Correlation of Sedline-generated variables and clinical signs with anaesthetic depth in experimental pigs receiving propofol

Abstract

Most of currently available electroencephalographic (EEG)-based tools to assess depth of anaesthesia have not been studied or have been judged unreliable in pigs. Our primary aim was to investigate the dose-effect relationship between increasing propofol dose and variables generated by the EEG-based depth of anaesthesia monitor Sedline in pigs. A secondary aim was to compare the anaesthetic doses with clinical outcomes commonly used to assess depth of anaesthesia in this species. Sixteen juvenile pigs were included. Propofol infusion was administered at 10 mg kg-1 h-1, increased by 10 mg kg-1 h-1 every 15 minutes, and stopped when an EEG Suppression ratio >80% was reached. Patient state index, suppression ratio, left and right spectral edge frequency 95%, and outcomes from commonly used clinical methods to assess depth of anaesthesia in pigs were recorded. The best pharmacodynamic model was assessed for Patient state index, suppression ratio, left and right spectral edge frequency 95% in response to propofol administration. The decrease of Patient state index best fitted to an inhibitory double-sigmoid model (including a plateau phase). The increase of suppression ratio fitted a typical sigmoid Emax model. No relevant relationship could be identified between spectral edge frequency 95% values and propofol administration. A large variability in clinical outcomes was observed among pigs, such that they did not provide a reliable evaluation of propofol dose. The relationship between propofol dose and Patient state index/suppression ratio described in the present study can be used for prediction in future investigations. The evaluation of depth of anaesthesia based on common clinical outcomes was not reliable.

Fig 1. Sedline-generated and clinical variables over time.

Median (interquartile range) for patient state index (black dots), suppression ratio (dark grey squares), and clinical depth of anaesthesia score (light grey triangles) during increasing propofol infusion in nine pigs. The individual durations of propofol administration were divided in 30 equal time intervals (TI) for each pig (TI₁ to TI₃₀). T₀ = baseline values. Box plots (median and 25-75th percentiles) represent the time intervals at which the different clinical outcome scores changed (in brackets).

Fig 2. Pharmacodynamic modeling for patient state index (PSI).

A. Observed data (circles) and prediction from the final population model (line) for PSI values during increasing propofol infusion rate. The time axis shows the normalized time intervals. B. Observed PSI data against individual-predicted values. C. Prediction-corrected visual predictive check (for 1000 simulations) of the final model for PSI; Blue dots = observations; red lines = 5^th, 50^th, and 95^th percentiles of the observed values; red/blue shaded area = 95% CI of the 5^th, 50^th and 95^th percentiles of the final prediction model.

Fig 3. Pharmacodynamic modeling of suppression ratio (SR).

A. Observed data (circles) and prediction from the final population model (line) for SR during increasing propofol infusion rate. The time axis shows the normalized time intervals. B. Observed SR data against individual-predicted values. C. Prediction-corrected visual predictive check (for 1000 simulations) of the final model for SR; Blue dots = observations; red lines = 5^th, 50^th, and 95^th percentiles of the observed values; red/blue shaded area = 95% CI of the 5^th, 50^th and 95^th percentiles of the final prediction model.

Fig 4. Spectral edge frequency 95% (median and interquartile ranges) -left (white dots) and -right (black dots) at each time interval (TI).