Scaling propofol doses for the obese: is lean body weight the answer?
Sarah McLeay, Glynn Morrish, Carl Kirkpatrick & Bruce Green
University of Queensland, Australia
Objectives: Longer awakening times have been observed in obese compared to normal-weight patients when propofol is dosed on total body weight (TBW), as per label recommendations. This could be explained by a non-linear relationship between TBW and clearance (CL) as previously demonstrated [1]. Lean body weight (LBW) has been shown to increase non-linearly with TBW and linearly with drug clearance [2]. This body size metric may therefore be more appropriate than TBW for propofol dose selection. The objectives of this study were to evaluate the substitution of a non-linear covariate model for propofol clearance [1] with a linear LBW covariate model and examine the effects of different propofol dosing regimens on awakening times of obese patients.
Methods: One-hundred “true” datasets with 198 subjects each, 7 optimal time-points per subject and weight stratified into 3 groups of 40-60kg, 60-80kg, and 80-100kg, were simulated from a prior population PK model [1], in which clearance was defined as CL=86.4L.h-1*((TBW/70)0.75) – 2.7*(age-60). Age, weight, and height for each subject were simulated from a covariate distribution model constructed from a medical patient dataset (n = 999). PK parameters were re-estimated with both the original model and a reduced LBW model in which LBW replaced covariates on clearance as the single covariate on CL: CL=θ*(LBW/55). The predictive performance of the LBW model was evaluated by calculating the mean error (ME) and root mean square error (RMSE) of individual PK estimates from the “true” simulated values. These values were compared to the ME and RMSE of the original model. A visual predictive check (VPC) of the LBW model was also performed in which concentration-time data from a “true” simulated dataset was overlaid with the 10th, 50th and 90th percentiles of 1000 simulated datasets from the LBW model. The LBW model was used to simulate PK profiles for 4000 male subjects (180cm, 30yrs) in 4 weight categories: 70, 100, 130, and 160kg, with a dosing regimen of (a) a 2mg/kg bolus dose followed by a 1h, 6mg/kg/h infusion based on TBW, and (b) a 2.5mg/kg bolus and 1h, 7.6mg/kg/h infusion based on LBW (equivalent to recommended dose per kg TBW for a 70kg subject). A prior PD model for probability of awakening [3] with an EC50 = 1.07ug/ml was used to determine the post-infusion median probability of awakening over time for each weight group.
Results: Re-estimation of individual CL values using the full multivariate and LBW model yielded a ME of -3.94L/h vs -5.03L/h, and RMSE of 8.25L/h vs 8.99L/h, respectively. ME and RMSE of the other PK parameter estimates were also comparable. Between subject variability (as %CV) for each of the parameters was within 10% of the simulation values for both models, with random unexplained variability within 2%. The VPC confirmed that the LBW model was able to describe the “true” data, with ~80% of the “true” data falling within the 10th and 90th prediction intervals at each time point. Dosing on TBW resulted in an increased probability of longer time to awakening in the larger weight groups, with the median probability of a subject still being asleep at 30min post-infusion being 68% for 160kg vs. 28% for 70kg. When dosed on LBW, there was no difference between groups (Pasleep(30min) = 30% for all groups).
Conclusions: The LBW covariate model has similar predictive properties to a multivariate covariate model that describes a non-linear increase in CL with TBW. This LBW model may explain the observed differences in patient times to awakening when dosing is based on TBW. As such, we propose that LBW rather than TBW is the most appropriate body size descriptor to dose propofol in the obese population. Observational data from both normal-weight and obese subjects will be collected to support these findings and a prospective clinical trial completed in order to develop a safer dosing regimen for obese patients.
References:
[1] Schuttler, J. & Ihmsen, H. (2000). Population pharmacokinetics of propofol: a multicenter study. Anesthesiology 92(3): 727-38
[2] Han, P.Y, Duffull, S.B., Kirkpatrick C.M. & Green B. (2007). Dosing in obesity: a simple solution to a big problem. Clin Pharmacol Ther 82(5):505-8
[3] Shafer, A., Doze, V.A., Shafer, S.L. & White, P.F. (1988). Pharmacokinetics and pharmacodynamics of propofol infusions during general anesthesia. Anesthesiology 69(3): 348-56