Reproducibility of on-line vectorcardiography measurements in patients with and without acute ischaemic heart disease.

Abstract

The aim of the present study was to determine inter and intraobserver variations of measurements with on-line vectorcardiography (VCG). The VCG registrations were evaluated by two independent observers. One observer also evaluated the VCG registrations on two separate occasions. Monitored VCG variables were: ST vector magnitude (ST-VM), ST vector lead X (ST-X), ST change vector magnitude (STC-VM) and QRS vector difference (QRS-VD). On-line VCG was performed for 24 hours in 60 patients (10 with low probability of ischaemic heart disease, 25 with unstable angina pectoris and 25 with acute myocardial infarction). A close correlation between the two observers and small coefficients of variation were found regarding the ST-VM initial value (r = 0.99, 4.7), the ST-X maximum depression (r = 0.99, 3.2) and the QRS-VD end value (r = 0.98, 5.6). A less close correlation and higher coefficients of variation were found regarding the number of QRS-VD episodes (r = 0.94, 41.5), ST-VM episodes (r = 0.89, 37.8) and STC-VM episodes (r = 0.87, 35.1). Correlation coefficients and coefficients of variations for VCG measurements performed on two separate occasions by one observer ranged from 0.97 to 0.99 and from 18.1 to 1.8 respectively. Three (12%) of 25 patients with acute myocardial infarction did not meet the VCG infarction criterion (QRS-VD > or = 15 microVs) by both observers. In addition, five (20%) of the 25 patients with unstable angina pectoris met the VCG infarction criterion by both observers. The inter and intraobserver variation for VCG interpretations was low, but the number of QRS-VD, ST-VM and STC-VM episodes varied between the two observers. This finding suggests that additional training may improve the results. Caution is also recommended in using VCG to rule out or establish the diagnosis of acute myocardial infarction.