Abstract
Introduction: Paracetamol is one of the most common antipyretic and analgesic medications that is administered to children. Paracetamol appears to work primarily through the cyclooxygenase pathway. However, its precise mechanism of action is still not fully understood. Fever is a common feature of many childhood illnesses and is an important component of the body’s response to infection. Fever is often treated in children to alleviate discomfort. Despite paracetamol’s extensive use, information on its pharmacokinetics, metabolism and pharmacodynamics in children is limited. Studies are often difficult to perform in children due to the nature of the studies, ethical considerations and limited samples. The effects of fever and illness on paracetamol pharmacokinetics and metabolism are also not described in the literature. The present study aimed to investigate the pharmacokinetics of oral paracetamol in a paediatric population, examine paracetamol metabolism in children and explore paracetamol pharmacodynamics in fever using population modelling. The effect of fever and other covariates on pharmacokinetic and pharmacodynamic parameters were also evaluated.
Methods: An observational prospective population pharmacokinetic study was performed to describe the pharmacokinetics of paracetamol and its metabolites in children from the Southern Region, New Zealand. A pharmacodynamic study was also subsequently performed to describe the paracetamol pharmacodynamics in children in relation to fever. Participants aged between 28 days and six years of age were recruited from the paediatric assessment unit and children’s ward at Dunedin Public Hospital. Two different groups were studied: a fever group who had a recorded temperature of greater than 37.5 °C and were administered paracetamol and a pain group who were afebrile and were administered paracetamol for analgesia. Blood samples were collected from all participants as part of routine care. Demographic and clinical information were recorded and left over blood samples were obtained. An assay for paracetamol and metabolites paracetamol sulphate and paracetamol glucuronide was developed and validated using liquid chromatography–mass spectrometry. The assay required 20 μL of left over plasma sample. Chromatographic separation was achieved using a Sunfire C18 column with a run time of 11.2 minutes. Pharmacokinetic and pharmacodynamic modelling was performed using Phoenix NLME (Certara, Version 8.3.4.295). A population pharmacokinetic model for paracetamol and a population pharmacokinetic model for paracetamol and metabolites paracetamol glucuronide and paracetamol sulphate were developed. A pharmacodynamic model was also developed for paracetamol in relation to fever using data from the fever group. Different error models were explored for each model and covariates were tested using forwards selection and backwards deletion.
Results: A one compartment population pharmacokinetic model with first order absorption, a time lag compartment and a log additive error model best described the data. Weight fixed as a covariate at 0.75 on clearance and 1 on volume of distribution, to account for allometry, improved the model. The estimates from the pharmacokinetic model were 14.36 L (1.0 L/kg) (95% CI 11.16 L – 17.6 L) for volume of distribution, 6.54 L/h (0.47 L/h/kg) (95% CI 5.89 L/h – 7.19 L/h) for clearance, 0.72 h -1 for absorption rate constant (95% CI 0.58 h - 1 – 0.85 h -1 ) and 0.25 h for time lag (95% CI of 0.25 – 0.25 h). A one compartment population pharmacokinetic metabolite model with first order formation and elimination for paracetamol glucuronide and paracetamol sulphate and a log additive error model best described the data. Weight as a fixed covariate at 0.75 on all clearances and 1 on all volumes of distribution significantly improved the model. Age as a covariate at 0.23 on clearance to paracetamol glucuronide also significantly improved the model. The estimates for the metabolite model were a paracetamol clearance of 2.0 L/h (0.14 L/kg/h) (95% CI 1.51 L/h – 2.42 L/h), paracetamol glucuronide clearance of 3.2 L/h (0.23 L/kg/h) (95% CI 2.62 L/h – 3.85 L/h), paracetamol sulphate clearance of 5.8 L/h (0.42 L/kg/kg) (95% CI 4.97 L/h – 6.67 L/h), clearance to paracetamol glucuronide of 1.5 L/h (0.11 L/kg/h) (95% CI 1.19 L/h – 1.73 L/h) and clearance to paracetamol sulphate of 4.5 L/h (0.33 L/kg/h) (95% CI 3.72 L/h – 5.24 L/h). An inhibitory Emax model with an effect compartment and a log additive error model best described the pharmacodynamic model. Age as a covariate at 0.005 on baseline effect iii significantly improved the model. The estimates for the pharmacodynamic model were a baseline effect of 38.05 °C (95% CI 37.94 °C – 38.16 °C), equilibrium constant of 0.51 h-1 (95% CI 0.50 h-1 – 0.53 h-1 ), IC50 of 2.99 mg/L (95% CI 2.93 mg/L – 3.04 mg/L) and maximum inhibitory effect (Imax) of 1.12 °C (95% CI 1.07 °C – 1.18 °C).
Conclusions: The volume of distribution estimated from the pharmacokinetic model was similar to previous studies but clearance was higher than previously estimated. This was thought to be due to the study age group consisting of infants and preschool children, who have higher liver blood flow compared to older children and adults. Weight was a significant covariate for clearance and volume of distribution and was fixed at the values of allometry. Weight has been shown consistently to affect paracetamol pharmacokinetics. Paracetamol sulfation was the predominant pathway in the metabolite model. This is consistent with development, where sulfation predominates in childhood. Age was a significant covariate on the clearance to paracetamol glucuronide and was thought to be due to maturational changes in glucuronidation. There was no effect of fever on the metabolic pathways examined. However, the minor oxidative pathway was unable to be described due to not meeting laboratory validation standards. Access to deuterated internal standards for each metabolite should be obtained in the future. The pharmacodynamic model of paracetamol and fever provided more modest estimates for the baseline effect and Imax than previous studies. This was thought to be due to the present study’s more inclusive criteria for fever as paracetamol is thought to produce a greater magnitude of temperature reduction in participants with higher fevers. Age was a significant covariate on baseline effect. No other covariates affected the pharmacodynamic response.