Background The Circle of Willis (CoW) is a circulatory anastomosis comprised of seven arterial segments distributing cerebral blood flow (CBF). The circle plays a critical role in supporting collateral circulation and diffusing pulsatile pressure to protect downstream microvasculature. Approximately 75% of adults have missing or hypoplastic CoW segments, which could adversely affect collateral flow and the homeostasis of microvasculature, resulting in small vessel disease. This study examined whether missing anterior cerebral artery (A1) or posterior cerebral artery (P1) segments of the CoW relates to global or regional white matter hyperintensities (WMHs), a purported neuroimaging marker of small vessel disease.
Methods Vanderbilt Memory & Aging Project participants with normal cognition and no history of clinical stroke were studied (n= 167, 72±7 years, 42% female). Multimodal 3T brain MRI included magnetic resonance angiography (MRA) for examining CoW patency and FLAIR for quantifying WMHs. Using MRA, A1 and P1 segments were manually coded as present or absent by a board-certified neuroradiologist. Linear regressions related A1 and P1 variants to WMHs, adjusting for age, sex, race/ethnicity, education, Framingham Stroke Risk Profile, diagnosis, apolipoprotein E4 status, and intracranial volume.
Results The absence of a P1 segment (present in 5% of the cohort) related to increased WMHs globally (b=12.5, p=0.0009) and regionally in the parietal (b=4.9, p=0.0003), temporal (b=1.2, p=0.0007), and occipital lobes (b=2.4, p=0.0008) but was unrelated to WMHs in the frontal lobe (b=3.8, p=0.10). In contrast, the absence of one A1 segment (present in only 1% of the cohort) was unrelated to WMHs regionally or globally (p-values>0.15).
Conclusions Among cognitively normal older adults, missing a P1 segment correlates with increased WMHs in the parietal, temporal, and occipital lobes, perhaps due to alterations in CBF homeostasis to the microvasculature of the posterior cortex. In contrast, a missing A1 segment did not relate to WMHs. This latter result may be due to insufficient power given the low prevalence of missing A1 segments in this cohort.