Evaluation Of Sample Entropy And Bispectral Index On Monitoring Sedation Depth During Target-controlled Infusion Of Propofol Anesthesia

Objective:Maintaining the depth of sedation in normal range is one of important factors to ensure the patients safety and surgical operation performing smoothly. Comparing with Approximate entropy (ApEn), sample entropy (SampEn) is an improved algorithm. To investigate relationship of SampEn of electroencephalogram and bispectral index (BIS) with the changes of target-controlled concentration of propofol and sedation depth in general anesthesia induction. To evaluate the feasibility and accuracy of SampEn which coulde monitor different levels of consciousness in patients. Measuring the SampEn as an index of sedation depth could provides the theoretical basis of safe, effective and individualized therapy.Methods:40 patients ASA I or II, aged from 18-65 years, male or female, BMI 20-28 kg/m2, scheduled for elective surgery with cancer undergoing general anesthesia. Anesthesia was induced with intravenous target-controlled infusion (TCI) of propofol. TCI of propofol was set initially at 0.5μg/ml, followed by increments of 0.5μg/ml, every propofol concentration level was maintained over 1 min. During stepwise increase (0.5μg/ml) of propofol, the level of sedation was assessed by Modified Observer's Assessment of Alertness/Sedation scale (MOAA/S) every 20 s until the MOAA/S score is 0. Electrocardiogram (ECG), noninvasive arterial blood pressure (NIBP) and pulse oxygen saturation (SpO2) were monitored as regular. EEG-derived indices (BIS A-1050, Aspect Medical System, USA), the special electrode and the authors-programmed software of electroencephalogram (EEG) acquisition were applied to capture the raw EEG signals and the value of BIS (sampled at 128 samples/s by computer). The datas were recorded and transferred into a computer hard disk for off-line analysis (256 samples of analysis sequence), and the SampEn of EEG was calculated off-line from recorded raw EEG data. Pharmacokinetic-Pharmacodynamic (PK/PD) model was described the relationship among the propofol concentration, SampEn of EEG and different levels of consciousness. The MOAA/S as an index was used to evaluate prediction ability of SampEn to the sedation depth. And compare with the bispectral index (BIS) at same state undergoing the self-controlled study. Statistical analyses were performed with SPSS 17.0. The coefficient of determination (R2) was calculated to describe the correlation of TCI of propofol dose-response with SampEn/BIS. We investigated the ability of the indicators (SampEn/BIS) to differentiate between consciousness and unconsciousness using prediction probability (Pk), which was calculated with PKMACRO provided by Smith and colleagues. The performance of SampEn and BIS as index of propofol target concentration and sedation levels on different MOAA/S was compared by R2 and Pk, and analyzed using paired t test. A value less than 0.05 was considered statistical significant.Results:With increasing of the target concentration of propofol, the values of SampEnere decreasing relatively as well as BIS. R2 of SampEn and BIS were 0.91±0.15 and 0.86±0.08, respectively(P>0.05). MOAA/S score decreased from 5 to 0 was associated with the increase of sedation depth. Compared with the previous score on MOAA/S, both SampEn and BIS were able to distinguish between MOAA/S levels 5 to 2 (P<0.05), whereas at deeper anesthesia, SampEn showed significance difference between MOAA/S levels 0 and 1 (P＜0.05). The PK values of MOAA/S for SampEn and BIS were 0.94±0.01 and 0.92±0.01, respectively. There were no significant differences in Pk values between SampEn and BIS(P>0.05)Conclusion:SampEn reports on similarity in time series. Comparing with ApEn, SampEn is a simpler and improved algorithm. SampEn does not count self-matches, reduces the bias, is largely independent of record length and displays high consistency. SampEn has the strong ability to anti-jamming, and is insensitive to missing datas. The SampEn of EEG has the good correlation with TCI of propofol dose-response, which was similar with the BIS performance. SampEn of EEG could predict well the sedation depth by differentiate the levels of MOAA/S accurately. In the field of the sedtion depth monitoring, application of SampEn has the good feasibility and high clinical value.