Abstract
Introduction: Paracetamol has been used as an analgesic and antipyretic drug in patient populations for decades. Recently, it has become available as an intravenous (IV) formulation for use in postoperative pain management. Inclusion of IV paracetamol in multimodal analgesia has raised concerns about patient safety due to its toxic effects in overdose. When compared to healthy volunteers, patient populations may exhibit variable physiology that can influence the pharmacokinetics (PK) of medicines. Changes to drug absorption and distribution are well known, but published research on postoperative drug metabolism is limited.
Paracetamol is extensively metabolised in the liver via conjugation and oxidation pathways; with oxidative metabolism being responsible for its hepatotoxicity. Altered postoperative metabolism of paracetamol is of interest to clinicians because 95% of paracetamol is metabolised. Increased oxidative metabolism may predispose an already vulnerable patient group to toxicity. Alternatively, an increase in conjugation metabolism may decrease the amount of active parent paracetamol available in the plasma and thus limit its clinical utility. The influence of patient demographics on paracetamol pharmacokinetics in the postoperative period is unknown.
The use of paracetamol in the postoperative period also raises concerns over its effect on prothrombin time (PT), clotting and potential implications for surgical patients. Paracetamol has been reported to cause an abnormal elevation in PT following overdose, but it is unknown if this effect may be seen at therapeutic doses (1). PT is a measure of hepatic injury and thus relevant in the prediction and monitoring of hepatotoxicity and is reported as the International Normalised Ratio (INR) (1-3). The mechanism for the increase in PT seen in paracetamol poisoning without hepatic injury, is the reduction in functional factor VII (1, 4). N-acetylcysteine (NAC) is the antidote of choice for paracetamol overdose and may have an independent effect on the INR (5). The second part of the research presented in this thesis was to develop a population pharmacokinetic-pharmacodynamic (PKPD) model to describe the effect of paracetamol and NAC on INR.
Methods: A prospective multiple dose pharmacokinetic study with two sampling intervals was investigated to describe the PK profile of paracetamol and its metabolites in a group of abdominal surgery patients in Dunedin, New Zealand. All 20 patients were given 1 g paracetamol by IV infusion at induction of anaesthesia (Interval 1) and every 6 hours thereafter, with the final dose given at 48 to 72 hours (Interval 2). Plasma and urine samples were collected for up to 8 h after infusion for both Intervals. Samples were analysed by high-performance liquid chromatography to determine the amount of paracetamol and its metabolites. The primary objective of this study was to describe the pharmacokinetic profile of paracetamol and its metabolites. The secondary objective was to identify changes in the pharmacokinetic parameters of volumes of distribution, metabolite formation and urinary clearance during the postoperative period. The data was modelled in Phoenix® WinNonlin® using a user-defined parent-metabolite model with linear disposition, to obtain estimates for volume of distribution, metabolic and urinary clearance.
The 20 patients from this dataset were then combined with another prospective multiple dose pharmacokinetic study with multiple sampling intervals. This dataset contained 33 patients from four study sub-groups from Limerick and Cork, Ireland, increasing the study size and incorporating a greater variety of paracetamol dosing regimens and surgical procedures. Within the combined dataset consisted of 53 patients; 28 were men, median age (range) 60 years (33-87), median weight (range) 74 kg (54-129). Patients received 1, 1.5 or 2 g of IV paracetamol. Plasma and urine samples were collected at various intervals for up to 6 days after surgery. The objective of this study was to describe the population pharmacokinetic profile of intravenous paracetamol and its metabolites in the postoperative period and to describe the inter-subject variability. The secondary objective was to use the population model to investigate the potential influence of patient demographics and clinical characteristics (covariates) on the PK of paracetamol in the postoperative period. Simultaneous modelling of parent paracetamol and its metabolites was conducted in Phoenix® NLMETM to estimate PK parameters, inter-subject variability and covariate effects.
A further study involved a post hoc analysis to quantify the effect of paracetamol in overdose on the INR. This was investigated by developing a population PKPD model using a data from multiple published clinical studies. The dataset included paracetamol plasma concentration and INR data from healthy clinical trial volunteers receiving high doses (60 and 90 mg·kg-1) and a retrospective group of paracetamol overdose patients. A total of 172 patients, contributing 198 cases, were included in the study dataset (151 paracetamol overdose cases, 8 control overdose cases, 39 cross-over clinical trial cases). Of the 198 cases, 96 (49%) were male, and the median age (range) was 22 years (13-71). A structural population PKPD model was developed in Phoenix® NLMETM. A total of 172 patients contributed to 908 paracetamol plasma and INR observations.
Result: From the initial PK analysis of abdominal surgery patients the mean (95% CI) metabolic clearance to paracetamol glucuronide increased from 4.2 (3.1-5.3) to 9.5 (6.9-12.2) L·h-1·70kg-1, P < 0.001 and urinary clearance increased from 5.6 (4.5-6.7) to 9.6 (7.0-12) L·h-1·70kg-1, P = 0.002. The mean (95% CI) volume of distribution of paracetamol increased from 11 (7.5-14) L·70kg-1 to 26 (18-33) L·70kg-1, P = 0.032. From the combined dataset analysis, the population mean estimate (95% CI) for central (plasma) volume of distribution of parent paracetamol (VP) was 13 (12-15) L, peripheral (tissue) volume of distribution (VT) 47 (43-51) L, and inter-compartmental clearance (Q) 72 (59-85) L·h-1. Mean (95% CI) metabolic clearances (L·h-1) for glucuronidation (CLPG), sulfation (CLPS), and oxidation (CLPO), were 14 (13-15), 0.42 (0.37-0.51), and 0.39 (0.23-0.33), respectively. Mean (95% CI) urinary clearances (L·h-1) of parent paracetamol (CLRP), paracetamol glucuronide (CLRG) paracetamol sulfate (CLRS) and paracetamol cysteine + mercapturate (CLRO) were 0.049 (0.038-0.060), 1.7 (1.5-1.9), 1.4 (1.2-1.7), and 1.7 (1.3-2.0), respectively. The only patient covariate that was significant was the effect of renal function on the renal excretion of unchanged parent paracetamol.
The PKPD of oral paracetamol in the high dose (60 and 90 mg·kg-1) and overdose dataset, were best described by a one-compartment model with first-order input and linear disposition. A baseline Emax model was modified by the addition of an effect compartment to describe the time course and effect of paracetamol on INR by inhibition of the activation of vitamin K-dependent coagulation factors. The population mean estimates (95% CI) of the PK parameters volume of distribution, clearance, absorption rate constant (Ka), and lag time (tlag) were 13 (11-15) L, 2.5 (2.2-2.8) L·h-1, 4.2 (2.1-6.4) h-1, and 0.38 (0.33-0.43) h, respectively. The population estimates (95% CI) of the PD parameters EmaxP and EC50P for paracetamol were 0.53 (-1.6-2.7) (increase in INR) and 1302 (-4919-7522) μM, respectively. The population estimate (95% CI) of the pharmacodynamic (PD) parameter Emax for NAC was 0.33 (0.27-0.38) (increase in INR). Covariates investigated included patient age and gender.
Conclusions: Following major abdominal surgery, there were significant increases in the metabolic conversion of paracetamol to its glucuronide and the glucuronide urinary clearance suggesting potential induction of the glucuronidation metabolic pathway. There are no significant covariate effects on the parent-metabolite PK parameters. Altered postoperative metabolism were not associated with increased risk of toxicity. The PKPD analysis of paracetamol and its effect on INR has further contributed to our understanding of the time course of INR after overdose by quantifying the effect of paracetamol on INR by estimating Emax and EC50. The estimate of EC50P for paracetamol was within the range of plasma paracetamol values in the study group, but was five-fold greater than therapeutic Cmax. This suggests that while an elevated INR occurs in overdose, it would not be expected with therapeutic doses of paracetamol in surgical patients.