Thorough QT studies are typically conducted for drugs with systemic bioavailability and include a positive control, typically moxifloxacin, with a well-described QTc effect. This study tested two hypotheses: that (i) re-measuring the QT intervals based on electrocardiogram (ECG) pattern similarity improves the moxifloxacin time profile, and (ii) that study conduct influences the ability to detect a typical moxifloxacin time profile. ECGs from 65 studies with available moxifloxacin plasma concentrations were obtained, including four studies with an unexpected moxifloxacin response. Residual error of a concentration-QT model was evaluated before and after re-measuring the QT interval based on ECG pattern similarity. Intra-replicate heart rate differences were calculated using the original heart rate measurements and the 10-s average heart rates. Similarity re-measurements reduced the residual error of the model (before vs. after of 8.43 ± 2.00 vs. 7.55 ± 1.86 ms; p < 0.001). For both original and averaged 10-s heart rate, intra-replicate heart rate differences were significantly lower (p < 0.001) in studies with the expected response than in those with an unexpected time profile. The pattern similarity measurement methodology reduces the residual error of the model, which influences the time profile of the moxifloxacin response. Accuracy of study conduct, represented by intra-replicate heart rate differences, separated studies with and without the expected moxifloxacin time profile.
Author(s): Lars Johannesen, Christine Garnett, Marek Malik