Haini Wen 1, Kai Lukas Wiater 1, Sandrine Marchand 2,3, Elin Svensson 1,4, Lena Friberg 1
1 Department of Pharmacy, Uppsala University (Uppsala, Sweden), 2 Université de Poitiers, Inserm U1070, PHAR2 (Poitiers, France), 3 CHU de Poitiers, Laboratoire de toxicologie et de pharmacocinétique (Poitiers, France), 4 Department of Pharmacy, Pharmacology and Toxicology, Radboud University Medical Center (Nijmegen, The Netherlands)
Introduction:
Ventilator-associated pneumonia (VAP) and hospital-acquired pneumonia (HAP) are serious infectious complications in hospitalized and critically ill patients, contributing to prolonged hospital stay, increased healthcare costs, and high mortality [1]. Intravenously (IV) administered antibiotics may fail to achieve adequate pulmonary concentrations [2]. Nebulised antibiotics have therefore been recommended to enhance local drug delivery, especially when systemic therapy is insufficient [3]. However, large randomized controlled trials have not demonstrated clear clinical benefits when nebulised therapy is added to standard regimens [4-6], potentially due to suboptimal dosing and inadequate infection-site drug exposure [7,8]. Therefore, evaluating dosing strategies for nebulised aminoglycosides is essential before initiating future clinical trials.
Objectives:
This study aimed to inform dosing of inhaled aminoglycosides through translational modelling and simulation for the treatment of ventilator-associated pneumonia (VAP) and hospital-acquired pneumonia (HAP).
Methods:
A published translational PBPK modelling framework [9] was adapted to predict epithelial lining fluid (ELF) and systemic concentrations of gentamicin and amikacin following intravenous and nebulised administration. The PBPK model divided the lung airways into 24 generations ( generation 1-16: tracheobronchial region; 17-24: alveolar region), with each generation further divided into ELF, epithelium, and interstitium compartments.
PBPK models were initially developed using rat plasma and ELF PK data and subsequently scaled and validated against plasma and ELF PK in patients with VAP. Drug deposition into each generation was predicted by typical-path models [10]. Total lung deposition fractions were based on values reported for the nebulisers used in the original clinical studies: 13% for gentamicin (vibrating mesh nebuliser) [11] and 45% for amikacin (breath-synchronized vibrating mesh nebuliser) [12] in intubated patients.
Infection sites of established pneumonia were defined as alveolar ELF and alveolar interstitium [13]. The gentamicin PBPK model was used to simulate apramycin PK given their PK similarities [14]. Predicted infection-site concentrations were connected to pharmacokinetic/pharmacodynamic (PK/PD) models to simulate bacterial killing under IV, nebulised, and combination regimens. All regimens were administered every 24 hours. To reflect clinical scenarios, we simulated nebulisation delivery conditions for non-intubated HAP and intubated VAP patients using a vibrating mesh nebuliser. All PBPK parameters were assumed identical between HAP and VAP, except for total lung deposition fractions: 34% for HAP [15] and 13% for VAP [11], applied uniformly across gentamicin, amikacin, and apramycin. Treatment success was defined as bacterial stasis at 24 hours. Plasma exposures (cumulative AUC and steady-state trough) served as surrogates for systemic toxicities, including nephrotoxicity and ototoxicity.
Results:
PBPK models well described plasma and ELF PK after IV bolus and intratracheal dosing of amikacin and gentamicin in rats, with most predictions within two-fold of observations. After model translation, human PBPK models captured ELF and plasma PK of nebulised gentamicin and amikacin in intubated VAP patients.
PK/PD simulations indicated that nebulisation alone at the standard IV dose (gentamicin 7 mg/kg, amikacin 30 mg/kg) was less effective than IV monotherapy for bacterial killing at 24 hours at the infection sites. To achieve efficacy equivalent to IV monotherapy, nebulised gentamicin required doses of 20 mg/kg for HAP and 60 mg/kg for VAP, while amikacin required 70 mg/kg for HAP and 170 mg/kg for VAP. Combining lower nebulised doses (e.g., gentamicin 15 mg/kg or amikacin 45 mg/kg for VAP) with 50% of the standard IV dose achieved treatment success against Pseudomonas aeruginosa and Staphylococcus aureus. For apramycin, simulated nebulised regimens showed limited additional bacterial killing against Pseudomonas aeruginosa compared to IV alone.
High-dose gentamicin nebulisation (20 mg/kg HAP; 60 mg/kg VAP) yielded higher troughs (0.69 and 0.60 mg/L) than IV 7 mg/kg (0.38 mg/L), yet AUC0-24h was lower with nebulisation (82.5 and 71.9 vs 85.6 h·mg/L) and even lower with halved IV plus nebulisation combinations. All simulated trough levels remained below established toxicity thresholds. The same trend was observed for amikacin.
Conclusion:
Our findings highlight the potential of translational PBPK/PD models to inform dosing for inhaled aminoglycosides in pulmonary infections. These models support the design of regimens that balance target-site efficacy with safety. Although simulated plasma exposures were within accepted limits across proposed regimens, the comparative safety of high-dose nebulised versus standard IV aminoglycosides has not been established and requires prospective investigation. Future clinical studies are needed to confirm these predictions and inform clinical guidelines.
References:
1. Gamazo JJ et al. Rev Esp Quimioter. 2023; 36(Suppl 1):9–14.
2. Wenzler E et al. Clin Microbiol Rev. 2016; 29(3):581–632.
3. Kalil AC et al. Clinical Infectious Diseases. 2016; 63(5):e61–e111.
4. Ehrmann S et al. Massachusetts Medical Society; 2023; 389(22):2052–2062.
5. Kollef MH et al. Chest. 2017; 151(6):1239–1246.
6. Stokker J et al. Intensive Care Med. 2020; 46(3):546–548.
7. Rouby J-J et al. The Lancet Infectious Diseases. Elsevier; 2020; 20(7):778–779.
8. Rouby J-J et al. Anaesth Crit Care Pain Med. 2020; 39(2):179–183.
9. Wen H et al. CPT:PSP. 2025; 14(4):796–806.
10. Boger E et al. CPT:PSP. 2018; 7(10):638–646.
11. https://www.hamilton-medical.com/dam/jcr:95d0c4cb-5f48-420d-b1d9-f9808cdacbfa/Aerogen-Pro_brochure_en.pdf
12. Luyt C-E et al. Crit Care. 2009; 13(6):R200.
13. Paal M et al. Antimicrobial Agents and Chemotherapy. 2021; 65(12).
14. Sou T et al. Clinical Pharmacology & Therapeutics. 2021; 109(4):1063–1073.
15. Dugernier J et al. Pharm Res. 2017; 34(2):290–300.
Reference: PAGE 34 (2026) Abstr 11899 [www.page-meeting.org/?abstract=11899]
Poster: Drug/Disease Modelling - Infection