Concentration-QTc modeling provides consistent high QTc sensitivity across species and study designs

Abstract

To support an integrated risk assessment for QTc prolongation under the ICH E14/S7B Q&As, the in-vivo QTc assay should follow best practice for demonstrating high drug exposure multiples and sufficient QTc detection sensitivity. Concentration-QTc (C-QTc) modeling can improve study power compared with a traditional by-time analysis and can be used to quantify drug concentration producing a 10 msec effect and any shift between human and pre-clinical species.

The objective was to compare by-time analysis and C-QTc modeling sensitivity in the 3 large animal species. ICH E14/S7B Q&A compliant in-vivo QTc validation studies were performed with oral moxifloxacin using both a Latin square and an ascending dose design in dog (n = 4 for Latin square and n = 6 for ascending dose design), NHP (n = 4), and minipig (n = 8 for Latin square and n = 5 for ascending dose design). For each study, a separate full PK phase was included.

For Dog and NHP modeling, basic assumptions were met for the use of a linear mixed effect (LME) model. Least significant difference (LSD) values for LME modeling in dog and NHP for the Latin square design were 3.4–3.5 msec versus 10 msec for the by-time analysis. For the ascending design, LME LSD values were 2.5–2.6 msec for dog and 3.7–4.4 msec for NHP versus respective values of 5.4 msec and 6.6 msec for the by-time analysis. Human multiple of the moxifloxacin critical concentration for dog and NHP was respectively 1.8× and 1.5× for the Latin square and 2.2× and 2.3× for the ascending design. For minipig modeling, basic assumptions for LME were not met, notably due to the high dose effects and a curvilinear C-QTc relationship. An Emax model was used producing LSD values of 3–5 msec versus 10–11 msec for the by-time analysis, and human multiple of 0.5–0.7×. When the high dose was omitted to apply an LME model, LSD values for the Latin square design were 3.9–4.3 msec with a human multiple of 0.5×.

Moxifloxacin C-QTc modeling provides consistent LSD sensitivity of 3–5 msec for all 3 species and both study designs, meeting requirements for ICH E14/S7B Q&As 5.1 and 6.1.