Determinação indireta e validação do valor energético de alimentos de origem animal para frangos de corte

Abstract

Among the methods used for the determination of energy from feed, stands out the indirect method using prediction equations. The objective of this research was to develop prediction equations, using the principle of meta-analysis, that accurately predict the metabolizable energy values apararente corrected for nitrogen balance (AMEn) content of by-product from animal origin for broilers and cross-validate these prediction equations using a set of meat and bone meal (MBM) samples. A literature review was performed of works developed in Brazil and abroad where, had been determined the chemical composition and the AMEn values of by-product from animal origin for broilers. These data were cataloged and coded in a standart metric, thus allowing statistical analysis. The stepwise method was adopted, using the Akaike Information Criterion (AIC) for the selection of the best fit models. The equations that were choosen for the prediction of AMEn values for by-product from animal origen were AMEn are: AMEn = 5263,335 – 18,35 (PB) + 175,281 (Ca) – 724,344 (P) e AMEn = 4459,27 + 1,06 (EB) – 91,25(MM) – 78,98(PB) + 40,12(Ca), for by-product meal and meat and bone meal for brazilians feedstuf, respectivelly, AMEn = 4298,91 – 83050 (Ca) – 185,07 (P) and EMAn = 2204,77248 + 0,3469 (EB) – 217,3786 (P) for abroad feedstuf and AMEn = - 2380,51 + 99,35 (PB) + 64,08 (Ca) - 255,28 (P), for meat end bone meal for brazilians or abroad meat and bone meals. It was determined the chemical composition of five meat and bones and AMEn content for broilers at the starter phase, in order to validate the prediction equations proposed. There was no similarity between the predicted and observed data. The validation of the prediction equations at the starter phase with data determined with broilers in the pre-initial phase was not sufficient, thus suggesting that validation should be done with values determined at all phases of rearing.