The Journal of Nutrition Vol. 128 No. 6 June 1998, pp. 1054-1057

Excess Dietary Protein May Not Adversely Affect Bone^1,2,3

Robert P. Heaney

Creighton University, Omaha, NE 68178

	ABSTRACT

Abstract References

Too little protein is always harmful for the skeleton. Increasing dietary protein increases endogenous calcium excretion. The ability to adapt depends upon the adequacy of an individual's calcium intake. At a population level, the effect of protein is often minimized because calcium intake rises with increasing protein intake. A dietary calcium-to-protein ratio 20:1 (mg:g) probably provides adequate protection for the skeleton. Excess protein will not harm the skeleton if the calcium intake is adequate.

Dietary protein affects bone in a variety of ways. Approximately one third of the mass of bone is protein, and, as such, bone is one of the most protein-dense tissues of the body. Dietary protein, with its content of essential amino acids, is necessary for new bone matrix synthesis. Bone growth is stunted in protein-energy malnutrition, and outcomes of hip fracture are dramatically improved with protein supplements in the typical elderly victim of osteoporotic fractures (Bastow et al. 1983, Delmi et al. 1990). Additionally, dietary protein comes in the form of foods that contain associated nutrients also important for bone building. One of these is zinc, which is a major determinant of serum insulin-like growth factor 1 (IGF-1) in normal adults (Devine et al. 1998). IGF-1 is osteotrophic; in several situations, it promotes bone gain, for example, after recovery from hip fracture (Bonjour et al. 1996). Clearly, these effects of protein on bone are all positive and underscore the importance of ensuring an adequate protein intake throughout life. Far from having an adverse effect, protein intake would seem to be good for bone.

However, the relationship between intake of a nutrient and production of harmful effects is typically biphasic. As noted, there is clearly demonstrable harm associated with low protein intakes; this is indicated schematically in the left-hand portion of Figure 1. The question addressed in this symposium relates to the right-hand side of the figure. Is there harm associated with high intakes? In one sense, the answer must be a positive one because too much of any nutrient may be toxic. But the question is better expressed as follows: are protein intakes at the upper end of the range likely to be found in the population harmful? At what level of intake in Figure 1 does the curve start to rise again?

View larger version (16K): [in this window] [in a new window]   Fig 1. Relationship between level of protein intake and production of biological harm. (Copyright Robert P. Heaney, 1997. Reproduced with permission.)

In theory, the principal mechanism by which harm might be produced is the increase in urinary calcium loss that has been extensively reported for high protein intakes. As Barzel (1995) has pointed out, this is related to the acid load commonly associated with proteins, either from the oxidation of sulfur-containing amino acids or from the relatively large number of inorganic ions associated with meat. The well-documented net effect of such loads is an increase in urinary calcium excretion. Other things being equal, increased calcium loss ought to be bad for bone. Sodium is an example of a nutrient that increases urinary calcium loss, with an effect of roughly the same magnitude as that for protein (Nordin et al. 1993). For sodium, there is reasonably solid evidence that high sodium intakes are associated with greater bone loss and reduced bone mass (Devine et al. 1995).

Despite the extensively documented effect of protein on urinary calcium loss when studied under controlled conditions, there is surprisingly little evidence in the real world outside of the metabolic unit that high protein intakes reduce bone mass or increase fracture risk. One bit of evidence, commonly cited as support for a harmful effect of protein, is the study of Abelow et al. (1992), which showed a direct correlation between hip fracture risk in populations and their corresponding protein intakes. However, such comparisons, involving a wide variety of ethnic groups and nations, are both inappropriate and unconvincing, if for no other reason than because there are so many other confounding variables that may well be more important. For example, in this instance, one can substitute other variables such as latitude or gross national product for protein intake, and the plot would look very much the same. Moreover, ethnic groups differ in other ways, such as in hip geometries, which are known to be associated with different intrinsic fragility; hence lumping such different populations together in a single plot is very much a matter of mixing apples with oranges. Only when differences in fracture rate can be found within an ethnic or national group as a function of varying protein intake will this type of observational study yield plausible conclusions. Here the evidence is decidedly mixed.

Table 1 lists eight observational studies that have been published within the last few years. Two showed no association between protein intake and either osteoporosis or bone loss; three showed a positive association between protein intake and bone mass, i.e., high protein was associated with better bones; three showed a negative association. One would have to say that the situation is unclear, at least on this evidence. However, it must be noted that it is often very difficult to assess intake of any nutrient adequately in observational studies because of the very large inaccuracies involved in estimating any nutrient intake from food-frequency questionnaires or diet records. Furthermore, protein intake is strongly positively associated with calcium intake in mixed diets (Holbrook and Barrett-Connor 1991). Because the two nutrients presumptively have countervailing effects on bone mass, it should not be surprising that the putative negative effect of protein might be obscured or neutralized in individuals consuming uncontrolled diets. As evidence of this, in most of the observational studies in which both protein and calcium intakes were estimated and in which a positive bivariate correlation between protein and bone mass was found, the association disappeared from the regression model when appropriate adjustment for calcium intake was made.

Because of the many confounding variables in observational studies, it will be more useful to evaluate the evidence derived from metabolic studies. Here the picture is much clearer. Increased protein intake does increase urinary calcium excretion. In several studies over many years, investigators at the Universities of California, Wisconsin and Chicago demonstrated conclusively that urine calcium rose as protein intake increased (Anand and Linkswiler 1974, Chu et al. 1975, Johnson et al. 1970, Margen et al. 1974, Schwartz et al. 1973, Walker and Linkswiler 1972). Perhaps one of the clearest demonstrations of the effect is seen when the amino acid load in a total parenteral nutrition solution is raised. The result is an immediate increase in urinary calcium loss (Bengoa et al 1983).

Figure 2 shows the relationship of relative urinary calcium excretion to relative protein intake, derived from 11 experiments summarized by Heaney and Recker (1982). In aggregating the data from various studies, baseline protein intakes and urinary calcium values in each report were normalized to a value of 1.0. Altered intakes and excretion values are expressed as relative values with respect to this baseline. Thus a value of 2.0 means a doubling of protein intake or urinary calcium relative to baseline values. The degree of excess protein intake was quite high in several of the experiments, and was probably greater than would be commonly found in free-living individuals. Nevertheless, it is evident that there is a quite good linear relationship between the change in protein intake and change in urinary calcium across the entire range of intakes from subnormal to excessive. The value for the slope of the relationship was +0.52, which means that a doubling of protein intake produces about a 50% increase in urinary calcium.

View larger version (12K): [in this window] [in a new window]   Fig 2. Scatterplot and regression of relative urine calcium on relative protein intake in studies analyzed by Heaney and Recker (1994). (Copyright Robert P. Heaney, 1997. Reproduced with permission.)

Because all of these studies were performed with isolated protein rather than with real foods, it has been objected that the effect would be different under more natural conditions. In fact, it is true that when protein is accompanied by phosphorus, as would be commonly the case with actual protein-rich foods, the increase in urinary calcium is blunted (Spencer et al. 1978). However, it is less well recognized that the extra phosphorus leads to an increase in the excretion of calcium through the gut because of increased calcium in the digestive secretions, not because of decreased calcium absorption. The net negative effect of the increased protein intake remains the same (Heaney and Recker 1994). The studies of Spencer et al. (1978) did not include measurement of endogenous fecal calcium; hence they would necessarily have missed the increased fecal loss. Moreover, because they and others have shown that increased phosphorus intake does not alter calcium balance by itself, it follows that the hypocalciuric effect of phosphorus must be offset by a corresponding increase in fecal calcium loss, even though it is difficult to detect that increase without tracer methods. Thus the reduced hypercalciuria seen with food sources of increased protein does not provide a solution to this problem.

Another approach leading to the same conclusion is provided by metabolic studies in a group of middle-aged women studied on their self-selected diets, performed in my laboratory (Heaney and Recker 1982). When protein intake and urinary calcium were analyzed similarly as for Figure 2, the slope of the relationship was virtually identical to the one derived from the studies summarized in Figure 2. However, in this case, all of the protein intake was in the form of real foods, with all of their associated co-ingested nutrients, including phosphorus. Specifically, in the women who were the subjects of our study, multiple regression modeling could be described by the following equation: In other words, urinary calcium was higher by about 0.85 mg for each extra gram of protein ingested. Although this value is relatively small, it still means that ~40-50% of daily urinary calcium loss in typical adult women is driven by their protein intakes.

Our study was not the only one showing this relationship with food sources of protein. A quantitatively similar result was produced by an intervention study of van Beresteijn et al. (1990), in which the effect of two different milk sources on urine calcium was described. Subjects were given either regular milk or a protein-reduced, calcium-augmented milk. Calcium and phosphorus intake from all sources was approximately the same in both groups. Urine calcium was lower in the individuals receiving the lower protein milk, by a value very similar to that predicted from our studies.

Thus, there is reasonable certainty that calcium excretion rises as protein intake increases. Therefore, the question must be the following: does this change adversely affect bone? I have already alluded to one reason why it may not, i.e., the fact that increased protein intake in self-selected diets is often associated with increased calcium intake as well. Therein lies the clue to answering this question.

A quantitative example will help illustrate what I mean. If one were to ingest an extra 140 g of beef every day, protein intake would increase by about 40 g, and endogenous calcium loss would increase by 36-40 mg/d. Other things being equal, this should result in measurable bone loss. But the body adapts to the associated drop in extracellular fluid [Ca⁺⁺]. This increases secretion of parathyroid hormone, which leads to increased 1--hydroxylation of calcidiol and hence increased absorption of food calcium. Calculations from data of Dawson-Hughes et al. (1988) allow one to estimate that a 40-mg increase in calcium loss would produce an increase in serum calcitriol with a slope of 0.0022 pmol/L. Thus the absorptive response to an 8 pmol/L increase in calcitriol would be an increase in absorption fraction of 8 × 0.0022, or 0.018. Not surprisingly, the amount of additional absorbed calcium that change would yield depends upon the calcium intake. At 1500 mg/d, an absorptive increase of this magnitude would yield about 27 mg of extra calcium absorbed from the diet, whereas at an intake of 250 mg/d, this same absorptive increase yields only 4.5 mg. Thus, whether protein has an appreciable adverse effect on the skeleton depends upon the body's ability to adapt, which in turn is dependent upon the adequacy of calcium intake.

Protein increases the obligatory loss of calcium. The importance of obligatory excretion lies in the fact that it constitutes a floor below which the body cannot reduce calcium loss when faced with dietary deficiency. But in the absence of deficiency, it creates no problem. Because high protein intakes tend to be associated with high calcium intakes, there will generally be no appreciable effect of protein intake on bone status at the population level.

As a consequence of these considerations, it will be more useful to evaluate diets not on their protein content per se, but on their calcium-to-protein ratios (Heaney 1993, van Beresteijn et al. 1990). An illustration of this utility is found in a longitudinal study of bone gain in women in their third decade (Recker et al. 1992), in which weak associations between bone gain and both calcium and protein intake were found. However, when the nutrient intakes were expressed as a calcium-to-protein ratio, the relationship became much stronger. The covariation of the two nutrients tended to minimize their effects in self-selected diets. (In standard statistical modeling, the same result is produced by adjusting the intake of one nutrient for that of another. However, such adjustment does not yield a tangible index the way a ratio does.)

What then can be estimated to be a "good" calcium:protein ratio? The 1997 recommended intakes for the two nutrients for middle-aged women (Food and Nutrition Board 1997) result in a calcium:protein ratio of 20:1 (mg calcium/g protein). Presumably that value is adequate; at least it can serve as a reference point against which other intakes can be compared. The data of NHANES III (Alaimo et al. 1994, McDowell et al. 1994), by contrast, yield a ratio of 9.3:1 for the actual median intakes of 50- to 59-y-old women, a value less than half the recommendation of the Food and Nutrition Board. Although median protein intakes are somewhat higher than they need to be, the real villain in this situation is not the protein intake, but the low calcium intake of this group. The population intakes of protein would have to be reduced to frank deficiency to reach a 20:1 ratio at prevailing calcium intakes.

	FOOTNOTES

Manuscript received 27 January 1998. Revision accepted 9 March 1998.

	LITERATURE CITED

• Abelow B. J., Holford T. R., Insogna K. L. Cross-cultural association between dietary animal protein and hip fracture: a hypothesis. Calcif. Tissue Int. 1992; 50:14-18[Medline]
• Alaimo, K., McDowell, M. A., Briefel, R. R., Bischof, A. M., Caughman, C. R., Loria, C. M. & Johnson, C. L. (1994) Dietary intakes of vitamins, minerals and fiber of persons ages 20 months and over in the United States: Third National Health and Nutrition Examination Survey, Phase 1, 1988-91. Advance data from vital and health statistics, no. 258. National Center for Health Statistics, Hyattsville, MD.
• Anand C. R., Linkswiler H. M. Effect of protein intake on calcium balance of young men given 500 mg calcium daily. J. Nutr. 1974; 104:695-697
• Barzel U. S. The skeleton as an ion exchange system: implications for the role of acid-base imbalance in the genesis of osteoporosis. J. Bone Miner. Res. 1995; 10:1431-1436[Medline]
• Bastow M. D., Rawlings J., Allison S. P. Benefits of supplementary tube feeding after fractured neck of femur. Br. Med. J. 1983; 287:1589-1592
• Bengoa J. M., Sitrin M. D., Wood R. J., Rosenberg I. H. Amino-acid induced hypercalciruia in patients on total parenteral nutrition. Am. J. Clin. Nutr. 1983; 38:264-269[Abstract/Free Full Text]
• Blaauw R., Albertse E. C., Beneke T., Lombard C. J., Laubscher R., Hough F. S. Risk factors for the development of osteoporosis in a South African population. S. Afr. Med. J. 1994; 84:328-332[Medline]
• Bonjour J.-P., Schurch M. A., Rizzoli R. Nutritional aspects of hip fractures. Bone 1996; 18:139S-144S[Medline]
• Chiu J.-F., Lan S.-J., Yang C.-Y., Wang P.-W., Yao W.-J., Su I.-H., Hsieh C.-C. Long-term vegetarian diet and bone mineral density in postmenopausal Taiwanese women. Calcif. Tissue Int. 1997; 60:245-249[Medline]
• Chu J.-Y, Margen S., Costa F. M. Studies in calcium metabolism. II. Effects of low calcium and variable protein intake on human calcium metabolism. Am. J. Clin. Nutr. 1975; 28:1028-1035[Abstract/Free Full Text]
• Cooper C., Atkinson E. J., Hensrud D. D., Wahner H. W., O'Fallon W. M., Riggs B. L., Melton L. J., III. Dietary protein intake and bone mass in women. Calcif. Tissue Int. 1996; 58:320-325[Medline]
• Dawson-Hughes B., Harris S. S., Dallal G. E. Plasma calcidiol, season, and serum parathyroid hormone concentrations in healthy elderly men and women. Am. J. Clin. Med. 1997; 65:67-71
• Dawson-Hughes B., Stern D. T., Shipp C. C., Rasmussen H. M. Effect of lowering dietary calcium intake on fractional whole body calcium retention. J. Clin. Endocrinol. Metab. 1988; 67:62-68[Abstract/Free Full Text]
• Delmi M., Rapin C.-H., Bengoa J. M., Delmas P. D., Vasey H., Bonjour J.-P. Dietary supplementation in elderly patients with fractured neck of the femur. Lancet 1990; 335:1013-1016[Medline]
• Devine A., Criddle R. A., Dick I. M., Kerr D. A., Prince R. L. A longitudinal study of the effect of sodium and calcium intakes on regional bone density in postmenopausal women. Am. J. Clin. Nutr. 1995; 62:740-745[Abstract/Free Full Text]
• Devine, A., Rosen, C., Baylink, D. & Prince, R. L. (1998) Effects of zinc and other nutritional factors on IGF-1 and IGF binding proteins in postmenopausal women. Am. J. Clin. Nutr. (in press).
• Feskanich D., Willett W. C., Stampfer M. J., Colditz G. A. Protein consumption and bone fractures in women. Am. J Epidemiol. 1996; 143:472-479[Abstract/Free Full Text]
• Food and Nutrition Board, Institute of Medicine (1997) Dietary Reference Intakes for Calcium, Magnesium, Phosphorus, Vitamin D and Fluoride. National Academy Press, Washington, DC.
• Heaney R. P. Protein intake and the calcium economy. J. Am. Diet. Assoc. 1993; 93:1261-1262[Medline]
• Heaney R. P., Barger-Lux M. J., Dowell M. S., Chen T. C., Holick M. F. Calcium absorptive effects of vitamin D and its major metabolites. J. Clin. Endocrinol. Metab. 1997; 82:4111-4116[Abstract/Free Full Text]
• Heaney R. P., Recker R. R. Effects of nitrogen, phosphorus, and caffeine on calcium balance in women. J. Lab. Clin. Med. 1982; 99:46-55[Medline]
• Heaney R. P., Recker R. R. Determinants of endogenous fecal calcium in healthy women. J. Bone Miner. Res. 1994; 9:1621-1627[Medline]
• Holbrook T. L., Barrett-Connor E. Calcium intake: covariates and confounders. Am. J. Clin. Nutr. 1991; 53:741-744[Abstract/Free Full Text]
• Johnson N. E., Alcantara E. N., Linkswiler H. M. Effect of protein intake on urinary and fecal calcium and calcium retention of young adult males. J. Nutr. 1970; 100:1425-1430
• Lacey J. M., Anderson J.J.B., Fujita T., Yoshimoto Y., Fukase M., Tsuchie S., Koch G. C. Correlates of cortical bone mass among premenopausal and postmenopausal Japanese women. J. Bone Miner. Res. 1991; 6:651-659[Medline]
• Margen S., Chu J.-Y., Kaufman N. A., Callaway D. H. The calciuretic effect of dietary protein. Am. J. Clin. Nutr. 1974; 27:548-558
• McDowell, M. A., Briefel, R. R., Alaimo, K., Bischof, A. M., Caughman, C. R., Carroll, M. D., Loria, C. M. & Johnson, C. L. (1994) Energy and macronutrient intakes of persons ages 2 months and over in the United States: Third National Health and Nutrition Examination Survey, Phase 1, 1988-91. Advance data from vital and health statistics, no. 255. National Center for Health Statistics, Hyattsville, MD.
• Metz J. A., Anderson J.J.B., Gallagher P. N. Jr. Intakes of calcium, phosphorus and protein and physical-activity level are related to radial bone mass in young adult women. Am. J. Clin. Nutr. 1993; 58:537-542[Abstract/Free Full Text]
• Nordin B.E.C., Need A. G., Morris H. A., Horowitz M. The nature and significance of the relationship between urinary sodium and urinary calcium in women. J. Nutr. 1993; 123:1615-1622
• Recker R. R., Davies K. M, Hinders S. M., Heaney R. P., Stegman M. R., Kimmel D. B. Bone gain in young adult women. J. Am. Med. Assoc. 1992; 268:2403-2408[Abstract/Free Full Text]
• Reid I. R., Ames R. W., Evans M. C., Sharpe S. J., Gamble G. D. Determinants of the rate of bone loss in normal postmenopausal women. J. Clin. Endocrinol. Metab. 1994; 79:950-954[Abstract]
• Schwartz R., Woodcock N. A., Blakely J. D., MacKellar I. Effect of magnesium and protein level on calcium balance. Am. J. Clin. Nutr. 1973; 26:519-523[Abstract]
• Spencer H., Kramer L., Osis D., Norris C. Effect of a high protein (meat) intake on calcium metabolism in man. Am. J. Clin. Nutr 1978; 31:2167-2180[Abstract/Free Full Text]
• van Beresteijn E.C.H., Brussaard J. H., van Schaik M. Relationship between the calcium-to-protein ratio in milk and the urinary calcium excretion in healthy adultsa controlled crossover study. Am. J. Clin. Nutr. 1990; 52:142-146[Abstract/Free Full Text]
• Walker R. M., Linkswiler H. M. Calcium retention in the adult human male as affected by protein intake. J. Nutr. 1972; 102:1297-1302

0022-3166/98 $3.00 ©1998 American Society for Nutritional Sciences