Musculoskeletal health in an ageing society: a role for dietary protein?

Globally, ageing is associated with increased healthcare costs and social service needs^{(Reference Beard, Officer and de Carvalho9)}. In addition, the gap between lifespan (i.e. total lived age) and health span (i.e. years of life free from disease)^{(Reference Garmany, Yamada and Terzic10,Reference Olshansky11)} continues to grow, compounded by a decrease in habitual physical activity levels and increased prevalence of diseases associated with advanced age^{(Reference Volpi, Nazemi and Fujita12,Reference Wilkinson, Piasecki and Atherton13)} . Indeed, lifelong engagement in exercise (e.g. master athletes) results in the better maintenance of skeletal muscle mass into older age and may be considered a more true model of inherent ageing (i.e. represents ageing, per se, rather than the detriments seen due to inactivity)^{(Reference Hawkins, Wiswell and Marcell14,Reference Harridge, Magnusson and Saltin15)} . Moreover, whilst the cause(s) of age-related muscle loss (otherwise termed ‘sarcopoenia’) is clearly multifaceted, a key contributor is malnutrition, and in particular a reduced dietary protein intake^{(Reference Narici and Maffulli16)}. Indeed, higher protein intakes have been associated with greater retention of lean mass in older individuals in some^{(Reference Houston, Nicklas and Ding17)}, but not all^{(Reference Ni Lochlainn, Bowyer and Welch18)}, studies. Hence, with advanced age, it seems prudent to tailor protein intake recommendations to counter age-related changes in the metabolic response of skeletal muscle to ingested protein, as well as reduced physical activity. Importantly, with regards to attenuating age-related muscle loss, the roles of skeletal muscle go beyond locomotion to critical actions such as chewing and swallowing, breathing, maintenance of body posture and thermogenesis. Combined with the misalignment of health- and lifespan, this highlights an urgent unmet need in an ageing society to comprehensively understand protein intake requirements and develop appropriate recommendations.

A brief historical perspective on devised protein requirements and recommendations for adults

According to published records, proteins were first recognised as a distinct class of biological molecules by French chemist Antoine-François Fourcroy in the 18th century and described by the Dutch chemist Gerardus Johannes Mulder as ‘unquestionably the most important of all known substances in the organic kingdom. Without it, no life appears possible on our planet. Through its means, the chief phenomena of life are produced’^{(Reference Mulder21,Reference Mendel22)} . Since the 18th century, or even before, many scientists have dedicated their professional careers to determining protein requirements and recommendations for human subjects (Fig. 1). The first recorded evidence of protein requirements and recommendations appeared in about 1877 and was credited to Carl von Voit who was a German physiologist and dietitian. von Voit made the recommendation that a 70 kg person who undertakes a ‘moderate’ level of work should consume 118 g protein daily and referred to this value as the ‘lowest limit’ of supply to avoid risk of ‘damage to health’^{(Reference Lusk23,Reference Sigerist24)} . This figure was devised despite a dietary survey carried out in Munich by von Voit, that suggested a protein intake of 52 g daily was sufficient for good health (later, in about 1900, von Voit would recommend a protein requirement of 1⋅0 g/kg of body weight daily based on the dietary intake of highly productive factory workers)^{(Reference Lusk23,Reference Sigerist24)} . In contrast, at the beginning of the 20th century, supporters of nutritional reform recommended a daily protein intake of <30 g. A key representative of nutritional reform was the Danish nutritionist, Mikkel Hindhede, who conducted experiments demonstrating long-term adherence to diets with a daily protein intake of <30 g^{(Reference Hindhede25)}. Hindhede also suggested that earlier estimates of >100 g daily were exaggerated and highlighted the observation that recommendations were based on non-animal foods that were considered ‘less protein dense and cheaper than a meat-based diet’. As such, these recommendations were claimed to have helped avoid famine during the First World War^{(Reference Hindhede25)}.

Fig. 1. A brief summary of the key landmarks in the historical evolution of dietary protein and amino acid (AA) requirements and recommendations for human subjects. Dietary recommendations are provided relative to body weight (i.e. kg). EAA, essential amino acids; EAR, estimated average requirement; FAO, Food and Agriculture Organisation; RDA, recommended daily/dietary allowance; UNU, UN University; WHO, World Health Organisation.

During the 20th century, with significant advances in science and communication, a concerted effort was made by international committees to devise universal guidelines for protein intake recommendations. Whilst the originally proposed daily allowance of 1⋅0 g protein per kg of body weight for adults represented a figure of appealing simplicity, this recommendation was not based on scientific evidence. Accordingly, in 1955 the FAO assembled a committee, led by Professor Emile Terroine, to define the average/minimum requirements and the recommended allowance for dietary protein (see later for definitions of each)^{(Reference Périssé26)}. The average requirement for protein intake was set at 0⋅35 g/kg/d for adults. Protein requirements and recommendations were revisited in 1963 by a Joint FAO/WHO Expert Committee^{(Reference Périssé26)}, with an average protein requirement of 0⋅59 g/kg/d agreed that factored in nitrogen losses and the additional requirements for growth.

The FAO/WHO Expert Committee reconvened on multiple occasions in the years that followed to continue to refine protein recommendations, which included, for a brief period, sex-specific guidance (0⋅44 and 0⋅40 g/kg/d for men and women, respectively). In 1981, a joint FAO/WHO/UN University Expert Committee calculated the mean protein requirement based on short- and longer-term nitrogen balance studies (this technique is discussed later) and concluded no clear evidence of sex differences in nitrogen losses and thus protein requirements or recommendations^{(Reference Périssé26)}. The average requirement for highly digestible, good-quality protein (e.g. meat, milk, fish, egg) was set at 0⋅60 g/kg/d for both sexes. To translate this estimate of the average protein requirement to a level sufficient to cover individual variation within a population group, an estimated value of two standard deviations above the average physiological requirement would be expected to meet the needs of the majority of the population. Hence, the lower end of the safe intake of good quality, highly digestible protein was therefore set at 0⋅75 g/kg/d. In 2007, and informed by a meta-analysis of nitrogen balance studies, a Joint FAO/WHO/UN University expert consultation recommended 0⋅83 g/kg/d protein to meet the requirements of most (97⋅5 %) healthy adults^{(Reference Millward27,28)} (also see Rand et al.^{(Reference Rand, Pellett and Young29)}). To this end, these data provide the fundamental evidence base which informs protein requirements and recommendations by relevant authoritative bodies today.

Amino acid requirements: taking dietary protein requirements and recommendations one step further?

The concept of devising AA, in addition to or instead of protein, requirements and providing specific recommendations for each EAA is appealing given that not all dietary protein sources contain an identical AA profile. However, this concept is challenging to implement in practice. Hence, recommendations for intake of specific AA have been limited, as discussed elsewhere^{(Reference Rodriguez and Miller30,Reference Rodriguez31)} . The concept of AA requirements is ostensibly based on knowledge that the EAA content of a protein source, rather than the gross protein per se, dictates the metabolic availability and ‘quality’ of a protein source, with implications for muscle anabolic potential, and must be ingested in the diet. A seminal rodent study in the early 20th century revealed low-survival rates in rats fed with a diet exclusively containing zein (derived from maize which constitutes an ‘incomplete’ low-quality protein, deficient in lysine and tryptophan) compared with rats fed with casein from cow's milk, a high-quality protein with a full complement of EAA. Through a series of investigations^{(Reference Rose, Haines and Warner32–Reference Rose and Wixom35)}, this led biochemist and nutritionist, Professor William Cumming Rose, to the discovery of the EAA threonine^{(Reference Rose, Haines and Warner32–Reference Rose and Wixom35)}. Through manipulation of rodent diets, Rose demonstrated that ten AA are essential for rats and have to be consumed via diet as they cannot be synthesised in sufficient amounts without dietary intervention. Follow-up work demonstrated that eight AA are essential for adult human subjects (isoleucine, leucine, lysine, methionine, phenylalanine, threonine, tryptophan and valine). Longer-term studies established histidine as essential for adult human subjects, bringing the total to nine (and eleven non-essential amino acids)^{(Reference Kopple and Swendseid36)}. In brief, Rose's human experiments involved the provision of rudimentary diets to healthy male graduate students, consisting of maize starch, sucrose, butterfat without protein, maize oil, inorganic salts, known vitamins, a large brown ‘candy’ made of liver extract flavoured with peppermint oil (to supply any unknown vitamins) and mixtures of highly purified individual AA. In addition to nitrogen balance data to support his conclusions, Rose also noted a higher prevalence of symptoms of nervousness, exhaustion and dizziness when participants were deprived of an EAA^{(Reference Rose, Haines and Warner33)}. Although Rose's work received some criticism including concerns over the validity of prescribed diets, his findings remain fundamental to our current understanding of human AA requirements and human protein metabolism. Accordingly, subsequent research revealed that only EAA are required to increase MPS^{(Reference Tipton, Ferrando and Phillips37)}. Notwithstanding, whilst all EAA must be obtained through the diet, even when not acquired acutely (i.e. during a single meal), a true AA deficiency is difficult to achieve longer-term via a habitual the diet which likely contains a variety of different proteins and wholefood sources to an extent that complete deficiency is avoided^{(Reference Mariotti and Gardner38)}. The key factor(s) that discerns an EAA from a non-essential amino acid in human subjects remains to be fully established, but is likely attributed to a combination of evolutionary mechanisms and as a means to regulate energetically expensive cellular processes (e.g. MPS)^{(Reference Mitchell, Phillips and Hill39,Reference Browne and Proud40)} . Moreover, there is no evolutionary advantage for the endogenous generation of EAA, as they are sufficiently available through a ‘standard diet’, and circumvent the need to use long, complicated and energy-consuming pathways that would be required to synthesise sufficient quantities of all EAA.

Nitrogen balance: determining protein requirements in human subjects

The requirements for EAA and thus dietary protein have been determined by multiple methods to inform protein requirements and recommendations. Historically, descriptive or gross measures including growth and nitrogen balance have been used. To this end, the estimated average requirement and recommended dietary allowance (RDA) (discussed later) have been determined by the single endpoint of the amount of protein intake required to maintain nitrogen equilibrium (namely food nitrogen intake minus nitrogen excreted [urine, faeces, sweat skin and hair]), otherwise referred to as ‘nitrogen balance’^{(Reference Wolfe, Cifelli and Kostas41)}. However, concerns have been raised regarding the use of this technique for determining protein requirements, not least that recommendations are based on good-quality protein^{(Reference Rand, Pellett and Young29)} and that readouts of nitrogen balance has limited utility beyond nitrogen balance itself which lacks sufficient physiological relevance to outcomes related to lean body mass^{(Reference Pencharz, Elango and Wolfe42)}. In brief, nitrogen balance requires a minimum of 3 d per level of test intake (i.e. amount of dietary intake of protein) and 7–10 d of adaptation are needed to each intake of protein^{(Reference Tinline-Goodfellow, West and Malowany43)}. In addition, complete collection and quantification of all sources of nitrogen excretion, mostly in urine and faeces, are required but this is practically challenging. Moreover, the nature of the nitrogen balance calculation is often associated with significant variability given that nitrogen intake and excretion are independently associated with significant error, thereby lacking sufficient sensitivity^{(Reference Pencharz, Elango and Wolfe42)}.

The validity of the nitrogen balance technique has also been criticised given that a zero nitrogen balance on a lower protein intake may reflect biological accommodation (i.e. individuals can adapt to insufficient/suboptimal protein intakes by reducing nitrogen excretion)^{(Reference Pencharz, Elango and Wolfe42,Reference Højfeldt, Bülow and Agergaard44–Reference Højfeldt, Bülow and Agergaard46)} . In addition, studies have demonstrated an apparent disconnect between positive nitrogen balance and projected improvements in lean body mass^{(Reference Wolfe, Cifelli and Kostas41,Reference Pencharz, Elango and Wolfe42)} . Clearly, there are several limitations and additional considerations associated with the nitrogen balance technique that question the validity of current estimates of protein recommendations^{(Reference Wolfe, Cifelli and Kostas41,Reference Pencharz, Elango and Wolfe42,Reference Weiler, Hertzler and Dvoretskiy47)} . Indeed, even as early as 2002 the ‘dietary reference intakes’ report from The Food and Nutrition Board of the Institute of Medicine (The National Academies) stated that ‘due to the shortcomings of the nitrogen balance method, it is recommended that the use of nitrogen balance should no longer be regarded as the “gold standard” for the assessment of the adequacy of protein intake and that alternative means should be sought’ (Institute of Medicine of the National Academies)^{(Reference Wolfe, Cifelli and Kostas41)}. In contrast, recent data suggest that nitrogen balance may be useful in detecting EAA deficiencies in low intake states given that consumption of the protein RDA (~0⋅80 g/kg/d) following a strict, low-quality protein, vegan diet for ≥1 year has been shown to be inadequate to achieve nitrogen balance^{(Reference Bartholomae and Johnston48)}. Furthermore, the reanalysis of previously published nitrogen balance data, when using a different analytical approach (via two-phased linear crossover analysis), revealed a higher population estimate of about 1⋅0 g/kg/d, which approaches the protein requirement determined using more contemporary methods^{(Reference Humayun, Elango and Ball49)}.

As a potential alternative to nitrogen balance for determining protein requirements, the nitrogen-15 (¹⁵N, a rare stable isotope of nitrogen) end-product method has also been proposed^{(Reference Hudson, Baum and Diaz50,Reference Wolfe, Kim and Church51)} , a technique that has been employed for >50 years to measure the turnover of the entire nitrogen pool of the body^{(Reference Wolfe, Kim and Church51)}. In brief, the nitrogen-15 end-product method involves the oral ingestion of a labelled nitrogen (e.g. ¹⁵N-glycine, ¹⁵N-alanine) to determine nitrogen flux, or nitrogen turnover at the whole-body level. This method is based on the assumption that metabolically active nitrogen is freely exchanged between nitrogen-containing tissues and the metabolic nitrogen pool (e.g. AA)^{(Reference Sprinson and Rittenberg52)}. Nitrogen appearance in the metabolic pool occurs exogenously via the diet and endogenously via protein breakdown with nitrogen disappearance occurring through protein synthesis and nitrogen excretion as end-products, primarily urea or ammonia in the urine^{(Reference Fern, Garlick and Waterlow53)}. Measurements of whole-body protein breakdown, in addition to synthesis, can also be calculated by measuring protein intake. However, similar to nitrogen balance, this technique is associated with measurement error and technical challenges. The calculation of nitrogen flux, protein synthesis, protein breakdown and net protein balance using this technique are described elsewhere^{(Reference Wolfe, Kim and Church51,Reference Ferrando, Lane and Stuart54,Reference Børsheim, Chinkes and McEntire55)} .

Contemporary approaches for determining whole-body protein requirements

A more contemporary and arguably comprehensive method to determine protein requirements is called the indicator amino acid oxidation (IAAO) technique^{(Reference Pencharz, Elango and Wolfe42,Reference Elango, Ball and Pencharz56–Reference Matsumoto, Narumi-Hyakutake and Kakutani58)} . The most common application of IAAO is to provide an oral AA mixture to human subjects. Using IAAO, an EAA is ‘labelled’ with a stable isotope (usually ¹³C) and the appearance of this label in the breath (carbon dioxide, ¹³CO₂) is used to quantify AA oxidation as an indicator of protein or a single EAA requirement. IAAO was developed based on the principle that all EAA are required in sufficient quantities for protein synthesis. In theory, if a single AA is limiting or provided in excess, AA oxidation will be observed. Stable isotopes are naturally occurring atoms (e.g. carbon, oxygen, nitrogen, sulphur) containing extra neutrons, whose metabolic fate replicates their more common isotope, permitting a distinction between common and rare isotopes that are detectable (or ‘traceable’) in biology. Similar to nitrogen balance, the IAAO technique provides subjects with graded protein (or AA) intakes across multiple trials during which the indicator AA is provided at a continuous, excess, amount and adaptation of only 3–4 h is required^{(Reference Elango, Ball and Pencharz59)}. When the intake of protein/AA is low, the availability of one or more EAA will be limiting for protein synthesis, and thus will be oxidised. As protein intake levels increase, the excess and thereby the oxidation of the indicator AA decreases, reflecting an increased incorporation of AA into body protein. The AA intake level at which AA oxidation becomes minimal is termed the ‘breakpoint’ and represents the intake level that maximises whole-body protein synthesis rates. The same concepts apply for the assessment of EAA requirements, except that graded amounts of the EAA are provided whilst all other AA are provided in excess^{(Reference Trommelen and van Loon3)}. Fundamentally, this technique is based on the principle that beyond lean tissue itself, there is no inactive compartment to serve as a reservoir for AA and therefore AA must be partitioned between incorporation into body protein or for oxidation.

Evidence from the application of IAAO suggests that current recommendations for dietary protein may underestimate minimum protein requirements for whole-body balance by as much as 50 %, including in older people^{(Reference Tinline-Goodfellow, West and Malowany43,Reference Tian, Liu and Zhang60–Reference Bandegan, Courtney-Martin and Rafii63)} . Indeed, a recent review of the literature suggests that protein requirement estimates using the IAAO method range from about 5–260 % greater than the RDA across a range of populations^{(Reference Matsumoto, Narumi-Hyakutake and Kakutani58)}. A key criticism of IAAO is that participants are only adapted to the test intake on the study day, however, adaptation to longer periods does not seem to impact estimates of dietary requirements^{(Reference Elango, Ball and Pencharz56,Reference Elango, Humayun and Ball64)} . In addition, it also is feasible that oxidation (and thus IAAO) reflects fluctuations in protein synthesis only rather than protein breakdown^{(Reference Elango, Ball and Pencharz56)} that serves as a key component in accurately determining net protein balance, albeit less critical in healthy adult populations^{(Reference Tipton, Hamilton and Gallagher65,Reference Carbone and Pasiakos66)} . Clearly, current protein recommendations warrant consideration in the context of best available tools to provide valid estimates of required intakes, and this may be achieved with the employment of multiple assessments including IAAO. Understanding the specific EAA requirements across the health- and lifespan continuum and the provision of easy-to-access resources relating to dietary protein sources is of particular interest, particularly in the context of healthy ageing. Moreover, other emerging methods to measure protein kinetics may be suitable for estimating protein requirements including the use of ²H oxide (heavy water) and D3-creatine, but require investigation to confirm their utility in accurately determining protein requirements and recommendations in a range of populations^{(Reference Wolfe, Kim and Church51,Reference Wilkinson, Brook and Smith67)} .

Current UK recommendations for dietary protein intake: a misunderstood concept?

Formalised dietary protein recommendations have been devised for >100 years^{(Reference Sherman68)}. Nonetheless, optimal and/or recommended protein intakes across the health- and lifespan remain unclear^{(Reference Burd, McKenna and Salvador4)}. The current UK RDA for protein intake is based on a normal distribution of population requirements and an estimated average requirement (satisfying the requirements of about 50 % of the population) of about 0⋅55–0⋅60 g/kg/d, and is set at 0⋅75 g/kg/d for healthy adults (~50–55 g daily for a 70–75 kg individual). The general purpose of the RDA, which is set at the EAR plus two standard deviations, is to meet basic nutritional requirements and avoid deficiencies in 97–98 % of the population. Nevertheless, the protein RDA can easily be misrepresented and misinterpreted^{(Reference Wolfe, Cifelli and Kostas41)}. Indeed, the protein RDA is not a ‘recommendation’ nor an ‘allowance’, but rather an ‘adequate intake amount’ to avoid a negative nitrogen balance in the majority of the population^{(Reference Wolfe, Cifelli and Kostas41)}. This notion creates a further problem in that, unlike other macronutrients, the RDA for protein is not based on a health outcome (e.g. association with disease, function, lean tissue mass). Based on its definition, the protein RDA is therefore not intended, nor does it provide, an estimation of ‘optimal’ intakes, or exclude the possibility that less than the RDA represents a sufficient or optimal intake for a given individual.

In addition to the RDA, the acceptable macronutrient distribution range for protein is set at 10–35 % of total energy intake and was developed to express dietary recommendations in the context of a complete diet. However, in isolation the acceptable macronutrient distribution range is not considered helpful for dietary guidance. Indeed, the lowest level of protein intake reflected in the acceptable macronutrient distribution range is higher than the RDA (when reference body weights of 57 and 70 kg are assumed for women and men, respectively)^{(Reference Wolfe, Cifelli and Kostas41,Reference Wolfe and Miller69)} . In addition, if an individual were to meet the RDA for all macronutrients, only about 40 % (depending on age, sex, activity level and other factors) of the total energy requirement would be met, highlighting a wider issue with macronutrient recommendations^{(Reference Wolfe, Cifelli and Kostas41)}. Moreover, protein recommendations are not typically further delineated on the basis of other characteristics (e.g. age, sex, activity level, health status [exceptions discussed later]), despite data suggesting specific health benefits at levels of protein intake that significantly exceed the RDA^{(Reference Wolfe, Miller and Miller70,Reference Moore, Churchward-Venne and Witard71)} . Based on its purpose and definition, the protein RDA may more appropriately be termed the ‘recommended minimum intake’, alongside recommendations to increase daily intake, as previously proposed^{(Reference Wolfe, Cifelli and Kostas41)}. However, we would apply some caution to this recommendation as the RDA, or below, may represent a level of intake that is optimal for a small proportion of the population. Indeed, a population-wide recommendation to increase the protein RDA, or at least a suggestion that the RDA is the absolute minimum, may not be sensible for individuals with existing kidney damage, whether this condition is formally diagnosed or is unknown. The discussion of personalised recommended v. optimal v. maximal protein intake(s) is an important consideration^{(Reference Burd, McKenna and Salvador4,Reference Wolfe, Miller and Miller70,Reference Moore72,Reference Lonnie, Hooker and Brunstrom73)} . Undoubtedly, numerous factors warrant consideration when devising protein recommendations across the health- and lifespan continuum and, where possible, a tailored approach to protein nutrition should be considered as part of a well-balanced diet to supply the increasing demand of specific nutrients associated with ageing to avoid malnutrition^{(Reference Paddon-Jones and Leidy74)}.

Refining per meal protein recommendations for skeletal muscle anabolism in older age

The primary metabolic regulator of skeletal muscle mass is the stimulation of MPS and has been shown to correlate with longer-term changes to skeletal muscle outcomes^{(Reference Abou Sawan, Hodson and Malowany75)}. The use of stable isotope methodology to measure the acute response of MPS to a single-protein bolus has provided the scientific foundation to refine protein recommendations on a per meal basis. In healthy young adults, close to a consensus has been reached that a per meal dose of about 20–30 g (~0⋅25–0⋅30 g/kg) of high-quality protein (equating to about 3 g leucine; about 10 g EAA; about 5 g branched-chain amino acids [BCAA]) is sufficient for the maximal (but transient; about 2–4 h) stimulation of MPS. However, the AA composition, specifically the EAA profile and leucine content (the intracellular appearance of which seems particularly important for the stimulation of MPS^{(Reference Garlick76)}) of the protein source will ultimately influence the required protein dose for the maximal acute stimulation of MPS^{(Reference Volpi, Kobayashi and Sheffield-Moore77)}. Further, whilst young individuals demonstrate a robust response of MPS to these anabolic stimuli, a blunted response has been observed in older adults, termed ‘anabolic resistance’, which likely underpins muscle loss observed with ageing^{(Reference Moore, Churchward-Venne and Witard71,Reference Shad, Thompson and Breen78)} . For example, Moore et al. ^{(Reference Moore, Churchward-Venne and Witard71)} performed biphasic linear regression and breakpoint analysis using data sets derived from multiple laboratories that measured the acute response of MPS after the ingestion of varying amounts (0–40 g) of high-quality dietary protein (as a single bolus) in healthy older (mean of 71 years) and younger (mean of 22 years) men when normalised to body mass^{(Reference Moore, Churchward-Venne and Witard71)}. Whilst no difference in basal postabsorptive MPS rates were observed between age groups, biphasic linear regression and breakpoint analysis revealed the slope of first-line segment was lower in older men and that MPS reached a plateau after ingestion of 0⋅40(sem 0⋅19) g/kg per meal (95 % CI: 0⋅21, 0⋅59 g/kg per meal) and 0⋅24(sem 0⋅06) g/kg per meal (95 % CI: 0⋅18, 0⋅30 g/kg per meal) in older and younger men, respectively. These data suggest that older adults may require almost 2 × the per meal dose of protein to achieve a comparable MPS response to their younger counterparts^{(Reference Moore, Churchward-Venne and Witard71)}. Moreover, the large overlapping CI (0⋅21, 0⋅59 and 0⋅18, 0⋅30 g/kg per meal for older and young, respectively) highlight the inherent biological variability in the MPS response to ingested protein, particularly with advancing age, suggesting personalised protein recommendations regardless of age, are warranted when devising future protein recommendations. However, it is worthy of note that whilst protein intake is an independent, albeit small, predictor of better retention of muscle mass in older age, exercise represents the main stimulus for muscle adaptive remodelling, particularly resistance exercise^{(Reference Houston, Nicklas and Ding17,Reference Traylor, Gorissen and Phillips79–Reference Egan and Zierath83)} . Therefore, even in scenarios where alternative protein recommendations are reached, this could elicit only a small effect on muscle anabolism and remodelling in the absence of resistance exercise^{(Reference Houston, Nicklas and Ding17,Reference Traylor, Gorissen and Phillips79–Reference Egan and Zierath83)} . In addition, it is important to caveat that these findings presented by Moore et al.^{(Reference Moore, Churchward-Venne and Witard71)}, and others, are predominantly isolated to skeletal muscle and, even more so, the myofibrillar (i.e. contractile) proteins within skeletal muscle (largely from quadriceps muscle). Hence, these observations typically reflect the acute, fasted response to high-quality liquid forms of isolated protein to skeletal muscle.

Optimising protein nutrition for muscle health can be more complex than simply recommending a daily total protein intake (e.g. source, type, quality, timing, pattern, nutrient co-ingestion). As a logical extension to per meal protein recommendations, the notion that daily protein intakes should be spread evenly between meals/servings (~2–4 h) is intuitive, particularly in older adults that typically consume the majority of their daytime protein intake within a single meal^{(Reference Smeuninx, Greig and Breen84)}. Indeed, a common proposal based on the ‘refractory period’ (or ‘muscle full effect’) of MPS^{(Reference Mitchell, Phillips and Hill39)} and that there is no inactive compartment to serve as a reservoir for protein, is that an even daily protein intake distribution across feeding events is superior to an uneven skewed distribution. However, conflicting findings have been reported from studies in older adults that have measured the response of MPS and lean mass outcomes to the manipulation of protein meal pattern^{(Reference Tieland, Beelen and Laan85–Reference Areta, Burke and Ross89)}, with some indications that meal 1 (i.e. breakfast) is when muscle seems to be the most receptive to protein provision, as during sleep recycled AA are directed towards more critical organs and away from skeletal muscle^{(Reference Tieland, Beelen and Laan85–Reference Areta, Burke and Ross89)}. Accumulating evidence, though, also suggests that bedtime protein feeding may increase overnight MPS rates and enhance skeletal muscle remodelling^{(Reference Trommelen, Holwerda and Kouw90)}. However, given that most of our understanding of MPS responses to protein provision is based on isolated protein sources, particularly in the acute postprandial phase, caution should be applied when translating to longer-term, habitual practices which consist predominantly of wholefoods of varying ‘quality’. Nevertheless, based on current understanding, it is generally accepted that recommended protein intakes for, especially active, older adults should exceed the current RDA and be raised to 1⋅0–1⋅2 g/kg/d based on 3× ~0⋅4 g/kg per meal^{(Reference Baum, Kim and Wolfe91)}. As an extension of advising an increase in protein intake, it is likely that this protein intake will intuitively be spread throughout a day. Further, wholefoods are typically nutrient-dense and better represent habitual dietary patterns than isolated protein sources. Unlike isolated sources, protein-rich wholefoods contain other non-protein-derived nutrients that theoretically may affect the stimulation of MPS (referred to as the 'food matrix'), although this area of research is in its infancy. Nevertheless, the preponderance of data suggests that protein-rich wholefoods do not inhibit the MPS response^{(Reference Burd, Beals and Martinez92)} and, combined with the pragmatism of having to account for ‘other’ nutritional needs, we would therefore recommend that the majority of an individuals' protein intake should be derived from wholefood sources, where possible.

For >20 years there has been suggestions that the RDA for protein may not be adequate for older people to maintain skeletal muscle^{(Reference Campbell, Trappe and Wolfe45)}. Whilst these guidelines markedly exceed the RDA, there is currently no evidence that high(er) protein diets are harmful to health (e.g. kidney, bone) in otherwise healthy individuals^{(Reference Devries, Sithamparapillai and Brimble93–Reference Shams-White, Chung and Du96)}. Numerous studies in older adults support the notion of longer-term higher (than the RDA) protein intakes on lean mass outcomes (e.g. lean body mass, muscle mass, bone health, metabolic health, body composition, strength, function)^{(Reference Houston, Nicklas and Ding17,Reference Børsheim, Bui and Tissier97–Reference Kerstetter, Mitnick and Gundberg103)} . Furthermore, a series of studies have observed no harmful effects on blood lipid profiles, metabolic health, liver or kidney function when prescribing very high (3⋅4–4⋅4 g/kg/d) protein diets for periods of up to 6 months, albeit in resistance-trained individuals^{(Reference Antonio, Ellerbroek and Silver104–Reference Antonio, Ellerbroek and Silver107)}. Notwithstanding, we acknowledge that achieving these high(er) protein intake recommendations can be challenging, particularly for older adults. Indeed, one in three older adults fail to consume even the protein RDA^{(Reference Paddon-Jones and Leidy74)}. This protein undernutrition is exaggerated in frail older adults owing to issues such as reduced appetite, dysphagia, medications and/or psycho-social barriers. Moreover, a low-protein intake is associated with frailty^{(Reference Coelho-Júnior, Rodrigues and Uchida108)}. The consumption of high-quality protein foods and liquids, protein supplementation and/or fortification of foods increases the peripheral availability of dietary AA and thus represents a potentially effective strategy for compromised older populations that warrants further exploration. Indeed, multiple factors can impact the likelihood of malnutrition and our nutritional (and specifically protein) needs and these must inform interventional dietary approaches and dietary protein intake recommendations in older adults^{(Reference Saunders and Smith109)}.

Alternative protein sources for muscle protein synthesis in the 21st century

To date, formal protein recommendations have almost exclusively focused on protein dose with relatively limited consideration to protein source or quality. In contrast, perhaps the most significant evolution in protein recommendations relates to the transition from typically higher-quality animal-based to typically lower-quality plant-based protein sources. This trend is driven, at least in part, by increasing ethical concerns and concerns surrounding the sustainability of animal-based protein production to meet growing global population demands^{(Reference van der Heijden, Monteyne and Stephens110)}. Protein quality is defined by a number of factors, including the AA content (particularly leucine), AA profile and AA bioavailability combined with protein and/or AA needs, and the digestion kinetics and delivery of AA to biological tissues for protein synthesis^{(Reference Morgan, Harris and Marshall111,Reference Wolfe, Rutherfurd and Kim112)} . Historically, animal proteins have been considered to stimulate a greater postprandial MPS response and thus superior for muscle anabolism, largely due to their relative high ‘quality’ (i.e. composition of EAA), high density of protein (i.e. proportion of protein per total weight) and high digestibility. Indeed, early records of protein recommendations refer almost exclusively to animal-based products as ‘highly digestible and good-quality protein’, whilst highlighting the need to consume more foods to reach protein requirements if derived from non-animal-based ‘less protein-dense’ sources. Consistent with this notion, some previous studies suggested that plant proteins were less potent in stimulating MPS compared with animal proteins at an equivalent dose^{(Reference Morgan, Harris and Marshall111)}. This notion was assumed to be attributed to the typically lower EAA content, limited content of a specific AA such as leucine, lower digestibility and/or higher splanchnic extraction of AA of plant proteins^{(Reference van Vliet, Burd and van Loon113,Reference Pinckaers, Trommelen and Snijders114)} . However, these potential issues can be overcome relatively simply via protein extraction, AA fortification, protein blends that exhibit complementary AA profiles and/or simply increasing protein intake to meet AA requirements^{(Reference van Vliet, Burd and van Loon113,Reference Pinckaers, Trommelen and Snijders114)} .

A growing body of research has demonstrated that animal-free protein sources can effectively stimulate MPS in a manner that is comparable to animal-based proteins^{(Reference van Vliet, Burd and van Loon113,Reference Monteyne, Coelho and Murton115–Reference Berrazaga, Micard and Gueugneau118)} , although this observation is likely to be context dependent. Indeed, at least in young ‘anabolically’ sensitive adults, even when a less favourable increase in plasma bioavailability (i.e. lower postprandial plasma AA) have been observed following the ingestion of non-animal compared with animal protein sources, markers of skeletal muscle anabolism are comparable^{(Reference Lanng, Oxfeldt and Pedersen119)}. However, the application of an exclusively plant-based lower-quality protein diet may be concerning if insufficient quantities of protein (and thus EAA) are consumed. This deficiency is exacerbated by the observations of reduced peripheral availability of AA with ageing (via increased splanchnic retention of AA^{(Reference Fujita and Volpi120)}) which likely contributes to age-related muscle loss^{(Reference Fujita and Volpi120)}. Indeed, increased splanchnic retention of AA is also associated with plant-based proteins, due to their lower digestibility^{(Reference Berrazaga, Micard and Gueugneau118,Reference Millward, Fereday and Gibson121,Reference Fouillet, Bos and Gaudichon122)} . It is, though, worthy of note that whilst the impact of insufficient provision of all EAA may be difficult to detect in tightly controlled acute metabolic studies, an accumulation of small AA deficiencies over an extended period of time may be important and result in a greater cumulative MPS deficit, with consequences for skeletal muscle health^{(Reference Brook, Wilkinson and Mitchell123)}, as muscle breakdown, and thus atrophy, will likely need to increase to provide an endogenous supply of EAA for critical physiological tissues and organs^{(Reference Carbone and Pasiakos66,Reference Wolfe124)} . Nevertheless, in practice, human subjects rarely consume foods in isolation and indeed mixed meals within a habitual diet likely contain sufficient amounts of all EAA. Based on current evidence, if protein intake is ≥1⋅6 g/kg/d, the long-term impact of protein source (within a mixed wholefoods diet) on muscle remodelling may be negligible^{(Reference Morgan, Harris and Marshall111)}. Indeed, for most people, the benefits of protein intake and different protein intake strategies seem to diminish greatly beyond about 1⋅6 g/kg/d^{(Reference Stokes, Hector and Morton7,Reference Morgan, Harris and Marshall111)} .

Although largely speculative, it should be considered whether there are metabolic and molecular consequences of switching to an exclusively plant-based lower-quality protein diet in older age, having followed an omnivorous diet throughout the majority of an individual's life. Indeed, individuals habituated to high protein, and thus high EAA, intakes may require a greater relative protein intake to account for an attenuated peripheral dietary AA appearance and/or enhanced AA oxidative capacity^{(Reference Højfeldt, Bülow and Agergaard44)} given that processes involved in the uptake of AA into muscle may be more efficient under scenarios of an impaired muscle anabolic potential^{(Reference Mazzulla, Hodson and Lees125)}. Whilst there is currently limited evidence to support any long-term detriment of a plant-based diet on musculoskeletal outcomes at an advanced stage of life^{(Reference Domić, Grootswagers and van Loon126,Reference Key, Papier and Tong127)} , it is important to note that human subjects possess inherent adaptive biology which provides an evolutionary advantage^{(Reference Fitch128,Reference Gluckman, Low and Buklijas129)} , and raises the question, is nature smarter than people think? Hence, we cannot exclude the possibility that the same cannot be true for longer-term exposure to types of protein source, under conditions of chronic protein ingestion from lower- or higher-quality sources, such that a sudden switch to an exclusively lower-quality, low protein, diet may be detrimental to musculoskeletal health in older age, but warrants further exploration.

Sustainability of different protein sources: a complex debate

Alternative protein sources cannot be discussed without an acknowledgement of and appreciation for environmental sustainability. Much controversy and misinformation surround the sustainability associated with our food choices. Undoubtedly, rapid growth in global population has contributed to stressors in food systems that have clear consequences for the environment and the continued existence of our planet^{(Reference Fanzo, Bellows and Spiker130)}. Indeed, concerns surrounding the sustainability of increased production of animal-based proteins to meet growing consumer demands is driving nutritional research into alternative protein sources (e.g. plant, fungal, algal, insect, laboratory grown ‘meat’, ‘animal-free animal proteins’), which will represent an area of intense research for many years to come^{(Reference van der Heijden, Monteyne and Stephens110)}. A reductionist approach to this issue is to advise a global population switch to excessive plant-based diets^{(Reference Rasmussen, Hall and Vansant131)}; however, the sustainability of different protein (and food) sources is a hugely complex debate for multiple reasons. First, dietary protein sources differ by many characteristics (e.g. AA composition, digestion characteristics, protein density, nutritional composition, form) that justifies the need for assessments of environmental impact to include nutritionally relevant functional units^{(Reference Beal, Gardner and Herrero132–Reference McAuliffe, Takahashi and Beal134)}. Indeed, a recent study suggests that, whilst their analysis revealed animal-source foods still tended to be associated with higher environmental impacts than plant-based foods, shifting to a nutritionally relevant functional unit in life cycle analyses confirms a lower relative environmental impact of nutrient-dense foods compared with when using conventional units (e.g. per total weight, energy), which may be particularly important in older, clinical, and malnourished populations^{(Reference Katz-Rosene, Ortenzi and McAuliffe135,Reference Beal and Ortenzi136)} . Further, when considering ‘ounce equivalents’ of protein food sources, which is a recommendation published by The Dietary Guidelines for Americans to help consumers meet protein requirements with a variety of protein food sources, consumption of ounce equivalents of animal-based protein food sources, such as beef, pork, eggs, result in a greater gain in whole-body net protein balance than the ounce equivalents of plant-based protein food sources, such as tofu, kidney beans, peanut butter, mixed nuts, with further inter-individual variations between protein food sources of various types^{(Reference Park, Church and Schutzler137)}. Therefore, protein source, and by extension quality, is an important consideration in the context of fully understanding the environmental consequences of a given food source, which is likely due to distinct differences in nutrient density (i.e. EAA profiles) and bioavailability of EAA for use by the body.

Secondly, environmental consequences are associated with every stage of the food chain from agricultural production (e.g. farming methods, land use), processing and manufacturing (e.g. packaging, transportation), consumer activities (e.g. storage, cooking) and food waste disposal, and these consequences are not mutually exclusive for protein sources across the spectrum of protein ‘quality’^{(Reference Jeswani, Figueroa-Torres and Azapagic138)}. In addition, lots of produce goes to waste during processing and transportation due to damage, with some forms of produce more vulnerable to damage than others^{(Reference Ishangulyyev, Kim and Lee139)}. According to the FAO, about one-third of all edible produced foods are wasted every year across the entire supply chain, accelerating environmental consequences associated with global food production, highlighting the need for immediate urgent alternative action^{(Reference Ishangulyyev, Kim and Lee139)}. There is growing consensus that food systems need to sustainably provide a diversity of both plant- and animal-source foods, not least for their protein (and more specifically, EAA) content but other vital nutrients^{(Reference Leroy, Beal and Gregorini140,Reference van Vliet, Provenza and Kronberg141)} , to meet global nutritional requirements whilst minimising environmental consequences^{(Reference Beal, Gardner and Herrero132,Reference Leroy, Beal and Gregorini140,Reference Beal142,Reference Leroy, Smith and Adesogan143)} . Accordingly, several early studies have investigated different means to increase the palatability and quality of protein sources that are disposed of during the food production process. For example, blue whiting and nile–tilapia are underutilised fish species containing high-quality protein and, following hydrolysation, have been investigated for their skeletal muscle anabolic properties using marine by-products that have traditionally been disposed of during production^{(Reference Lees, Nolan and Amigo-Benavent144,Reference Shekoohi, Amigo-Benavent and Wesley Peixoto da Fonseca145)} . In addition, the use of other food sources, including insects, have been proposed as an alternative approach to developing high-quality protein with a lower carbon footprint to support skeletal muscle health^{(Reference Vangsoe, Joergensen and Heckmann146,Reference Hermans, Senden and Churchward-Venne147)} . Indeed, the consumption of insects is already common, predominantly in Asia, Africa and South America, and has gained huge interest in recent years as an alternative dietary protein source that may be produced on a more viable and sustainable scale and, as such, may contribute to global sustainability and food security^{(Reference Vangsoe, Joergensen and Heckmann146–Reference Churchward-Venne, Pinckaers and van Loon148)}. Cell- (or lab-) based meat, sometimes referred to as ‘cellular agriculture’, is also receiving increasing attention^{(Reference Wood, Thorrez and Hocquette149,Reference Rubio, Xiang and Kaplan150)} . However, the current energy cost associated with cellular agriculture is significantly higher than more traditional approaches and the feasibility of this concept to support global demand for food has been questioned^{(Reference Wood, Thorrez and Hocquette149,Reference Rubio, Xiang and Kaplan150)} . Undoubtedly, though, some of these approaches do have the potential to maximise sustainability of our food systems to support environmental longevity.

Finally, malnutrition is widespread globally (including protein deficiency^{(Reference Wu, Fanzo and Miller151)}) affecting billions of people, with higher deficiencies in lower-income countries^{(Reference Beal and Ortenzi136)}. Diets in higher-income countries are typically high in nutrient poor ultra-processed foods, whereas lower-income countries diets are dominated by starchy staple (low-protein quality and density) foods that lack diversity, each creating their own unique challenges that likely require a nation-specific approach to sustainability and malnutrition^{(Reference Beal142,Reference Beal, Massiot and Arsenault152)} . Further, there is strong evidence to suggest that specific types of foods, including animal foods, are rich in unique nutrients that can otherwise be challenging to consume in sufficient amounts to promote optimal human health in their absence^{(Reference Beal, Gardner and Herrero132,Reference Beal and Ortenzi136,Reference van Vliet, Provenza and Kronberg141)} . Indeed, in some of the most prominent ‘blue zones’ across the globe (i.e. regions where people live significantly longer than the average, often with an extraordinary number of centenarians), whilst diets are often composed predominantly of plant-based foods, they also consist of varying amounts of animal foods that provide vital nutrients that seemingly contribute to extending longevity and vitality. Though beyond the scope of this review, an important consideration in our food choices for sustainability and malnutrition, as well as whole-body metabolic health and longevity, is also how the food is prepared and the impacts of modern civilisation on food production, regardless of the source. In addition, approaches such as food fortification may also represent important strategies to combat population nutrient deficiencies^{(Reference Ohanenye, Emenike and Mensi153,Reference Kaur, Agarwal and Sabharwal154)} . Undoubtedly, home and/or local produce, land use, food availability, food diversity, less (ultra) processed foods and acknowledging the nutritional value of all foods are all important considerations when addressing food systems in a more holistic manner in line with food demand.

Priority future research directions: where next?

This review has explored some of the most prevalent areas for future research in the field of protein nutrition and put forth some of the key issues and dilemmas that require further research endeavour. Indeed, it is important to recognise the nutritional value of all food types and advocate for foods supported by rigorous, high-quality research that is communicated with policy makers, rather than engaging in polarised public debates. Future research in the field of protein nutrition will likely be dominated by the exploration of novel, alternative, sustainable protein sources that can effectively support skeletal muscle remodelling across the health- and lifespan continuum. Undoubtedly, this new knowledge will encapsulate novel nutrition strategies (e.g. parenteral nutrition, AA fortification) to achieve higher protein (or EAA) intakes in progressively aged and diseased populations. However, as much of our understanding of skeletal muscle anabolic responses to protein are based on isolated liquid-form protein sources, this raises questions over the applicability of current consensuses to habitual practices. Hence, more research is needed into wholefood approaches, including the consumption of ultra-processed foods, that more closely reflect current typical habitual practices. Finally, there is preliminary evidence suggesting that sexual dimorphism to protein provision exists with advancing age. Given the clear gap in female-based research, future work should clarify the sex-specific requirements and recommendations for dietary protein. Undoubtedly, dietary requirements are likely to substantially vary across the globe and indeed across and within clinical populations, and this also must not be ignored when devising future recommendations.