Article Text

Download PDFPDF

Original research
Comparative analysis of coil embolization in posterior and anterior communicating artery aneurysms
  1. Hyun Ho Choi1,
  2. Young Dae Cho2,
  3. Dong Hyun Yoo2,
  4. Su Hwan Lee2,
  5. Eung Koo Yeon2,
  6. Hyun-Seung Kang3,
  7. Won-Sang Cho3,
  8. Jeong Eun Kim3,
  9. Moon Hee Han4
  1. 1 Department of Neurosurgery, Chung-Ang University Hospital, Chung-Ang University College of Medicine, Seoul, Korea, Seoul, The Republic of Korea
  2. 2 Department of Radiology, Seoul National University Hospital, Seoul National University College of Medicine, Seoul, The Republic of Korea
  3. 3 Department of Neurosurgery, Seoul National University Hospital, Seoul National University College of Medicine, Seoul, The Republic of Korea
  4. 4 Department of Radiology, Korea Veterans Hospital Medical Center, Seoul, The Republic of Korea
  1. Correspondence to Dr Young Dae Cho, Department of Radiology, Seoul National University Hospital, Seoul National University College of Medicine, Seoul 110-744, The Republic of Korea; aronnn{at}


Background Recanalization rates after coil embolization are known to be higher in cerebral aneurysms of the posterior (vs anterior) circulation. Although often grouped with anterior lesions, aneurysms of the posterior communicating artery (PcoA) may nevertheless behave differently.

Objective We performed a comparative analysis to explore differences in recanalization rates of PcoA and anterior communicating artery (AcoA) aneurysms, both integral to the circle of Willis.

Methods Between October 2012 and July 2017, 699 AcoA (n=427) and PcoA (n=272) aneurysms were treated by endovascular coil embolization, monitoring 667 (95.4%) via radiologic imaging for ≥ 6 months. Cumulative recordings of medical and imaging data were retrospectively reviewed, conducting propensity score matching and binary logistic regression analysis.

Results In the 667 aneurysms followed longer term, recanalization occurred in 111 (16.6%; minor 72; major 39) and was significantly more frequent in PcoA (25.5%) than in AcoA (11.0%; P<0.01) aneurysms during similar follow-up periods. After 1:1 propensity score matching, an even greater proclivity for recanalization was evident at PcoA sites (PcoA 23.0%; AcoA 12.2%; P<0.01). Although A1 segment dominance was linked to recanalization in AcoA aneurysms (18.2% vs 7.6%; P=0.01), the PcoA counterpart had no bearing on recanalization (27.7% vs 24.1%; P=0.51).

Conclusions Despite a clear preponderance of AcoA aneurysms, recanalization of PcoA aneurysms proved significantly greater, attesting to posterior circulation behavior.

  • aneurysm
  • coil
  • hemorrhage
View Full Text

Statistics from


The posterior communicating artery (PcoA) provides a critical link between the internal carotid artery (ICA) and the posterior cerebral artery (PCA) in the circle of Willis. Although generally deemed to be part of the anterior circulation (AC), the behavior of PcoA aneurysms after coil embolization seems to differ in terms of subsequent recanalization.1 2 Various risk factors thereof cited in previous studies include large sized or wide necked aneurysms, rupture, and stent usage.1 3–5 However, the impact of location on recanalization of such aneurysms has yet to be properly addressed, given the many variables (ie, sites, classifications) that confound accurate comparisons.1 3 At present, rates of recanalization in aneurysms of the posterior circulation (PC) are known to exceed those of AC aneurysms, without a clear basis for this imbalance.6–8

Another critical link in the circle of Willis is the anterior communicating artery (AcoA), which bridges the internal carotid arteries. The AcoA is among the most common sites for development of AC aneurysms and accounts for >50% of all ruptures.9–11 Aneurysms of AcoA and PcoA are also subject to similar branching (A1, A2, ICA, and PCA) and perforator (medial lenticulostriatal arteries, including recurrent artery of Huebner, and anterior thalamoperforating arteries) anomalies, posing technical challenges to endovascular coiling.12 13 The hemodynamics of these neighboring vessels may well have bearing on effective packing and recanalization of coiled aneurysms.14 Thus direct comparison of PcoA and AcoA aneurysms in this setting may offer a reasonable means of profiling PcoA aneurysms after coil embolization.

Considering the respective clinical ramifications and unique characteristics of these lesions, we performed a comparative analysis of PcoA and AcoA aneurysms, focusing on procedural outcomes and recanalization rates during follow-up surveillance.

Patients and methods

Study population and data collection

Between October 2012 and July 2017, 1793 patients harboring 2268 aneurysms were treated by coil embolization at our institution. Aneurysms arising from the AcoA (427/2268; 18.8%) or PcoA (272/2268; 12.0%) contributed 699 lesions to a consecutively populated and prospectively collected database. Corresponding medical records were reviewed in retrospect, assessing the following patient related factors: gender, age, hypertension, diabetes mellitus, hyperlipidemia, smoking history, clinical presentation (subarachnoid hemorrhage (SAH) or unruptured intracranial aneurysm), and retreatment. Angiographic parameters, including initial occlusive states, overall and neck sizes of aneurysms, depth to neck (D/N) ratios, and modes of endovascular treatment were also recorded. Configurations and arterial architectures of all aneurysms were depicted via cerebral angiography and rotational angiography with three-dimensional image reconstruction, using Innova IGS 630 (GE Healthcare, Wauwatosa, Wisconsin, USA), Integris V, or Allura Clarity (both Philips Medical Systems, Amsterdam, The Netherlands) systems. Maximal overall and neck dimensions of aneurysms were determined in three-dimensional angiographic images, assessing D/N ratios on working projections of DSA. In referencing the hemodynamic environment, we classified each PcoA as dominant or hypoplastic (PcoA/P1 ≤1) and assessed concurrent contralateral Al segment hypoplasia (ie, dominant A1 diameter/contralateral A1 diameter >2 or absence of contralateral A1).12 13 Therapeutic strategies for endovascular treatment were planned accordingly, based on consensus of the cerebrovascular team in multidisciplinary deliberation. All patients provided informed consent for treatment. This study adhered to principles stipulated in the Declaration of Helsinki and received the approval of our institutional review board. Required patient consent for this retrospective investigation was waived.

Endovascular procedures

Most coiling procedures were performed under general anesthesia. Patients with unruptured aneurysms received antiplatelet premedication, checking for clopidogrel resistance (VerifyNow P2Y12 assay; Accriva Diagnostics (Werfen), San Diego, California, USA). Dual antiplatelet agents (aspirin and clopidogrel) were given when anticipating stent deployment. In poor responders to clopidogrel (P2Y12 reaction units >285), we added cilostazol. Starting from November 2014, we switched to low dose prasugrel as premedication, using this exclusively from June 2016 to replace clopidogrel.15 Intravenous bolus injections of unfractionated heparin (3000 IU) were given after placement of femoral arterial sheaths, delivering additional increments (1000 IU) at hourly intervals and monitoring activated clotting times. Patients with acute rupture were exempt from antiplatelet premedication, and we deferred systemic heparinization until adequate protection was achieved. Dual antiplatelet therapy continued for at least 3 months post-embolization in patients with stent protected aneurysms, using single agent maintenance for at least 1 year thereafter. Antiplatelet maintenance was similarly advised in patients experiencing procedural thromboembolism, coil loop protrusion, and steno-occlusive disease.

Angiographic results and clinical outcomes

Angiographic evidence of occlusion was initially assessed by two experienced neurointerventionists (with 10 years' and 15 years' experience), classifying outcomes according to the Raymond scale as complete occlusion, residual neck, or residual sac.3 Complete occlusion and residual neck equally represented therapeutic success. Reviewer differences were resolved through consensus input by a third neurointerventionist (>25 years' experience). All neurointerventionists were also involved in the initial aneurysm treatment.

Time of flight MR angiography was recommended at 6, 18, and 36 months after coil embolization. In instances of suspected recanalization detected by MR angiography, DSA was advised to determine the need for further coil embolization. Follow-up angiographic outcomes were derived from last performed DSA or MR angiography, qualifying as either complete occlusion or recanalization (minor vs major) designations by the Raymond scale. Repeat coil embolization was recommended for major recanalization.

Extravasation of contrast or device visibility beyond saccular contours on fluoroscopic angiography was indicative of procedural rupture. Stagnation of contrast within vessels, luminal filling defects, or non-visualization of distal arteries, as well as clinically apparent ischemic symptoms with infarction in a treatment related vascular territory, were deemed thromboembolic complications. Any neurologic deterioration occurring after procedures warranted CT or MR scans to detect hemorrhagic and ischemic lesions. Poor functional outcomes were defined as 6 month postprocedural scores >2 by the modified Rankin Scale, excluding any created by underlying disease or other disease treatments.

Statistical analysis

The χ2and Fisher’s exact tests or unpaired t test were used to assess categorical and continuous variables, respectively. Categorical data were conveyed as frequencies and group percentages, expressing continuous variables as mean±SD. Propensity score matching was undertaken to compensate for group wise imbalances in baseline characteristics with a potential to skew post-embolization recanalization risk. Score matching of each patient, reflecting the probability of aneurysm location, was conducted within the R environment (R Foundation for Statistical Computing, Vienna, Austria) using the nearest neighbor package for optimal 1:1 matching. Univariate analysis served to examine factors impacting recanalization during follow-up monitoring. Multivariate analysis incorporated variables with P values <0.05 in univariate analysis. Results of a binary logistic regression model were expressed as odds ratios (ORs) with 95% confidence intervals (CIs) and P values. Statistical significance was set at a two tailed P value <0.05. All computations relied on SPSS V.22 (IBM, Armonk, New York, USA) or the R freeware platform.


Baseline characteristics of patients and aneurysms

Ultimately, 699 aneurysms were selected for study. Mean age of the study population was 59.9±10.0 years (range 14–85 years), and most patients (459/699, 65.7%) were women. The mean maximum diameter of the aneurysms was 5.4±2.5 mm (median 4.8 mm; range 1.6–16.8 mm), and mean neck size was 3.4±1.5 mm (median 3.0 mm; range 1.2–12.0 mm). Separate subsets of ruptured (n=92) and recurrent (n=88) aneurysms were identified, as were medical subsets of hypertension (n=379, 54.2%), diabetes mellitus (n=89, 12.7%), and hyperlipidemia (n=300, 42.9%). Baseline characteristics of the patients and treated aneurysms are summarized in table 1.

Table 1

Characteristics and treatments of anterior and posterior communicating artery aneurysms

Comparison of procedural outcomes in PcoA and AcoA aneurysms

Aneurysm subsets and treatments undertaken are compared in table 1. Patients with PcoA (vs AcoA) aneurysms were significantly older (P=0.048), and women predominated (P<0.01). The lesions were also larger by comparison (maximal size: PcoA 5.8±2.9 mm; AcoA 5.1±2.2 mm; P<0.01), with broader necks (PcoA 3.7±1.7 mm; AcoA 3.3±1.4 mm; P<0.01) and higher D/N ratios (PcoA 1.4±0.6; AcoA 1.3±0.5; P<0.01). In the PcoA subset, 45 aneurysms (16.5%) recanalized (recurred) after the initial treatment, significantly exceeding the recanalization rate of AcoA aneurysms (10.1%; P=0.012).

Embolization techniques also differed significantly by location (P<0.01). In particular, a single microcatheter technique sufficed more often (40.5%) in AcoA aneurysms than in PcoA aneurysms (25.0%; P<0.01), requiring dual microcatheters with greater frequency in PcoA aneurysms (31.6%) than in AcoA aneurysms (25.1%; P=0.06). However, the two groups were similar in terms of stent protection; procedure related complications, namely procedural rupture (P=1.00) and thromboembolism (P=0.23), were also comparable. Initial assessments of occlusion showed significant differences (P<0.01), especially immediate post-embolization persistence of residual sacs (PcoA 26.5%; AcoA 32.6%; P=0.09).

Comparison of follow-up results in PcoA and AcoA aneurysms

Overall, 667 (95.4%) of 699 aneurysms were monitored via angiography for >6 months (PcoA 95.2%; AcoA 95.6%; P=0.84). The mean follow-up period was 20.4±12.7 months (median 18 months; range 6–60 months), and respective follow-up durations did not differ significantly (PcoA 19.8±12.2 months; AcoA 20.8±13.0 months; P=0.34). Among the 667 aneurysms treated, 556 (83.4%) showed complete occlusion, with recanalization in 111 (16.6%; minor 72; major 39). The recanalization rate was significantly higher in PcoA than in AcoA aneurysms (PcoA 25.5%; AcoA 11.0%; P<0.01) (table 2).

Table 2

Recanalization rates after coil embolization shown by location (anterior vs posterior communicating artery)

In 1:1 propensity score matched subjects, each followed for >6 months by angiographic imaging, group characteristics were comparable. The recanalization rate of PcoA aneurysms remained significantly higher than that of AcoA aneurysms (PcoA 23.0%; AcoA 12.2%; P<0.01) (table 3).

Table 3

Characteristics before and after propensity score matching of aneurysms followed via imaging (≥6 months)

Variables potentially implicated in recanalization were first subjected to univariate analysis, determining that female gender, large aneurysms (>7 mm), wide necked lesions (>4 mm), SAH at presentation, high D/N ratio (>1.5), retreatment for recanalization, and PcoA location significantly impacted recanalization. In multivariate analysis, large aneurysms (>7 mm; OR=2.530, 95% CI 1.501 to 4.264; P<0.01), wide neck (>4 mm; OR=1.742, 95% CI 1.034 to 2.935; P=0.04), SAH at presentation (OR=1.806, 95% CI 1.102 to 2.961; P=0.02), retreatment for recanalization (OR=1.876, 95% CI 1.192 to 2.953; P<0.01), and PcoA location (OR=1.984, 95% CI 1.318 to 2.986; P<0.01) emerged as independent risk factors for recanalization (table 4).

Table 4

Analysis of risk factors for recanalization during follow-up surveillance of all study participants

Kaplan-Meier estimates of cumulative survival without recanalization are shown for both groups in figure 1. Overall patient survival at 24 months without recanalization was 80.7% (PcoA, 74.4%; AcoA, 88.1%). Cumulative survival without recanalization reached significance for PcoA location (P<0.01) via generalized Wilcoxon analysis.

Figure 1

Kaplan–Meier estimates of stable occlusion at 24 months post-embolization in posterior communicating artery (PcoA) and anterior communicating artery (AcoA) aneurysms.

Risk factors for recanalization in PcoA and AcoA aneurysms are addressed in the supplementary table 1. Aside from the above mentioned risk factors, A1 dominance (vs symmetry) was linked to recanalization in AcoA aneurysms (18.2% vs 7.6%; P=0.01), although PcoA variants (dominance vs hypoplasia) had no bearing on recanalization (27.7% vs 24.1%; P=0.51). Groupwise Kaplan–Meier estimates of cumulative survival without recanalization are shown according to flow dominance in figure 2. Patient survival at 24 months without recanalization differed significantly in AcoA aneurysms, showing a significant reduction in instances of A1 dominance (76.8%), compared with symmetry of A1 (91.5%; P<0.01). However, survival in PcoA aneurysms was unrelated to flow (dominant PcoA 71.5%; hypoplastic PcoA 75.3%; P=0.69).

Figure 2

Kaplan–Meier estimates of stable occlusion at 24 months post-embolization in (A) anterior communicating artery (AcoA) aneurysms (dominant vs symmetric A1 segment) and (B) posterior communicating artery (PcoA) aneurysms (dominant vs hypoplastic PcoA).

Ultimately, 20 patients (10 in each group) showed poor functional outcomes 6 months after coil embolization. All patients had ruptured aneurysms, and their unfavorable outcomes reflected poor clinical conditions after initial bleeding. No aneurysms ruptured during post-embolization follow-up surveillance.


In this series of treated PcoA and AcoA aneurysms, endovascular coil embolization yielded low complication rates and excellent clinical outcomes. Nevertheless, PcoA aneurysms tended to be larger than AcoA counterparts, displaying broader necks and a higher rate of recanalization. Even after propensity score matching to balance differing aspects of each subset, the recanalization rate of PcoA aneurysms still proved comparatively higher; and PcoA location emerged as an independent risk factor for recanalization in study participants overall. Multivariate analysis further implicated the following factors in recanalization: large aneurysm size (>7 mm maximally, P<0.01), broad neck (>4 mm, P=0.04), SAH at presentation (P=0.02), and retreatment (P<0.01). Although A1 segment dominance was linked with recanalization of AcoA aneurysms, there were no similar implications for PcoA.

The significantly disparate recanalization rates we documented for aneurysms of PcoA (25.5%) and AcoA (11.0%; P<0.01) serve to corroborate similar outcomes in past studies. Raymond et al observed a higher rate of recanalization in PcoA aneurysms (37.2%) compared with AcoA aneurysms (25.0%), although statistical significance was not reached.3 In the ISAT trial, recoiling procedures were significantly more frequent in PcoA aneurysms than in aneurysms at other locations (P=0.03), which was not the case for anterior cerebral artery aneurysms (including AcoA lesions).1 Because recanalization rates may be skewed by baseline characteristics, we used propensity score matching to confer balance, thus imparting a degree of homogeneity. Even so, recanalization proved distinctly more frequent in PcoA than in AcoA aneurysms (23.0% vs 12.2%; P<0.01). This finding suggests that hemodynamic conditions likely differ at these locations and stand to impact recanalization. We also used multivariate analysis to assess independent predictors of recanalization in study participants overall. PcoA location was subsequently confirmed as an independent risk factor, along with other previously established indices, such as large aneurysm size (maximum dimension >7 mm), a broad neck (>4 mm), SAH at presentation, and retreatment for recanalization.

There is already evidence that anatomic variants of communicating arteries may increase the risk of rupture in developing aneurysms. Lazzaro et al have shown that circle of Willis anomalies, including AcoA with A1 hypoplasia (diameter ≤50% of dominant A1 or absence of one of both A1 segments) and PcoA with fetal origins from PCA (PcoA diameter ≥P1 segment of ipsilateral PCA), are more often found in instances of rupture.16 Furthermore, a spectrum of hemodynamic stresses (eg, high wall shear stress, flow impingement, and blood flow velocity) are similarly implicated in recanalization of coiled AcoA or PcoA aneurysms.17 Findings of various studies indicate that hemodynamic stress is affected by local geometrics (inclusive of communicating vessels), which are critical in recanalization of coiled cerebral aneurysms. Although the relation between formation or rupture of aneurysms and dominance of communicating channels has been amply studied, there are limited data on dominance and recanalization. According to Choi et al, recanalization of AcoA aneurysms after stent assisted coil embolization is associated with contralateral A1 segment hypoplasia,12 and AcoA aneurysms with A1 dominance are more inclined to recanalize, compared with aneurysms marked by A1 symmetry. However, this is not the case for PcoA. In the present series, we found no difference between dominant and hypoplastic PcoA variants in terms of recanalization (27.7% vs 24.1%; P=0.51). On the other hand, Songsaeng et al have noted that recanalization of coiled PcoA aneurysms is prevented by small sized PcoAs.14 This suggests that hemodynamic stress in PcoA aneurysms is determined by factors such as ICA–PcoA angle or dome direction and not PcoA size alone.

In previous studies, aneurysms of PcoA have demonstrated unique characteristics, often warranting classification as AC aneurysms, while sometimes deviating from archetypal behavior. The International Study of Unruptured Intracranial Aneurysms (ISUIA) assigns PcoA aneurysms to the PC domain, based on clinical profiles more aligned with PC behavior18 whereas another source remains steadfast on the AC characteristics of PcoA aneurysms.19 In one particular study, PC is cited as a risk factor for recanalization after coil embolization.7 20–23 Sandra et al have also conducted a systematic review of recanalization and retreatment rates after coil embolization, documenting a higher recanalization rate in 862 PC aneurysms than in 1901 largely (>85%) AC aneurysms (PC 22.5%; AC 15.5%).6 The latter compare favorably with our data, attesting to the PC behavior of PcoA aneurysms.

Recently, a new technical option to treat AcoA and PcoA aneurysms has been introduced.24–27 Choi et al reported that 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.24 Moreover, Carvalho et al have used flow diversion for 18 PcoA aneurysms. They showed a low ischemic complication rate and 90% of PcoA flow change in aneurysms with a PcoA/P1 <1.00.25 Therefore, we could apply these techniques to prevent recanalization in PcoA aneurysm with a size ratio of PcoA/P1 <1.0. An intra-saccular flow disruptive device (ex, WEB) may also be an alternative.27

Regardless of observed differences in recanalization rates, endovascular treatment is a safe and effective means of treating aneurysms at either location. On detecting recanalization, we weigh the need for recoiling in our multidisciplinary decision making environment, discussing options at length with patients and families. Repeat embolization is encouraged in instances of major recanalization, because retreatment in this setting is equally safe and effective.28 Overall, 18 of our patients with recanalized aneurysms (AcoA 8; PcoA 10) elected repeat coil embolization, and none experienced hemorrhage after coil embolization procedures or during appropriate follow-up imaging after retreatment.

This study has the inherent bias of a retrospective design and a single center effort. Moreover, our focus was limited, comparing only PcoA and AcoA aneurysms. Although sufficient in scale to expose statistical differences in these groups, the results must be interpreted with caution. The observed disparity in recanalization may not directly apply to aneurysms at other sites. Another limitation is the lack of hemodynamic analysis, pursuing causes for these differing recanalization rates. A large, prospective, and multicenter trial, encompassing aneurysms at other locations and suitable hemodynamic parameters, is needed for validation.


The present findings indicate a greater tendency for recanalization after endovascular coil embolization in PcoA (vs AcoA) aneurysms, attesting their PC behavior. In monitoring patients after endovascular treatment, such rate differences must be considered.


View Abstract


  • Contributors HHC conceived and conducted the review of this series, analyzed the data, drafted and revised the manuscript, and approved the final version. DHY, SHL, EKY, H-SK, W-SC, JEK, and MHH performed the operations, assisted in conducting the review of the series, revised the manuscript, and approved the final draft. YDC conceived and conducted the project, performed the operations, analyzed 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 study was approved by the local institutional review board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

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.