Marsh, Schnider and Eleveld Propofol Models

Marsh Model

• One of the first developed (Gavin Kenny and colleagues). A relatively simple but good model that has been used in clinical practice for > 30 years
  • Familiar with extensive practical experience and evaluation
• Adapted from the Gepts three-compartmental model
• Used a dataset of 150 patients and then validated this in another 30 patients
• Bodyweight is the only covariate in this model i.e. no allometric scaling
• Central compartment is weight proportional (i.e. the loading dose increases with weight) and is expressed in ml per kg
• The 5 transfer rate constants (i.e. in and out of the 3 compartments) are fixed
• It was Initially used only in the DiprifusorTM machine but is now commonly available in most TCI pumps (some don’t include the effect site model – only plasma – but I wouldn’t buy one of those!)
• Ke0 was not included in the initial development of this model
• Different researchers have attempted to derive a Keo value for this model
• The original ke0 used was 0.26 min-1
• Many commercial TCI pumps used a larger (faster) ke0 of 1.2 min-1 and this is referred to as the modified Marsh model. However, this cause an over prediction of the effect site which is not as bad as Schnider but should be kept in mind during use.

Schnider Model

• More complex model and developed during combined PK-PD studies (Thomas Schnider and colleagues)
• Based on arterial samples from 24 volunteers using bolus and infusion of propofol. Note arterial samples will show higher drug concentrations than venous and there is some controversy over which is best for PK accuracy.
• Allometric scaling. Covariates
  • Weight
  • Height
  • Age
  • Sex
  • Lean body mass
• V1 and V3 (and thus k13 and k31) are fixed.
• V2 (and thus k12 and k21) is influenced only by age
  • It becomes smaller with advancing age
• Limitations of the model
  • Fixed central compartment volume of distribution, meaning the initial bolus will be the same for patients of all body weight and age i.e. the loading dose will be the same for a 90 kg as a 30 kg patient which doesn’t “seem” correct.
  • After a bolus of a given size, the same predicted peak plasma concentration is achieved for all patients
  • Uses the James formula for weight. As BMI increases beyond 42 kg m-2 in males, and 37 kg m-2 in females, there would be a paradoxically low calculated LBM that causes a large increase in calculated k10
  • TCI pumps have implemented different solutions to handle this e.g. they won’t allow input of high BMI
  • To achieve the same plasma concentration, the dose administered is considerably smaller than with the Marsh. This inaccuracy has led some to believe, erroneously, that Schnider is safer in more frail patients. The key issue, of course, is to slowly titrate up the Marsh in such patients. We do emphasise the importance of titration over and over again……

Eleveld Model

• This is a “general purpose model” with allometric scaling. Developed for a broad population range: adults, children, older subjects, and obese adults
• It is an important advance as this model incorporates PK data derived from 30 previously published studies containing data collected from children of all ages, adults, obese adults and elderly individuals (no more choice such as Marsh v Schnider, different model for kids etc.)
• Constructed a 3-compartment model where weight, age, sex and administration of comedication were covariates
• Retrospective analysis of this model performance shows that the model performs at least as good as, or better, than dedicated population models in all populations
• The model showed a bias < ±20% in children, adults, and obese adults, but a greater bias (−27%) in older subjects. Precision was < 30% in all groups. For PD, the bias and wobble were < 5 BIS units and the precision was close to 10 BIS units in all groups

Which model should I use?

• Traditional teaching was to use Marsh for Cp target and Schnider for Ce target
  • Marsh is associated with a higher dose for the same Cp or Ce
  • Ce target is associated with a higher initial dose
  • The effect-site concentrations predicted by the original Schnider model were shown to have no relationship to the BIS or any other measure of propofol PD effect
  • This may give rise to inaccurate information on the level of Ce of loss of consciousness
• The Marsh model has more accurate effect-site targeting
  • We recommend starting at a low concentration and titrating slowly – Large inter-individual and intra-individual differences exist
• Despite the difference between the Marsh and Schnider models, both can achieve and maintain reasonably stable plasma concentrations of propofol
• As long as you understand its limitations, it does not matter which model you choose and it may be best to get used to one particular model and stick with it
• We believe that the Eleveld model will supersede both of these and become the model of choice in all populations.

Titration is key to the safe delivery of propofol TCI