Dynamic Echocardiographic Trajectory Phenotypes During Veno-Arterial Extracorporeal Membrane Oxygenation and Their Association With In-Hospital Mortality: A Multicenter Cohort Study.
Li Z, Chen W, et al. • Journal of the American Heart Association • 2026
Dynamic echocardiographic trajectory phenotypes during veno-arterial extracorporeal membrane oxygenation identify distinct cardiac response patterns independently associated with in-hospital mortality, whereas static baseline phenotyping failed to stratify survival.
Key Findings
Results
Four distinct dynamic echocardiographic trajectory phenotypes were identified in patients receiving VA-ECMO for cardiogenic shock.
The four phenotypes were: Dyn-A ('Refractory Stagnation,' 17.0%), Dyn-B ('Isolated LV Recovery,' 14.9%), Dyn-C ('Biventricular Failure,' 6.1%), and Dyn-D ('Gradual Recovery,' 61.9%).
Latent class mixed models and probabilistic template-matching were applied to serial echocardiographic data including left ventricular ejection fraction, end-diastolic diameter, and right ventricular parameters.
The study analyzed 2034 adults receiving VA-ECMO from a multicenter registry.
Dyn-A was characterized by persistent systolic suppression and progressive dilation, while Dyn-D represented the largest group with gradual recovery.
Results
The high-risk Dyn-A phenotype ('Refractory Stagnation') was associated with significantly higher in-hospital mortality compared with the Dyn-B phenotype ('Isolated LV Recovery').
The adjusted hazard ratio for Dyn-A versus Dyn-B was 1.71 (95% CI 1.35–2.16).
The association remained significant after adjusting for baseline illness severity.
Associations with mortality were evaluated using multivariable Cox and Fine-Gray competing risk models.
Dyn-C ('Biventricular Failure') represented the smallest but presumably highest-risk group at 6.1% of the cohort.
Results
Static baseline echocardiographic phenotyping failed to stratify survival in VA-ECMO patients.
In contrast to dynamic trajectory phenotypes, static assessments at admission did not differentiate survival outcomes.
This finding highlights that single time-point assessments ignore dynamic cardiac evolution during ECMO support.
The authors conclude that 'longitudinal monitoring provides essential prognostic utility beyond static admission assessments.'
Discussion
Dynamic echocardiographic trajectories reflect divergent cardiac responses to VA-ECMO, ranging from recovery to progressive maladaptation.
The phenotypes captured both left ventricular and right ventricular responses over time, reflecting biventricular dynamics.
Dyn-C ('Biventricular Failure') represented a distinct phenotype involving failure of both ventricles, occurring in 6.1% of patients.
The trajectory approach captured the temporal evolution of cardiac function rather than a single snapshot, enabling differentiation of phenotypes with similar baseline profiles but divergent outcomes.
Methods
The study population consisted of 2034 adults with cardiogenic shock treated with VA-ECMO drawn from a multicenter registry.
This is described as a multicenter cohort study.
All patients received VA-ECMO specifically for cardiogenic shock.
Serial echocardiographic data were collected and analyzed longitudinally using latent class mixed models.
The large multicenter sample enabled identification of even relatively rare phenotypes such as Dyn-C (6.1% of patients).
What This Means
This research suggests that tracking how the heart changes over time during a form of life support called veno-arterial extracorporeal membrane oxygenation (VA-ECMO) — rather than just looking at a single measurement when a patient first arrives — provides critical information about who is likely to survive. Using repeated heart ultrasound (echocardiographic) measurements from over 2,000 patients at multiple hospitals, researchers identified four distinct patterns of how the heart responds to this support: a 'Gradual Recovery' group (the majority, about 62%), an 'Isolated Left Ventricle Recovery' group (about 15%), a 'Refractory Stagnation' group where the heart remained suppressed and continued to enlarge (about 17%), and a 'Biventricular Failure' group where both sides of the heart failed (about 6%). Patients whose hearts showed the 'Refractory Stagnation' pattern had 71% higher odds of dying in the hospital compared to those in the 'Isolated LV Recovery' group, even after accounting for how sick patients were at the start.
Importantly, looking only at how the heart appeared when patients were first admitted did not predict survival — only the dynamic, evolving patterns over time had prognostic value. This finding challenges current clinical practice, which often relies on a single assessment at admission to guide decision-making. The research suggests that continuous, serial monitoring of heart function during VA-ECMO support may be essential to identifying patients who are deteriorating and who might need escalated or altered treatment strategies, as opposed to those on a path to recovery.
Li Z, Chen W, Li C, Qiu G, Shi L, Duan W, et al.. (2026). Dynamic Echocardiographic Trajectory Phenotypes During Veno-Arterial Extracorporeal Membrane Oxygenation and Their Association With In-Hospital Mortality: A Multicenter Cohort Study.. Journal of the American Heart Association. https://doi.org/10.1161/JAHA.126.049942