Demonstrating the statistical and pharmacological sensitivity of nonclinical QTc analysis using a dofetilide dose–response in nonhuman primates

Abstract

Nonclinical QTc studies can augment clinical QTc assessments in regulatory submissions provided they are of sufficient quality and sensitivity. Both the statistical performance and species translation play a role in determining the sensitivity of the model.

The current analyses examine the effects of dofetilide or vehicle on the QT interval in nonhuman primate (NHP; n = 16) using a one-step estimated marginal means method where both treatment and animal ID are used in regression models to avoid a separate rate correction step, in comparison to other commonly utilized methods. The doses of dofetilide were chosen to span a threshold dose with exposure only just exceeding the concentration associated with 10 ms QTc prolongation in man, to a dose where exposures exceed the Emax for QTc prolongation. The primary objective was an evaluation of which doses and exposures can be detected as eliciting a statistically significant change in QTc.

A group size of 8 for cross-over analysis was insufficient to detect, as statistically significant, the effects of the threshold dose of 0.01 mg/kg dofetilide using common correction and statistical analysis methods and hourly time intervals. Higher doses were all detected as causing a statistically significant effect using the same techniques. The ‘One-Step’ method was able to detect as statistically significant effects at all doses of dofetilide across a wide range of time and exposure. There were also temporal differences between the mean effects observed using the common and ‘One-Step’ methods. Preliminary concentration-QTc assessment suggests a higher maximum prolongation in concentration QTc with the ‘One-Step’ method. Furthermore, this analysis suggests that at exposures associated with a 10 ms QTc prolongation in man a 10 ms prolongation is also observed in NHP. The observed ED₅₀ concentration (0.85 ng/ml unbound) is close to that described in man (0.98 ng/ml).

These analyses demonstrate the statistical sensitivity of the ‘One-Step’ method of QTc assessment in NHP. The pharmacological sensitivity was also demonstrated and a detection threshold of 10 ms was consistent in terms of exposure between NHP and man. Overall, QTc assessment using the ‘One-Step’ method in NHP is a robust and sensitive model to supplement clinical QTc assessment.