Doppler echocardiography in the dog: measurement variability and reproducibility.

Serial Doppler echocardiographic examinations were carried out in random order on six boxer dogs, on 3 separate days, by two experienced Doppler echocardiographers, to assess measurement variability and reproducibility of 65 parameters. Large numbers of parameters exhibited significant differences for each of the categories of intraobserver, interobserver, interday and interoperator. The coefficients of variation for all parameters measured ranged from 5.03 to 46.43%, but most were less than 20%. In general, least variation was found for the intraobserver category, and the best reproducibility for M-mode and left ventricular volumetric data. The worst reproducibility was found for tricuspid inflow and pulmonary venous flow measurements. The results of this study suggest differences greater than 20% for serial scans must be achieved to document genuine change, although the specific data should be consulted. Furthermore, variability and reproducibility are improved if a single experienced operator/observer acquires and measures serial scans.