Prediction of fluid responsiveness in septic shock patients

Abstract

The aim of this study was to assess and compare the ability of the automatically and continuously measured stroke volume variation (SVV) obtained by FloTrac/Vigileo, and pulse pressure variation (PPV) measured by an IntelliVue MP monitor, to predict fluid responsiveness in mechanically ventilated septic shock patients. We conducted a prospective study on 42 septic shock patients. SVV, PPV and other haemodynamic data were recorded before and after fluid administration of 500 ml of 6% hydroxyethyl starch. Responders were defined as patients with an increase in stroke volume index of at least 15% after fluid loading. Twenty-four (57.1%) patients were classified as fluid responders. The baseline SVV correlated with the baseline PPV (r=0.96, P<0.001). SVV and PPV were significantly higher in responders than in nonresponders (15.5±4.5 vs. 8.2±3.3% and 16.4±5.2 vs. 8.3±3.5, respectively, P<0.001 for both). There was no difference between the area under the receiver operating characteristic curves of SVV [0.92, 95% confidence interval 0.832-1.00] and PPV (0.916, 95% confidence interval 0.829-1.00). The optimal threshold values in predicting fluid responsiveness were 10% for SVV (sensitivity 91.7% and specificity 83.3%) and 12% for PPV (sensitivity 83.3% and specificity 83.3%). Our results were independent of the site of arterial catheterisation. The SVV, obtained by FloTrac/Vigileo, and the automated PPV, obtained by the IntelliVue MP monitor, showed comparable performance in terms of predicting fluid responsiveness in passively ventilated septic shock patients, with a regular cardiac rhythm and a tidal volume not less than 8 ml kg(-1).