COLOR-FLOW DOPPLER DETERMINATION OF TRANSMITRAL FLOW AND ORIFICE AREAIN MITRAL-STENOSIS - EXPERIMENTAL EVALUATION OF THE PROXIMAL FLOW-CONVERGENCE METHOD

To evaluate the in vivo accuracy of color Doppler flow-convergence met hods for determining transmitral flow volumes and effective orifice ar eas in mitral stenosis, we studied two models for flow-convergence sur face geometry, a hemispheric (HS) model and an oblate hemispheroid (OH ) modes in a chronic animal model with quantifiable mitral flows. Colo r Doppler flow mapping of the proximal flow-convergence region has bee n reported to be useful for evaluation of intracardiac flows. Flow-con vergence methods in patients with mitral stenosis that use HS assumpti on for the isovelocity surface have resulted in underestimation of act ual flows. Chronic mitral stenosis was created surgically in six sheep with annuloplasty rings (group 1) and in 11 sheep with bioprosthetic porcine valves (group 2). Hemodynamic and echocardiographic/Doppler st udies (n = 18 in group 1; n = 21 in group 2) were performed 20 to 34 w eeks later. Left ventricular inflow obstruction was of varied severity , with mean transmitral valve gradients in group 1 ranging from 1.3 to 18 mm Hg and in group 2 ranging from 6.3 to 25.6 mm Hg. Although tran smitral flows derived by both geometric flow convergence models showed significant correlations with reference cardiac outputs, the correlat ions for the OH model were better than those for the HS model (group 1 , r = 0.86 for the OH model vs r = 0.72 for the HS model; group 2; r = 0.84 for the OH model vs r = 0.62 for the HS model). The OH model was also superior to the HS model in determining effective orifice areas compared to reference orifice areas determined by postmortem planimetr y of anatomic orifices (group 1 only, r = 0.64 for OH vs 0.58 for HS), by the Gorlin and Gorlin formula (group 1, r = 0.63 for OH vs 0.72 fo r HS; group 2, r = 0.82 for OH vs 0.76 for HS), and by the Doppler pre ssure half-time method (group 1, r = 0.76 for OH vs 0.69 for HS; group 2, r = 0.84 for OH vs 0.62 far HS). The percentage differences betwee n the reference values and calculated data with the OH model were sign ificantly smaller for transmitral flows (-5.8% vs -59% in group 1 and -13% vs -63% in group 2) and for effective orifice areas in both group s, p < 0.0001 (-21% vs -56% for mitral orifice area by postmortem plan imetry in group 1; -28% vs -61% in group 1 and -24% vs -68% in group 2 for mitral orifice area by the Gorlin and Gorlin formula; and -18% vs -57% in group 1 and -27% vs -71% in group 2 for the pressure half-tim e method, p < 0.0001 for each comparison). These studies demonstrate t hat flow convergence principles applied to color Doppler flow mapping permit estimation of transvalvular flow volumes and orifice areas in m itral stenosis. However, in the presence of orifices that are not infi nitesimally small and when it is not clinically feasible to sample at substantial distances proximal to the orifice, determination of flows and areas with flow convergence principles results in underestimation relative to reference standards. This underestimation can be minimized by applying an OH geometric model to the flow convergence surface are a rather than a strictly HS model.