Background
Currently, there are no established echocardiographic (echo) or hemodynamic predictors of mortality after weaning venoarterial extracorporeal membrane oxygenation (ECMO) in children. We wished to determine which measurements predict mortality.
Methods
Over 3 years, we prospectively assessed six echo and six hemodynamic variables at 3-5 circuit rates while weaning ECMO flow. Hemodynamic measurements were: heart rate, inotropic score, arterio-venous oxygen difference (AV-O2), pulse pressure, oxygenation index (OI), and lactate. Echo variables included: shortening/ejection fraction, outflow tract Doppler-derived stroke distance (VTI), degree of atrioventricular valve regurgitation, longitudinal strain (GLS), and circumferential strain (GCS). Patients were stratified into those who died or required heart transplant (Gr1), and those who did not (Gr2). For each patient, we compared the change for each variable between full versus minimum ECMO flow for each group.
Results
We enrolled 21 patients ranging in age 0.02-15 years. Five had dilated cardiomyopathy, and 16 had structural heart disease with severe ventricular dysfunction. Thirteen of 21 patients (62%) comprised Gr1, including two patients with heart transplants. Eight patients constituted Gr2. Gr1 patients had a significantly greater increase in OI (35% mean increase, p<0.01) off ECMO compared to full flow, but no change in VTI or AV-O2. In Gr2 VTI increased (31% mean increase, p<0.01) with no change in AV-O2 or OI. Pulse pressure increased modestly with flow reduction only in Gr1 (p<0.01).
Conclusion
Failure to augment VTI or an increase in OI during the ECMO weaning is associated with poor outcomes in children. We propose that these measurements should be performed during ECMO wean, as they may discriminate who will require alternative methods of circulatory support for survival.