Calibrating Cardiac Output Measurement of Electrical Cardiometry With Transthoracic Echocardiography

Cardiac output (CO) is a key variable when describing and treating the cardiovascular system. The ideal cardiac output (CO) monitor would be safe, non-invasive, low-cost, painless, easy to interpret and would allow for the continuous, hands-free acquisition of accurate data. Electrical cardiometry (EC), a more recent version of impedance cardiography (ICG), meets many of the criteria for an ideal monitor and previously published research suggests that EC successfully trends CO1-3, but its accuracy in measuring absolute values of stroke volume (SV) has been inconsistent in published research4-8. Electrical cardiometry works by sending an insensible high frequency alternating current through the thorax and measures changes in thoracic impedance that are attributed to systolic blood flow acceleration over the cardiac cycle. Using fiducial points in the tracing of the first derivative of impedance, EC calculates left ventricular ejection time (LVET) and mean velocity of blood during systole. By further estimating a patient constant that is primarily based on body mass ('volume of electrically participating tissue' (VEPT)), stroke volume is estimated in the following way:

• SVEC = VEPT * Mean blood velocity * corrected Flow Time (FTc).

Cardiac output monitors are typically validated against a reference standard, which has traditionally been thermodilution using a pulmonary artery catheter (PAC). Given the risks associated with the use of pulmonary artery catheters9 an alternative method, such as transthoracic echocardiography (TTE), lends itself as valuable substitute technique for measuring CO. Transthoracic echocardiography is safe, painless and non-invasive. Similarly to EC, SV estimate by TTE is based on a measurement of mean blood velocity and ejection time (mean blood velocity x ejection time = velocity time integral, VTI; (Figure 3)). However, in contrast to EC, the parameters for blood velocity and left ventricular ejection time are not multiplied with a normative patient constant, but with a measured parameter: the left ventricular outflow tract area (LVOT area):

• SVTTE = LVOT area * mean blood velocity * ejection time. Main disadvantage of using TTE for CO-measurement is advanced training that is needed to obtain VTI measurements and that it is highly labor intense to obtain repeated VTIs for continuous or repeated CO measurements. Measurement of LVOT area, by contrast, is a one-time measurement obtained from parasternal long axis view, which necessitates only basic training of TTE.

Recently Martin and colleagues10 compared SV parameters derived from EC and TTE in a sample of 44 healthy pregnant volunteers and found excellent agreement in mean left ventricular ejection time and heart rate. However, agreement in SV was poor with a percentage error of 42% in Bland-Altman analysis. Hypothesizing that both methodologies in estimating SV, EC and TTE, use measures for blood velocity and systolic ejection time, Martin et al. hypothesized that agreement between both methods could be significantly improved by calibrating SV-EC with the LVOT area that is derived from one time TTE measurement. Based on a multiplicative linear regression model and following log-linear transformation Martin et al. derived the following formula for calibration:

SVEC_Modified = 2.2 * LVOT_area(0.705) * SV_EC(0.388) * Weight(0.21) By applying this formula, agreement in SV measurements significantly improved and percentage error decreased to 22 %. However, data was derived from a very specific patient population (pregnant women) and accuracy of the modified SV-assessment was not validated in a prospective manner.

Therefore, this study proposes to compare SV-estimates derived from EC and TTE in non-pregnant outpatients undergoing routine echocardiography assessment. Primary aim of this study is to evaluate if agreement between both methods can be improved by calibrating SV-measurements obtained from electrocardiometry with a one-time LVOT area measurement obtained from TTE.

Secondary aim will be to perform passive leg raising test and compare sensitivity in both identifying fluid responsiveness between both monitoring methodologies.

Third aim is to compare CO derived measurements from EC and TTE with gold standard measurements derived from patients undergoing routine right heart catheterisation for evaluation of cardiac and pulmonary function.