Protein reduction by up to 1.5% in commercial broiler production did not affect performance but had several benefits

A series of 5 commercial feeding trials was performed from which 4 trials took place on commercial broiler farms growing between 300,000 and 420,000 broilers per cycle.

Dietary protein was gradually reduced in subsequent trial while standard diets well represented German broiler feed. Basically, average crude protein levels were decreased by 0.3 % up to 1.5 %.

For the reduced protein diets amino acid profiles according to AMINOChick® were applied. Protein levels of starter diets were not changed but those of grower 1, grower 2 and finisher diets were reduced. Moreover, in the last 3 trials, digestible Thr:Lys and digestible Arg:Lys were increased at protein reduction.

While DL-Met, L-Lys, L-Thr and L-Val were supplemented in standard feed, L-Arg and L-Ile were added in protein reduced diets as well. In one trial also glycine was added.

Protein reduction reduced soybean meal inclusion by up to 33 %. Also inclusion of oil was reduced while grains were increased.

Dietary protein reduction did not compromise growth performance, feed conversion ratio, carcass yield and breast meat yield. Therefore, productivity index was not affected as well.

Results suggested that 1 %-point crude protein reduction reduced nitrogen excretions by 16 % on average. Nitrogen utilization for deposition increased up to 69 % while standard feed allowed for about 60 % nitrogen utilization.

Water intake appeared to be lower with dietary protein reduction which tended to increase litter dry matter content. Protein reduction clearly reduced litter quantity and improved litter quality indicated by considerable improvements of footpad health.

Comparison of indirect and direct nitrogen balance indicated substantial nitrogen losses by ammonia. Such losses also play a role for the environmental impact categories acidification and eutrophication potential. Basically, protein reduction allowed for reduction of these categories of up to 12 %.

Global warming potential was decreased by dietary protein reduction by up to 10 %. However, if certified (deforestation-free) soybean meal was used, global warming potential would not be affected by dietary protein reduction.

Germany and other central European countries belong to hotspots of N2O and NH3 emissions and also of NO3- output by animal production [1]. Because of problems related to that, e.g. quality of ground and drinking water, legislation in Germany has been revised in order to monitor, control and finally reduce N-emission from livestock [2–4]. N-emission from livestock farming can be reduced by feeding low-protein diets with balanced amino acid profiles.

Reduction of protein in broiler diets has many benefits. Greenhalgh et al. [5] pronounced that dietary protein reduction is even an imperative. The list of benefits ranges from reducing emissions with all consequences on sustainability impact factors including rainforest protection to improved production conditions including better litter quality. Also animal health can be improved – all together optimizing profitability which is directly achieved by reduced diet costs but also by avoiding penalties due to footpad lesion or exceeding legal allowances for nitrogen emissions [5–8].

However, for various reasons the broiler industry is not implementing the reduced protein strategy to the utmost [8–10]. While sometimes cost-efficiency of and accessibility to supplemental amino acids plays a role, there are also doubts whether scientific findings generated under ideal and well controlled small-scale pen-facility conditions can directly be implemented into commercial practice.

Therefore, a series of feeding experiments was conducted on commercial farms in Lower Saxonia in the North part of Germany. Compared to typical German broiler feeds, crude protein (CP) contents were gradually reduced in subsequent trials.

As summarized in Table 1 five feeding experiments were conducted on two commercial broiler farms and one facility with rather large replicate pens of 250 broilers each. Farm A comprised 10 barns (1800 m2 per barn, 23 birds/m2; 4 feeding and 6 nipple drinker lines per barn) and Farm B used 8 barns (1800 m2; 21 birds/m2; 4 feeding lines; 8 nipple drinker lines). In the pen facility 10 pens (17 m2; 15 birds/m2; 4 automatic feeders, 1 nipple drinker line) were available. Dried corn silage (0.56 kg/m2) was used as bedding material on Farm A whereas straw granules were used on Farm B (0.56 kg/m2) and lignocellulose in the pen-facility (quantity was not measured but material was added when bedding became damp).

Mixed-sex, day old Ross 308 chicks were placed at the same day in each experiment in all locations. Houses were fully climate controlled. Temperature, humidity, and light programs were in close agreement with breeder recommendations. However, during last week of the Trial 5 (Farm B) ambient temperatures increased due to high outside temperatures. Basically, broilers received a standard vaccination program against infectious bronchitis and Newcastle disease.

Broilers were cropped at three (Farm A) or two (Farm B) dates. On Farm A, first thinning took place at day 29 (~1500 g/bird), a second thinning at day 34 (~2000 g/bird) and the main crop took place at days 41-42 (~2700 g/bird). On Farm B the thinning happened at days 32/33 (~1900 g/bird) and main harvest at days 40-42 (~2750 g/bird). In the pen facility Trial 4 all birds were grown to final age of 34 days.

In Table 2 analysed CP (according to Dumas method) levels of all diets are shown. These values are weighted averages of analyses per feed and phase as well as overall (WACP). For trials 1, 2, 3, and 5 a total of 97, 95, 96, and 54 feed batches were produced at a commercial feed production site. The smaller amounts of feed used in Trial 4 were also produced there. As outlined in Table 1, all feeds except starter diets contained unground wheat which was added during truck loading, and which was considered in feed formulation. All feed productions as well as all wheat batches were sampled and analysed for amino acids (AA) and crude nutrients. Analysed AA and nutrients in feeds within treatment and phase showed very little coefficients of variation of 3 % and often even lower than 2 %. In addition, analysed AA levels were very close to expected values. Altogether, this demonstrates that precise and accurate feed production is possible in a commercial environment if ingredient composition is known. In very few cases analysed values did not match expectations. Unfortunately, those cases were observed in Trial 4 and might be associated with large mixer size and small required volumes. Methionine (Met) levels of grower 2 diet reduced in protein were 16 % lower than expected (-0.09 %-Pts; -11 % lower Met+Cys). In the finisher diet reduced in protein, threonine (Thr) levels were 6 % too low (-0.05 %-Pts). It was assumed that these deviations affected at least meat yield.

In all trials, standard diets were formulated according to German commercial practice, using ingredients available at that time and adjusting diet composition to the respective ingredient price scenario (Tables in Annex). Moreover, analysed CP levels were in very good agreement with a recent assessment of German broiler feeds [11] indicating that standard diets were very representative.

Crude protein levels in German starter diets are at about 21.5 %. It is not recommended to reduce CP below this level to avoid the risk of marginal supply of nutrients during the important first days of life. While protein reduction strategy was basically according to Evonik’s AA recommendations (AMINOChick® 3.0 [12]), there were few adjustments made during the course of the project. For CP reduction only registered and commercially available amino acids were used particularly for trials on commercial farms. While the feed manufacturer used DL-Met, L-Lys, L-Thr and L-Val on regular basis for standard feeds in trials 1-4, CP reduction required using L-Arg and L-Ile (see feed compositions in Annex). However, at the time of Trial 5 the feed compounder used L-Arg already in the regular feed production. Trial 4 took place in a pen-facility because glycine (Gly) was added to meet AMINOChick® 3.0 recommendations. With respect to AMINOChick® 3.0 recommendations, particularly the minimum standardized ileal digestible (SID) essential AA and Gly-equivalents to SID lysine (Lys) ratios were considered while SID Lys were in line with standard feeds.

During this project, we learned from other projects that it might be necessary to increase the Thr and Gly supply. CP reduction per se affects the Thr metabolism. Liu et al. [13] speculated that an increased metabolic availability of acetyl-CoA due to higher starch and glucose load in low CP diets down-regulates the activity of Thr-3-dehydrogenase (TDH). This would result in reduced Thr metabolism and increased plasma Thr concentrations [13] but also suggests, that Thr as precursor for de-novo synthesis of Gly would not be effective. However, increasing dietary Thr supply under low CP conditions would counterbalance this effect on TDH while the addition of Gly would be needed to meet the Gly-equivalent requirement [14,15]. Gly equivalents are considered conditionally essential AA especially in young broilers. Therefore, SID Thr : SID Lys ratios were increased in CP-reduced diets of Trials 3, 4, and 5 (Tables Annex). In Trial 4 also Gly was added. Suggestions how to adjust the AA profile at protein reduction are implemented in the updated AMINOChick® 3.1 version.

In all trials dietary CP reduction did neither affect growth performance nor feed conversion ratio. This is true for overall final averages but also for all single crops in the large-scale trials (Trial 1, 2, 3, and 5) as well as for the pen trial (Table 3). Moreover, manual weighing of samples of 50 birds per barn or pen at feed change dates clearly confirmed this finding (Table 4). For Trial 5, average body weight of main crop in the CP-reduced treatment was 66 g behind that of the standard treatment. This can be traced back to one house. Interestingly, the direct neighbor barn showed excellent performance supplied from the same feed silo. Manual weight determinations did not indicate such a difference. It is concluded that an error in data collection and respective impact on calculations led to this deviation. Excluding this particular house would move the average body weight of the main crop to 2764 g/bird. Overall, it is concluded that dietary WACP can be reduced up to 1.5 %-points without compromising broiler performance compared to standard feeds. This is remarkable as protein content of standard feeds was already reduced during the last years but allowed for high performance. Moreover, while similar CP reduction in an earlier trial at the same pen-facility [6] failed to maintain performance, results of Trial 4 revealed same performance compared to standard. Indeed, the earlier trial lasted 41 days, while the current trial ended at day 34, but dietary CP levels were similar. One reason for the success laid likely in the concept of enhanced SID Thr: SID Lys ratios.

Water intake did not indicate a clear pattern, but water-feed ratio was reduced by 4 to 11 points in protein reduced treatments (Table 3). While there was no obvious problem with litter quality in any of the trials, improvement in footpad health indicated clear improvements of litter quality. Particularly proportion of best lesion score (no lesions) increased (Table 5). While this effect was less in first thinning, it became more pronounced at final age [16]. Basically, these findings are important as footpad health is an important welfare indicator in European broiler production. In addition, chicken feet are valuable food products which are exported to a large degree.

Growth performance, feed conversion and mortality are considered in the European Efficiency Index providing an indication on productivity. The results shown in Table 3 confirm that CP-reduction by up to 1.5 % compared to German standard does not compromise productivity in broiler production. Profitability strongly depends on feed ingredient and diet cost. The recent months showed high volatility of macro ingredient prices such as for soybean meal, oil sources and grains. Protein reduction has the potential to improve overall profitability particularly when protein sources are expensive, [8]. Oil additions usually decrease with dietary CP-reduction. Thus, also high cost for oil inclusions can be counterbalanced by reduced protein strategy.

 

Obtaining carcass evaluation data from commercial slaughter plants was a challenge in the presented projects. However, some examinations were possible in Trial 4 and 5 (Table 5). Both data sets have their limitations as only 5 male and 5 female broilers per pen were selected in Trial 4 and slaughter results of only two barns were available in Trial 5. The latter did not allow for statistical evaluation. Still, there were about 51,000 broilers behind the results of Trial 5 indicating that CP reduction did not compromise carcass or meat yield. However, for Trial 4 a lower breast meat yield was reported for the low CP treatment which would be of economic relevance. This is interesting insofar as body weights and feed conversion ratio did not indicate differences. Meat deposition is very sensitive to AA supply and any AA shortage compared to minimum specification potentially limits protein synthesis. As mentioned earlier, particularly Met (-0.09 %; grower 2) and Thr (-0.05 %; finisher) levels were much lower than specified for the reduced CP-diets. As breast meat deposition occurs especially in second half of the growth period, a negative impact of the lower AA supply on meat deposition appears likely or can – at least – not be excluded. It is concluded that CP reduction as practiced in this project does not affect slaughter traits.

Main driver for dietary protein reduction in Germany is reduction of N-output. According to German legislation, farmers have to deliver farm gate nitrogen and phosphorous balances [3]. The results presented in the upper part of Table 6 are based on indirect N-balance assuming 30 g N/kg body weight and are in line with German legislation [3,17]. In Trial 2, 20 broilers were used for whole body analyses (Table 7). Accordingly, there was no difference between treatments and an average N-content of 29.7 g/ kg body weight is in good agreement with German legislation. Muesse et al. [18] reported an average of 30.3 g/kg body weight with 30.0 for standard and 30.7 g N/kg body weight for broilers fed protein reduced diets. Interestingly, N-content was slightly but significantly higher in females than in males in that study (30.5 vs 30.2 g N/kg; P < 0.05) being in line with our results. Therefore, calculating N balances with 30 g N/kg body weight for a mixed sex flock is representative and recommended for farm gate balances.

W assuming 30g N/kg liveweight X in % of intake Y n.m.: not measured Z litter was weighted per treatment, not per pen, no stats; Within trial differences with P < 0.05 or P < 0.10 are indicated by superscripts (a,b) or (A,B)

As expected, reduction of CP in feed reduced N-intake by 1.7 % to 8.1 % which was partly significant (Table 6). N-deposition according to the indirect N-balance did not differ between treatments except for Trial 3 which is a consequence of higher average final body weights in the CP-reduced treatment. However, N-excretions were significantly (Trial 5 by trend) reduced by 7 % (Trial 2) to 22 % (Trial 4) in CP-reduced treatments. These effects suggest per %-point average CP reduction an average reduction of N-excretion of 16 % (or 13 % if Trial 1 is excluded, which showed an extraordinarily high reduction of 29 %). This is well above the rule of thumb that 1%-point of protein reduction would reduce N-excretions by 10 % and demonstrates the huge potential for avoiding N-emissions in broiler production by small dietary adjustments.

According to Uwizeye et al. [1], N-utilization in German broiler production was 60 % and on the top of the international ranking. According to a recent evaluation in Germany, N-utilization of 60 % was confirmed for 2020 and it was speculated that it would increase to 65 % by 2030 [11]. Results of Trials 3, 4, and 5 impressively indicate that this goal can be achieved or even exceeded already today. Resulting reductions in N-excretions would be an important and valuable contribution to reducing N-emissions.

Between sex significant differences with P < 0.05 are indicated by superscripts (a,b)

In commercial trials (1, 3, 5) litter quantity was reduced by 9 % on average due to CP-reduction in feed. For the pen facility trial quantity was even 31 % lower. This is of practical relevance for manure management. For instance, the effect observed in Trial 1 would mean 67 t less litter to handle for the 420,000 broiler farm and is, therefore, a welcomed side effect. Litter samples were taken in the project. However, taking representative samples from a commercial barn is a challenge and a sampling plan was developed both for the barns as well as for the pen trial [16]. Dry matter (DM) content varied considerably between different areas such as habilitation zone, drinking zone and feeding trough zone. While many samples were taken per barn and area and partly pooled for analysis, it was estimated that about 60 % of the litter is located in the habilitation zone, 30 % in the drinking zone and 10 % at feeding zone. Therefore, DM and N-content as shown in Table 6 are weighted averages. DM-content of litter increased with dietary CP-reduction whereas N-content decreased. Still, litter quantity on DM-basis was reduced in reduced CP-treatments.

Although not as statistically significant as with indirect N-balance, N-excretions based on litter analyses were reduced with lower dietary CP levels. It is remarkable that the direct N-balance always revealed lower values compared to the indirect method. Related to the N-excretion by indirect N-balance, these losses accounted for 2.2 % (standard, Trial 4) and 9.6 % (Trial 5) and are very likely gaseous N-losses mainly in form of ammonia. Maybe, high ambient temperatures during last weeks of Trial 5 amplified microbial activity in the litter and respective ammonia production.

The use of soybean meal (SBM) in animal nutrition is in discussion as especially products from Latin America are often associated with deforestation of rainforests. Therefore, Germany and other European countries signed a declaration to support and target zero net deforestation [19]. FEFAC, the European Feed Manufacturers’ Federation, released soy sourcing guidelines for the members to move the industry towards this goal [20]. In 2020 about 57 % of the domestic SBM consumption in Germany was defined as “conversion-free” according to the FEFAC definition and 37 % of the consumption was even certified [21]. Use of non-certified soy products is still high which also will have implications on sustainability aspects (see below).

1 determined with Opteinics™ ; Trial 3 calculations done for all barns, thus, statistics were possible (Lemme et al., 2023) otherwise results based on treatment averages; Within trial differences with P < 0.05 are indicated by superscripts (a,b), for soybean meal, soybean oil and wheat inclusions no statistics

The large-scale trials demonstrated that a SBM reduction by 17 % is possible with currently registered feed additives (Table 8). However, CP reduction in Trial 4 reduced SBM by even 34 %. Indeed, Selle et al. [7] indicated a potential SBM reduction of even 66 % but referred to a standard diet with an inclusion level of 33 % SBM which is 10-13 %-points higher than practiced in German standard diets. Dietary CP reduction did not only reduce the use of SBM but also that of oil. Oil additions were considerably reduced by up to 31 % in CP reduced diets (Table 8). On the other hand, wheat inclusion was increased to basically fill the gap of reduced SBM and vegetable oils (also higher inclusions of AA). Therefore, higher proportions of locally produced ingredients had been used in CP reduced diets.

Opteinics™ (Version 2.0.6 [22]) is a digital solution offering dynamic environmental footprinting and life cycle assessment (LCA). It was used to calculate various sustainability indicators (Table 8) based on the GFLI data base (GFLI v2). Details on diet formulations (ingredient composition, analysed CP content of diets) and on broiler performance (body weight gain, feed conversion ratio, mortality) were entered into Opteinics™. A German scenario was defined for feed manufacturing. Origin of grains and rapeseed products was Lower Saxonia whereas soybean meal and oil originated from Brazil. Sustainability impact factor values were available for DL-Met, L-Lys, L-Thr, L-Trp and L-Val. In case of L-Ile and Gly, figures from L-Val and in case of L-Arg figures from L-Trp were considered best estimates. Default values by Opteinics™ were used for all other input options, e.g. energy mix for feed manufacturing or animal housing. For manure management, litter was regarded as waste assuming an application to land which is already suffering from nutrient overburden.

Global warming potential (GWP) of standard treatments varied between 3717 and 3939 kg CO2 per ton of broiler carcass (Table 8). This variation is mainly explained by feed ingredients. Contribution of feed ingredients to overall GWP was 71 % in all trials. Within diet composition, SBM was the major contributor to GWP accounting for 58-62 %. Soybean oil accounted for 9-11 %, respectively. In Trial 1, contribution of SBM to GWP (58 %) was slightly lower compared to the other standard treatments as field peas were used as additional protein source.

a determined with Opteinics™ bonly the definition of soybean meal was changed in Opteinics™ c calculations done in 2024 with updated data base particularly for feed additives which may explain general differences between Trial 5 and trials 3 or 4

Dietary CP reduction decreased GWP per ton of carcass especially in Trial 3 (-9 %) and Trial 4 (-10 %). This effect is mainly attributable to lower SBM inclusions and, to a lesser effect, lower soybean oil inclusions. It should be noted that body weight gain, and thus carcass weight, was increased with CP reduction in Trial 3 (see Table 1). The reason for this effect remains unclear, but introduction of blending finisher 1 with finisher 2 diets at daily changing proportions might have allowed to be closer to the daily changing nutrient requirements of broilers. The WACP levels achieved in Trial 4 are similar to an earlier desktop LCA for European broiler production conducted by Evonik suggesting a 14 % reduction of GWP with CP reduction [23]. The outcome of that study was related to one ton of live weight, but the similarity of the magnitude of response to CP reduction in Trial 4 is remarkable (360 vs 400 kg CO2) [23].

It is a target of the German government and the feed industry [19,20] to exclusively use deforestation-free SBM for animal feed in future. Considering deforestation free SBM in Opteinics™ instead of conventional SBM, which carries a land use change (LUC) burden, had a huge impact on GWP (Table 9). GWP per ton of carcass declined by 39 % in standard diets of trials 3 and 4 by switching to deforestation-free SBM. This decreases in GWP by using deforestation-free SBM is similar to previous LCA calculations [24]. On the other hand, reduction of SBM due to dietary CP reduction had no further beneficial effect on GWP in the deforestation-free SBM scenarios. Even the contrary can be observed for Trial 4. The majority of this effect is explained by high usage of supplemental amino acids (with uncertain footprints for some of them) being in line with previous LCA [23,24]. It is concluded that the GWP per ton of carcass can be reduced by about 10 % with WACP reduction of 1.1 to 1.5 % under already efficient German production conditions. However, this effect disappears with using deforestation-free SBM.

The effects on acidification potential (AP) and terrestrial eutrophication potential (tEP) need different explanation compared to GWP (Tables 8 and 9). Despite there was some variation between standard treatments, the largest treatment effects occurred in Trials 3 and 4 (-8 to-9 % for AP; -9 to -11 % for tEP). These effects were mainly caused by reduced ammonia emissions from animal housing, manure storage and manure spreading. This is associated to improved N-utilization and respective N-excretions in treatments with reduced dietary protein. Particularly outcome of Trial 4 confirms Stubbusch and Binder [23] who reported improvements of 12 % and 11 % for AP and EP with WACP reduction of 1.5 %. Unlike for GWP, replacement of conventional SBM by deforestation-free SBM, did not affect responses on AP and tEP (Table 9). This mechanism is confirmed by others [23,24].

In a series of 5 feeding trials being conducted under German commercial conditions dietary average crude protein reductions of 0.3 % up to 1.5 % were tested in broilers. Basically, the amino acid profile was according to AMINOChick® while for protein reduced diets increased SID Thr:SID Lys and SID Arg:SID Lys ratios are recommended. For achieving reasonable protein reductions not only DL-Met, L-Lys, L-Thr and L-Val but also L-Arg and L-Ile are needed. Further protein reductions require use of glycine. Dietary protein reduction resulted in considerable reduction in soybean meal and oil usage whereas wheat inclusions increased.

Trial results suggest that average protein content in broiler feeds can be reduced by up to 1.5 %-points without compromising growth performance, feed conversion ratio, carcass yield, breast meat yield as well as productivity index. Protein reduction clearly reduced litter quantity and improved litter quality indicated by considerable improvements of footpad health.

Results further suggest that 1 %-point crude protein reduction in feed reduces nitrogen excretions by 16 % on average. Nitrogen utilization for deposition increased up to 69 % while standard feed allowed for about 60 % nitrogen utilization. Moreover, nitrogen balance calculations indicated substantial nitrogen losses by ammonia. Such losses also play a role for the environmental impact categories acidification and eutrophication potential which were reduced by up to 12 % with dietary crude protein reduction. Global warming potential was decreased by dietary protein reduction by up to 10 %. However, if deforestation-free soybean meal was used, global warming potential would not be affected by dietary protein reduction.