R. Gomeni(1), I. Cognet(1) and F. Servin(2)
1) Zeneca Pharma, Cergy (France), 2)Hôpital Bichat, Paris (France)
Introduction: The ‘Diprifusor’ target controlled infusion (TCI) device is currently used as a standardised computer-assisted infusion system for the administration of propofol, an hypnotic agent. The recommended initial target propofol plasma concentration (CT) for induction of anaesthesia has been shown to lie between 4 and 6 µg/ml. However individual adjustments are usually done in relation to a series of factors such as age of the patient, presence of a premedication, concomitant administration of opioid and ASA status.
Objectives: Estimate the probability to use a pre-defined propofol concentration value to induce anaesthesia (loss of consciousness) in a patient with given demographic characteristics, with and without the presence of pre-treatments and concomitant medications.
Methods: 238 patients, aged from 18 to 70 years, undergoing general anaesthesia have been enrolled in an open multicenter study to receive propofol using the Diprifusor TCI device. The investigators had to decide the plasma target CT to be used for induction. Each patient received propofol according to the same administration strategy an initial bolus administered at the constant rate of 20 ml/min until the calculated concentration reaches the CT value followed by the Diprifusor computed infusion rate suitable to maintain the target CT setting up to the induction.
A concentration-to-event analysis, based on a time-to-event modelling approach, has been used to estimate the probability of the choice by the investigator of a CT value in order to reach a loss of consciousness status given the values of some prognostic factors (covariates). The model included fixed effect parameters to describe the Probability Density Function (PDF) and the influence of the covariates and random effect parameters to account for the interpatient variability. The NONMEM (Non-linear Mixed Effect Modelling) program (version V) has been used for model building and parameter estimation. The PDF of CT was estimated using both Weibull and Exponential models. A stepwise procedure was used to include the covariates in the model and the criteria for selecting the ‘best’ model was based on the likelihood ratio test. Goodness of fit has been evaluated comparing the model-estimated hazard to the one obtained from a nonparametric estimator.
Results: The Weibull model was retained as the optimal one to describe our data. The results indicate that the target CT was reduced in relation to increasing age as well as in relation to the use of opioid and in presence of a premedication as illustrated in Table 1 by the increase of the percentage of elderly patients achieving induction with a CT of 4 µg/ml. Finally the time necessary to reach induction was also retained in the final model showing that the induction time for a same target CT increases with age.
Table 1: Probability of achieving induction with a CT of 4 µg/ml
| Probability (%) | Opioid*= 2 µg/kg | Opioid*= 6 µg/kg | ||
| 30 years | 60 years | 30 years | 60 years | |
| Without premed. | 11% | 55% | 16% | 70% |
| With premed. | 20% | 80% | 29% | 91% |
*expressed in fentanyl equivalents
Conclusion: The proposed methodology only describes the behaviour of anaesthetists involved in this trial but may contribute to further define practical strategies for an optimal use of propofol in TCI mode.
Reference: PAGE 8 () Abstr 130 [www.page-meeting.org/?abstract=130]
Poster: oral presentation