Comparison of Echocardiographic Indices Used to Predict Fluid Responsiveness in Ventilated Patients

RATIONALE: Assessment of fluid responsiveness relies on dynamic echocardiographic parameters that have not yet been compared in large cohorts. OBJECTIVES: To determine the diagnostic accuracy of dynamic parameters used to predict fluid responsiveness in ventilated patients with a circulatory failure of any cause. METHODS: In this multicenter prospective study, respiratory variations of superior vena cava diameter (∆SVC) measured using transesophageal echocardiography, of inferior vena cava diameter (∆IVC) measured using transthoracic echocardiography, of the maximal Doppler velocity in left ventricular outflow tract (∆VmaxAo) measured using either approach, and pulse pressure variations (∆PP) were recorded with the patient in the semirecumbent position. In each patient, a passive leg raise was performed and an increase of aortic velocity time integral greater than or equal to 10% defined fluid responsiveness. MEASUREMENTS AND MAIN RESULTS: Among 540 patients (379 men; age, 65 ± 13 yr; Simplified Acute Physiological Score II, 59 ± 18; Sequential Organ Failure Assessment, 10 ± 3), 229 exhibited fluid responsiveness (42%). ∆PP, ∆VmaxAo, ∆SVC, and ∆IVC could be measured in 78.5%, 78.0%, 99.6%, and 78.1% of cases, respectively. ∆SVC greater than or equal to 21%, ∆VmaxAo greater than or equal to 10%, and ∆IVC greater than or equal to 8% had a sensitivity of 61% (95% confidence interval, 57-66%), 79% (75-83%), and 55% (50-59%), respectively, and a specificity of 84% (81-87%), 64% (59-69%), and 70% (66-75%), respectively. The area under the receiver operating characteristic curve of ∆SVC was significantly greater than that of ∆IVC (P = 0.02) and ∆PP (P = 0.01). CONCLUSIONS: ∆VmaxAo had the best sensitivity and ∆SVC the best specificity in predicting fluid responsiveness. ∆SVC had a greater diagnostic accuracy than ∆IVC and ∆PP, but its measurement requires transesophageal echocardiography.