Dietary protein and bone health: towards a synthesised view

The present paper reviews published literature on the relationship between dietary protein and bone health. It will include arguments both for and against the anabolic and catabolic effects of dietary protein on bone health. Adequate protein intake provides the amino acids used in building and maintaining bone tissue, as well as stimulating the action of insulin-like growth factor 1, which in turn promotes bone growth and increases calcium absorption. However, the metabolism of dietary sulphur amino acids, mainly from animal protein, can lead to increased physiological acidity, which may be detrimental for bone health in the long term. Similarly, cereal foods contain dietary phytate, which in turn contains phosphate. It is known that phosphate consumption can also lead to increased physiological acidity. Therefore, cereal products may produce as much acid as do animal proteins that contain sulphur amino acids. The overall effect of dietary protein on physiological acidity, and its consequent impact on bone health, is extremely complex and somewhat controversial. The consensus is now moving towards a synthesised approach. Particularly, how anabolic and catabolic mechanisms interact; as well as how the context of the whole diet and the type of protein consumed is important.

Dietary protein is crucial for the maintenance of bone tissue as well as for bone growth. Bone is 35 % protein and requires a supply of amino acids to be used for protein turnover. The mechanostat is a process whereby bone remodels itself in response to elastic deformation acting on it, one large provider of this force is muscle mass (11) . This explains observations that higher muscle mass is associated with increased bone mass (12) . Adequate protein ensures an adequate muscle mass, so is an important determinant of bone health. The current internationally recommended protein intake guideline for adults of all ages is 0⋅83 g/kg/d (13) although values of 0⋅8 g/d are used by many agencies with some recommending higher values for the elderly (14)(15)(16)(17)(18) . Requirements for infants, children and adolescents vary by country, but are higher than that in adults due to the need for growth. For example, in the UK, the recommended nutrient intake is 12⋅5 g/d in those aged 0-3 months (17) which translates to over 2 g/kg/d. Western populations are generally dietary protein sufficient. For example, in the UK National Diet and Nutrition Survey (2014-2016, aged 19-64 years), the median protein intake was 74 g/d. Based on a UK average body weight of 72 kg for women and 85 kg for men (18) , this suggests a median intake of over 1 g/kg/d.
There are some groups in western societies, such as frail older people, who are at risk of a low-protein intake. For example, in one study, 32 % of frail older people did not meet the 0⋅8 g/kg/d requirement (19) . Conversely, in another study older care home residents had sufficient protein intake, with 95 % attaining 0⋅8-1 g/kg/d (20) , and an analysis of the UK National Diet and Nutrition Survey of protein intakes of the elderly after trimming for under-reporting indicated median intakes of 1⋅24 g/kg/d with a negligible prevalence of deficiency (21) . Low-protein intakes are important due to the association of low-protein intakes and frailty in older people (22) .
Also, protein-energy malnutrition is still very common throughout the developing world. For example, 22⋅2 % of children aged 0-59 months globally have stunted growth and 7⋅5 % of children have wasting (23) , although actual protein deficiency per se is rare with growth deficits more likely to reflect enteric infections from a poor environment (24) .
The present paper will now discuss the proposed anabolic and catabolic actions of protein on bone health. It will exclude discussion of weight-loss studies as protein metabolism may differ in this situation.

Anabolic associations of dietary protein with bone health
Protein intake stimulates the release of the hormone IGF-1 (25) , which increases muscle mass (25) and bone growth (26) . Accordingly, lower protein intake leads to lower IGF-1 (25,27) which in turn leads to a lower bone mass (25)(26)(27)(28) . This could result in a higher fracture risk, with studies finding a negative association between IGF-1 concentration and predicted fracture risk (25,29,30) . Correcting low-protein intake theoretically leads to a variety of musculoskeletal health benefits in older individuals ( Fig. 3) (25) .
Observational studies have shown a beneficial association between a higher protein intake and improved bone health. For example, in children and adolescents, cross-sectional analyses have associated a higher protein intake with a higher bone mineral content (BMC) (31,32) and a larger bone area (32,33) . In longitudinal research, studying children with high physical activity levels, a higher protein intake was associated with an increase in femoral neck bone mineral density (BMD) z score A. L. Darling et al. 166 between age 7 and 15 years (34) . However, lower protein intake was associated with a reduction in femoral neck BMD z score during the same time period (34) . In older adults (over 60 years), higher protein intake has been associated, in cross-sectional studies, with higher spinal BMD (35,36) , total body BMD (36) and femoral neck BMD in women (37) . Higher protein intake has also been associated with higher total hip BMD in men and women (37,38) . Conversely, studies have found no difference in protein intake between women with normal BMD and women with osteopoenia or osteoporosis (39) , and no association between protein intake and spinal or femoral neck BMD in older women (40) .
In premenopausal women, some studies have found that increased protein intake is associated with higher hip or spine BMD (41)(42)(43) or BMC (41,44) . However, other studies have found no association with radial, spinal or femoral neck BMC (45) or lumbar spine or femoral neck BMD (42,43,45,46) . The few studies assessing younger to middle-aged men have found a positive association between protein intake and BMD in black men (47) and vertebral BMC in all men (48) . However, studies have also found no association between protein intake and BMD in white men (47) and no association for all men for total hip and spine BMD (49) or radial BMC (48) . However, it must be borne in mind that not all observational analyses are multivariate adjusted. Some associations between dietary protein and bone health will be due to confounding from dietary, lifestyle and demographic factors. The type of protein consumed, and the adequacy of calcium intake may also vary between studies. These factors could explain differing results.
Protein supplementation studies have shown an improvement in BMD, BMC or other indices or bone size or strength in some studies but not others. For example, one study found improved bone growth after protein supplementation in malnourished children (50) . However, there have been no trials to date in nonmalnourished children. In terms of older people, in a study of hospitalised adults with a hip fracture, there was a reduced femoral shaft bone loss in those supplemented with 20 g/d protein (51) . Similarly, a study of older patients' post-hip fracture found that 20 g/d protein supplementation was associated with reduced proximal femur bone loss (52) . However, a study in community-dwelling adults aged 70-80 years, found no effect of whey protein supplementation (30 g/d) on bone mass or strength (53) . Therefore, benefits of supplemental protein on bone may be confined to frailer older people post-hip fracture.
In terms of bone markers, over all age groups, evidence from trials is also mixed. Some studies have found no difference in bone markers in participants allocated to high-or low-protein diets (54) or participants allocated to a protein supplement compared to placebo (55) . However, some studies have found lower bone resorption in those supplemented with protein (56,57) .

Catabolic associations of dietary protein with bone health
To maintain life, extracellular fluid must strictly stay within the limits of pH 7⋅35-7⋅45 (hydrogen ions between 0⋅035 and 0⋅045 mEq). Each day, human subjects on a typical western diet produce 1 mEq/kg body weight (58) . This increased physiological acidity leads to a series of physiological responses to neutralise the acid ( Fig. 4) (59) . The body instigates buffering of body fluids, including increased bicarbonate production. The lungs increase carbon dioxide loss, the kidneys excrete more acid and bone loses alkaline mineral into the body fluids (59) . The latter is achieved via increased activity of osteoclast cells (60) , which break down and remodel bone tissue. There is also evidence for a direct dissolution of bone calcium carbonate under exposure to acidity (61) . Studies of acidic states such as ammonium chloride ingestion (62) and starvation (63) have demonstrated a negative calcium balance and increased calciuria (59,64) . This negative calcium balance could have a negative impact on bone health if it occurs over the long term.
Dietary composition influences the acid-base status of the body. The consumption of sulphur amino acids from animal protein increases physiological acidity, as does phosphate from dietary phytates in grains. This means some cereal proteins produce as much, or more physiological acidity than animal proteins. For example, oatmeal, walnuts and whole wheat are higher producers of acidity than are chicken, beef and cheddar (65) . Reprinted from Heaney (8) . Copyright (2020), with permission from Elsevier. https://www.sciencedirect.com/science/article/pii/S8756328203002369?via %3Dihub Dietary protein and bone health 167 Consumption of green vegetables and fruit leads to increased alkalinity. This is because they contain alkaline potassium salts of weak organic acids such as citrate, lactate and malate. A higher protein:potassium ratio is undesirable, as demonstrated by the finding that it is associated with increased higher renal net acid excretion (66) . A higher protein:potassium ratio is associated with higher potential renal acid load (66) . Therefore, high protein, without adequate protective potassium, will increase physiological acidity. The net endogenous acid production in modern western diets could have negative implications for bone health, if the acidity is large enough and for long enough. An analysis of the net endogenous acid production of modern and preagricultural diets found that modern diets had an average of +48 mEq/d compared with −88 mEq/d for the preagricultural diets (67) . Therefore, today we consume more acidic diets than was previously the case.
In terms of epidemiology, some ecological studies in the 1990s have suggested that higher protein intakes are associated with a detriment to bone health. For example, two studies found a positive association between animal protein intake per capita and hip fracture incidence (68,69) . However, ecological studies are prone to bias due to the methodology used. Moreover, few, if any, crosssectional, cohort studies or randomised controlled trials have found an association between higher protein intake and poorer indices of bone health.
It is known that calcium excretion may rise with increased protein intake suggesting a detriment to bone mass. However, evidence shows that calcium absorption may increase, offsetting calcium loss. One study, using a within-subjects study design, gave research participants a low-protein diet (0⋅7 g/kg/d) and a high-protein diet (2⋅1 g/kg/d). They found increased urinary calcium during the high-protein diet, but calcium absorption also increased (70) . However, another intervention trial showed no difference in calcium absorption, urinary calcium excretion or level of bone resorption markers when consuming the RDA of protein compared with consuming three times the RDA (71) . This suggests no detrimental effect of higher protein intake on calcium metabolism and bone markers. However, this was only a short-term trial in only a small sample size, and it is unclear what the effect would be on bone metabolism in the long term.
Baseline calcium intake may also be important. For example, in the Framingham study, the increased fracture risk associated with higher animal protein intake was only present in the participants with lower calcium intake (<800 mg/d) (72) . There was no association between higher animal protein intake and fracture risk when calcium intake was sufficient (≥800 mg/d) (72) . This suggests adequate calcium intake may offset any detrimental effects of a high animal protein diet.

Systematic reviews and meta-analyses on protein intake and bone
There are conflicting findings from systematic reviews and meta-analyses on dietary protein and bone health. Meta-analyses of protein supplementation have found either no overall effect (73) or a tiny beneficial effect (74,75) on bone health, with no evidence of a detrimental effect in any of the systematic review and meta-analyses published to date. Meta-analyses of cross-sectional studies assessing the relationship between dietary protein and bone health generally show a positive association (73,74) , although the association is often not present when analysing only multivariate-adjusted studies (73) . Meta-analyses of cohort studies have found either a beneficial association with fracture risk (76,77) or no association with fracture risk (73,74) . Therefore, any small gains in BMD may not translate into fracture risk in the long term (73) . The association between protein intake and bone health in observational studies is stronger in case-control studies compared with cohort studies (73) . This could be due to case-control studies having significant inherent bias (78) . Overall, the message across these meta-analyses is that there is no evidence of a detrimental association between protein intake and bone health. As evidenced earlier, some meta-analyses suggest a benefit of protein to bone health, but others suggest no association.

Towards a synthesised view of dietary protein and bone health
There have been recent efforts to synthesise the anabolic and catabolic mechanisms of dietary protein on bone health. A key review (79) discusses how the positive aspects of dietary protein intake, including increased A. L. Darling et al. 168 calcium absorption and IGF-1 induced bone formation, work in tandem with the negative effects. Particularly, they discuss how protein may benefit bone health if consumed as part of a diet containing enough dietary calcium, and alkalising fruit and vegetables (79) .
This synthesised approach may explain some complex findings of research studies. For example, in one study, higher dietary protein was associated with larger bone size (periosteal circumference and cortical area), and higher BMC and polar strength strain index (80) . However, children in the same study with a high dietary potential renal acid load had a lower BMC and cortical area than those with a lower dietary potential renal acid load (80) .
A low protein:potassium ratio is likely to be ensured by consuming a balanced diet. Indeed, there is an argument for a whole diet approach for bone health (65) , which includes a balanced intake of nutrients such as protein, potassium, calcium and phosphate. As discussed earlier, one way of increasing potassium intake is to consume more fruit and vegetables. Adequate calcium intake may also help compensate for any sulphur amino acidinduced bone loss (81) . Adequate protein intake ensures enough amino acids for growth and repair of body tissues but should not be in excess. Other food constituents such as soya isoflavones and caffeine may also have potential effects on bone health (65) . Soya isoflavones are known to have oestrogen-like effects on the body. Therefore, theoretically they may have beneficial effects on bone. Some studies have found a benefit of soya isoflavone supplementation on BMD (82,83) , but most studies have found no benefit (84)(85)(86) . Higher caffeine intake has been associated with poorer bone health (87) , which could be due to a small caffeine-induced reduction in calcium absorption (88) . However, this could also be due to consumption of caffeinated beverages being higher in individuals who have low calcium intakes (89) .

Conclusion
There is a long-standing debate as to whether high dietary protein intakes are beneficial or detrimental for bone health. We know that adequate dietary protein intake is essential to provide amino acids for building and maintaining bone tissue. It also has anabolic effects on bone by stimulating the release of IGF-1 and calcium absorption from the gut. However, some forms of dietary proteins may increase net physiological acidity because of their sulphur amino acid or phytate content. This could lead to increased bone loss in the long term in order to provide a source of alkaline mineral.
Research over the past 40 years has supported both anabolic and catabolic associations between protein intake and bone health. Data from cross-sectional studies support a positive association. However, cohort studies assessing fracture risk show both positive and negative associations, leading to null associations in metaanalyses. Intervention studies assessing BMD show no effect (or a tiny benefit) of protein intake for bone health in adults. There is a lack of research on this topic assessing children and adolescents, as well as adults with very low or very high intakes of dietary protein.
To make sense of the opposing effects of dietary protein on bone we are moving towards a synthesised view whereby dietary protein has both anabolic and catabolic effects on bone. The overall effect depends on the whole diet, as food components modify the net physiological pH. For example, calcium-containing foods, or the consumption of fruit and vegetables, may contribute to reduced physiological acidity from a higher protein diet.