Recommendations for Ca supply of adult dogs for maintenance have undergone enormous changes over the last decades. In 1974 the National Research Council (NRC) recommended a Ca supply of approximately 3·6 g for a 15 kg dog^{( 1 )}. The German Society of Nutritional Physiology (GfE) recommended less than half of this amount^{( 2 )}, and in 2006 the NRC reduced the recommended intake of a 15 kg dog to approximately 1 g Ca^{( 3 )}, and the minimum requirements to half of this. The NRC discusses many uncertainties in the bioavailability of Ca in dogs^{( 3 )}. A recent meta-analysis⁽ Reference Mack, Alexander and Morris ^{4 )} plotted literature data on Ca intake against faecal Ca excretion in adult dogs in a modified Lukas test. The relationship between intake and faecal excretion showed a remarkable uniformity, a result which does not suggest a pronounced high variability of Ca availability. In trials with a low Ca intake of ≤30 mg/kg metabolic body weight (BW) there was, however, a high variation of faecal Ca excretion which ranged from 2 to 75 mg/kg metabolic BW. The data with very low faecal Ca excretion came exclusively from diets with meat and/or fat and/or cooked starch. These diets are likely to have a very high digestibility^{( 3 )}. This suggests that digestibility of DM may affect faecal Ca losses. In a previous study with a fibre model Kienzle et al. ⁽ Reference Kienzle, Dobenecker and Wichert ^{5 )} showed that in diets with the same Ca intake faecal Ca excretion increased with faecal DM excretion, i.e. with a decrease in DM digestibility due to fibre addition. A predictive equation was established with the independent variables faecal DM excretion and Ca intake and the dependent variable faecal Ca losses on a BW basis. The predictive equation for faecal Ca losses in mg/kg BW was: faecal losses = −33·8 + (13·6 faecal DM excretion (g/kg BW)) + (0·78 Ca intake (mg/kg BW)) (equation 1). In the present study we investigated whether this equation from a fibre model⁽ Reference Kienzle, Dobenecker and Wichert ^{5 )} would be suitable to predict faecal Ca losses in pet food from Ca intake and faecal DM excretion in a retrospective evaluation of an already existing dataset from a pet food digestion trial. The hypothesis was, that if this was true, then faecal Ca losses and consequently Ca requirements of dogs would vary with DM digestibility and food DM intake.

Materials and methods

Standard procedure digestion trials (3 d dietary change to test diet, 7 d adaptation, 5 d trial, dogs separated for feeding, one meal per d, food intake measured daily, food allowance aimed to maintain BW according to historical data, dogs individually housed during 5 d trial for total faecal collection, water ad libitum, ≥2 weeks break between trials) with twenty-five dry and fifteen moist complete and commercially produced pet foods (Table 1) were carried out. A total of forty dogs (10–35 kg BW, nineteen beagles and twenty-one foxhound-crossbreds, age 1·5–5 years, thirty-three females, seven males, majority of animals intact, all products tested in both sexes, males in more trials than females, only one breed per trial) took part in the study. In total there were 326 single observations on Ca intake and faecal excretion (275 in beagles and forty-seven in foxhound-crossbreds). All procedures were approved by the Regierung von Oberbayern, the appropriate authority in Bavaria. For each trial the dogs were adapted to the food for a total of 10 d. This was followed by total faecal collection for 5 d. Faeces were lyophylised and ground. Ca was determined by flame photometry after wet digestion. Ca intake, faecal Ca and faecal DM excretion were calculated per kg BW to allow the application of the predictive equation of Kienzle et al.⁽ Reference Kienzle, Dobenecker and Wichert ^{5 )} (equation 1). Linear regression analysis was carried out for Ca intake and faecal Ca excretion as well as between the experimentally determined faecal Ca excretion and the faecal Ca excretion as predicted by equation 1. A multiple linear regression with the independent variables faecal DM excretion and Ca intake and the dependent variable faecal Ca losses was calculated with the pooled data from the fibre model⁽ Reference Kienzle, Dobenecker and Wichert ^{5 )} and the present study. All regressions were calculated using Sigmaplot 12.5 (Systat Software GmbH).

Table 1. Crude nutrient content of diets (% DM)

Results

Ca intake and faecal Ca losses correlated significantly (r ² 0·86; P < 0·001; Fig. 1(a)). Even though Ca intake was clearly above the current recommended intake in all experiments there was a considerable number of trials with negative apparent Ca digestibility (see dots above the X = Y line in Fig. 1(a)). There was a highly significant correlation (r ² 0·87; P < 0·001) between the experimentally determined faecal Ca excretion and the faecal Ca excretion predicted by the equation of Kienzle et al.⁽ Reference Kienzle, Dobenecker and Wichert ^{5 )} (equation 1) as shown in Fig. 1(b). The regression equation calculated from the pooled data of the present study and the investigation of Kienzle et al.⁽ Reference Kienzle, Dobenecker and Wichert ^{5 )} between the independent variables faecal DM excretion (g/kg BW) and Ca intake (mg/kg BW) and the dependent variable faecal Ca losses (mg/kg BW) was: faecal Ca excretion = −9·4 + (0·7 Ca intake) + (16·8 DM excretion) (equation 2).

Fig. 1. (a) Relationship between calcium intake and faecal calcium excretion. ----, Regression line: y = 31·66 + 0·75x; sem 45·7; r ² 0·741, n 326; ––––, x = y. (b) Relationship between experimentally determined and predicted faecal Ca excretion (by use of equation 1). ––––, Regression line: y = 6·47 + 0·91x; sem = 33·6; r ² 0·80; n 326. BW, body weight.