Luong Vuong 1, Joseph Standing 2,3, Omar Elkayal 1, Sophida Boonsathorn 2,4, Sian Bentley 6,7, Athanasios Tragiannidis 8,9, Andreas Groll 8, Jenna Nickless 10, Adam Brothers 10, Anne Uyttebroeck 11, Romina Valenzuela 12, Jorge Morales 12, Isabel Spriet 1,13, Erwin Dreesen 1
1 Department of Pharmaceutical and Pharmacological Sciences, KU Leuven (Leuven, Belgium), 2 Great Ormond Street Institute of Child Health, University College London (London, United Kingdom), 3 Department of Pharmacy, Great Ormond Street Hospital (London, United Kingdom), 4 Ramathibodi Hospital, Mahidol University (Bangkok, Thailand), 5 Quotient Sciences (Nottingham, United Kingdom), 6 Pharmacy Department, Royal Brompton Hospital (London, United Kingdom), 7 National Heart and Lung Institute, Imperial College London (London, United Kingdom), 8 Infectious Disease Research Program, Centre for Bone Marrow Transplantation and Department of Paediatric Haematology and Oncology, University Children's Hospital Müenster (Münster, Germany), 9 Haematology Oncology Unit, and Department of Paediatrics, Aristotle University of Thessaloniki, AHEPA Hospital (Thessaloniki, Greece), 10 Department of Pharmacy, Seattle Children's Hospital (Washington DC, USA), 11 Department of Paediatric Haematology and Oncology, University Hospitals Leuven (Leuven, Belgium), 12 Hospital Dr. Luis Calvo Mackenna, Universidad de Chile (Santiago, Chile), 13 Pharmacy Department, University Hospitals Leuven (Leuven, Belgium)
Introduction: Invasive aspergillosis remains a major cause of morbidity and mortality in paediatric patients with immunodeficiencies and haematological cancers. Posaconazole is used for prophylaxis and treatment, but despite the availability of three formulations (oral suspension and gastro-resistant tablets, and a solution for intravenous administration), optimal dosing remains unclear [1].
Objectives:
• To develop a population pharmacokinetic (popPK) model that best describes the pharmacokinetics of posaconazole, and
• to conduct simulations identifying posaconazole dosing regimen(s) that are potentially safe and effective in paediatric patients.
Methods: We developed a popPK model using data from seven studies documenting the pharmacokinetics of all three posaconazole formulations in paediatric patients [1-7]. We evaluated different fixed- and random-effects structures with linear and non-linear (Michaelis–Menten) clearance using state-of-the-art model-building and diagnostic tools. Covariate model building was done via two-way stepwise covariate modelling (α_forward = 0.05, α_backward = 0.01).
We tested time-varying body weight, concomitant proton-pump inhibitor (PPI; yes/no) use, and diarrhoea (yes/no) as covariates. We also tested the posaconazole dose as a covariate on the absolute nonlinear suspension bioavailability (Fsus) using the formula Fsus = 1 – Dose/(Dose+ED50), with the dose expressed as mg/m2 body surface area (BSA), and ED50 the estimated dose in mg/m2 resulting in an Fsus of 50% [5]. We used the next observation carried backward to impute the missing PPI and diarrhoea status. We assessed the final model using prediction-corrected visual predictive checks (pcVPC; 1,000 replicates). We also used bootstrapping to obtain nonparametric estimates of parameter uncertainty (n = 2,000 bootstraps).
We then performed stochastic simulations accounting for parameter uncertainty. We created virtual patient cohorts with body weights ranging from 5 kg to 50 kg in 5-kg increments, reflecting the weight range of most paediatric patients [1]. We subsequently estimated BSA (m2) from weight (g) alone using the Boyd method: BSA = 0.0004688∙weight^(0.8168–0.0154∙Log(weight)) [8]. Each virtual patient was uniquely identified by body weight, diarrhoea, and PPI status. Hence, for each diarrhoea and PPI status combination (yes/yes, yes/no, no/yes, no/no), 10 patients were generated. We simulated each patient 10,000 times in two stages (100 × 100): first sampling parameters from their variance–covariance matrix, followed by sampling from the inter-individual variability distribution.
We evaluated several dosing regimens for the intravenous and tablet formulations. Specifically, each regimen included a loading dose administered twice daily (bid) on day 1 followed by the same maintenance dose administered once daily qd (x mg/kg or y mg; x = {6, 12, 15, 20}, y = {100, 200, 300, 600} for tablets, y = {100, 200} for intravenous). For the suspension, we only tested 200 mg four times daily (qid) and 400 mg qid dosing. To assess efficacy and safety, we calculated the probability of attaining a trough concentration (C_trough) >1 mg/L (P_efficacy) [2] and the probability of not exceeding a C_trough of 4 mg/L (P_safety), a threshold associated with symptomatic adverse drug reactions [9]. We evaluated both metrics on days 2 and 8 of therapy.
PopPK analysis was performed in NONMEM 7.5.0, while simulations were conducted using the rxode2 R package (version 4.1.1).
Results: A total of 309 paediatric patients contributed 1050 total posaconazole plasma concentrations. The missingness percentage for PPI and diarrhoea status was 1.5%. A one-compartment popPK model with nonlinear elimination best described the data. Concomitant PPI use and having diarrhoea significantly reduced the absolute Fsus by 40% (9.8% RSE) and 27% (24.4% RSE), respectively. At a posaconazole dose of 22 mg/m2 (9.3% RSE), the Fsus equals 50%.
The suspension showed low P_efficacy (<80%), even with the doubled adult dosing of 400 mg qid in the absence of diarrhoea and PPI use. Both tablet and intravenous formulations, when used as 12–15 mg/kg/dose, attained reasonably effective and safe exposure on day 2, with P_efficacy >75% and P_safety >55% for the intravenous formulation and P_efficacy >70% and P_safety >80% for the tablet formulation. By day 8, as posaconazole accumulated, exposure remained effective but became less safe, as P_efficacy increased and P_safety decreased.
Conclusions: We proposed a loading dose of 12–15 mg/kg/dose bid for both intravenous and tablet formulations on day 1, followed by (model-informed) therapeutic drug monitoring to ensure safety. We did not recommend the oral suspension, consistent with the guidance in the Summary of Product Characteristics [10].
References:
[1] Kane Z et al. Antimicrobial Agents and Chemotherapy (2023) 7, e00077-23.
[2] Vanstraelen K et al. The Pediatric Infectious Disease Journal (2016) 2, 183-8.
[3] Nickless JR et al. Journal of the Pediatric Infectious Diseases Society (2018) 4, 365-7.
[4] Valenzuela R et al. Revista chilena de infectologÃa (2018), 15-21.
[5] Boonsathorn S et al. Clinical Pharmacokinetics (2019) 1, 53-61.
[6] Tragiannidis A et al. Journal of Antimicrobial Chemotherapy (2019) 12, 3573-8.
[7] Bentley S et al. Journal of Antimicrobial Chemotherapy (2021) 12, 3247-54.
[8] Sharkey I et al. British Journal of Cancer (2001) 1, 23-8.
[9] Jensen K et al. Med Mycol (2023) 8.
[10] https://www.merck.com/product/usa/pi_circulars/n/noxafil/noxafil_pi.pdf
Reference: PAGE 34 (2026) Abstr 12145 [www.page-meeting.org/?abstract=12145]
Poster: Drug/Disease Modelling - Other Topics