Is colour flow imaging needed to exclude clinically significant valvular regurgitation in adult patients undergoing transthoracic echocardiography?

OBJECTIVE: To establish whether clinically significant aortic and mitral valvular regurgitation can be excluded in adult patients undergoing transthoracic echocardiography without using colour flow imaging. SETTING: Sunnybrook Health Science Centre, a tertiary referral centre with full cardiovascular services affiliated with the University of Toronto, Toronto, Ontario. DESIGN: Logistic regression models were developed from a retrospective review of 14,051 unselected consecutive echocardiograms from 1991 through 1994. The dependent variable was more than mild aortic or mitral valvular regurgitation. Independent variables included age, sex and various functional and structural measures obtained during routine two dimensional echocardiography. The negative predictive values and sensitivity of the models were estimated. INTERVENTION: The number of patients correctly classified by these models, as well as the proportion for whom the colour flow imaging did not add to the baseline echocardiogram, was determined. Nonparametric bootstrapping was used to obtain confidence intervals for these statistics. MAIN RESULTS: Ten models were developed, with negative predictive values ranging from 96.2% to 100%. Incorporation of such decision aids into the software of echocardiographic machinery would help echocardiographers to rule out significant aortic or mitral regurgitant lesions. In practices where colour flow imaging is routinely employed, 40% fewer procedures could be performed. CONCLUSIONS: Models based on simple demographic and two dimensional echocardiographic variables can reliably exclude significant valvular regurgitation and could potentially reduce the volumes and costs of colour flow imaging. Given the widespread diffusion of colour Doppler imaging, the models may also be helpful to avoid misinterpretation of flow imaging results, by defining subgroups in whom the prior probability of significant aortic or mitral regurgitation is extremely low.