PK of acetaminophen and its metabolites in preterm and term neonates using relevant external background information in a Bayesian approach with Stan
Sebastian Weber (2), Willi Weber (1), Franziska D. Weber (1), Henning Schmidt (2), Marc Pfister (1,3), Johannes N. van den Anker (1,4)
(1) Department of Paediatric Pharmacology and Pharmacometrics Research Center, University Children’s Hospital Basel, Switzerland, (2) Novartis Pharma AG, Basel, Switzerland, (3) Quantitative Solutions LP, Menlo Park, CA, USA (4) Division of Clinical Pharmacology, Children’s National Medical Center, Washington, DC, USA
Objectives: Drugs used in the adult population are urgently required in the treatment of pre- and term neonates. Acetaminophen (APAP) and its metabolites are well investigated in adults. In term and preterm neonates the PK information of APAP is still insufficient for choosing optimal dose regimens. Currently, iv administration of APAP is clinically investigated in neonates [1]. We obtained such study data, including concentrations of APAP and its glucuronide and sulfate metabolites in plasma. Without knowing the dose fraction formed to a specific metabolite, its volume of distribution remains unidentifiable in a NONMEM model. We employed a Bayesian approach to include extensive literature data from adults [2] into the model via the prior. This enabled calculation of all model parameters, including formation rates and volumes. The main objective was to quantify development changes in the neonate metabolism on the PK of APAP and its metabolites.
Methods:
A one compartment model with linear elimination was used for APAP and its metabolites. Both metabolite compartments received their input as metabolized dose fraction from the parent compartment. Identified relevant external background information:
- in adults a dose fraction of 60% and 30% of APAP are renally excreted as APAP-glucuronide and -sulfate, respectively. Both of their clearances correspond to the GFR.
- ‘1/4’ allometric scaling using body weight and maturation dependend on postmenstrual age (pma) were applied on all clearances and volumes [3].
- exponential increase with pma of glucuronide-formation, glucuronide & sulfat elimination clearance were incorporated into the model.
- exponential decrease with pma of the APAP volume of distribution
The study included in total of N=10 neonates with gestational age (GA) <=28 weeks who received 5x15mg/kg/12h and N=25 neonates with GA > 28, receiving 7x 15mg/kg/8h. Per subject on average 7.2 PK samples were obtained. We use the Bayesian software Stan for model parameter inference [4].
Results:
Full parameter sets for the parent drug and its metabolites were estimated for a reference neonate with a postmenstrual age of 34 weeks:
- elimination clearance CL[l/h/70kg]: APAP=17.1(15.7,18.7)95, APAPgluc=8.15(6.0,10.6)95, APAPsulf=7.0(5.9,8.24)95
- formation clearance CL[l/h/70kg]: APAPgluc=2.6(1.9,3.4)95, APAPsulf=12.9(11.1,14.7)95
- volume of distribution [l/70kg]: APAP=78.2(70.2,86.7)95 , APAPgluc=49.5(35.8,65.8)95, APAPsulf=28.5(24.8,32.3)95
- the maturation process of the glucuronide formation was about twice, 2.06(1.6,2.7)95, as fast as the GFR maturation process (11.9 weeks doubling time).
- the G/S formation ratio was found to increase rapidly from 1:10 for very early (pma=25) to about 1:2 for later (pma=45) neonates, drastically smaller than the adult 2:1 G/S ratio.
Conclusions: With a Bayesian approach we were able to interpret a sparse data set in the context of available background knowledge. This enabled us to obtain quantitatve results on the changing metabolism of neonates for APAP, APAPgluc & APAPsulf PK. The Stan model has been written to use NONMEM data sets, enabling rapid application in future analyses.
References:
[1] G.M. Pacifici, K. Allegaert. 2014., Curr. Therap. Res., Clin. and Exp. 77 (Dec): 24-40
[2] M.M. Rhodin et al., 2009., Pedi. Neph. (Berlin, Germany) 24 (1): 67–76
[3] A.L. Sumpter, and N. H. G. Holford. (2011), Paed. Anaes. 21 (3): 309–15.
[4] Stan Dev Team (2015), Version 2.6, http://mc-stan.org
