| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Research Institute of Child Nutrition, Dortmund, Germany (UA, TR, and FM), and the Children's Hospital, University of Cologne, Cologne, Germany (CMN and ES)
2 Supported by the Ministry of Science and Research North Rhine–Westphalia, Germany, and by a research grant from Protina Pharm GmbH (to TR).
3 Reprints not available. Address correspondence to T Remer, Research Institute of Child Nutrition, Department of Nutrition and Health, Heinstueck 11, D-44225 Dortmund, Germany. E-mail: remer{at}fke-do.de.
See corresponding editorial on page 921.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: We examined whether the long-term dietary protein intake and diet net acid load are associated with bone status in children.
Design: In a prospective study design in 229 healthy children and adolescents aged 6–18 y, long-term dietary intakes were calculated from 3-d weighed dietary records that were collected yearly over the 4-y period before a one-time bone analysis. Dietary acid load was characterized as potential renal acid load (PRAL) by using an algorithm including dietary protein, phosphorus, magnesium, and potassium. Proximal forearm bone variables were measured by peripheral quantitative computed tomography.
Results: After adjustment for age, sex, and energy intake and control for forearm muscularity, BMI, growth velocity, and pubertal development, we observed that long-term dietary protein intake was significantly positively associated with periosteal circumference (P < 0.01), which reflected bone modeling, and with cortical area (P < 0.001), bone mineral content (P < 0.01), and polar strength strain index (P < 0.0001), which reflected a combination of modeling and remodeling. Children with a higher dietary PRAL had significantly less cortical area (P < 0.05) and bone mineral content (P < 0.01). Long-term calcium intake had no significant effect on any bone variable.
Conclusions: Long-term dietary protein intake appears to act anabolically on diaphyseal bone strength during growth, and this may be negated, at least partly, if dietary PRAL is high, ie, if the intake of alkalizing minerals is low.
Key Words: Children • bone health • modeling • remodeling • peripheral quantitative computed tomography • dietary protein • potential renal acid load
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
90% of total bone mineral content is attained by age 17 y (5). However, the calcium and mineral contents of the skeleton appear to be markedly influenced by nutrients other than calcium, specifically protein (6–8) and alkalizing minerals (9, 10), which are increasingly described as playing a major role. Clinical studies have provided convincing evidence that protein supplements can have substantial positive effects on bone health in the elderly (11, 12). However, the findings of larger epidemiologic studies are less clear. Evidence for both a negative and a positive effect of protein on bone health exists. An overview of this topic was given by Ginty (8).
We examined the association of long-term protein intake and dietary potential renal acid load (PRAL) with diaphyseal radial bone in a sample of healthy children and adolescents with the use of peripheral quantitative computed tomography (pQCT).
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
As a spinoff project, a single pQCT analysis of the forearm was undertaken in 1998–1999 in 371 DONALD participants aged 6–18 y (14, 15). For the present study, we selected 229 (115 boys, 114 girls) of these participants who had 
4 of the possible 5 three-day weighed dietary records (4 records: n = 63 subjects; 5 records: n = 166 subjects) and valid reported energy intakes during the 4 y preceding bone analysis.
Ethical permission was obtained from the institutional review board of the Research Institute of Child Nutrition in Dortmund, the ethics committee of the medical faculty of the University of Cologne, and the Federal Office for Radiation Protection (Salzgitter, Germany). Parental informed consent and child's assent were obtained before entry into the study.
pQCT of forearm bone and forearm muscle area
An XCT-2000 device (Stratec Inc, Pforzheim, Germany) equipped with a low-energy (38 keV) X-ray tube was used to conduct the pQCT analysis (14–16) on the nondominant forearm. The effective radiation was 0.1 µSv from a radiation source of 45 kV at 15 µA. The scanner was placed on the forearm, where the distance from the ulnar styloid process was 65% of the forearm length. A 2-mm thick single tomographic slice was sampled at a voxel size of 0.4 mm. The speed of the translational scan movement was 15 mm/s. The resulting time for a measurement run was 2–3 min in the younger and 4–5 min in the older subjects, depending on the cross-sectional size of the forearm. Image processing and calculation of numerical values were performed by using the manufacturer's software package (version 5.40; Statex Inc, Paris, France). The cross-sectional area of cortical bone was determined by detecting the outer and inner cortical bone contour at a threshold of 710 mg/cm3. Periosteal circumference was determined under the assumption that the bone is cylindrical, whereby the outer bone radius was calculated as follows:
 | (1) |
These measurements allowed the evaluation of bone modeling and remodeling, which are 2 mechanisms used to construct and reconstruct the skeleton (17). Modeling characterizes the expansion process of the bone's cross-section assessed by determination of the periosteal circumference. In a cross-sectional study design, higher periosteal circumference values indicate that more modeling, ie, more skeletal construction has taken place in the respective subjects. Remodeling indicates the changes in cortical density or cortical porosity, ie, reconstruction. Bone mineral content, cortical area, and polar strength strain index reflect a combination of modeling and remodeling. Cross-sectional forearm muscle area was also determined with the XCT-2000 device at 65% of the ulnar length as previously described (15, 16, 18).
Anthropometric measurements and Tanner stages
Body weight was measured with an electronic scale to the nearest 0.1 kg and standing height to the nearest 0.1 cm with a digital telescopic wall-mounted stadiometer. From these measurements, body mass index [BMI; weight (kg)/height2 (m)] was calculated and converted into SD scores of BMI (SDS-BMI) by using Cole's LMS method (19), which allows the assessment of individual BMI in relation to a reference population. The recent data derived from 17 regional German surveys published by Kromeyer-Hauschild et al (20) were used as a reference.
Growth velocity (GV) was calculated in 2 ways: 1) as the 4-y GV taken from height measurements over the 4 y before pQCT and 2) as the mean GV at pQCT measurement, averaged from GV 1 y before and 1 y after pQCT.
Tanner stage 1-5 of the study participants were determined by a pediatrician. On the basis of pubic hair as a clinical marker of the beginning of adrenal androgen secretion (18), the subjects were assigned to 2 groups: without (prepubescent) and with (pubescent) pubic hair. The age of menarche in girls and of voice change in boys were reported by the children via standardized questioning.
Dietary survey
The parents of the children or of the older subjects themselves weighed and recorded all foods and fluids consumed, ingredients of home-prepared meals, as well as leftovers using electronic food scales (± 1 g) on 3 consecutive d. Semiquantitative recording (eg, number of spoons, scoops) was allowed if weighing was not possible. However, in 75% of the completed records, >90% of the food items were weighed (21).
Energy and nutrient intakes, including food fortification and nutrition supplements, were calculated as individual means on the recorded days by using our nutrient database LEBTAB (22).
To check for the validity of the dietary measurements, we used the reported energy intake as a surrogate measure of the general quality of the dietary data (23). For this, according to Goldberg et al (24), the ratio of reported energy intake and predicted basal metabolic rate were used. Basal metabolic rate was calculated by using the equations of Schofield (25), which included the measured height and weight of the individuals. For identification of implausible records, which were excluded from further analysis, sex- and age-specific cutoffs were used, considering CVs for energy intake and physical activity levels of light physical activity (26).
Dietary PRAL: estimation method and background
The dietary component (PRAL) of net endogenous acid production was estimated according to Remer et al (27, 28):
 | (2) |
For a maximum possible exact estimation of dietary PRAL, the use of the summated sulfur-containing amino acid content of the individual foods would be preferable (34, 35, 37). However, in our nutrient database LEBTAB, as in several other databases, food-specific data on methionine and cysteine contents are not available. Therefore, for the calculation of PRAL we used an average value for sulfur-containing amino acids. This proved to be an almost accurate estimate of measured urinary sulfate excretion in mixed diets: a lactovegetarian, a moderate-protein omnivorous diet, and a high-protein omnivorous diet (30) and, as mentioned above, the corresponding PRAL correlated highly significantly with the NAE analyzed in 24-h urine samples (28).
To check whether the intake of foods containing proteins with higher amounts of sulfur-containing amino acids might be associated with the analyzed bone variables, we additionally used the food grouping published by Massey (37) to aggregate those foods with the highest potential to generate acid as sulfate (meat, fish, eggs, and grain).
Statistical analysis
SAS procedures (version 8.2; SAS Institute Inc, Cary, NC) were used for data analysis. Data are represented as means ± SDs, unless indicated otherwise. Dietary intakes of protein (g/d), calcium (mg/d), and PRAL (mEq/d) were adjusted for age, sex, and total energy by using the residual method. Muscle area and bone variables were also adjusted for age and sex by using residuals.
With this procedure, the examined variables (nutrient intakes or bone variables) of the individuals were regressed on their total energy intakes or age. For nutrient intakes, this was done for each of the 4 or 5 study time points. The residuals from the regression represent the differences between each individual's variable size and the variable size predicted by their total energy intake or age. The variable residual is uncorrelated with total energy intake or age, and this allows the variation due to the examined variable to be evaluated directly.
Because dietary protein, calcium, and PRAL significantly correlated with each other (P < 0.0001, Pearson's correlation coefficient)—even after adjustment for age, sex, and energy intake—collinearity diagnostics according to Belsley, Kuh, and Welsch (38) were performed. No evidence of collinearity between any of the dietary factors (expressed as residuals) could be detected. Therefore, the residuals were included in subsequent multivariate analysis. Each of these residuals was the arithmetic mean of the 4 or 5 individual residuals obtained from the yearly diet records, thus reflecting long-term dietary intakes.
Preliminary analysis of covariance (ANCOVA) was used to test for interactions between diet and sex or developmental group. Because diet-by-sex and diet-by–developmental group interactions exclusively were nonsignificant (P > 0.1) for all bone variables studied, the subsequent analyses were performed with the total sample of 229 subjects.
For final analysis of long-term dietary intakes on measures of pQCT, stepwise multiple regression was used. Apart from dietary protein, PRAL, and calcium, muscle area, BMI, GV, menarche (voice change), and Tanner stages (calculated as dummy variables) were also included in the model as potential confounders. Regression analyses were run separately with GV at pQCT and 4-y GV (see Anthropometric measurements and Tanner stages).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and 2. The study sample was almost equally divided into subgroups of males and females and likewise for the developmental stages of prepubescence and pubescence.
|
|
). Muscle area was 1.6-fold higher in pubescent than in prepubescent boys; the respective difference for girls was 1.5 fold. The bone variables increased from the prepubescent to the pubescent stages, whereby only a slight increase in cortical density was found. Age and BMI did not differ significantly between boys and girls.
Daily intakes of energy, protein, minerals, and PRAL (Table 2) were higher in the pubescent than in the prepubescent group and higher in boys than in girls. However, when calculated as nutrient densities, intakes of protein, calcium, potassium, and phosphorus were independent of developmental stage and sex.
Protein intakes in prepubescent children were 2 g · kg–1 · d–1. In pubescent children, protein intakes were lower (1.6 g · kg–1 · d–1 in boys, 1.4 g · kg–1 · d–1 in girls). These intakes represented an overall higher protein intake than the recently published recommended dietary allowances of 0.95 g · kg–1 · d–1 in 4–13-y-old boys and girls and of 0.85 g · kg–1 · d–1 in 14–18-y old boys and girls.
Mean calcium intakes were modestly below the adequate intakes proposed by the National Institute of Medicine (39), ie, 800 mg/d for 3–8-y-old boys and girls and 1300 mg/d for 9–18-y-old boys and girls). Highly significant positive associations were seen between muscle area and pQCT measures of periosteal circumference, cortical area, bone mineral content, and polar strength strain index (Table 3). Also, protein was significantly and positively associated with these bone variables. For PRAL, the results were significant only for cortical area and bone mineral content, showing a negative association. Of all the examined independent variables, only mean GV at pQCT showed a significant negative association with cortical density (r2 = 0.04, P = 0.0015). However, GV at pQCT was not associated with any of the other bone variables, whereas 4-y GV showed significant positive associations with cortical area (r2 = 0.03, P = 0.0016) and bone mineral content (r2 = 0.02, P = 0.0091) (data not shown). The associations of the bone variables with protein intake and dietary PRAL remained unchanged when the regression analyses were run with 4-y GV instead of GV at pQCT. Calcium did not enter the model for any bone variable. Only sporadic associations with bone variables were seen for BMI and menarche or voice change (Table 3). Additionally, Tanner stages (especially Tanner stage 5) positively predict the already age-adjusted cortical area and bone mineral content.
|
). For each bone variable, the standardized regression coefficients were highest for muscle area (Table 3).
The associations between age-adjusted bone variables and long-term dietary protein intake (calculated as a percentage of energy intake) and PRAL (adjusted for protein intake, age, and sex) are shown in Figures 1 and 2. Subjects with a higher or a lower proportion of sulfur-containing amino acids in their dietary protein did not show any consistent bias.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In general, our results agreed with findings in elderly patients who had bone anabolic effects after protein supplementation of their initially low-protein diets (11, 12). Our study provides evidence of a consistent positive association of dietary protein with periosteal circumference, cortical area, bone mineral content, and polar strength strain index at the proximal diaphyseal radius in children and youth. This potential protein anabolism was found for habitual Western diets with higher protein intakes and explained 3–4% of the examined variability in bone variables. This is clearly less than what is explained by muscularity (Table 3) and adrenarchal hormones (r2 
0.1) (18).
In line with recent findings in a juvenile longitudinal cohort (41), which reported a very poor correlation of GV with diaphyseal bone, we also found a negative association of GV measured by pQCT only with cortical density. Whether the observed positive association of long-term growth (4-year GV) with cortical area and bone mineral content might reflect a common anabolic cause deserves further research.
Interestingly, variations in protein intake did not associate with cortical density, in line with the explanation that the metabolic activity in cortical bone (remodeling) is influenced more by estrogens than by androgens (14) and that muscularity, which interacts with the growth hormone–insulin-like growth factor (IGF) system, has almost no or only a modest effect on cortical density (18). However, the extent of protein intake seems to stimulate modeling, ie, the main process for increasing bone strength during childhood and adolescence (1).
A protein-induced increase in IGF-I is strongly assumed to be the most likely explanation for an osteotrophic effect of protein (8). IGF-I is a major determinant of bone growth and mineral content (42). Until now, associations between dietary protein intake and IGF-I were primarily studied in elderly or malnourished children. However, in a recent study, Hoppe et al (43) showed a significant positive association between protein intake, growth, and circulating IGF-I concentrations in healthy young children. This may underline the in vivo potential of protein for tissue anabolism via IGF-I. Similarly, though more bone-related, Cadogan et al (44) found that supplementation in 12-y-old girls with 570 mL milk/d for 18 mo was associated with an increase in plasma IGF-I and bone mineral status compared with control subjects. As discussed by the authors (44) and summarized by Ginty (8), the higher protein content of milk could have mediated an increase in plasma IGF-I that, in turn, may have been stimulatory for osteoblast activity or promoted bone mineralization. Additionally, the daily amount of protein ingested may also influence calcium homeostasis together with parathyroid hormone secretion and 1,25-dihydroxy vitamin D status (45). However, in a recent study, parathyroid hormone and vitamin D status remained unaffected by corresponding changes in protein intake (46).
Until now, many studies concerning dietary protein intake on bone health focused on its potential negative effect. The primary assumed mechanism by which bone resorption may be increased in response to higher dietary protein intakes is the metabolic oxidation of the S-containing amino acids methionine and cysteine to H2SO4 with a consecutive reduction of blood pH (47). However, the acidifying effect of protein cannot be regarded as isolated, because other alkalizing nutrients (eg, potassium, magnesium) can counterbalance it. So far, there is only predominantly indirect evidence for such an acid-base homeostatic effect on bone: increasing intakes of fruit and vegetables, ie, alkali-forming foods (48, 49), or alkali-forming diets (50) decrease urinary calcium excretion. Additionally, observational, clinical, and intervention studies found a positive effect of alkali-forming foods, ie, fruit and vegetables, on bone health (51) in elderly (52, 53) and in early pubertal children (3, 54). Only one recent study associated directly the estimated dietary net acid production with indexes of bone health, finding that lower estimates of net endogenous non–carbonic acid production were correlated with higher bone mass and a tendency to less bone resorption in premenopausal and perimenopausal women (10). Because it has postulated and supported by measurements of serum bicarbonate and blood pH levels that the capability to excrete protons gradually decreases with age as the glomerular filtration rate drops (55, 56), our findings of a negative association of PRAL with bone variables, even during childhood and adolescence, when renal function should be near its optimum, are all the more remarkable. However, definite biochemical data on the age-dependency of the renal function in eliminating acidity have still to be established.
In line with our findings, Cadogan et al (44), who examined the effects of milk supplementation in 12-y-old girls, also found no association between calcium intake and bone variables. Although intervention trials in children and adolescents have regularly shown positive effects of calcium or dairy supplementation on bone mass acquisition (57–59), observational studies—especially those that have examined long bones (60, 61)—failed to detect associations.
One study with a comparable study design to ours exists; however, the results of the 2 studies are conflicting (62). In this study, neither positive nor negative relations between long-term protein intake and bone mineral densities at different sites were found. Several reasons may have accounted for this. First, different bone sites may be differently susceptible to metabolic influences (14). Second, the dual-energy X-ray absorptiometry (DXA) method used for the measurements may not have been accurate enough to specifically identify association with protein because it yields only a 2-dimensional projection (areal bone density), which tends to underestimate volumetric density in smaller and overestimate in larger subjects (1, 63). The PQCT method, however, provides a 3-dimensional assessment of the structural and geometric properties of the skeleton and thus allows a more sensitive measurement of bone quality (63, 64). In this context, the periosteal circumference and cortical density determined by pQCT more realistically reflect modeling and remodeling, respectively, than the corresponding variables calculated from DXA. This applies also to those variables reflecting a combination of modeling and remodeling.
The limitations of the current study also warrant mention. First, although weighed dietary records are regarded to be particularly reliable (65), some skepticism against the dietary assessment tool might remain. Second, a more specific analysis using individual data on methionine and cysteine intakes would be desirable, although in our present evaluation we did not see any association between bone variables and the intake of food groups with a higher content of sulfur-containing amino acids. Third, an assessment of serum IGF-I could support the hypothesis that bone anabolism by protein is driven by this hormone.
In conclusion, our data provide evidence of a positive link between long-term dietary protein intake and diaphyseal bone stability in healthy children and adolescents. Hereby, 2 seemingly contradictory mechanisms appear to be effective: an anabolic effect (probably mediated by IGF-I) on periosteal circumference, cortical area, bone mineral content, and strength strain index and a catabolic effect mediated by dietary acid load and characterizable by PRAL. A high PRAL, which indicates an inadequate intake of alkalizing minerals, can at least partly negate an osteotrophic protein effect. Our findings support the health benefit of a diet rich in base-yielding vegetables and fruit, which is in accordance with the "5-A-Day" campaign. In children, an adequate alkali intake should be achieved through appropriate nutrition, and only if this is not possible with alkalizing supplements, eg, potassium bicarbonate or citrate. These findings provide further evidence that an appropriate evaluation of dietary influences on bone health should involve an integrative approach, because a focus on single nutrients is not sufficient.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Related articles in AJCN:
This article has been cited by other articles:
|
|
 |
|
  L. Libuda, U. Alexy, T. Remer, P. Stehle, E. Schoenau, and M. Kersting Association between long-term consumption of soft drinks and variables of bone modeling and remodeling in a sample of healthy German children and adolescents Am. J. Clinical Nutrition, December 1, 2008; 88(6): 1670 - 1677. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. R Fenton, M. Eliasziw, A. W Lyon, S. C Tough, and D. A Hanley Meta-analysis of the quantity of calcium excretion associated with the net acid excretion of the modern diet under the acid-ash diet hypothesis Am. J. Clinical Nutrition, October 1, 2008; 88(4): 1159 - 1166. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. P. Thorpe, E. H. Jacobson, D. K. Layman, X. He, P. M. Kris-Etherton, and E. M. Evans A Diet High in Protein, Dairy, and Calcium Attenuates Bone Loss over Twelve Months of Weight Loss and Maintenance Relative to a Conventional High-Carbohydrate Diet in Adults J. Nutr., June 1, 2008; 138(6): 1096 - 1100. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Vormann and T. Remer Dietary, Metabolic, Physiologic, and Disease-Related Aspects of Acid-Base Balance: Foreword to the Contributions of the Second International Acid-Base Symposium J. Nutr., February 1, 2008; 138(2): 413S - 414S. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Remer Dietary calcium, the 16{alpha}-hydroxyl metabolic pathway of steroids, and sex hormones in blood and urine Am. J. Clinical Nutrition, January 1, 2008; 87(1): 192 - 193. [Full Text] [PDF]
|
|
|
|
|
|
M. Thorpe, M. C. Mojtahedi, K. Chapman-Novakofski, E. McAuley, and E. M. Evans A Positive Association of Lumbar Spine Bone Mineral Density with Dietary Protein Is Suppressed by a Negative Association with Protein Sulfur J. Nutr., January 1, 2008; 138(1): 80 - 85. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. A. Lanham-New The Balance of Bone Health: Tipping the Scales in Favor of Potassium-Rich, Bicarbonate-Rich Foods J. Nutr., January 1, 2008; 138(1): 172S - 177S. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Vatanparast, D. A. Bailey, A. D. G. Baxter-Jones, and S. J. Whiting The Effects of Dietary Protein on Bone Mineral Mass in Young Adults May Be Modulated by Adolescent Calcium Intake J. Nutr., December 1, 2007; 137(12): 2674 - 2679. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. A. Frassetto, S. A. Lanham-New, H. M. Macdonald, T. Remer, A. Sebastian, K. L. Tucker, and F. A. Tylavsky Standardizing Terminology for Estimating the Diet-Dependent Net Acid Load to the Metabolic System J. Nutr., June 1, 2007; 137(6): 1491 - 1492. [Full Text] [PDF]
|
|
|
|
|
|
A. A Welch, S. A Bingham, J. Reeve, and K. Khaw More acidic dietary acid-base load is associated with reduced calcaneal broadband ultrasound attenuation in women but not in men: results from the EPIC-Norfolk cohort study Am. J. Clinical Nutrition, April 1, 2007; 85(4): 1134 - 1141. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Z. Budek, C. Hoppe, K. F. Michaelsen, S. Bugel, and C. Molgaard Associations of Total, Dairy, and Meat Protein with Markers for Bone Turnover in Healthy, Prepubertal Boys J. Nutr., April 1, 2007; 137(4): 930 - 934. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Misra, P. Tsai, E. J Anderson, J. L Hubbard, K. Gallagher, L. A Soyka, K. K Miller, D. B Herzog, and A. Klibanski Nutrient intake in community-dwelling adolescent girls with anorexia nervosa and in healthy adolescents. Am. J. Clinical Nutrition, October 1, 2006; 84(4): 698 - 706. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Berkemeyer and T. Remer Anthropometrics Provide a Better Estimate of Urinary Organic Acid Anion Excretion than a Dietary Mineral Intake-Based Estimate in Children, Adolescents, and Young Adults J. Nutr., May 1, 2006; 136(5): 1203 - 1208. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||
); the mean (±SD) percentages of protein from these food groups were 61.7 ± 3.6%, 52.4 ± 2.6%, and 45.5 ± 3.9%, respectively.