To the best of our knowledge, there has been a limited exploration of metabolomic analyses in children undergoing surgical repair of VSD under CPB. In a study by Davidson et al. [5], the perioperative metabolism of 165 children undergoing cardiac surgery was analyzed. The most significant changes in amino acid metabolism, such as arginine and proline metabolism, were found in children during the perioperative period of CPB. This finding follows the results of our study. Of these, aspartate, glutamate, nicotinamide methyl ester, tritiated alkaloids, and kynurenine were identified as useful in determining the postoperative prognosis of the children. Meanwhile, Wu et al. [6] observed significant changes in both lipid metabolism and amino acid metabolism by studying metabolic changes in fetal sheep undergoing CPB, a finding that is also mentioned in our study.

Our study represents the first comprehensive examination of metabolomic changes in serum samples from children with VSD during the perioperative correction of defects under CPB using molecular docking techniques. In our patient analysis, we observed significant alterations in 623 metabolites at different time points (Tp, T0, and T24). Subsequent KEGG enrichment analysis revealed three recurrent metabolic pathways—glycerophospholipid metabolism, arginine and proline metabolism, and retrograde endogenous cannabinoid signaling—in various comparisons between the two groups. These findings suggest that the perioperative metabolic profile changes in children undergoing surgical repair of VSD under CPB predominantly involve lipid and amino acid metabolism. Alterations in lipid metabolism concurrently affect the digestive, immune, and nervous systems. These changes within metabolic pathways may play a crucial role in the perioperative pathophysiological processes of children undergoing surgical repair of VSD under CPB.

Glycerophospholipids (GPL) are crucial components of cell membranes and their metabolism produces various bioactive lipid molecules, including inositol triphosphate, diacylglycerol, arachidonic acid, phosphatidic acid, and lysophosphatidic acid. These molecules regulate diverse signaling pathways within cells [9]. Substantial evidence supports the notion that glycerophospholipid metabolism plays a pivotal role in systemic immunity and low-grade inflammatory states, thereby implicating GPL as a potential mediator of inflammation [10, 11].

At the systemic level, the metabolism of arginine and proline ultimately yields various biochemically diverse products including proline, glutamate, creatine, urea, polyamines, nitric oxide, etc. Arginine plays a pivotal role in the urea cycle and has immunomodulatory and anti-inflammatory properties [10]. Other members of the urea cycle, such as ornithine and citrulline, are negatively associated with inflammation [11]. In contrast, arginine serves as a precursor of NO synthesis. NO functions as a crucial modulator of vascular endothelial dysfunction by influencing vasodilation and platelet aggregation and inhibiting smooth muscle proliferation [12]. While our findings suggest a potential defect in NO production through L-arginine, reduced NO production in endothelial cells and elevated levels of reactive oxygen species contribute to the decreased resistance of endothelial cells to neutrophil-induced cytotoxicity [13].

The endogenous cannabinoid system (ECS) is a lipid signaling system comprising endogenous cannabinoid-like ligands derived from arachidonic acid, including anandamide (AEA) and 2-arachidonoylglycerol (2-AG) [14]. The enrichment of retrograde endogenous cannabinoid signaling observed in our study may be attributable to alterations in perioperative lipid metabolism in children undergoing surgical repair of VSD under CPB. Alternatively, perioperative fasting can potentially modulate retrograde endogenous cannabinoid signaling.

Following our metabolomic analysis, we identified an arginine-glycine amidinotransferase from the arginine and proline metabolic pathways as a crucial focus for further investigation. In children undergoing surgical repair of VSD under CPB, bosentan and AdipoRon were found to modulate perioperative metabolic changes by binding to arginine-glycine amidinotransferase, thereby influencing perioperative injury.

Endothelin-1 (ET-1) is a potent vasoconstrictor that plays key roles in mediating cell proliferation, fibrosis, and inflammation [15]. Activation of endothelin receptor A induces pulmonary vasoconstriction and smooth muscle cell proliferation, whereas endothelin receptor B eliminates ET-1, mediating endothelial cell vasodilation, and release of prostacyclin and nitric oxide. Bosentan, an endothelin receptor antagonist, competitively inhibits ET-1 expression, and reduces pulmonary vascular resistance [16]. Previous studies in rats demonstrated that bosentan lowers blood pressure, prevents cardiac hypertrophy, and exhibits anti-inflammatory and antithrombotic effects [17]. Our identification of the binding site of bosentan on arginine-glycine amidinotransferase provides valuable insights into its potential mechanism of action.

Lipocalin is secreted by adipose tissue and binds to the lipocalin receptor, exerting its antidiabetic effects in the liver and skeletal muscle by activating the AMP-activated protein kinase (AMPK) and peroxisome proliferator-activated receptor (PPAR) alpha pathways. It ameliorates vascular dysfunction by activating endothelial NO production and exhibits anti-atherosclerotic effects on the vasculature by inhibiting various inflammatory conditions [18, 19]. It was shown in a recent animal study by Jenke et al. [8] that AdipoRon, a lipocalin receptor agonist, can attenuate CPB-induced cardiac impairment and inflammation. This attenuation occurred through AMPK-mediated inhibition of proinflammatory TLR4 and TNF-α signaling. In addition, the immunosuppressive IL-10 is upregulated in cardiac cells. These findings are consistent with those of the present study.

However, this study is subject to certain limitations. First, the sample size was restricted to eight patients, warranting the inclusion of a larger cohort for metabolomic analysis to enhance the robustness of the study. Second, the absence of fundamental experiments to assess the activity and expression of key enzymes within the metabolic pathway may have influenced the reliability of our findings. Meanwhile, to control the variables we selected children with VSD for the study, so the generalisability of the results needs to be further explored. In addition, it is crucial to acknowledge that the molecular docking results hold only a predictive value, necessitating supplementary experiments for further validation. However, drug applicability and potential adverse effects require further investigation.