Background and Purpose-We sought to determine the relations between infarct
subtype and white matter hyperintensities (WMHIs) on MRI.
Materials and Methods-We studied 395 ischemic stroke patients with 1.0-T MR
T. The number of lacunar, border-zone, and cortical infarcts was registered
. WMHIs were analyzed in 6 areas. Univariate and multivariate statistical a
nalyses were used to find the risk factors for different infarct subtypes a
nd to study the connections between WMHIs and brain infarcts.
Results-Lacunar infarcts were associated with hypertension (odds ratio [OR]
, 1.79; 95% CI, 1.17 to 2.73), alcohol consumption (OR, 1.96; 95% CI, 1.17
to 3.28), and age (OR, 1.03; 95% CI, 1.00 to 1.66). Border-zone infarcts we
re associated with carotid atherosclerosis (OR, 2.20; 95% CH, 1.15 to 4.19)
. Atrial fibrillation (OR, 3.02; 9.5% CI, 1.66 to 5.50) and carotid atheros
clerosis (OR, 1.94; 95% CI, 1.12 to 3.36) were independent positive predict
ors, and history of hyperlipidemia (OR, 0.44; 95% CI, 0.26 to 0.75) and mig
raine (OR, 0.48; 95% CI, 0.25 to 0.93) were negative predictors for cortica
l infarcts. Patients with lacunar infarcts had more severe WMHIs than patie
nts with nonlacunar infarcts in all WM areas (P less than or equal to 0.001
). Patients with border-zone infarcts showed severe periventricular lesions
(P=0.002), especially around posterior horns (P=0.003). The extent of WMHI
s in patients with cortical infarcts did not differ from that in those with
out cortical infarcts.
Conclusions-Various infarct subtypes have different risk profiles. The asso
ciation between lacunar infarcts and WMHIs supports the concept of small-ve
ssel disease underlying these 2 phenomena. The connection between border-zo
ne infarcts and periventricular WMHIs again raises the question of the disp
uted periventricular vascular border zone.