Comparison of three systems for predicting the digestible energy value of natural grassland and lucerne hays for horses.

This research evaluated and compared the accuracy and precision of three different systems—the National Research Council (NRC), the Gesellschaft für Ernährungsphysiologie (GfE), and the Institut National de la Recherche Agronomique (INRA)—in predicting the digestible energy (DE) value of natural grassland and lucerne hays for horses. The results show that the INRA system was more accurate than the other two, even though all three systems had similar levels of precision.

Research Methodology

• The research was precipitated on a number of digestibility trials; 15 trials with natural grassland hays and 9 trials with lucerne hays. The hay was harvested from the temperate zone and maintained a varied range in terms of chemical composition and DE values.
• This study took under consideration those trials which met strict experimental standards and a comprehensive standard of methods.

Results

• The INRA system showed the highest level of accuracy compared to the NRC and GfE systems. The average bias (difference) between predicted and measured DE values of grassland and lucerne hays was less with the INRA system.
• Despite the accuracy distinctions, the precision of the three systems was notably similar. The standard errors of prediction, after correcting for bias, were not significantly different amongst the systems.
• The GfE system tended to underestimate the DE value of the hays, and this underestimation was quite pronounced for lucerne hays.

Sources of Error

• The observed differences between the predicted and measured DE values were largely resulting from errors in predicting the organic matter digestibility and energy digestibility for both types of hays.
• The study suggests that categorization of hays according to botanical families (lucerne vs grassland) could improve the accuracy of DE value predictions.

Predictive Variables and Conclusions

• Appropriate predictive variables are crucial to minimize the error in DE value prediction. The differing chemical compositions and digestibility of the various cell wall components of lucerne and grassland hays mandates the choice of suitable predictive variables.
• Neutral detergent fiber (NDF) may be preferable over acid detergent fiber (ADF) for predicting the DE value of lucerne hays, as opposed to both ADF and NDF being relevant predictors of DE value for natural grassland hays.