The translation of QTc across species - Impact of subject number and exposure multiple tested on discriminatory sensitivity

Abstract

Small, relatively insensitive studies can be useful in safety assessment when multiples of the therapeutic clinical concentration are tested. This is a fundamental principle in safety testing in animals which is equally valid for early clinical evaluations in healthy volunteers. It is often less practical to increase the number of test subjects than to increase the exposure tested. Both, when combined with the analysis method, can have an impact on the sensitivity to detect an effect. The objective is that the relationship between statistical power, analysis method, number of animals and exposure multiple explored can be illustrated using the example of QTc assessment in animals. The statistical power to detect an effect on the electrocardiogram QTc interval in nonhuman primates (NHP) for different analyses methods was known. The concentration-QTc relationship was also known for reference agents in NHP. Lastly, the critical concentration associated with a 10 ms QTc interval change in man was known for these same reference agents. This information was combined to illustrate how doubling the number of NHP used or increasing the exposure tested would support a conclusion concerning the presence or absence of an effect on the QTc interval for a test agent. In NHP, the most sensitive analysis methods have >80 % power (at p < 0.05) to detect an effect of the reference agent at the critical concentration using only 4 animals. Less sensitive techniques can detect an effect with the same power when either more animals are used or where higher multiples of the critical concentration are tested. This illustrates the principle that even with only 4 animals and an insensitive technique an effect can be detected provided higher exposures are tested. Conversely, a study using more animals, or a more sensitive analysis needn't require a higher exposure in animals to exclude an effect in man. Rather than focusing on a fixed QTc threshold sensitivity regardless of experimental design these analyses demonstrate that investigators have the flexibility to use the simplest available combination of exposures, animal numbers and analysis to achieve an effective QTc assessment.