Integration of individual preclinical and clinical anti-infective PKPD data to predict clinical study outcomes

Abstract

The AIDA randomized clinical trial found no significant difference in clinical failure or survival between colistin monotherapy and colistin-meropenem combination therapy in carbapenem-resistant Gram-negative infections. The aim of this reverse translational study was to integrate all individual preclinical and clinical pharmacokinetic-pharmacodynamic (PKPD) data from the AIDA trial in a pharmacometric framework to explore whether individualized predictions of bacterial burden were associated with the trial outcomes. The compiled dataset included for each of the 207 patients was (i) information on the infecting Acinetobacter baumannii isolate (minimum inhibitory concentration, checkerboard assay data, and fitness in a murine model), (ii) colistin plasma concentrations and colistin and meropenem dosing history, and (iii) disease scores and demographics. The individual information was integrated into PKPD models, and the predicted change in bacterial count at 24 h for each patient, as well as patient characteristics, was correlated with clinical outcomes using logistic regression. The in vivo fitness was the most important factor for change in bacterial count. A model-predicted growth at 24 h of ≥2-log₁₀ (164/207) correlated positively with clinical failure (adjusted odds ratio, aOR = 2.01). The aOR for one unit increase of other significant predictors were 1.24 for SOFA score, 1.19 for Charlson comorbidity index, and 1.01 for age. This study exemplifies how preclinical and clinical anti-infective PKPD data can be integrated through pharmacodynamic modeling and identify patient- and pathogen-specific factors related to clinical outcomes - an approach that may improve understanding of study outcomes.

FIGURE 1

Illustration of the inputs and workflow of the analysis. Blue rectangles denote published data used as input for models. Blue diamonds denote published models. Orange ovals denote predicted data from developed models. The orange diamond represents the developed multivariable logistic regression model. Superscripts refer to the source from which data or models were extracted and match the reference number in the bibliography.

FIGURE 2

Box plot visualizing the distribution of the individual drug interaction (INT_i) estimates. Estimation was performed with parameters for both the AB1845 and AB2092 strains.

FIGURE 3

Change in bacterial density over time, predicted for patients in the colistin monotherapy arm (top rows) and colistin–meropenem combination therapy arm (bottom rows) using model averaging. (a) Isolates grouped by colistin MIC (columns). (b) Isolates grouped by their fitness (columns) in the mouse thigh infection model, i.e., the change in bacterial density measured at 24 h in untreated mice.

FIGURE 4

Predicted probabilities of clinical failure at day 14 by logistic regression when predicted change in bacterial density at 24 h post first CMS dose was <2‐log₁₀ growth (top row) and ≥2‐log₁₀ growth (bottom row). The probabilities of clinical failure at day 14 are plotted against SOFA score (a and d), Charlson comorbidity index (b and e), and age (c and f). The variables that are not illustrated in a panel were set to their median in the population (e.g., SOFA score = 6 and Charlson comorbidity index = 2 in the rightmost panel). The gray areas are the 95% confidence intervals around the prediction.

FIGURE 5

Logistic regression predicted probabilities of death at day 14 plotted against SOFA score (a), Charlson comorbidity index (b), and age (c). Variables that are not illustrated in a panel were set to their median in the population (e.g., SOFA score = 6 and Charlson comorbidity index = 2 in the right panel). The gray areas are the 95% confidence intervals around the predictions.