Migrating from target-controlled infusion to closed-loop control in general anaesthesia

https://doi.org/10.1016/j.cmpb.2003.11.005Get rights and content

Abstract

The target-controlled infusion (TCI) technique has been successfully and commercially used in clinical general anaesthesia with the intravenous anaesthetic agent propofol. The technique is based on a population pharmacokinetic model and is an open-loop control system. Closed-loop control requires a reliable and consistent signal for feedback utilisation. With all anaesthetic agents the somatosensory evoked potentials (SEP) have been shown to give increased latency as anaesthetic depth is increased. Using infusion rate and SEP response data from rats anaesthetised with propofol a mathematical model was derived to describe the anaesthetic process. This model was used as a design reference to develop a proportional integral (PI) closed-loop control system using SEP as the feedback measure. A serials of 10 trials were conducted to investigate the difference between continuous bolus injection and infusion, all under closed-loop control. The trials showed that the use of SEPs in closed-loop control of anaesthesia is feasible.

Introduction

Using pharmacokinetic models, physicians monitor drug effects on patients and suggest dosage regimes which will keep the concentration of drug within the desired range at its site of action and hence achieve the desired therapeutic effect. In anaesthesia, this method is known as target-controlled infusion (TCI) [1]. TCI systems aim to maintain the central compartment of a three-compartment pharmacokinetic model at determined target plasma concentrations. The time course of concentration is predicted by pharmokinetic population models. Furthermore, the pharmacodynamic effects of anaesthetic agents on the nervous system are sigmoidally related to the plasma concentrations, and are commonly represented by the Hill equation [2]. However, when using TCI, this pharmacodynamic behaviour is assessed by the anaesthetist from the patient’s overall external physiological state and then used to set the desired target concentrations. This scenario has been clinically and commercially used in the operating theatre for maintaining general anaesthesia with propofol [3], [4]. The success of the TCI system implies: (i) as yet, there is no unified and easy means for automatic assessment of anaesthetic state and, therefore, skilled human intervention is necessary; (ii) large tolerance in physiological therapy is acceptable from use of population modelling and (iii) the three-compartment model gives a reliable prediction of drug concentration at its site of action.

TCI systems provide model-based open-loop control for drug delivery and may, therefore, be inaccurate in drug administration because of model mismatch between the population model and an individual patient’s behaviour. The model discrepancy can be overcome by the use of closed-loop control techniques if there exists a reliable feedback signal. In closed-loop systems for drug delivery, the effect of a drug is measured and used to adjust drug administration to achieve actual therapeutic states rather than target drug concentrations [5]. In principle, the use of feedback in closed-loop systems may offset the sources of variability in the processes between drug delivery and drug response, of which the most obvious is the inaccuracy of population pharmacokinetic models as applied to individuals in open-loop TCI systems [6].

The concept of “depth of anaesthesia” (DOA) implies a state of the central nervous system resulting from a balance between the depression caused by anaesthetic drugs and arousal caused by surgical and other stimuli [7]. However, DOA is too difficult to be defined either qualitatively or quantitatively. Instead, the concept of “anaesthetic effect” is adopted in this study to describe the extent to which an anaesthetic agent affects the nervous system and attenuates signal transmission along the nervous pathways [8], [9]. In physiological studies, it has been found that the responses recorded from the primary somatosensory cortex reflect both anaesthetic depth and its modulation in response to painful stimuli [8], [10]. Thus, somatosensory evoked potentials (SEP) give direct insight into the macro response of the somatosensory cortex to external stimuli and can be used as an effective indicator of DOA [11], [12]. With all anaesthetic agents the SEP has been shown to give increased latency as anaesthetic depth is increased [12], [13]. Thus, a real-time SEP recording and monitoring system has been developed for laboratory studies undertaken in this project [14]. Numerous studies have utilised EEG signals and audio evoked potentials (AEP) to indicate DOA. However, AEP are buried in EEG with much worse signal-to-noise ratio than SEP. Motivated by the implications behind TCI, this study has modelled the pharmacological effects of propofol for anaesthesia with a linear autoregressive moving average (ARMA) equation and, accordingly, developed a closed-loop drug delivery control system using SEP as the feedback signal. The TCI and the two-term proportional integral (PI) control algorithms were employed as the control kernels within the SEP recording and monitoring system. The PI control system aimed to regulate the anaesthetic effect of propofol via the latency change in SEP. Anaesthesia experiments on rats were conducted to evaluate the use of closed-loop systems with the intravenous anaesthetic propofol in the absence of a pharmacokinetic model to determine if such a system might, in future, provide improved DOA regulation, particularly in terms of accuracy and stability.

Section snippets

Target-controlled infusion

The pharmacokinetics of propofol anaesthesia is best described by a three-compartment model [15] as shown in Fig. 1. Usually, the central compartment describes the blood vessels and organs abundant with vessels, such as the liver and kidneys. One of the peripheral compartment represents well-perfused tissues such as muscle and brain. The other peripheral compartment includes the remaining peripheral tissues, such as skin and bone [16]. Although drug elimination may be from either the central

Monitoring and control

Major challenges were encountered in the development of a reliable data-logging and control suite of programmes. In addition to acquisition of signals from small electrodes, the signal conditioning required novel algorithm development for latency quantification, while the controller algorithms had to be interfaced to specially-adapted, initially controlled syringe drives for minute drug adjustments. Furthermore, the animal preparations were intensive, highly-skilled and time-critical. Thus, the

Experimental results

TCI has been well recognised because of its success in clinical general anaesthesia. Hence, the pharmacokinetic model of propofol obtained from experimental data in humans and in clinical TCI was applied to rats with inter-species scaling of the weight difference [26]. Two experiments of TCI control for propofol anaesthesia in rats were conducted to validate anaesthetic effects on the latency change in SEPs. The TCI module of the experimental apparatus shown in Fig. 5 was employed. When a TCI

Discussion and conclusions

As stated in Section 1, while AEP have been studied for feedback control of DOA, the AEP is buried in the spontaneous EEG and has a much worse signal-to-noise ratio than SEP. EEG has obvious potential for measuring DOA, but because of sigmoidal saturation effects being dangerous at deep anaesthesia levels of control, inter-patient variability, intra-patient time variation, instabilities and unclear anatomical location, it is uncertain for safe and reliable indications. Thus, SEP has been the

Acknowledgements

C.H. Ting was supported by a postgraduate scholarship from the University of Sheffield.

References (31)

  • M. Morgan

    Intravenous anaesthesia: a step forward

    Anaesthesia

    (1998)
  • J. Vuyk

    TCI: supplementation and drug interactions

    Anaesthesia

    (1998)
  • A.J. Asbury

    Feedback control in anaesthesia

    Int. J. Clin. Monit. Comput.

    (1997)
  • P.S.A. Glass, S.L. Shafer, J.R. Jacobs, J.G. Reves, Intravenous drug delivery systems, in: R.D. Miller (Ed.),...
  • A. Angel

    Adventures in anaesthesia

    Exp. Physiol.

    (1991)
  • Cited by (29)

    • Injectable anesthetics

      2023, Anesthesia and Analgesia in Laboratory Animals
    • Intelligent decision support for lung ventilation

      2020, Control Applications for Biomedical Engineering Systems
    • Optimized PID control of depth of hypnosis in anesthesia

      2017, Computer Methods and Programs in Biomedicine
      Citation Excerpt :

      It was one of the first feedback suitable signals available for closed loop regulation of DoH and became very popular in clinical onset. Other devices and respective signals are available [5], e.g. the auditory evoked potential (AEP) index [6], the wavelet-based index (WAVCNS)) [7,8], the M-entropy [9] and the somatosensory evoked potentials (SEP) [10]. There are three phases of anesthesia during a standard surgery procedure.

    • Pharmacology of Injectable Anesthetics, Sedatives, and Tranquilizers

      2008, Anesthesia and Analgesia in Laboratory Animals
    • Closed loops in anaesthesia

      2006, Best Practice and Research: Clinical Anaesthesiology
    View all citing articles on Scopus
    View full text