Prediction equations for metabolizable and digestible energy in feline diets

Knowledge of the energy values and digestibility of food is important to correctly balance feline diets. Moreover, the dietary energy content directly affects its consumption, the amount of the diet offered to the cat and the ability to meet the nutritional requirements of these animals. In this sense, several methods have been developed to determine the dietary energy content, including the use of prediction equations from the chemical composition of the food. However, the prediction equations used to calculate the dietary energy values require improved accuracy due to changes in processing and chemical composition of foods and also in the emergence of new commercial diets. The objective of this study was to develop and validate the metabolizable energy prediction equations (ME) and digestible energy (DE) of complete and balanced feline diets. The chemical and energy composition of 38 complete and balanced diets for cats were obtained from different studies performed between 2010 and 2015. The equations were estimated using multiple linear regression via the Stepwise method. They were then evaluated along with equations from the literature. In order to select the best equations, the probability value (Student's t) for the null hypothesis, the coefficient of determination (R-2), error values (mean absolute deviation, mean absolute percentage error and mean square error) and bias, were taken into consideration. Although most of the equations presented had a tendency to overestimate values, they can all be used to predict dietary energy values. The best-evaluated equations were DE = 47.24 x crude protein + 32.52 x ether extract - 81.27 x mineral matter + 2532.76 presented in this study and the equation ME = 75 x ether extract + 2766 presented in the literature.