Usefulness of Echocardiography/Doppler to Reliably Predict Elevated Left Ventricular End-Diastolic Pressure in Patients With Pulmonary Hypertension.

The ability of echocardiography (echo)/Doppler to predict elevated left ventricular (LV) end-diastolic pressure (EDP) specifically among patients with pulmonary hypertension is not well defined. This was a retrospective analysis of 161 patients referred to a specialized pulmonary hypertension clinic. A model based on an American Society of Echocardiography (ASE)/European Association of Echocardiography (EAE) joint statement was evaluated, and a new model was developed using univariate linear regression and multivariable logistic regression for potentially better prediction of elevated LVEDP. The study cohort had a median pulmonary arterial pressure was 34.0 mm Hg and pulmonary vascular resistance was 3.7 Wood units; 81 patients (51%) had LVEDP >15 mm Hg on invasive testing. Doppler E/A, E/e' (septal, lateral, and average), e'/a' (lateral and average), and left atrial volume and diameter all had significant correlation with LVEDP (p <0.05). The ASE/EAE model performed poorly (sensitivity 54% and specificity 66%) for detecting elevated LVEDP. Only echo/Doppler grade 3 diastolic dysfunction had an LVEDP significantly different from other grades (grade 0 to 2, median 15 mm Hg, interquartile range 13 to 22 mm Hg; grade 3, median 22 mm Hg, interquartile range 19 to 32 mm Hg; p <0.01). An experimental model was statistically significant in its prediction of elevated LVEDP (area under the receiver operating characteristic curve 0.7, p <0.001) but demonstrated poor performance (sensitivity 67% and specificity 61%). In conclusion, numerous echo/Doppler measurements correlate with elevated LV filling pressure. However, both the ASE/EAE model and our experimental model had poor test performance that did not permit confident identification of elevated LVEDP.