Underestimation of left atrial volume by three-dimensional echocardiography validated by magnetic resonance imaging: a meta-analysis and investigation of the source of bias.

BACKGROUND

Left atrial volume (LAV) is a fundamental prognostic factor in a variety of cardiac diseases including atrial fibrillation, heart failure, and valvular diseases. Developing a repeatable, noninvasive, and accurate method of measuring LAV is crucial. Three-dimensional echocardiography (3DE) has achieved better estimation of LAV than two-dimensional echocardiography. However, underestimation of LAV by 3DE has often been reported and no previous study has synthesized these data. The present study aimed at revealing existence and extent of bias in LAV measurement by 3DE and investigating related factors affecting the bias.

METHODS

Studies comparing LAV between 3DE and magnetic resonance imaging published before August 15, 2011 were eligible. A meta-analysis with random effects model was performed to evaluate the systematic bias. Factors affecting the bias were investigated by univariate followed by multivariate analysis.

RESULTS

A total of 14 studies including 395 subjects revealed underestimation of LAV by 3DE (-9.4 mL; 95% confidence interval, -13.2 to -5.6 mL; P < 0.00001, Fig. 2). Existence of cardiac disease led to more underestimation, whereas increasing the number of plane included in the analysis counteracted the underestimation (P < 0.00001 for each comparison).

CONCLUSIONS

Only by synthesizing a number of small studies as a meta-analysis could we display underestimation of LAV by 3DE and factors influencing the systematic bias. These data provide a more detailed basis for analyzing and improving the accuracy of 3DE, an indispensable step toward further clinical application in LAV assessment.