ECMO after the Norwood procedure identifies a particularly vulnerable group of infants with single-ventricle heart disease. A systematic review and meta-analysis by Almehandi and colleagues found substantially higher early mortality and longer intensive care stays among infants requiring postoperative extracorporeal support. These findings describe prognosis after clinical deterioration; they do not establish that ECMO itself causes the excess risk.
What the Review Included
The review included eight observational studies reporting 2,612 patients, approximately 15% of whom received ECMO after Norwood palliation. The Norwood operation is the first stage of palliation for hypoplastic left heart syndrome and related single-ventricle conditions. Subsequent stages include the bidirectional Glenn operation and Fontan completion.
The investigators compared postoperative ECMO recipients with infants managed without ECMO. Preoperative ECMO, hybrid stage-one palliation, and studies without a comparison group were excluded. The authors recognized overlapping institutional cohorts and avoided combining those overlapping reports within the same forest plot. Different outcomes drew on different subsets of studies, so the full review population should not be used as the denominator for every result.
ECMO After the Norwood Procedure: Early Outcomes
The short-term mortality analysis included four studies and 895 patients. The detailed results reported a pooled risk ratio of 5.53, with a 95% confidence interval of 4.30–7.11. Statistical heterogeneity for that analysis was low, but similar effect estimates across studies do not eliminate confounding by illness severity.
ECMO is generally initiated because an infant cannot separate from bypass or develops profound circulatory instability, refractory hypoxemia, or cardiac arrest. These indications carry different risks, and the review could not reliably separate their outcomes. The comparison therefore does not answer what would have happened to the same critically ill infant if ECMO had been withheld.
| Outcome | Reported finding | Interpretation |
|---|---|---|
| Short-term mortality | Risk ratio 5.53 in the detailed results | Strong association with high-risk clinical status; not a causal estimate of ECMO harm. |
| Cardiopulmonary bypass duration | Approximately 28 minutes longer in ECMO recipients | May reflect surgical complexity and difficulty separating from bypass. |
| Intensive care stay | Approximately 10.6 days longer | Greater resource use, with substantial variation across studies. |
| Total hospital stay | No statistically clear pooled difference | Wide uncertainty and competing effects of early death limit interpretation. |
Progression to Glenn and Fontan
The authors described subsequent palliation rates as comparable, but the estimates were imprecise. Glenn completion was assessed in three studies with 611 patients, with an odds ratio of 0.15 and a 95% confidence interval of 0.02–1.06. Fontan completion was assessed in three studies with 497 patients, with an odds ratio of 0.18 and a confidence interval of 0.01–2.98.
These wide intervals do not demonstrate equivalence between groups. Differences in follow-up, inclusion of early nonsurvivors, and selection of survivors can materially affect later-stage comparisons. The practical message is that progression to further palliation remains possible after ECMO, while the likelihood for an individual infant cannot be reliably estimated from these pooled comparisons alone.
Relevance for Perfusion and Cardiac Intensive Care Teams
The review emphasizes the importance of documenting why support was initiated and assessing potentially correctable residual lesions. Failure to separate from CPB, extracorporeal cardiopulmonary resuscitation, and postoperative hypoxemia should not be treated as interchangeable patient populations when evaluating program outcomes.
Longer bypass duration may identify a more difficult operation or persistent hemodynamic problems, but this association does not prove that reducing bypass time alone would prevent ECMO or improve survival. Likewise, the pooled evidence does not establish an optimal cannulation time, shunt-management strategy, flow target, or anticoagulation protocol.
For quality improvement and family discussions, the most useful interpretation combines the reason for ECMO, residual anatomy, organ injury, recovery trajectory, and institutional experience. Crude comparisons with infants who never required rescue support cannot replace that clinical context.
Limitations and Interpretation
All included studies were observational, and the evidence was largely retrospective and drawn from small, single-center cohorts. The authors identified serious risk of bias in two studies and moderate concerns in most others. Variation in ECMO indications, shunt configuration, treatment era, outcome definitions, and follow-up restricted meaningful subgroup analysis.
Several outcomes showed substantial heterogeneity, and the authors reported evidence suggestive of small-study effects. With few studies, publication-bias testing itself is uncertain. The report also contains inconsistencies between its abstract and detailed results, including the confidence interval for mortality, and some sensitivity-analysis wording conflicts with the accompanying statistics. The mortality interval presented above follows the detailed Results section; borderline secondary findings should be interpreted cautiously.
The review is useful for describing the severity and resource needs of postoperative ECMO patients. It cannot establish the survival benefit of ECMO, prove that later-stage outcomes are equivalent, or determine the best rescue strategy.
Original Article
Almehandi A, Altarkait A, Ali Y, et al. Outcomes after extracorporeal membrane oxygenation following the Norwood procedure: a systematic review and meta-analysis. Journal of Cardiothoracic Surgery. 2026;21:514. doi:10.1186/s13019-026-04232-4.
This independent educational summary accompanies the unchanged source article, available under the Creative Commons Attribution 4.0 license. The cover is a conceptual editorial illustration, not a technical circuit guide.
