Grigory Nesvijevski 1, Friederike L. Born 2, Benjamin Sellner 3, Edgar I. Campos-Madueno 2, Petra S. Dittrich 2, Sven Panke 2, Elisabet I. Nielsen 1, Lena Friberg 1
1 Uppsala University (Uppsala , Sweden), 2 Federal Institute of Technology Zurich (Basel, Swtizerland), 3 University of Basel (Basel, Switzerland)
INTRODUCTION
Antimicrobial resistance continues to rise globally [1], yet research and development in anti‑infective therapeutics face persistent hurdles, reflected in a marked slowdown in the approval of novel‑class antibiotics [2]. The NCCR AntiResist initiative argues that physiologically relevant assays, those that reproduce in vivo‑like conditions, represent an underexplored but promising strategy to accelerate the discovery of effective antimicrobial interventions [3]. The current work focuses on urinary tract infections caused by Escherichia coli. This pathogen displays substantial clinical manifestation heterogeneity [4], including urinary infections. E. coli is the most prevalent causative agent and exhibits notable resistance to multiple critical antimicrobials, including a global resistance rate of 39.4% to ciprofloxacin [1] Here, we developed models to describe nitrofurantoin and ciprofloxacin concentrations in urine, and their implementation in axenic in vitro systems, specifically the µ‑hollow‑fibre platform, for assessing antimicrobial efficacy.
OBJECTIVES
Explore physiology-based model structures describing the urine voiding mechanism, impactful parameters such as volume synthesis speed, voiding interval and minimal residual volume, and their impact on predicted drug-of-interest exposure.
Describe E. coli behaviour, their growth and killing kinetics in in vitro experiments, when undergoing urinary-relevant drug-of-interest exposures.
METHODS
Published pharmacokinetic population models or non-compartmental data [5][8], combined with renal excretion fractions [6][7][9] were used to estimate changes in drug amounts over time. Urinary concentration profiles were generated by implementing a simulated time-cyclic bladder volume using mrgsolve [11]. For nitrofurantoin, the simulation was performed for repeated oral administration of 100 mg q12 daily and 50 mg q6 daily for 5 days. For ciprofloxacin, simulation was performed for IV administration of a 400 mg dose q12 for 7 days. In vitro experiments were performed, where bacteria were exposed to the predicted ciprofloxacin concentration-time profile in the bladder in a µ-hollow-fibre system.
RESULTS
For nitrofurantoin, prediction of 100 mg q12 with a urine synthesis rate of 2 L per day, 50 mL of residual urine, and micturition every 4 hours, reached an AUC of 1084 mg · h/L, in line with the observed [5] mean of 943.49 [228.2–2212.0] mg · h/L. For Cmax, we simulated a value of 63.8 mg/L, matching the observed range (mean of 94.1 [40.1-209.4] mg/L).
Prediction of 50 mg q6 yielded an AUC of 1076 mg · h/L and a Cmax of 49.1 mg/L, in line with the observed data: 855.95 [378.6-2098.9] mg · h/L and 94.4 [26.8-176.3] mg/L. Predicted Tmax was 4.5 hours, in line with the observed Tmax of 5.1 [3.3 – 5.5] hours and 6.8 [1.8 – 8.1] hours for both dosing regimens.
For ciprofloxacin, prediction of IV 400 mg q12 with a urine synthesis rate of 1.5 litres per day, 50mL of residual urine, and micturition every 4 hours, reached 480.22 µg/mL at 1.42 hours. This simulation aligned with observed concentrations after oral administration of 500 mg, [10], with a maximal concentration observed in the 0-2 hours bracket at 367 (242-650) mg/L and 394 (162-681) mg/L in the 2-4 hours bracket.
Qualitatively, after reaching their respective maximum, profiles show a continuous decrease in concentration driven by urine dilution, which was accentuated by the urine synthesis rate. During micturition, concentration remained unchanged, but the residual volume after micturition determined the steepness of dilution post-micturition. The micturition interval affected the accumulation of the drug, combined with residual volume.
The ciprofloxacin bladder concentration-time profile was mimicked in a µ-hollow-fibre experimental setup, where the high concentrations compared to plasma exposures exhibited rapid and pronounced bacterial killing with no regrowth, whereas the untreated control chambers showed growth throughout the experiment.
CONCLUSIONS
This simulation framework shows potential for predicting urinary concentration profiles by integrating urine synthesis rate, residual urinary volume, and micturition interval with published pharmacokinetic data. Mimicking the profile in an experimental setting demonstrated the suitability of axenic models for replicating complex drug profiles and showed that in vivo urinary persistence is unlikely to be due solely to insufficient drug activity. This work lays the groundwork for investigating the translatability of bladder-mimicking in vitro experiments.
References:
[1] GLASS report 2025, WHO https://www.who.int/publications/i/item/9789240116337
[2] Challenges of Antibacterial Discovery, Lynn L. Silver
[3] Revitalizing antibiotic discovery and development through in vitro modelling of in-patient conditions, Julie Solier
[4] Pathogenic Escherichia coli, James B. Kaper
[5] The pharmacokinetics of nitrofurantoin in healthy female volunteers: a randomised crossover study, Angela Huttner
[6] Nitrofurantoin: properties and potential in treatment of urinary tract infection: a narrative review, Mahdizade
[7] Pubchem https://pubchem.ncbi.nlm.nih.gov/compound/6604200#section=ATC-Code
[8] Optimizing ciprofloxacin dosing in intensive care unit patients through the use of population pharmacokinetic-pharmacodynamic analysis and Monte Carlo simulations, David Khachman
[9] Transintestinal elimination of ciprofloxacin, R.Rohwedder
[10] The pharmacokinetics and serum and urine bactericidal activity of ciprofloxacin, David C. Brittain
[11] https://mrgsolve.org/user-guide/
Reference: PAGE 34 (2026) Abstr 12166 [www.page-meeting.org/?abstract=12166]
Poster: Drug/Disease Modelling - Infection