Dynamic variables to predict fluid responsiveness in young children

The evidence that dynamic variables predict fluid responsiveness in young children is limited by conflicting research results.

Sixty patients, 1-3 years of age, undergoing major neurosurgery, received 10 mL/kg of Ringer's solution over 10 min after anesthesia induction. Respiratory variation in aortic blood flow peak velocity (∆Vpeak), plethysmographic variability index (PVI), FloTrac/Vigileo-derived stroke volume variation (SVV), dynamic arterial elastance (Ea_dyn ), and pulse pressure variation (PPV) were measured before and following fluid loading. An increase in the cardiac index (CI) of ≥10% following fluid loading identified fluid "responders."

Twenty-six patients (43.3%) were fluid responders. Baseline ∆Vpeak was an excellent predictor of a CI increase following fluid loading with an area under the receiver operating characteristic curve (AUROC) of 0.982 (p < 0.001). The PVI showed fair diagnostic accuracy for CI-fluid responsiveness (AUROC 0.775, p < 0.001). Baseline ∆Vpeak and PVI cutoff values were 9.6% and 15%, respectively. PPV, SVV, and Ea_dyn were not predictors or were poor predictors for CI-fluid responsiveness (AUROC 0.669, 0.653, and 0.533, respectively).

Volume-based PVI and ∆Vpeak showed acceptable reliability for fluid responsiveness prediction in young children undergoing major neurosurgery, whereas pressure-based SVV using FloTrac/Vigileo, Ea_dyn , and PPV did not.