I. INTRODUCTION
Estimating the available energy of feed ingredients is fundamental to formulating a well-balanced diet and lower the production cost. Measurement of apparent metabolisable energy (AME) is the accepted standard procedure for evaluating the available energy value of ingredients for poultry, as it is simple and straightforward and considers most of the energy losses after digestion and metabolism (Carré et al., 2014).
The measurement of AME content of feed ingredients for broilers, regardless of the methods (direct, difference and regression), is normally conducted using mash diets and the impact of feed form (FF; mash vs. pellets) on the AME and nitrogen-corrected AME (AMEn) estimates is largely overlooked. To the authors’ knowledge, no study has previously investigated the effect of FF on the AME and AMEn of individual ingredients. However, published data on the effects of FF on AME and AMEn of complete diets in broilers have been equivocal (Svihus et al., 2004; Amerah et al., 2007; Roza et al., 2018). Hussar and Robblee (1962) reported that pelleting had no effect on dietary AME or AMEn in a wheat-based diet. Svihus et al. (2004) reported an increase in AME value of a wheat-based diet as a result of pelleting compared to the same diet in mash form. In contrast, negative effects of pelleting on the AME and AMEn have also been reported in wheat- and sorghum-based diets (Amerah et al., 2007; Abdollahi et al., 2014).
These findings suggest that the estimates determined in assays using mash diets might over- or under-estimate the AME of individual ingredients when used in complete pelleted diets. This calls into question the application of AME values obtained with mash diets to commercial situations where the majority of broiler feeds is pelleted. Therefore, the objective of the present study was to investigate the impact of FF in energy evaluation of individual feed ingredients for broiler chickens.
II. MATERIALS AND METHODS
The present study was divided into two experiments; the first experimental design was a 4 × 2 factorial arrangement of four cereal grains (maize, sorghum, wheat and barley) and two FF (mash vs. pellets), and the second experimental design was a 3 × 2 factorial arrangement of three protein sources (meat and bone meal [MBM], soybean meal [SBM] and canola meal [CM]) and two FF (mash vs. pellets).
In the first experiment, the AME of test grains were determined using the direct method. Four basal diets were formulated to contain the same inclusion level (962 g/kg) of each grain. In the second experiment, the AME of test protein sources were determined by the difference method. A maize-soybean basal diet was formulated and the test diets, each containing a different protein source, were developed by replacing (w/w) 30% of the basal diet with one of the protein sources. In both experiments, each diet was divided into two equal batches and, one was offered in mash form and the second was pelleted. On d 18, a total of 288 birds were individually weighed and randomly allocated to 48 cages with six replicates per treatment (six birds/cage). Birds were fed the experimental diets from d 21 until 28 d of age, with the first three d serving as an adaptation period. Diets were offered ad libitum. For the determination of AME, feed intake and total excreta output for each replicate were recorded over the last four d of the assay. Sub-samples of excreta were lyophilised, ground and analysed for dry matter (DM), gross energy (GE), and nitrogen (N). Appropriate formulas were used for the calculation of AME and AMEn.
III. RESULTS AND DISCUSSION
The influence of grain type and FF on N retention, AME and AMEn for broiler chickens is shown in Table 1. The influence of treatments on the AMEn followed a similar pattern to that of the AME. Neither the AMEn nor the N retention were subject to an interaction (P > 0.05) between grain type and FF. However, grain type (P < 0.001) and FF (P < 0.01) had a significant effect on the AMEn. The AMEn of maize and sorghum were similar (P > 0.05) and higher (P < 0.05) than those of wheat and barley. Barley showed the lowest (P < 0.05) AMEn. Pelleting increased (P < 0.05) the AMEn values, regardless of the grain type. Grain type influenced (P < 0.001) the N retention, with bird fed maize having higher N retention compared to those fed the other grains.
Table 1 - Influence of grain type and feed form on nitrogen (N) retention, apparent metabolisable energy (AME) and nitrogen-corrected AME (AMEn) (MJ/kg DM) in broilers measured from 25 to 28 d posthatch1.
The increase in AME and AMEn of cereal grains by 0.25 MJ/kg and 0.22 MJ/kg, respectively, due to pelleting in the current study is in agreement with a previous study (Roza et al., 2018), reporting that pelleting increased AME and AMEn values of a maize-based diet by 0.27 and 0.26 MJ/kg, respectively, compared to mash form. In contrast, pelleting was reported to reduce the AME of a wheat-based diet by 0.46 MJ/kg (Abdollahi et al., 2011) and AMEn of a maize-based diet by 0.17 MJ/kg (Abdollahi et al., 2018). Amerah et al. (2007) reported a significant negative effect of pelleting wheat-based diet on AMEn (11.81 MJ/kg) compared to mash diet (12.54 MJ/kg). These contradictory results could be related to differences in diet composition, as the above studies have been conducted with complete feeds.
The lack of FF effect on N retention is in agreement with the study by Selle et al. (2012), who reported no effect of FF on N retention in sorghum-based diets. These findings are also in agreement with those by Woyengo et al. (2010) who found no effects of FF on the N retention of broilers fed a maize-soybean meal-based diet. Similarly, Favero et al. (2012) reported that FF had no influence on the N retention in turkeys fed a maize-soybean meal-based diet. In contrast, other studies showed that pelleting increased the N retention in wheat- (Pirgozliev et al., 2016) and maize-based diets (Zatari and Sell, 1990). These contradictory results could be related to differences in pelleting conditions and diet composition.
The influence of FF on N retention, AME and AMEn of protein sources for broiler chickens is shown in Table 2. Significant (P < 0.05) protein source × FF interactions were observed for the AMEn. Pelleting did not have any effect (P > 0.05) on the AMEn of SBM, reduced (P < 0.05) that of MBM, and increased (P < 0.05) the AMEn of CM. Significant (P < 0.05) protein source × FF interactions were observed for N retention. Feeding pelleted diets increased (P < 0.05) the N retention for MBM, but had no effects on those for SBM and CM.
Table 2 - Influence of protein source and feed form on nitrogen (N) retention, apparent metabolisable energy (AME) and nitrogen-corrected AME (AMEn) (MJ/kg DMs) in broilers measured from 25 to 28 d posthatch1.
Pelleting reduced the AME and AMEn by 3.58 and 3.94%, respectively, only for MBM. However, pelleting increased AMEn of CM by 7.25% compared to the mash form, which may be attributed to the effect of heat and pressure in disrupting the structure of the cell walls, thus releasing the lipids contained in the oil bodies (Jiménez-Moreno et al., 2009). No previous studies have compared the effect of FF on the AME or AMEn of protein sources for broilers. The influence of pelleting on N retention was pronounced only for MBM with an increase of 8.9% compared to the mash form.
To our knowledge, there are no reports available on the effect of FF on AME or AMEn of individual cereal grains or protein sources and the current findings suggest that the application of AME or AMEn values determined based on assays using mash diets might result in over- or under-estimation. Therefore, energy evaluation of individual ingredients under different feed processing conditions is crucial for formulating well-balanced diets for broilers.
Presented at the 30th Annual Australian Poultry Science Symposium 2020. For information on the next edition, click here. 
.jpg&w=3840&q=75)