DDMODEL00000173: Executable_ColistinMeropenem_Interaction

  public model
Short description:
The model describes the in vitro effect and interaction of colistin and meropenem on wild-type and resistant strain of P. aeruginosa. Model comprised of files: ColistinMeropenem_Interaction_simulated.csv, ColistinMeropenem_Interaction_original_simulated.lst, Command_target.txt, ColistinMeropenem_interaction_original_real.lst, ColistinMeropenem_Interaction_original_simulated.mod, ColistinMeropenem_Interaction_original.mod
Original code
  • Dynamic interaction of colistin and meropenem on a WTand a resistant
  • Mohamed AF, Kristoffersson AN, Karvanen M, Nielsen EI, Cars O, Friberg LE
  • J Antimicrob Chemother, 2/2016
  • 1Department of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden; 2Institute for Medical Research, Kuala Lumpur, Malaysia; 3Department of Medical Sciences, Section of Infectious Diseases, Uppsala University, Uppsala, Sweden
  • Objectives: Combination therapy can be a strategy to ensure effective bacterial killing when treating Pseudomonas aeruginosa, a Gram-negative bacterium with high potential for developing resistance. The aim of this study was to develop a pharmacokinetic/pharmacodynamic (PK/PD) model that describes the in vitro bacterial time–kill curves of colistin and meropenem alone and in combination for one WT and one meropenemresistant strain of P. aeruginosa. Methods: In vitro time–kill curve experiments were conducted with a P. aeruginosa WT (ATCC 27853) (MICs: meropenem 1 mg/L; colistin 1 mg/L) and a meropenem-resistant type (ARU552) (MICs: meropenem 16 mg/L; colistin 1.5 mg/L). PK/PD models characterizing resistance were fitted to the observed bacterial counts in NONMEM. The final model was applied to predict the bacterial killing of ARU552 for different combination dosages of colistin and meropenem. Results: A model with compartments for growing and resting bacteria, where the bacterial killing by colistin reduced with continued exposure and a small fraction (0.15%) of the start inoculum was resistant to meropenem, characterized the bactericidal effect and resistance development of the two antibiotics. For a typical patient, a loading dose of colistin combined with a high dose of meropenem (2000 mg q8h) was predicted to result in a pronounced kill of the meropenem-resistant strain over 24 h. Conclusions: The developed PK/PD model successfully described the time course of bacterial counts following exposures to colistin and meropenem, alone and in combination, for both strains, and identified a dynamic drug interaction. The study illustrates the application of a PK/PD model and supports high-dose combination therapy of colistin and meropenem to overcome meropenem resistance.
Bruna Torres
Context of model development: Combination Therapy Dose Selection;
Long technical model description: The study was carried out in steps: first a PK/PD model for meropenem alone was developed, then this model was combined with a PK/PD model for colistin alone (reported previously) using shared (but strain-specific) bacterial parameters for the two drugs. Static drug experiments on colistin and meropenem in combination for concentrations predicted to provide information on the interaction were conducted and the model was updated based on this data plus previously reported dynamic time–kill data replicating the first dose interval of the meropenem and colistin combination. Lastly, predictions of dosages were made from the developed model. All bacterial count data were transformed into logarithms in the data analysis. The residual error was split into two different components: one consistent difference common for all replicates at the same time point (RES) and one replicate-specific difference (RRES) to avoid bias due to correlations between the replicates. For time points where all dilutions had bacterial counts below the limit of detection (LOD) the probability for the observation to be below the LOD was estimated using the M3 method. The average tendencies in the population (typical parameter values) were estimated along with random effects described by the residual errors. No inter-experimental variability was estimated. Sampling importance resampling (SIR) was used to assess the parameter uncertainty distribution and obtain standard errors. The population analysis software NONMEM7 with the Laplacian method and ADVAN13 was used to analyse the data. The model structures consisted of two main parts: PD related compartments (bacterial system) and PK compartments (drug concentrations). In the bacterial system, the bacteria are either in a proliferating and drug-susceptible compartment (S) or in a resting and drug-insusceptible compartment (R). The bacteria multiply with a first-order rate constant in the susceptible compartment (kgrowth) and all bacteria have a first-order natural death rate (kdeath). The total bacterial content in the system (S+R) stimulates the transfer from the proliferating stage to the resting stage (kSR) as the number of bacteria increases. For PK, the drug concentrations in the compartments for colistin (C) and meropenem (M) drove the killing of the bacteria. The decline in drug concentration was determined by first-order elimination rate constants (ke). The PD model for colistin included a resistance (Re) development model with rate constants for development of resistance (kon) described by Emax functions and a rate constant for return to susceptibility (koff). The PD model for meropenem included a pre-existing resistant subpopulation with a reduced susceptibility to meropenem as a shift in concentration (ShiftM). For ARU522 the second subpopulations had a decreased growth rate. Drug effect of colistin (kdrug,C) for ATCC 27853 was described by a basic Emax model and for ARU552 by a linear model. Drug effect of meropenem (kdrug,M) for ATCC 27853 was described by a power model and for ARU552 by a sigmoid Emax model. In the predictions for colistin regimens, a two-compartment model for CMS disposition in combination with a one-compartment colistin PK model was used. For meropenem, a two-compartment PK model was applied.;
Model compliance with original publication: Yes;
Model implementation requiring submitter’s additional knowledge: No;
Modelling context description: The model was developed to characterize the pharmacokinetic/pharmacodynamic (PK/PD) relationship of colistin and meropenem alone and in combination against a WT P. aeruginosa (ATCC 27853) and a clinical meropenem-resistant isolate of P. aeruginosa (ARU552). The PK/PD model was built on static and dynamic time–kill curve in vitro data and then combined with reported PK models for colistin and meropenem and applied to show how a model-based approach could support the selection of dosages resulting in a pronounced bacterial killing.;
Modelling task in scope: estimation; simulation;
Nature of research: In vitro;
Therapeutic/disease area: Anti-infectives;
Annotations are correct.
This model is not certified.
  • Model owner: Bruna Torres
  • Submitted: Oct 10, 2016 3:55:58 PM
  • Last Modified: Oct 10, 2016 3:55:58 PM
  • Version: 12 public model Download this version
    • Submitted on: Oct 10, 2016 3:55:58 PM
    • Submitted by: Bruna Torres
    • With comment: Model revised without commit message