| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Endocrine Division (B.A.I., R.M.N.) and Mallinckrodt General Clinical Research Center (E.J.A.), Massachusetts General Hospital, Boston, Massachusetts 02114
Address all correspondence and requests for reprints to: B. Avery Ince M.D., Ph.D., Endocrine Division, Bulfinch 327, 55 Fruit Street, Boston, Massachusetts 02114. E-mail: bince{at}partners.org.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Thirty-nine healthy, premenopausal women consuming ad libitum diets [mean, 1.1 g/kg protein, 819 mg (20.5 mmol) Ca, 1152 mg (37 mmol) P, 129 mmol Na] were switched to isocaloric diets containing the U.S. recommended dietary allowance (RDA) of protein (0.8 g/kg) and similar amounts of calcium, phosphorus, and sodium. Bone resorption and related endpoints were assessed before and 1 wk after the switch.
As dietary protein changed from ad libitum to RDA levels, mean urine nitrogen decreased 26% (2.4 g/d; P < 0.001) and mean blood urea nitrogen decreased 15% (1.9 mg/dl; P < 0.001). Mean urine pH increased from 6.3 to 6.8 (P < 0.001), and net renal acid excretion (NRAE = urine ammonium plus titratable acids minus bicarbonate) decreased 68% (21.4 mEq/d; P < 0.001). Mean urinary calcium decreased 32% [42 mg (1 mmol)/d; P < 0.001], and bone resorption urine N-telopeptides) decreased 17% (74 µmol bovine collagen equivalents/d; P < 0.001). Mean serum calcium, PTH, and 1,25 dihydroxy vitamin D remained unchanged.
In this 2-wk study, decreasing dietary protein from ad libitum to RDA levels decreased NRAE, calciuria and estimates of bone resorption, suggesting that decreased U.S. protein consumption might reduce bone loss. Inasmuch as other dietary modifications, such as increasing vegetable and fruit intake, can result in sustained reductions in NRAE without reducing protein intake, the advisability of reducing protein intake for skeletal protection from acid attack requires further investigation.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The typical American diet includes 70–100 g of dietary protein daily (7, 8), far exceeding the U.S. recommended dietary allowance (RDA) of 0.8 g/kg/d, or 44 g/d for the idealized 55-kg woman (9). Dietary protein generates nonvolatile or fixed acids, and subjects eating typical American diets have blood pH and serum bicarbonate concentrations that decrease progressively (within the normal range) as endogenous nonvolatile acid production increases (10, 11, 12, 13). In such analyses, the increases in net renal acid excretion (NRAE) and in urinary calcium excretion are directly related (12).
The aforementioned observations can be explained by invoking the notion of a skeletal base reservoir (13, 14) consisting primarily of alkaline salts of calcium, which is mobilized to neutralize acid production. Calcium salts can, in theory, be mobilized from the skeleton by increases in bone resorption or by decreases in skeletal accretion. Either would reduce bone mineral content and thus might play an etiological role in osteoporosis or other metabolic bone diseases. A possible illustration of such a phenomenon might be hip fracture, the incidence of which is much higher among Western, industrialized nations with elevated protein intakes than it is among developing nations with low protein consumption, even after correction for numerous confounders (15). Although some recent physiological trials have demonstrated increased bone collagenolysis with increased protein intake (16), and other trials have demonstrated an effect of acid-base balance on bone remodeling (17, 18), still other clinical studies have demonstrated a beneficial effect of protein supplementation on bone density in malnourished, elderly patients having suffered a recent hip fracture (19), and a protective effect of dietary protein on bone density in the elderly (20). Furthermore, the human genome likely evolved in primates ingesting a high-protein diet (21). In part due to difficulties inherent in making broadly generalizable inferences about the skeletal effect of dietary protein from studies using special populations (such as the malnourished elderly), and in part because no studies have yet investigated the effects of the ad libitum North American diet on bone resorption, we conducted this General Clinical Research Center (GCRC)-based physiological trial.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Forty-two healthy, ambulatory 22- to 39-yr-old women (33 Caucasians, 4 African-Americans, 4 Asians, and 1 Hispanic) were recruited by advertisement and paid for their participation. Women were excluded if significant illness, prescription drug use, abnormal menses, or abnormal complete blood count, renal and liver function tests, or blood TSH levels were present. After written, informed consent was obtained and randomization was performed, three women withdrew prematurely: one for personal reasons, one due to an intercurrent illness precluding further participation, and one due to exclusion for noncompliance. The mean age and weight of participants who completed the study were 27.3 ± 1.8 yr and 60.5 ± 1.6 kg, respectively.
Study design
Four-day food records were obtained during wk 1 to estimate ad libitum dietary intake. During wk 2, women received all food and beverages from the Massachusetts General Hospital (MGH) GCRC while continuing usual daily activities. Each woman served as her own control. The wk-2 diet was calculated to contain the same calories, calcium, phosphorus, and sodium as the ad libitum diets, but only 0.8 g/kg protein. Typical foodstuffs included cereal, milk, orange juice, tea, coffee, canned fruits and vegetables, bread, butter, chicken, meat, pasta, and cookies. Reductions in dietary protein were accomplished by decreasing allotments of chicken and meat. The amino acid compositions of the ad libitum and study (RDA protein) diets were very similar to one another and contained equivalent proportions of sulfur-containing amino acids, which are known to generate fixed acids disproportionately (22 ; Table 1
). In the RDA protein diets, total calories were kept constant by increasing dietary fat (through addition of butter and plain cookies) and by increasing dietary carbohydrates (through addition of simple sugar-rich foods such as jelly beans, candy corn, and plain cookies). Intake of cereal grains (breakfast cereal, pasta, bread, and bagels) remained unchanged on both diets because dietary starch may affect urinary calcium (23) and because these foods are thought to be potentially acid-forming (24). The carbohydrate compositions of the ad libitum and study diets are recorded in Table 2. Phosphorus intake was kept constant by use of NeutraPhos (Beach Pharmaceuticals, Tampa, FL), a neutral pH mixture of potassium and sodium phosphates, as necessary. The sodium and potassium in NeutraPhos were included in all calculations of intake. Dietary compliance was determined by 24-h urine nitrogen, sodium, and potassium measurements, food-item checkoff lists, weighing uneaten food for estimation of nutrient content, and anonymous subject questionnaires.
|
|
Data collection and laboratory procedures
Each urine specimen was divided in half by the subjects. Subjects (who were all coached in the gathering of divided collections) were given a "hat" for initial collection of urine. After each void, the contents of the hat were then transferred in equal amounts to each of two clear plastic cups. The contents of each cup were then transferred either to the mineral oil or to the acid collection container. Half of each voided urine sample in every 24-h specimen was collected under mineral oil for measurement of pH, bicarbonate, ammonium, titratable acids, and type 1 collagen NTx. The remaining half was collected with HCl for measurement of other analytes. Most measurements were performed using techniques referenced previously (25, 26). Urine NTx were determined by ELISA (Osteomark, Ostex International Inc., Seattle, WA). Serum osteocalcin was determined using an immunoradiometric sandwich assay (Nichols Institute, San Juan Capistrano, CA). Urinary titratable acids were measured by titration to pH 7.4 with NaOH.
Nutrient calculations
Nutrient content of the diets was calculated using the Nutrition Data System for Research (NDS-R) software version 4.02 (developed by the Nutrition Coordinating Center at the University of Minnesota, Minneapolis, MN, Food and Nutrient Database 30, released November 1999).
Statistics
All group data are expressed as mean ± SEM. For each woman, we calculated the mean of each variable at the ends of wk 1 and 2 and assessed the statistical significance of changes using paired t tests. We performed univariate regression analyses of NRAE against all other measured variables using SAS PROC REG (SAS Institute, Cary, NC) and then did the same for urinary calcium and NTx. All independent variables for which the univariate regression had a P value < 0.05 were then used in a multivariate analysis (SAS PROC GLM) to assess their relative contributions to variations in NRAE and, separately, urinary calcium and urine NTx. To assess the statistical significance of these multivariate relationships, we used the repeated-measures regression approach (SAS PROC GEN MOD) from which the SE values of the regression coefficients were provided by the generalized estimating equations.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Analysis of 4-d food records kept during wk 1 revealed mean intakes of 67.3 g protein, 819 mg (20.5 mmol) calcium, 1152 mg (37 mmol) phosphorus, and 129 mmol sodium. Average U.S. daily intakes of these nutrients are 70–100 g protein, 713 mg (18 mmol) calcium, 1025 mg (33 mmol) phosphorus, and 142 mmol sodium (7, 27, 28, 29, 30). As such, the ad libitum diets consumed by our subjects were very typical of those consumed by young, adult women in the general U.S. population. Week-2 measurements of urine nitrogen (6.9 ± 0.2 g/d), phosphorus [525 ± 18 mg (17 ± 0.6 mmol)/d], and sodium (98 ± 5 mmol/d) correlated well with calculated values, as shown in Table 3.
|
Mean urine nitrogen decreased from 9.3 ± 0.5 g/d (wk 1) to 6.9 ± 0.2 g/d (wk 2) (P < 0.001; Fig. 1A). As such, the mean change was 2.4 ± 0.4 g/d. Mean urine sodium decreased from 129 ± 8 mmol/d (wk 1) to 98 ± 5 mmol/d (wk 2) (P < 0.001), a mean change of 31 ± 7 mmol/d. Mean urine K was 58 ± 3 mmol/d in wk 1 and 62 ± 2 mmol/d in wk 2, but this difference was not statistically significant (P = 0.36). Judging by these objective data, review of the subject food-item checkoff lists, and the anonymous questionnaires (showing that >99% of all study food was consumed by all subjects), overall dietary compliance was excellent.
|
Mean NRAE decreased from 31.4 ± 2.4 mEq/d in wk 1 to 9.9 ± 1.7 mEq/d in wk 2 (P < 0.001; Fig. 1B
). The mean change was 21.5 ± 2.2 mEq/d (67%). Univariate regression of mEq/d of NRAE upon grams per day of urine nitrogen (the best proxy for protein consumed) revealed a significant, positive relationship defined by the equation NRAE = 3.9 N – 11; R = 0.60 (data not shown). In contrast, univariate regressions of NRAE on total carbohydrate intake and total starch intake revealed a negative relationship between NRAE and grams per day of total carbohydrates [NRAE = –0.12 total carbohydrates + 55; R = 0.41 (data not shown)], and a weak, positive relationship between NRAE and grams per day of starch intake [NRAE = 0.06 starch + 14; R = 0.09 (data not shown)]. Repeated-measures multivariate analysis revealed that NRAE increases by 2.12 mEq/d for every gram of urine nitrogen (P = 0.05), decreases by 0.08 mEq/d for every gram of total carbohydrate consumed (P < 0.001), and is unrelated to starch intake. Mean urine pH also increased from 6.26 ± 0.1 in wk 1 to 6.75 ± 0.1 in wk 2 (P < 0.001), with a mean change of 0.5 ± 0.1. There were subtle but nonetheless statistically significant changes in serum bicarbonate and also in venous blood pH. Mean serum bicarbonate increased slightly from 25.1 ± 0.9 to 25.5 ± 1.0 mEq/liter (P = 0.003), and mean venous blood pH increased from 7.39 ± 0.1 to 7.40 ± 0.1 (P = 0.003).
Urinary calcium excretion
During wk 1 (ad libitum intake), the mean 24-h urinary calcium was 132 ± 7 mg (3.3 ± 0.2 mmol). During wk 2 (RDA protein intake), urinary calcium was 89 ± 5 mg (2.2 ± 0.1 mmol), having decreased by 42 ± 7 mg (1.0 ± 0.2 mmol/d) (P < 0.001; Fig. 1C). Urinary calcium excretion varied as a function of NRAE in a direct, positive manner (Fig. 2). Regression of 80 observations revealed a linear relationship in which Ca = 84.89 + 1.09 NRAE. The relationship had an R value of 0.35; 95% confidence intervals were 0.49–2.00 for the slope and 66–105 for the intercept. Univariate regression of milligrams per day of urinary calcium on grams per day of urine nitrogen also revealed a significant positive relationship defined by the equation Ca = 11.9 N + 15; R = 0.51 (data not shown), but urinary calcium was not significantly related to dietary carbohydrate or starch. Since urinary calcium was also significantly related to urinary sodium, potassium, and phosphorus in univariate regressions, we performed a repeated-measures multivariate analysis, which revealed that urine calcium increased by 10 mg/d for every gram of urine nitrogen (P < 0.001), 0.45 mg/d for every mEq of urine Na (P < 0.001), 0.9 mg/d for every gram of urine K (P = 0.004), and decreased by 0.1 mg/d for every milligram increase in urine phosphorus (P = 0.03). Urinary calcium is no longer significantly related to NRAE once urine nitrogen is taken into account, probably because NRAE and urine nitrogen are highly correlated.
|
Mean urine NTx decreased from 442 ± 20 µmol bovine collagen equivalents (BCE)/d at the end of wk 1 to 368 ± 16 µmol BCE/d at the end of wk 2 (P < 0.001; Fig. 1D). As such, the mean change was 74 ± 8 µmol BCE/d. Urine NTx varied as a function of NRAE in a direct, positive manner (Fig. 3). Univariate regression on 80 observations revealed a linear relationship in which NTx = 332.52 + 3.54 NRAE. The relationship had an R of 0.36; 95% confidence intervals were 1.5–5.6 for the slope and 280–385 for the intercept. Univariate regression of urine NTx on urine nitrogen also revealed a direct, positive relationship defined by the equation NTx = 15.9 N + 277; R = 0.25 (data not shown). Repeated-measures multivariate analysis revealed that urine NTx increased 3.1 µmol BCE/d for every milliequivalent of NRAE (P = 0.001), and 0.15 µmol BCE/d for every milligram per day of urine creatinine (P = 0.01) but was not related to any other variable.
|
As expected, because of the slight (<5%) reduction in dietary phosphorus intake (Table 3), as well as the presumed reduction in bone resorption, urinary phosphorus decreased slightly, but significantly, from 638 ± 34 mg/d (20.6 ± 1.1 mmol/d) to 525 ± 18 mg/d (17 ± 0.6 mmol/d) (P < 0.001; Table 3). The mean change therein was 113 ± 30 mg/d (3.6 ± 1.0 mmol/d).
Osteoblast activity
Mean serum osteocalcin decreased slightly from 15.8 ± 1.4 ng/ml at the end of wk 1 to 13.4 ± 1.3 ng/ml at the end of wk 2, but this change was not statistically significant (P = 0.166).
Hormonal regulators of calcium homeostasis
Mean serum PTH trended upward from 29.6 ± 2.3 (wk 1) to 32.2 ± 2.1 pg/ml (wk 2), but this slight increase was not significant (P = 0.07). There was a 90% probability that the study could detect a treatment difference at a two-sided 5% significance level, if the true difference between the groups was 9.8 pg/ml.
Mean 1,25 dihydroxy vitamin D levels did not change significantly (35.6 ± 1.9 vs. 34.3 ± 1.7 ng/ml; P = 0.65); mean serum calcium levels were also unchanged (8.88 ± 0.03 vs. 8.90 ± 0.04; P = 0.51). There was a 90% probability of detection of a treatment difference in 1,25 dihydroxy vitamin D levels at a two-sided 5% significance level, if the true difference between the groups was 12.4 ng/ml.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
). Similarly, both Heaney et al. (32) and Insogna and Broadus (33) suggest an upper limit for "normal" urinary calcium in estrogen-replete women of approximately 250 mg/d. This value is consistent with the upper limit of 242 mg/d in our participants when they consumed ad libitum diets, but it is substantially higher than the upper limit of 164 mg/d observed in our participants while consuming RDA amounts of dietary protein (Table 3). Current normative U.S. and western European data (32, 33, 34) were likely gathered in subjects consuming typical (large) amounts of dietary protein (7, 8, 27), which increased urinary calcium (1, 2, 3, 4, 5, 6). As such, current concepts about the amount of urinary calcium that should be considered normal are biased by the diets typical of the U.S. and western Europe. In contrast, on diets approximating the RDA protein consumption, the mean urinary calcium in these 39 healthy women was 89 ± 5 mg/d (2.2 ± 0.1 mmol/d), an amount that represents hypocalciuria by most current standards. Further examination of this issue is warranted.
|
Dietary phosphorus reportedly mitigates protein-induced hypercalciuria (36), and some have suggested that the high phosphorus content of most foodstuffs containing dietary protein may prevent protein-induced bone loss (37). We show here that bone resorption is higher in subjects receiving ad libitum amounts of dietary protein, although phosphorus intake is slightly higher (if at all different) on this diet. Because the reduction in dietary protein prescribed was not accompanied by large decreases in phosphorus, our data neither support nor refute this possibility.
Our findings show that both urinary calcium and estimates of bone resorption vary in direct proportion to 1) urine nitrogen (the best proxy for protein intake) (data not shown) and 2) NRAE. NRAE is demonstrated to increase by 3.9 mEq/d with each gram of urinary nitrogen excreted (data not shown). In contrast, NRAE decreases by 0.12 mEq/d for every gram of total carbohydrate intake and increases by 0.06 mEq/d for each gram of starch consumed (data not shown). When metabolic acid loading with ammonium chloride is undertaken experimentally in healthy human volunteers, increases in urinary calcium are accompanied by decreased serum bicarbonate (13). There is, however, a lengthy plateau below which bicarbonate does not fall despite a continued, steady rise in urine hydrogen excretion (13). Because serum pH remains within normal limits during these experiments (13), it appears that a "nonbicarbonate" base reservoir large enough to offset considerable acid loading is being accessed. The skeleton, a massive reservoir of labile alkaline calcium salts that can be mobilized in defense of pH homeostasis, seems a plausible candidate. Our experiment, although short in duration, underscores the plausibility of this hypothesis and is supported by carefully executed animal studies demonstrating the long-term, adverse skeletal effects of chronic acid ingestion (38).
Alarming data reveal that 50% of U.S. women above the age of 55 yr have either osteopenia or osteoporosis (39). Our overarching hypothesis is that osteoporosis in this country may in part be the inevitable culmination of a lifelong assault on pH homeostasis that occurs mainly because of excess dietary acid precursors relative to dietary base precursors (14, 17, 18, 40). The resultant metabolic acid load routinely triggers not only well-described renal adaptive mechanisms including the primary buffer, serum bicarbonate, but also causes mobilization of skeletal base, likely calcium hydrogen phosphate (precursor of calcium hydroxy-apatite), from bone. In this way, bone loses calcium and bone mass is diminished in the defense of pH homeostasis, acid-base balance being more centrally important to survival than skeletal health.
Although the aforementioned hypothesis was not directly tested in this short-term study, data from other experiments are consistent with this idea (16, 17). Our findings show that reductions in U.S. daily dietary protein consumption can decrease bone resorption and may favorably affect calcium balance. Although not explored in these experiments, it may also be possible to counter the effects of elevated NRAE without altering protein intake, for example, by providing exogenous alkali (17, 18, 35) or by increasing intake of fruits and vegetables (24). We are currently conducting experiments to test this hypothesis. Such considerations may prove important in patients who have elevated NRAE despite ingestion of RDA levels of dietary protein or despite consumption of insufficient dietary protein. Whether NRAE, calcium excretion, or bone resorption would be reduced in free-living women who have switched to diets with RDA levels of protein, and who are not conforming to experimentally determined caloric and macronutrient quotas, is a very interesting question outside the scope of the present study. To the extent to which women reducing their nitrogen intake replace lost calories with dietary fat and/or dietary simple sugars without changing other nutrients (starch and phosphate, for example), our data suggest that NRAE, calcium excretion, and bone resorption will decrease. If the lost calories are instead replaced with cereal grains or other products thought to be acid-forming (24), the effect on NRAE and bone is unclear. Furthermore, whether bone gain ensues from decreased bone resorption depends on the relative contribution of concurrent changes in bone formation. Given that changes in bone turnover indices do not quantitatively predict changes in bone mass, long-term and larger studies should be conducted to explore the possibility that low-cost, nontoxic dietary interventions might mitigate adult bone loss. This is particularly important and timely because of recent data casting doubt upon the advisability of hormone replacement therapy to prevent bone loss (41).
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: BCE, Bovine collagen equivalents; NRAE, net renal acid excretion; NTx, type I collagen N-telopeptide; RDA, recommended dietary allowance.
Received November 19, 2003.
Accepted April 16, 2004.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  U. Alexy, T. Remer, F. Manz, C. M Neu, and E. Schoenau Long-term protein intake and dietary potential renal acid load are associated with bone modeling and remodeling at the proximal radius in healthy children Am. J. Clinical Nutrition, November 1, 2005; 82(5): 1107 - 1114. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. R. Zwart, J. E. Davis-Street, D. Paddon-Jones, A. A. Ferrando, R. R. Wolfe, and S. M. Smith Amino acid supplementation alters bone metabolism during simulated weightlessness J Appl Physiol, July 1, 2005; 99(1): 134 - 140. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Sakhaee, N. M. Maalouf, S. A. Abrams, and C. Y. C. Pak Effects of Potassium Alkali and Calcium Supplementation on Bone Turnover in Postmenopausal Women J. Clin. Endocrinol. Metab., June 1, 2005; 90(6): 3528 - 3533. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
