**Background:** Integrating warm-season grasses into traditional cool-season grass rotational grazing systems can increase pasture availability during hot, dry months and bridge the "summer slump" forage gap. Warm- and cool-season grasses differ in non-structural carbohydrate (NSC) storage mechanisms, with cool-season grasses typically having higher NSC concentrations. This is clinically relevant because current feeding recommendations for horses with metabolic dysfunction include limiting dietary NSC. Despite reported benefits for pasture yield, the impacts of this practice on equine metabolic health and the hindgut microbiome had not been previously investigated. Prior studies on diet and the equine microbiome focused primarily on concentrate vs. forage diets, with limited research on grazing horses and only one previous study on cool- vs. warm-season pasture grasses.
**Methods:** Research was conducted in 2018 at the Ryders Lane Environmental Best Management Practices Demonstration Horse Farm (Rutgers University). Eight adult Standardbred mares (age: 18 ± 0.71 yr; body weight: 537 ± 17 kg; body condition score: 5-7) were randomly assigned to two 1.5 ha integrated warm- and cool-season rotational grazing systems. Fecal samples were collected after 21 d adaptation to: initial hay diet in spring (HAY-SP), cool-season grass pasture in spring (CSG-SP), warm-season grass pasture (WSG; bermudagrass or crabgrass), cool-season grass in fall (CSG-FA), and final hay diet (HAY-FA). Oral sugar tests (OST) were conducted following adaptation to CSG-SP, WSG, and HAY-FA, with blood samples collected at baseline and 30, 60, 90, 120, 180, and 240 min post-administration. Fecal pH was measured, and short-chain fatty acids (SCFA: acetate, propionate, butyrate, valerate) and branched-chain fatty acids (BCFA: isobutyrate, isovalerate, isocaproate) were quantified by GC-MS. The V4-V5 region of the 16S rRNA gene was sequenced (Illumina MiSeq). Sequence analysis was performed in QIIME 2 and R, using bacterial co-abundance groups (BCG) identified by SCNIC. Random forest classification and regression with nested cross validation were applied, along with Spearman correlations.
**Key Results:** Random forest classification predicted forage type based on microbial composition with accuracy of 0.90 ± 0.09. Forage crude protein (CP) and NSC concentrations were predicted from microbial composition (p < 0.0001). Shannon Diversity was greater when horses adapted to WSG than CSG-SP (p < 0.05). Fecal pH was greater in horses adapted to WSG (7.56 ± 0.18) and HAY-FA (7.57 ± 0.18) than HAY-SP (6.73 ± 0.18), CSG-SP (6.58 ± 0.18), or CSG-FA (6.53 ± 0.18; p ≤ 0.02). Total BCFA were greater in CSG-SP, WSG, and CSG-FA than either hay diet (p < 0.05). Total SCFA were greater in CSG-SP and CSG-FA than hay diets; WSG did not differ from cool-season pastures. Akkermansia and Clostridium butyricum were enriched in horses grazing WSG and were positively correlated with CP and negatively with NSC. Clostridium butyricum was negatively correlated with peak plasma glucose following OST (p ≤ 0.05). Glucose AUC was lowest for WSG (42.4 ± 6.8 mg/dL*h) vs. CSG-SP (70.0 ± 6.8 mg/dL*h) and HAY-FA (76.3 ± 6.8 mg/dL*h; p ≤ 0.03). Peak plasma glucose was lower for WSG (110 ± 3 mg/dL) vs. HAY-FA (123 ± 3 mg/dL; p = 0.01), with a trend vs. CSG-SP (120 ± 3 mg/dL; p = 0.06). Forage NSC, WSC, and ESC were positively correlated with AUC (rs ≥ 0.53; p ≤ 0.007) and peak glucose (rs ≥ 0.41; p ≤ 0.04). Forage CP was negatively correlated with glucose responses (rs ≤ −0.50; p ≤ 0.01). Insulin responses did not differ by forage.
**Clinical Implications:** This study demonstrates that integrating warm-season grasses into cool-season equine grazing systems causes distinct shifts in the fecal microbiome but is unlikely to substantially impact glucose metabolism in healthy adult horses. Forage NSC, WSC, and CP were the most influential nutrients driving microbial shifts, while fiber had less impact. The higher fecal pH in horses adapted to WSG (7.56 vs. 6.53-6.73 for cool-season pastures) may suggest some benefit, though all values were within normal ranges for healthy forage-fed horses. The lower glucose AUC and peak glucose in horses adapted to WSG suggest more rapid glucose clearance, but longer adaptation periods would be needed to confirm this. The identification of Akkermansia and Clostridium butyricum as bacteria enriched in horses grazing warm-season grasses and their relationships with forage nutrients and glycemic responses provide targets for future research into equine metabolic health modulation. These bacteria have been linked to anti-inflammatory responses and improved gut barrier function in other species, but their function in the equine hindgut requires further investigation.