Comparison of Cardiac Output Determined by Bioimpedance, Thermodilution, and the Fick Method

Abstract

• Background Cardiac output can be determined by using a variety of methods.• Objectives To determine the precision and bias between 3 methods for determining cardiac output: bioimpedance, thermodilution, and the Fick method.• Methods Cardiac output was determined by using bioimpedance via neck and thorax patches and thermodilution via pulmonary artery catheter in 46 patients in the intensive care unit. A subset of 15 patients also had cardiac output determined by using the Fick method.• Results Mean (SD) cardiac output in all patients was 6.3 (2.2) L/min by thermodilution and 5.6 (2.0) L/min by bioimpedance. In the 15 patients in whom all 3 methods were used, mean cardiac output was 6.0 (1.7) L/min by thermodilution, 5.3 (1.7) L/min by bioimpedance, and 8.6 (4.5) L/min by the Fick method. Bias and precision (mean difference ± 2 SDs) were 0.7 ± 2.9 L/min between thermodilution and bioimpedance, 1.7 ± 3.8 L/min between the Fick method and thermodilution, and 2.4 ± 4.7 L/min between the Fick method and bioimpedance.• Conclusion Bioimpedance, thermodilution, and Fick determinations of cardiac outputs are not interchangeable in a heterogenous population of critically ill patients.

Changes in tissue oxygen tension, venous saturation, and Fick‐based assessments of cardiac output during hyperoxia

“…We noted that Fick O2 overestimates CO during normoxic ventilation, and that this overestimation increases progressively during hyperoxia due to its effects on SvO 2 . Several groups have noted a similar degree of overestimation by Fick O2 as compared to a reference standard during normoxia, at times to the same degree that we did . A fraction of the error may be that we neglected to measure coronary blood flow, though as mentioned previously we believe this to be a negligible fraction of total CO. More likely, this phenomenon is related to compounding measurement errors in several of the components of Fick O2 .…”

“…In response to changes in PaO 2 , SvO 2 may change to a greater magnitude than arterial oxyhemoglobin saturation, narrowing the arteriovenous oxygen content difference and increasing Fick O2 estimates of CO. Fick O2 remains commonly used in critically ill infants and children with congenital heart disease, as well as in the cardiac catheterization laboratory, where it is used to estimate pulmonary vascular resistance and evaluate candidacy for cavopulmonary anastomosis or for heart transplantation . The majority of studies comparing Fick estimates to reference standard measurements, such as thermodilution, describe an overestimation of CO by Fick relative to reference standards, at times to a clinically and statistically significant degree . This phenomenon appears to be more pronounced in patients with elevated arterial oxygen tension.…”

Diurnal and position-induced variability of impedance cardiography measurements in healthy subjects

“…In the last decades, many reports and reviews have been published on ICG with conflicting results and conclusions. Valuable ICG measurements have been reported (Woltjer et al 1997;Ono et al 2004;Sodolski & Kutarski 2007), with values comparable to those obtained by conventional methods such as thermodilution or Fick method (Woltjer et al 1996;Spiess et al 2001;Drazner et al 2002;Van De Water et al 2003;Albert et al 2004;Engoren & Barbee 2005), as well as inaccurate measurements with substantial variation between different algorythms (Woltjer et al 1997;de Waal et al 2008). Due to this, ICG as a useful clinical device has not yet gained wide acceptance (Donovan et al 1986;Wang et al 2006;de Waal et al 2008).…”

“…Measurements by ICG were reported to be reliable as they correlate highly with clinical standard methods (Spiess et al 2001;Drazner et al 2002;Van De Water et al 2003;Albert et al 2004;Cotter et al 2004), however these results are debated by others (Drazner et al 2002;Sodolski & Kutarski 2007;Mathews & Singh 2008;Engoren & Barbee 2005;de Waal et al 2008). After more than 40 years of comparitive studies and improvements in hardware and software (Woltjer et al 1996;Ono et al 2004), the third generation ICG has now been introduced for clinical use, allowing measurement of multiple CV parameters.…”

Noninvasive continuous cardiac output monitoring in perioperative and intensive care medicine

“…31 -34 On the other hand, there are also data showing high bias and wide limits of agreement when comparing thoracic electrical bioimpedance with pulmonary artery thermodilution in cardiac surgery patients 35 and with pulmonary artery thermodilution in mixed ICU patients. 36 In a meta-analysis published in 2010, 19 more than 400 patients from 13 thoracic electrical bioimpedance validation studies were analysed. Compared with thermodilution cardiac output measurements, the authors reported a pooled bias of 20.10 litre min 21 with a pooled weighted precision (one standard deviation) of agreement of 1.14 litre min 21 resulting in an overall percentage error of 42.9% for thoracic electrical bioimpedance measurements.…”