Data Quality in Small QTc Studies. Marek Malik, London

Transcription

1 Data Quality in Small QTc Studies Marek Malik, London

2 Data Quality in Small QTc Studies DISCLAIMER Views and opinions presented are only my own

3 Data Quality in Small QTc Studies The E14 Guidance on conducting thorough QT studies needs revisiting. Different replacement and/or modification concepts have been proposed. Most expect the QT/QTc investigations to be moved into the very early clinical tests, such as the SAD and MAD FIM studies. This presentation is devoted to some aspects that should make the QT/QTc investigations as part of FIM studies practical and viable.

4 Data Quality in Small QTc Studies Electrocardiographic challenges of small clinical studies such as the first-in-man investigations Metrics of QTc data quality, power considerations of small studies Replacement and/or avoidance of standard positive control in small QTc studies and in conc-qtc analyses Approaches to PK/PD QTc modelling with the distinction of model types Reducing QTc variability in small studies

5 Data Quality in Small QTc Studies Electrocardiographic challenges of small clinical studies such as the first-in-man investigations Metrics of QTc data quality, power considerations of small studies Replacement and/or avoidance of standard positive control in small QTc studies and in conc-qtc analyses Approaches to PK/PD QTc modelling with the distinction of model types Reducing QTc variability in small studies

6 Data quality Small clinical studies

7 Small clinical studies Automatic ECG measurement

8 Error of an automatic measurement by a modern machine [ms] Average error [ms] (mean +SD) Treatment bias in automatic measurement QTcF in a quionolone study courtesy of Dr Sarapa 14 Placebo Treatment A 12 Treatment B P = 4 x P = 2 x Placebo Trearment A Treatment B Global median manual QT measurement [ms]

9 QT interval Small clinical studies Drug-induced heart rate changes Predose Tmax RR interval

10 QT interval QT interval QT interval Small clinical studies Drug-induced heart rate changes RR interval RR interval RR interval

11 QT interval QT interval QT interval Small clinical studies Drug-induced heart rate changes RR interval RR interval RR interval

12 QT interval [ms] QT interval [ms] QT interval [ms] QT interval [ms] Small clinical studies Baseline 1 Baseline Baseline 3 Baseline RR interval [ms] 320 Baseline 1 Baseline Baseline 3 Baseline RR interval [ms] Baseline 1 Baseline Baseline 3 Baseline RR interval [ms] 320 Baseline 1 Baseline Baseline 3 Baseline RR interval [ms]

13 Data Quality in Small QTc Studies Electrocardiographic challenges of small clinical studies such as the first-in-man investigations Metrics of QTc data quality, power considerations of small studies Replacement and/or avoidance of standard positive control in small QTc studies and in conc-qtc analyses Approaches to PK/PD QTc modelling with the distinction of model types Reducing QTc variability in small studies

14 Required number of subjects in a X-over study Standard deviation of QTc difference between placebo and period baseline [ms] Power of small TQT studies Comparisons 24 comparisons 40 comparisons s = 5.5 ms Time of the day [h] Expected QTc difference between Rx and placebo [ms]

15 Precision in QTc studies Metrics for comparison of data quality Intra-subject standard deviation of QTc on placebo

16 Data Quality in Small QTc Studies Electrocardiographic challenges of small clinical studies such as the first-in-man investigations Metrics of QTc data quality, power considerations of small studies Replacement and/or avoidance of standard positive control in small QTc studies and in conc-qtc analyses Approaches to PK/PD QTc modelling with the distinction of model types Reducing QTc variability in small studies

17 Assay sensitivity in QTc studies Electrocalrdiogram??? II Electrocardiogram??? V2 correlation coefficient of descending T waves difference 106 ms

18 Assay sensitivity in QTc studies Shift

19 Assay sensitivity in QTc studies

20 Assay sensitivity in QTc studies Reproducibility between subjects on placebo: Distribution of values (A t1 B t1 ) (A t2 B t2 ) where X t is QTc measured in placebo subject X at time t, for all pairs of placebo subjects A and B (A B) and all pairs of time-points t1 and t2 (t1 t2).

21 Data Quality in Small QTc Studies Electrocardiographic challenges of small clinical studies such as the first-in-man investigations Metrics of QTc data quality, power considerations of small studies Replacement and/or avoidance of standard positive control in small QTc studies and in conc-qtc analyses Approaches to PK/PD QTc modelling with the distinction of model types Reducing QTc variability in small studies

22 PK/PD models

23 PK/PD models

24 PK/PD models Linear Sigmoid E max E max Logarithmic

25 Data Quality in Small QTc Studies Electrocardiographic challenges of small clinical studies such as the first-in-man investigations Metrics of QTc data quality, power considerations of small studies Replacement and/or avoidance of standard positive control in small QTc studies and in conc-qtc analyses Approaches to PK/PD QTc modelling with the distinction of model types Reducing QTc variability in small studies

26 Reduction of QTc variability Submitted for publication on 6 th February 2015

27 APD/HR hysteresis Franz MR at al, J Clin Invest 1988; 82:

28 I QT/RR hysteresis II I II III avr avl avf V1 V2 V3 V4 V5 V6 III avr avl avf V1 V2 V3 V4 V5 V6

29 QT/RR hysteresis interval based interval based time based

30 Contribution of individual RR intervals QT/RR hysteresis 3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% History of QT reading [cardiac cycles]

31 QT/RR hysteresis

32 QT/RR hysteresis

33 QT/RR hysteresis

34 QT/RR hysteresis

35 Correlation coefficient RR vs QTc=QT/RRa Correlation coefficient RR vs QTc=QT/RRa QT/RR hysteresis Correction coefficient a Correction coefficient a Fridericia correction

36 QT/RR hysteresis

37 QT/RR hysteresis

38 QT/RR hysteresis

39 QT/RR hysteresis It is pointless to try to stabilize HR by supine positions Correcting the QT intervals to the hysteresis modeled RR leads to substantial reduction of the QTc variability Correcting QT interval for simultaneously measured RR interval or an average of small number of RR intervals should be avoided Study analysis windows may include variable HR episodes

40 QT/RR hysteresis Preservation of the time course for hysteresis could serve the role of the positive control in TQTS, i.e., verification that the QT and RR data are adequate to detect clinically relevant repolarization effects of drugs where they exist.

41 Thank You