Cerebral blood flow (CBF) was measured in 39 men at normocapnia and after 5% CO2 inhalation using the xenon-133 technique. Twenty-three patients had unilateral carotid artery occlusion with no angiographic evidence of contralateral carotid artery stenosis or ophthalmic collateral flow. Eleven of these patients had undergone extracranial-intracranial (EC-IC) bypass surgery. Sixteen age-matched normal men underwent CBF measurements at normocapnia and hypercapnia to provide control data. Mean hemispheric CBF was not different between hemispheres ipsilateral and contralateral to the carotid artery occlusion either in the patients who had undergone bypass surgery or in those with carotid artery occlusion alone. Considering all patients with carotid artery occlusion, mean CO2 reactivity was decreased in the hemisphere ipsilateral to the occlusion as compared to the contralateral hemisphere in both groups. Based on data from normal individuals, a hemispheric difference in CO2 reactivity of more than 0.94%/mm Hg PaCO2 or a global CO2 reactivity of less than 0.66%/mm Hg PaCO2 was considered abnormal for an individual patient. Six of 23 patients with carotid artery occlusion (three with an EC-IC bypass) had global or hemispheric abnormalities in CO2 reactivity. Patients with impaired CO2 reactivity were not distinguishable from other patients by neurological examination, presence of transient ischemic attacks, or evidence of infarction on computerized tomography scanning. This test was safe and simple to perform and may be a useful means of detecting impaired cerebrovascular collateral reserve capacity. If impaired CO2 reactivity after carotid artery occlusion proves to be associated with a high risk of subsequent stroke, the test would provide a physiological basis for selecting a subgroup of patients who could be helped by cerebral revascularization.