Neonatal cardiac surgery is associated with high morbidity and mortality. Perioperative metabolic stress may result in distinct metabolic phenotypes associated with adverse clinical outcomes.

The purpose of this study was to identify perioperative metabolomic phenotypes associated with adverse clinical outcomes in neonates undergoing cardiac surgery with cardiopulmonary bypass (CPB).

Plasma samples were collected from 230 neonates at 3 institutions at preoperative, post-CPB, and 24-hour time points. Targeted metabolomics was performed via mass spectrometry. Feature selection used the Boruta algorithm, and unsupervised hierarchical clustering (Euclidean distance, Ward’s linkage) identified phenotypes. The primary endpoint was a composite of morbidity, including death, extracorporeal membrane oxygenation, cardiac arrest, renal replacement therapy, hepatic injury, elevated lactate, and low cardiac output syndrome.

Fifty-one of 76 metabolites passed quality control. Unsupervised clustering identified 3 metabolomic groups with composite outcome rates of 40%, 6.1%, and 58%. For clinical interpretation, these corresponded to intermediate-, low-, and high-risk phenotypes, respectively. The high-risk phenotype was characterized by elevated 24-hour tyrosine, alanine, proline, and methionine, metabolites linked to amino acid catabolism and inflammation. The low-risk phenotype demonstrated higher post-CPB cystine and greater pre/post ratios of galactitol and acetylcarnitine, consistent with preserved redox balance and fatty acid oxidation. Multiple comparison adjustments confirmed associations between these metabolic pathways and clinical outcomes.

Metabolomic profiling identified phenotypes associated with risk for adverse outcomes after neonatal cardiac surgery. Amino acid catabolism and inflammatory pathways were linked to morbidity, supporting further evaluation of metabolomic biomarkers for perioperative risk stratification and targeted therapeutic intervention.