What is the optimum ke0 for use with the Marsh model for propofol?: Proposed methodology for a clinical study.

John B Glen, Research Department, Glen Pharma Ltd, Knutsford, Cheshire WA16 0DZ, United Kingdom

Introduction: Previous studies, using surrogate EEG markers of effect, have proposed values of ke0 for propofol ranging from 0.20 [1] to 1.21 [2]. The Diprifusor (AstraZeneca) TCI system which uses the Marsh [3] PK model for propofol uses a value of 0.26 for prediction only. As interest in control of the effect-site concentration increases, it is important that the optimum ke0 for use with the Marsh PK model is identified.

Methods: This computer simulation study (PK-SIM) investigated the influence of ke0 values of 0.26, 0.52 and 1.04 on predicted blood and effect-site concentrations and the cumulative amount (mg) of propofol delivered in a 70 kg patient over time with effect site TCI. An input target profile (Ce_T) relevant to conscious sedation was used with an initial setting of 0.4 μg.ml^-1. This was increased in increments of 0.2 μg.ml^-1, as soon as each target was reached, until a final value of 2.0 μg.ml^-1.

Results: The lowest ke0 value gave the greatest initial bolus, but thereafter there was little difference in the amount of propofol delivered over time. However the rate of increase in calculated effect site concentration (Ce_CALC) is fastest with the largest ke0.  Assuming a clinical study indicates that a desired depth of sedation is achieved after 8 min, if a ke0 of 0.26 is used, Ce _CALC would be 0.9 μg.ml^-1 and if ke0 1.04, the Ce_CALC would be 1.8 μg.ml^-1. If at 8 min Ce_T is adjusted to Ce_CALC at this time and continued up to 30 min, a system with a ke0 of 0.26 will deliver 108 mg propofol and one with a ke0 of 1.04 will deliver 198 mg.

Fig: Influence of ke0 used with Marsh PK model for propofol on cumulative amount delivered to maintain a desired level of sedation.

Discussion: These results suggest that simple endpoints could be used to compare the clinical utility of effect controlled TCI systems with different ke0 values. The system with the optimum ke0 should facilitate the achievement and maintenance of a desired effect with the minimum likelihood of over or under sedation. If the ke0 value is ‘too small’ insufficient drug will be delivered and upward titration of Ce _T will be required and if ke0 is ‘too large’ downward titration will be necessary to maintain the desired effect. A sedation score able to detect minor differences in levels of conscious sedation is proposed: 0, fully awake;1, drowsy but brisk response to normal voice; 2, drowsy with sluggish response to normal voice; 3, responds only to loud voice; 4, no response to voice.

These observations are also relevant to the situation with plasma control TCI where the plasma concentration is adjusted to a displayed, predicted effect site concentration at the time when a desired depth of sedation or anaesthesia is reached.

References: 1. Br J Anaesth.1999; 82:333.  2. Anesthesiol. 2000; 92:399.  3. Br J Anaesth. 1991; 67: 41.