Prediction of metabolizable energy for pigs on the basis of crude nutrients in the feeds

Abstract

A number of correlation and regression analyses were performed on data from digestibility and N-balance experiments with growing pigs (carried out during the years 1966–1982), in order to investigate whether the metabolizable energy content of pig feeds could be estimated with a reasonable degree of accuracy on the basis of their crude nutrient content (chemical analysis). The data comprise 321 diets varying widely in chemical composition and 331 feedstuff samples including 92 different feedstuff sources. Three or more proximate analyses and usually five or more digestibility and N-balance experiments were conducted with each feed sample. The statistical analyses were performed on average values of proximate analyses and average metabolizable energy values from the repeated digestibility and N-balance experiments. Crude fibre accounted for 74 and 62% of the variation in the digestability of gross energy in diets and feedstuffs, respectively. Stepwise regression analyses showed that metabolizable energy in the diets could be estimated on basis of crude fibre and gross energy with R 2 = 0.75 and CV = 5.2. The best prediction ( R 2 = 0.77 CV = 5.0) was obtained including crude protein, crude fibre, NFE substances and organic matter. A common regression analysis with the intercept fixed at zero showed that metabolizable energy could be estimated on the basis of crude protein, crude fat, crude fibre and NFE substances with R 2 = 0.77 and CV = 5.0. Estimation of metabolizable energy in the individual feedstuffs on basis of crude fibre and organic matter was much poorer ( R 2 = 0.73, CV = 10.5). The best prediction ( R 2 = 0.87, CV = 7.2) was obtained by including crude protein, crude fibre, NFE substances, NDF, gross energy and organic matter. Deleting 31 feedstuff samples of animal origin improved the accuracy of the prediction equation ( R 2 = 0.86, CV = 6.6).