Abstract
Quizartinib prolongs QT interval through inhibition of the slow delayed rectifier potassium current (I(Ks)). We used non-linear mixed-effects modeling to explore the relationship between quizartinib and its pharmacologically active metabolite AC886 and the Fridericia-corrected QT interval (QTcF) in newly diagnosed acute myeloid leukemia (AML) patients. We evaluated linear and non-linear drug effect models, using triplicate QTcF measurements with available time-matched pharmacokinetic samples from the Phase 3 QuANTUM-First trial. The effect of intrinsic and extrinsic factors on model parameters was tested using stepwise covariate model building. Simulations were conducted to predict the change from baseline in QTcF (ΔQTcF) at the maximum concentration at steady-state (C(max,ss)) for quizartinib maintenance daily doses of 30 and 60 mg. The concentration-QTcF (C-QTcF) relationship was best described by a sigmoidal maximum effect model. After accounting for the effect of quizartinib, including AC886 concentrations did not further explain changes in QTcF. Circadian variations in QTcF were described using an empirical change from baseline based on clock times. Age and hypokalaemia were identified as statistically significant covariates on baseline QTcF; no covariates were found to impact the C-QTcF relationship. The median model-predicted ΔQTcF at C(max,ss) was 18.4 ms (90% confidence interval (CI): 16.3-20.5) at 30 mg and 24.1 ms (90% CI: 21.4-26.6) at 60 mg. In conclusion, in newly diagnosed AML patients, ΔQTcF increased non-linearly with increasing quizartinib concentrations. The predicted ΔQTcF increase at C(max,ss) supports the proposed dose adaptation based on observed QTcF and the dose reduction in case of strong cytochrome P450 3A (CYP3A) inhibitors coadministration.