Hypothermia and the approximate entropy of the electroencepbalogram

Background: The electroencephalogram is commonly used to monitor the brain during hypothermic cardiopulmonary bypass and circulatory arrest. No quantitative relationship between the electroencephalogram and temperature has been elucidated, even though the qualitative changes are well known. This study was undertaken to define a dose-response relationship for hypothermia. and the approximate entropy of the electroencephalogram. Methods: The electroencephalogram was recorded during cooling and rewarming in 14 patients undergoing hypothermic cardiopulmonary bypass and circulatory arrest. Data were digitized at 128 Hz, and approximate entropy was calculated from 8-s intervals. The dose-response relationship was derived using sigmoidal curve-fitting techniques, and statistical analysis was performed using analysis of variance techniques. Results: The approximate entropy of the electroencephalogram changed in a sigmoidal fashion during cooling and rewarming. The midpoint of the curve averaged 24.7degreesC during cooling and 28degreesC (not significant) during rewarming. The temperature corresponding to 5% entropy (T0.05) was 18.7degreesC. The temperature corresponding to 95% entropy (T-0.95) was 31.3degreesC during cooling and 38.2degreesC during rewarming (P < 0.02). Conclusions. Approximate entropy is a suitable analysis technique to quantify the electroencephalographic changes that occur with cooling and rewarming. It demonstrates a delay in recovery that is of the same magnitude as that seen with conventional interpretation of the analog electroencephalogram and extends these observations over a greater range of temperatures.