Hospital-level volume in extracorporeal membrane oxygenation cases and death or disability at 6 months.

Summary

In a binational registry-embedded cohort (n=1,232), ECMO center volume (>30 vs ≤30 cases/year) was not associated with death or new disability at 6 months across VV-, VA-, and ECPR-ECMO after multivariable adjustment. Sensitivity analyses, including exclusion of inter-hospital transfers, confirmed the null association, implicating coordinated regional care and training.

Key Findings

• No difference in 6‑month death or new disability between high- and low-volume centers for VV-ECMO (OR 1.09, 95% CI 0.65–1.83; p=0.744).
• No difference for VA-ECMO (OR 1.10, 95% CI 0.66–1.84; p=0.708) or ECPR (OR 1.38, 95% CI 0.37–5.08; p=0.629).
• Findings were consistent across sensitivity analyses, including exclusion of transferred patients; total n=1,232 (663 high-volume, 569 low-volume).

Clinical Implications

Health systems should prioritize coordinated networks, standardized protocols, and inter-hospital transfers/training over rigid volume thresholds when organizing ECMO services and benchmarking outcomes.

Limitations

• Observational design leaves potential residual confounding.
• Generalizability may be limited to regions with coordinated ECMO networks (Australia/New Zealand).

Future Directions

Prospective multicenter studies to identify which components of coordinated networks (e.g., protocols, simulation training, transfer timing) drive outcome equivalence; incorporation of patient-reported outcomes and cost-effectiveness.