Measurement of cardiac output (CO) and, sometimes more importantly, changes in CO can be extremely useful when assessing circulatory function. An easy, fast and reliable method to measure the CO is not only required for research purposes but also for clinical settings and currently missing in equine medicine.

Thermodilution technique as a measure of cardiac output (CO) was first introduced in the 1950's and is since then widely used in all kinds of species and recognized as a gold standard for evaluating and determination of the CO (1, 2). Notably in the clinical setting, however, it is considered being relatively invasive due to the necessity of placing intracardiac and -pulmonary artery catheters and therefore often not utilized in equine patients (3). As a result, there is a continuing search for a method of CO measurement that is less invasive than its predecessors.

A number of non-invasive methods of assessing CO have been studied in horses in the past, with transesophageal or transthoracic Doppler echocardiography (4, 5), Lithium dilution techniques with and without pulse contour analysis (6, 7), and carbon dioxide breath analysis currently available (8, 9). In human medicine thoracic bioimpedance was one of the first and currently most widely used non-invasive method for CO monitoring due to its easy application and non-invasiveness (10, 11).

Bioelectrical impedance analysis or bioimpedance is a widely used method for estimating fluid shifts in a biological system. In bioimpedance, a weak electric current flow through the body and the voltage is measured in order to calculate impedance (resistance) of the body (12). Thoracic electrical bioimpedance uses change in impedance by an alternating current applied across the thorax to determine various hemodynamic parameters, including stroke volume, cardiac output, and thoracic fluid content (12). Since the advent of the first available devices in the mid-1980's the method has become popular due to its ease of use (13). However, disadvantages of this non-invasive methodology are its inaccuracy impacted by electric noise and movement artifacts as well as its dependency of particular and accurate electrode placement (14).

To minimize the error rate and improve accuracy and reliability the bioimpedance methodology underwent further modification. While bioimpedance analyses only on the measured changes in signal amplitude the newly introduced bioreactance analyses the time delay or relative phase shift of the oscillating currents making the signal and analysis more reliable and less prone to motion artifacts and positioning errors (15–17). When compared to thermodilution the bioreactance methodology showed good agreement in swine and humans (15, 16). Further, validation studies of this technology in human and dogs have shown good correlation and agreement of bioreactance values compared with CO values obtained using other devices (16, 18–20) while other studies have reported a large limit of agreement and high percentage errors of up to 80% (21, 22). Validation studies for equine patients are lacking.

The goal of this study was to evaluate the performance of a bioreactance device for cardiac output measurements in adult horses. We hypothesized that there would be good agreement between TD CO measurements and bioreactance CO measurements across a range of CO states but that due to the anatomical differences between humans and horses a correction factor would be necessary to obtain valid values.

A total of 156 sets of paired NICOM and TDCO measurements were obtained and compared. No significant change in heart rate or respiratory rate was appreciated throughout the experiment.

The repeatability of the NICOM measurements was satisfactory with a MPE of 8.7% for repeated baseline measurements.

The administered drugs xylazine and dobutamine resulted in a CO range from 22 to 75 L/min for TD and CO range from 10.2 to 37.4 L/min for uncorrected NICOM over time during baseline measurements, after xylazine bolus and during dobutamine CRI (Figure 2).

The trend analysis for NICOM and TDCO revealed a concordance of 89% with significant correlation (r² = 0.9, p < 0.001) (exclusion zone <1 L/min).

The linear regression analysis performed for all data points at all CO situations calculated a relationship between NICOM and TDCO of Y = 0.4874 · X + 0.5936 with a goodness of fit of r = 0.96 and 95% confidence intervals for the slope from 0.4743 to 0.5006. For the low CO measurements, a factor of 0.496 and for the high CO measurements a factor of 0.437 was determined.

The four-quadrant plot describing the agreement between two methods and the change in CO is shown in Figure 3.

The NICOM and TDCO measurements showed a significant correlation for low, “normal” and high values (p < 0.001) with a correlation coefficient of r² = 0.77 for low, r² = 0.93 for “normal,” and r² = 0.89 for high CO measurements.

For the Bland Altman agreement the NICOM values were corrected by the factor 1:0.4874 (= 2.05) determined by the regression analysis for all measurements. The mean bias, lower, and upper limit of agreement were 0.44 L/min and −4.34 to 5.2 L/min for low, −0.41 L/min and −4.18 to 3.36 L/min for normal and 1.13 L/min and −4.98 to 7.25 L/min for high CO measurements, respectively. Bland and Altman plots are shown in Figure 4.

No adverse events occurred in association with either NICOM or TDCO measurements in any horse at any time.

The bioreactance cardiac output measurements is a relatively new way for continuous and non-invasive measurement of CO that has been validated for accuracy in different species (15, 16, 18–20, 26). The present study involved simultaneous bioreactance and TDCO measurements and was designed to determine whether the continuous data provided by the bioreactance would be able to accurately measure CO in horses and further quantify changes or detect trends.

Thermodilution technique was used as gold standard measurement to evaluate accuracy of the NICOM. While pulmonary artery catheterization and thermodilution measurement of CO has been widely used in experimental settings, the routine use of TDCO determination in clinical equine practice has been limited due to its invasiveness and necessity for extensive equipment such as the CO computer and a high-speed injector to ensure that the large volumes of thermal indicator are delivered consistently (3).

Our data indicate that bioreactance measurements and the TDCO measurements show a strong correlation, and the linear regression analysis as well as concordance calculation indicates that bioreactance measurements needed to be corrected by the factor 2 (1:0.4874). The reason for the almost 50% underestimation of the NICOM could be the difference in cardio-vascular anatomy and physiology between human and equine patients as the used monitoring system is based on an algorithm developed for human patients. A recent study developed a 1D computer model for the equine arterial circulation comprising all major vessels of the arterial tree indicating significant differences in arterial pressure waves and wave forms as well as vessel performance for the horse (27). Furthermore, the significant differences in anatomy and physiology of the vascular structures between horses and humans and differences and cardiac performance such as stroke index and heart rate (28, 29) make the translation of inbuilt algorithms difficult. Therefore, it is not surprising that a correction factor was necessary to obtain accurate values. Another important factor to consider is the difference in cardiothoracic anatomical structure of human and horse (bipedal vs. quadrupedal) as well as the resultant difference in location of sensor placement. While the proportionality constant in the used NICOM device is derived from characteristics of humans (15) such equine based calibration constant is unknown. More studies with horses of different sizes, weights, and ages are needed to determine if other correction factors are necessary or the factor determined in this study represents an adequate constant for all studies in horses. The very good accuracy and performance of the NICOM, despite being described in the literature (16, 18–20), was not expected in that degree by the authors. However, despite the small bias we appreciated a wider LAO, particularly when higher CO states were tested. We can only speculate that the unique cardiovascular anatomy of the horse with a relatively large heart placed in the center of the chest contributes to the good performance of the NICOM. Further studies using different cardiovascular conditions are needed to confirm our findings.

For clinical application the detection of a trend in CO changes is of great importance as the measurement of actual values. Therefore, the second goal of this study was to evaluate whether the NICOM would detect changes in CO and provide accurate trends. For sufficient evaluation a wide range of different cardiovascular stages with high and low cardiac output values was induced by drugs that have known effect on the cardiovascular system and the cardiac output. Xylazine at comparable dosages as used in this study has shown to decrease the cardiac output by ~50% (30) and thermodilution as well as bioreactance measurements showed expected decreases in CO values for up to 35 min after xylazine bolus administration. To induce high cardiac output states dobutamine was administered at dosages of 1.5 and 3 mcg/kg/min. In anesthetized and in standing horses dobutamine infusion rates of 0.5 to 4 mcg/kg/min resulted in an increase of cardiac output up to 35% (31–33) which was also seen in our horses. These drug protocols resulted in CO values ranging from 22 to 75 L/min measured by TD and these trends and changes were detected by NICOM with great correlation (r² = 0.978).

After correcting the values obtained by bioreactance using the regression factor 1.96, the Bland Altman analysis calculated a small bias for all three induced and investigated CO stages but a large LOA ranging from +5 to −7 L/min. Although, this LAO can still be considered acceptable being in the range of 15% it needs to be considered particularly in clinical settings in which decisions could be based on the measured CO values (24, 26). This also emphasized the importance of careful interpretation of these results obtained on a relatively small number of healthy horses.

No major adverse events were observed in the animals throughout the course of the study when a bioreactance system was used. This can be expected as the system is non-invasive.

The current study shows that this device can perform well in standing, healthy horses.

The results of this study showed a good agreement of the bioreactance-based system for measurement of CO compared to thermodilution technique, but also a relatively wide limit of agreement (2SD). As a next step this methodology needs to be introduced into clinical settings and tested on a larger number of cardiovascularly compromised patients and on a larger range of age, size, and bodyweight to assess not only reliability of this technique but also performance under clinical conditions including practicability. The calibration factor we found in this study will need further evaluation in a larger and more heterogeneous dataset. Further, developing an algorithm which would account for patient-specific parameters will be necessary to improve clinical utility.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.

The animal study was reviewed and approved by the protocol was approved by the University's Institutional Animal and Use Committee (#807004).

KH: study design, data collection, analysis and interpretation of data, and writing up of the first draft of the manuscript. SH: study design, data collection, analysis and interpretation of data, and critical revision of the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.