Accuracy of a multiparametric score based on pulse wave analysis for prediction of fluid responsiveness: ancillary analysis of an observational study

PURPOSE: The pressure recording analytical method (PRAM) monitor is a non-invasive pulse contour cardiac output (CO) device that cannot be considered interchangeable with the gold standard for CO estimation. It, however, generates additional hemodynamic indices that need to be evaluated. Our objective was to investigate the performance of a multiparametric predictive score based on a combination of several parameters generated by the PRAM monitor to predict fluid responsiveness. METHODS: Secondary analysis of a prospective observational study from April 2016 to December 2017 in two French teaching hospitals. We included critically ill patients who were monitored by esophageal Doppler monitoring and an invasive arterial line, and received a 250–500 mL crystalloid fluid challenge. The main outcome measure was the predictive score discrimination evaluated by the area under the receiver operating characteristics curve. RESULTS: The three baseline PRAM-derived parameters associated with fluid responsiveness in univariate analysis were pulse pressure variation, cardiac cycle efficiency, and arterial elastance (P < 0.01, P = 0.03, and P < 0.01, respectively). The median [interquartile range] predictive score, calculated after discretization of these parameters according to their optimal threshold value was 3 [2–3] in fluid responders and 1 [1–2] in fluid non-responders, respectively (P < 0.001). The area under the curve of the predictive score was 0.807 (95% confidence interval, 0.662 to 0.909; P < 0.001). CONCLUSION: A multiparametric score combining three parameters generated by the PRAM monitor can predict fluid responsiveness with good positive and negative predictive values in intensive care unit patients. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (10.1007/s12630-020-01736-y) contains supplementary material, which is available to authorized users.