J. Anim. Physiol. a. Anim. Nutr. zyxwvu 79 (1998), 69-79 zyxwv 0 1998 Blackwell Wissenschafts-Verlag, Berlin ISSN 0931-2439 zyxwvutsr Receipt of ms.: zy 11. 11. zy 1997 ‘Institutfur Physiologic, Physiologische Chemie und Tierernahrung, Ludwig- zy Maximilians-Universitat, Muncben, Germany, zyxw ’ WALTHAM Centrefor Pet Nutrition, Waltham-on-the-Wolds, Melton Mowbray, UK and ’Institut fur Ernahrung der Veterinarmedizinischen Universitat, Wien, Austria The development of an improved method of predicting the energy content in prepared dog and cat food By E. KIENZLE’, B. OPITZ’, K. E. EARLE’, P. M. SMrTt+’, I. E. MASKELL’ and c. IBEN’ Introduction The equations for predicting the metabolizable energy content (ME) of prepared dog and cat foods recommended by the National Research Council (NRC 1985, 1986) and the Association of American Feed Control Officials (AAFCO 1997) are sufficient for many commercially available foods. However, problems arise with products that are high in fat, such as convalescent diets, where energy content is often underestimated and high fibre diets such as weight-loss diets, where energy is often overestimated. Therefore in this study two different approaches were investigated for the prediction of digestible energy (DE), and subsequently ME, in a wide range of dog and cat foods. Firstly the proximate analysis of food [protein (CP), acid ether extract (AEE) and carbohydrates including fibre] were used to provide independent variables for comparison with the DE values from feeding trials (dependent variable). Secondly the prediction of apparent digestibility of gross energy (ad GE) calculated from the fibre content in the dry matter of the diet as described earlier by KIENZLE et al. (1998), was compared to the experimental DE values. Finally, a method for predicting ME by N-correction of DE is discussed. Materials and methods Two sets of data from dog digestion trials, which differed only in their method of dietary fibre analysis [crude fibre (CF, database A) and total dietary fibre (TF) after ENGLYST and CUMMINGS 1988 for database B] were evaluated. Data from cat digestion trials, where dietary CF analysis only were available, were compiled and are hereafter known as database C. Database zyxwvuts A contains results from 27 dog digestibility trials as previously described (OPITZ et al. 1998; KIENZLE et al. 1998) and additional data from other dog digestion trials reported in the literature (WAGNER-GEBAUER 1966; GOCKE 1970; OSTERBERG-GEBAUFA 1972; RIKLIN 1973; MEYER et al. 1980; APEL 1981; MEYER and MUNIIT 1983; MOSLINGER 1983; ELBERS 1985; ARNDT 1986; KOCH-ERHORN 1987; MUHLUM 1987; BEHI:I:LD 1988; INGWERSEN 1988; MEYER and SCHCJNEMANN 1989; STRUCKMEYER 1989; ZENTEK and MEYER 1993; ZI:NTEK 1996). Thus the total number of observations was increased to 128. Database B contains the results from 107 dog digestion trials carried out by the authors (KIENZLE et al. 1998). Database C includes data from 24 cat digestion trials carried out by the authors and 34 from reports in the literature (NORVELL 1976; MOSLINGER 1983; ZENTEK 1987; SCHNEIDER 1988; FIGGE 1989; RADICKE 1995). In all databases values for proximates, fibre and energy were based on the dry matter content of the diets. As different methods of fibre analysis (CF or TF) were used between sets of data, the calculation of the remaining carbohydrate fractions are not fully U. S. Copyright Clearance Center Code Statement: 093 1-2439/98/7902-0069 $ 14.00/0