Historical flaws in bioassays used to generate metabolizable energy values for poultry feed formulation: a critical review.

Dietary energy available to animals is key for formulating feed as it is required for all aspects of the animal's life. In poultry, apparent (AME) and true (TME) metabolizable energy (ME) values have been used for feed formulation with (AME or TME) or without correction for nitrogen balance. For the past 50 yr, the accuracy of ME has been an ongoing debate, and the comparability of data produced using different bioassay systems is often questionable. Overall, the ingredient matric ME values used in feed formulation are not consistent, and to some extent, confusing. This review was to examine ME data published in the past century to elucidate the accuracy of different bioassay systems and examine the values for accuracy and useability. A variety of flaws are identified in the literature, suggesting a thorough re-thinking of feedstuff ME values currently used in feed formulation and in developing prediction equations. Two protocols, namely multiple linear regression and basal diet substitution methods, are proposed as more accurate bioassays for feedstuff ME values. AME aligns more closely with the actual energy levels of feed ingredients likely available to growing birds, which should be used for poultry feed formulations instead of AME. It is suggested that nutritionists need to carefully apply any reported AME values and only use those in formulation practice after careful scrutinizing. Any in vitro, NIR or table values must be calibrated or computed based on the values produced from flawless bioassays so as to apply the derived values accurately. Flaws identified in this literature review can be avoided with care to achieve more accurate AME. However, the assumption that the energy of individual ingredients is additive in a complete diet is still untrue at least under some circumstances. This may require efforts from industry and researchers to investigate relations among the main ingredients in a complete diet so that more accurate formulation can be performed based on the outcomes that may fine-tune the additivity assumption.