Arteriovenous malformations (AVMs) are vascular lesions characterized by direct connections between feeding arteries and draining veins without an intervening capillary network. Two hypotheses, normal perfusion pressure breakthrough (NPPB) and occlusive hyperemia, prevail in the literature regarding the occasional development of hemorrhage and edema following AVM resection. The NPPB hypothesis was introduced in 1978. Since the occlusive hyperemia hypothesis was first postulated in 1993, however, a debate has persisted within the cerebrovascular community concerning which hypothesis better explains the complications of edema and hemorrhage seen after AVM resection. Recent advances in cerebrovascular imaging and hemodynamic analysis have allowed a better evaluation of intracerebral changes following AVM resection. It is likely that these 2 hypotheses are not mutually exclusive and perhaps exist in a spectrum of hemodynamic alteration following AVM resection.
Cerebral vascular malformations occur in 0.1% to 4.0% of the general population. Arteriovenous malformations (AVMs), vascular lesions characterized by direct connections between feeding arteries and draining veins without an intervening capillary network, are some of the most dangerous congenital vascular malformations, and occur in approximately 15 of every 100,000 adults. Patients with AVMs tend to present symptomatically between the ages of 10 and 40, most commonly as a result of intracranial hemorrhage. Complete obliteration of the AVM nidus, which can be accomplished by microsurgical resection, endovascular embolization, or radiosurgery, alone or in combination, eliminates the risk of hemorrhage from these lesions. An infrequent but potentially devastating consequence following excision or occlusion of an intracranial AVM involves hemorrhage into the surrounding parenchyma or edema of the surrounding parenchyma. Two hypotheses, normal perfusion pressure breakthrough (NPPB) and occlusive hyperemia, prevail in the literature regarding postoperative hemorrhage following AVM resection. Since the occlusive hyperemia hypothesis was first postulated in 1993, a debate has persisted within the cerebrovascular community. This article discusses recent advances in cerebrovascular imaging and analysis that have allowed a more complete evaluation of intracerebral changes following AVM resection.
Cerebral perfusion pressure (CPP) is the net pressure gradient causing blood flow to the brain, and it is defined as the mean arterial pressure (MAP) minus the intracranial pressure (ICP). Cerebral blood flow (CBF) is dependent on both perfusion pressure and cerebrovascular resistance (CVR), such that:
In a single vessel level, cerebral blood flow can be summarized according to Hagen-Poiseuille’s formula ( Fig. 1 ). The magnitude of blood flow is proportional to the vessel resistance, and the blood velocity is directly proportional to the pressure difference along the blood vessel. The resistance in a vessel is inversely related to the fourth power of the vessel diameter. Thus even marginal changes of the vessel’s radius result in significant alterations of resistance. The ability of intracranial vessels to alter their radius in response to alterations in CPP is termed cerebrovascular autoregulation and is an essential mechanism by which the brain maintains CBF in the face of changes in systemic blood pressure or ICP.
Basic AVM physiology
A basic knowledge of flow and pressure alterations induced by AVMs is important in understanding the mechanisms dictating postoperative complications following AVM resection. Studies regarding flow in and around AVMs, however, are controversial. While some believe that AVMs lead to impaired autoregulation, others contend that loss of autoregulation may be the root cause of AVM formation. For instance, flow regulation can be impaired when perfusion pressures are above or below the limits of autoregulation. Such is the case with venous hypertension, which is often seen in AVMs. The idea of autoregulatory dysfunction has become common thinking in the physiology of AVMs, yet studies have consistently demonstrated preserved CO 2 responsiveness both before and after AVM resection. An alternative hypothesis, adaptive autoregulatory displacement, contends that vascular territories adjacent to AVMs shift the autoregulatory curve to the left, thus placing the lower pressure limit at a level lower than that postulated for the normal brain.
Brain AVMs are characterized by low resistance within the nidus secondary to the lack of an interposing capillary bed, thus depriving surrounding parenchyma of blood flow. Vascular steal is a phenomenon often associated with AVMs. Single-photon emission computed tomography (CT) studies have demonstrated decreased flow in areas surrounding the malformation. Local CBF measured on Xenon-CT has also demonstrated impairment in areas surrounding the AVM nidus with improvement following AVM resection. Using magnetic resonance (MR) perfusion imaging, Guo and colleagues demonstrated perinidal perfusion abnormalities within AVMs that gradually reversed following radiosurgery. Few individuals appear to suffer ischemic consequences as a result of steal, which may be explained by adaptive autoregulatory displacement.
Postresection Hemorrhage and Edema
The pioneering hypothesis attempting to explain the phenomenon of postoperative hemorrhage and edema was offered by Spetzler and colleagues in 1978. The hypothesis, termed normal perfusion pressure breakthrough (NPPB), suggests that the parenchyma surrounding a high-flow AVM is chronically hypoperfused; as a result, it has impaired autoregulation, rendering it vulnerable to the normal perfusion pressure likely to be seen following AVM resection. Thus, following removal of an AVM, the local capillary beds and arterioles in the remaining normal parenchyma experience increased perfusion but lack the ability to vasoconstrict and autoregulate. This could lead, in some cases, to hyperemia, compromise of the capillary beds, and resultant edema and/or hemorrhage.
For more than a decade NPPB remained the primary explanation for postoperative hemorrhage and edema following AVM resection. An alternative explanation termed occlusive hyperemia was offered by Al-Rodhan and colleagues in 1993 based on their retrospective review of 295 operative AVM cases and mounting evidence against the NPPB hypothesis. Multiple studies before 1993 had demonstrated decreased perinidal CBF that normalized following excision, suggesting that there was perhaps some mechanism maintaining CBF within the normal range despite increased perfusion pressures. The ability to autoregulate was tested with CO 2 reactivity by multiple groups that almost universally demonstrated restoration of normal reactivity after excision. The authors concluded that hemorrhage and edema associated with resection of high-flow AVMs was the result of two interrelated mechanisms. First, stagnant arterial flow in former AVM feeders worsens the existing hypoperfusion and ischemia. Postoperative angiograms have frequently demonstrated reduced or stagnant arterial flow in former AVM-feeding vessels, often leading to retrograde thrombosis. Al-Rodhan attributed this stagnation to increased resistance to flow, endothelial abnormalities, and a reflex vasoconstriction compensating for normal or increased perfusion pressures.
Second, obstruction of the venous outflow of surrounding parenchyma leads to passive hyperemia and further arterial stagnation, the end result of which is cerebral edema and possible hemorrhage. Al-Rodhan cites prior studies demonstrating abnormalities of the venous system in AVMs including stenosis, agenesis, or occlusion of major venous sinuses.
Barnett and colleagues observed rapid reduction in elevated draining vein pressure following excision, changes that likely predispose to increased incidence of thrombotic occlusion.
The Current Debate
The basic assumption of NPPB is that a dramatic increase in perinidal pressure reroutes flow toward territories incapable of vasoconstriction in response to the increase in CBF, and several earlier studies support this hypothesis. For example, Muraszko and colleagues tested resected AVM samples with vasoactive substances in vitro; 4 out 24 AVMs were considered nonreactive and lacked spontaneous activity. They found that patients with nonreactive vessels in vitro developed more frequent postoperative edema and hemorrhage. Moreover, using orthogonal polarization spectral (OPS) imaging in 2 patients, Pennings and colleagues assessed arteriolar pulsatility before surgical resection of an AVM. While visualizing microvessel flow, they observed a significant decrease in arteriolar pulsatility immediately after AVM exclusion. Furthermore, they were also able to show a drastic increase in microvascular flow in the perinidal brain tissue following surgery. In 1997, Chyatte used transcranial Doppler ultrasound in conjunction with acetazolamide and CO 2 to challenge vessel reactivity in patients before AVM surgery. While vasomotor paralysis was observed in only 2 of the 35 patients, those patients went on to develop postoperative edema/hemorrhage following surgical resection. De Salles and colleagues used angiographic and transcranial doppler ultrasound imaging and also demonstrated impaired vasoreactivity in response to hyperventilation in AVM feeding vessels.
Though the previously mentioned studies using various methodologies have offered evidence in support of NPPB, the hypothesis has been the subject of significant controversy. More recent investigations have contradicted many aspects of the NPPB hypothesis, casting doubt on the link between impaired autoregulation and postoperative complications. As discussed previously, multiple studies demonstrated intact CO 2 vasoreactivity following AVM resection. Young and colleagues demonstrated improved perfusion in the ipsilateral hemisphere following resection, but no increase in CBF in response to increasing MAP, suggesting intact autoregulation. They demonstrated this phenomenon both in cases without postoperative complications and in those with presumed NPPB. The culmination of these findings led Young and colleagues to postulate adaptive autoregulation as a possible explanation for the hemodynamic alterations seen in AVMs. A condition of the NPPB hypothesis is that complications occur adjacent to the malformation, as these are the areas placed under chronic stress from the AVM. Barnett and colleagues demonstrated that the worst vascular steal effect actually occurs 2 to 4 cm distal to the malformation, and Young and colleagues demonstrated global increases in CBF following resection, suggesting that focal mechanisms may not predominate.
Recent studies have also shed doubt on the mechanisms implicated in the occlusive hyperemia hypothesis. The occlusive hyperemia hypothesis suggests that arterial stagnation leads to hypoperfusion and ischemia in surrounding brain tissue. Meyer and colleagues first noted that arterial stagnation is a common observation following AVM resection. They also noted that stagnating flow is often seen in former feeding arteries, but not their smaller branches. Additionally, they noted that slow transit within the vessels likely reflects a reduction in flow velocity rather than a linear reduction in blood flow. Meyer and colleagues went on to show that postoperative brain tissue oxygenation levels are highest in patients with excessive angiographically confirmed stagnation of flow. Asgari and colleagues confirmed these significant elevations in oxygen saturation in patients with postoperative hyperemic complications, and they felt this finding was significant enough to dispute the possibility of a venous mechanism for postoperative hemorrhagic complications.
As an increasing number of AVMs are being treated with radiosurgery, reports are slowly beginning to come out of similar hemorrhagic complications after treatment. Pollock described 2 patients with abrupt neurologic deterioration within months of AVM radiosurgery, and in both instances, there was radiographic evidence of venous outflow occlusion. Chapman and colleagues followed this with a report of 2 patients who suffered edema and hemorrhagic complications after radiosurgery and demonstrated evidence of venous occlusion. Chapman concluded that a venous occlusive mechanism was likely causative in all 4 cases. Celix and colleagues added further to the evidence regarding postradiosurgery venous occlusion in a 57-year-old man who experienced hemorrhage within 9 days of treatment and had radiographic evidence of a thrombus in the primary draining vein.