Detection and impact of hysteresis when evaluating a drug's QTc effect using concentration-QTc analysis

Early-phase studies quantifying the QTc prolongation potential for a new drug often use linear concentration-QTc (C-QTc) models, assuming no delay between plasma concentrations and QTc changes. However, that assumption is not always correct. The term "hysteresis" has been utilized to describe a time lag present between a measurable concentration and a measurable effect. To detect and quantify hysteresis and its impact on study interpretation, studies with hysteresis of 0.25-4 h were simulated with different doses, half-lives, and sampling schedules in a crossover design. Hysteresis was quantified using a novel method termed exposure-normalized GRI (enGRI), a proposed modification of the Glomb-Ring Index (GRI), to account for delay and magnitude of QTc effects. With realistic sampling, the rate of false negative studies (FN) increased proportionally to the delay, even for delays shorter than 1 h. Using an enGRI threshold (γ) of 2 ms resulted in FN with undetected delay and FN without hysteresis at approximately the same rate. For γ = 2 ms, the specificity of enGRI was > 90% throughout the investigated scenarios. We therefore propose the incorporation of enGRI when interpreting results from C-QTc analysis with the intent of characterizing QTc effects.