Resolving Early Targets and Metabolomic Profile of Congenital Heart Disease Through Tandem Mass Spectrometry Screening in Neonates.

BACKGROUND

A good prognosis of congenital heart disease (CHD) depends on early diagnosis and intervention. Under the current screening conditions, a significant proportion still go undetected. Metabolomics, as a phenotype-correlated research methodology, remains underused in the study of CHD, which could provide the possibility to screen neonatal CHD efficiently.

METHODS

Data for the analysis are from >22 000 neonates captured in the Network Platform for CHD from April 2020 to November 2021 in 11 cities in China. After data matching and quality control, a total of 22 674 neonates were finally included and divided into the CHD group (n=1823), nonsignificant CHD group (n=17 968), and normal group (n=2748). Demographic and clinical characteristics and tandem mass spectrometry-based metabolic data for genetic and metabolic disease screening were gathered and compared for all groups. Machine learning models based on metabolic biomarkers were constructed to screen CHD in neonates.

RESULTS

After quality control, 22 539 neonates were ultimately included. Among them, 1823 were diagnosed with CHD, 17 968 were nonsignificant CHD, and 2748 were normal. A total of 46 distinguishing metabolic biomarkers were identified, and we found that the CHD group had significantly lower levels of 17-hydroxyprogesterone (CHD versus nonsignificant CHD, P<0.001, log2 fold change=-0.16; CHD versus normal, P<0.001, log2 fold change=-0.15). We constructed CHD and ventricular septal disease screening models based on metabolic biomarkers. The best fitting model achieved an area under the receiver operating characteristic curve of 0.745 (95% CI, 0.696-0.791).

CONCLUSIONS

This study reveals the unique metabolic profile of neonates with CHD. The screening model demonstrates considerable potential in early neonatal CHD screening and reflects significant value from a health economics perspective.