Modeling finishing pig growth performance and carcass response to dietary net energy using ingredient nutrient values from four different databases

Data from 91 grow-finish pig experiments with diets differing in net energy (NE) and other nutrients were used in a meta-analysis to model growth performance and carcass characteristics based on the diet's NE, crude protein (CP), standardized ileal digestible (SID) Lys, neutral detergent fiber (NDF), and lysine to NE (Lys:NE) ratio. Energy and nutrient concentrations in all diets were determined using values from four different databases (NRC, 2012 [NRC]; CVB, 2022 [CVB]; INRA, 2024 [INRA]; Rostagno et al., 2024 [BR]). Stepwise selection using Akaike information criteria was used for model selection considering the linear, quadratic, and interactive effects of NE, SID Lys, CP, NDF, Lys:NE ratio, and body weight. Model variable predictors were generated using a linear mixed model. Selected models indicated that ADG and G:F increased as dietary NE increases in the four databases evaluated. However, the magnitude of response in G:F differed between ingredient databases at the early and late stages of the finishing period. The model using BR database appears to overestimate ADG when diets with concentrations of NDF are used. The 4 models predicted greater changes in ADG and G:F when the Lys:NE ratio was maintained compared to when a constant SID Lys level was used. Dressing percentage (DP), backfat depth (BF), and loin depth (LD) increased and percentage lean (PL) decreased as NE increased with different magnitudes of response depending on the database. However, when a constant Lys:NE ratio was not maintained, increasing NE did not change DP predictions using the BR database, with the NRC and INRA models predicting a slight increase only at high NE levels. The CVB model suggested a similar response in DP whether Lys:NE ratio was maintained or not. For all the ingredient databases, BF and PL were predicted to linearly increase and decrease, respectively, as NE increased whether the Lys:NE ratio was maintained or not, but the changes were greater without maintaining the Lys:NE ratio. The models predicted LD to increase quadratically, with a greater response at low NE levels and the response plateauing at high dietary NE, regardless of whether the Lys to NE was maintained or not. In conclusion, the models generated provide estimates of the magnitude of the change for growth and carcass parameters of grow-finish pigs when dietary NE and other nutrients changed, with the magnitude of the changes being greater when Lys:NE ratio are maintained compared to constant levels of Lys.