Smith KL, Liddle DW, Schiff M, Fishbein KT, Ghassemzadeh R, Feingold B, Olivieri LJ. Utility of Echocardiography in 173 VA ECMO Clamp Trials in Babies and Children. Pediatric Cardiology. 2026. doi:10.1007/s00246-026-04245-9.
PubMed: https://pubmed.ncbi.nlm.nih.gov/41925852/
Take-Home Points
- This retrospective single-center study evaluated 173 VA-ECMO clamp-trial echocardiograms in 103 pediatric patients undergoing assessment for ECMO decannulation.
- In children with cardiomyopathy/myocarditis, higher left ventricular ejection fraction (LVEF) predicted successful decannulation. Every 5% decrease in LVEF reduced the odds of successful decannulation by 29%.
- In persistent pulmonary hypertension of the newborn (PPHN), lower tricuspid regurgitation (TR) velocity and less than mild atrioventricular valve regurgitation were strongly associated with successful decannulation.
- In the congenital heart disease (CHD) cohort, no echocardiographic parameter reliably predicted successful decannulation.
- Among biventricular CHD patients, longer ECMO duration was associated with reduced likelihood of successful decannulation.
- The study reinforces that echocardiographic predictors of ECMO readiness are highly diagnosis-specific and may need to be interpreted differently across pediatric populations.
Commentary from Dr. Varun Aggarwal (Wilmington, DE, USA), editor-in-chief of Congenital Heart Disease Journal Watch:
Few moments in pediatric cardiac intensive care generate more anxiety than the decision to discontinue extracorporeal membrane oxygenation (ECMO). Decannulate too early and cardiopulmonary collapse may follow. Wait too long and patients become increasingly exposed to bleeding, thrombosis, infection, neurologic injury, and end-organ dysfunction. Despite decades of ECMO experience, clinicians continue to struggle with a deceptively simple question: how do we know when a child is ready?
Smith and colleagues tackle this challenge by examining one of the most used tools in the decannulation process, echocardiography. Their study represents the largest single-center analysis to date evaluating echocardiographic findings obtained during pediatric VA-ECMO clamp trials and provides important insight into how physiology-specific assessment may guide decision-making.
The investigators reviewed 173 clamp-trial echocardiograms performed in 103 children supported with VA-ECMO over a 12-year period. Rather than treating all ECMO patients as a homogeneous group, they categorized patients according to underlying physiology: cardiomyopathy/myocarditis (CM), persistent pulmonary hypertension of the newborn (PPHN), and congenital heart disease (CHD). This distinction may represent the study’s greatest strength.
The ECMO population is remarkably heterogeneous. A neonate with severe PPHN, a child with fulminant myocarditis, and an infant recovering from stage-one palliation all arrive on ECMO for fundamentally different reasons. It is therefore logical that the echocardiographic parameters predicting recovery might differ as well. Yet prior studies have often pooled these populations together, potentially obscuring clinically useful signals.
The findings in the cardiomyopathy group are perhaps the most intuitive. Children who tolerated decannulation demonstrated better left ventricular systolic function, and every 5% reduction in LVEF significantly lowered the likelihood of successful cardiac support withdrawal. This reinforces what clinicians have long suspected: for myocarditis and cardiomyopathy patients, recovery of pump function remains the central determinant of successful ECMO liberation.
Interestingly, no right ventricular parameter predicted outcome in this group. This contrasts with portions of the adult ECMO literature, where right ventricular performance often carries prognostic significance. The findings suggest that in pediatric myocarditis and cardiomyopathy, traditional left ventricular recovery metrics may remain the most clinically useful echocardiographic endpoint when considering decannulation.
The PPHN cohort yielded an entirely different set of predictors. Here, lower tricuspid regurgitation velocity and less atrioventricular valve regurgitation strongly correlated with successful decannulation. Physiologically, this makes excellent sense. Elevated tricuspid regurgitation velocity reflects increased right ventricular pressure and persistent pulmonary vascular disease. Likewise, worsening atrioventricular valve insufficiency may represent ongoing right ventricular strain. Together, these measurements act as indirect markers of unresolved pulmonary hypertension.
The practical implication is important. For neonates with PPHN, successful ECMO decannulation may depend less on left ventricular function and more on evidence of pulmonary vascular recovery. Rather than focusing primarily on ventricular performance, clinicians may gain greater insight by carefully monitoring right-sided hemodynamic markers.
The most intriguing result, however, may be the absence of significant findings in congenital heart disease.
Despite extensive analysis, the authors could not identify a reproducible echocardiographic predictor of successful decannulation in either biventricular or single-ventricle CHD patients. At first glance, this seems disappointing. Yet the observation probably reflects the extraordinary complexity of congenital heart disease rather than a failure of echocardiography itself.
The CHD population in this study included a wide variety of diagnoses, surgical repairs, ventricular morphologies, and postoperative physiologies. The factors that determine successful ECMO separation following Norwood palliation may differ dramatically from those relevant after tetralogy of Fallot repair, Ebstein anomaly surgery, or neonatal arterial switch procedures. Expecting a single echocardiographic parameter to predict outcomes across such diverse anatomy may be unrealistic.
Indeed, this finding highlights a broader truth in congenital heart disease. Standardized metrics often work poorly in heterogeneous congenital populations. What matters is not merely ventricular function but the interaction among ventricular performance, loading conditions, residual lesions, pulmonary vascular resistance, atrioventricular valve competence, shunting physiology, and surgical anatomy. The absence of a universal predictor may itself be a clinically important finding.
For ACHD and congenital programs, the study raises an important research priority. Future investigations may need to move beyond broad CHD categories and instead evaluate lesion-specific predictors. Postoperative hypoplastic left heart syndrome, tetralogy of Fallot, transposition of the great arteries, and atrioventricular septal defects likely require distinct decannulation frameworks.
Another valuable aspect of this analysis is its challenge to indiscriminate echocardiographic utilization. The authors correctly note that current guidelines provide limited direction regarding frequency and timing of ECMO-weaning echocardiograms. By identifying physiology-specific markers associated with successful decannulation, this work may help develop more focused and efficient imaging protocols.
Importantly, the study should not be interpreted as suggesting that echocardiography alone determines ECMO readiness. Decannulation decisions integrate hemodynamics, laboratory trends, end-organ function, vasoactive requirements, respiratory status, and overall clinical trajectory. Echocardiography remains one component of a broader assessment. Nevertheless, objective markers that improve confidence in decision-making are particularly valuable when the consequences of error are so high.
Ultimately, this study advances the field by demonstrating that successful ECMO liberation is highly dependent on underlying disease physiology. The right echocardiographic question varies according to the diagnosis. For myocarditis and cardiomyopathy, clinicians should focus on myocardial recovery. For PPHN, pulmonary vascular unloading appears paramount. For congenital heart disease, however, the search for reliable decannulation predictors remains unfinished.
Clinical Perspective
Echocardiography provides meaningful guidance during pediatric VA-ECMO clamp trials, but predictive parameters differ according to diagnosis. Higher LVEF predicts successful decannulation in cardiomyopathy/myocarditis, whereas lower TR velocity and minimal AV valve regurgitation are favorable markers in PPHN. No consistent echocardiographic predictor emerged for congenital heart disease patients, underscoring the need for lesion-specific investigations and individualized decannulation strategies.
