The Predictive Value of Trendelenburg Position and Neck Ultrasound for Fluid Responsiveness in Prone ARDS Patients with VV-ECMO

Abstract Background In clinical practice, fluid administration is widely used to treat hypotension in patients undergoing veno-venous extracorporeal membrane oxygenation (VV-ECMO). However, volume expansion (VE) may aggravate acute respiratory distress syndrome (ARDS) and increase patient mortality, predicting fluid responsiveness is of great significance in the treatment of hypotension in patients undergoing VV-ECMO. Methods This prospective single-center study was conducted in a medical intensive care unit and included 51 VV-ECMO patients with ARDS in the prone position who required volume expansion due to hypotension. Stroke volume index variation (△SVI), carotid artery corrected flow time (FTc), and artery peak velocity variation (ΔVpeak) were taken before and during the Trendelenburg position or Volume expason is given. Fluid responsiveness was defined as a volume expansion-induced increase in ΔSVI of ≥ 15%. Results 33 patients (64.7%) were identified as fluid responders. The area under the receiver operating characteristic curve (ROC) for FTc and ΔVpeak induced by the Trendelenburg position to predict fluid responsiveness were 0.866 (95% confidence interval [CI] 0.755–0.977) and 0.833 (95% CI 0.716–0.949), respectively. The sensitivity at the optimal threshold of 331.5 ms for FTc was 84.85% (95% CI 69.1–93.4%), with a specificity of 83.33% (95% CI 60.8–94.2%). For ΔVpeak, the sensitivity at the optimal threshold of 10.1% was 81.82% (95% CI 65.9–91.4%), with a specificity of 77.78% (95% CI 54.9–91.0%). The grey zone for FTc and ΔVpeak included 29% and 45% of patients, respectively. Conclusions Changes in FTc and ΔVpeak, monitored through neck ultrasound and induced by the Trendelenburg position, are reliable indicators for predicting fluid responsiveness in VV-ECMO patients with ARDS in the prone position. Furthermore, FTc demonstrates superior predictive value compared to ΔVpeak.