Critical Review of the Procedures Used for Estimation of the Energy Content of Diets and Ingredients in Poultry.

The energy content of ingredients is estimated from tabulated values, predictive equations, and in vivo bioassays. Numerous institutions and research centers have edited comprehensive tables to evaluate the nutritive value of ingredients in poultry diets. However, the energy values provided in these tables vary widely for most traditional raw materials, including protein meals, cereals, and lipid sources. Various reasons help to explain some of the discrepancies among sources but in most cases, the differences in energy reported are not justified. Predictive equations based on near-infrared reflectance (NIRS) technology are gaining popularity for energy estimation of dietary ingredients. Online regression equations facilitate feed formulation but often the equations available are not suitable for use under many practical conditions. In vivo trials conducted at research institutions and feed companies are valid sources of information, especially for non-traditional ingredients. However, in vivo tests are of limited use under most practical conditions. In summary, each of the methods described has advantages and disadvantages. Two priorities in poultry research are the standardization of the procedures used in the in vivo trials and the online implementation of simple methods, based on NIRS technology, to predict accurately the energy content of ingredients and feeds. Nutritionists and feed mill managers should be aware of the methodology used and their applicability before selecting any of the procedures reported in this review. [ABSTRACT FROM AUTHOR]