To investigate whether the observed association of leukoaraiosis with intracerebral hemorrhage is direct or mediated by risk factors, we compared 116 patients with intracerebral hemorrhage confirmed by computed tomography and 155 controls without intracerebral hemorrhage, evaluating the prevalence of leukoaraiosis and vascular risk factors. Leukoaraiosis was observed in 21 (18%) of the 116 patients and in 12 (8%) of the 155 controls (p<0.01). Only two (6%) of the 31 patients with lobar hemorrhage had leukoaraiosis on computed tomograms, compared with 17 (24%) of the 71 patients with basal ganglionic hemorrhage (p<0.05). Leukoaraiosis was significantly correlated with intracerebral hemorrhage after controlling for age and sex by using multiple logistic regression analysis, while the correlation disappeared after controlling for hypertension. Our results indicate that leukoaraiosis is not an independent risk factor for intracerebral hemorrhage. (Stroke 1990;21:1419-1423) T he term "leukoaraiosis" was introduced byHachinski et al 1 to describe computed tomographic (CT) images of symmetric patchy or diffuse areas of low density in the hemispheric white matter. Although leukoaraiosis has been reported in association with intracerebral hemorrhage (ICH), 2 it is not yet clear whether leukoaraiosis itself is a risk factor for ICH or whether this association is related to the presence of shared vascular risk factors. It is known that leukoaraiosis shares with ICH arterial hypertension as a possible risk factor 34 and that the parenchymal vasculopathy (hypertensive or amyloid) present in ICH has also been reported in leukoaraiosis. 3 We sought to determine whether vascular risk factors, in particular arterial hypertension, explain the presence of leukoaraiosis on cranial CT scans of patients with ICH.
Subjects and MethodsAll cranial CT scans were carried out by the Careggi Hospital Neuroradiology Service, the only city emergency neuroradiology service during the study period. From January 1,1985, to December 31, 1986, a CT diagnosis of spontaneous ICH was made Received December 15, 1989; accepted June 22, 1990. in 172 patients aged >40 years. Of these 172 patients, 41 were excluded from the study because of the presence of movement artifacts, large lesions with a mass effect, previous neurosurgery, or other aspects interfering with a proper evaluation of leukoaraiosis on CT pictures and a further 15 were excluded because of insufficient data on the risk factors to be evaluated. This selection produced the 116 patients with ICH involved in this study. We randomly selected 224 from among 7,304 cranial CT studies with no evidence of ICH performed during the same period in patients aged >40 years. Among these 224 patients, the 155 who met the criteria for proper definition of study variables were chosen as controls. They presented with transient ischemic attack or ischemic stroke (31), trauma (28), cerebral neoplasms (28), nonvascular or vascular dementia (18), epilepsy (11), headache (10), vertigo or syncope (8), depressi...