Reply

Is the deep white matter of adults more vulnerable to greater ischemia than the overlying cortex because of hemodynamic factors? Is the concept of an internal arterial borderzone valid? These important questions remain unanswered. We recently reported that structurally normal (on CT or T2-weighted MR images) white matter regions in patients with chronic carotid occlusive disease were not subject to a greater amount of ischemia than the overlying cortex (1). Yonas et al contend that this conclusion is valid for only a small group of patients, because of selection bias. In addition, they provide an anecdotal example of a patient with reduced white matter blood flow after blood pressure reduction to support their argument that the white matter is vulnerable to ischemia from hemodynamic mechanisms.

The patients included in our study were, in fact, quite representative of patients with carotid occlusion. These patients were enrolled in the St. Louis Carotid Occlusion Study, a longitudinal study of hemodynamic factors and stroke risk (2). Baseline epidemiologic and stroke risk factors for the 117 patients enrolled in this study were very similar to those reported from other studies of stroke risk, such as the North American Symptomatic Carotid Endarterectomy Trial (3). Patients with atherosclerotic carotid occlusion were invited to participate, regardless of whether the occlusion was bilateral or whether ischemic ocular or cerebral events had occurred.

We limited our analysis of white matter hemodynamics to structurally normal white matter. Consequently, we excluded patients for whom no imaging studies were available for review, as well as those having regions of interest located in structurally abnormal tissue. Whereas it is true that white matter hypodensity on CT or T2 hyperintensity on MR images does not always indicate tissue infarction, it often does. Damaged tissue may have impaired autoregulation in the absence of reduced perfusion pressure. Consequently, it is not valid, as Yonas et al suggest, to analyze such white matter regions and draw firm conclusions regarding the hemodynamic status of that tissue.

The patient illustrated in the figure in Yonas et al's letter is a case in point. The significance of the low flow in the deep white matter with blood pressure reduction is unknown. There are several possible explanations for the reduced white matter blood flow that do not necessarily invoke selective white matter ischemia caused by hemodynamic factors. The tissue may be dead, with no metabolic demand (hence no need for flow), or the tissue may be injured and unable to autoregulate normally. Furthermore, even if their observation of an apparent flow reduction actually represents viable tissue and failure of autoregulation to maintain blood flow, this alone does not prove the existence of a selectively vulnerable, deep white matter region. The degree of flow reduction in the white matter may be to the same degree as the overlying cortex. Normal baseline cerebral blood flow to the white matter is a fraction of that to the cortex, reflecting inherent differences in metabolic demand.

A longitudinal study of cerebral hemodynamics documenting that selective ischemia is present in the deep white matter (requiring the establishment of normal control values), and that this is associated with future white matter infarction, may be a better approach to addressing the question of an internal borderzone. Our conclusion that structurally normal white matter regions in patients with chronic carotid occlusive disease is not subject to a greater amount of ischemia than the overlying cortex is valid. It remains to be determined if selective ischemia (greater in the deep white matter than in the overlying cortex) caused by hemodynamic factors plays a role in the pathogenesis of deep white matter infarction in patients with acute carotid occlusion. Our data suggest that if selective ischemia does exist, it is unlikely to be a common phenomenon of chronically reduced perfusion pressure. We agree that further study regarding the presence of selective deep white matter ischemia is necessary.