Original Article

Access this article online

Quick Response Code:
Demographic age‑related variation in
Circle of Willis completeness assessed
by digital subtraction angiography
Ryan G. Eaton, Varun S. Shah1, David Dornbos III, Orel A. Zaninovich,
Website: Nicole Wenger1, Travis M. Dumont2, Ciarán J. Powers
http://www.braincirculation.org

DOI:
10.4103/bc.bc_43_19 Abstract
OBJECTIVE: Incomplete circle of Willis  (CoW) configuration is an important risk factor for
cerebrovascular pathology, namely aneurysm formation and ischemic stroke. This study was
performed to characterize CoW variation using digital subtraction angiography and to identify
demographic and physiologic features that may influence the risk of having an incomplete CoW
configuration.
MATERIALS AND METHODS: A retrospective review of 274 patients who underwent cerebral
angiography by a single surgeon for any indication was conducted. Each CoW branch was graded
as normal, hypoplastic, or aplastic. Univariate and multivariate regression analyses were conducted
to assess the impact of age, gender, race, and certain comorbidities on CoW configuration.
RESULTS: A  complete CoW was identified in 37.23% of patients. In univariate analysis,
patients  <40  years old were more likely to have a complete CoW  (odds ratio  [OR]: 4.973, 95%
confidence interval  [CI]: 2.610–9.476, P <  0.001) as were patients  <70  years old  (OR: 2.849,
95% CI: 1.131–7.194, P < 0.05). Univariate analysis on demographic factors and comorbidities
revealed CoW completeness to decrease with hypertension  (OR: 0.575, 95% CI: 0.347–0.951,
P = 0.031) and diabetes mellitus (OR: 0.368, 95% CI: 0.180–0.754, P = 0.006). Multivariable logistic
regression analysis used to assess the impact of age on CoW completeness showed age to be an
independent predictor of complete CoW, with an inverse correlation between increasing age and
CoW completeness  (OR: 0.955, 95% CI: 0.937–0.973, P < 0.001) after controlling for potential
confounders including hypertension and diabetes mellitus.
Department of
CONCLUSIONS: CoW configuration shows considerable variation with age; however, further
Neurological Surgery,
investigation is required to elucidate the full impact of other demographic and vascular risk factors
The Ohio State University
on CoW anatomy.
Wexner Medical Center,
1
College of Medicine, The Keywords:
Ohio State University Age, cerebral arteries, circle of Willis, demographic, digital subtraction angiography
Wexner Medical
Center, Columbus, OH,
2
Department of Surgery, Introduction anatomy of this eminent structure can
Division of Neurosurgery, have clinical repercussions. Experimental

T
The University of Arizona,
he Circle of Willis (CoW), the and computational models suggest that
Arizona, USA
epicenter of cerebral blood flow, increased flow through an abnormal
Address for provides crucial collateral circulation configuration of patent branches increases
correspondence: through its seven major branches. Given its wall sheer stress leading to aneurysmal
Dr. Ciarán J. Powers, importance in neurovascular physiology, formation. [1‑4] Similarly, incomplete
Department of Neurological
Surgery, The Ohio State
even slight variations to the normal configuration increases the risk of ischemic
University Wexner Medical stroke, likely due to reduced collateral
Center, 410 W 10th Ave., circulation.[4,5]
Columbus, OH 43210, This is an open access journal, and articles are
USA. distributed under the terms of the Creative Commons
E‑mail: ciaran.powers@ Attribution‑NonCommercial‑ShareAlike 4.0 License, which
osumc.edu allows others to remix, tweak, and build upon the work How to cite this article: Eaton RG, Shah VS,
non‑commercially, as long as appropriate credit is given and the Dornbos III D, Zaninovich OA, Wenger N, Dumont TM,
Submission: 30‑09‑2019 new creations are licensed under the identical terms. Powers CJ. Demographic age-related variation in
Revised: 31‑10‑2019
Circle of Willis completeness assessed by digital
Accepted: 26‑11‑2019
subtraction angiography. Brain Circ 2020;6:31-7.
Published:18-02-2020 For reprints contact: [email protected]

© 2020 Brain Circulation | Published by Wolters Kluwer Health – Medknow 31

 Eaton, et al.: Angiographic assessment of circle of Willis completeness

An incomplete CoW as a risk factor for stroke and Materials and Methods
aneurysm formation continues to be a focus of research
in cerebrovascular disease. In 1984, Kayembe et  al.[6] Patients
found aneurysms to be more common in autopsy studies The study was approved by the institutional review
in patients with CoW anatomic variations, a finding board with a waiver of consent, given that all data
suggested by CoW assessment with advanced imaging. were gathered retrospectively and anonymously. From
Individuals with variations in the CoW are more likely to January 2016 to March 2018, 322 consecutive patients
have aneurysmal recurrence after endovascular treatment[7] who underwent DSA for any indication by a single
and may be at higher risk for aneurysmal rupture.[8,9] surgeon were included in the study. Clinical indications
Variation of the first segment of the anterior cerebral for DSA included ischemic stroke, cerebral aneurysm,
artery increases the likelihood of having an anterior suspected vasospasm, intracerebral hemorrhage, vessel
communicating artery aneurysm[10] and puts individuals at dissection, transverse sinus stenosis, severe epistaxis,
risk for infarct after aneurysmal rupture.[11] Acute ischemic arteriovenous malformation, and dural arteriovenous
stroke risk is similarly elevated in individuals with variant fistula. Of the initial 322 patients, 48 were excluded due
CoW configurations.[5,12‑18] Individuals with good cerebral to (1) incomplete imaging due to either missing internal
collateral circulation tend to have smaller infarct size at carotid artery or vertebral artery injections, (2) occlusion
baseline and higher rates of neurological improvement or severe stenosis of vasculature proximal to the CoW,
following acute ischemic stroke.[19,20] or (3) patients with moderate or severe vasospasm.
Patients diagnosed with Moyamoya disease were also
A recent work by Zaninovich et  al.[21] supports the excluded from the study.
emerging theory that age and gender have a significant
effect on the rates of complete CoW, which subsequently Patient information was gathered via a retrospective
impacts these disease states. Women have a higher review of personal health information. Demographic
lifetime risk of stroke, more frequent recurrences, information was used to compile age, gender, and
and higher mortality, whereas men have the risk of race. The International Classification of Diseases‑9
first stroke at a younger age.[22] Notably, stroke risk and International Classification of Diseases‑10 codes
strongly correlates with age in both genders.[22] Ghods were used to describe whether patients had any of the
et  al. showed that gender affects aneurysmal location, comorbidities of interest, including Type II diabetes
in which women are more likely to have multiple mellitus, hypertension, coronary artery disease,
aneurysms, aneurysms of the internal carotid artery, and peripheral artery disease, prior cerebrovascular
tend to present with subarachnoid hemorrhage.[23] With accident or transient ischemic accident, and obstructive
regard to age, middle cerebral artery aneurysms are less sleep apnea at the time of angiography. Body mass
common with age >55 years, and women are more likely index was calculated using the patients’ height and
to present later in life.[23,24] weight measurements. Hyperlipidemia was defined
as any preangiography low‑density lipoprotein
Digital subtraction angiography (DSA) remains the gold value ≥100 mg/dl. Smoking status was determined by
standard for assessing cerebral arteries and is particularly evaluating documentation obtained preangiography,
successful in identifying hypoplastic vessels when with any documented previous smoking history being
compared to computed tomography angiography (CTA) considered positive.
and magnetic resonance angiography (MRA).[25‑27] The
primary objective of this study was to describe CoW Image acquisition
completeness with respect to age, gender, and race as DSA images were acquired via transfemoral approach
assessed by DSA, rather than noninvasive imaging, with a 4 Fr or 5 Fr catheter in a fully equipped DSA
which tends to be less specific. To our knowledge, this is unit. A catheter was advanced into the right and left
the largest patient cohort to be assessed by DSA, allowing internal carotid arteries and a vertebral artery. Contrast
for characterization of anatomic variations of each major was injected manually, and images were captured at
vessel, including fetal posterior communicating arteries. 3 frames/s. Anteroposterior and lateral images were
The mechanism underlying apparent changes in CoW obtained with oblique images variably obtained based
configuration with age has yet to be elucidated, although on pathology identified during angiography.
this is hypothesized to be secondary to a combination of
factors including atherosclerosis.[28,29] As such, this study Image analysis
also seeks to describe the population and patient factors All DSA images were reviewed by three blinded
associated with incomplete CoW with respect to known reviewers. To minimize differences between prior studies
atherosclerotic risk factors, namely blood pressure, on CoW configuration, the major branches of the CoW
low‑density lipoprotein, smoking, diabetes mellitus, and were considered to be the anterior communicating
body mass index.[30] artery, the first segments of both anterior cerebral
32 Brain Circulation ‑ Volume 6, Issue 1, January‑March 2020
 Eaton, et al.: Angiographic assessment of circle of Willis completeness

arteries, the first segments of both posterior cerebral most common variants were an aplastic left posterior
arteries, and both posterior communicating arteries.[21,25] communicating artery (31.75%) and a hypoplastic right
Each segment was graded as normal, hypoplastic, or posterior communicating artery (29.56%). The incidence
aplastic [Figure  1]. Hypoplastic was defined as vessel of normal vessels in the anterior circulation was 68.25%
size <30% of the size of the ipsilateral distal vessel. For the for anterior communicating artery and 95.26% and
posterior communicating artery and the first segment of 91.24% for the first segments of left and right anterior
posterior cerebral artery, this was the second segment of cerebral arteries, respectively. The left and right posterior
the posterior cerebral artery, and for anterior circulation communicating arteries were normal in 25.18% and
arteries, this was the second segment of the anterior 25.91% of cases, respectively, whereas the left and right
cerebral artery.[25] Aplastic vessels were defined as a posterior cerebral arteries were normal in caliber in
vessel that could not be identified on any DSA image. 88.32% and 85.77% of cases, respectively. Hypoplastic
and aplastic vessels were seen in greatest frequency
Statistical analysis in the anterior communicating artery and posterior
The incidence of each vessel segment of the CoW communicating arteries. Fetal posterior communicating
was tabulated for the entire cohort. The frequency of arteries were present in 14.60% of patients on the left and
incomplete CoW was defined as any CoW with at least 16.79% on the right.
one aplastic vessel and was calculated for three age
groups (ages <40, 40–69, and ≥70 years). The incidence Patients aged <40, 40–69, and ≥70 years had a complete
of incomplete CoW was found within each demographic CoW in 67.92%, 31.75%, and 18.75%, respectively [Table 2].
and comorbid disease group of interest including gender, On univariate analysis, patients <40 years old were more
race, diabetes mellitus, smoking status, hypertension, likely to have a complete CoW (odds ratio [OR]: 4.973,
hyperlipidemia, coronary artery disease or peripheral 95% confidence interval [CI]: 2.610–9.476, P < 0.001) as
artery disease, prior cerebrovascular accident including were patients <70 years old (OR: 2.849, 95% CI: 1.131–
transient ischemic accident, obstructive sleep apnea, and 7.194, P < 0.05) compared to patients aged ≥70 years.
obesity. Univariate binary logistic and linear regression Overall, increasing age correlates with total aplastic
analyses were used to assess the impact of categorical and and hypoplastic vessels (β = 0.017, 95% CI: 0.011–0.024,
continuous dependent variables, respectively. Finally, a P < 0.001).
multivariable logistic regression analysis was conducted
to assess the potential predictors of CoW completeness Table 1: Original data on the incidence of normal,
with respect to age using co‑variants that approached hypoplastic, aplastic, and fetal circle of Willis branches
statistical significance (P < 0.20) on univariate analysis. Normal, Hypoplastic, Aplastic, Fetal,
n (%) n (%) n (%) n (%)
Results A‑CommA 187 (68.25) 56 (20.44) 31 (11.31) ‑
Left A1 261 (95.26) 10 (3.65) 3 (1.09) ‑
The overall incidence of patent, hypoplastic, and aplastic Right A1 250 (91.24) 12 (4.38) 12 (4.38) ‑
vessels for each of the seven major branches of the CoW is Left P‑CommA 69 (25.18) 78 (28.47) 87 (31.75) 40 (14.60)
listed in Table 1. The overall incidence of a complete CoW Right P‑CommA 71 (25.91) 81 (29.56) 76 (27.74) 46 (16.79)
in the entire patient cohort was 37.23% when incomplete Left P1 242 (88.32) 11 (4.01) 21 (7.66) ‑
CoW was defined as any aplastic vessels. The incidence Right P1 235 (85.77) 20 (7.30) 19 (6.93) ‑
A‑CommA: Anterior communicating artery, A1: First segment of anterior
of a complete CoW was 8.03% when an incomplete CoW cerebral artery, P‑CommA: Posterior communicating artery, P1: First segment
was defined as any hypoplastic or aplastic vessels. The of posterior cerebral artery

Figure 1: Original digital subtraction angiography image of posterior communicating artery variation. Examples of hypoplastic (left), normal (middle), and aplastic (right)
posterior communicating artery

Brain Circulation ‑ Volume 6, Issue 1, January‑March 2020 33

 Eaton, et al.: Angiographic assessment of circle of Willis completeness

There was no significant difference in the frequency which has been cited as a potential risk factor for ischemic
of CoW completeness with respect to gender or strokes.[16] The incidence of approximately 15% overall in
race [Table 3]. On univariate regression analysis [Table 4], this study was comparable to an incidence of 11%–46%
both hypertension (OR: 0.575, 95% CI: 0.347–0.951, documented in a recent review.[51]
P  = 0.031) and diabetes mellitus (OR: 0.368, 95% CI:
0.180–0.754, P = 0.006) correlated with decreased CoW The degree of environmental and genetic contributions
completeness. Multivariate regression analysis showed to CoW configuration remains unclear.[52] In a mouse
age to be an independent predictor of complete CoW model,[28] both aging and hypertension were found
with an inverse correlation between increasing age and to reduce posterior communicating artery diameter.
CoW completeness (OR: 0.955, 95% CI: 0.937–0.973, However, it has been demonstrated that posterior
P < 0.001) when controlling for potential confounding communicating artery variations occur more often within
variables including African‑American race, other race, a family.[29] In this study, younger patient age was highly
hypertension, history of cerebrovascular accident, and predictive of CoW completeness, a relationship becoming
diabetes mellitus. Importantly, when controlling for age, increasingly well established in literature.[21,32,37‑39] The
diabetes mellitus and hypertension did not significantly results of this study strengthen this association, finding
correlate with CoW completeness status. age to be an independent predictor of complete CoW, even
after controlling for hypertension and diabetes mellitus.
Discussion
The two other demographic variables assessed (gender
CoW completeness has varied considerably in imaging[31‑42] and race) showed no significant difference in CoW
and autopsy[6,43‑50] studies. The completeness rate in this completeness. Previous studies have found a higher
study replicates a recent large sample study with a female preponderance;[32,34,38] however, others[39,53] have
similar focus using CTA.[21] The functional assessment reported opposite findings, with men having a slightly
afforded by DSA allows for accurate characterization of higher completeness. Racial and ethnic differences in
small‑diameter vessels into aplastic versus hypoplastic. configuration have been a recent research focus, in which
This nuanced evaluation of hypoplastic vessels altered a Sri Lankan autopsy study[46] demonstrated that CoW
the completion frequency considerably when hypoplastic configuration varied among ethnicity. MRA evaluation
vessels were considered incomplete. The largest DSA of CoW in Ecuadorian Mexicans,[54] Turkish,[42] and
CoW analysis, a 117‑patient series conducted by Han Pakistani[41] populations yielded variable completeness
et al.[25] previously with an analogous vascular grading incidence of 65.10%, 85%, and 22.2%, respectively. Due
scale, showed high variability in the communicating to challenges in assessing ethnic background via the
segments relative to the first segments of both the personal health information system, this study estimated
anterior and posterior cerebral arteries. This high degree differences in ethnicity using race, and thus it is possible
of variation was replicated in this study, with the that a more refined categorization with respect to ethnic
anterior and posterior communicating arteries showing background may show a difference and should be a
considerable variation, an anomaly that has known consideration for similar studies in future.
significant clinical ramifications in cerebrovascular
pathology. [10,11] To our knowledge, this is the first Atherosclerosis follows an analogous age‑based
dedicated DSA assessment of fetal posterior circulation, trajectory to CoW completeness frequency changes

Table 2: Original data on the impact of age on circle of Willis completeness

Age Patients Completion Mean age Univariate analysis for Univariate analysis for total hypoplastic and
(years) (n) frequency (%) (SD) complete CoW, OR (95% CI); P aplastic vessels, β coefficient (95% CI); P
<40 53 67.92 28.94 (6.08) 4.975 (2.611-9.434); <0.001 −0.571 (−0.833-−0.309); <0.001
40-69.99 189 31.75 55.77 (8.15) 2.101 (1.244-3.546); <0.01 0.315 (0.087-1.848); <0.01
≥70 32 18.75 74.72 (4.19) 0.351 (0.139-0.884); 0.026 0.211 (−0.121-0.543); 0.211
Total 274 37.23 52.79 (15.08) 0.954 (0.937-0.972); <0.001 0.017 (0.011-0.024); <0.001
SD: Standard deviation, OR: Odds ratio, CI: Confidence interval, CoW: Circle of Willis

Table  3: Original data on the influence of demographic factors on circle of Willis completeness
Patients (n) Age, mean (SD) Completion frequency (%) OR (95% CI); P
Female 175 51.21 (14.84) 38.29 1.134 (0.679-1.894); 0.630
Male 99 55.59 (15.10) 35.35 ‑
Caucasian 217 52.98 (15.42) 36.41 1.165 (0.635-2.135); 0.622
African‑American 55 52.14 (13.50) 40.00 0.570 (0.295-1.098); 0.093
OR: Odds ratio, CI: Confidence interval, SD: Standard deviation

34 Brain Circulation ‑ Volume 6, Issue 1, January‑March 2020

 Eaton, et al.: Angiographic assessment of circle of Willis completeness

Table  4: Original data on the impact of risk factors is of descriptive and retrospective nature and future
for atherosclerotic vascular disease on circle of Willis work should focus on prospectively following a health
completeness population over time to further solidify the effect of age
Patients Completion OR (95% CI); P on CoW configuration.
(n) frequency (%)
Smokers 161 34.78 0.801 (0.486-
1.318); 0.382
Conclusions
Nonsmokers 108 40.74
Hyperlipidemia 42 30.95 0.614 (0.270-
CoW anatomy, as assessed by DSA, shows considerable
LDL <100 mg/dL 64 42.19 1.396); 0.245
variation. Complete CoW is closely and inversely related
Hypertension 170 32.35 0.575 (0.347-
0.951); 0.031
to age. The relationship between CoW completeness and
No hypertension 103 45.63
other demographic factors, such as race and gender,
CAD or PAD 49 40.82 1.186 (0.631-
2.232); 0.597 did not reach significance in this study. Furthermore,
No CAD or PAD 224 36.61
Prior CVA 56 46.43 1.597 (0.881-
comorbid conditions commonly associated with
No prior CVA 217 35.02 2.890); 0.123 atherosclerosis and cerebrovascular disease did not
Diabetes mellitus 53 20.75 0.368 (0.180-
significantly affect CoW completeness when controlling
No diabetes mellitus 220 41.36 0.754); 0.006 for age. Further work is needed to elucidate the
OSA 20 35.00 0.895 (0.335- mechanism behind the age‑related decline in CoW
No OSA 253 37.55 2.326); 0.821 completeness and to anticipate which individuals are at
BMI >30 106 36.79 0.986 (0.595- risk of CoW incompleteness and its associated pathologic
Normal BMI 167 37.13 1.631); 0.956 effects.
LDL: Low‑density lipoprotein, CAD: Coronary artery disease, PAD: Peripheral
artery disease, CVA: Cerebrovascular accident, OSA: Obstructive sleep
apnea, BMI: Body mass index, OR: Odds ratio, CI: Confidence interval
Financial support and sponsorship
Nil.

and is heavily affected by blood pressure, low‑density Conflicts of interest

lipoprotein, smoking, diabetes mellitus, and body mass There are no conflicts of interest.
index.[30] One study found no relationship between
cerebral small‑vessel disease and CoW completeness,[55] References
supporting a potential hypothesis that configuration
changes require continuous vascular demand on 1. Fahy P, McCarthy P, Sultan S, Hynes N, Delassus P, Morris L. An
experimental investigation of the hemodynamic variations due
large‑diameter arteries. This study showed that to aplastic vessels within three‑dimensional phantom models of
hypertension and diabetes mellitus were associated the circle of Willis. Ann Biomed Eng 2014;42:123‑38.
with an increased odds of having an incomplete CoW. 2. Alnaes MS, Isaksen J, Mardal KA, Romner B, Morgan MK,
However, when included in a multivariate analysis, Ingebrigtsen T. Computation of hemodynamics in the circle of
Willis. Stroke 2007;38:2500‑5.
age was found to be the primary mechanistic driver
3. Nam SW, Choi S, Cheong Y, Kim YH, Park HK. Evaluation of
impacting CoW vessel anatomy, rather than these aneurysm‑Associated wall shear stress related to morphological
confounding comorbidities. Interestingly, other variations of circle of Willis using a microfluidic device. J Biomech
comorbidities analyzed which are also strongly 2015;48:348‑53.
associated with age such as hyperlipidemia, coronary 4. Pascalau R, Padurean VA, Bartos D, Bartos A, Szabo BA. The
geometry of the circle of Willis anatomical variants as a potential
artery disease or peripheral artery disease, and prior cerebrovascular risk factor. Turk Neurosurg 2019;29:151-8.
cerebrovascular accident did not reach significance in the 5. Hoksbergen AW, Legemate DA, Csiba L, Csáti G, Síró P, Fülesdi B.
regression analysis. It is certainly possible that this study Absent collateral function of the circle of Willis as risk factor for
was underpowered to adequately evaluate the impact of ischemic stroke. Cerebrovasc Dis 2003;16:191‑8.
the conditions and thus a larger review may be helpful 6. Kayembe KN, Sasahara M, Hazama F. Cerebral aneurysms and
variations in the circle of Willis. Stroke 1984;15:846‑50.
in further establishing this relationship. 7. Songsaeng D, Geibprasert S, Willinsky R, Tymianski M,
TerBrugge KG, Krings T. Impact of anatomical variations of
The population of patients undergoing cerebral the circle of Willis on the incidence of aneurysms and their
angiography is weighed toward individuals with recurrence rate following endovascular treatment. Clin Radiol
2010;65:895‑901.
cerebrovascular conditions, specifically evaluation
8. de Rooij NK, Velthuis BK, Algra A, Rinkel GJ. Configuration of
of stroke and aneurysm, which is known to decrease the circle of Willis, direction of flow, and shape of the aneurysm
CoW completeness.[21,56] Unfortunately, this problem as risk factors for rupture of intracranial aneurysms. J Neurol
is inherent to assessments made by DSA which is 2009;256:45‑50.
an invasive test, requiring appropriate indications. 9. Stojanović N, Stefanović I, Randjelović S, Mitić R, Bosnjaković P,
Stojanov D. Presence of anatomical variations of the circle of
Using a study population more representative of a Willis in patients undergoing surgical treatment for ruptured
healthy population using CTA or MRA imaging would intracranial aneurysms. Vojnosanit Pregl 2009;66:711‑7.
potentially produce different results. Finally, this study 10. Krasny A, Nensa F, Sandalcioglu IE, Göricke SL, Wanke I,

Brain Circulation ‑ Volume 6, Issue 1, January‑March 2020 35

 Eaton, et al.: Angiographic assessment of circle of Willis completeness

Gramsch C, et al. Association of aneurysms and variation of the posterior communicating collaterals of the circle of Willis. Transl
A1 segment. J Neurointerv Surg 2014;6:178‑83. Stroke Res 2019;10:189‑203.
11. Rinaldo L, McCutcheon BA, Snyder KA, Porter AL, Bydon M, 29. Sánchez van Kammen M, Moomaw CJ, van der Schaaf IC,
Lanzino G, et al. A1 segment hypoplasia associated with cerebral Brown RD Jr, Woo D, Broderick JP, et al. Heritability of circle of
infarction after anterior communicating artery aneurysm rupture. Willis variations in families with intracranial aneurysms. PLoS
J Neurosurg Sci 2019;63:359‑64. One 2018;13:e0191974.
12. Moritz A, Koci G, Steinlechner B, Hölzenbein T, Nasel C, 30. Pencina MJ, D’Agostino RB Sr., Larson MG, Massaro JM,
Grubhofer G, et al. Contralateral stroke during carotid Vasan RS. Predicting the 30‑year risk of cardiovascular disease:
endarterectomy due to abnormalities in the circle of Willis. Wien The Framingham Heart Study. Circulation 2009;119:3078‑84.
Klin Wochenschr 2007;119:669‑73. 31. Macchi C, Catini C, Federico C, Gulisano M, Pacini P, Cecchi F,
13. Chuang YM, Liu CY, Pan PJ, Lin CP. Posterior communicating et al. Magnetic resonance angiographic evaluation of circulus
artery hypoplasia as a risk factor for acute ischemic stroke arteriosus cerebri (circle of Willis): A morphologic study in 100
in the absence of carotid artery occlusion. J Clin Neurosci human healthy subjects. Ital J Anat Embryol 1996;101:115‑23.
2008;15:1376‑81. 32. Krabbe‑Hartkamp MJ, van der Grond J, de Leeuw FE, de Groot JC,
14. Lochner P, Golaszewski S, Caleri F, Ladurner G, Tezzon F, Algra A, Hillen B, et al. Circle of Willis: Morphologic variation
Zuccoli G, et al. Posterior circulation ischemia in patients with on three‑dimensional time‑of‑flight MR angiograms.
fetal‑type circle of Willis and hypoplastic vertebrobasilar system. Radiology 1998;207:103‑11.
Neurol Sci 2011;32:1143‑6. 33. Hartkamp MJ, van Der Grond J, van Everdingen KJ, Hillen B,
15. Miyazawa N, Shinohara T, Yamagata Z. Association of Mali WP. Circle of Willis collateral flow investigated by magnetic
incompleteness of the anterior part of the circle of Willis with resonance angiography. Stroke 1999;30:2671‑8.
the occurrence of lacunes in the basal ganglia. Eur J Neurol 34. Horikoshi T, Akiyama I, Yamagata Z, Sugita M, Nukui H.
2011;18:1358‑60. Magnetic resonance angiographic evidence of sex‑linked
16. Arjal RK, Zhu T, Zhou Y. The study of fetal‑type posterior cerebral variations in the circle of Willis and the occurrence of cerebral
circulation on multislice CT angiography and its influence on aneurysms. J Neurosurg 2002;96:697‑703.
cerebral ischemic strokes. Clin Imaging 2014;38:221‑5. 35. Malamateniou C, Adams ME, Srinivasan L, Allsop JM, Counsell SJ,
17. Ryan DJ, Byrne S, Dunne R, Harmon M, Harbison J. White Cowan FM, et al. The anatomic variations of the circle of Willis
matter disease and an incomplete circle of Willis. Int J Stroke in preterm‑at‑term and term‑born infants: An MR angiography
2015;10:547‑52. study at 3T. AJNR Am J Neuroradiol 2009;30:1955‑62.
18. van Seeters T, Hendrikse J, Biessels GJ, Velthuis BK, Mali WP, 36. Li Q, Li J, Lv F, Li K, Luo T, Xie P. A multidetector CT angiography
Kappelle LJ, et al. Completeness of the circle of Willis and risk study of variations in the circle of Willis in a Chinese population.
of ischemic stroke in patients without cerebrovascular disease. J Clin Neurosci 2011;18:379‑83.
Neuroradiology 2015;57:1247‑51. 37. Tanaka H, Fujita N, Enoki T, Matsumoto K, Watanabe Y,
19. Leng X, Lan L, Liu L, Leung TW, Wong KS. Good collateral Murase K, et al. Relationship between variations in the circle
circulation predicts favorable outcomes in intravenous of Willis and flow rates in internal carotid and basilar arteries
thrombolysis: A systematic review and meta‑analysis. Eur J determined by means of magnetic resonance imaging with
Neurol 2016;23:1738‑49. semiautomated lumen segmentation: Reference data from 125
20. Fanou EM, Knight J, Aviv RI, Hojjat SP, Symons SP, Zhang L, healthy volunteers. AJNR Am J Neuroradiol 2006;27:1770‑5.
et al. Effect of collaterals on clinical presentation, Baseline 38. Naveen SR, Bhat V, Karthik GA. Magnetic resonance angiographic
Imaging, complications, and outcome in acute stroke. AJNR Am evaluation of circle of Willis: A morphologic study in a tertiary
J Neuroradiol 2015;36:2285‑91. hospital set up. Ann Indian Acad Neurol 2015;18:391‑7.
21. Zaninovich OA, Ramey WL, Walter CM, Dumont TM. Completion 39. Jin ZN, Dong WT, Cai XW, Zhang Z, Zhang LT, Gao F, et al. CTA
of the Circle of Willis varies by gender, age, and indication characteristics of the circle of Willis and intracranial aneurysm in
for computed tomography angiography. World Neurosurg a Chinese crowd with family history of stroke. Biomed Res Int
2017;106:953‑63. 2016;2016:1-11.
22. Writing Group Members, Mozaffarian D, Benjamin EJ, Go AS, 40. Karatas A, Coban G, Cinar C, Oran I, Uz A. Assessment of the
Arnett DK, Blaha MJ, et al. Heart disease and stroke statistics‑2016 circle of Willis with cranial tomography angiography. Med Sci
update: A Report from the American Heart Association. Monit 2015;21:2647‑52.
Circulation 2016;133:e38‑360. 41. Shaikh R, Sohail S. MRA‑based evaluation of anatomical
23. Ghods AJ, Lopes D, Chen M. Gender differences in cerebral variation of circle of Willis in adult Pakistanis. J Pak Med Assoc
aneurysm location. Front Neurol 2012;3:78. 2018;68:187‑91.
24. Lindner SH, Bor AS, Rinkel GJ. Differences in risk factors 42. Yeniçeri IÖ, Çullu N, Deveer M, Yeniçeri EN. Circle of Willis
according to the site of intracranial aneurysms. J Neurol variations and artery diameter measurements in the Turkish
Neurosurg Psychiatry 2010;81:116‑8. population. Folia Morphol (Warsz) 2017;76:420‑5.
25. Han A, Yoon DY, Chang SK, Lim KJ, Cho BM, Shin YC, et al. 43. Orlandini GE, Ruggiero C, Orlandini SZ, Gulisano M. Blood
Accuracy of CT angiography in the assessment of the circle of vessel size of circulus arteriosus cerebri (circle of Willis):
Willis: Comparison of volume‑rendered images and digital A statistical research on 100 human subjects. Acta Anat (Basel)
subtraction angiography. Acta Radiol 2011;52:889‑93. 1985;123:72‑6.
26. Stock KW, Wetzel S, Kirsch E, Bongartz G, Steinbrich W, 44. Kapoor K, Singh B, Dewan LI. Variations in the configuration of
Radue EW. Anatomic evaluation of the circle of Willis: MR the circle of Willis. Anat Sci Int 2008;83:96‑106.
angiography versus intraarterial digital subtraction angiography. 45. Eftekhar B, Dadmehr M, Ansari S, Ghodsi M, Nazparvar B,
AJNR Am J Neuroradiol 1996;17:1495‑9. Ketabchi E. Are the distributions of variations of circle of Willis
27. Patrux B, Laissy JP, Jouini S, Kawiecki W, Coty P, Thiébot J. different in different populations? – Results of an anatomical
Magnetic resonance angiography (MRA) of the circle of Willis: study and review of literature. BMC Neurol 2006;6:22.
A prospective comparison with conventional angiography in 54 46. De Silva KR, Silva R, Amaratunga D, Gunasekera WS,
subjects. Neuroradiology 1994;36:193‑7. Jayesekera RW. Types of the cerebral arterial circle (circle of Willis)
28. Faber JE, Zhang H, Rzechorzek W, Dai KZ, Summers BT, Blazek C, in a Sri Lankan population. BMC Neurol 2011;11:5.
et al. Genetic and environmental contributions to variation in the 47. Hashemi SM, Mahmoodi R, Amirjamshidi A. Variations in the

36 Brain Circulation ‑ Volume 6, Issue 1, January‑March 2020

 Eaton, et al.: Angiographic assessment of circle of Willis completeness

Anatomy of the Willis’ circle: A 3‑year cross‑sectional study from 52. Forgo B, Tarnoki AD, Tarnoki DL, Kovacs DT, Szalontai L,
Iran (2006‑2009). Are the distributions of variations of circle of Persely A, et al. Are the Variants of the circle of Willis determined
Willis different in different populations? Result of an anatomical by genetic or environmental factors? Results of a twin study and
study and review of literature. Surg Neurol Int 2013;4:65. review of the literature. Twin Res Hum Genet 2018;21:384‑93.
48. Gunnal SA, Farooqui MS, Wabale RN. Anatomical variations of 53. Macchi C, Lova RM, Miniati B, Gulisano M, Pratesi C, Conti AA,
the circulus arteriosus in cadaveric human brains. Neurol Res Int et al. The circle of Willis in healthy older persons. J Cardiovasc
2014;2014:1-16. Surg (Torino) 2002;43:887‑90.
49. Klimek‑Piotrowska W, Rybicka M, Wojnarska A, 54. Del Brutto OH, Lama J. Variants in the circle of Willis and
Wójtowicz A, Koziej M, Hołda MK. A multitude of variations in white matter disease in Ecuadorian mestizos. J Neuroimaging
the configuration of the circle of Willis: An autopsy study. Anat 2015;25:124‑6.
Sci Int 2016;91:325‑33. 55. Del Brutto OH, Mera RM, Zambrano M, Lama J. Incompleteness
50. Karatas A, Yilmaz H, Coban G, Koker M, Uz A. The Anatomy of of the circle of Willis correlates poorly with imaging evidence
circulus arteriosus cerebri (Circle of Willis): A study in Turkish of small vessel disease. A population‑based study in rural
population. Turk Neurosurg 2016;26:54‑61. Ecuador (the Atahualpa project). J Stroke Cerebrovasc Dis
51. Lambert SL, Williams FJ, Oganisyan ZZ, Branch LA, Mader EC 2015;24:73‑7.
Jr. Fetal‑type variants of the posterior cerebral Artery and 56. Varga A, Di Leo G, Banga PV, Csobay‑Novák C, Kolossváry M,
concurrent infarction in the major arterial territories of the Maurovich‑Horvat P, et al. Multidetector CT angiography of the
cerebral hemisphere. J Investig Med High Impact Case Rep circle of Willis: Association of its variants with carotid artery
2016;4:1-3. disease and brain ischemia. Eur Radiol 2019;29:46‑56.

Brain Circulation ‑ Volume 6, Issue 1, January‑March 2020 37