Utility of dynamic preload indices of hypovolemia in isoflurane-anesthetized horses

Objective: To evaluate the utility of the dynamic indices, systolic pressure variation (SPV%), pulse pressure variation (PPV%) and plethysmographic variability index (PVI%) during induced hypovolemia, and to determine their correlation with cardiac output (CO) in anesthetized horses. Study design: Prospective experimental study. Animals: A group of eight healthy adult research horses. Methods: After sedation with intravenous (IV) xylazine, anesthesia was induced with IV ketamine and diazepam and maintained with constant-dose isoflurane delivered in 100% oxygen. Horses were positioned in dorsal recumbency, paralyzed with IV atracurium infusion, mechanically ventilated to normocapnia and instrumented with two 7 Fr Swan-Ganz catheters for measurement of thermodilution CO and a facial artery catheter to measure arterial blood pressure. Dobutamine was used to maintain normotension. All three dynamic indices of hypovolemia were recorded before (baseline), and 10 minutes after graded hemorrhage (5 mL kg–1 increments) up to 25 mL kg–1 and subsequent autotransfusion (5 mL kg–1 increments). A cut-off value that best distinguished nonbaseline time points while optimizing sensitivity and specificity was determined for each dynamic index. Relationships between dynamic preload indices and CO were determined using Spearman’s correlation. Results: Baseline median (range) for SPV%, PPV% and PVI% were 9% (5–15%), 14% (8–37%) and 12% (7–22%), respectively. Maximal accuracies for SPV%, PPV% and PVI%, respectively, were obtained at 25 mL kg–1 hemorrhage with cut-off values of 16.1% (96% accuracy), 28.5% (90% accuracy) and 15.4% (72% accuracy). Accuracy of cut-off values increased proportionally with blood volume removed for all three dynamic indices. Inverse associations were observed between CO and PPV% [–0.49; 95% confidence interval (CI): –0.74 to –0.15; p < 0.007], SPV% [–0.34; 95% CI: –0.69 to 0.10; p = 0.131] and PVI% [–0.20; 95% CI: –0.57 to 0.13; p = 0.253]. Conclusions and clinical relevance: SPV% followed by PPV%, but not PVI%, are useful dynamic preload indices with negative correlation to CO in anesthetized horses.