Janneke M. Brussee (1), Nienke J. Vet (2), Pyry A.J. Välitalo (1), Elke H.J. Krekels (1), Mariska Y. Peeters (3), Abraham J. Valkenburg (4), Evelyne Jacqz-Aigrain (5), Joop M.A. van Gerven (6), Matthijs de Hoog (2), Saskia N. de Wildt (4), Catherijne A.J. Knibbe (1,3)
(1) Division of Pharmacology, Leiden Academic Centre for Drug Research (LACDR), Leiden University, Leiden, The Netherlands, (2) Department of Pediatrics, Erasmus MC - Sophia Children’s Hospital, Rotterdam, the Netherlands, (3) Department of Clinical Pharmacy, St. Antonius Hospital, Nieuwegein, the Netherlands, (4) Department of Pediatric Surgery, Erasmus MC - Sophia Children’s Hospital, Rotterdam, the Netherlands, (5) Department of Paediatric Pharmacology and Pharmacogenetics, Hôpital Robert Debré, Paris, France, (6) Centre for Human Drug Research, Leiden, the Netherlands
Objectives: Midazolam is often used for sedation in pediatric intensive cares and for surgical procedures. A recent study quantified the effect of inflammation and organ failure on midazolam clearance in critically ill term neonates and children [1]. This project aims to evaluate the predictive performance of this recently developed model across a wide range in (preterm) neonatal, pediatric and adult populations.
Methods: Concentration predictions by the population pharmacokinetic model for midazolam in which CRP concentrations and organ failure are taken into account [1], were evaluated in n=55 preterm neonates [2,3], n=18 critically ill children [4], n=26 children undergoing open-heart surgery [5], n=24 children undergoing craniofacial surgery [6], n=6 pediatric oncology patients [7], and n=20 healthy adults [8]. In total n=149 subjects (aged 1 day – 31 years, body weight 0.77 – 89 kg) receiving intravenous midazolam were included. In case of missing covariate values (i.e. CRP concentrations [2,3,6-8]), typical values without covariate effects were assumed.
Results: The model predictions were adequate for critically ill children [4] and children undergoing open-heart surgery [5] (median prediction error (MPE) 14% and -3.1% respectively). In the pediatric oncology patients [7], overall MPE was 50.2% and the trough concentrations were more accurately predicted than the peak concentrations. For patients undergoing craniofacial surgery [6], plasma concentrations were overpredicted (MPE -250%). In preterm neonates [2,3], the plasma concentrations were underpredicted (MPE 63.5 and 68.2%). In the densely sampled adult study [8], a trend in residuals over time (CWRES) was apparent, with under- and overprediction of the peak and trough concentrations, respectively.
Conclusions: Based on the data from these 7 studies, it can be concluded that the model cannot be used for direct extrapolation to other populations such as preterm neonates and adults. However, for critically ill children in the same age and body weight range as the subjects in the learning dataset and to a lesser extent for children undergoing surgical procedures, midazolam plasma concentrations could be accurately predicted.
References:
[1] Vet NJ, Brussee JM, et al. 2016 Am J Respir Crit Care Med. 2016 Jan 21
[2] De Wildt SN et al. Clin Pharmacol Ther. 2001 Dec; 70(6):525-31
[3] Jacqz-Agrain E et al. Lancet. 1994 Sep; 344(8923):646-50
[4] De Wildt SN et al. Critical care medicine. 2003 Jul;31(7):1952-8
[5] Valkenburg AJ et al. Pain. 2011 Sep;152(9):2059-64
[6] Peeters MY et al. Anesthesiology. 2006 Dec; 105(6):1135-46
[7] De Wildt SN et al. Clin Pharmacol Ther. 2000 Mar;67:104
[8] Van Gerven JM et al. Br J Clin Pharmacol. 1997 Nov; 44(5):487-93
Reference: PAGE 25 (2016) Abstr 5931 [www.page-meeting.org/?abstract=5931]
Poster: Drug/Disease modeling - Paediatrics