Search

Introduction. Metabolizable energy (ME) of petfoods can be determined by animal trials or predicted using published equations.1 To date, the most commonly used equation is the modified Atwater factorial equation (3.5 kcal/g of protein and carbohydrates, 8.5 kcal/g of fat) proposed by the National Research Council (NRC) in 1985. This equation is still recommended by the Association of American Feed Control Officials (AAFCO).2 In 2006, however, the NRC has recommended newer predictive equations of ME based on digestible energy estimated as a function of dietary fiber (total dietary fiber (TDF) or crude fiber (CF)). Our objective was to compare ME determined by those different equations versus ME determined by animal trials (gold standard) in dry cat food. Material and methods. A dataset containing 243 observations with dry expanded cat foods was included in the study. For each dietary trial, proximate analysis, gross energy of the diet and feces and digestion trials on 7 cats were available. ME was predicted by modified Atwater, NRC2006 with CF or TDF equations and compared with the experimentally determined ME by simple linear regression analysis. The mean absolute difference between measured and predicted ME was also calculated as an index of error of prediction. Results and discussion. The mean (±SD) measured ME of diets was 3930±317 kcal/kg as fed (range: 3089-4654 kcal/kg). A good correlation was found between ME measured and predicted by modified Atwater equation (R²=0.809). However, mean error of prediction was quite high (293±148 kcal/kg, 7.4%) and this equation gave a general underestimate of ME. ME predicted by NRC2006 with CF and TDF equations showed a good correlation with measured ME (R²=0.803 and 0.870 respectively). Despite a tendency toward a slight underestimation of ME for energy dense diets, the precision of the prediction of NRC2006 equation with CF was improved with a mean error was 141±114 kcal/kg (3.6%). The precision of prediction was best (95±76kcal/kg, 2.5%) using TDF. Conclusion. Our results suggest that the equations recommended by the NRC in 2006 (with CF or TDF) are better than modified Atwater equation to predict ME in dry cat food.

Metabolizable energy in dry dog food is best predicted by NRC 2006 equation Calvez J, Weber M, Ecochard C, Biourge V, Royal Canin Research Center, Aimargues, France. Metabolizable energy (ME) of petfoods can be determined by animal trials or predicted using published equations. To date, the most commonly used equation is the modified Atwater factorial equation (3.5 kcal/g of protein and carbohydrates, 8.5 kcal/g of fat) proposed by the National Research Council (NRC) in 1985. This equation is still recommended by the American Association of Feed Control Officials (AAFCO). In 2006, however, NRC has recommended newer predictive equations of ME based on digestible energy estimated as a function of dietary fiber (total dietary fiber (TDF) or crude fiber (CF)). Our objective was to compare ME determined by those different equations versus ME determined by animal trials (gold standard) in dry dog food. A dataset containing 389 observations with dry expanded dog foods was included in the study. For each dietary trial, proximate analysis, gross energy of the diet and the feces and digestion trials on 6 dogs were available. ME was predicted by modified Atwater, NRC2006 with CF or with TDF equations and compared with the experimentally determined ME by simple linear regression analysis. The mean absolute difference between measured and predicted ME was also calculated as an index of error of prediction. The mean (±SD) measured ME was 3872±298 kcal/kg of diets as fed (range: 2871-4352 kcal/kg). A good correlation was found between ME measured and predicted by modified Atwater equation (R²=0.838). However, mean error of prediction was quite high (205±111 kcal/kg, 5.3%) and this equation did underestimate ME, especially for diets with an energy density above 3500 kcal/kg. ME predicted by NRC 2006 equations using CF or TDF showed a good correlation with measured ME (R²=0.817 and 0.875 respectively) and the precision of the prediction was improved as the mean error were 105±84 kcal/kg (2.7%) and 95±76 kcal/kg (2.5%) respectively. Our results suggest that the equations recommended by the NRC in 2006 (using CF or TDF) are better than the modified Atwater equation to predict ME in dry dog food, especially for diet above 3500 kcal/kg as fed.