Influence of recording scale on determination of left ventricular end-diastolic pressure

Left ventricular (LV) pressure waveforms for 20 beats, originally recorded at 200 scale (200 torr/10 cm) were rescaled by a computer system to represent the identical beats recorded at 40 scale (40 torr/10 cm). Both the original and the rescaled beats were independently evaluated by eight cardiologists who were unaware of the nature of the manipulated data. For each beat the average left ventricular end-diastolic pressure (LVEDP) identified by the eight cardiologists was higher when analyzed on the 200 scale, with an average difference of 2.8 torr (P Less Than 0.001). This scale-dependence of pressure analysis has implications for the development of automated pressure analysis algorithms and for physician review of computer-determined LVEDP data points.