Interaction models often need an estimate of the drug concentration available to inhibit an enzyme. When almost all of the drug is bound to plasma proteins, small uncertainty in the unbound fraction can matter.
Cannady and colleagues reported that evacetrapib’s unbound fraction in human plasma was below 0.1% and difficult to estimate. Their screening calculation used a 1% surrogate instead. That input belonged to the model. A separate study measured midazolam exposure after repeated evacetrapib dosing and found area-under-the-curve ratios of 1.19 and 1.44 at the two doses studied.
The assumed free fraction is not a laboratory measurement of free drug in a particular patient. When interpreting an interaction prediction, identify which concentrations were measured, which were assumed and which clinical data were used to test the prediction.
Cannady et al., evacetrapib interaction study with laboratory predictions and clinical testing (full text) · DOI