Meta-analysis of phosphorus balance data from growing pigs

Abstract

Many studies have highlighted concerns over current methods of determining endogenous P losses and P requirements in growing pigs. Therefore, a database containing observations on 350 pigs was assembled from various studies. Four functions for analyzing P balance data were considered: 1) a straight line, 2) a diminishing returns function (monomolecular), 3) a sigmoidal function with a fixed point of inflection (Gompertz), and 4) a sigmoidal function with a flexible point of inflection (Richards). The nonlinear functions were specifically reparameterized to assign biological meaning to the parameters. Meta-analysis of the data was conducted to estimate endogenous P excretion, maintenance requirement, and efficiency of utilization. Phosphorus retention was regressed against either available P intake or total P intake [all variables scaled by metabolic BW (BW(0.75))]. There was evidence of non-linearity in the data, and the monomolecular function provided the best fit to the data. The Richards equation did not fit the data well and appeared overparameterized. Estimates of endogenous P excretion of 14 and 17 mg/kg of BW(0.75) x d based on available and total P analysis, respectively, were predicted by the monomolecular equation, which were within the range reported in the literature. Maintenance requirement values of 15 mg of available P/kg of BW(0.75) x d and 37 mg of total P/kg of BW(0.75) x d were obtained, based on the monomolecular equation. Average efficiencies of conversion of dietary P to retained P were 65 and 36% for available and total P, respectively, with greater efficiency values calculated for low P intakes. Although the monomolecular equation fitted the data best, more observations at high P intakes/kg of BW(0.75) are required to determine conclusively whether P retention scaled by metabolic BW is linearly related to available or total P intake.

Aspects of nitrogen metabolism in sheep-fed mixed diets containing tree and shrub foliages

“…In support of the monomolecular model that predicts N excretion better, Kebreab et al (35) noted that the monomolecular equation was best at determining efficiencies of phosphorus utilisation for body weight gain compared with the sigmoidal model. Schulin-Zeuthen et al (23) also showed that the monomolecular model was the best predictor of the relationship between phosphorus retention and available phosphorus intake compared with linear and Gompertz models. In their study (23) , the Gompertz model was better than the linear one as observed for the relationship between UN excretion and NI in the present study.…”

“…These were a relationship exhibiting diminishing returns behaviour (monomolecular) and a relationship exhibiting sigmoidal behaviour (Gompertz) (22) . The PROC NLMIXED of SAS was used to parameterise the monomolecular and Gompertz functions with a little modification of the equation used by Schulin-Zeuthen et al (23) and the exponential model in the following forms:…”

“…A non-linear mixed-model procedure based on Schinckel & Craig (24) and Schulin-Zeuthen et al (23) was followed for these analyses. Study including parameters a, b and c was considered random in the models.…”

“…Non-linear mixed-model analysis was also used to determine the MFN and EUN excretion in three types of diets, which were estimated by the intercept on the Y-axis (i.e. b) (23) . Table 1 provides a summary of the data.…”

“…The non-linearity of regression could be expected for any kind of biological response (23) . It has been noted that monomolecular model improved the prediction of N excretion compared with either linear or Gompertz model.…”