Methods of prediction of the digestible energy content of dog foods from gross energy value, proximate analysis and digestive nutrient content

Abstract

The data from 106 digestibility trials with dogs were analysed with the objective of developing a predictive relationship between in vivo digestible energy (DE) content of foods and analytical components. The foods were commercial formulations arranged into three categories (wet, intermediate-moisture and dry) on the basis of water content and ingredient profile. Mean (n= 106) apparent crude protein, acid ether extract and nitrogen-free extract digestibility coefficients were 0.81, 0.85 and 0.79, respectively. The use of modified Atwater factors to predict food metabolisable energy (ME) gave values 104, 99 and 109% of in-vivo DE values for wet, intermediatemoisture and dry foods and thus over-estimated food energy value. Sixteen single function or multiple regression equations were generated predicting DE content of each dog food category from gross energy (GE), Atwater ME, proximate components and digestible nutrient content. The DE content of wet and intermediate-moisture foods was estimated with high precision using the single function of GE. Coefficients of determination (R2) for significant (P < 0.001) linear regression equations were 0.96 and 0.89, respectively. Other useful multiple regression equations were identified between in-vivo DE and combinations of total and digestible proximate components ( R 2 > 0.80). Less reliable relationships were obtained for the prediction of DE content of dry dog foods, probably because estimates of fibre content were not included; greatest precision (R2 = 0.60) was obtained for the linear regression between in-vivo determined DE and analysed GE content.

Protein and amino acid bioavailability estimates for canine foods

“…The use of 80% as an average aFD N for commercial pet foods is supported by Hervera (2011) and , who reported mean values for commercial dry foods close to 80% (82.4 and 81.9%, respectively). Kendall et al (1982) reported mean values (n = 106) of 83, 84, and 77% for wet, intermediate, and dry commercial canine foods. Hall et al (2013) reported standardized total tract protein digestibility coefficients for 331 dry and moist canine foods of 89.7 and 88.0%, respectively.…”

Using Gross Energy Improves Metabolizable Energy Predictive Equations for Pet Foods Whereas Undigested Protein and Fiber Content Predict Stool Quality

“…The AFFCO [4] recommends a predictive equation based primarily on fixed energy values and digestibility coefficients for dietary components (crude protein, crude fat, and carbohydrate) for estimating the ME content of dog and cat foods. The original factors in the equation described by Atwater [8] were modified by Kendall et al [9] for dogs and Kendall et al [5] for cats. The modified Atwater factors for dogs and cats (3.5 kcal/g protein, 8.5 kcal/g fat, and 3.5 kcal/g carbohydrate) provide reasonable estimates of ME for commercial pet foods with digestibilities in the range of 75 to 85% [3].…”

Prediction of digestible energy content across feed ingredients and fish species by linear regression

“…However, De and Fernández-Carmona et al (1996 reported that the acid detergent fibre fraction, compared to crude fibre content, was better correlated to ADE content in mixed diets evaluated with rabbits. Anderson et al (1991) obtained an R 2 -value of 0.87 for the accuracy of neutral detergent fibre content to predict the ADE Together with a low value as predictor found in the present study, calculation through difference, with all the accumulated errors associated with several other chemical determinations (Kendall et al 1982;De Blas et al 1992), rendered NFE unsuitable as an independent variable.…”

“…additivity) among protocols applied by different laboratories (see Glencross et al 2007a). Prediction of the ''usable'' energy value of feeds and feed ingredients from its chemical composition has been attempted for many years in numerous studies for ruminants (Wainman et al 1981;Weiss 1993), dogs (Kendall et al 1982), poultry (Carpenter and Clegg 1956;Sibbald et al 1963;Fisher 1982;National Research Council 1994), pigs (Wiseman and Cole 1979;Just et al 1984; Morgan and Whittemore 1982;Morgan et al 1987;Noblet and Perez 1993) and rabbits (Maertens et al 1988;De Blas et al 1992;Wiseman et al 1992). Prediction equations in the above studies were derived under controlled experimental conditions and analytical procedures.…”

“…However, except for giving a close approximation of the determined energy value, it must allow greatest utility with the fewest convenient predictors (Kendall et al 1982). The question was raised by Just et al (1984) whether prediction equations, that are sometimes quite accurate, could be generally applied.…”