Hemodynamic consequences of cerebral vasospasm on perforating arteries: a phantom model study.

BACKGROUND AND PURPOSE

Hemodynamics of cerebral vasospasm after subarachnoid hemorrhage remain unclear, and the discrepancy between ultrasonographic or angiographic evidence of arterial narrowing and neurological ischemic deficit is still debated. Most blood flow studies have been involved with large arteries, and thus, very little is known regarding the hemodynamic behavior of small perforating vessels. Patients with symptomatic vasospasm, however, often present with neurological signs suggesting involvement of deep-sited areas of the brain supplied by perforating arteries.

METHODS

A pulsatile pump was set to provide an outflow of 350 mL/min through a 10-mm-diameter C-flex tube at a perfusion pressure of 130/80 mm Hg. The perfusion fluid used was prepared to approximate blood viscosity. Perforating arteries were simulated by a 1-mm tube connected to the parent tube at a 90 degrees angle. Cylindrical stenotic devices of decreasing diameters were then introduced into the parent tube at the level of the aperture of the secondary tube and 1.5 diameters upstream of it. Velocity profiles both proximal and distal to the stenosis in the parent tube were obtained with a newly developed ultrasonographic flowmeter that allows for high spatial resolution.

RESULTS

Increasing stenosis resulted in decreased outflow in the main tube, although it was significant only with severe stenosis. Whenever the simulated stenosis was placed upstream of the secondary tube, flow reduction was associated with a progressive change in the velocity profile, which gradually changed from laminar conditions to a jet stream limited to the center of the lumen. Further diameter reduction was responsible for the occurrence of flow separation with retrograde flow velocities in the periphery of the lumen. In the secondary tube, flow reduction was much more pronounced and began at a lesser degree of stenosis. Increasing fluid viscosity and decreasing perfusion pressure enhanced flow separation and prominently affected the outflow in the secondary tube. Conversely, whenever the simulated stenosis involved the branching area of the secondary tube, there was a slightly progressive decrease in the relative flow in the main tube as the stenosis became tighter. When the stenosis equaled the diameter of the secondary tube, the relative contribution of the secondary tube increased markedly at the expense of the main tube outflow.

CONCLUSIONS

The present results show that local cerebral vasospasm induces changes in postvasospastic velocity profile affecting the shear rate and may eventually lead to flow separation. This phenomenon may, in turn, result in a venturi-like effect over the aperture of perforating arteries branching out of the postvasospastic portion of the affected parent artery. These alterations of cerebral hemodynamics may account for at least part of the vasospasm symptomatology, especially in the vertebrobasilar system, where vasospasm is commonly focal rather than diffuse. Furthermore, these changes proved to be affected significantly by manipulations of pressure and viscosity, supporting the use of hyperdynamic therapy in the management of cerebral vasospasm.