Chestnut–quebracho tannins shape the fecal metabolome of weaned pigs through microbiota-dependent urolithin production

Abstract Background Chestnut and quebracho tannins may support gut health in weaned piglets, but their efficacy likely depends on farm environment and microbial capacity to convert tannins into bioactive metabolites. This study evaluated the effects of a blend of chestnut-derived hydrolysable tannins and quebracho-derived condensed tannins on growth performance, diarrhea occurrence, gut metagenome, and fecal metabolome in piglets reared under two commercial farm conditions. A total of 160 weaned piglets (initial body weight 6.53 ± 0.13 kg) were assigned to a 2 × 2 factorial design comprising 2 weaning units (W1 and W2) and 2 dietary treatments (control diet or control diet + 2 g/kg tannin blend; 40 piglets per treatment × farm). The trial lasted 76 d, with weighings at T0 (entry into weaning, median 28 d), T1 (d 35), T2 (d 49), and T3 (d 76). Fecal scores and samples were collected at T1, T2, and T3 for shotgun metagenomics (155 samples per time point) and untargeted metabolomics, and targeted urolithin quantification was performed in tannin-treated piglets at T1 and T2. Results Tannin supplementation did not affect overall growth performance throughout the weaning phase, but reduced diarrhea occurrence at T2 in W2 ( P = 0.032). Species-level beta diversity was consistently affected by treatment across time points ( R 2 = 0.01–0.03; P < 0.01), whereas alpha diversity was mainly farm-driven during the early post-weaning phase. Linear discriminant analysis effect size identified farm-dependent taxonomic markers, including Escherichia coli in control pigs from W1 at T1 and Megasphaera elsdenii , Faecalibacterium prausnitzii , and Lactobacillus amylovorus in tannin-treated pigs at later time points. Untargeted metabolomics revealed treatment-related fecal signatures, with isourolithin A and urolithin B among the most discriminant metabolites in treated pigs, especially in W2. Targeted analysis identified 3 urolithin metabotypes (metabotype B in 50.0% of samples), with higher isourolithin A, urolithin B, and urolithin A in W2 than W1 at T2 ( P < 0.01). Network integration identified candidate cooperative microbial consortia associated with urolithin production, including Ellagibacter isourolithinifaciens ; these co-abundance associations are correlative and remain to be functionally validated. Conclusions Overall, chestnut–quebracho tannins were associated with a farm-dependent reduction in post-weaning diarrhea and modulated the gut ecosystem through microbiota-dependent polyphenol metabolism.