Extracorporeal Membrane Oxygenation in the Interstage Period After Norwood and Related Stage I Palliations in Infants With Single Ventricle: A Systematic Review of Incidence, Risk Factors, and Outcomes Authors: Dongzhe Li, Stephen J Bottrell, Stephen B Horton, Christian P Brizard, Salvatore Pepe Published: 2026 ABSTRACT AbstractBackgroundNeonates with congenital univentricular conditions require staged palliation culminating in a Fontan circulation to support survival. The first‐stage Norwood procedure carries substantial risk of myocardial dysfunction, frequently necessitating extracorporeal membrane oxygenation (ECMO) support. The aim of this systematic review was to examine the incidence, risk factors, and outcomes of post‐Norwood ECMO use during the interstage period.MethodsWe systematically searched MEDLINE, Embase, Cochrane Library, and ClinicalTrials.gov for studies published between January 1990 and March 2026 reporting ECMO use following Norwood procedures. Studies evaluating ECMO as part of postoperative management strategies were included.ResultsFrom 1509 identified studies, 15 met inclusion criteria, encompassing 4030 patients. The pooled incidence of post‐Norwood ECMO was 18.1% (685/3775 patients), ranging from 15.5% to 29.4% across studies. In‐hospital survival among ECMO‐supported patients ranged from 28.6% to 50%, with no clear temporal improvement. Risk factors for ECMO initiation included low birth weight (<2.5 kg), prolonged cardiopulmonary bypass time, moderate‐to‐severe preoperative atrioventricular valve regurgitation, and complex cardiac anatomy (aortic atresia, diminutive ascending aorta). Factors associated with poor outcomes included prolonged ECMO duration and ECMO‐related complications, particularly neurological events, renal failure, and bleeding.ConclusionsDespite increasing post‐Norwood ECMO use, survival outcomes remain poor. Identified risk factors may inform prospective risk stratification strategies, early identification of high‐risk patients, and evidence‐based decision‐making regarding ECMO initiation, continuation, or withdrawal in this vulnerable population. Significant heterogeneity across studies precluded meta‐analysis, highlighting the need for standardized reporting in future research.Keywords: extracorporeal membrane oxygenation, Fontan, hypoplastic left heart syndrome, Norwood procedureSubject Categories: Cardiovascular Surgery, Cardiopulmonary Resuscitation and Emergency Cardiac Care, Congenital Heart Disease Nonstandard Abbreviations and Acronyms atrioventricular valve regurgitation hypoplastic left heart syndrome Clinical Perspective What Is New? This systematic review consolidates three decades of studies on extracorporeal membrane oxygenation use during the interstage period after the Norwood procedure in neonates with single ventricle physiology; the incidence of extracorporeal membrane oxygenation in this vulnerable population has increased over years despite the survival remains poor despite technological and surgical advances. This systematic review consolidates three decades of studies on extracorporeal membrane oxygenation use during the interstage period after the Norwood procedure in neonates with single ventricle physiology; the incidence of extracorporeal membrane oxygenation in this vulnerable population has increased over years despite the survival remains poor despite technological and surgical advances. What Are the Clinical Implications? The findings from the review have clinical implications for perioperative risk stratification using physiological (eg, birth weight) and anatomical characteristics (eg, hypoplastic left heart syndrome variants, ascending aorta diameter). The findings from the review have clinical implications for perioperative risk stratification using physiological (eg, birth weight) and anatomical characteristics (eg, hypoplastic left heart syndrome variants, ascending aorta diameter). Early and proactive extracorporeal membrane oxygenation initiation may have survival benefits, but further research is needed to define optimal timing and patient selection; the review underscores the need for standardized definitions and multicenter collaboration to refine patient selection and guide withdrawal or transition decisions. Early and proactive extracorporeal membrane oxygenation initiation may have survival benefits, but further research is needed to define optimal timing and patient selection; the review underscores the need for standardized definitions and multicenter collaboration to refine patient selection and guide withdrawal or transition decisions. Neonates with congenital single functional ventricle conditions, such as hypoplastic left heart syndrome (HLHS) and related variants, require palliative surgery to establish the Fontan circulation over 3 stages for survival. The interstage period between the first‐stage Norwood operation and second‐stage bidirectional cavopulmonary shunt carries the highest mortality risk of all palliation stages, with in‐hospital mortality rising substantially from ∼11% in patients without extracorporeal membrane oxygenation to as high as 58% in those requiring ECMO support. 1 , 2 Postoperative ECMO requirement reflects underlying hemodynamic dysfunction, which includes low cardiac output syndrome, coronary hypoperfusion, pulmonary overcirculation, and renal dysfunction, and serves as a strong predictor of mortality in this vulnerable population. However, the clinical characteristics and critical risk factors affecting ECMO requirement after Norwood procedures remain incompletely defined. Current research is limited by small sample sizes, institutional heterogeneity, and variable surgical volumes, leading to significant outcome variations across centers. Recent reports demonstrate marked survival variability, ranging from 68% 3 to 33% 4 across institutions, and registry data from the ELSO (Extracorporeal Life Support Organization) reported overall survival of 31%. 5 This heterogeneity likely reflects differences in patient selection criteria, institutional protocols, and surgical expertise. Therefore, this systematic review aims to comprehensively assess the reported incidence of post‐Norwood ECMO support, associated mortality rates, risk factors predicting ECMO requirement, and adverse outcomes during the interstage period. Due to inconsistent or absent reporting of the specific timing of ECMO deployment across studies, “interstage” ECMO following the Norwood procedure refers to ECMO initiated only during the time from the operating room to before Stage II palliation. This specifically covers ECMO initiated within the same hospitalization as the Norwood procedure, typically due to failure to separate from bypass, low cardiac output syndrome, or cardiac arrest in the early postoperative period, and later ECMO initiated after intensive care discharge after the Norwood procedure but before Stage II palliation. By examining current evidence, we seek to identify knowledge gaps, inform clinical practice, and guide future research to improve outcomes during this critical phase of single ventricle palliation. METHODS This systematic review was conducted according to Preferred Reporting Items for Systematic Reviews and Meta‐Analyses guidelines (Table S1) and prospectively registered with the International Prospective Register of Systematic Reviews (CRD420251124759). Institutional human research ethics review board approval for this systematic review of published studies was not required. We searched Embase, MEDLINE (PubMed), Cochrane Library, and ClinicalTrials.gov to March 31, 2026, using systematic keyword (Medical Subject Headings) search strategies (Table S2). The authors declare that all supporting data are available within the article and Supplemental Material. Eligibility Criteria Inclusion of articles meeting PICOST criteria (Population, Intervention, Comparison, Outcome, Study design, Time frame) was as follows: Population: infants <6 months old with univentricular congenital condition requiring staged surgical palliation; Intervention: ECMO support following the Norwood procedure and related Stage I palliations (including hybrid procedures and other forms of arch reconstruction for single ventricle physiology), during the interstage period up to (but not including) Stage II or subsequent interventions; Comparison: Standard care without ECMO; survivors versus nonsurvivors; Outcomes: ECMO incidence, survival, anatomical/procedural details, postoperative morbidities; Study design: randomized controlled trials and nonrandomized studies (including cohort studies); Time frame: publications from 1990, coinciding with implementation of ECMO technology in the setting of pediatric cardiac surgery. Exclusion criteria were conference papers and abstracts; review articles, case reports or case series with n<10; non‐English language; stage 2 or stage 3 operations; when >1 study used the same database, the study with the largest cohort was selected for data extraction. Identified references were exported into Covidence for screening of title, abstracts, and full texts by 2 independent reviewers (S.P. and D.L.), based on eligibility criteria. Disagreements between reviewers regarding inclusion were resolved by consensus. Study Selection and Data Extraction Two independent reviewers (S.P., D.L.) screened titles, abstracts, and full texts using Covidence (www.covidence.org), Veritas Health Innovation, Melbourne, Australia. Data extraction encompassed study characteristics, patient demographics, cardiac diagnoses, procedural details (shunt types, cardiopulmonary bypass time, aortic cross clamp time, hypothermic circulatory arrest time), and clinical outcomes (incidence and duration of post‐Norwood ECMO support, complications, length of intensive care unit stay, hospital stay, mortality). Studies were categorized into 2 groups: those comparing ECMO versus non‐ECMO patients and those examining survivors versus nonsurvivors within ECMO cohorts. Statistical Analysis Study quality was assessed using the Newcastle‐Ottawa Scale for cohort and case–control studies (converted to Agency for Healthcare Research and Quality standards) and Cochrane Risk of Bias 2 tool for randomized trials. Continuous variables are presented as mean±SD or median with interquartile range, and categorical variables as counts or percentages. Where appropriate, Fisher's exact test and chi‐square test were used to calculate P values for categorical variables. Meta‐analysis was not performed. This was due to 4 major issues related to extensive data heterogeneity and missing data: (1) clinical heterogeneity over varied surgical eras (1990s–2020s) and populations (some studies excluded hybrid procedures, 6 , 7 others included them 3 ); (2) study design heterogeneity (retrospective cohorts, randomized clinical trial, case–control) and inconsistent ECMO definitions (ie, cardiopulmonary bypass (CPB) weaning failure excluded, 7 only extracorporeal cardiopulmonary resuscitation included 8 ); (3) outcome time points defined differently (ie, in‐hospital versus interstage versus 30‐day mortality) and with wide outcome ranges (28.6%–68% survival); and (4) different statistical data formats (ie, means versus medians, categorical versus continuous). RESULTS Study Characteristics From 1506 identified studies, 15 met inclusion criteria, encompassing 4930 patients from predominantly single‐center retrospective cohorts (Figure 1). Study characteristics and preprocedural details are summarized in Tables 1 and 2. Twelve studies were single‐center institutional reviews, and 3 used multicenter registries: the SVR (Single Ventricular Reconstruction) trial, 10 ELSO registry, 5 and ECHSA (European Congenital Heart Surgeons Association) database. 12 Study designs comprised 13 retrospective cohorts, 1 case–control study, 6 and 1 randomized controlled trial. 10 Preferred Reporting Items for Systematic Reviews and Meta‐Analyses flow diagram of data search, screening, and final inclusions. Table 1Characteristics of Included StudiesAuthorYSample sizeSource of dataComparing groupsRecruitment periodStudy designCardiac diagnosisRelated stage I palliation (inclusion/exclusion)ECMO definitionMortality end pointsFriedland‐Little et al 6 2014255University of MichiganECMO and non‐ECMO2001–2010Retrospective case–controlHLHS, UAVC, DORV/HLV, TA‐dextro‐transposition of the great arteries, DILV, TA/normally related great arteries/hypoplastic aortic archHybrid procedure excluded; DKS/arch repair includedAll post‐Norwood ECMONot reportedDe Jesus‐Brugman et al 3 201985Riley Hospital for ChildrenECMO and non‐ECMO2009–2017Retrospective cohortHLHS, UAVC, DILV, Heterotaxy syndromeHybrid procedure included; DKS excludedAll post‐Norwood ECMOIn‐hospital mortality and 30‐d mortalityMayr et al 9 2022257German Heart Centre MunichECMO and non‐ECMO2007–2019Retrospective cohortHLHS, UAVC, DORV/HLV, DILVOnly NorwoodAll post‐Norwood ECMOIn‐hospital mortality and 30‐d mortalityBeshish et al 7 , * 2023269Children's Healthcare of AtlantaECMO and non‐ECMO2010–2013Retrospective cohortHLHS, DORV/HLV, DILVHybrid procedure excludedECMO due to failure to wean off from CPB excludedIn‐hospital mortality and 30‐d mortalityTabbutt et al 10 2012549Single Ventricle Reconstruction TrialECMO and non‐ECMO2005–2008Randomized controlled trial (Blalock‐Taussig and Sano shunt)Not specified in detailOnly NorwoodAll post‐Norwood ECMONot reportedSengupta et al 11 2022500Boston Children's HospitalECMO and non‐ECMO1997–2017Retrospective cohortHLHS, DORV, double outlet left ventricle, UAVC, double inlet right ventricleOnly NorwoodAll post‐Norwood ECMONot reportedCattapan et al 12 20252115European Congenital Heart Surgeons AssociationECMO and non‐ECMO2013–2022Retrospective cohortNot specified in detailOnly NorwoodAll post‐Norwood ECMONot reportedDeBrunner et al 13 201232Riley Hospital for ChildrenSurvivors and non‐survivors in ECMO population1994–2008Retrospective cohortHLHS, UAVCOnly NorwoodAll post‐Norwood ECMOIn‐hospital and 30‐d mortality† Hoskote et al 14 200616Toronto Hospital for Sick ChildrenSurvivors and nonsurvivors in ECMO population1997–2003Retrospective cohortHLHS, dilated cardiomyopathyOnly NorwoodAll post‐Norwood ECMOIn‐hospital mortalityPizarro et al 15 200112Not reportedSurvivors and nonsurvivors in ECMO population1998–1999Retrospective cohortHLHS, TA/TGA, DILV/TGA/aortic hypoplasia/mitral stenosis, dextrocardia/left atrial isomerim/TGATotal APVC repair; Arch repair/PABAll post‐Norwood ECMOIn‐hospital mortalityHoashi et al 16 201414National Cerebral and Cardiovascular Centre, JapanSurvivors and nonsurvivors in ECMO population2007–2013Retrospective cohortNot specified in detailBilateral PAB includedAll post‐Norwood ECMOIn‐hospital mortalityAlsoufi et al 2 , * 201438Children's Healthcare of AtlantaSurvivors and nonsurvivors in ECMO population2002–2012Retrospective cohortNot specified in detailOnly NorwoodAll post‐Norwood ECMOIn‐hospital mortalitySherwin et al 3 2012738Extracorporeal Life Support OrganizationSurvivors and nonsurvivors in ECMO population2000–2009Retrospective cohortOnly HLHSOnly NorwoodELSO registry data, all typesIn‐hospital mortalityRavishankar et al 17 200636The Children's Hospital of PhiladelphiaSurvivors and nonsurvivors in ECMO population1998–2005Retrospective cohortHLHS, UAVC, DILV/coarctation, TA/TGA, TGA/hypoplastic right ventricle, coarctationOnly NorwoodAll post‐Norwood ECMOIn‐hospital mortalityPolimenkos et al 8 201114The Heart Institution for ChildrenSurvivors and nonsurvivors in ECMO population2007–2009Retrospective cohortHLHS, DORV/mitral atresia/aortic atresia/interrupted aortic arch, DILV/TGA/MA/hypoplastic aortic arch, UAVC/pulmonary atresia/right isomerism/obstructed APVCDKS; TAPVC repair: total anomalous pulmonary venous connectionExtracorporeal cardiopulmonary resuscitation onlyIn‐hospital mortalityAPVC indicates anomalous pulmonary venous connection; DILV, double inlet right ventricle; DKS, Damus‐Kanye‐Stansel procedure; DORV, double outlet right ventricle; ECMO, extracorporeal membrane oxygenation; HLHS, hypoplastic left heart syndrome; HLV, hypoplastic left ventricle; PAB, pulmonary artery banding; TA, trucuspid atresia; TGA, transposition of great arteries; and UAVC, unbalanced atrioventricular canal.*Alsoufi et al 2 and Beshish et al 7 used the same data source. For ECMO vs non‐ECMO comparisons, data from Beshish et al were used due to its larger sample size, and Alsoufi et al was included for survivor vs nonsurvivor analysis. † Only in‐hospital mortality data was used for cross‐comparison. Characteristics of Included Studies APVC indicates anomalous pulmonary venous connection; DILV, double inlet right ventricle; DKS, Damus‐Kanye‐Stansel procedure; DORV, double outlet right ventricle; ECMO, extracorporeal membrane oxygenation; HLHS, hypoplastic left heart syndrome; HLV, hypoplastic left ventricle; PAB, pulmonary artery banding; TA, trucuspid atresia; TGA, transposition of great arteries; and UAVC, unbalanced atrioventricular canal. Alsoufi et al 2 and Beshish et al 7 used the same data source. For ECMO vs non‐ECMO comparisons, data from Beshish et al were used due to its larger sample size, and Alsoufi et al was included for survivor vs nonsurvivor analysis. Only in‐hospital mortality data was used for cross‐comparison. Table 2Baseline Characteristics and Preprocedural Details of ECMO and Non‐ECMO Patients Following the Norwood ProcedureVariableFriedland‐Little 6 , * De Jesus‐Brugman 3 Mayr 9 Beshish 7 Tabbutt 10 , * Sengupta 11 Cattapan 12 Post‐Norwood ECMO, n642541658578391No ECMO, n191602162044644221724Male sex, n/N (%)ECMO41/64 (64.0)13/25 (52.0)23/41 (56.1)40/65 (61.5)NR49/78 (62.8)NRNo ECMO124/191 (64.9)38/60 (63.3)147/216 (68.1)127/204 (62.3)NR270/422 (64.0)NR P values0.900.330.141.00NR0.90NRAge at surgery, d, median (IQR), or mean±SDECMO6 (5–8)9 (6–11)8 (7–12.5)6 (4–7)4.9±2.95 (3–6)NRNo ECMO7 (5–8)8 (6–11)9 (7–12)5 (4–7)5.92±4.244 (3–6)NR P values0.180.200.680.295NR0.75NRWeight, kg, median (IQR), or mean±SD* ECMO3.1±0.63.1 (2.8–3.5)3.1 (2.8–3.4)3.2 (2.8–3.5)3.05±0.63.1 (2.7–3.5)3.2 (2.9–3.6)No ECMO3.3±0.53.2 (2.9–3.5)3.2 (2.9–3.5)3.2 (2.8–3.5)3.11±0.523.2 (2.9–3.5)3.36 (3–3.7) P values0.01† 0.120.410.888NR0.23<0.001† Low birth weight, <2.5 kg, n/N (%)ECMO10/64 (15.6)NR5/41 (12.2)NR20/85 (23.5)14/78 (17.9)NRNo ECMO16/191 (8.4)NR11/216 (5.1)NR56/464 (12.1)50/422 (11.8)NR P values0.1NR0.08NR0.005† 0.14NRHypoplastic left heart syndrome, n/N (%)ECMONR23/25 (92.0)33/41 (80.5)50/65 (76.9)NRNRNRNo ECMONR50/60 (83.0)157/216 (72.7)152/204 (74.5)NRNRNR P valuesNR0.50.2970.369NRNRNRAortic atresia, n/N (%)ECMO36/64 (56.3)14/25 (56.0)22/41 (53.7)40/65 (61.5)NR47/78 (60.3)NRNo ECMO96/191 (50.2)29/60 (48.3)79/216 (36.6)106/204 (52.0)NR222/422 (52.6)NR P values0.4070.5190.04† 0.177NR0.22NRRight dominant ventricle, n/N (%)ECMO63/64 (98.4)23/25 (92.0)NR52/65 (80.0)NRNRNRNo ECMO170/191 (89.0)57/60 (95.0)NR173/204 (84.8)NRNRNR P values0.02† 0.27NR0.362NRNRNRRestrictive atrial septal defect, n/N (%)ECMONR8/25 (32.0)22/41 (53.7)NRNRNRNRNo ECMONR19/60 (31.7)75/216 (34.7)NRNRNRNR P valuesNR0.980.03† NRNRNRNRPreoperative atrial septostomy, n/N (%)ECMO6/64 (9.4)2/25 (8.0)NRNRNRNRNRNo ECMO9/191 (4.7)4/60 (6.7)NRNRNRNRNR P values0.221.00NRNRNRNRNRExtracardiac or chromosomal anomaly, n/N (%)ECMO16/64 (25.0)NR8/41 (19.5)‡ 14/65 (21.5)§ NR20/78 (25.6)NRNo ECMO29/191 (15.2)NR23/216 (10.6)‡ 47/204 (23.0)§ NR82/422 (19.4)NR P values0.075NR0.110.78NR0.22NRAscending aorta diameter, mm, median (IQR), or mean±SDECMO2.7 (1.8–4.7)NRNR2.5 (1.9–4.5)2.84±1.62NRNRNo ECMO3.0 (2.1–4.6)NRNR2.9 (2.0–5.03)3.19±1.77NRNR P values0.34NRNR0.304NRNRNRPreoperative atrioventricular valve regurgitation (≥ moderate), n/N (%)ECMO9/64 (14.1)NR13/41 (31.7)11/65 (16.9)NR5/78 (6.4)NRNo ECMO28/191 (14.7)NR37/216 (17.1)24/204 (11.7)NR23/422 (5.5)NR P values0.91NR0.04† 0.282NR0.79NRReduced ventricular function, n/N (%)ECMO24/64 (37.5)NR2/41 (4.9)29/65 (44.6)NRNRNRNo ECMO48/191 (25.1)NR21/216 (9.7)25/204 (12.3)NRNRNR P values0.06NR0.55<0.001† NRNRNRECMO indicates extracorporeal membrane oxygenation; IQR, interquartile range; and NR, not reported.*Friedland‐Little et al 6 and Tabbutt et al 10 reported birth weight; all other studies reported weight at time of surgery. † Indicates significance difference ECMO vs No ECMO. ‡ Extracardiac anomalies only. § Chromosomal anomalies only. Baseline Characteristics and Preprocedural Details of ECMO and Non‐ECMO Patients Following the Norwood Procedure ECMO indicates extracorporeal membrane oxygenation; IQR, interquartile range; and NR, not reported. Friedland‐Little et al 6 and Tabbutt et al 10 reported birth weight; all other studies reported weight at time of surgery. Indicates significance difference ECMO vs No ECMO. Extracardiac anomalies only. Chromosomal anomalies only. Seven studies compared ECMO versus non‐ECMO patients 3 , 6 , 7 , 9 , 10 , 11 , 12 (Table 3), and 8 examined survivors versus nonsurvivors within ECMO cohorts 2 , 5 , 8 , 13 , 14 , 15 , 16 , 17 (Table 4). Inconsistent eligibility criteria were observed: some studies excluded hybrid procedures 6 , 7 or Damus‐Kaye‐Stansel operations, 3 whereas others included these patients, 3 , 6 these differences are noted in Table 1. One study excluded patients requiring ECMO for cardiopulmonary bypass weaning failure, 7 and another included only extracorporeal cardiopulmonary resuscitation cases. 8 Table 3Procedural Details and Clinical Outcomes of ECMO and Non‐ECMO Patients Following the Norwood ProcedureVariableFriedland‐Little 6 De Jesus‐Brugman 3 Mayr 9 Beshish 7 Tabbutt 10 Sengupta 11 Cattapan 12 Post‐Norwood ECMO, n642541658578391No ECMO, n191602162044644221724Blalock‐Taussig shunt, n/N (%)ECMO40/64 (62.5%)NR18/41 (43.9%)28/65 (43.1%)49/85 (57.6%)41/78 (52.6%)NRNo ECMO149/191 (78%)NR104/216 (48.1%)56/204 (27.5%)219/464 (47.2%)201/422 (47.6%)NR P values0.04* NR0.620.027* 0.0760.46NRRight ventricle‐to‐pulmonary artery/Sano shunt, n/N (%)ECMO18/64 (28.1%)NR23/41 (56.1%)37/65 (56.9%)36/85 (42.4%)37/78 (47.4%)NRNo ECMO31/191 (16.2%)NR112/216 (51.9%)148/204 (72.5%)245/464 (52.8%)221/422 (52.4%)NR P values0.04* NR0.620.027* 0.0760.64NRCPB time, min, median (IQR)ECMO133 (109–181)188 (171–255)156 (131.5–259.3)179 (157–219)NR164 (142–238)205 (151–237)No ECMO91 (73–120)167 (153–179)142 (112–168)160 (139.8–182)NR145 (125–174)167 (130–280) P values<0.001* 0.004* 0.001* 0.001* NR<0.001* <0.001* Aortic cross clamp time, min, median (IQR)ECMO43 (36–48.5)51 (29–68)51.5 (42.8–70.5)69.5 (57–83.5)NR74 (63–105)72 (53–94)No ECMO39 (35–46)48 (20–56)47 (40–57)73.5 (63–86)NR73 (55–101)67 (50–87) P values0.190.060.005* 0.206NR0.120.002* Hypothermic circulatory arrest time, min, median (IQR)ECMO39 (33.5–44)NR41.5 (30.3–48.5)NRNR22 (7–48)NRNo ECMO36 (31–43)NR40 (33–46.5)NRNR23 (6–46)NR P values0.10NR0.66NRNR0.76NRDelayed sternal closure, n/N (%)ECMONR23/25 (92%)NR54/65 (83.1%)NR59/78 (75.6%)NRNo ECMONR59/60 (98.3%)NR115/204 (56.4%)NR307/422 (72.7%)NR P valuesNR0.21NR0.001* NR0.75NRChest open time, d, median (IQR)ECMONR6 (5–9)NRNRNRNRNRNo ECMONR3 (2–3)NRNRNRNRNR P valuesNR<0.001* NRNRNRNRNRReintervention, n/N (%)ECMO11/64 (17.2%)NRNR49/65 (75.4%)NRNRNRNo ECMO12/191 (6.3%)NRNR42/204 (20.6%)NRNRNR P values 0.01 NRNR<0.001* NRNRNRIntensive care unit stay, d, median (IQR)ECMONR38 (25–74)20 (11–27.5)22 (13–41)NRNRNRNo ECMONR18 (13–28)13 (8–20)14 (9–30.3)NRNRNR P valuesNR<0.001* 0.001* 0.002* NRNRNRHospital stay, d, median (IQR)ECMONR64 (43–101)23 (13.5–60.5)39 (23–64)NRNRNRNo ECMONR37 (27–49)23 (15–36)26 (18–48)NRNRNR P valuesNR<0.001* 0.870.025* NRNRNRIn‐hospital transplant‐free survival, n/N (%)ECMONR17/25 (68%)14/41 (34.1%)27/65 (41.5%)NR34/78 (43.5%)NRNo ECMONR56/60 (93.3%)188/216 (87%)186/204 (91.2%)NRNRNR P valuesNR0.0046* <0.001* <0.001* NRNRNRInterstage transplant‐free survival, n/N (%)ECMONR14/25 (56%)10/41 (24.4%)20/65 (30.8%)NRNRNRNo ECMONR49/60 (81.7%)182/216 (84.3%)142/204 (70%)NRNRNR P valuesNR0.014* <0.001* <0.001* NRNRNRECMO indicates extracorporeal membrane oxygenation; IQR, interquartile range; and NR, not reported.*Indicates. Procedural Details and Clinical Outcomes of ECMO and Non‐ECMO Patients Following the Norwood Procedure ECMO indicates extracorporeal membrane oxygenation; IQR, interquartile range; and NR, not reported. Indicates. Table 4Clinical Characteristics and Outcomes of Survivors and Nonsurvivors Requiring Post‐Norwood ECMO SupportVariableDeBrunner 13 , ‡ Hoskote 14 , * Pizarro 15 , * Hoashi 16 , † Alsoufi 2 Sherwin 5 Ravishankar 17 Polimenakos 8 , ‡ , § Survivors, n1086416226146Nonsurvivors, n22861022512228Male sex, n/N (%)Survivors5/10 (50%)NRNR3/4 (75%)12/16 (75%)142/226 (62.8%)10/14 (71.4%)5/6 (83.3%)Nonsurvivors12/22 (55%)NRNR5/10 (50%)9/22 (40.9%)302/512 (59%)16/22 (72.7%)7/8 (87.5%) P values1.00NRNR0.390.050.510.600.9Weight at surgery, kg, median (IQR) or mean±SD‡ Survivors3.07±0.393.6, 3.2, 3.1, 3.2, 2.0, 3.8, 4.9, 4.82.6, 3.1, 2.8, 3.0, 3.3, 2.03.2 (2.6–3.6)3.3 (2.9–3.8)3.2 (2.8–3.4)3.1±0.73.41±1.71Nonsurvivors3.10±0.513.5, 3.0, 3.7, 2.7, 3.3, 3.2, 4.3, 2.43.8, 1.5, 2.6, 1.4, 2.9, 3.32.8 (2.1–3.3)3.1 (2.6–3.4)3.0 (2.6–3.4)2.7±0.93.51±1.88 P values0.88NRNR0.230.670.03ǁ 0.700.2Age at surgery, d, median (IQR) or mean±SDSurvivorsNR4, 6, 15, 17, 37, 42, 112, 1396, 1, 4, 5, 1, 149 (40–68)6 (4–7)7 (5–10)NR7.6±2.52NonsurvivorsNR3, 9, 4, 4, 8, 8, 14, 281, 3, 2, 14, 4, 535 (5–108)6 (4–8)7 (4–11)NR7.9±2.9 P valuesNRNRNR0.300.310.68NR0.2HLHS, n/N (%)Survivors9/10 (90%)6/8 (75%)5/6 (83.3%)3/4 (75%)12/16 (75%)Only HLHS7/14 (50%)4/6 (66.7%)Nonsurvivors16/22 (72.7%)7/8 (87.5%)5/6 (83.3%)4/10 (40%)17/22 (77.3%)—15/22 (68.2%)6/8 (75%) P values0.3871.001.000.240.870.200.9Extracardiac/genetic abnormality, n/N (%)SurvivorsNRNRNRNR1/16 (6.3%)10/226 (4.4%)1/14 (7.1%)NRNonsurvivorsNRNRNRNR4/22 (18.2%)26/512 (5.1%)3/22 (13.6%)NR P valuesNRNRNRNR0.050.700.50NRBlalock‐Taussig shunt, n/N (%)SurvivorsNR7/8 (87.5%)NR1/4 (25%)5/16 (31.3%)NRNRNRNonsurvivorsNR6/8 (75%)NR2/10 (20%)6/22 (27.3%)NRNRNR P valuesNR1.00NR1.000.79NRNRNRRight ventricle‐to‐pulmonary artery shunt, n/N (%)SurvivorsNR1/8 (12.5%)NR3/4 (75%)11/16 (68.8%)NRNRNRNonsurvivorsNR2/8 (25%)NR8/10 (80%)16/22 (72.7%)NRNRNR P valuesNR1.00NR0.840.79NRNRNRShunt size, mmSurvivorsNR5, 3.5, 3.5, 3.5, 3.0, 3.5, 4, 44, 4, 4, 4, 4, 4NRNRNRNRNRNonsurvivorsNR5, 5, 3.5, 3, 3.5, 3.5, 3, 3.54, 3.5, 3.5, 3.5, 4, 4NRNRNRNRNR P valuesNRNRNRNRNRNRNRNRCardiopulmonary bypass time, min, median (IQR) or mean±SDSurvivors184.9±42.1145, 139, 112, 155, 171, 138, 437, 206NR294 (233–381)159 (142–187)NR103.9±30NRNonsurvivors222.0±80.9153, 180, 181, 312, 166, 174, 253, 227NR236 (183–420)174 (150–212)NR150.1±70NR P values0.10NRNR0.470.29NR0.01ǁ NRAortic cross clamp time, min, median (IQR) or mean±SDSurvivorsNR65, NR, 58, 37, 76, 79, 233, 76NR50 (38–107)67 (52–75)NRNRNRNonsurvivorsNR65, 79, 55, 167, 67, 56, 86, 75NR69 (49–85)67 (54–81)NRNRNR P valuesNRNRNR0.960.58NRNRNRReoperation, n/N (%)§ SurvivorsNRNRNRNRNRNR1/14 (7.1%)2/6 (33.3%)NonsurvivorsNRNRNRNRNRNR6/22 (27.3%)5/8 (62.5%) P values0.2090.592Neurological events, n/N (%)Survivors3/10 (30%)NR0/6 (0%)NR2/16 (14.3%)33/226 (15%)NR1/6 (16.7%)Nonsurvivors11/22 (50%)NR3/6 (50%)NR4/22 (20%)132/512 (26%)NR2/8 (25%) P values0.45NR0.182NR0.67<0.001ǁ NR0.8Renal failure, n/N (%)Survivors0/10 (0%)NR0/6 (0%)NRNR24/226 (10.6%)NR3/6 (50%)Nonsurvivors8/22 (36.4%)NR5/6 (83.3%)NRNR126/512 (24.6%)NR4/8 (50%) P values0.04NR<0.05NRNR<0.001ǁ NR1.00Bleeding, n/N (%)Survivors5/10 (50%)NR0/6 (0%)NR5/16 (35.7%)79/226 (35%)1/14 (7%)1/6 (16.7%)Nonsurvivors7/22 (31.8%)NR2/6 (33.3%)NR8/22 (38.1%)239/512 (48%)6/22 (27%)2/8 (25%) P values0.44NR0.455NR0.890.003ǁ 0.200.6ECMO duration, h, median (IQR) or mean±SDSurvivors124.5±75.795, 13, 134, 120, 72, 96, 240, 5848, 24, 24, 120, 24, 24NR72 (48–120)71 (47–118)50.1±12.596 (72–156)Nonsurvivors208.7±148.1422, 208, 27, 40, 203, 275, 298, 204112, 102, 70, 48, 192, 90NR168 (96–216)122 (69–189)125.2±25192 (120–276) P values0.05NRNRNR0.008ǁ <0.001ǁ 0.01ǁ 0.02ǁ ECMO indicates extracorporeal membrane oxygenation; HLHS, Hypoplastic left heart syndrome; IQR, interquartile range; Neurological events, intraventricular hemorrhage, cerebral hemorrhage, cerebral ischemia, seizure; and NR, not reported.Values are presented as mean±SD, median (range), or individual data.*Hoskote et al, 14 and Pizarro et al, 15 reported individual patient‐level data. † Hoashi et al, 16 reported medians with full range. ‡ Debrunner et al, 13 reported birth weight and Polimenkos et al, 8 reported weight at time of extracorporeal cardiopulmonary resuscitation. § Polimenkos et al, 8 reported the number of reoperations, multiple procedures may have been performed on some patients. ǁ Indicates. Clinical Characteristics and Outcomes of Survivors and Nonsurvivors Requiring Post‐Norwood ECMO Support ECMO indicates extracorporeal membrane oxygenation; HLHS, Hypoplastic left heart syndrome; IQR, interquartile range; Neurological events, intraventricular hemorrhage, cerebral hemorrhage, cerebral ischemia, seizure; and NR, not reported. Values are presented as mean±SD, median (range), or individual data. Hoskote et al, 14 and Pizarro et al, 15 reported individual patient‐level data. Hoashi et al, 16 reported medians with full range. Debrunner et al, 13 reported birth weight and Polimenkos et al, 8 reported weight at time of extracorporeal cardiopulmonary resuscitation. Polimenkos et al, 8 reported the number of reoperations, multiple procedures may have been performed on some patients. Indicates. Quality Assessment Of 15 studies assessed, the single randomized trial had high risk of bias due to post hoc ECMO comparisons (Table S3). Among remaining studies, 8 were rated good quality and 6 poor quality due to insufficient confounder adjustment (Table S4). Temporal Trends in ECMO Use Overall post‐Norwood ECMO incidence was 18.1% (685/3775 patients), ranging from 15.5% (85/549) to 29.4% (25/85) across studies. 3 , 7 , 9 , 10 , 11 , 12 In considering change in perioperative practice and stratifying studies into 3 eras (based on midpoint of recruitment periods: early era [1990–2000, 2 studies]; middle era [2001–2010, 8 studies]; and contemporary era [2011–2025, 4 studies]), there appeared to be an increasing trend in ECMO incidence. For the middle era incidence was 15.5% (range 15.48%–15.6%) and for the contemporary era incidence was 19.15% (range 15.9%–29.41%). For the contemporary era 24.2% (65/269) in Beshish et al, 7 29.4% (25/85) in De Jesus‐Brugman et al, 3 and 18.5% (391/2115) in Cattapan et al 12 Note that data from Friedland‐Little et al 6 were excluded from incidence calculations due to 3:1 matched case–control design (Figure 2). These trends suggest increasing ECMO use over time, potentially reflecting expanded indications, newer devices, and availability. However, the lack of substantial improvement in survival rates may indicate persistent challenges in patient selection and management. Extracorporeal membrane oxygenation incidence by study recruitment periods reported by Beshish et al, 7 De Jesus‐Brugman et al, 3 Mayr et al, 9 Sengupta et al, 11 Tabbutt et al, 10 and Cattapan et al 12 ECMO indicates extracorporeal membrane oxygenation. Indications for ECMO ECMO indications were broadly categorized as: primary respiratory failure (eg, refractory hypoxemia, pulmonary hypertension, and pulmonary‐to‐systemic flow imbalance); primary circulatory failure (eg, low cardiac output syndrome, arrhythmias, cardiac arrest); combined circulatory‐respiratory failure; and failure to wean from cardiopulmonary bypass or shunt‐related complications. Clinical indications were reported for post‐Norwood ECMO support in 7 studies (n=268). 3 , 6 , 9 , 11 , 13 , 14 , 15 The most common indications were low cardiac output (32%, 85/268), and cardiac arrest (24%, 65/268). Indications for ECMO support were categorized as respiratory failure (20%, 54 of 268 cases) or circulatory failure (80%, 214 of 268). Risk Factors for ECMO Requirement Patient Demographics Age at surgery was similar between ECMO and non‐ECMO groups. 3 , 6 , 7 , 9 , 10 , 11 Four studies reported higher proportions of low birth weight (<2.5 kg) patients in ECMO groups 6 , 9 , 10 , 11 (Figure 3). Incidence of low birth weight (<2.5 kg) by extracorporeal membrane oxygenation vs no ECMO. Points represent study level proportions, with error bars indicating 95% CIs. Reported by Friedland‐Little et al, 6 Mayr et al, 9 Sengupta et al, 11 and Tabbutt et al 10 ECMO indicates extracorporeal membrane oxygenation. Anatomical and Functional Features Cardiac Diagnosis Three studies showed greater HLHS prevalence requiring ECMO support 3 , 7 , 9 (Figure 4A). Aortic atresia was more prevalent in ECMO patients across 5 studies, 3 , 6 , 7 , 11 with Mayr et al 9 reporting significantly more aortic atresia with mitral atresia (P=0.04) but not mitral stenosis (P=0.72) (Figure 4B). Ventricular morphology was inconsistently reported across studies. Where specified, the majority of patients had systemic right ventricles (predominantly HLHS variants). Three studies reported proportions of right versus left dominant ventricles, with right dominant ventricles comprising 80% to 98.4% of ECMO patients compared with 84.8% to 95% of non‐ECMO patients. 3 , 6 , 7 Incidence of hypoplastic left heart syndrome diagnosis by ECMO vs no ECMO (A), and incidence of aortic atresia variant by ECMO vs no ECMO (B). Error bars denote 95% CIs. Reported by Beshish et al, 7 De Jesus‐Brugman et al, 3 Friedland‐Little et al, 6 Mayr et al, 9 and Sengupta et al 11 ECMO indicates extracorporeal membrane oxygenation. Ascending Aorta Ascending aorta diameter was consistently smaller in ECMO groups across studies, though measurement and reporting methods varied. Diameters in the ECMO group were smaller by 0.35 mm (mean), 0.3 mm (median), and 0.4 mm (median), respectively. 6 , 7 , 10 In addition, Beshish et al used a 4 mm threshold, finding more patients with ECMO below this cutoff (67.7% [44/65] versus 63.2% [129/204]). 7 Sengupta et al found higher prevalence of diameter <2 mm in ECMO patients (15.4%, 12/78 versus 12.6%, 53/422). 11 Mayr et al reported more patients with diameter <3 mm (39%, 16/41 versus 32.4%, 70/216). 9 Atrioventricular Valve Regurgitation Preoperative echocardiographic assessment of atrioventricular valve regurgitation (AVVR) and ventricular function was reported in 4 studies. 6 , 7 , 9 , 11 All assessments were based on preoperative studies, though exact timing relative to surgery was not consistently reported. Moderate‐to‐severe AVVR was more prevalent in ECMO patients across 4 studies, 6 , 7 , 9 , 11 with Mayr et al 9 reporting significant differences (31.7%, 13/41 versus 17.1%, 37/216, P=0.04) (Figure 5). Incidence of preoperative atrioventricular valve regurgitation assessed as greater than “moderate” by ECMO vs No ECMO subjects. Error bars denote 95% CIs. Reported by Beshish et al, 7 Friedland‐Little et al, 6 Mayr et al, 9 and Sengupta et al 11 ECMO indicates extracorporeal membrane oxygenation. Ventricular Function Results were inconsistent across studies. Friedland‐Little et al 6 and Beshish et al 7 observed higher proportions of reduced ventricular function in ECMO groups (37.5%, 24/64 versus 25.1%, 48/191, P=0.06 and 44.6%, 29/65 versus 12.3%, 25/204, P<0.001 respectively), whereas Mayr et al 9 found the opposite (4.9%, 2/41 versus 9.7%, 21/216, P=0.55). Procedural Factors Cardiopulmonary Bypass Time CPB time was consistently longer in ECMO groups across all 6 reporting studies, 3 , 6 , 7 , 9 , 11 , 12 with significant differences in all studies (Figure 6). Aortic cross‐clamp time was significantly longer only in studies by Mayr et al 9 (P=0.005) and Cattapan et al 12 (P=0.002). Box plots of cardiopulmonary bypass time by ECMO vs no ECMO subjects, median with interquartile range, as reported by De Jesus‐Brugman et al, 3 Friedland‐Little et al, 6 Mayr et al, 9 Beshish et al, 7 Sengupta et al, 11 and Cattapan et al 12 ECMO indicates extracorporeal membrane oxygenation. Delayed Sternal Closure Two studies reported higher delayed sternal closure prevalence in ECMO patients. 3 , 7 , 11 De Jesus‐Brugman et al 3 reported longer median chest‐open duration in ECMO patients (6 versus 3 days, P<0.001). Clinical Outcomes: ECMO Versus Non‐ECMO Patients requiring ECMO had significantly longer intensive care unit and hospital stays, 3 , 7 , 9 except hospital stay in Mayr et al 9 (P=0.87). Postoperative reinterventions were more frequent in ECMO patients: shunt revision (17.2%, 11/64 versus 6.3%, 12/191, P=0.01), 6 catheter‐based interventions (70.8%, 46/65 versus 18.3%, 37/204, P<0.001), and aortic arch revision (4.6%, 3/65 versus 2.5%, 5/204, P=0.814). 7 In‐hospital transplant‐free survival was significantly lower in ECMO patients: 44% (58/131) versus 89.5% (430/480) in non‐ECMO patients. Interstage transplant‐free survival was 33.5% (44/131) versus 77.3% (373/480), respectively. 3 , 7 , 9 Outcomes in ECMO‐Supported Patients Eight studies examined 900 ECMO patients, with 290 survivors and 610 deaths, yielding 32.2% in‐hospital survival rate. 2 , 5 , 8 , 13 , 14 , 15 , 16 , 17 Survival rates ranged from 28.6% (4/14) to 50% (8/16), with no clear temporal improvement over the 15‐year study period (Figure 7). Mortality on extracorporeal membrane oxygenation reported by study recruitment period, from DeBrunner et al, 13 Hoskote et al, 14 Pizarro et al, 15 Hoashi et al, 16 Alsoufi et al, 2 Sherwin et al, 5 Ravishankar et al, 17 Polimenkos et al 8 ECMO indicates extracorporeal membrane oxygenation. Predictors of Mortality Patient Characteristics Survivors had higher surgical weight across all studies. 2 , 5 , 8 , 13 , 14 , 15 , 16 , 17 Most patients underwent surgery within 2 weeks of life, 2 , 5 , 8 , 14 , 15 , 16 except Hoashi et al 16 who reported later surgery (median 49 versus 35 days, P=0.30) with lower survival rates (28.6%, 4/14 versus 32.3%, 286/886). Three studies reported higher prevalence of extracardiac or genetic abnormalities in nonsurvivors. 2 , 5 , 17 Procedural Factors Nonsurvivors had longer CPB times in 4 of 5 reporting studies, 2 , 13 , 14 , 17 except the small study by Hoashi et al 16 Shunt type showed no survival advantage across 3 reporting studies. 2 , 14 , 16 ECMO‐Related Factors ECMO duration was consistently longer in nonsurvivors across 7 studies, 2 , 5 , 8 , 13 , 14 , 15 , 17 with significant differences in 4 studies. 2 , 5 , 8 , 17 Median durations ranged from 122 to 208.7 hours in nonsurvivors versus 50.1 to 124.5 hours in survivors. Complications ECMO complications were more prevalent in nonsurvivors across 6 studies. 2 , 5 , 8 , 13 , 15 , 17 Neurological events, renal failure, and bleeding were most common. Neurological complications were more frequent in nonsurvivors in all 5 reporting studies. 2 , 5 , 8 , 13 , 15 Renal failure 5 , 8 , 13 , 15 and bleeding complications 2 , 5 , 8 , 13 , 15 , 17 were similarly more prevalent in nonsurvivors. DISCUSSION This systematic review demonstrates increasing post‐Norwood ECMO use with persistently poor survival outcomes. The overall ECMO incidence of 18.1% aligns with contemporary reports showing temporal increases from 12% (178/1478) pre‐2020 13 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 to 19.2% (1067/5548) post‐2020, 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 likely reflecting both advancing ECMO technology and application to higher‐risk neonates previously considered unsuitable for intervention. Risk Factors for ECMO Patient Characteristics Low birth weight (<2.5 kg) emerged as a marker associated with worse outcomes across multiple studies, 6 , 9 , 10 , 11 supporting recent findings by Averin et al, 39 linking low weight to increased ECMO use, mortality, and poor neurodevelopmental outcomes. This pattern may reflect the underlying association with overall physiological vulnerability and reduced physiological reserve. Anatomical Complexity Specific cardiac anatomies showed significant association with ECMO requirement. Aortic atresia variants, 3 , 6 , 7 , 9 , 11 and diminutive ascending aorta consistently increased ECMO risk. 6 , 7 , 9 , 10 , 11 These findings of ascending aorta align with Carvajal et al, 40 who demonstrated stepwise ECMO incidence increases as ascending aorta diameter decreased from >4 to <1.5 mm. Although consistent in direction, these findings are limited by differing thresholds and potential confounding by higher HLHS prevalence in ECMO groups. Future studies should standardize aortic measurement protocols and adjust for cardiac diagnosis in multivariate analyses. Mayr et al, 9 further refined the association between aortic atresia and need for post‐Norwood ECMO by reporting a significantly higher prevalence of aortic atresia with mitral atresia in ECMO patients (34.1%, 14/41 versus 19.4%, 42/216, P=0.04), but not aortic atresia with mitral stenosis (19.5%, 8/41 versus 17.1%, 37/216, P=0.72). This distinction is noteworthy, as aortic atresia with mitral stenosis is typically associated with poorer outcomes due to factors such as ventriculo‐coronary dependence. 41 , 42 However, other studies did not report statistical distinction between these subtypes, limiting broader comparisons. Such anatomical variants may increase surgical complexity and impair coronary perfusion, necessitating mechanical circulatory support. Moderate‐to‐severe atrioventricular valve regurgitation was more prevalent in ECMO patients, 6 , 7 , 9 , 11 consistent with AVVR being a known risk factor for clinical deterioration in single ventricle physiology. 43 However, inconsistent grading criteria across studies limit evidence strength, reflecting current clinical challenges in standardized AVVR assessment due to complex valve morphology and eccentric regurgitant jets. 43 Procedural Factors Prolonged cardiopulmonary bypass time was consistently associated with increased ECMO requirement across studies. 3 , 6 , 7 , 9 , 11 , 12 This association may reflect more complex procedures or technical challenges; however, causal relationships cannot be definitively established from these observational data. Although prolonged bypass itself may contribute to postoperative morbidity through inflammatory mechanisms, it likely also serves as a marker for overall case complexity and patient risk profile. Shuhaiber et al, 44 demonstrated that prolonged CPB times indicate suboptimal surgical efficacy, and extended bypass introduces systemic inflammation, coagulopathy, and end‐organ dysfunction, 45 compounding vulnerability in this high‐risk cohort. Surgical strategies evolved significantly over the study period, including transitions from modified Blalock‐Taussig to Sano shunts and introduction of hybrid approaches. ECMO management has also changed, particularly regarding shunt handling (eg, clamping versus banding). Although granular data on these practices were limited, studies from the contemporary era more frequently reported Sano shunt use and strategies to limit pulmonary overcirculation on ECMO. Future research should explicitly examine how these evolving practices impact ECMO indications and outcomes. Mortality and Predictors Despite technological advances, ECMO survival remains poor (32.2% overall, 290/900), with no temporal improvement over 15 years. This contrasts with higher survival rates after Stage II (41%, 42/103) 46 and Fontan‐stage ECMO (35%, 81/230), 47 highlighting unique Norwood cohort vulnerability. The absence of survival improvement may reflect increasing ECMO application to previously unsuitable high‐risk patients, masking potential clinical advances. Predictors of Mortality Several factors overlapped between ECMO requirement and mortality prediction, including low weight and prolonged CPB time, suggesting a risk continuum from perioperative through ECMO course. ECMO duration was consistently associated with mortality. 2 , 5 , 8 , 13 , 14 , 15 , 17 Though causality remains unclear, longer support may reflect greater illness severity rather than representing an independent risk factor. Complications Neurological events, renal failure, and bleeding were more prevalent in nonsurvivors. 2 , 5 , 8 , 13 , 15 , 17 Neurological complications raise particular concern given growing evidence of long‐term neurodevelopmental impairments in pediatric cardiac ECMO survivors. 48 , 49 , 50 Renal failure likewise demonstrated a strong association with mortality, consistent with broader pediatric post‐cardiotomy ECMO populations where acute kidney injury independently predicts poor outcomes. 51 , 52 Clinical Implications Risk Stratification The identified risk factors, which included low birth weight, complex anatomical variants (aortic atresia, diminutive ascending aorta), moderate‐to‐severe AVVR, and prolonged CPB time, together may inform prospective risk stratification strategies. Early identification of high‐risk patients could enable proactive ECMO planning and potentially improve outcomes through earlier intervention. Timing Considerations Limited data suggest early ECMO initiation may improve survival. Sengupta et al 11 reported superior outcomes with initiation within 3 days (52%, 26/50 versus 28.6%, 8/28 survival, P=0.038), and Perry et al 53 demonstrated higher survival rates with 2‐day cutoffs (59.6%, 90/152 versus 43.4%, 72/167). These findings emphasize the importance of early recognition and prediction of patients likely requiring ECMO support. Pathophysiological Framework of Post‐Norwood ECMO Based on the perioperative risk factors, ECMO indications, and outcomes, identified in this systematic review, we propose the following integrated conceptual premises to underpin more targeted future work, together with more detailed and uniform reporting formats. Preoperative Vulnerability Patients entering Norwood procedures often face significant challenges. Low birth weight diminishes physiological reserve, leaving little margin for complications. Complex anatomy, particularly in cases like HLHS with aortic atresia or small ascending aorta, not only increases surgical complexity but also elevates the risk of coronary insufficiency. Additionally, significant AVVR contributes to volume loading and ventricular dysfunction, further compromising cardiac function before surgery. Intraoperative Stress The surgical procedure itself introduces additional stressors. Prolonged CPB time triggers an inflammatory response and leads to myocardial edema, potentially compromising postoperative cardiac function. Moreover, residual lesions such as shunt stenosis or arch obstruction may persist, setting the stage for early postoperative complications. Early Postoperative Decompensation The combination of preoperative vulnerability and intraoperative stress often manifests as early postoperative decompensation. Low cardiac output syndrome, resulting from myocardial dysfunction and coronary insufficiency, is a common presentation. Pulmonary overcirculation can lead to systemic hypoperfusion and acidosis, further straining the compromised cardiovascular system. In severe cases, these factors culminate in cardiac arrest, necessitating immediate intervention. ECMO‐Related Complications When ECMO becomes necessary, it introduces its own set of complications. Bleeding, a significant concern, stems from both the required anticoagulation and surgical site issues. Neurological events, including embolic phenomena and hypoperfusion, pose serious risks to long‐term outcomes. Renal failure, often resulting from pre‐ECMO hypoperfusion and exacerbated by ECMO‐related hemolysis, further complicates patient management. Determinants of Outcome Several factors influence the ultimate outcome for these patients. ECMO duration serves as a proxy for the severity of the underlying pathology, with longer durations generally indicating more severe conditions. The ability to address the root cause, whether through surgical revision or supporting myocardial recovery, is crucial. Finally, the cumulative burden of complications, particularly neurological and renal sequelae, significantly impacts both short‐term survival and long‐term quality of life. This framework highlights the continuum from preoperative risk to postoperative decompensation and ECMO‐related morbidity. It emphasizes the importance of optimizing preoperative status, refining surgical techniques, and early recognition of decompensation to potentially improve outcomes. Limitations Substantial heterogeneity in study populations, designs, and outcome reporting precluded formal meta‐analysis. Small sample sizes, predominantly retrospective designs, and inconsistent inclusion criteria (hybrid procedures, Damus‐Kaye‐Stansel operations) limit cross‐study comparability. Inconsistent reporting of critical parameters (reoperation rates, shunt sizes, AVVR grading criteria) prevented comprehensive risk factor assessment. Statistical comparisons were limited by inconsistent reporting across studies. Where available, we have included P values. However, many studies, particularly those with smaller sample sizes, did not report statistical significance for some subgroup comparisons. In addition, heterogeneous reporting of variables across studies required simple data harmonization to enable cross‐study comparisons (eg, summing or reformatting reported values to match a common reporting structure). This limited our ability to generate pooled effect estimates and draw firm conclusions about the clinical relevance of some observed differences. Heterogeneity in ECMO definitions, timing, and mortality end points limits direct comparisons between studies. Some excluded failure to wean from bypass or only included extracorporeal cardiopulmonary resuscitation cases, whereas others included all ECMO use. Similarly, mortality was variably reported as in‐hospital, 30‐day, or interstage. The predominant focus on in‐hospital mortality may not fully capture the clinical trajectory, with interstage survival (between the Norwood operation and Stage II operation) potentially representing a more meaningful end point during this highly vulnerable period. The long inclusion period (1990–2025) spans significant changes in surgical and ECMO practices. Although we attempted to analyze temporal trends to account for changes in surgical and ECMO practices, inconsistent reporting of time‐dependent outcomes limited our ability to draw firm conclusions about changes over time. Future Directions Future research should use standardized definitions and outcome measures. Development of a Norwood‐specific risk stratification framework incorporating the identified predictors could guide clinical decision‐making. Standardized assessment protocols for AVVR and ventricular function would improve risk prediction accuracy and greater capacity for data comparison between cohorts. Long‐term neurodevelopmental follow‐up studies are essential given the high neurological complication rates observed. Although large registry studies provide robust data on overall trends, the present systematic review complements those findings by synthesizing granular clinical details and highlighting important reporting variations across institutions and eras. In order to facilitate more robust meta‐analyses and generate higher‐quality evidence to guide clinical decision‐making, we propose the following key areas for standardization in future research: Development of consensus definitions for post‐Norwood ECMO, including clear delineation of immediate postoperative versus later interstage support before Stage II intervention. Development of consensus definitions for post‐Norwood ECMO, including clear delineation of immediate postoperative versus later interstage support before Stage II intervention. Standardized reporting of key time points (eg, 30‐day, hospital discharge, and interstage survival). Standardized reporting of key time points (eg, 30‐day, hospital discharge, and interstage survival). Consistent categorization of anatomical subtypes and surgical approaches. Consistent categorization of anatomical subtypes and surgical approaches. Multicenter prospective studies using uniform data collection protocols contributing to large registry studies. Multicenter prospective studies using uniform data collection protocols contributing to large registry studies. CONCLUSIONS Post‐Norwood ECMO incidence is increasing while survival outcomes remain poor. Risk factors including low birth weight, complex cardiac anatomy, prolonged CPB time, and ECMO duration contribute to adverse outcomes. Early identification of high‐risk patients may enable proactive management strategies, though the persistent high mortality underscores the need for continued research into optimal patient selection, timing of intervention, and risk stratification in this vulnerable population. Sources of Funding Christian P. Brizard and Salvatore Pepe are supported by grant funding from the National Health and Medical Council of Australia (GNT1186861). Disclosures None. Supporting information Tables S1–S4 Acknowledgments We thank Dominic Morelli for clerical assistance in the final formatting of figures and text. This article was sent to John L. Jefferies, MD, MPH, Guest Editor, for review by expert referees, editorial decision, and final disposition. Supplemental Material is available at https://www.ahajournals.org/doi/suppl/10.1161/JAHA.125.047642 For Sources of Funding and Disclosures, see page 18. References Associated Data Supplementary Materials Tables S1–S4