COMPOSITION ANALYSIS OF PORK CARCASSES BY DUAL-ENERGY X-RAY ABSORPTIOMETRY

Dual-energy x-ray absorptiometry (DXA) was used as a noninvasive metho d to measure the composition of pig carcasses. A total of 181 half-car casses (10 to 51 kg, from pigs slaughtered at approximately 30, 60, 90 , and 120 kg) were scanned using a Lunar (Madison, WI) DPX-L densitome ter. The DXA measurements of fat, lean, bone mineral, and total tissue mass were compared with chemical analysis for fat, water, protein, to tal ash, and scale weight. The mean value for total tissue mass by DXA was slightly less than the mean carcass weight (32.3 kg vs 33.6 kg, P > .05, R-2 = .998). Although highly correlated (R-2 = .81), the DXA m easurement of the percentage of fat in the half-carcass was less (P < .001) than the chemical measurement (19.5 vs 24.9%). The DXA measureme nt of lean tissue mass (total mass less fat and bone mineral) was corr elated with carcass protein (R-2 = .97) and water (R-2 = .99) content. The correlation (R-2) between DXA bone mineral content and carcass as h content was only .68; however, DXA bone mineral content was more hig hly correlated with carcass weight (R-2 = .93) than was carcass ash co ntent (R-2 = .70). When we used the DXA R value (ratio of the attenuat ion coefficients for fat and lean) to predict percentage of fat in the carcass, the mean value for predicted carcass fat was 25.9% (P > .05) . Similarly, carcass protein and water content were predicted from DXA lean. Using DXA region of interest analysis, estimates of the fat con tent of the shoulder and ham regions were close to chemical values; ho wever, DXA underestimated the fat content of the loin and side regions by 20 and 28%, respectively. When prediction equations were used to e valuate DXA measurements of the half-carcasses of 28 gilts and 37 boar s slaughtered at approximately 120 kg, the half-carcasses of gilts con tained more fat (33.9 vs 27.8%, P < .001), less protein (14.1 vs 16.1% , P < .001), and less water (45.9 vs 52.1%, P < .001) than those of bo ars. These results indicate that DXA could be a valuable research tool for measuring the composition of pig carcasses. On the basis of the r esults of this study, prediction equations were revised for the DXA es timation of fat, protein, and water content of the half-carcass: Fat ( %) = 450 - (315 . DXA R value), Protein (g) = -145 + (.23 . DXA lean), and Water (g) = 150 + (.73 . DXA lean). Furthermore, it seems that se parate prediction equations are needed for regional analysis.