[Reliability of digital subtraction angiocardiography for the evaluation of left ventricular function and physical stress].

The reliability of determinations of left ventricular dimensions from digital subtraction angiocardiographies during exercise using intravenous injection of contrast material was proved in 20 patients. All angiocardiograms could be analyzed qualitatively and quantitatively only by means of the time-interval-difference (TID-) mode. Compared with other mask modes the superiority of the TID-mode results from the close temporal relationship between mask and contrasted frame. Thus, changes in position and brightness of the background structures are small, and left ventricular contours can be determined accurately despite intense respiratory movements. The intra-observer variability of determinations of end-diastolic and end-systolic volumes was +/- 12 ml (i.e. 7.5% of the mean value), corresponding inter-observer variability was +/- 24 ml (i.e. 17% of the mean value), respectively. Intra-observer deviation in determining ejection fraction was 2.8% (i.e. 4.5% of the mean value). The intra-observer variability in determining left ventricular axes was basal +/- 0.40 cm (i.e. 7% of the mean value), equatorial +/- 0.73 cm (i.e. 13% of the mean value) and apical +/- 0.34 cm (i.e. 8% of the mean value). Deviations of corresponding circumferential fiber shortening rates were basal +/- 0.22 s-1 (i.e. 14% of the mean value), equatorial +/- 0.32 s-1 (i.e. 22% of the mean value), and apical +/- 0.18 s-1 (i.e. 13% of the mean value). The results show that left ventricular dimensions can be determined accurately from intravenous angiocardiograms during exercise using digital subtraction angiocardiography.