A validation study of electrical cardiometry in pregnant patients using transthoracic echocardiography as the reference standard

Abstract

To validate electrical cardiometry (EC) in pregnant patients using transthoracic echocardiography (TTE) as the reference standard. To improve EC accuracy via a one-time, measurement of left ventricular outflow tract (LVOT) diameter. 44 non-laboring, resting women with singleton, viable pregnancies underwent simultaneous EC and TTE measurements. Data were analyzed using Bland-Altman analysis. Entry multiple regression with stepwise elimination was used to develop a model for improved prediction of stroke volume by TTE (SVTTE) using EC. Bootstrapping and an 11-fold cross validation were used to test the model. Heart rate by TTE and EC had a mean bias of 3.3beats/min and mean percentage error of 10.7%. Envelope time and left ventricular ejection time had a mean bias of -4.9ms and mean percentage error 12.7%. Stroke volumes by the two techniques had a mean bias of 15.6mL and mean percentage error of 43.7%. A model, SVEC_Modified, predicting SVTTE was developed using LVOT area, stroke volume by electrical cardiometry and weight. SVTTE and SVEC_Modified had a mean bias of -0.83mL and mean percentage error of 22%. EC accurately measures heart rate and duration of systole when compared with TTE. Stroke volume measurements correlate but have a high bias and percentage error. Knowledge of LVOT area, by a one-time, measurement with TTE, could improve prediction of stroke volume by EC.