SICE Annual Conference 2010 August 18-21, 2010, The Grand Hotel, Taipei, Taiwan ¥400 © 2010 SICE A Hypnosis and Analgesia Control System Using a Model Predictive Controller in Total Intravenous Anesthesia During Day-case Surgery Eiko Furutani, Keigo Tsuruoka, and Shogo Kusudo Department of Electrical Engineering Kyoto University Katsura, Nishikyo-ku, Kyoto, 615-8510 E-mail: furutani@kuee.kyoto-u.ac.jp Gotaro Shirakami Department of Anesthesiology Kagawa University Ikedo, Miki-cho, Kagawa 761-0793 E-mail: gshi@med.kagawa-u.ac.jp Kazuhiko Fukuda Department of Anesthesia Kyoto University Shogoinkawahara-cho, Sakyo-ku, Kyoto 606-8507 E-mail: kfukuda@kuhp.kyoto-u.ac.jp Abstract—In this paper, a hypnosis and analgesia control system using a model predictive controller during total intra- venous anesthesia for day-case surgery is developed. A new analgesia index for analgesia control is constructed from pulse rate and Entropy Difference (ED) obtained by spectral analysis of electroencephalogram (EEG). As the anesthetics propofol and remifentanil are used, and the Bispectral Index (BIS) and the proposed analgesia index are used as hypnosis and analgesia in- dices, respectively. The model of hypnosis and analgesia responses to propofol and remifentanil consists of pharmacokinetic models, pharmacodynamic models considering interaction of drugs, and dead times. The control system has an identification function of individual parameters of pharmacodynamic models and dead times, and adjusts infusion rates of propofol and remifentanil simultaneously using model predictive control after the identifi- cation. Simulation results show that it has a good performance, but the identification method of individual parameters has to be improved. Keywords—anesthesia control analgesia index model predic- tive control propofol remifentanil. I. INTRODUCTION Day-case surgery is becoming more common because of its advantages of shortening hospital stay, releasing back patients into society early, reducing economical burden of patients, and so on. It is desirable to minimize doses of anesthetics during day-case surgery for avoiding postoperative adverse reactions such as postoperative nausea and vomiting. For this purpose many researches on anesthesia control systems [1]–[8], which keep anesthesia level of patients adequately by adjusting infusion rates of anesthetic agents, have been done. Recently, propofol, an anesthetic drug, and remifentanil, an analgesic drug, are often used simultaneously to reduce the amounts of total doses by utilizing their potentiation. For such cases, hypnosis and/or analgesia control systems have also been developed [4]–[6]. However, no reliable index of analgesia has been established, and no control system considering both the potentiation of propofol and remifentanil and dead times from drug infusions to hypnosis and analgesia indices has been developed. In this paper, an analgesia index for feedback control during day-case surgery based on pulse rate and Entropy Difference (ED) [4], obtained by spectral analysis of electroencephalogram (EEG), is proposed, and a model predictive control system of hypnosis and analgesia is constructed using the Bispectral Index (BIS) [9] and the proposed analgesia index considering the potentiation and the dead times of propofol and remifentanil. The system has an identification function of individual pharmacodynamic parameters and dead times from the measurements in the first 10 minutes of anesthesia. The performance of the system is confirmed by simulation. This paper is organized as follows. An analgesia index for day-case surgery is proposed in Section II. Section III gives the model of hypnosis and analgesia responses to propofol and remifentanil considering drug potentiation and dead times. In Section IV, design of model predictive controller and an identification function of individual pharmacodynamic param- eters and dead times are explained. Section V gives simulation results of the hypnosis and analgesia system and evaluates the controller performance. II. ANALGESIA INDEX In this section, an analgesia index available during day- case surgery for adjusting infusion rate of remifentanil is studied. Pain stimulation produces physiological responses such as elevation of blood pressure, increases in heart rate and respiratory rate, vascular constriction, perspiration, and grimace. These responses can be detected by measurements such as arterial pressure, variation of heart rate, electrocardio- gram (ECG), blood flow volume, or electromyogram (EMG). Many analgesia indices using the above measurements have been proposed [3], [4], [7], [8], [10]. For example, Mathews et al. [4] proposed Entropy Difference (ED) obtained by spectral analysis of EEG from electrodes on patients’ forehead, Huiku et al. [7] proposed Surgical Stress Index (SSI) obtained from pulse rate and pulse wave, and Ushiyama et al. [10] proposed heart rate variability (HRV) obtained from ECG. The existing indices, however, have insufficient reliability or use measurements unavailable during day-case surgery. PR0001/10/0000-0223 - 223 -