///2014 Abstract Details
2014 Abstract Details2019-07-18T14:34:47-06:00

Modified electrical cardiometry predicts velocity time integral, stroke volume and cardiac output when compared to transthoracic echocardiography in pregnant patients

Abstract Number: S-38
Abstract Type: Original Research

Erin I Martin MD1 ; Thomas Archer MD, MBA2; Jerasimos Ballas MD, MPH3; Adanna Anyikam MD, MPH4; Kristin Mantell MD5; Kristen Buono MD6

Electrical cardiometry (EC) is a simple, non-invasive and continuous monitor that measures cardiac output (CO) via impedance cardiography. In our previous studies, EC showed useful CO trends within individual pregnant patients, but did not give accurate absolute values when compared to transthoracic echocardiography (TTE). Since heart rate is easily measured, stroke volume (SV) is the variable requiring validation.

A certified cardiac sonographer, blinded to simultaneous EC data acquisition, performed TTE on 26 non-laboring patients positioned left side down and weighing 100 kg or less. SV by TTE (SV_TTE) equals the product of velocity time integral (VTI_TTE) and left ventricular outflow tract (LVOT) area. SV by EC (SV_EC) equals the product of square root of ICON (SQRT_ICON), left ventricular ejection time (LVET), and a patient constant. Unmodified, SV_EC had a 48.2 percentage error when compared to SV_TTE by the Bland-Altman technique.

Our new model (EC_A) predicts VTI (a measurement normally obtained by echocardiography) from EC data alone. Our key innovation is that Vmean_EC_A = SQRT_ICON/(22.7 - 0.16 x Weight), where Vmean_EC_A is the EC_A-derived prediction of mean systolic blood flow velocity in the LVOT. Vmean_EC_A multiplied by LVET gives VTI_EC_A, which we call a “virtual VTI.”

Obtaining a “virtual VTI” from EC alone is appealing since measuring VTI by TTE is difficult and time consuming, especially for repeated measurements. With our new approach, continuous VTI_EC_As are easily trended. When absolute SV is required, a one-time measurement of LVOT diameter using hand-held TTE could be obtained (although LVOT diameter was obtained by formal TTE in this study). LVOT area multiplied by VTI_EC_A gives SV_EC_A.

Applying EC_A to data from 4 laboring patients, Vmean_EC_A, LVET, VTI_EC_A and SV_EC_A agree with Vmean_TTE, Envelope Time, VTI_TTE and SV_TTE with percentage errors of 22.2, 15.4, 19.9 and 23.3%, respectively. Mean biases are -0.03m/sec, -0.2msec, -1.1cm and -4.0mL, respectively. Corresponding values for 6 new, non-laboring patients were 24.5, 8.8, 24.0 and 25.5%, with mean biases of -0.05m/sec, -5.0msec, -1.8cm and -7.2mL (Figure1).

Our new model not only allows for increased absolute accuracy in SV measurement, but also makes the equation for SV by EC mirror that for SV by TTE:

SV_TTE = (Vmean_TTE) x (Envelope Time) x (LVOT area), and

SV_EC_A = (Vmean_EC_A) x (LVET) x (LVOT area).

Further testing of our model is required.

SOAP 2014