However, treatment with the PED comes with a higher risk of impeding flow to the ophthalmic artery, although this did not result in clinical sequelae in the current study.
We would like to acknowledge the role of Claire McKinley and Thomas Tandy in helping to complete this project. Both were instrumental in obtaining regulatory approval and keeping the project on time.
Contributors CRD, KCL, RWC, MEJ, AJE, and JG were involved in the design and conception of the paper. CRD performed the literature search. CRD, DC, HRH, PJS, JMG, and DD obtained the data. CRD and RMS performed the statistical analysis. All authors were involved in the data analysis. CRD compiled the primary manuscript with contributions from each of the authors. All authors critically revised the manuscript and approved it in its final form.
Competing interests JG is a consultant for Covidien, Microvention, and Stryker. MEJ receives payment for lectures from Stryker and Covidien. AJE is a consultant for Stryker and Covidien. He received a research grant from Stryker. He receives royalties from Cook and CareFusion.
Ethics approval Ethics approval was obtained from the University of Virginia Institutional Review Board.
Provenance and peer review Not commissioned; externally peer reviewed.