Net protein contribution from an intensive Australian pork supply chain

Introduction

With an increasing demand on our agricultural resources, it is imperative that all food production systems are making a net contribution to food supply, particularly livestock systems that in some cases consume nutrients that could otherwise be fed directly to humans. Current food consumption patterns indicate that 70% more food will need to be produced by 2050 (FAO 2018) from available land resources that have recently peaked (Taylor and Rising 2021). This is also the case for meat, with meat consumption expected to increase by 50% by 2050 (FAO 2018).

The significant challenges confronting our food systems have turned the focus increasingly to their respective sustainability. Often, popular commentary purports the virtues of plant-based protein over animal-based protein as being more sustainable. For the most part, these comparisons are based on environmental impacts or life-cycle assessments (LCA) of various foodstuffs that typically use single-score comparisons based on water, energy, greenhouse-gas emission-intensity, land-use and land use-change parameters. Whereas it is generally agreed that plant-based food stuffs have lower greenhouse-gas emissions than do animal food stuffs, there has been much discussion about the appropriateness of current LCA comparisons between plant- and animal-based proteins, with variations in methodology making it difficult to achieve an accurate comparison. Also, meals are eaten rather than individual foodstuffs and the gap in GHG emissions from a pork meal and a vegetarian meal is not as big as current commentary would suggest (Davis et al. 2010). Most reported LCA studies comparing meat and plant protein foodstuffs do not account for the GHG emissions for the waste proportion of feedstuffs (production/manufacture, retail and household), with considerably higher waste attributed to plant foodstuffs (Conrad et al. 2018).

As society developed, livestock systems thrived by producing nutrient-rich food from low-value inputs that were not used for human consumption, together with a myriad of co-products that can be used for shelter, clothing, cooking and health. In the case of ruminants, inedible fodder from generally non-arable land was the primary input, while for monogastric animals, this was undesired human food by-products, excess grain that could not be stored, dairy co-products and milling-waste streams. In modern times, livestock systems have become increasingly reliant on feed crops grown on arable land, either to change the properties of the meat products for more premium markets (as is the case with feedlot Wagyu and Hanwoo beef, Greenwood 2021), or to improve efficiency and reduce cost of production, particularly for intensively reared monogastric animals (Zampiga et al. 2021). This is argued to increase the feed-food competition, as arable land is being diverted to produce a lower quantity of food from livestock production, thereby further affecting the global supply of food (Schader et al. 2015). The use of human-edible food in livestock accounts for one-third of global grain production (Mottet et al. 2017), with projections of the growth in cereal grain use as livestock feed of 1.4% per annum (CAST 1999), whereas almost two-thirds of protein in the diets of pigs and poultry in northern Europe is derived from human-edible feedstuffs (Wilkinson 2011). What this argument overlooks is the amount of edible grain from arable land that does not meet human food standards, and hence needs to be disposed of elsewhere, nor does it measure the comparative nutritive value of grain versus animal products for humans. Regardless, the sustainability and net food contribution from livestock systems is in the spotlight and needs to be addressed.

Numerous strategies to increase sustainability in livestock production have been suggested, and these generally fall into the following three categories (Schader et al. 2015):

In reality, it is not one of these strategies that will make livestock production more sustainable, but a combination of all three; production systems need to be efficient with minimal environmental impacts, where possible the systems should be utilising feed components that are not utilised directly by humans, and there should be a more equitable distribution of livestock products for consumers worldwide (as opposed to distribution for the more affluent nations, which arguably need to consume less). Of all three, it is the consistency strategy that also moves the focus from livestock’s role as a source of high-quality protein within the food system, to another role, which is utilising resources that cannot be otherwise used for human food production. These resources are (1) grasslands, areas covering approximately two-thirds of the global agricultural area that are less suitable for arable crop production but suited for food production via ruminants, and (2) excess grain that cannot be stored, downgraded grains that are not suitable for human food or processing, food waste and by-products/co-products of food production, such as brans, whey, and oilseed cakes/meals. In addition, while many plant-based proponents are targeting livestock production as being unsustainable, 1 in 12 people still go hungry, and one-third of all food produced is lost or wasted, which equates to ~1.3 billion tonnes of food per year (FAO 2021).

To progress the discussion on the contribution that the respective plant or animal production systems make to the human food chain, and their relative sustainability, the comparison beyond consumption of feed or food stock from arable land to the contribution of the production systems to the nutrient requirements of the human population needs to be elevated, with a key focus being protein and amino acid supply. While an obvious solution would be to simply divert human-edible feedstuffs to direct human consumption, all proteins are not the same, varying in their nutritional profile, digestibility and bioavailability, environmental implications and consumer acceptance. In contrast to most plant proteins, animal proteins are of higher quality and are a complete source of indispensable amino acids (IAA) for humans (Ertl et al. 2016a). Food is also more than just a nutrient supply, and the social motivations for the consumption of animal protein, including pleasure, luxury, and indication of social status, are powerful and difficult to alter (Wyngaarden et al. 2020).

Net protein contribution (NPC) describes the contribution of a production system to meeting human nutritional requirements. If the NPC is greater than 1.0, the system is a net contributor, producing more protein than is being consumed during production. If it is below 1.0, the production system is not contributing positively and competing for protein with human nutritional requirements (Ertl et al. 2016b). To determine NPC of a livestock production system, the attributes of the protein and its quality, the efficiency of protein conversion in the animal and the contribution it makes to meeting the nutrient requirements of the consumer need to be understood.

Dietary protein quality is primarily characterised by the content of IAA. Indispensable amino acids cannot be synthesised by the human body and must be obtained from the diet (Herreman et al. 2020). The digestible indispensable amino acid score (DIAAS) is the current recommended standard to evaluate the nutritional quality of proteins (FAO 2011), and recommends the classification of proteins by using quality categories based on the DIAAS value, as follows: <75 (no quality claim); 75–99 (high-quality protein); and ≥100 (excellent-quality protein). Most cereals, legumes and forages have proteins of low nutritional quality, soybeans are considered high-quality protein, whereas potato protein and animal protein such as eggs, milk, meat and their derivatives are sources of excellent-quality protein. Animals can convert low-quality protein and store these nutrients that would otherwise be inaccessible or lost to the agri-food system and make them nutritionally accessible to humans (Wyngaarden et al. 2020).

The most obvious and classical way of measuring the efficiency of the conversion of feed to food is through a feed conversion ratio (FCR), which simply indicates the amount of feed consumed by the animal for each unit of liveweight it gains. Monogastric species (pigs and broiler chickens) convert feed more efficiently than do ruminants, with ruminants fed diets higher in grain content converting more efficiently than those consuming just forages. A lower FCR is often translated as being a smaller burden on the environment; however, land usage and the alternative uses of feedstuffs should also be considered (Peters et al. 2014).

If we consider the alternative uses for feedstuffs, in particular those suitable for humans, then conversion ratios can look very different, leading to different interpretations. On the basis of global estimates of meat production and feed use, Galloway et al. (2007) calculated the ‘total feed to meat’ ratio in ruminant and monogastric systems to be 20:1 and 3.8:1 respectively. However, when they adjusted the data to include only ‘feeds from arable land’, the ratios shifted to 3:1 and 3.4:1 respectively. This knowledge has led people to consider the human-edible protein conversion efficiency (HePCE) as a comparator of food systems. Others have continued this refinement by considering the quality of protein being consumed, and subsequently produced, resulting in being able to define the NPC of a food.

The reported NPC studies for livestock are limited to a handful of studies mainly conducted in Europe, which is in sharp contrast to LCA studies reported for various foodstuffs. In Australia, there has only been one reported NPC study for beef from rangeland and feedlot systems (Thomas et al. 2021). This paper describes the process for calculating NPC and then applies the methodology to calculate the NPC from an intensive pork supply chain in Australia.

Calculating net protein contribution

The NPC describes the contribution of a production system to meeting human nutritional needs (Ertl et al. 2016b). The NPC is calculated in several steps that consider the efficiency of turning human-edible feed into human-edible protein and the quality of the protein fed.

The first step is the calculation of the human-edible protein conversion efficiency (HePCE), the ratio of human-edible protein produced (HePp) and human-edible protein consumed (HePf).

As not all proteins are created equal, in terms of amino acid composition and ratio, the second step is to consider the ratio of the quality of the protein (PQR) that is being produced (DIAASp) and the quality of the protein being fed (DIAASf).

The third, and final, step is the multiplication of the efficiency of human-edible protein conversion (HePCE) by the change in protein quality (PQR) that occurs when this human-edible protein is fed to animals. This is the net protein contribution (NPC) of the system being examined.

Although these three equations seem straightforward, there is considerable complexity and interpretation involved in determining the amount of human-edible protein and its quality.

Human-edible protein determination – feed (HePf)

When it comes to the definition of the ‘human-edible’ fraction of a raw material, differences among authors show that estimation methodology is not clear and consistent (Laisse et al. 2016). The Council for Agricultural Science and Technology (CAST 1999) calculated estimates of the proportions of human-edible energy and protein in feedstuffs without providing any methodology behind their estimates. Wilkinson (2011) reportedly expanded these CAST (1999) proportions to broader categories, and assumed that 80% of cereal, pulse grains and soybean meal were edible for humans, while only 20% of other oilseed meals and by-products of all other grains were edible. The term human-digestible protein (HDP) is also used.

Determining an accurate measure of the potential human-edible protein of feedstuffs is difficult, as the human-edible fraction of raw materials is not one fixed and generally applicable value, but differs from region to region. It depends on the technology available to, potentially, transform the food, as well as the degree of food availability, and the eating habits of consumers (Laisse et al. 2016). A good example of this is bread wheat entering a flour mill. If wholemeal bread is being produced, then the human-edible proportion is close to 100%, falling to 85% for brown bread and 70% for white bread (Wilkinson 2011). Ertl et al. (2015) took this into account by looking at the scientific literature to support the following three scenarios for estimating the human-edible fraction (Table 1):

Similarly, Laisse et al. (2016) looked at industrial processes used in France to produce food from plant products. The proportion of human-edible protein corresponded to the ratio of the amount of protein actually edible and used in human food after processing to the total protein of the raw material prior to processing. This study also looked at potential scenarios based on strong growth in protein demand, which would lead to greater enhancement of the value of plant proteins, increasing the competition between animal and human consumption and changing eating habits. Estimates for the human-edible fraction of a range of raw materials are shown in Table 2, including our estimates for human-edible fractions of feedstuffs for Australia used in the calculations presented in this paper. A key difference between Australian and overseas values is the fact the Australian pork industry utilises processed animal proteins (PAPs) that are notable omissions from most pig diets in Europe and to some extent in North America. It should be noted that the higher the human-edible fraction within a raw material, the harder it will be for the livestock system to make a positive NPC. To this end, the values attributed for the calculations in this paper could be considered conservative, and over time, more knowledge of the ingredient source could allow us to yield a higher NPC from pork production.

Determining the human-edible protein in feed is clearly complicated by social, cultural, and regional specifics. These influence the assumptions that are made; therefore, results need to be interpreted carefully and comparisons among production systems made with caution. Values attributed to ingredient inputs into the supply chain being assessed in this paper take into account the actual end uses historically, now and in the future. For example, barley has been attributed a value of 60% (which is not inconsistent with other estimates) on the basis that the supply chain seeks parcels of grain that have not made malting specifications or are weather-damaged. Very little sorghum is used for human consumption in Australia, downgraded parcels are sought for diets, and it has an important role as a rotation crop with other more human-edible cereals (the same applies to faba beans grown in Queensland). Soybean meal is more contentious, with some contending that soybeans are grown as much for the meal as a livestock feed as they are for oil for human food (Goldsmith 2008). With vegetable-oil pricing now heavily influenced by US renewable energy mandates, the meal is more and more becoming a by-product (supply exceeding demand for human consumption only, Lusk 2022), so a conservative position of 50% has been adopted.

Meat and bone meal (MBM), as with most PAPs, is not considered to be edible by humans and therefore is given a zero value. Although human-edible protein may have been utilised in the production of PAPs, an approach similar to the allocation of GHG emissions to the primary product when MBM is being utilised as a renewable-energy source (Ariyaratne et al. 2015; Kowalski et al. 2021) has been applied. If economic allocation is utilised, Beer et al. (2007) showed that MBM comprised just 0.55% of the total economic value of the beef supply chain in the year 2000. If we included MBM at this 0.55% level, it would increase the total human-edible protein fed by 0.18%.

Human-edible protein determination – produced (HePp)

The amount of edible protein produced can be used to assess the efficiency of different production systems, allowing comparison among dairy, egg and meat production. When it comes to the definition of edible fractions, some production systems are easy to define, the edible fraction of milk and eggs is almost 100%, but meat production is more complex.

Large differences in yield exist among the species themselves on multiple levels. The primary step of slaughtering an animal is the production of a carcase, where offal, some fats and inedible products are removed. When compared with the liveweight of the animal, this dressing percentage has a large range among species, from 40% for fish to 75% for pork (Nijdam et al. 2012), whereas the yield of retail cuts from these carcases varies less, 70% for beef to 100% for fish. However, if we multiply these numbers through to a lean meat yield of the live animal, only approximately one-third of a live ruminant makes it onto the consumers table, compared with roughly 56% of a pig or a bird (Table 3).

A large component of meat and other animal protein consumption is cultural (Chiles and Fitzgerald 2018), with consumption of different species and their components having long historical origins. Therefore, the definition of the ‘edible’ fraction differs among countries, and among population groups within a country (Flachowsky et al. 2017). Pork production does not produce only meat, with a not insignificant contribution of edible offal from the ‘fifth quarter’ contributing to human nutrition in many diverse markets. Although consumption of this ‘fifth quarter’ is often low domestically within western countries, significant export opportunities exist into communities where nose-to-tail consumption is the norm, not a novelty. Edible-offal estimates range from 2.4% to 8.0% of the live market pig (Table 4).

The final component to calculating the human-edible protein of pork is determining the protein content of the pork products. There are various sources (Table 5) that have determined values for pork products. An intact pork carcase, the simplest pork product, ranges in value from 114 to 173 g protein/kg (CAST 1999; Schinckel et al. 2001; Ertl et al. 2016a). We could further refine this to the meat produced, which would allow us to use a protein content of 215 g/kg (Laisse et al. 2018b); however, we would also need to accurately determine the meat yield of a pork carcase. Offal has also been similarly defined, with Seong et al. (2014) determining the protein content of individual organs. The intestines were the lowest, with 85 g/kg for large intestine, 120 g/kg for the small intestine, with the pancreas (210 g/kg) and the liver (221 g/kg) being the highest. If all the fancy meats (heart, liver, kidney, tongue) are grouped together, then the protein content ranges from 151 to 181 g/kg (Ertl et al. 2016a; Laisse et al. 2018b). For the purpose of the calculation reported in this paper, we have estimated the protein content for the carcase and offal to be 140 g/kg and 180 g/kg respectively, on the basis of the ranges in the reviewed literature (Table 5).

Digestible indispensable amino acid score

The match between dietary supply and human protein needs is vital to support the health and wellbeing of human populations (FAO 2011). Feeding the world’s growing population in a time when constraints on land, water and food resources are increasing, requires an accurate method to define the amount and quality of protein required to meet human nutritional needs. Over 30 years ago, a joint FAO/WHO expert group recommended the use of the protein digestibility corrected amino acid score (PDCAAS) for evaluating protein quality.

In calculating the PDCAAS, the limiting amino acid score (the ratio of the first-limiting amino acid in a gram of the target food protein to that of the reference protein) is multiplied by protein digestibility, with the aim to assess how well the dietary protein can match the amino acid demand. Many reviews were undertaken by stakeholders that raised concerns and limitations of the PDCAAS method, and after 20 years of use, a new expert panel was established to address these concerns, the major criticisms being as follows (FAO 2011):

The result of this expert consultation was the establishment of a new protein-quality measure, the digestible indispensable amino acid score (DIAAS). As protein digestibility does not always reflect the digestibility of individual dietary IAAs, using a score based on the individual dietary IAA digestibility is preferable, and is defined as

Digestibility should be based on the true ileal digestibility (i.e. determined at the end of the small intestine) of each amino acid, preferably determined in humans, or, if not possible, in the growing pig or in the growing rat, in that order. The panel recognised that amino acid digestibility may vary greatly among batches of food or food ingredients and that it is impractical to assay individual batches, such that the use of tabulated data is permitted.

The panel also established recommended amino acid scoring patterns (i.e. the reference protein) to be used for calculating protein quality for dietary assessment (Table 6):

In effect, only two of these scoring patterns were used, namely (1) the infant pattern for the regulatory assessment of infant formulas, and (2) the young children pattern for the regulatory assessment of all other foods and population groups. The ratio is calculated for each individual amino acid and the lowest value is designated at the DIAAS and used as an indicator of dietary-protein quality. Independent of the absolute variation in DIAAS among raw materials that can be observed (Table 7), the dataset shows clear disparity within plant proteins.

As the calculation of DIAAS values is dependent on both amino acid content and digestibility within a food stuff, it is therefore subject to variation. Herreman et al. (2020) assessed the DIAAS of five animal- and 12 plant-protein sources where multiple datasets existed on IAA composition, crude protein content and IAA standardised ileal digestibility. When assessing the generated DIAAS value from each dataset, different datasets of the same food item fell across multiple quality categories. For consistency, the use of standard reference databases such as the USDA National Nutrient Database (https://fdc.nal.sda.gocv/ndb/), feed ingredient compositions from the Nutrient Requirements of Swine (NRC 2012) or INRAE-CIRAD-AFZ feed tables (https://www.feedtables.com) should allow for consistent values.

Whereas the DIAAS values are obtained from analysis of single protein materials, meals for humans and diets for livestock are often complex, combining raw materials with amino acid profiles that work in complement. Cereal grains are generally high in the sulfur amino acids but low in lysine, but can be combined with leguminous proteins that are generally higher in lysine, but lower in methionine and cysteine, increasing the DIAAS values of the mix. The calculation of a DIAAS for a diet of mixed ingredients is possible given that the quantities of each ingredient are known. The methodology (FAO 2011) utilises the individual raw material IAA composition and digestibility and their inclusion rate in the diet to calculate the true digestible IAA content of the mix to determine its DIAAS value. However, no account can be taken for any changes in digestibility of individual ingredients that may arise from any complementary combinations or antagonistic interactions.

Net protein contribution for pork

The NPC from an Australian intensive pork supply chain is presented in Table 8. The data used for these analyses were based on diet formulations, production and slaughter data from July 2021 until June 2022. The NPC for an intensive Australian pork supply chain is estimated to be 3.26, with this value being higher than the pork estimates reported in the literature (Table 9). This higher NPC for Australian pork is both a factor of a higher HePCE and PQR. Although there is some variance in the estimated percentage of human-edible protein available within raw materials, as a result of regional differences, which are important to consider in assessing environmental impacts (Rodríguez et al. 2014), the greater contributor is the inclusion of alternative raw materials including PAPs such as meat and bone meal, which is a key component of pig diets in Australia, contributing 20% of all digestible protein in the studied supply chain, and is not considered to be human edible. The largest contributors of human-edible protein to diets in this supply chain were of cereal grain (68.2%) and legume (30.6%) origin; however, they accounted only for 63.1% of all digestible protein. The NPC for Australian pork also compares favourably to that reported for other livestock industries, with the exception of Australian grass-fed beef (Thomas et al. 2021), which has been reported to have a very high NPC value of 1597 (noting a 100% grass-fed system should technically have an undefined value with zero human-edible protein being fed). As such, this value should be treated with some caution in the overall context of NPC values for livestock.

Food waste into landfill is a considerable issue both from the sheer volumes of food that is dumped into landfill (7.6 million tonnes per annum in Australia; FIAL 2021) and the considerable greenhouse gas emissions that are associated with this waste (17.5 million tonnes CO₂-e per annum, 3.5% of Australia’s national emissions; FIAL 2021). The omnivorous nature of the pig enables the recycling of many food-waste streams and also manufacturing by-products in their diets, and it is here that their advantage over broiler production, ‘given the huge dependence of the Australian chicken-meat industry on imported, expensive soybean meal’ (Selle et al. 2023) can see pig production contributing positively to net protein production for human consumption (Laisse et al. 2018b). It is estimated, for instance, that the French pork sector utilised nearly 1.4 million tonnes of by-products (including oilseed cakes and distillers’ grains) in 2015 in commercial stockfeed manufacturing alone. In addition, many farms manufacturing their own feed also buy meals and, often liquid by-products to incorporate into diets noting there are significant logistical and biosecurity/health challenges to overcome before these more decentralised food waste streams are readily utilised more widely by the pig industry.

Conclusions

The challenges facing our food systems, with increased demand for food from diminishing resources, the significant amount of food wasted and providing food to the people that need it most, are significant. Whereas the populist commentary constantly questions the livestock industries for environmental damage and competing for valuable food resources affecting the global supply of food, the data suggest that livestock industries can most definitely make a positive contribution to our food systems by utilising by-products/co-products of other food stuffs, and this will continue. The NPC as a measure allows the livestock industries to demonstrate this contribution. The NPC for an Australian intensive pork supply chain clearly demonstrates this positive contribution. The significant amount of food wasted globally provides the pork industries opportunities to further enhance this protein contribution and lower its carbon footprint, provided some of the logistical and health/safety challenges are overcome.

Data availability

The data that support this study will be shared upon reasonable request to the corresponding author.

Conflicts of interest

The authors are all employees of the SunPork Group and the data pertain to the operations of the SunPork Group. The authors declare no conflicts of interest.

Declaration of funding

This research did not receive any specific funding.