The decision algorithm for managing patients in cardiogenic shock depends on cardiac index (CI) estimates. Cardiac index estimation via thermodilution (CI-TD) using a pulmonary artery catheter is used commonly for obtaining CI in these patients. Minimally invasive methods of estimating CI, such as multibeat analysis (CI-MBA), may be an alternative in this population. A prospective, observational study. Cardiac intensive care unit. Twenty-two subjects in cardiogenic shock provided 101 paired CI measurements. Measurements were obtained concomitantly by intermittent CI-TD and CI-MBA (Argos Cardiac Output Monitor; Retia Medical, Valhalla, NY). For each CI-TD, CI-MBA estimates were averaged over 1 minute to provide paired values. Bland-Altman and 4-quadrant analyses were performed by plotting changes between successive CI measurements (ΔCI) from each of the 2 methods. Concordance was calculated as a percentage using ΔCI data points from the 2 methods, outside an exclusion zone of 15%. The correlation coefficient between CI-MBA and CI-TD was 0.78 across patients. Mean CI-TD was 2.19 ± 0.46 L/min/m2 and mean CI-MBA was 2.38 ± 0.59 L/min/m2. The mean difference between CI-MBA and CI-TD (bias ± SD) was 0.20 ± 0.47 L/min/m2, and the limits of agreement were -0.72 to 1.11 L/min/m2. The percentage error was 40.0%. The concordance rate was 94%. A secondary analysis of a subgroup of patients during periods of arrhythmia demonstrated a similar accuracy of performance of CI-MBA. Cardiac index-MBA is not interchangeable with CI-TD. However, CI-MBA provides reasonable correlation and clinically acceptable trending ability compared with CI-TD. Cardiac output-MBA may be useful in trending changes in CI in patients with cardiogenic shock, especially in those whose pulmonary artery catheterization placement carries a high risk or is unobtainable.