Hazard Structure and Repair Kinetics: A 2x2 Classification that Resolves the Stochastic-Death / Individual-Tolerance Dilemma

The General Unified Threshold model of Survival, or GUTS, presents two contrasting approaches to understanding hazard structure and repair kinetics. These two variants, stochastic death and individual tolerance, differ in their predictions about no-effect concentrations under changing exposure over time. Unfortunately, no single test exists to distinguish between these two models.

Each variant of GUTS relies on two independent assumptions for classification: whether the hazard accumulates while damage remains above a certain threshold (cumulative) or is determined by the highest damage reached (peak-based), and whether repair is proportional or saturating. Remarkably, calibration does not determine which structure fits best; instead, each structure adjusts its own damage recovery. In twelve out of fourteen compounds, the cumulative hazard model proves faster in recovering damage.

To further differentiate these models, researchers introduced the concept of a carry-over ratio, which compares the survival after two successive identical exposures separated by complete recovery. If this ratio exceeds one, the cumulative hazard model is rejected. In the case of Gammarus pulex, fourteen out of fifteen carry-over ratios are at or above one, with seven ratios being significantly above one. This indicates that the cumulative hazard model may be more appropriate in these instances.

On the other hand, the peak-based hazard model predicts a maximum ratio of survival when intervals between exposures are extended in five out of six compounds. This suggests that for longer intervals, the peak-based hazard model could provide a more accurate prediction of survival.

Finally, when considering organophosphates, the peak-based hazard model emerges as the preferred choice for toxicokinetic-toxicodynamic risk assessment. This preference is based on a single test involving measured damage after diazinon exposure in Daphnia magna. After the inhibition of acetylcholinesterase, which was still present, the immobilization stopped.

This observation aligns with the predictions made by the peak-based hazard model, making it the appropriate default for assessing the risks associated with narcotics and uncouplers.