The translation of hERG block and APD90 in vitro to clinical QTc prolongation in man

Abstract

The core ICH S7B in vitro assay is a hERG assessment which has associated best practice recommendations. Three reference agents are being used to compare margins based on concentrations associated with 10 ms QTc prolongation in man. The objective of the current analyses was to compare the margins for best practice hERG patch-clamp data with those for other in vitro assessments. The unbound critical concentrations shared as part of the Training Materials were used as the denominator throughout. The numerators were: the hERG IC₅₀ generated in a study conforming to current best practice guidance, the pKi from published dofetilide-binding data, the concentrations associated with either published rate corrected APD₉₀ prolongation in isolated IPD-derived human cardiomyocytes, or with a 25 ms change in APD₉₀ in human ventricular trabeculae tissue. The respective margins for dofetilide, moxifloxacin and ondansetron in the patch-clamp study were 45, 32 and 4.2. The dofetilide-binding displacement pKi, gave margins of 11, 15 and 10. In isolated cardiomyocytes the margins were 0.5, 5.2 and 1.3. In human trabeculae tissue the margins were 14, 4.5 and 9. There is an underlying assumption that for selective hERG blocking drugs the amount of hERG block (and associated hERG margin) would be similar for the same 10 ms change in QTc. The consistent margin determined using the dofetilide-binding displacement pKi and the concentration associated with a 25 ms APD90 prolongation in human trabeculae would tend to support that assessment. There was more variability in isolated cardiomyocytes where there was an 11-fold difference between dofetilide and moxifloxacin margins while the ondansetron margin lay in between these margins. The hERG patch-clamp study had skewed variability with a margin for ondansetron 11-fold and 8-fold different from the margins for dofetilide and moxifloxacin, respectively. The small margin for ondansetron caused wider confidence intervals for the pooled margin of 18-fold. Overall, these data suggest that any of these assays might realistically be used to predict a QTc prolongation in man. The hERG patch-clamp study is the required study. Where there are ‘ondansetron-like’ blocking characteristics an additional assay may be warranted to clarify the prediction.