Comparison of pulse pressure variation versus echocardiography-derived stroke volume variation for prediction of fluid responsiveness in mechanically ventilated anesthetized dogs

Abstract

ObjectiveTo evaluate the ability and accuracy of aortic flow velocity–time integral variation (ΔVTI) and peak aortic velocity variation (ΔVpeak) compared with pulse pressure variation (PPV) to predict fluid responsiveness in mechanically ventilated dogs. Study designProspective clinical study. AnimalsA group of 50 mechanically ventilated dogs with spontaneous hypotension during orthopedic or oncologic surgery. MethodsInvestigations were performed in the surgery room. When mean arterial pressure (MAP) decreased to <65 mmHg, measurements were performed before and after a fluid challenge (lactated Ringer’s solution 5 mL kg−1 over 15 minutes). Responders were defined as a change in stroke volume (SV; transesophageal Doppler) ≥15%. Data were analyzed using paired/unpaired t test or Mann–Whitney/Wilcoxon test when appropriate and receiver operating characteristics (ROC) curves; a p value <0.05 was considered statistically significant. ResultsAfter the fluid challenge, 35 (70%) of 50 dogs were responders with significant increases in SV and decreases in PPV; 15 dogs were nonresponders. ΔVTI and ΔVpeak correlated with a 15% increase in SV. The optimum cut-off value for PPV was 15.6% (sensitivity, 88%; specificity, 100%), for ΔVTI was 10.65% (sensitivity, 65%; specificity, 100%) and for ΔVpeak was 10.15% (sensitivity, 80%; specificity, 100%). The area under the ROC curve for PPV was (0.93 ± 0.08) and for ΔVpeak was (0.89 ± 0.09), before fluid challenge. The gray zone area spread from 6.15% to 15.6% for PPV (18 dogs), 2.45% to 10.65% for ΔVTI (22 dogs) and 0.6% to 10.15% for ΔVpeak (25 dogs). ConclusionsWhen using mechanical ventilation, ΔVTI and ΔVpeak predicted fluid responsiveness with the same ability as PPV, based on the area under the ROC curve analysis. However, PPV showed great accuracy demonstrated by a narrower gray zone that included fewer individuals. Clinical relevanceΔVTI and ΔVpeak can be used as indices of fluid responsiveness in anesthetized dogs.