&NA;Three cases of large cerebral arteriovenous fistulae are presented in which surgical ablation was complicated by brain swelling from hyperperfusion breakthrough believed to be caused by acute intraoperative hypoperfusion superimposed on chronic preoperative hypoperfusion. On the basis of these cases, experimental data, and theoretical considerations, we seriously question the wisdom of using staged surgical resection of cerebral arteriovenous malformation to prevent complications related to alterations in cerebral hemodynamics. The reasons for this concern are: the repeated occurrence of acute‐on‐chronic hypoperfusion during staged resection; a lack of understanding of the time course for the correction of a disordered autoregulation; risk of hemorrhage between the initial and final resection; difficulty in assessing and substantiating flow reduction after subtotal resection; the rapidity of collateralization; the divergence of flow from large, readily accessible feeding arteries to deep penetrating vessels; and attenuation of the wall thickness in collateral vessels as a consequence of increased flow. (Neurosurgery25:429‐436, 1989)