Second-trimester PM2.5 and PM10 exposure is associated with neonatal metabolome alterations and elevated risk of pregnancy loss: a large-scale matched cohort study with metabolomic profiling.

Abstract

Prenatal air pollution has been linked to pregnancy loss, but the critical windows and underlying mechanisms remain uncertain. In a propensity score-matched cohort of 30,654 singletons (106 losses, 30,548 live births), we assessed seven exposure windows from pre-pregnancy to peripartum, built a metabolic risk score from five metabolites, and performed multi-window regression and enrichment analyses. Second-trimester PM2.5 was associated with an increased risk of pregnancy loss (OR = 2.32, 95% CI: 1.12 to 4.78), and PM10 showed a similar but weaker association (OR = 1.86, 95% CI: 1.18 to 2.95). In single-window models, the metabolic risk score was associated with second-trimester PM2.5 (β = 0.052, P < 0.001), but this association was attenuated in the multi-window model (β = 0.003, P = 0.287). Enrichment analysis pointed to the amino acid metabolism pathway (4/7 metabolites, FDR P = 0.002), with second-trimester PM2.5 associated with lower levels of glycine, proline, valine, and tyrosine; fatty acid oxidation and mitochondrial function showed no enrichment (0/5; FDR P = 1.000). Together, these findings suggest that second-trimester PM2.5 and PM10 exposure are associated with both pregnancy loss and neonatal metabolic changes, with amino acid metabolism prominently affected. Glycine, proline, valine, and tyrosine may serve as postnatal indicators of in utero exposure-related metabolic disruption. The second trimester may represent a vulnerable window for intervention; however, single-window and sensitivity analyses did not consistently reach statistical significance, and external validation is needed.