BackgroundVeno-venous extracorporeal membrane oxygenation (VV ECMO) has become standard of care in patients with the most severe forms of acute respiratory distress syndrome. However, hemolysis and bleeding are one of the most frequent side effects, affecting mortality. Despite the widespread use of VV ECMO, current protocols lack detailed, in-vivo data-based recommendations for safe ECMO pump operating conditions. This study aims to comprehensively analyze the impact of VV ECMO pump operating conditions on hemolysis by combining in-silico modeling and clinical data analysis.MethodsWe combined data from 580 patients treated with VV ECMO in conjunction with numerical predictions of hemolysis using computational fluid dynamics and reduced order modeling of the Rotaflow (Getinge) and DP3 (Xenios) pumps. Blood trauma parameters across 94,779 pump operating points were associated with numerical predictions of shear induced hemolysis.ResultsMinimal hemolysis was observed at low pump pressures and low circuit resistance across all flow rates, whereas high pump pressures and circuit resistance consistently precipitated substantial hemolysis, irrespective of flow rate. However, the lower the flow rate, the more pronounced the influence of circuit resistance on hemolysis became. Numerical models validated against clinical data demonstrated a strong association (Spearman’s r = 0.8) between simulated and observed hemolysis, irrespective of the pump type.ConclusionsIntegrating in-silico predictions with clinical data provided a novel approach in understanding and potentially reducing blood trauma in VV ECMO. This study further demonstrated that a key factor in lowering side effects of ECMO support is the maintenance of low circuit resistance, including oxygenators with the lowest possible resistance, the shortest feasible circuit tubing, and cannulae with an optimal diameter.
Read full abstract