Background Posterior communicating artery (PcoA) compromise may serve as adjunctive treatment in patients with hypoplastic variants of PcoA who undergo coil embolization of PcoA aneurysms. However, procedural safety and the propensity for later recanalization are still unclear.
Objective To evaluate clinical and radiologic outcomes of coil embolization in this setting, focusing on compromise of PcoA.
Methods As a retrospective review, we examined 250 patients harboring 291 aneurysms of hypoplastic PcoAs, all consecutively treated by coil embolization between January 2004 and June 2016. PcoA compromise was undertaken in conjunction with 81 of the treated aneurysms (27.8%; incomplete 53; complete 28). Medical records and radiologic data were assessed during extended monitoring.
Results During the mean follow-up of 33.9±24.6 months (median 36 months), a total of 107 (36.8%) coiled aneurysms showed recanalization (minor 50; major 57). Recanalization rates were as follows: PcoA preservation 40.5% (85/210); incomplete PcoA occlusion 34.0% (18/53); complete PcoA occlusion 14.3% (4/28). Aneurysms >7 mm (HR 3.40; P<0.01), retreatment for recanalization (HR 3.23; P<0.01), and compromise of PcoA (P<0.01) emerged from multivariate analysis as significant risk factors for recanalization. Compared with PcoA preservation, complete PcoA compromise conferred more favorable outcomes (HR 0.160), whereas incomplete compromise of PcoA fell short of statistical significance. Thromboembolic infarction related to PcoA compromise did not occur in any patient.
Conclusion PcoA compromise in conjunction with coil embolization of PcoA aneurysms appears safe in hypoplastic variants of PcoA, helping to prevent recanalization if complete occlusion is achieved.
- coil embolization
- posterior communicating artery
Statistics from Altmetric.com
The posterior communicating artery (PcoA) provides a critical vascular link between the anterior and posterior circle of Willis, supplying also the anterior thalamus via the lamoperforating arteries. However, it is one of the most common sites of intracranial aneurysms and poses the greatest risk of rupture.1 Owing to the anatomic anomalies encountered in parent arteries and the array of configurations shown by these aneurysms, such lesions may be among the easiest or the most difficult to treat by any means, whether surgical or endovascular.2
In some wide-neck PcoA aneurysms, preserving PcoA flow during treatment may be particularly problematic. The PcoA is frequently incorporated at its origin, and affected branches are often acutely angled at saccular necks.3 Endo et al were the first to address PcoA compromise in PcoA aneurysms, finding that 50% of patients thus treated suffered infarction after coil embolization. They attributed the above outcomes to compromise in instances of PcoA dominance.4 Cho et al have nevertheless argued that compromise of hypoplastic PcoA variants, in conjunction with coil embolization of PcoA aneurysms, appears to be safe with no major consequences.3 It thus seems that PcoA compromise may serve an adjunctive role in this therapeutic context, especially in complex lesions that widely incorporate the PcoA orifice, where PcoA originates from the aneurysm dome, or in repeatedly recurrent aneurysms. It is also feasible that PcoA compromise may promote stability, helping to prevent subsequent recanalization by reducing hemodynamic stress. To our knowledge, this premise has yet to be investigated.
In this study we evaluated the clinical and radiologic outcomes of coil embolization in patients with PcoA aneurysms and PcoA hypoplasia, focusing on prevention of recanalization after therapeutic PcoA compromise.
Study population and data collection
Between January 2004 and June 2016, coil embolization was undertaken at our institute in 3453 patients with 4119 aneurysms, including 494 (12.0%) arising at PcoA orifices. Patients presenting with so-called ‘true’ (non-junctional) PcoA aneurysms were excluded as study participants, as were those with dissecting, fusiform, blood blister-like or false aneurysms. PcoA compromise was selectively undertaken in patients with oddly configured aneurysms involving hypoplastic PcoA variants (PcoA/P1 ≤1). A total of 138 aneurysms marked by PcoA dominance (PcoA/P1 >1) and 32 aneurysms of perforator vessels (devoid of P1 connections) were thereby excluded, leaving 324 aneurysms eligible for study.
The primary endpoint of this study was recanalization in coiled aneurysms during follow-up monitoring, thus eliminating another 33 aneurysms lacking or deficient (<6 months) in follow-up data. Ultimately, 250 patients harboring 291 PcoA aneurysms and displaying PcoA hypoplasia were accepted for this investigation. Therapeutic alternatives were formulated by multidisciplinary deliberation of both neurosurgical and neurointerventional teams. Informed consent was obtained after discussion with each patient and family. This study protocol adhered to principles set forth by the Declaration of Helsinki and was approved by our Institutional Review Board.
The following patient parameters were considered: age; gender; hypertension (HTN), diabetes mellitus (DM) or hyperlipidemia (HL) as comorbidities; smoking history; clinical presentation (subarachnoid hemorrhage (SAH) vs unruptured intracranial aneurysm (UIA)); and retreatment for recanalization. Angiographic features were also noted, including size of aneurysm, depth-to-neck ratio (D/N ratio), occlusion status at baseline, stent usage, PcoA angulation, PcoA/P1 ratio, and degree of PcoA compromise during treatment. PcoA angulation at its origin from the distal internal carotid artery (ICA) was estimated for assignment of subsets (>90o vs ≤90o), and PcoA/P1 ratios were grouped as ≤0.5 or >0.5 (all PcoA/P1≤1). In terms of PcoA status, all lesions were stratified as preservation versus compromise (complete or incomplete).3
Cerebral angiography and rotational angiography with three dimensional (3D) image reconstruction were generally performed using Integris V, Allura Clarity (Philips Medical Systems, Amsterdam, Netherlands) or Innova IGS 630 (GE Healthcare, Chicago, IL, USA) systems to depict both configurations and arterial architectures of aneurysms under study. The broadest dimensions of aneurysms and neck diameters were recorded in 3D angiographic images. D/N ratios were measured on working projections of digital subtraction angiography.
Endovascular procedure and angiographic follow-up
Nearly all procedures were performed under general anesthesia. Single-agent antiplatelet medication (clopidogrel, 300 mg loading dose) was given the day before the procedures to each patient undergoing UIA treatment, again administering clopidogrel (75 mg) on the morning the procedures were done. Based on clopidogrel resistance testing (VerifyNow P2Y12; Accriva Diagnostics, San Diego, CA, USA), poor responders (P2Y12 reaction units >285) also received aspirin.5 If stent deployment was likely, dual antiplatelet agents (aspirin and clopidogrel, loading doses of 300 mg each) were administered, adding cilostazol (200 mg) in poor clopidogrel responders. As of November 2014, low-dose prasugrel was substituted as premedication. Prasugrel eventually replaced clopidogrel altogether, beginning in June 2016.6
A bolus of unfractionated heparin (3000 IU) was delivered intravenously after femoral artery sheath placement, followed thereafter by hourly bolus dosing (1000 IU) and testing of activated clotting times. In patients with SAH, pretreatment antiplatelet premedication was withheld in advance of endovascular treatment, administering systemic heparin once ruptured aneurysms were adequately protected. Following embolization procedures, use of antiplatelet agents was limited and maintained only as needed (ie, for atherosclerotic steno-occlusive disease, coil protrusion or thromboembolic complications). Patients with stents deployed were advised to continue dual antiplatelet therapy for at least 3 months, followed by single-agent maintenance for at least 1 year.
Based on arterial diameter ratios (PcoA/P1), the following PcoA variants were stipulated: fetal (no P1); dominant (>1.0); hypoplastic (≤1.0); and perforator (thalamoperforator origin; no connection to P1).3 Only patients with hypoplastic variants of PcoA qualified for this study. PcoA compromise was designated as complete or incomplete, according to integrity of flow on post-embolization carotid and vertebral angiographic studies (figure 1 and figure 2). Any persistence of PcoA flow discovered by postprocedural ICA angiography, despite measures being taken to selectively intercede, signified incomplete PcoA compromise. Initial angiographic assessments of occlusion after coil embolization were interpreted by two experienced neurointerventionists and categorized according to the Raymond classification as complete occlusion, residual neck and residual sac.7 Follow-up diagnostics were scheduled arbitrarily and were independently reviewed by two experienced neurointerventionists (YDC, HSK) blinded to pertinent clinical and radiologic details. All follow-up tests, including magnetic resonance angiography (MRA) (maximum intensity projection (MIP) and source images) and conventional angiographic documentation were scrutinized. In the event of disagreement, a third interventional neuroradiologist (MHH) helped to reach a consensus. Degree of recanalization was also assessed at 6, 12, 24 and 36 months after endovascular treatment using time-of-flight MRA with 3D reconstruction and source images. Conventional angiography was recommended if MRA was unfeasible or if recanalization was suspected by MRA and a decision for further treatment was needed. Using the 3-point Raymond scale, recanalization on follow-up MRA was categorized as complete occlusion, minor recanalization or major recanalization. In instances of major recanalization, repeat embolization was advised.
Continuous data were expressed as mean±SD. χ2 or Fisher’s exact test and Student’s t-test were used to analyze categorical and continuous variables, respectively. A binary logistic regression model was used in univariate analysis to evaluate risk factors of recanalization during follow-up monitoring. Multivariate analysis was also performed, invoking variables with P values <0.10 in univariate analysis. All results of binary logistic regression were expressed as odd ratios (ORs), each with 95% confidence interval (95% CI) and P value. Two-tailed P values <0.05 were considered significant. All computations relied on standard software (SPSS v22; IBM, Armonk, NY, USA).
Baseline characteristics of patients and aneurysms
A total of 250 patients with 291 aneurysms involving hypoplastic PcoAs were selected for study, including 48 males with 50 aneurysms (17.2%) and 202 females with 241 aneurysms (82.8%). Mean age of the subjects was 58.3±10.1 years (median 59 years; range 14–85 years); mean maximum diameter of the aneurysms was 5.5±2.4 mm (median 5.0 mm; range 1.8–15.3 mm); and mean neck diameter was 3.7±1.5 mm (median 3.3 mm, range 1.6–10.0 mm). Angles of PcoA origin ranged from 7.8° to 180° (median 41.0°; mean 53.0±32.7°), and mean PcoA/P1 ratio was 0.60±0.20 (median 0.60). Sixty-eight aneurysms were accompanied by SAH at patient presentation; in 42 aneurysms, recanalization occurred. Baseline characteristics of patients with PcoA aneurysms are detailed in table 1.
During coil embolization, a balloon protection technique was applied to 29 aneurysms and stent placement was undertaken in 63 aneurysms (46 Enterprise (Codman Neuro/Integra Lifesciences, Plainsboro, NJ, USA); nine LVIS (MicroVention/Terumo Group, Aliso Viejo, CA, USA); eight Neuroform (Stryker Corp, Kalamazoo, MI, USA)). Stents were primarily deployed from the distal ICA to the ICA bifurcation (n=58). All others were seated from the distal ICA to the PcoA (n=2); within PcoA only, aligning the proximal stent with the saccular neck (n=2); or from the PcoA to the ICA bifurcation, retrograde (n=1). In 81 aneurysms, attempted PcoA compromise succeeded completely (n=28) or in part (n=53). Stent deployment (from the distal ICA to the ICA bifurcation) took place in 34 of the 81 compromised aneurysms, using a balloon in nine instances.
In the immediate aftermath of coil embolization, 99 aneurysms (34.0%) showed complete occlusion, whereas 118 (40.6%) retained residual necks, and 74 (25.4%) displayed residual sacs. Procedural complications occurred in 28 patients (thromboembolism 22; procedural leakage 6). Six months later, 26 of these patients had recovered well. Only two patients with severe SAH at presentation did poorly. Procedural complications developed in 10 (thromboembolism 7; hemorrhage 3; incomplete 9/53; complete 1/28) of 81 patients subjected to PcoA compromise, all having recovered well at 6 months. Cerebral ischemia (ie, thalamic or posterior cerebral artery (PCA) infarction) due to PcoA compromise was not encountered in any patient.
The mean follow-up period was 33.9±24.6 months (median 36 months; range 6–140 months). Within this time frame, inter-observer agreement in assessing occlusion was excellent (k=0.815; range 0.784–0.846). Recanalization occurred in 107 (36.8%) coiled aneurysms (minor recanalization 50; major recanalization 57) over a mean period of 15.2±17.2 months (median 6 months; range 6–96 months). The other 184 (63.2%) showed complete occlusion during follow-up monitoring, which did not differ significantly in duration (recanalization 35.0±25.4 months; complete occlusion 33.3±24.1 months; P=0.58). The overall annual recanalization rate of coiled aneurysm was 13.0% (107/822.4) per aneurysm-year.
Recanalization rates by degree of PcoA compromise were as follows: PcoA preservation 40.5% (85/210); incomplete occlusion 34.0% (18/53); complete occlusion 14.3% (4/28). Major recanalization was recorded in 21.9% (46/210) and 20.8% (11/53) of patients with PcoA preservation and incomplete occlusion, respectively. There were no instances of major recanalization in patients with complete PcoA occlusion (0/28) (table 2).
Analysis of risk factors for recanalization involved the following variables: gender, age (>65 years), HTN, DM, HL, smoking history, clinical presentation (SAH vs UIA), PcoA/P1 ratio (>0.5), PcoA angulation (>90°), maximum aneurysm size (>7 mm), neck size (>4 mm), D/N ratio (>1.5), retreatment for recanalization, stent usage, postprocedural occlusion status, and degree of PcoA flow compromise. In univariate analysis, SAH presentation, aneurysm size (>7 mm), neck size (>4 mm), and PcoA flow compromise showed significant associations with recanalization. Aneurysm size >7 mm (OR 3.40; P<0.01), retreatment (OR 3.23; P<0.01), and compromise of PcoA (P<0.01) emerged from multivariate analysis as significant factors in recanalization. Compared with PcoA preservation, outcomes of complete PcoA compromise proved more favorable (OR 0.160). Incomplete compromise of PcoA fell short of statistical significance (table 3).
Kaplan-Meier estimates of cumulative survival without recanalization are shown in figure 3. The overall survival rate of patients without recanalization at the 36 month follow-up point was 63.3%, but 36 month estimates differed by aneurysm size (≤7 mm, 70.7%; >7 mm, 27.0%), retreatment (initial treatment 65.3%; retreatment 51.2%), and PcoA compromise (preservation 59.8%; incomplete compromise 69.1%; complete compromise 84.8%). In generalized Wilcoxon analysis, cumulative survival rates without recanalization reached significance for aneurysms >7 mm (P<0.01), retreated aneurysms (P<0.01) and treated lesions involving complete PcoA compromise (P=0.033).
The International Study of Unruptured Intracranial Aneurysms (ISUIA) has established 5 year cumulative rupture rates in patients without histories of SAH.8 In terms of posterior circulation (including PcoA), there is a 2.5% risk of rupture for aneurysms <7 mm. However, that risk increases in accord with overall size (7–12 mm, 14.5%; 13–24 mm, 18.4%; ≥25 mm, 50%). Furthermore, the likelihood (HR) of rupture in PcoA aneurysms is significantly greater than that observed in middle cerebral artery (MCA) aneurysms (HR 1.90; P=0.02); and annual risks of rupture in PcoA aneurysms appear to gradually increase as lesion size increases (3–4 mm, HR 0.41; 5–6 mm, HR 1.00; 7–9 mm, HR 3.19; 10–24 mm, HR 6.12; ≥25 mm, HR 126.97).9 The need to treat aneurysms is thus a function of size, becoming more urgent in larger lesions.
Recanalization after endovascular coil embolization likewise is more apt to occur in larger aneurysms.10 11 In a study by Shimizu et al,12 aneurysm size >7 mm proved to be the sole independent risk factor for recanalization (P=0.005), whereas small-sized aneurysms (≤7 mm) were more likely to remain completely occluded long-term.13 Moreover, the PcoA location was shown to have a significant relation with retreatment after coil embolization in the International Subarachnoid Aneurysm Trial.14 Jeon et al further reported that PcoA aneurysms display the highest rate (60%) of progression from minor to major recanalization at 6 months after coil embolization, exceeding comparable rates for aneurysms of ICA (31.3%), anterior cerebral artery (47.6%), MCA (22.2%) and posterior circulation (11.1%).15 In the current study of aneurysms involving hypoplastic PcoA variants, aneurysms of larger size (>7 mm, OR 3.40, 95% CI 1.61 to 7.21; P<0.01) similarly conferred significant risk of recanalization after coil embolization.
Recanalization rates after repeat embolization of recanalized aneurysms have also been documented in previous reports. Cho et al cite a 44.2% rate of recanalization (minor 9.9%; major 34.3%) after repeat embolization,16 and the findings of Henkes et al indicate that recanalization rates increase after repeat embolization,17 having performed second-round treatments in 12.3% (350/2759 aneurysms) and third attempts in 26.9% (94/350 aneurysms). Not surprisingly, recanalization rates determined herein were significantly higher in aneurysms requiring repeated treatment for recanalization, compared with initial treatment successes (54.8% vs 33.7%; P<0.01); in our analysis, retreatment for recanalization was identified as a significant risk factor of recanalization (OR 3.23, 95% CI 1.51 to 6.92; P<0.01).
Angulation of vascular side branches and geometric properties of curvatures implicit in this setting may influence wall shear stress and wall pressure gradients as well.18 Although the long-term stability of small-sized coiled PcoA aneurysms is already known,19 the impact of PcoA compromise remains unclear in terms of altering flow patterns and hemodynamic parameters of aneurysms that influence growth and rupture.20 21 We suspected that PcoA compromise might reduce saccular flow and thereby protect against recanalization, so PcoA compromise was attempted in 81 of 291 aneurysms involving hypoplastic PcoA variants. As shown by post-procedural angiography, PcoA flow from the ICA completely ceased in 28 patients, leading to retrograde filling of the PcoA from the PCA (complete compromise), while persisting in 53 others (incomplete compromise), despite packing the PcoA orifice with coils. Ultimately, recanalization rates in instances of PcoA preservation (40.5%) and incomplete occlusion (34.0%) were similar, whereas the recanalization rate after complete PcoA occlusion (14.3%) proved to be half that recorded for PcoA preservation; none of the patients with complete PcoA compromise developed major recanalization. In multivariate analysis, complete compromise (vs preservation) of PcoA was associated with significantly less risk of recanalization after coil embolization (HR 0.16, 95% CI 0.05 to 0.55; P<0.01). Incomplete PcoA occlusion lacked statistical significance.
Clinical outcomes of PcoA compromise have also been chronicled in earlier publications. Endo et al used coil embolization and PcoA occlusion to treat 14 patients with ruptured PcoA aneurysms,4 confirming retrograde filling of PcoA through P1 by the Allcock test before each procedure. Nevertheless, half of their patients developed postprocedural infarcts in territories of tuberothalamic arteries. Ipsilateral P1 segments were not visualized by preoperative vertebral angiography in any of the patients with subsequent infarcts, and in patients who developed infarcts (vs those who did not), PcoA diameter and PcoA/P1 ratio means were significantly greater.4
Two mechanisms may thus be responsible for ischemic events after PcoA compromise. One entails immediate hypoperfusion, whereas the other is thromboembolic in nature.3 Regarding the patients above, hypoperfusion may be implicated in thalamic infarction if occluded PcoAs were dominant, leaving insufficient retrograde PcoA flow from diminutive P1 vessels. Cho et al achieved differing results from compromise of 46 hypoplastic PcoAs (PcoA/P1 ≤1) in 44 patients during coil embolization.3 In instances of complete occlusion (23 patients with 23 aneurysms), no compromise-related infarction ensued. However, two patients experienced delayed cerebral ischemia (including one transient ischemic attack) following incomplete occlusion (21 patients with 23 aneurysms). Although purely speculative at this juncture, incompletely disrupted PcoA flow may serve as a thromboembolic source, precipitating delayed infarction.3
Consequently, our goal in every attempt at PcoA compromise was complete occlusion, thus preventing any related thromboembolism. This proved technically prohibitive in some patients, due to PcoA flow demands. Still, none of our patients suffered cerebral infarction as a result, dispelling such theoretical concerns. As mentioned earlier, intentional compromise of hypoplastic PcoAs is selectively performed at our center, especially in challenging cases with few options and in repeatedly recurring aneurysms. Hence, PcoA compromise during coil embolization of PcoA aneurysms may be a safe and effective therapeutic adjunct in the context of PcoA hypoplasia. In terms of reducing hemodynamic stress and recanalization risk, complete rather than partial compromise seems more beneficial.
Use of a flow diverter is perhaps another option for treating aneurysms that develop in hypoplastic PcoA variants. According to related reports,22 23 impeded or interrupted PcoA flow was evident in 30–50% of lesions after diverter placement or in later angiographic monitoring, with no apparent clinical sequelae. In addition, most of the aneurysms showed complete or near-complete occlusion on follow-up angiography.
Our study has certain limitations, including its retrospective design and confinement to a single center. In addition, PcoA compromise was performed as part of a tailored treatment strategy rather than mandated by specific indications. The merit of hypoplastic PcoA compromise in preventing recanalization must be corroborated through further prospective studies.
Our efforts have shown that compromise of hypoplastic PcoAs may yield favorable clinical outcomes, helping to prevent recanalization of PcoA aneurysms if complete occlusion is achieved. Aneurysm size (>7 mm) and retreatment for recanalization were also identified as significant contributors to recanalization of PcoA aneurysms in instances of PcoA hypoplasia.
Contributors HHC conceived and conducted the review of this series, analysed the data, drafted and revised the manuscript, and approved the final version. JL, H-SK, MHH, W-SC, JEK, SHL, EKY, and DHY assisted in conducting the review of the series, performed the operations, revised the manuscript, and approved the final draft. YDC conceived and conducted the project, performed the operations, analysed the data, revised the manuscript, and approved the final draft.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent Not required.
Ethics approval The Seoul National University Hospital, Seoul National University.
Provenance and peer review Not commissioned; externally peer reviewed.
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.