Quantitative Electroencephalography for Detection of Cerebral Ischemia

Abstract

In Response: Drs. Edmonds and Singer demonstrate, as have others [1], that changes in the electroencephalogram (EEG) processed with power spectral analysis can be documented during ventricular tachyarrhythmias induced for implantable cardioverter defibrillator (ICD) testing. Likewise, as previously established [2], they suggest that changes in near-infrared measurements of cerebral oxygen saturation are observed during ICD testing. These investigators correctly appreciate the impact of anesthetic variables on the EEG and have emphasized the need for caution in the use of automated intraoperative EEG assessment. We, however, disagree with their criticism of our study design and the interpretation of our results. Our experimental protocol was designed to examine the raw and quantitative EEG for evidence of persistent cerebral ischemia that might result from multiple ventricular fibrillations during ICD testing as compared to a control population undergoing abdominal surgery without ICD testing [3]. Since no evidence of cerebral ischemia was observed in the raw EEG recorded from either the ICD or the control patients, we agree that it is improbable that cerebral ischemia existed during the sampling period. Therefore, it is particularly disturbing that the automated technique of quantitative EEG interpretation that we used (Cardiovascular Intraoperative Monitoring [CIMON]; Cadwell Laboratories, Kennewick, WA) found frequent evidence of ischemic change. In the absence of an independent measure of cerebral perfusion, however, it is at least possible that the CIMON technique is more sensitive to cerebral ischemia than traditional visual interpretation of the raw EEG. We consider this extremely unlikely and, in contrast to Drs. Edmonds and Singer, are also troubled by this method's lack of specificity. Fundamental to the clinical utility of intraoperative EEG assessment is the reliable distinction between cerebral ischemia and other confounding variables encountered in the operating room. It is of interest that Dr. Edmonds, in his own cited article [4], failed to mention the significant influence that an anesthetic technique that used bolus dosing of opioids might have had on the frequency content of EEG interpreted using the CIMON technique, especially since "therapeutic interventions" during cardiopulmonary bypass were made on the basis of this automated EEG assessment. With well controlled and replicated studies, automated techniques of quantitative EEG assessment might one day be used as a basis for appropriate intraoperative interventions. Until such time, we again support the position of the American Electroencephalographic Society that "the clinical application of quantitative EEG analysis is considered to be limited and adjunctive" [5]. David C. Adams, MD Eric J. Heyer, MD, PhD Ronald G. Emerson, MD Departments of Anesthesiology and Neurology College of Physicians and Surgeons of Columbia University New York, NY 10032