An energy system is needed to optimize feed formulation and influence economics and sustainability. This study assessed the sensitivity of productive energy (PE) and classic net energy (CNE) to BW gain (BWG), feed conversion ratio (FCR), and net energy for gain (NEg) and maintenance (NEm), and developed models to predict BWG, FCR, and protein accretion (PAC). 1920 chicks in 96 pens were assigned to one of 8 blocks and 12 experimental diets, which varied in total digestible amino acids (TDAA; T1-T3) or digestible (dig.) starch and TDAA (T4-T6), or increasing (oil, TDAA; T7-T9) or reducing (soy hulls; T10-T12) densities. Blocks received a standard diet before transitioning to successive one-week treatment intervals, starting after the previous block. Birds were control-fed to control energy intake. In each block, BWG and FCR were assessed, body protein, fat, and energy gain (NEg) were determined with dual-energy X-ray absorptiometry, and heat production (NEm, fasting one) in calorimetry chambers. Diet N-corrected apparent metabolizable energy (AMEn), non-starch polysaccharides and dig. fat, dig. starch, dig. crude protein (dCP), and TDAA were determined. CNE was calculated as AMEn – heat increment and PE as NEg + NEm. A Completely Randomized Block Design with 12 treatments, 8 blocks, and 1 replication per block was used. ANOVA and Tukey tests were run. The sensitivity of the energy systems was determined by comparing Tukey outcomes for PE and CNE with those of BWG, FCR, NEg, and NEm. Linear mixed modeling used JMP. BWG, FCR, PAC, NEg, and PE were positively influenced (P<0.05) by TDAA or diet density (P<0.05) and negatively by diet dilu- tion. The PE (not AMEn or CNE) efficiency to produce BWG and PAC was stable (P>0.05) across treatments. PE was 2.3, 1.8, 1.8, and 1.8 times more sensitive to BWG, FCR, NEg, and NEm, respectively, than CNE. Models to predict BWG, FCR, and PAC based on digestible nutrients were validated (adjR2>0.95). Models to predict BWG (adjR2=0.99), FCR (adjR2=0.86), and PAC (adjR2=0.98) based on PE were validated. The one for FCR also included dCP.
In conclusion, PE is 90% more sensitive than CNE to changes in performance and actual net energy constituents (NEg, NEm) than CNE, and models to predict BWG, FCR, and PAC based on PE were developed.
