First Eawag toxicokinetic-toxicodynamic modeling workshop
Roman Ashauer, Eawag - Aquatic Research
The 1st Eawag1 toxicokinetic-toxicodynamic (TKTD) modeling workshop was held from 3—7 May at Eawag Kastanienbaum in Switzerland. This technical workshop aimed at exchanging know-how between researchers using and developing TKTD models for ecotoxicology, with a focus on aquatic species. TKTD models translate exposure concentrations to internal concentrations (TK) and link the internal concentration to an effect on the individual (TD).
These models are the only method to mechanistically extrapolate toxic effects, for example, between different time-variable exposure patterns, chemicals or species. There was confusion about assumptions and interpretation of existing TKTD models―such as CBR, DEBtox survival, DAM or TDM―which has hampered development in the field for years. Thus we aimed to resolve this major obstacle.
We discussed and defined state variables in TKTD models, their underlying assumptions and terminology as well as how previously disparate models relate to each other. These insights facilitated a unification of various existing TKTD models for survival within a new, more general framework. We agreed that the model output should be described by a likelihood function for survival. We are preparing this “Grand Unified Theory for Survival (GUTS)” for publication.
Sophisticated physiologically based toxicokinetic (PBTK) models may be required if internal distribution rates dominate toxicokinetics and their parameterization can be aided by in silico methods.
Experimental design was discussed from different points of view (e.g., parameter estimation, practical issues and mechanistic investigations). We concluded that the optimal test design can considerably deviate from standard test designs depending on prior knowledge and objective. Raw data from standard tests can be used but may not suffice.
If sub-lethal effects vary in time and/or depend on exposure history of the individual then a toxicokinetic model is needed. The TD may best be modeled using a DEB approach for those endpoints but not necessarily all. TD can be modeled using mechanistic or descriptive approaches. Approaches that are mechanistic in their structure and/or assumptions (e.g., DEB theory) are needed to extrapolate beyond test conditions.
TKTD models can be linked to individual based models in a straightforward way to extrapolate effects on individuals to the population level.
The unified framework serves as a reference model for survival and allows for quantitative interpretation of patterns in data that are best explained by various assumptions (e.g., stochastic death vs. individual tolerance). We believe that integration of sub-lethal effects into TKTD models requires further discussion.
Author contact information: email@example.com
Return to the Globe