Catarina S. Nunes

Modelling and multivariable control in anaesthesia using neural-fuzzy paradigms Part II. Closed-loop control of simultaneous administration of propofol and remifentanil

OBJECTIVE Part II of this research study is concerned with the development of a closed-loop simulation linking the patient model as well as the fuzzy relational classifier already introduced in Part I with a control algorithm. The overall architecture is in fact a system advisor, which provides information to the anaesthetist about the adequate infusion-rates of propofol and remifentanil simultaneously. METHODS AND MATERIAL The developed fuzzy multivariable controller includes three rule-bases and takes into account the synergetic interactions between the above drugs and uses such knowledge to achieve rapidly the desired depth of anaesthesia (DOA) level. RESULTS The result of the study is a closed-loop control scheme, which adjusts efficiently the infusion-rates of two drugs in response to DOA changes. This controller can either be used in an advisory mode or closed-loop feedback mode in the operating theatre during surgery. CONCLUSION It is hoped that this control scheme coupled with the patient model presented in Part I of this study will be used routinely in the operating theatre in the very near future.

The effect of a remifentanil bolus on the bispectral index of the EEG (BIS) in anaesthetized patients independently from intubation and surgical stimuli.

Background and objective: Remifentanil boluses are used in different clinical situations and the effects on bispectral index monitoring are unclear. We analysed the effect of a remifentanil bolus on the bispectral index of the electroencephalogram (bispectral index) under total intravenous anaesthesia with propofol and remifentanil. Methods: ASA I–III patients were included in this study. All patients received a 2 μg k g−1 remifentanil bolus in a period free from stimuli. Bispectral index and haemodynamic data were collected from an A‐2000XP bispectral index monitor (every second) and an AS/3 Datex monitor (every 5 s). Bispectral index data were analysed using the area under the curve. Mean arterial pressure and heart rate were averaged at each 30‐s period and analysed using analysis of variance. Results: A total of 240 bispectral index values were obtained per patient. The area under the curve between 90 and 120 s after the bolus was significantly lower than the basal area under the curve (average of all areas before the bolus, P < 0.05). Mean arterial pressure and heart rate were significantly reduced from 96.4 ± 19.9 mmHg at the time of the bolus to 74.2 ± 16.6 mmHg 120 s after, and from 70 ± 16.4 bpm at the time of the bolus to 61 ± 13.6 bpm after (P < 0.001), respectively. Conclusions: There was a significant reduction in the areas under the curve between 90–120 s following the bolus. Heart rate and blood pressure also showed significant reductions. Thus, remifentanil bolus given under total intravenous anaesthesia with propofol and remifentanil decreases bispectral index, an effect independent of intubation and surgical stimuli.

Clinical variables related to propofol effect-site concentrations at recovery of consciousness after neurosurgical procedures.

Target controlled infusion (TCI) systems and computer data acquisition software are increasingly used in anesthesia. It was hypothesized that the use of such systems might allow retrieval of information useful to anticipate the effect-site concentrations of propofol at which patients would recover from anesthesia. The goal of the study was to identify variables related to propofol effect-site concentrations at recovery of consciousness (ROC). Sixteen patients with a Glasgow of 15, ASA 1 or 2, subjected to neurosurgical procedures, received TIVA with TCI propofol and remifentanil. Data were collected every 5 seconds from Datex AS3 and Aspect A200XP (BIS). Effect-site TCI was used for propofol (initial effect target 5.0 μg/ml) and for remifentanil (initial plasma target 2.5 ng/ml). All clinical events were noted. Variables possibly related to propofol effect-site concentration at ROC were selected. Data are expressed as mean ± SD. Effect-site propofol concentration at ROC was 1.3 ± 0.5 μg/ml. A positive correlation was found between propofol effect-site concentration at ROC and: age (49.3 ± 17 years) (P = 0.003); mean remifentanil dose during surgery (0.11 ± 0.05 μg/kg/min) (P = 0.003); mean propofol dose during surgery (0.12 ± 0.03 mg/kg/min) (P = 0.046); and remifentanil effect-site concentration at ROC (2.85 ± 2.06 ng/ml) (P = 0.002). Propofol effect-site concentrations were not correlated with: weight, height, LBM, duration of anesthesia, minimum BIS at induction (30.4 ± 6.8), time till minimum BIS (4.7 ± 2.2 min), mean and median BIS during surgery (38.2 ± 4.5 and 37.8 ± 5.3). BIS-related variables were not useful as ROC predictors. Only drug variables and age correlated with propofol effect-site concentrations at ROC.

Modelling the dynamics of depth of anaesthesia: Cerebral state index in dogs

The goal of this study was to obtain models that described the relation between the anaesthetic drug infusions (propofol) and an electroencephalogram (EEG) derived index (Cerebral State Index -CSI) during general anaesthesia in dogs. The first phase integrated the adaptation of hardware for EEG acquisition and exploration for the best electrodes position in dogs skull. The clinical protocol implementation and data collection were the next steps followed by CSI modeling. CSI showed adequate response to changes in drug infusion, reflecting the changes of depth of anaesthesia in dogs. The models obtained adjusted well to the original CSI data and also predicted the CSI trend during surgery. Using this monitor in current practice might improve quality in the anaesthesia procedure providing a useful tool to administer a correct sedation.

Comparison of Neural Networks, Fuzzy and Stochastic Prediction Models for return of consciousness after general anesthesia

This paper presents three modeling techniques to predict return of consciousness (ROC) after general anesthesia, considering the effect concentration of the anesthetic drug at awakening. First, several clinical variables were statistically analysed to determine their correlation with the awakening concentration. The anesthetic and the analgesic mean dose during surgery, and the age of the patient, proved to have significantly high correlation coefficients. Variables like the mean bispectral index value during surgery, duration of surgery did not present a statistical relation with ROC. Stochastic regression models were built using the variables with higher correlation. Secondly, fuzzy models were built using an Adaptive Network-Based Fuzzy Inference System (ANFIS) also relating different sets of variables. Thirdly, radial basis function (RBF) neural networks were trained relating different sets of clinical values with the anesthetic drug effect concentration at awakening. Clinical data was used to train and test the models. The stochastic models and the fuzzy models proved to have good prediction properties. The RBF network models were more biased towards the training set. The best balanced performance was achieved with the fuzzy models.

Predictive adaptive control of the bispectral index of the EEG (BIS): using the intravenous anaesthetic drug propofol

The problem of controlling the level of unconsciousness measured by the Bispectral Index of the EEG (BIS) of patients under anaesthesia, is considered. It is assumed that the manipulated variable is the infusion rate of the hypnotic drug propofol, while the drug remifentanil is also administered for analgesia. Since these two drugs interact, the administration rate of remifentanil is considered as an accessible disturbance. In order to tackle the high uncertain present on the system, the predictive adaptive controller MUSMAR is used. The performance of the controller is illustrated by means of simulation with 45 patient individual adjusted models, which incorporate the effect of the drugs interaction on BIS. This controller structure proved to be robust to the remifentanil disturbance, different reference values and noise. A reduction of propofol consumption was also observed when comparing to the real clinical dose used for a similar BIS trend.

Nonlinear Modeling of Cerebral State Index in Dogs

The Cerebral State Index (CSI) is an electroencephalogram derived signal representing the depression of central nervous activity produced by anesthetic drugs. In this study, a nonlinear model was developed to describe the CSI tendency during general anesthesia in dogs, by evaluating the effect of the anesthetic drug propofol. The model was based on a compartmental and Hill Equation structure with individually identified parameters. The clinical data of 14 dog surgeries were collected and used for modeling and testing. The model presented good results, following the CSI trend. A model for drug-effect for veterinarian anesthesia is an important step when developing advisory, educational and control systems. The overall aim is to improve animal safety and comfort.

Predictive adaptive control of unconsciousness - exploiting remifentanil as an accessible disturbance

The problem of controlling the level of unconsciousness measured by the BIS index of patients under anesthesia, is considered. It is assumed that the manipulated variable is the administration rate of propofol, while remifentanil is also administered for analgesia. Since these two drugs interact, the administration rate of remifentanil is considered as an accessible disturbance. A predictive adaptive controller structure that explores this fact is proposed and illustrated by means of simulation.

Insulin Exhibits an Antiproliferative and Hypertrophic Effect in First Trimester Human Extravillous Trophoblasts.

Our aim was to investigate the effect of high levels of glucose, insulin, leptin, and tumor necrosis factor alpha, biomarkers of diabetes in pregnancy, in the process of placentation, using as a cell model a first trimester extravillous human trophoblast cell line (HTR8/SVneo cells). Exposure of HTR8/SVneo cells for 24 hours to either glucose (20 mmol/L) or leptin (25-100 ng/mL) did not cause significant changes in cell proliferation and viability. Tumor necrosis factor alpha (24 hours; 10-100 ng/L) caused a small decrease (10%) in cell proliferation and an increase (9%) in cell viability; however, both effects disappeared when exposure time was increased. Insulin (24 hours; 1-10 nmol/L) caused a concentration- and time-dependent decrease (10%-20%) in cell proliferation; the effect of insulin (10 nmol/L) was more pronounced after a 48 hours exposure (35%). In contrast, exposure to insulin (10 nmol/L; 48 hours) showed no significant effect on cell viability, apoptosis, and migration capacity. Insulin appears to cause hypertrophy of HTR8/SVneo cells as it reduces the cell mitotic index while increasing the culture protein content. The antiproliferative effect of insulin seems to involve activation of mammalian target of rapamycin, phosphoinositide 3-kinase, and p38 mitogen-activated protein kinase. Finally, simvastatin and the polyphenol quercetin potentiated the antiproliferative effect of insulin; on the contrary, the polyphenol resveratrol, the polyunsaturated fatty acids eicosapentaenoic and docosahexaenoic acids, and folic acid were not able to change it. In conclusion, we show that insulin has an antiproliferative and hypertrophic effect on a first trimester extravillous human trophoblast cell line. So insulin might affect the process of placentation.Our aim was to investigate the effect of high levels of glucose, insulin, leptin, and tumor necrosis factor alpha, biomarkers of diabetes in pregnancy, in the process of placentation, using as a cell model a first trimester extravillous human trophoblast cell line (HTR8/SVneo cells). Exposure of HTR8/SVneo cells for 24 hours to either glucose (20 mmol/L) or leptin (25-100 ng/mL) did not cause significant changes in cell proliferation and viability. Tumor necrosis factor alpha (24 hours; 10-100 ng/L) caused a small decrease (10%) in cell proliferation and an increase (9%) in cell viability; however, both effects disappeared when exposure time was increased. Insulin (24 hours; 1-10 nmol/L) caused a concentration- and time-dependent decrease (10%-20%) in cell proliferation; the effect of insulin (10 nmol/L) was more pronounced after a 48 hours exposure (35%). In contrast, exposure to insulin (10 nmol/L; 48 hours) showed no significant effect on cell viability, apoptosis, and migration capacity. Insulin appears to cause hypertrophy of HTR8/SVneo cells as it reduces the cell mitotic index while increasing the culture protein content. The anti-proliferative effect of insulin seems to involve activation of mammalian target of rapamycin, phosphoinositide 3-kinase, and p38 mitogen-activated protein kinase. Finally, simvastatin and the polyphenol quercetin potentiated the antiproliferative effect of insulin; on the contrary, the polyphenol resveratrol, the polyunsaturated fatty acids eicosapentaenoic and docosahexaenoic acids, and folic acid were not able to change it. In conclusion, we show that insulin has an antiproliferative and hypertrophic effect on a first trimester extravillous human trophoblast cell line. So insulin might affect the process of placentation.

Modeling State Entropy of the EEG and Auditory Evoked Potentials - Hypnotic and Analgesic Interactions

Because of the complexity of raw electroencephalogram (EEG), for the anesthesiologist it is very difficult to evaluate the patients hypnosis state. Because of this, several depth of anesthesia monitors have been developed, and are in current use at the operating room (OR). These monitors convert the information supplied by the EEG or derived signals into a simple, easy to understand index. Nowadays, general anesthesia is controlled only by the clinician, which decides what is the best drug combination for the patient, regarding all information given by monitors and sensors in the OR. In this work, we collected data from two study groups with auditory evoked potentials (AEP) monitoring, and Entropy (SE) monitoring. A model was fitted to the signals and the Hill equation parameters adjusted, in both study groups. The objective was to predict hypnosis indices, regarding only the drugs administered to a patient, and capture the initial individual patient characteristics that might influence the drugs interaction in the human body. Hypnotic and analgesic drugs interact in different ways throughout the anaesthesia stages. The models obtained captured the different dynamic interaction of drugs, during the induction and maintenance phases, demonstrating that the model must have incorporated all this information in order to perform satisfactorily. Other information like haemodynamic variables might be included in the search for the optimum model.