Abstract
BACKGROUND: Cardiac output (CO) estimation in patients in intensive care units (ICUs) by a non-invasive, automated, oscillometric, cuff-based apparatus (Mobil-O-Graph [MG]) is reproducible with acceptable accuracy versus thermodilution. In this pilot study, we tested the hypothesis that clinical decisions based on the MG device are in agreement with those based on invasive measurements using a Swan-Ganz catheter (SGC). METHODS: Hemodynamic monitoring using an SGC and an MG was performed on 20 consenting critically ill patients in shock and under treatment, hospitalized in ICU. Retrospectively, three ICU physicians were asked to determine the need for blood transfusion, inotropes, fluids, diuretics, oxygen, and vasoconstrictive agents. Decisions (defined as "need for action" or "no action") were based: (i) on SGC-acquired data and standard ICU monitoring (SIM); (ii) on MG-acquired data and SIM; (iii) SIM only. The decisions were compared using Cohen's kappa agreement coefficient and Wilcoxon's nonparametric test. RESULTS: The overall number of decisions, as well as the subanalysis of "need for action" decisions, based either on information from an SGC or MG, were comparable. The significant positive kappa agreement coefficients indicated moderate to strong agreement. MG-derived decisions agreed with SGC-derived decisions to a significantly higher degree compared with SIM-based decisions. CONCLUSIONS: Clinical decisions in the ICU setting based on MG data were in acceptable agreement with SGC-based decisions. Larger studies are required to confirm this finding. MG devices may provide a simple, operator-independent, low-cost, first-line bedside method for simultaneous continuous monitoring of blood pressure and CO levels in critically ill patients outside the ICU.