Frequency of QT recording and the reliability of QT prolongation assessments

Abstract

With the known high intra-subject chaotic, circadian and day-to-day variabilities of QT intervals, we investigated the frequency (n) of QT replications per time point, T (or per dose/concentration) and the corresponding detectable limit for QT prolongation assessment (QTPA). Baseline only QT data were obtained from several prospectively designed QTPA trials in healthy young male and/or female subjects (N≥40 in all trials), with n up to 6 at each of the multiple T per day (intervals ranged from minutes to hours) for 2-6 treatment periods. For each trial, 1 to 6 Jackknife randomly re-sampled QTs were drawn from each of the T and the means, standard deviation (SD), changes of means from T to T, and rate of QT outliers were determined. The inter-/intra- subject QT variabilities (ITERV, ITRAV) and the corresponding lowest detectable limit for a reliable QTPA were estimated from 1000 Jackknife datasets. Results show that with n = 1 to 6 per T, the ITRAV SD nonlinearly decreased from 6.82±4.31 (range 1.52-13) to 3.08±1.85 (range 0.44–8.35) ms, respectively. At any n, the mean absolute QT changes from T to T never equaled zero, with the min/max ranged 0.8/49 ms., respectively. When n=1, mean absolute change is 16±6.64. The ITERV SD is less affected by the n when N≥40. At best, the rate of QT changes >30 ms is about 3%, and that of QT changes >60 ms about 1% from beat-to-beat, T to T, or day to day. This study suggests n >3 per T is the limit for a reliable QTPA. Clinical Pharmacology & Therapeutics (2004) 75, P48–P48; doi: 10.1016/j.clpt.2003.11.182