Pirana

Parametric population PK modeling and simulation was performed using NONMEM (version 7.2, ICON Development Solutions, Ellicott City, MD, USA), Intel Visual Fortran Compiler XE 14.0 (Santa Clara, CA, USA), RStudio (version 1.1.456; RStudio, Boston, MA, USA, 2018), R (version 3.5.1; R foundation, Vienna, Austria, 2018), XPose (version 4.6.1; Uppsala University, Department of Pharmaceutical Biosciences, Uppsala, Sweden, 2018), PsN (version 4.6.0; Uppsala University, Department of Pharmaceutical Biosciences, Uppsala, Sweden, 2016), and Pirana [27 (link)] (version 2.9.4; Certara, Princeton, NJ, USA, 2018). The fT _{> MIC} and PTA were calculated using Excel 2013.
Nonparametric population PK modeling, simulation, and calculation of fT _{> MIC} and PTA was performed using Pmetrics version 1.5.2 (Laboratory of Applied Pharmacokinetics and Bioinformatics, Los Angeles, CA, USA) [28 (link)], Intel Visual Fortran Compiler XE 14.0 (Santa Clara, CA, USA), RStudio (version 1.1.456), and R (version 3.5.1). The raw VPC data were imported from Pmetrics into PsN (version 4.6.0) using the Pirana interface [27 (link)] to generate VPCs with the same layout as NONMEM. VPC plots were subsequently created using XPose (version 4.6.1) within RStudio (version 1.1.456).

Data management, plotting, and postprocessing of results were performed in R (R Foundation for Statistical Computing, Vienna, Austria), partially using the Xpose package (Department of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden). The modeling and simulations were performed in NONMEM 7.4 (Icon Development Solutions, Ellicott City, Maryland), aided by PsN (Department of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden) and Pirana (Certara, Princeton) [24 (link)].

The population pharmacokinetic model was developed using nonlinear mixed effect modeling (NONMEM version 7.4.4) software provided by Icon Clinical Research LLC, New York, NY, USA, along with the Perl-speaks-NONMEM (PsN) toolkit (Lindbom et al., 2005 (link)). The execution and management of the model as well as report generation were performed with the aid of Pirana (Certara, Princeton, NJ, USA) (Keizer et al., 2011 (link)). The actual process of modeling began with the development of one-compartment and two-compartment base models using the first-order conditional estimation method with interaction (FOCE-I) to obtain pharmacokinetic parameter estimates without any covariates. The allometric scaling function was used to evaluate the relationship between body weight and vancomycin clearance (CL) as well as volume of distribution (Vd), with a coefficient of 0.75 for CL and 1 for Vd. The exponential random-effect model was used to describe the between-subject variability (BSV) on pharmacokinetic (PK) parameters, while additive, proportional, and combined residual error models were tested to describe the residual error between the observed and predicted concentrations of vancomycin.

The population PK/PD analysis was performed with the NLME modeling program NONMEM 7.4 (ICON Plc., Dublin, Irland) (Beal et al., 2009 ). There exist several other software packages for parameter estimation of NLME models providing the same or similar algorithms. A variety of algorithms are provided in R Core Team (2019 , version 3.6.1). The software Monolix (version 2019R1. Antony, France: Lixoft SAS, 2019) and Diffmem (see https://bitbucket.org/tomhaber/diffmem/src/master/, Melicher et al., 2017 (link)) are based on the stochastic approximation expectation maximization algorithm and the recently published package Pumas (based on Julia, see https://pumas.ai/) contains several deterministic and stochastic algorithms. Standard errors were computed with the $COVARIANCE step in NONMEM. Pirana (Certara, Princeton, USA) was used for the generation of the visual predictive check with auto_bin option. The simulations in section out-of-sample validation and simulation study were performed with the ODE integrator CVodes (Sundials, Lawrence Livermore National Laboratory, Livermore) (Hindmarsh et al., 2005 (link)) interfaced to CasADi (Optimization in Engineering Center [OPTEC], K.U. Leuven) (Andersson et al., 2019 (link)).