Protein quality as a complementary functional unit in life cycle assessment (LCA)

Abstract

Goal and theoretical commentary: A number of recent life cycle assessment (LCA) studies have concluded that animal-sourced foods should be restricted-or even avoided-within the human diet due to their relatively high environmental impacts (particularly those from ruminants) compared with other protein-rich foods (mainly protein-rich plant foods). From a nutritional point of view, however, issues such as broad nutrient bioavailability, amino acid balances, digestibility and even non-protein nutrient density (e.g., micronutrients) need to be accounted for before making such recommendations to the global population. This is especially important given the contribution of animal sourced foods to nutrient adequacy in the global South and vulnerable populations of high-income countries (e.g., children, women of reproductive age and elderly). Often, however, LCAs simplify this reality by using 'protein' as a functional unit in their models and basing their analyses on generic nutritional requirements. Even if a 'nutritional functional unit' (nFU) is utilised, it is unlikely to consider the complexities of amino acid composition and subsequent protein accretion. The discussion herein focuses on nutritional LCA (nLCA), particularly on the usefulness of nFUs such as 'protein,' and whether protein quality should be considered when adopting the nutrient as an (n)FU. Further, a novel and informative case study is provided to demonstrate the strengths and weaknesses of protein-quality adjustment.

Case study methods: To complement current discussions, we present an exploratory virtual experiment to determine how Digestible Indispensable Amino Acid Scores (DIAAS) might play a role in nLCA development by correcting for amino acid quality and digestibility. DIAAS is a scoring mechanism which considers the limiting indispensable amino acids (IAAs) within an IAA balance of a given food (or meal) and provides a percentage contribution relative to recommended daily intakes for IAA and subsequent protein anabolism; for clarity, we focus only on single food items (4 × animal-based products and 4 × plant-based products) in the current case exemplar. Further, we take beef as a sensitivity analysis example (which we particularly recommend when considering IAA complementarity at the meal-level) to elucidate how various cuts of the same intermediary product could affect the interpretation of nLCA results of the end-product(s).

Recommendations: First, we provide a list of suggestions which are intended to (a) assist with deciding whether protein-quality correction is necessary for a specific research question and (b) acknowledge additional uncertainties by providing mitigating opportunities to avoid misinterpretation (or worse, dis-interpretation) of protein-focused nLCA studies. We conclude that as relevant (primary) data availability from supply chain 'gatekeepers' (e.g., international agri-food distributors and processors) becomes more prevalent, detailed consideration of IAA provision of contrasting protein sources needs to be acknowledged-ideally quantitatively with DIAAS being one example-in nLCA studies utilising protein as a nFU. We also contend that future nLCA studies should discuss the complementarity of amino acid balances at the meal-level, as a minimum, rather than the product level when assessing protein metabolic responses of consumers. Additionally, a broader set of nutrients should ideally be included when evaluating "protein-rich foods" which provide nutrients that extend beyond amino acids, which is of particular importance when exploring dietary-level nLCA.

Keywords: Amino acids; Digestibility; Environmental footprints; Food; Health; Nutrition; Nutritional LCA.

Fig. 1

Differences in A global warming potential (GWP₁₀₀; kg CO₂-eq / 100 g protein) and B land use (LU; m.²*year) / 100 g protein) per product according to internationally weighted averages calculated by Poore and Nemecek (2018a) when products are either uncorrected for protein quality (No DIAAS) or corrected using untruncated DIAAS (as labelled in both graphs), based on DIAAS values reported in Table 2. Whilst protein values are reported in Poore and Nemecek (2018a) for most products, the value used for wheat was unidentifiable (cereals were simply reported as ‘variable protein’); as a result, we adopted the protein value from the same food commodity used in Table 1 which equated to 11.2% protein (USDA ; product code 08,144); however, the GWP value was transformed from bread to wheat and may therefore be slightly misaligned with primary processing into consumable wheat, though the protein content was not sensitive following a check with similar food items in USDA (2019)