Abstract 13715: Poor Concordance Between Respiratory Variation in Inferior Vena Cava and Surrogates of Cardiac Output Measurements for Determination of Fluid Responsiveness

IntroductionFluid responsiveness in critically ill patients is paramount but often challenging to determine. Point-of-care ultrasound (POC-US) is increasingly being used as a tool to for its assessment. Several POC-US parameters including respiratory variation in inferior vena cava diameter (IVCd) and surrogates of cardiac output variability, such as carotid artery flow variation with passive leg raise (PLR) maneuver have been proposed.HypothesisThe purpose of this study was to determine if these measurements are in agreement when assessing for fluid responsiveness.MethodsRespiratory variation in IVCd and change in carotid blood flow with PLR was measured in 14 consecutive ICU patients. All measurements were performed by a critical care attending or fellow trained in POC-US. For spontaneously breathing patients IVC collapsibility index (IVCci) of > 40%, and in patients on controlled mechanical ventilation IVC dispensability index (Divc) of > 18% were used as markers of fluid responsiveness. An increase in carotid blood flow ≥ 20% with passive leg raise was used for comparison. Concordance of data between the two tests was determined using Cohen’s kappa coefficient (k). This project was approved by Einstein IRB # 2016-7070.ResultsMeasurements were performed on 14 ICU patients. Two patients were mechanically ventilated. Three had moderate to severe pulmonary hypertension. Three patient’s IVC measurements were not performed due to limited ultrasound windows. With IVCd variation assessment 10 met criteria for fluid responsiveness. Carotid flow variation with PLR was obtained in 12 patients, of them 5 met criteria for fluid responsiveness. Comparing to carotid blood flow, IVCd variation had a sensitivity, positive predictive value, and negative predictive value of 80%, 40% and 0% respectively. There was extremely poor concordance between the two measurements as assessed by k measurement (k = -0.18).ConclusionsPreload based determinations of fluid responsiveness are not reliable when used for ICU patients, instead cardiac output dependent assessments should be used and training in these measurements emphasized.