Population Pharmacokinetic Pharmacodynamic Modelling to Improve Anaesthesia in Children and Adults

Abstract

Pharmacokinetic pharmacodynamic models are used in adult and paediatric anaesthesia. The pharmacokinetic parameter volume is used to calculate loading dose while maintenance dose rates can be determined from clearance. The major differences (covariates) between children and adults concern growth (size) and development (age). Body composition is a component of size; size descriptors including fat-free mass and normal fat mass have been proposed to account for differences between individuals. Knowledge of these covariates allows prediction of drug time-concentration and concentration-effect relationships in the individual patient. In this thesis I develop pharmacokinetic pharmacodynamic models for the anaesthetic and analgesic drugs propofol, sevoflurane, dexmedetomidine, oxycodone, diamorphine, acetaminophen and ibuprofen. A paediatric propofol pharmacokinetic parameter set was derived and used to suggest dose regimens for anaesthesia. Clearance increased with age to reach 92% of the typical adult value by 6 months postnatal age. A target concentration of 3 mg/L in children aged 3-12 months was achieved with a 2.5 mg/kg bolus and stepwise infusion 12-11-10 mg/kg/h for 0-15, 15-30 and 30-60 minutes. A compartmental model was developed to describe sevoflurane pharmacokinetics and included a direct measure of sevoflurane plasma concentration. Sevoflurane pharmacodynamics were examined with bispectral index (BIS) as a measure of drug effect. The combined drug effects of sevoflurane, propofol and remifentanil on BIS were accounted for using a drug interaction model. The maximal reduction in BIS (EMAX ) was 61 (PPV 22%) and the concentration producing 50% of E MAX (C50) was 17 mg/L (PPV 10%). The arterial sevoflurane concentration-effect relationship, quantified with an equilibration half-time was 1.3 minutes. The C50 for propofol (1.7 mg/L) and remifentanil (12 μg/L) were estimated. Pharmacokinetic models often fail to account for changes in drug disposition attributed to obesity. Normal fat mass or fat-free mass may be superior size descriptors to total body weight to account for obesity related changes. Fat-free mass was the best size descriptor for dexmedetomidine clearances and normal fat mass for dexmedetomidine volumes of distribution. The factor of fat contributing to normal fat mass for volume (FfatV) was 0.293. Conversely, total body weight was the best size descriptor for oxycodone pharmacokinetics given via intravenous, intramuscular, buccal, epidural and nasogastric routes. The absorption half-time (TABS) was longest with buccal administration (160 minutes). Nasogastric administration was associated with a lag-time of 9 minutes. Bioavailability with buccal was 99% compared with 67% for nasogastric. This parameter set was used to design an oxycodone dosing regimen to achieve a target concentration of 35 μg/L for adequate and safe analgesia. The intravenous dose required to achieve that target in a neonate was 100 μg/kg followed by 14 μg/kg/h. The diamorphine exposure-analgesia relationship was explored using a pharmacokinetic model developed from an adult diamorphine model and a model for morphine in children. Simulation demonstrated the rapid formation of 6-monoacetylmorphine from diamorphine and the subsequent formation of morphine. This model was used to suggest an intravenous diamorphine dose to achieve a target concentration of morphine (30 μg/L). That dose in neonates was 28 μg/kg followed by infusion 14 μg/kg/h. Pharmacokinetic models for intravenous drugs (e.g., propofol, dexmedetomidine) can be programmed into target-controlled infusion pumps for use in children and adults. Models for these drugs in this thesis are developed from broader populations than those currently available in target-controlled infusion pumps. This may provide more accurate plasma and effect-site targeting in a wider range of individuals. Models for analgesics (oxycodone, diamorphine) can be used to identify doses associated with a target effect. Pharmacodynamic models can provide information about concentration- dependent adverse effects. Fat mass is a useful covariate to describe variability in pharmacokinetic parameters for some drugs. This must be considered when developing models for clinical use in a broad population of ages and weights.