The quality of collateral blood supply in carotid disease is pivotal for the resulting hemodynamic compromise. However, the interrelation between different patterns of collateral blood flow and actual impairment of cerebral autoregulation (CAR) has not been analyzed so far. Dynamic CAR was assessed noninvasively by the phase shift between respiratory-induced 0.1-Hz oscillations of arterial blood pressure (ABP) and cerebral blood flow velocity (CBFV) in 101 patients with severe unilateral carotid stenosis (> or = 75%) or occlusion. CO(2)-vasomotor reactivity was assessed via inhalation of 7% CO(2). Spontaneously activated collaterals via primary (anterior/posterior communicating artery, type I, n = 65) and secondary (ophthalmic artery / leptomeningeal with or without primary pathways, group II, n = 24) pathways were assessed by transcranial Doppler/duplex sonography. Signs of functional stenosis in the anterior collateral pathways were subsumed under type III (n = 12). Best dynamic CAR (phase shift) on affected sides was observed for type I (n = 65), in which values did not differ significantly from contralateral sides. Reduced phase shift values were present in type II; poorest values were observed for type III. CO(2)-reactivity differed mainly between type I and the other types. A less distinct differentiation of autoregulatory impairment was found when dividing patients into groups of different degrees of stenosis. Symptomatic patients (previous TIA/stroke) were significantly less frequent in the group with type I collateral flow and had significantly lower phase shift and CO(2)-reactivity values. In conclusion, we found that dynamic CAR is substantially impaired if secondary collateral pathways are activated or if functional stenosis in the activated anterior collateral pathway is present. These hemodynamic constellations are also associated with a higher proportion of clinically symptomatic patients. Determination of dynamic CAR by transfer function analysis represents a convenient, sensitive method for detection of cerebral hemodynamic compromise in obstructive carotid disease.