Hemoglobin Concentration Affects Electroencephalogram During Cardiopulmonary Bypass: An Indication for...

Hemodilution during cardiopulmonary bypass (CPB) is widely used to decrease transfusion and improve microcirculation but has drawbacks, such as diminished hemoglobin levels. Among others, reduced brain oxygenation accounts for neurological adverse outcomes after CPB. The aim of the present study was to ascertain if and how continuous electroencephalogram (EEG) during CPB is affected by hematocrit level and what should be the minimum value to avoid significant frequency band shifts on the EEG. A comparative study design was used with 16 subjects undergoing elective mitral valve repair/replacement. EEG was continuously recorded during the surgical procedure (from anesthesia induction to 20 min after CPB end). Data were marked at relevant time points (T0: before CPB start; T1: after 30 min from CPB beginning; T2: at CPB end), and the following 2 min EEG analyzed with a fast Fourier transform to obtain relative power for delta, theta, alpha, and beta bands. A general linear model for repeated measure was used to study interactions of time (T0, T1, and T2, EEG frequency band, and topographical distribution. The relative powers for each electrode were calculated and represented using topographic maps. Power spectrum differences between time points (T2-T1; T2-T0; T1-T0) were calculated for each electrode, and differences >10%, considered indicative of neuronal sufferance, were included in further analysis. Cutoff hemoglobin values that maximize the proportion of correctly classified EEG band shifts were obtained by previous definition were obtained. At T2, diffuse EEG slowing in delta and theta bands was detected; a minor slowing over anterior regions was evident at T1 for the theta band. Decrements in EEG power greater than 10% were detected only for the delta band at T2. Hemoglobin concentration levels at which no slowing increase was evident were 9.4 mg/dL (Ht: 28.2%) at T1 and 9.2 mg/dL (Ht: 27.6%) at T2. EEG burst-suppression pattern related to a lesser degree of slowing at T2. In conclusion, we propose hemoglobin cutoff levels that prevent EEG slowing indicative of neuronal sufferance. In addition, burst-suppression EEG patterns offer higher central nervous system protection as measured on EEG.