In vivo and modeling approaches to improve prediction of phosphorus availability in ruminants

Improving prediction of P availability necessitates understanding of P digestion and absorption mechanism in ruminants. Greater knowledge of the interaction of P with other nutrients and the utilization of dietary P in the digestive tract will improve our ability to optimize P feeding and reduce P runoff in agricultural areas. In vivo experiments were performed and the data were used to reparamterize a model regarding P digestion and metabolism.

The interaction of P and iron was investigated in lactating dairy cows by infusing 0, 200, 500, or 1250 mg/d Fe (equivalent to 0, 2, 5, or 12.5 mg Fe/L in drinking water) in the form of ferrous lactate solution into the abomasum of lactating cows. Phosphorus absorption was not negatively influenced by abomasally infused ferrous lactate, and the highest infusion (1250 mg Fe/d) approximates a drinking water iron content far above that found in most samples from the field. In the second study the effects of dietary P intake on intestinal P absorption was evaluated in eight growing Holstein steers fitted with permanent duodenal and ileal cannulas. Diets varying in P content (0.15%, 0.27%, 0.36% and 0.45%, DM basis) were fed , and increasing P intake increased the quantity of P absorbed from the small intestine linearly without affecting the absorption efficiency (mean = 59.6%). Only a small portion of P absorption occurred in large intestine and this was not affected by dietary P concentration. An absence of change of salivary P secretion at low dietary P suggested rumen function was prioritized during short-term P deficiency.

Finally the data from these experiments along with four other studies were used to parameterize the P digestion and metabolism model of Hill et al. (2008) to provide a better understanding of the digestion and metabolism of P fractions in cattle. The data used were adequate to parameterize the digestive elements of the model with good precision, and the model structure appears to be appropriate with no significant mean or slope bias. The resulting model could be used to derive P bio-availabilities of commonly used feedstuffs in cattle production. Although the model explained the data used with no apparent bias, this does not guarantee that the model parameters are valid for all conditions. Additional data are needed to evaluate this model in a wider range of scenarios.