Sessions

Maintaining the health and digestive physiological balance of horses requires precise knowledge of the nutrient composition of forages, which form the foundation of their daily rations. Although traditional wet chemistry procedures are considered the gold standard in forage analysis, their time and cost requirements often limit their use in everyday practice. Near-infrared spectroscopy (NIRS) offers a faster and cheaper alternative; however, the reliability of the method depends heavily on the applied calibration models, the number and quality of the reference samples. The objective of our study was to investigate the nutrient composition of meadow hay samples from Hungary intended for horses (n=42) by comparing dry matter, crude protein (CP), neutral detergent fiber (NDF), acid detergent fiber (ADF), and acid detergent lignin (ADL) values measured by wet chemistry and NIRS methods. Statistical analyses (paired t-test, Pearson correlation, linear regression, and Bland-Altman analysis) revealed significant differences (P < 0.001) between the two methods for all four evaluated parameters. NIRS technology showed the most reliable predictive performance for crude protein, demonstrating a strong linear relationship (r = 0.89) and accurately reflecting true differences (R² = 0.78). In contrast, the detection of the NDF fraction showed only a moderate correlation (r = 0.63; R² = 0.40) alongside extremely wide limits of agreement (0.59% – 8.33%), indicating significant practical uncertainty. The weakest result was yielded by ADL, where the correlation was practically non-existent (r = 0.09). The Bland-Altman analysis of ADL revealed a severe proportional bias, proving that the applied NIRS model failed to differentiate true lignin content. This may be attributed to the limitations of the calibration database, which is optimized for foreign samples with typically lower lignin contents. Based on our results, NIRS is a promising technology, and its application in daily practice is undoubtedly better than performing no forage analysis at all. However, without further validation specifically targeted at equine forages and domestic hays, its reliability currently falls significantly short of wet chemistry reference methods. Therefore, when evaluating NIRS results the current inaccuracies of the method—particularly regarding fiber fractions—must be strictly taken into consideration.