Chemical composition of growing pigs and its relationship with body tissue composition assessed by X-ray-computed...

Ninety hybrid (mainly Large White × Landrace) pigs from 2 experimental replicates were used to study the potential use of computed tomography (CT) as a nondestructive technology for estimating the chemical body composition of growing pigs. Body tissue components (lean, fat, and bone) of 6 live pigs from each sex (boars, gilts, and barrows) were assessed by CT imaging before slaughter at approximately 30, 60, 90, 120, and 150 kg of BW. After slaughter, the empty body components were ground and frozen until analyzed for protein, lipid, ash, and moisture content. Several growth functions were evaluated and the allometric function (Y = aBW(b)), which was evaluated as log(10)chemical component weight = b(0) + b(1)log(10)BW, provided the best fit to the data. For each sex, the allometric coefficient (b(1)) for protein (0.92 to 0.99) was close to but less than 1; for ash (1.03 to 1.12), it was close to but greater than 1; for moisture (0.82 to 0.86), it was less than 1, and for lipid (1.61 to 1.71), it was greater than 1. Deposition rates (change in component weight per unit change in BW) for each chemical component were predicted using derivatives of the function. The mean deposition rates for protein and lipid were 0.141 and 0.286 kg/kg of BW gain, respectively. The deposition rate for protein was generally stable across different BW, whereas that for lipid increased as BW increased. In addition, linear, quadratic, exponential, and logistic functions were fitted to the data to study the relationship between the CT data and chemical components. The linear function was assessed to be the best equation, based on the Bayesian information criterion. The prediction equation for protein (kg) = -1.64 + 0.28 × CT lean (kg), and for lipid (kg) = -0.69 + 1.09 × CT fat (kg), had R(2) values of 0.924 and 0.987, respectively. Sex had no effect (P > 0.05) on the prediction of protein and lipid. The effect of BW was not significant (P > 0.05) for the prediction of lipid, but it was significant (P > 0 0.05) for the prediction of protein. However, the addition of BW to the base prediction equation for protein resulted in an increase of only 0.013 in the R(2) value. It was concluded from this study that CT scanning has great potential as a nondestructive technology for estimating the physical and chemical body composition of pigs. Additional research is required to validate the utility and accuracy of the prediction equations.