Sun JX, Reisner AT, Saeed M et al. Crit Care Med 2009;37:72–80.
Different algorithms used to estimate cardiac output were tested in a large database of arterial curves signals from patients equipped with invasive monitoring devices. Most algorithms had excellent agreement with regard to directional changes in cardiac output; however, according to the algorithm used, agreement with reference cardiac output varied markedly. Commercial devices could not be tested because their algorithm is not public and/or due to missing information in the database (such as height and weight).
Cardiac output is a key component of haemodynamic monitoring. It is often measured using a pulmonary artery catheter but this method is invasive and its use may be limited. Less invasive techniques have been developed, several of which are based on analysis of arterial blood pressure traces. Indeed, irruption of stroke volume into the aorta is responsible for the systolic increase in blood pressure, and accordingly, pulse pressure (defined as systolic pressure minus diastolic pressure) is proportional to stroke volume. However, aortic elastance and systemic vascular resistances (vascular tone) also affect this relationship. Hence, the relationship between pressure and volume varies between individuals, and for each individual it also varies as a non-linear function of arterial blood pressure. Several algorithms have been developed, aimed at better computing stroke volume from arterial pulse. In this manuscript the authors compared the performances of several algorithms; some were commercial algorithms but most were in the public domain.