The effects of intracranial arteriovenous malformations on cerebral hemodynamics

Significantly decreased perfusion pressure is common in vascular territories irrigating neuronal tissue in patients with AVMs. Without an adaptive mechanism, many of these patients would suffer from ischemic symptoms because their perfusion pressure would fall well below the lower limit of the "normal" autoregulation curve. There is evidence that an "adaptive autoregulatory displacement" occurs in these patients which maintains CBF above ischemic levels in spite of sometimes severe arterial hypotension. This "adaptive autoregulatory displacement" may explain why only a minority of patients appear to suffer from ischemic neurologic deficits attributed to "steal." Although CBF may be mildly reduced in hypotensive regions, function is maintained and there is no increase in CBV which would be expected as a result of ischemic arteriolar vasodilatation. The mechanism for the coupled reduction in CBF, CBV, and cerebral metabolic rate in AVM patients remains to be determined, but diaschisis and decreased neuronal mass may play a role. Adaptive changes in autoregulation may also explain, in part, why NPPB is only rarely encountered after AVM resection, despite the presence of significant arterial hypotension in the majority of patients. Chronic arteriolar vasodilation does not usually lead to vasomotor paralysis because most patients maintain a constant CBF in spite of increase in systemic arterial pressure. Our studies revealed bihemispheric changes in CBF after AVM removal which do not correlate with pre-resection arterial hypotension. Therefore, arterial hypotension may be a necessary but not sufficient condition for the development of "steal" and NPPB. Other nonhemodynamic mechanisms such as abnormal sensory and autonomic innervation of the cerebral vasculature may play a role in the pathogenesis of cerebral hyperemia. Finally, AVM hemodynamics appear to play an important role in the etiology of spontaneous ICH. There is accumulating evidence that lesions with severe arterial hypotension (and greatest "buffering effect") are least likely to hemorrhage. This is important in clinical practice because these lesions are usually large, high-flow AVMs, that is, those with the highest risk from treatment. Hemodynamic and anatomic characteristics of the venous drainage as well as interactions of AVM flow dynamics with the coagulation system also appear to play an important role in determining the natural history of what is probably a heterogeneous disease process with distinct prognostic risk groups.