Physiologically based pharmacokinetic modelling of acute digoxin toxicity and the effect of digoxin-specific antibody fragments

Abstract

Context: Recommended doses of digoxin-specific antibody fragments (digoxin-Fab) for treatment of acute digoxin poisoning are pharmacokinetically unsubstantiated and theoretically excessive. Physiologically based pharmacokinetic (PBPK) modelling creates clinical simulations which are closely related to physiological and pharmacokinetic behaviour. This paper details the formulation of a PBPK model of digoxin and explores its use as a simulation tool for acute digoxin toxicity and its management.

Materials and methods: A PBPK model of digoxin was constructed and validated for acute digoxin poisoning management by comparing simulations with observed individual acute overdose patients. These simulations were compared with standard two-compartment PK model simulations.

Results: PBPK model simulations showed good agreement with post-absorption plasma concentrations of digoxin measured in 6 acute overdose patients. PBPK predictions were accurate to 1.5-fold or less of observed clinical values, proving to be more accurate than two-compartment simulations of the same patients which produced up to a 4.9-fold change.

Conclusions: Compared to conventional two-compartment modelling, PBPK modelling is superior in generating realistic simulations of acute digoxin toxicity and the response to digoxin-Fab. Simulation capacity provides realistic, continuous data which has the potential to substantiate alternative, less expensive, and safer digoxin-Fab dosing strategies for the treatment of acute digoxin toxicity.

Keywords:

• digoxin
• acute
• toxicity
• physiologically based pharmacokinetic modelling
• digoxin-Fab

Acknowledgements

Thank you to Sibylle Neuhoff for providing us with additional parameters to be used in our PBPK model.

Disclosure statement

The authors have no declaration of interest to report. The authors alone are responsible for the content and writing of the paper.

Additional information

Funding

This research was partially supported by an NHMRC Program Grant 1055176.