The determination of retainable phosphorus, relative biological availability, and relative biological value of phosphorus sources for broilers.

A 10- to 21-d chick bioassay was conducted to determine the absolute retention value (ARV) for 2 different defluorinated phosphates (DF-1 and DF-2) and a reagent grade dicalcium phosphate (DCP). The total and test P in excreta regressed on feed P levels were subjected to general straight-line (linear), 1-slope broken-line, 2-slope broken-line, and polynomial regression methods to find the best analysis model. The relative biological availability (RBA) and relative biological value (RBV) for P from the 2 different defluorinated phosphates (DF-1 and DF-2) were obtained by the slope ratio method using 3 different bone measurements (% tibia ash, tibia breaking force, tibia weight) and RBV calculated using percentage tibia ash, weight gain, and feed/gain. The DCP was used as reference standard for RBA and RBV. The ARV measured at the breakpoints for test P by 2-slope analysis were determined to be 82.99% for DCP, 76.34% for DF-1, and 70.30% for DF-2. The ARV of test P determined at 0.45% NPP was 62.41% for DCP, 63.58% for DF-1, and 59.25% for DF-2. The relationship of ARV and RBA were similar in that DCP was 6% higher in ARV at the breakpoint compared with DF-1 and the RBA of DF-1 was 71 and 91% from tibia weight and tibia breaking force, respectively, compared with the bone parameters from chicks fed DCP. The DF-1 phosphate had 3 and 7% higher ARV at the breakpoint and 0.45% NPP, respectively, compared with DF-2. The RBA of DF-2 was 59 and 80% from tibia weight and bone-breaking force. The ARV of phosphate sources were independent of an arbitrary reference. The ARV for P sources provide retainable P information for industry-based feed formulation that can reduce excess P in poultry waste. The excreta P data from broilers fed increasing levels of DCP indicates that the data are best described statistically with a 1-slope broken-line regression, 2-slope broken-line regression, or polynomial regression.