Protein, Dense Bones and Even Denser Dieticians
Opinion is divided concerning the effect of dietary protein on bone health. On the one hand, lots of strong, robust men recommend eating lots of protein (a bear minimum of a gram per pound of bodyweight per day). On the other hand, lots of mainstream nutritionists and dieticians tell the general public that a gram per kilogram per day (1kg equals 2.2lbs) is perfectly sufficient. Calcium, the most important component of robust bones, tends to be excreted in the urine at an increased rate when extra protein is consumed, and the latter group confidently assures the public that the aforesaid strong, robust men will inevitably metamorphose into fragile weaklings once they get old.

Who's right? When I began writing this article, I had little doubt about the answer. The explanation was simple: (a) dietary protein increases the amount of calcium in the urine because it initially allows more calcium to be absorbed in the gut; (b) low-protein diets minimise the amount of calcium absorbed in the gut and therefore force the body to hold on to as much of its existing store of calcium as possible (hence less excreted in urine). Correct or not, this illustrates an important point that must be borne in mind: calcium (or any other nutrient) has to be absorbed in the gut before it can be excreted in the urine (unless it is given intravenously), and urinary excretion of calcium (or any other nutrient) could reflect increased absorption, the degradation of existing stores, or some combination of the two.

My simple explanation encountered several challenges when I started reading through early feeding experiments examining the relationship between protein, calcium balance and bone health. I only concerned myself with experiments in which the subjects were being fed on-site or provided with all of their food pre-packaged (thus minimising the potential for interference by other dietary factors), but the oft-observed effect of higher-protein intakes increasing urinary calcium output was not being adequately countered by evidence of greater absorption in the gut. Higher protein intakes were resulting in markedly more negative calcium balances in many cases.

I needed to dig deeper. I started looking at the sources of the extra protein, the effect of particular amino acids from those protein sources, the amount of different macro-minerals in the diets, and the whole-package effect of different protein-rich foods. My eventual explanation was not too far from my initial one, and permitted me to retain the provocative provisional title, but the end verdict is less dogmatic, more specific and more laden with caveats than its embryo.

The vast majority of the early experiments, along with a couple of later ones, examined the effect of purified proteins. This is informative in that it isolates the effect of the particular proteins, but it would be erroneous to assume that consuming those proteins via their natural context in protein-rich foods necessarily leads to the same outcome. What was that outcome? Here are the relevant experiments (full-text pdf files are available for almost all of them):


?(1) Higher-protein (141g per day) negatively affected calcium balance (judged from fecal and urine samples), and different magnesium intakes had no effect, in males aged 18-20 eating 1400mg per day of calcium and phosphorus 
?(2) 142g and 95g per day negatively affected calcium balance (judged from fecal and urine samples) in a similar group of males eating 800mg and 1000mg per day of calcium and phosphorus 
?(3) Similar diets with similar subjects to the above, with lower quantities of the minerals, produced negative calcium balances across the board that were more negative as protein intake went higher 
?(4) Higher-protein (93g per day) seemed to increase gut absorption more than urine excretion, and extra magnesium also seemed to improve balance (though this was downplayed), in teenaged males aged 13-14 
?(5) Various protein intakes with various subjects, always with equal intakes of calcium and phosphorus, always produced more negative calcium balances (judged from fecal and urine samples) on higher-protein intakes 
?(6) 142g of protein per day negatively affected calcium balance (judged from fecal and urine samples) compared to 47g per day, and parathyroid hormone (PTH) was unchanged, in males in their 20s eating 515mg and 1110mg of calcium and phosphorus per day
?(7) Calcium balance (judged from fecal and urine samples) was even more negative on 225g of protein per day than on 75g per day (despite high intakes on both), and PTH and hydroxyproline (a marker of bone collagen turnover) were unaltered, in males aged 23-30 who ate more phosphorus on the lower-protein diet
?(8) Calcium balance (with equal mineral intakes and judged from fecal and urine samples) was negatively affected on 112g compared to 47g of protein per day (despite apparent increased gut absorption in nearly half the higher-protein subjects), and increased urinary calcium correlated with increased glomerular filtration rate and decreased fractional renal tubular reabsorption of calcium, in males aged 44-86 and post-menopausal women
?(9, abstract) 2g of protein per kg per day negatively affected calcium balance compared to 0.8g per kg per day in osteoporotic subjects
?(10) Calcium balance (judged from fecal and urine samples) was even worse on 123g rather than 46g per day in females in their 20s consuming 500mg and 900mg of calcium and phosphorus per day
?(11) Adding the sulphuric amino acids (methionine and cysteine, SAAs) to a lower-protein diet increased urinary calcium to 50% of the extent seen on a higher-protein diet, and added phosphorus antagonised the effect of the SAAs, in males aged 18-26
?(12) Manipulating phosphorus intake (1010mg versus 2525mg per day) on diets high (150g per day) or low (50g per day) in protein left calcium balance (from 500mg per day) unaffected except for a worse balance (judged from fecal and urine samples) on the higher-protein, lower-phosphorus diet in males aged 19-25
?(13) Consuming 50g or 110g of protein per day, with 713mg and 1078mg of calcium and phosphorus, had no consistent effect on calcium balance (judged from fecal and urine samples) in post-menopausal women (half of them osteoporotic)
?(14) Increasing daily protein from 65g to 94g via spray-dried egg white did not affect calcium balance (judged from fecal and urine samples) in males aged 19-64 eating 650mg and 1000mg of calcium and phosphorus per day
?(15) Increasing protein from 16% to 27% of the diet via gluten added to bread in replacement of starch, with the diets otherwise being identical, significantly increased urinary calcium and non-significantly increased gut absorption (the "non-significant" increase was greater in absolute terms than the "significant" one, and thus resulted in an also "non-significant" trend towards improved calcium balance, but that's statistics for you) in post-menopausal women and similarly aged men (who had previously had the misfortune to be subjected to an NCEP Step 2 diet)
?(16) Adding 40-50g of protein to a daily base of 40g, via whey-casein isolate from milk or soy protein isolate (SPI) with or without isoflavones, had no overall effect on calcium balance (judged from intravenous calcium radioisotopes as well as fecal and urine samples) in post-menopausal women

In almost all of the above experiments, higher protein intakes negatively affected calcium balance (judged from fecal and urine samples) compared to lower protein intakes with the same quantities of calcium and phosphorus. In a few cases they left it unaltered, and in no case did they clearly improve it. However, there was a hint of improved calcium balance on a higher protein intake in experiment (4), the subjects of which were young teenagers - an age group that was not featured in any of the other experiments. This may be significant.

Other factors that emerge from the above are that the sulphur-containing amino acids (methionine and cysteine) are partly (but not entirely) responsible for increasing urinary calcium excretion on higher-protein diets (11), that phosphorus reduces urinary calcium excretion (7, 11, 12), and that calcium balance was often negative on higher and lower protein intakes despite high intakes of the mineral. The last factor brings into question the general quality of the experimental diets. The purified proteins most commonly fed were gluten, casein (from milk), lactalbumin from whey (from milk) and egg white albumin. Sometimes soy protein isolate (lower in the sulphuric amino acids) was given, with no noticeable different effect on overall balance. In fact, judging by the above, it seems that the purified proteins from animal sources have a very small positive effect on calcium absorption from the gut, which is cancelled out (or completely overturned) by increased urinary excretion. Soy protein isolate, lower in the sulphuric amino acids, has less effect on urinary calcium excretion but had a worse effect on gut calcium absorption in study (16). In study (7), in which calcium balance was markedly negative on both diets despite a high calcium intake, the fat in the diets was provided almost entirely by corn oil and margarine.

The above experiments are artificial because they do not gauge the effect of adding protein to the diet from natural food sources, all of which happen to be high in phosphorus (and potassium and, in some cases, calcium). What happens when protein intake is increased via food?

The earliest such experiment (17) appeared to support the findings of the ones using purified proteins. A diet providing 150g of protein via low-fat ground beef resulted in an even more negative calcium balance than one providing 75g of protein - and even than one containing barely any protein at all! All balances were very negative, but they were less so on the diet providing barely any protein. The catch is that the subjects (males aged 22-32) were being fed a measly 100mg of calcium per day. The source of the protein on the 75g and 150g diets was low-fat ground beef, which is very low in calcium and therefore ill-equipped to reverse calcium deficiency. When the subjects were later given calcium supplements along with 75g of protein per day, calcium balance became positive in most of them. The researchers did not examine the effect of 150g of protein with calcium supplements.

The next experiment using protein-rich food (red meat from beef) was also (18) the first to administer oral and intravenous calcium tracers. Unlike intravenous calcium tracers, oral calcium tracers need to be absorbed in the gut before they can be excreted in the urine, and the proportions of the two excreted will therefore give a good indication of how much calcium has been absorbed in the gut and of how much of that in the urine is derived from endogenous (internal) sources (i.e. bone). This experiment also involved an initial low-calcium intake (200mg per day), and those eating 0.5, 1 and 2g of protein per kg of bodyweight per day (from 50g, 200g and 300g of beef) were all in negative calcium balance thereon, with no differences between groups, but went into positive balance (again with no differences between groups) on higher intakes ranging from 800mg to 2000mg per day. Phosphorus intake was higher as protein intake increased, this being due to the high phosphorus content of the meat. It was a controlled experiment not of protein intake but of the whole-package effect of adding a protein-rich food (beef) to the diet. The subjects were males aged 40-67.

Another experiment also (19, pdf) gauged the effect of meat (beef and turkey), but this also gauged the effect of cereal bran on colonic function. The subjects (males in their early 20s) were generally in mildly positive calcium balance (from a high intake of close to 1000mg per day) on the diets low (63g per day) and high (136g per day) in protein, although the high-protein diet increased absorption of calcium from the gut (and excretion via urine). Then they were switched a high-fibre, high-phytate diet that was also higher in protein (164g per day) and calcium (1302mg per day) due to the protein and calcium present in the added foods (calcium-fortified whole-wheat biscuits, wholemeal bread, bran and bran crispbread). The result of following this "healthy" high-fibre, high-phytate diet was that calcium balance went down the toilet - quite literally!

The effect of beef and chicken was gauged in another experiment, but this one (20) replicated the conditions of those using purified proteins by equalising intakes of calcium (600-800mg per day) and phosphorus (1400-1500mg per day). The higher protein intake (1.5-2g per kg per day) increased urinary calcium excretion compared to the lower intake (0.5-0.8g per kg per day), but total calcium balance wasn't even estimated. Blood samples were taken, and these showed that levels of parathyroid hormone (PTH) - which usually increase when calcium status is poor - significantly decreased on the higher protein intake (especially in two hyperparathyroid subjects). These results - less urinary calcium but elevated PTH on a low-protein diet - are consistent with studies (26) and (30), which also assessed calcium absorption and observed it to be greater on the higher protein intake.

A later experiment which also equalised calcium and phosphorus intakes on low-protein (0.6g per kg per day) and higher-protein (2g per kg per day) diets found no differences in urinary calcium excretion (21, abstract) when the extra protein was provided by an even mixture of high-SAA animal and low-SAA plant foods. Gut absorption of calcium was not assessed. The subjects were males aged 66-88.

Meat is high in phosphorus but not calcium; milk products are high in both. It stands to reason, therefore, that increasing protein intake via milk (or yogurt or cheese) would have a positive impact on calcium balance rather than the neutral one seen with meat. This was supported by an experiment (on males aged 19-26) using (22) mixtures of purified proteins, protein-rich foods and mineral supplements to compare the effects of meat (ground beef), combined meat and milk products (cottage cheese and semi-skimmed milk), a simulated high-meat diet and a simulated meat- and milk-enriched diet. Also compared were low-protein diets (55g per day) enriched (1660mg per day) or not enriched (890mg per day) with phosphorus to match the intake on the high-protein-from-meat diet (146g of protein per day). Calcium balance was close to neutral on, and not significantly different between, the high-protein-from-meat and the low-protein diet, but phosphorus balance was significantly different, being positive on high-protein-from-meat and negative on low-protein. The low-protein-plus-phosphorus diet reduced urinary calcium excretion but also reduced gut absorption of calcium and left overall balance the same as the other two diets, and it produced a phosphorus balance as negative as that on the low-protein diet. The high-protein-from-meat-and-milk diet (146g of protein, 1370mg of calcium and 2060mg of phosphorus per day) resulted in positive calcium and phosphorus balances that were significantly higher than those on the other food-based diets. Casein, lactalbumin, dried egg white and gluten were added along with calcium and phosphorus supplements to simulate the high-protein diets enriched with meat or meat and milk, and these produced calcium balances matching those of the diets simulated. Thus, consuming purified proteins does not produce negative calcium balances in the presence of adequate quantities of calcium. However, phosphorus balances were significantly worse on the diets using purified proteins, as they had been on the low-protein diet enriched with phosphorus - effects that may have been due to the quality of the phosphorus supplement (although the balances were negative due to urinary excretion rather than absorptive failure).

Another experiment (on males aged 30-64) compared a naturally higher-protein (2g per kg per day), higher-phosphorus diet via beef to a low-protein diet (1g per kg per day) and found that (23) eating around 500g of red meat per day had no effect on calcium balance. The calcium intake was close to 1000mg per day on both diets. In this instance neither gut absorption nor urinary excretion of calcium was affected by the high protein intake. It is possible that phosphorus both opposes calcium absorption (they have to compete for it) and spares it from waste once absorbed.

Phosphorus is not the only other mineral that affects calcium retention. This abstract suggests that (24, abstract) potassium spares calcium on a diet enriched with sulphuric amino acids. The subjects (young Japanese women) were fed 50g and then 100g of protein per day, the higher amount being made up by meat (high-phosphorus but high-SAA food) or soy protein isolate (low-SAA purified protein). The meat-enriched diet increased calcium absorption and urinary excretion compared to the SPI-enriched one, with balance being the same. Adding apple to the meat-enriched diet reversed both of the trends and left the balance the same. Adding the SAAs to the SPI-enriched diet had a negative impact on calcium balance, which was restored by adding potassium.

When purified proteins were added to protein-rich foods (casein in low-fat milk and gluten in bread, along with soyabean added to minced meat), on higher-protein diets (21% of calories) that contained a little more calcium and phosphorus than the low-protein alternative (12% of calories), (25) urinary calcium increased slightly with more protein but calcium balance was marginally significantly better, though none of the differences were sexily different from neutral. The subjects were both men and women, and both in their 20s or 30s and their 60s or 70s.

The previous naturally protein-and-phosphorus-enriched-via-meat diets had a neutral effect on calcium balance, but the subjects were always males. A later experiment was carried out on pre-menopausal women aged 21-39, who ate 0.7g or 2.1g of protein per kg per day following a baseline period of 1g per kg per day. The extra protein and phosphorus were provided mainly by fish, poultry and undried egg whites, but the higher phosphorus content of the higher-protein diet was countered by a higher potassium content in the low-protein diet. The other minerals were equal between the diets. Fibre and phytate were higher in the main low-protein diet, but another low-protein group was fed the lower quantity of these for comparison. Oral and intravenous calcium tracers were administered to help gauge absorption. The result was that (26) eating the so-called "recommended daily allowance/amount" (RDA) of protein induced hyperparathyroidism within 4 days whereas thrice the RDA increased calcium absorption. Hyperparathyroidism was equally quick to present itself on the low-protein diet containing the lower amount of fibre and phytate.

In an experiment on post-menopausal women aged 50-75 eating 12% or 20% of their calories as protein (with calcium intakes equal and phosphorus, potassium and sodium naturally higher on the higher-protein diet due to their presence in beef, pork and poultry), (27) calcium retention (judged via oral tracers) was at least unaffected by the higher protein intake, and there was a "trend" (of borderline significance - see table 2) for it to be greater. 

Conversely, another experiment on pre-menopausal women found that (28) reducing protein intake to the RDA from a "high" (i.e. borderline deficient) baseline intake of 1.1g per kg per day resulted in less urinary calcium. It wasn't that different from study (26), however. Parathyroid hormone (PTH) increased on the even-lower-protein diet, as it had in study (26), although in this case it was deemed to be non-significant. Urinary calcium had also been elevated on the higher-protein diet in study (26), but calcium absorption had been greater. Study (28) did not assess calcium absorption. The even-lower protein intake was achieved by reducing intakes of chicken "and meat" (I thought that chicken was a type of meat). Phosphorus, potassium and sodium were added to the even-lower-protein diet to compensate for their depletion via the removal of meat, making this experiment similar to study (20) - on which PTH was also higher on the lower protein intake (again despite less urinary calcium excretion). The one interesting thing that study (28) did do was equalise intakes of the SAAs on the even-lower-protein diet. Study (11) added the SAAs to a low-protein diet and found that they accounted for about half of the increase in urinary calcium on a higher-protein diet containing purified proteins (an effect that was countered by also adding phosphorus to the SAA-enriched low-protein diet). Study (24) found that potassium also countered the effect of the SAAs on urinary calcium. The baseline diet in study (28) was slightly higher in protein thanks to phosphorus-rich meat rather than purified proteins, but phosphorus was added to the even-lower-protein diet. Adding the SAAs to the phosphorus-and-potassium-enriched even-lower-protein diet in study (28) therefore matched the conditions of studies (11) and (24), with similar results. Protein-rich animal-source foods clearly do increase urinary calcium through another mechanism in addition to the SAAs, but the best candidate for that mechanism is increased absorption, which wasn't assessed in study (28).

Study (24) had compared higher-protein diets enriched with meat or soy protein isolate (SPI). SPI does contain some calcium and phosphorus, but it also contains phytate that hinders mineral absorption. It contains less of the SAAs and reduced urinary calcium but also absorption in study (24). A different experiment on post-menopausal women compared two otherwise near-identical 15%-protein diets that were enriched with either meat or SPI. Sodium and potassium were marginally higher on the meat-enriched diet; calcium was slightly higher on the SPI-enriched diet, as was phytate, and the SAAs were lower. Absorption and retention of calcium were assessed via oral tracers, and (29) no noteworthy differences were observed. 

Similar subjects (pre-menopausal women with a few post-menopausal ones) and diets to those used in study (26) were used again in a later experiment comparing 1g of protein per kg per day to 2.1g per kg per day. Mineral intakes were the same except for a slightly higher phosphorus intake on the higher-protein diet due to the phosphorus in the protein-rich foods (fish, poultry and undried egg whites). Every effort was made to equalise phosphorus intakes by selecting higher- and lower-phosphorus foods to make up the rest of the lower- and higher-protein diets. Oral and intravenous calcium tracers were administered, and the result was that (30) urinary calcium increased on the higher-protein diet but absorption increased even more and less of the urinary calcium was derived from bone! Quite a bit was derived from bone, because calcium balance was negative in both groups, and more so in the low-protein one.

Taking all of the above into account, it is clear that natural protein-rich foods do not negatively impact calcium balance (and sometimes positively impact it) compared to the alternatives, that low-protein intakes (i.e. the "daily recommended allowance") are themselves a cause for concern, and that a high absorbable intake of calcium and other minerals is the surest way of keeping calcium balance positive. Rather than writing a detailed concluding paragraph, I shall summarise the main points with bullets.

Summary of the Main Points:

?Methionine and cysteine, two of the amino acids found in protein, are sulphur-containing amino acids (SAAs) and increase the amount of calcium excreted in the urine
?Phosphorus and potassium spare calcium from SAA-induced urinary excretion
?Animal sources of protein contain higher quantities of the SAAs but are also high in phosphorus and potassium, and milk products are additionally high in calcium
?If intakes of minerals and all other nutrients are equalised, diets higher in protein due to the addition of purified SAA-rich proteins might increase calcium absorption but (at least in regular subjects not regularly partaking in vigorous physical training) will increase urinary calcium excretion so that calcium balance is the same as or worse than on a lower-protein diet - with the possible exception of when they are given to young, growing teenagers
?If diets higher in protein due to the addition of purified SAA-rich proteins are also higher in phosphorus and/or calcium, thereby mimicking diets higher in protein due to the addition of meat and/or milk products, calcium absorption should match or exceed excretion, but it is not clear how phosphorus balance will be affected
?Gluten, the protein from grains such as wheat, barley and rye, was also used in many of the experiments adding purified proteins, but its independent effect on calcium balance was not tested except in study (15) - when it was substituted for starch
?If intakes of minerals and all other nutrients are equalised, diets higher in protein due to the addition of low-SAA, high-phytate part-purified plant proteins (e.g. soy protein isolate) will (compared to animal proteins) decrease calcium absorption (due to the phytate) but will not increase excretion and will leave calcium balance about the same
?Protein per se appears to increase calcium absorption, but phytate (occurring alongside plant proteins) opposes absorption and the SAAs (higher in animal protein) increase urinary excretion, whereas phosphorus and potassium (naturally present alongside animal and plant proteins, although their absorption can be hindered by phytate alongside the latter) spare calcium from excretion 
?From a practical standpoint, the whole-product effect of adding high-SAA phytate-free phosphorus- and potassium-rich red meat, poultry, fish and/or eggs to the diet is neutral on calcium balance - and might be slightly positive in women
?The whole-product effect of red meat, poultry, fish and/or eggs is positive on phosphorus balance
?From a practical standpoint, the whole-product effect of adding low-SAA mineral- and phytate-rich protein-rich plant foods to the diet is neutral on calcium balance
?From a practical standpoint, the whole-product effect of adding high-SAA phytate-free phosphorus-, potassium- and calcium-rich milk products (milk, cheese, yogurt) to the diet is positive on calcium and phosphorus balance
?Compared to diets containing more than twice as much protein but only a little more phosphorus, the so-called RDA of protein reduces calcium absorption, induces hyperparathyroidism and increases the amount of urinary calcium that is derived from bone rather than the diet

Written: Jan-May 2009metastatic calcification