Comparison of cerebral blood flow by radionuclide cerebral angiography and by microspheres in cats.

Abstract

Radionuclide cerebral angiography is commonly used as an adjunct to the diagnosis of brain death. Despite its acceptance as a diagnostic tool, it is not clear whether the absence of cerebral blood flow by radionuclide cerebral angiography denotes a complete lack of cerebral blood flow. To compare cerebral blood flow estimated by radionuclide cerebral angiography with cerebral blood flow measured by the radiolabeled microsphere technique, we systematically varied cerebral perfusion pressure (mean arterial BP minus intracranial pressure) in anesthetized cats by infusing artificial cerebral spinal fluid into the lateral ventricle to increase intracranial pressure. We measured cerebral blood flow with both techniques as cerebral perfusion pressure was decreased from its baseline of 111 +/- 10 mm Hg to 20, 10, 5, 0, and less than 0 mm Hg, causing a stepwise decrease in cerebral blood flow. We found a correlation by regression analysis (r2 = .47, p less than .05) between radionuclide cerebral angiography and microsphere measurements of cerebral blood flow, when both blood flow measurements were expressed as a percentage of baseline values. However, if 20% of baseline flow was assigned as a cut-off point for critically low cerebral blood flow (based on human studies), radionuclide cerebral angiography was only 33% sensitive to detect critically reduced cerebral blood flow and had a positive predictive accuracy (of low-flow interpretation) of only 60%. Radionuclide cerebral angiography was unable to demonstrate a complete lack of cerebral blood flow, even in two instances when cerebral blood flow by microspheres was less than 0.1% of baseline. We conclude that the ability of radionuclide cerebral angiography to quantify low cerebral blood flow is poor, and that this technique may not identify severely reduced cerebral blood flow.