Left and right atrial flow dynamics were compared by means of color and pulsed Doppler in order to study whether color Doppler could reliably provide differentiation between normals [15] and patients without atrial shunt at catheterization [12], vs patients with confirmed atrial septal defect [12]. The procedure consisted of sequential analysis of colored images throughout the cardiac cycle using an apical approach. In addition pulsed Doppler indices were calculated from both annular traces, relating diastolic early (E) and late (A) filling waves at each annulus (E/A); E and A waves were also summed (E + A), and the sum was related between both annuli (Tricuspid/Mitral ratio). Sequential analysis had a 100% sensitivity and specificity for the diagnosis of atrial septal defect, showing an asymmetrical pattern with predominant images in the right atrium, from the 2nd half of systole till End-diastole, vs the symmetrical 'Horseshoe' pattern found over both atria for control subjects. It avoided diagnostic errors due to overriding septal images in systole in 44% of controls. There also was a significant increase of the Tricuspid/Mitral ratios, (for duration and velocity time integral of waves) in patients with atrial septal defect. The correlation coefficient between ratios and values of the Pulmonary/Systemic flow ratio invasively calculated for 10 patients was respectively 0.6 and 0.7 (p less than 0.01). Sequential analysis of colored images appears highly reliable for the diagnosis of atrial septal defect; anomalies of ratios, although of moderate value for predicting shunt magnitude, substantiate the inequality of atrial fillings.
