Objective: To investigate sources of variability in serial echocardiographic recordings in a core laboratory we assessed the impact of repeated echo recordings, repeated video measurements and measurements made by different investigators.
Patients, methods: Two investigators each recorded and analysed two-dimensional echos in 12 individuals (n = 24 in total) three times at one week intervals. Left ventricular end-diastolic and end-systolic volumes were measured using the biplane modified Simpson's rule. Ejection fraction was derived from these volumes and left ventricular mass estimated using the area-length method. The left ventricular spherity index was expressed as the ratio of the short axis area and the long axis area at end-diastole. A video recording from each examination was reexamined twice by both investigators.
Results: Deviations between repeated echo recordings and repeated video measurements ranged from -5 to +5% between investigators. A three-way repeated analysis of variance indicated a small, but systematic difference between investigators. Reproducibility, measured by coefficients of variation, ranged from 3-9% for different investigators, 3-6% for repeated video measurements and 7-19% for repeated echo recordings across the different variables. The total variability of all three factors should be considered when the smallest detectable significant change in a variable is assessed. These ranged from 16-28% across the five variables studied, when a 10% error of classification was accepted for a one-sided change in a variable.
Conclusion: Repeated echo recordings were the dominant component of variation. Two-dimensional echo measurements are reproducible and accurate, but the same investigator should follow the same patients.