Article Text

Download PDFPDF

Original research
Posterior communicating and vertebral artery configuration and outcome in endovascular treatment of acute basilar artery occlusion
  1. Diogo C Haussen1,2,
  2. Sushrut S Dharmadhikari1,
  3. Brian Snelling1,
  4. Vasileios-Arsenios Lioutas2,
  5. Ajith Thomas2,
  6. Eric C Peterson1,
  7. Mohamed Samy Elhammady1,
  8. Mohammad Ali Aziz-Sultan3,
  9. Dileep R Yavagal1
  1. 1University of Miami Miller School of Medicine/Jackson Memorial Hospital, Miami, Florida, USA
  2. 2Harvard Medical School/Beth Israel Deaconess Medical Center, Miami, Florida, USA
  3. 3Harvard Medical School/Brigham and Women's Hospital, Miami, Florida, USA
  1. Correspondence to Dr D R Yavagal, University of Miami Miller School of Medicine/Jackson Memorial Hospital, 1095 NW 14th Terrace, 2nd Floor, (D4-6), Miami, FL 33136-1060, USA; dyavagal{at}


Background We aimed to evaluate if vertebrobasilar anatomic variations impact reperfusion and outcome in intra-arterial therapy (IAT) for basilar artery occlusion (BAO).

Methods Consecutive BAO patients with symptom onset <24 h treated with IAT were included. Vertebral artery (VA) V3 and posterior communicating artery (PCoA) diameters were measured (CT angiography or MR angiography). The presence of PCoA atresia, VA hypoplasia, VAs that end in the posterior inferior cerebellar artery (PICA), and extracranial VA occlusion was recorded.

Results 38 BAO patients were included. Mean age was 63±15 years; 52% were men. Baseline National Institutes of Health Stroke Scale score was 21±9, and mean/median time from symptom onset to IAT were 10/7 h. First generation thrombectomy devices were mostly used. Overall Treatment in Cerebral Ischemia 2b-3 reperfusion was 68.4%. Good outcome (modified Rankin Scale score ≤2) was observed in 17.8% and mortality in 64.3% of cases at 90 days. 55% of patients had an atretic PCoA while 47% had a hypoplastic VA. The mean sum of the bilateral PCoA and VA diameters were 2.3±1.2 and 5.2±5.2 mm, respectively. VAs that end in the PICA was noted in 23% of patients, and extracranial VA occlusion in 42%. BAO was proximal/mid/distal in 36%/29%/34%. Multivariate linear regression analysis indicated hypertensive disease (β=2.97; 95% CI 1.15 to 4.79; p<0.01) and reperfusion rate (β=−0.40; 95% CI −0.74 to −0.70; p=0.02) independently associated with outcome. Multivariate analysis for predictors of reperfusion failed to identify other associations. A trend for better reperfusion with stent retrievers was noted (β=1.82; 95% CI −0.24 to 3.88; p=0.08).

Conclusions Reperfusion emerged as a predictor of good outcome in patients that underwent IAT for BAO. Angioarchitectural variations of the posterior circulation were not found to impact reperfusion or clinical outcome.

  • Thrombectomy
  • Stroke
  • Intervention
  • Magnetic Resonance Angiography
  • CT Angiography

Statistics from

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.


Acute basilar artery occlusion (BAO) is associated with extremely high morbidity and fatality rates.1 Predictors of outcome have not been consistently demonstrated,2–4 likely due to the unique vascular anatomy of the vertebrobasilar system and the structural complexity of the brainstem. The size and configuration of the posterior circulation arteries have been suggested to influence posterior circulation flow volumes and to impact collateral flow patterns in the case of BAO.5–9

We aimed to evaluate if vascular anatomic features and variations of the vertebrobasilar arterial system influence the chances of successful reperfusion and the clinical outcome in intra-arterial therapy (IAT) for acute ischemic stroke (AIS) patients with BAO.


We performed a retrospective review of patients with AIS and BAO who underwent endovascular intervention in two academic medical centers from 2007 to 2012 and from 2009 to 2012, respectively. The study was approved by the local institutional review boards.

Clinical data

All consecutive patients treated with IAT for BAO within 24 h of symptom onset and with baseline non-invasive vascular imaging confirming large vessel occlusion were included. Baseline demographics were collected. Pre-intervention variables recorded included: initial National Institutes of Health Stroke Scale score, intravenous thrombolytic administration, and time from symptom onset to the start of endovascular therapy. BAO was classified into caudal (confluence of vertebral arteries (VAs) to the takeoff of the anterior inferior cerebellar arteries (AICAs)), mid-basilar (from the AICAs to the origin of the superior cerebellar arteries (SCAs)) or distal (SCAs and distal).10 Thrombolysis method used and vessel reperfusion score on the modified Treatment in Cerebral Ischemia (mTICI) scale were recorded.11 Safety measures included: procedural vessel rupture and reperfusion hemorrhages by the European Cooperative Stroke Study (ECASS) criteria on postprocedure CT within 48 h.12 Outcomes were defined using the modified Rankin Scale (mRS).

Radiologic data

Patients’ records were evaluated for the presence of brain MR angiography (MRA) or CT angiography (CTA). If both were available, CTA images were preferred for review. Image data were transferred to a workstation for analysis. CTA or MRA time of flight (TOF) source images were utilized to visualize the VA and posterior communicating artery (PCoA). The VA V3 segment was identified, and cuts that provided axial visualization of the vessel were utilized to establish the VA diameter. If an appropriate axial slice was not present, a slice that allowed the longitudinal visualization of a V3 segment with parallel outline was used (figure 1). The VA V3 diameter was calculated independently by two authors. Inter-rater agreement with CTA was very strong (Person's correlation coefficient r=0.83; p<0.01) and with MRA TOF VA was strong (r=0.70; p<0.01). Due to the potentially less reliable measurements by MRA TOF compared with CTA due to decreased spatial resolution, the diameter of the VA injected during the conventional angiogram was correlated with the ipsilateral VA diameter established by MRA, validating the measurements with a very strong correlation (r=0.81; p<0.01). The presence of hypoplastic VA (defined by a diameter <2 mm)13 and VAs that end in the posterior inferior cerebellar artery (PICA) were recorded. VA occlusion was defined by the presence of an occlusive lesion in the extracranial VA.

Figure 1

Vertebral artery V3 segment measurement (arrow: maximum diameter). Left: CT angiography. Right: MR angiography.

The diameter of each PCoA was measured on axial source images in the slice that would provide the largest vessel diameter and summed. The inter-rater correlation for PCoA diameter sum was very strong by CTA (r=0.81; p<0.01) and strong by MRA (r=0.70; p<0.01). PCoA was considered atretic if the diameter was <1 mm.14 ,15

In one institution (University of Miami/Jackson Memorial Hospital), a 128 section CTA protocol was used from the base of the arch through the cerebral vertex. Scanning parameters were: number of detectors 128; gantry rotation time 0.5 s; caudocranial acquisition; collimation 0.6 mm; pitch 0.8:1; 120 kV; 210 mA average dose; and 0.75 mm thickness sections. Maximum intensity projection algorithms were generated. The MRA head was a flow compensated three-dimensional TOF of the circle of Willis on a 1.5 T scanner. Scan parameters were as follows: flip angle=20, TR=22 ms, TE=6.9 ms, number of excitations=1, field of view=210 mm, field of view phase=80%, slice thickness=0.7 mm, number of slices per slab=44, slabs=6, acquisition matrix=320×320, and receiver bandwidth=130 Hz/Px. In the other institution (Harvard Medical School/Beth Israel Deaconess Medical Center), CTA were uniformly used, and a 64 section protocol was used from the level of the carina to the cerebral vertex. Scanning parameters were: number of detectors 64; no gantry tilt; craniocaudal acquisition; collimation 40 mm; pitch 0.984; 120 kV; 203 mA effective dose; reconstruction; and 1.25 mm×1 mm thickness sections. Maximum intensity projection algorithms were generated.

Statistical analysis

Results are reported as mean±SD. The primary analysis was performed using univariate linear regression analysis with 3 month mRS score as the conditional variable. Multivariate linear regression analysis with the variable selection method was performed and included variables with p<0.2 on univariate analysis. Secondary analysis evaluated factors associated with reperfusion. Statistical significance was set at p<0.05.


Of the combined 266 consecutive interventions for AIS performed at both institutions during the study period, 38 patients with BAO were identified. Mean age was 63.5 (±15.9) years and 52% were men. Demographics, treatment profile, and radiological/clinical outcomes are described in table 1. Mean National Institutes of Health Stroke Scale score at presentation was 21.3±9.3 and mean/median time from symptom onset to groin puncture were 10.5/7 h (25–75th percentile: 4–11 h). Seventy-one per cent of patients were selected for IAT based on CTA, compared with 28.9% by MRI. Most patients were treated with first generation devices, and TICI 2b-3 reperfusion was achieved in 68.4%. At 3 months, good outcomes (mRS ≤2) were observed in only 17.8% of cases, while the mortality rate was 64.3%.

Table 1

Univariate linear regression of baseline characteristics, treatment variables, and outcomes associated with reperfusion

The posterior circulation anatomical configuration variables are listed in table 2. Twenty-one (55%) patients had an atretic PCoA while 18 (47%) had a hypoplastic VA. The mean sum of the bilateral PCoA diameters was 2.3±1.2 mm, while the sum of the bilateral VA diameters was 5.2±5.2 mm. Twenty-three per cent of patients had a VA that ended in the PICA, while 42% had an extracranial VA occlusion. Basilar access in patients with VA occlusion was uniformly performed through the contralateral VA, and therefore no cervical VA angioplasty or stenting was performed. The BAO site was proximal in 36%, mid in 29%, and distal in 34%. None of these factors was statistically associated with reperfusion or outcome on univariate analysis.

Table 2

Univariate linear regression of baseline angioarchitectural variables (n=38)

Because of the possibility that the diagnosis of a VA <2 mm could represent a VA with distal collapse due to decreased perfusion pressure and not necessarily a hypoplastic VA, we performed sensitivity analyses. The correlation between VA hypoplasia and extracranial VA occlusion was weak (r=0.35; p=0.04), indicating that VA <2 mm potentially represents an inherently small artery.

On multivariate analysis, underlying hypertensive disease (unstandardized β=2.97; 95% CI 1.15 to 4.79; p<0.01) and reperfusion rate (unstandardized β=−0.40; 95% CI −0.74 to −0.70; p=0.02) were independently associated with outcomes. Multivariate analysis for predictors of reperfusion failed to identify other significant associations. A trend for better reperfusion with stent retriever devices was noted (unstandardized β=1.82; 95% CI −0.24 to 3.88; p=0.08).

Sensitivity analyses evaluating if PCoA bilateral atresia and/or agenesis could impact mRS or reperfusion rates were performed; however, no significant association was observed (data not shown).


Our primary finding is that variations in the posterior circulation angioarchitecture do not impact reperfusion or outcomes in patients undergoing IAT for BAO. Clot location, size of the VA or PCoA, and presence of extracranial VA occlusion or VA ending in the PICA were not observed to affect angiographic or clinical outcome.

The presence of a hypoplastic VA or vertebrobasilar system has been correlated with increased risk of ischemic stroke.5 ,16–18 Chuang et al17 and Perren et al9 demonstrated that the incidence of unilateral hypoplastic VA is higher in cases of brainstem/cerebellar infarction compared with patients with hemispherical stroke. Moreover, an association between the posterior circulation infarct side with ipsilateral topographic VA was noted.17 These findings have been suggested to be a result of a low flow state,18 which may be explained by the fact that the volume of blood flow through an artery is directly related to its diameter (proportional to the fourth power of its radius).14 In fact, VA hypoplasia has been reported to be associated with a lower net VA flow volume, and to lead to hemodynamic insufficiency.17 ,19 Moreover, a small VA could increase the susceptibility to thrombosis and poor clearance of proximal or distal thrombi.19 We quantified the size of the VA and recorded the presence of extracranial VA occlusion and/or VA ending in the PICA. However, none of the aforementioned VA anatomical variables were shown to impact reperfusion rates or overall outcomes in cases of large vessel occlusion undergoing IAT.

An inverse relationship between the diameter of the PCoA and ipsilateral PCA has been demonstrated.14 Therefore, we evaluated if having a larger PCoA, and consequently a smaller PCA P1 segment, would lead to a lower chance of reperfusion. Moreover, we aimed to assess if the presence of a larger PCoA would be associated with improved outcomes due to better collateralization to the top of the basilar artery. The size of the PCoA has been demonstrated to correlate with long term clinical outcomes in patients that undergo vertebral or basilar artery surgical occlusions to treat giant vertebrobasilar aneurysms. Steinberg et al demonstrated a clear correlation between ability to tolerate proximal or distal basilar artery ligation and the size of the PCoA.8 However, the impact of a small PCoA may not simply be explained by an enhanced potential for collateral flow due to the presence of compensatory routes, as Chuang et al demonstrated that even in the absence of internal carotid occlusion, PCoA hypoplasia is an independent contributor to the risk of ischemic stroke.6

The impact of PCoA on the endovascular management of BAO has not been well established. Cross et al investigated the potential implications of the number of PCoA in the setting of IAT for BAO.7 The authors found no correlation between neurological status at the time of treatment and the number of PCoAs, and there was no clear correlation between the size of the PCoA and neurologic outcome or survival. Our study included quantification of PCoA dimensions, and corroborated the absence of association between PCoA size and clinical outcome or chances of reperfusion.

A previous study has shown that survival was higher in patients with thrombus located in the distal basilar artery.7 However, our results indicate that the location of the clot within the basilar artery was not found to significantly affect outcome. Our findings corroborate those of Arnold et al who previously demonstrated that neither the location of the clot nor the presence of collateral circulation had a significant effect on recanalization rates or any clinical impact.2

There were several factors that limited our analysis. Non-contrast CT and CT perfusion were utilized for patient selection on the majority of cases, and these modalities are known to be suboptimal for the evaluation of the posterior fossa parenchyma compared with MRI. Clot length was not measured, which is likely a very important variable given BAO is a disease where perforators are involved.7 ,20 Most of the recently treated patients in this series did not have angiographic runs of the anterior circulation; hence, systematic collateral evaluation of retrograde basilar flow through the PCoA could not be performed. Leptomeningeal PICA to AICA, PICA to SCA, or AICA to SCA collateral flow was also not evaluated, although these are likely not a physiologically critical variable in BAO.8 The etiology of basilar artery strokes impacts outcome, and was not included in the analysis.21 Finally, our sample comprised patients presenting at a relatively late time window, and the benefit of reperfusion may have been dampened.

In summary, reperfusion emerged as an independent predictor of good outcome in patients that underwent endovascular therapy for BAO. The investigated angioarchitectural variations of the posterior circulation were not found to impact reperfusion rates or clinical outcomes.



  • Contributors DCH: Study conception, design of the work, acquisition of the data, statistical analysis, interpretation of the data, and drafting of the manuscript. SSD and DRY: Data acquisition, data interpretation, and critical revision of the manuscript. BS: Data interpretation and critical revision of the manuscript. V-AL, AT, ECP, MSE, and MAA-S: Data acquisition and critical revision of the manuscript. All authors gave final approval of the version to be published, and are in agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

  • Competing interests MAA-S is a proctor for ev3/Covidien Vascular Therapies (Mansfield, Massachusetts, USA) and Codman (Raynham, Massachusetts, USA). DRY is a consultant to Boston Scientific, Micrus, Abbott Vascular, and Coaxia, and has received travel support from Abbott Vascular Inc.

  • Ethics approval The study was approved by the University of Miami institutional review board and Beth Israel Deaconess institutional review board.

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

  • Data sharing statement Unpublished data from this dataset are held by Jackson Memorial Hospital/University of Miami School of Medicine and DRY. Requests for data sharing would be required to be discussed with DRY directly.