Reliability of the pulse contour analysis for cardiac output measurement for assessing the fluid responsiveness]

Background. - The cardiac output (CO) is classically measured in intensive care unit patients, pulse contour (PC) method allows monitoring of CO. Objective. - The aim of the present study was to assess the ability of PC to assess the fluid responsiveness defined as an increase in CO more than or equal to 15% after 500 mL hydroxyethyl starch over 20 minutes. Patients and methods. - In this observational prospective study, patients in shock in whom a PC method was inserted were included. CO was measure using the PC and thermodilution methods before and after a fluid challenge indicated by the physician. The correlation coefficient was measured, the diagram of Bland and Altman was built and the percentage of error (Critchley and Critchley method) was calculated. The ability of PC to diagnose fluid responsiveness was assessed using a receiver operating characteristics (ROC) curve. Results. - Sixty-two fluid challenges were performed in 37 included patients. After fluid challenge, r(2) was 0.05 (P < 0.01), the bias between PC and thermodilution was 0.3 +/- 1.2 L/min and the percentage of error was 36%. The area of the ROC curve was 0.601 [0.468-0.723]. Conclusion. - In ICU patients with shock, PC cannot replace thermodilution to diagnose fluid responsiveness. (C) 2011 Published by Elsevier Masson SAS.