Objective: To evaluate the role of pleth variability index ( PVI ) by passive leg raising ( PLR ) test in volume responsiveness and volume status prediction in patients with septic shock.
Methods: A prospective randomized controlled trial ( RCT ) was conducted. Eighty-seven patients suffering from septic shock undergoing mechanical ventilation in Department of Critical Care Medicine of Subei People's Hospital from June 2012 to September 2014 were enrolled. The hemodynamic changes before and after PLR were monitored by pulse indicated continuous cardiac output ( PiCCO ) and PVI monitoring. Responsive group: positive fluid response was defined as an increase in cardiac index ( CI )≥10% after PLR. Unresponsive group: negative fluid response was defined as an increase in CI<10% after PLR. The hemodynamic parameters, including heart rate ( HR ), mean arterial pressure ( MAP ), central venous pressure ( CVP ), stroke volume variation ( SVV ), CI and PVI, and the changes in cardiac parameters (ΔHR, ΔMAP, ΔCVP, ΔSVV, ΔCI, and ΔPVI ) before and after PLR were determined. The relations between hemodynamic parameters and their changes with ΔCI were analyzed by the Pearson analysis. The role of the parameters for volume responsiveness prediction was evaluated by receiver operating characteristic ( ROC ) curves.
Results: 145 PLRs in 87 patients with septic shock were conducted, with 67 in responsive group and 78 in unresponsive group. There were no statistically significant differences in HR, MAP, CVP and CI before PLR between the responsive and unresponsive groups. SVV and PVI in responsive group were significantly higher than those in the unresponsive group [ SVV: ( 16.9±3.1 )% vs. ( 8.4±2.2 ) %, t = 9.078, P = 0.031; PVI: ( 20.6±4.3 )% vs. ( 11.1±3.2 )%, t = 19.189, P = 0.022 ]. There were no statistically significant differences in HR, MAP, CVP, SVV, and PVI after PLR between the responsive group and unresponsive group. CI in the responsive group was significantly higher than that in the unresponsive group ( mL×s(-1)×m(-2): 78.3±6.7 vs. 60.0±8.3, t = 2.902, P = 0.025 ). There were no statistically significant differences in ΔHR, ΔMAP, ΔCVP between responsive group and unresponsive group. ΔSVV, ΔCI and ΔPVI in responsive group were significantly higher than those in the unresponsive group [ ΔSVV: ( 4.6±1.5 )% vs. ( 1.8±0.9 )%, t = 11.187, P = 0.022; ΔCI ( mL×s(-1)×m(-2) ): 18.3±1.7 vs. 1.7±0.5, t = 3.696, P = 0.014; ΔPVI: ( 6.4±1.1 )% vs. ( 1.3±0.2 )%, t = 19.563, P = 0.013 ]. No significant correlation between HR, MAP or CVP before PLR and ΔCI was found. SVV ( r = 0.850, P = 0.015 ) and PVI ( r = 0.867, P = 0.001 ) before PLR were correlated with ΔCI. It was shown by ROC curve that the area under ROC curve ( AUC ) for SVV fluid responsiveness prediction was 0.948, and cut-off of SVV was 12.4%, the sensitivity was 85.4%, and specificity was 86.6%. The AUC for PVI fluid responsiveness prediction was 0.957, and cut-off was 14.8%, the sensitivity was 87.5%, and specificity was 84.8%. It was higher than other hemodynamic parameters ( HR, MAP, CVP ).
Conclusions: PVI and SVV can better predict fluid responsiveness in mechanically ventilating patients with septic shock after PLR. PVI as a new continuous, noninvasive and functional hemodynamic parameter has the same accuracy as SVV.