Prediction equations for energy values of animal meals obtained using meta-analysis

ABSTRACT The objective of this study was to determine prediction equations to estimate the nitrogen-corrected apparent metabolizable energy (AMEn) values of animal meals used in broiler diets through meta-analysis. A bibliographic review was undertaken with studies conducted in Brazil from 2000 to 2016 to catalogue information on AMEn values and the following chemical elements in the composition of the feedstuffs: crude protein (CP), ether extract (EE), gross energy (GE), mineral matter (MM), calcium (Ca), and phosphorus (P). Groups were also catalogued and formed according to sex and age of birds. Chemical correlations were analyzed, and a multiple linear regression model with the stepwise procedure was used to examine the association between the variables, which were included in the equation as a function of their importance. High and significant correlation coefficients between the independent (GE, MM, CP, EE, Ca, and P) and dependent variable (AMEn) contribute to the understanding of variations in the energy values of these feedstuffs. According to the coefficients of determination, the best equations to estimate AMEn of poultry offal meal and meat and bone meal are AMEn = 6139 − 45.5 CP + 0.356 GE − 123.5 MM ( R 2 = 0.8302 ) and AMEn = 2267 + 19.9 CP + 67.9 EE − 44.4 MM ( R 2 = 0.9021 ) , respectively.