Automatic adjustment of manually measured QT intervals in digital electrocardiograms improves precision of electrocardiographic drug studies

Drug investigation studies require similar patterns of T/U waves to be measured at the same fiducial points. However, such a systematic measurement is difficult to achieve with manual measurement as compulsory for regulatory review. A new system was developed that adjusts manually measured QT intervals in sets of ECGs obtained from the same drug study. For a given threshold w, the system organizes all the repolarisation patterns in groups of ECG leads that correlate each with each other with abs(r)>w. Within these groups, the system assigns a shift of the measured T wave end in each ECG lead to synchronise the patterns of the manual measurements. The same approach is used to align the QRS onset measurements. The system was used in a drug-study of 4090 ECGs obtained in healthy volunteers off- and on-treatment with an investigational drug. The difference between manual measurements and the measurements adjusted by the new software was characterised by calculating the regression residua between QT and RR intervals and by comparing the QT dispersion. After applying the system, the regression residua were reduced from 15.06 to 14.32 ms (p<0.002). The QT dispersion values were reduced from 38.3 (SD=172) to 30.2 (SD=15.4) ms (p<10^-200). Automatic adjustment of manual measurements of QT intervals in electronically recorded ECGs improves the precision of QT interval assessment in electrocardiographic drug studies.