Background: Neonatal hyperbilirubinemia (HB) is common in the first postnatal week, and early gut ecological and metabolic features may influence bilirubin homeostasis via the gut–liver axis. We sought birth meconium microbiome–metabolome features linked to HB and evaluated a candidate metabolite in vivo.

Methods: First-pass meconium was collected within 24 h from 219 neonates. A nested case–control subset was analyzed (HB, n = 21; Healthy Controls (HC), n = 39). Microbiota were profiled by 16S rRNA gene sequencing (V3–V4), and metabolites were quantified using a targeted absolute-quantification LC–MS panel (Q300). Community structure was assessed using weighted and unweighted UniFrac distances and tested by analysis of similarities (ANOSIM), with PERMDISP used to evaluate dispersion. Genus-level differences were explored by LEfSe and interpreted cautiously given the low-biomass nature of meconium. Metabolites were compared by univariate testing, visualized by volcano plots, and adjusted using false discovery rate (FDR). Azelaic acid (AzA) was tested in an acetylphenylhydrazine-induced rat model with serum total bilirubin, colonic occludin, colonic β-glucuronidase activity, and fecal 16S profiling.

Results: HB and HC showed modest but statistically significant differences in weighted and unweighted UniFrac distances (weighted R = 0.1965, p = 0.0030; unweighted R = 0.1225, p = 0.0100). Metabolomics showed limited global separation; AzA was lower in HB on nominal testing, but no metabolite remained significant after FDR correction. In rats, AzA was associated with lower serum total bilirubin, directionally higher colonic occludin, decreased colonic β-glucuronidase activity, and suggested a partial shift in fecal community structure toward controls.

Conclusions: Birth meconium profiling nominated lower AzA as a hypothesis-generating signal; the rat experiment provides complementary, biologically plausible evidence. Replication in independent neonatal datasets is warranted.