Pulse pressure variation as a guide for volume expansion in dogs undergoing orthopedic surgery

To investigate whether pulse pressure variation (PPV) can predict fluid responsiveness in healthy dogs during clinical surgery.

Prospective clinical study.

Thirty-three isoflurane-anesthetized dogs with arterial hypotension during orthopedic surgery.

Fluid challenge with lactated Ringer's solution (15 mL kg^-1 in 15 minutes) was administered in mechanically ventilated dogs (tidal volume 10 mL kg^-1) with hypotension [mean arterial pressure (MAP) < 65 mmHg]. The volume expansion was considered effective if cardiac output (CO; transesophageal Doppler) increased by ≥ 15%. Cardiopulmonary data were analyzed using two-way ANOVA, receiver operating characteristics (ROC) curves and Spearman coefficient; p < 0.05 was considered significant.

Effective volume expansion, mean ± standard deviation 42 ± 4% increase in CO (p < 0.0001) was observed in 76% of the dogs, resulting in a decrease in PPV (p < 0.0001) and increase in MAP (p < 0.0001), central venous pressure (CVP; p = 0.02) and ejection fraction (p < 0.0001) compared with before the fluid challenge. None of these changes occurred when volume expansion resulted in a nonsignificant CO increase of 4 ± 5%. No significant differences were observed in blood gas analysis between responsive and nonresponsive dogs. The increase in CO was correlated with the decrease in PPV (r = -0.65; p < 0.0001) but absolute values of CO and PPV were not correlated. The PPV performance (ROC curve area: 0.89 ± 0.06, p = 0.0011) was better than that of CVP (ROC curve area: 0.54 ± 0.12) and MAP (ROC curve area: 0.59 ± 0.13) to predict fluid responsiveness. The best cut-off for PPV to distinguish responders and nonresponders was 15% (50% sensitivity and 96% specificity).

In mechanically ventilated, healthy, isoflurane-anesthetized dogs, PPV predicted fluid responsiveness to volume expansion, and MAP and CVP did not show such applicability.