Computerised measurements of QT dispersion in healthy subjects.

OBJECTIVE To assess the stability and reproducibility of computerised QT dispersion (QTd) measurement in healthy subjects, as this is presently being incorporated into commercial electrocardiographic systems. METHODS 70 healthy volunteers (mean (SD) age 38 (10) years, 35 men, 35 women) with a normal 12 lead electrocardiogram (ECG) were studied. From each subject, 70 ECG recordings were taken using the MAC VU ECG recorder (Marquette). In study A, 50 ECGs were recorded in each subject: 10 supine, 10 sitting, 10 standing, 10 holding breath in maximum inspiration, and 10 holding breath in maximum expiration. After a mean interval of 8 (3) days (range 7 to 23), 10 recordings in supine and 10 in the standing position were repeated in each subject (study B). On measurements made using a research version of the commercial software without manual modification, the reproducibility of QTd was assessed by coefficient of variance (CV) and relative error, and comparisons made with other ECG indices. RESULTS (1) QTd measurements were stable and not influenced by changes in posture and respiratory cycle; (2) there was no difference in QTd measurements between men and women, or between age groups dichotomised at 35 years; (3) no correlation was found between QTd and heart rate or QT interval; (4) short term reproducibility of all QTd measurements (CV 15.6% to 43.8%) was worse than that of conventional ECG indices (CV 1.4% to 5.3%); (5) long term reproducibility of QTd measurements (relative error 27.4% to 31.0%) was also worse than that of conventional ECG indices (relative error 1.8% to 7.9%) (p < 0.0001); (6) the reproducibility of QTd measurements tended to increase when several serial recordings were averaged. CONCLUSIONS Computerised measurements of global QTd and global QT-SD from 12 lead ECG by the MAC VU/QT Guard system are not significantly altered by changes in posture and respiration. The reproducibility of all QTd measurements is inferior to that of conventional ECG indices in healthy subjects.