The authors state that they have no conflicts of interest.
High-Protein Intake Enhances the Positive Impact of Physical Activity on BMC in Prepubertal Boys†
Article first published online: 17 SEP 2007
Copyright © 2008 ASBMR
Journal of Bone and Mineral Research
Volume 23, Issue 1, pages 131–142, January 2008
How to Cite
Chevalley, T., Bonjour, J.-P., Ferrari, S. and Rizzoli, R. (2008), High-Protein Intake Enhances the Positive Impact of Physical Activity on BMC in Prepubertal Boys. J Bone Miner Res, 23: 131–142. doi: 10.1359/jbmr.070907
- Issue published online: 4 DEC 2009
- Article first published online: 17 SEP 2007
- Manuscript Accepted: 12 SEP 2007
- Manuscript Revised: 13 AUG 2007
- Manuscript Received: 6 JUL 2007
- bone growth;
- prepubertal boys;
- physical activity;
- calcium intake;
- protein intake;
- recommended dietary protein allowances
In 232 healthy prepubertal boys, increased physical activity was associated with greater BMC at both axial and appendicular sites under high-protein intake.
Introduction: Physical activity is an important lifestyle determinant of bone mineral mass acquisition. Its impact during childhood can be modulated by nutrition, particularly by protein and calcium intakes. We analyzed the relationship between physical activity levels and protein compared with calcium intake on BMC.
Materials and Methods: In 232 healthy prepubertal boys (age: 7.4 ± 0.4 [SD] yr; standing height: 125.7 ± 5.9 cm; body weight: 25.3 ± 4.6 kg), physical activity and protein and calcium intakes were recorded. BMC was measured by DXA at the radial metaphysis, radial diaphysis, total radius, femoral neck, total hip, femoral diaphysis, and L2-L4 vertebrae.
Results: In univariate analysis, the correlation coefficients r with BMC of the various skeletal sites were as follows: physical activity, from 0.26 (p = 0.0001) to 0.40 (p = 0.0001); protein intake, from 0.18 (p = 0.005) to 0.27 (p = 0.0001); calcium intake, from 0.09 (p = 0.181) to 0.17 (p = 0.007). By multiple regression analysis, the β-adjusted values remained correlated with BMC, ranging as follows: physical activity, from 0.219 (p = 0.0007) to 0.340 (p < 0.0001); protein intake, from 0.120 (p = 0.146) to 0.217 (p = 0.009). In contrast, it was not correlated for calcium intake: from −0.069 (p = 0.410) to 0.001 (p = 0.986). With protein intake (mean = 2.0 g/kg body weight/d) above the median, increased physical activity from 168 to 321 kcal/d was associated with greater mean BMC Z-score (+0.6, p = 0.0005). In contrast with protein intake (mean = 1.5 g/kg body weight/d) below the median, increased physical activity from 167 to 312 kcal/d was not associated with a significantly greater mean BMC Z-score (+0.2, p = 0.371). The interaction between physical activity and protein intake was close to statistical significance for mean BMC Z-score (p = 0.055) and significant for femoral neck BMC (p = 0.012). In keeping with the results derived from multiple regression analysis, the increased physical activity on mean BMC Z-score was not influenced by difference in calcium intake above (mean = 945 mg/d) and below (mean = 555 mg/d) the median.
Conclusion: In healthy prepubertal boys, the impact in increased physical activity on BMC seems to be enhanced by protein intake within limits above the usual recommended allowance.
Among environmental factors, physical activity and nutrition are key players for the acquisition of bone mass during growth. Growing bones are usually more responsive to mechanical loading than adult bones.(1) Physical activity increases bone mineral mass accumulation in children and adolescents. However, the impact seems to be stronger before than during or after the period of pubertal maturation.(2) Among nutrients that can specifically interact with bone metabolism, the influence of calcium has been extensively studied from infancy to the end of pubertal maturation.(3-5) Much less consideration has been given to protein intake, although this macronutrient is essential for bone accumulation during growth, as well as maintenance of the skeletal structural integrity throughout the adult life.(6)
The recommended dietary allowances (RDAs) for proteins in healthy children and adolescents have mainly been derived from nitrogen balance studies of short duration carried out in adults.(7-9) From these adult data, a factorial procedure was used so that values were assigned for maintenance requirements and an increment representing growth.(7-9) This latter estimate was increased for use efficiency and unevenness of daily growth rate, taking into account the inability of infants and children to store amino acids against intermittent needs.(7-9) The average requirements obtained by this factorial procedure were still raised by 2 SD (25%) in consideration of individual variability. Thus, in children 5-10 years of age, this derivation of daily protein allowances has been theoretically estimated to be ∼1.0 g/kg body weight/d.(7-9) Whether this level of protein intake is optimal for bone mineral mass accumulation in children is not known. In this report, we analyze in a homogenous cohort of healthy prepubertal boys the interaction between physical activity and protein compared with calcium intakes on bone mineral mass at several sites of the skeleton. This study strongly suggests that the positive effect of physical activity on bone mass acquisition before the onset of pubertal maturation is enhanced by protein intake higher than the usual recommended dietary allowance.
MATERIALS AND METHODS
The analysis presented in this report has been carried out on data obtained in 232 boys before their entry into an intervention trial and their splitting into two groups for testing the effect of calcium supplementation.(10)
Healthy prepubertal white boys with a mean age of 7.44 ± 0.39 (SD) yr (range: 6.5-8.5 yr) were recruited through the Public Health Youth Service of the Geneva region from September 1999 to September 2000. The protocol was approved by the Ethics Committee of the Department of Pediatrics of the University Hospitals of Geneva. Informed consent was obtained from the parents and their children. The following exclusion criteria were applied: ratio weight/height <3rd or >97th percentile according to Geneva reference values; presence of physical signs of puberty; chronic disease, gastrointestinal disease with malabsorption, congenital or acquired bone disease; and regular use of medication.
Protein and calcium intakes assessment
Protein and calcium intakes were assessed by frequency questionnaire.(11,12) The total animal protein intake was expressed either in grams per day or grams per kilogram body weight per day. It included dairy, meat, fish, and egg proteins. The calcium intake was essentially assessed from dairy sources.
Physical activity assessment
Physical activity was assessed by questionnaire based on self-reported time spent on physical education classes, organized sports, recreational activity, and usual walking and cycling.(13) Subsequently, the collected data were converted and expressed as physical activity energy expenditure (PAEE kcal/d) using established conversion equations.(14)
Measurement of anthropometric and bone variables
Participant's weight and standing height using a stapediometer were measured, and body mass index (BMI = body weight/standing height2) was calculated. BMC and areal BMD (aBMD) were determined at the radial metaphysis, radial diaphysis, total radius, lumbar spine (L2-L4), femoral neck, total hip, and femoral diaphysis by DXA using a Hologic QDR 4500 instrument (Waltham, MA, USA) and according to the manufacturer's software. The coefficient of variation (CV%) in vivo ranged from 1.0% to 1.6% for BMD and from 0.3% to 3.0% for BMC in young healthy adults.
Expression of the results and statistical analysis
Pearson's correlation coefficient r was calculated for the relationships between the physical activity, protein and calcium intakes, and BMC of the skeletal sites. The variance of BMC r2 accounted for by the three independent variables (i.e., physical activity and protein and calcium intakes) was calculated by multiple regression analysis. This computation provided β-adjusted values, as an estimate of the respective contribution by the three independent variables to the prediction of BMC. The BMC data were also analyzed after segregation according to the median of physical activity and protein and calcium intakes by ANOVA analysis. The values of BMC recorded at the skeletal sites were expressed in both absolute terms and in SD score (Z-scores) as computed from the measurements made on the whole cohort. A two-way ANOVA was used to analyze the interaction between protein or calcium intake and the influence of physical activity on BMC. The significance level for two-sided p values was 0.05 for all tests. The data were analyzed using STATA software, version 7.0.
The anthropometric variables of the 232 healthy prepubertal boys (Table 1) are within the reference values obtained in a population of similar ethnic and socioeconomic background.(15) The daily protein intake from animal food source is expressed both in absolute terms and related to body weight. With a mean of 1.78 g/d, this value is well above the allowance of reference protein intakes that are set for children of the same age category at 0.9 and 1.0 g/d in the French(7) and U.S.(8) recommendations, respectively. About one half of the animal protein intakes were provided by dairy products and the other half by meat, fish, and egg sources (Table 1). In this food group, ∼75% of the proteins come from meat consumption. On average, 70% of recorded total physical activity was estimated resulting from weight-bearing exercises. In this homogenous cohort of healthy boys, the CVs (CV = SD/mean × 100) of BMC measured in the various skeletal sites ranged from 12.9% (radial diaphysis) to 20.3% (femoral diaphysis). This variability in BMC is 3- to 5-fold larger than that computed for standing height, of which the CV was 4.7%.
Relationships between physical activity, protein intakes, calcium intakes, and BMC
Univariate analysis indicated that, for all measured skeletal sites, the most significant correlation was with total physical activity, followed by protein intake (Table 2). The weakest association was with calcium intake, particularly in its relation with femoral neck BMC, where the coefficient of correlation did not reach statistical significance, in contrast to physical activity and protein intake. Overall, the correlation coefficients were similar for dairy and meat-fish-egg proteins, but both were lower than that calculated for the total protein intake (data not shown). At all skeletal sites, total physical activity was somewhat better correlated than weight-bearing activity (Table 2). Total, rather than weight-bearing, physical activity was taken into account in the following analysis in relation with protein and calcium intakes.
Multiple regression analysis revealed that total physical activity was the strongest predictor of BMC recorded at the various scanned sites (Table 3). Nevertheless, protein intake remained a statistically significant predictor of bone mass in five of the seven scanned sites (Table 3). In contrast, at none of the seven sites did calcium intake remain predictive of BMC after adjustment for both physical activity and protein intakes (Table 3).
Distribution according to the median of protein or calcium intake and physical activity
The analysis showed that more intense physical activity (312 versus 167 kcal/d) under relatively low (38.6 g/d) protein intake was not associated with increased BMC (Table 4) or aBMD (Table 6) at any of the measured skeletal sites. In sharp contrast, a similar increase in physical activity (321 versus 168 kcal/d) under relatively high protein intake (55.9 g/d) was associated with very significant greater BMC at all measured skeletal sites (Table 4) and significant greater aBMD at several skeletal sites (Table 6).
More intense physical activity under relatively high (945 mg/d) or low (556 mg/d) calcium intakes was associated with increased BMC (Table 5) or aBMD (Table 6) of similar magnitude at several sites of the skeleton.
Figures 1A and 1B show the differential impact of increased physical activity in relation with variations in protein compared with calcium intakes. In these figures, the mean of the BMC data presented in detail in Tables 4 and 5 are expressed as Z-scores. The response to increased physical activity was markedly influenced by the protein intake: Δ Z-score of +0.15 and +0.64 for below and above the median of protein intake, respectively (Fig. 1A). In contrast, the response to increased physical activity was barely influenced by calcium intake: Δ Z-score of +0.43 and +0.51 for below and above the median of calcium intake, respectively (Fig. 1B). The data shown in Figs. 1A and 1B were still analyzed by ANOVA for an interaction between physical activity and either protein or calcium intakes. Whereas the interaction with physical activity was very close to statistical significance (p = 0.055) with protein intake, no interaction was found with calcium intake (p = 0.754). The statistical analysis of the interaction at the various skeletal sites between physical activity and protein and calcium intakes are indicated in Tables 4 and 5, respectively.
The impact of physical activity under high-protein intake on mean BMC Z-score (Fig. 1A) (+10.6%, p = 0.004) was associated with lower increment in both the corresponding scanned area (AREA, cm2: +6.5%, p = 0.0008) and BMD (aBMD, mg/cm2: +3.2%, p = 0.018). Similar results were observed at each skeletal site; those obtained at the femoral neck level are shown in Fig. 2. Thus, the response to increased physical activity at the femoral neck markedly differed whether protein intake was below or above the median (Δ Z-score BMC: +0.01 versus +0.66; Δ Z-score AREA: +0.05 versus +0.59; Δ Z-score aBMD: −0.01 versus +0.39). The greater AREA in response to increased physical activity under high- compared with low-protein intake reflects a larger width of the femoral neck, because for all scans, the height of the region of interest parallel to the femoral axis was constant. Analyzed by ANOVA, the interaction between physical activity and protein intake was p = 0.012 at the FN BMC, p = 0.040 at the FN area, and p = 0.132 at the FN aBMD.
Predictors of bone mineral mass in prepubertal boys
The foregoing analysis indicates that, in healthy prepubertal boys, two environmental or lifestyle factors, namely physical activity and protein intake, are positively associated with bone mineral mass at most measured skeletal sites. In multivariate regression analysis, physical activity with protein intake would account from 9% to 18% of the variance in BMC measured at the various skeletal sites. Calcium intake, which is also significantly associated with BMC in univariate analysis, does not seem to be an independent predictor when physical activity and protein intake are entered into the multiple regression equation. The apparent greater beneficial “effect” of physical activity than nutrition on bone mass acquisition has been suggested in two previous studies(16,17) on cohorts including both females and males and a wider age range (6-18 and 7-20 yr) than the foregoing report and consequently various stages of pubertal maturation. As in our study, the calcium intake did not remain an independent predictor of BMC.(16,17)
These observations do not mean that calcium is not an important nutrient in bone mineral mass accrual. Several randomized placebo-controlled trials tested the effect of calcium supplementation on bone mineral mass acquisition from before to after puberty.(10,18-28) However, overall, positive responses seem to be dependent on the amount of calcium spontaneously ingested, the stage of pubertal maturation, and the skeletal site examined. In our study, as in other similar reports,(16,17) calcium intake might not have been low enough to remain an independent predictor of BMC.
During childhood and/or adolescence, positive relationships were found between increased calcium intake from dairy foods and bone mineral mass acquisition.(29-33) The associated increase in protein intake from dairy foods may have substantially contributed to such positive relationships.(29-33) Likewise, in the often-quoted study carried out in two Yugoslav populations, the difference in bone mass in young adults was usually ascribed to calcium intake.(34) However, both protein intake and physical activity were also positively associated to bone mass.(34) The reasons why protein intake was not considered as an essential nutrient for calcium economy and bone health may stem from longterm prejudice or “belief,”(35) as well as data misinterpretation, resulting in questionable claims against dietary proteins, particularly against those provided by animal foods.(6,36)
The positive influence of increased physical activity was quite similar under relatively low (mean Δ Z-score = +0.43) and high (mean Δ Z-score = +0.51) calcium intake (Fig. 1B). In infants, children, and adolescents, physical activity seems to interact with dietary calcium supplementation.(10,17,37-40) A wider range in both physical activity level and calcium intake than recorded in our study may be required to observe a substantial positive interaction between the two factors.(39)
Recorded protein intakes compared with RDAs for children
According to several national and international agencies, protein requirements for children in the age range of our cohort are estimated to be 0.9-1.0 g/kg body weight/d.(7-9,41) The mean protein intakes of the healthy prepubertal boys investigated in this study amounted to nearly 80% above this dietary reference value. This level is not exceptional, because a survey of protein intakes in French children was found to surpass 2.0 g/kg body weight/d.(7) In a recent German study, it was also found to be 2.0 g/kg body weight/d in 8- to 9-yr-old healthy prepubertal boys and girls.(16) It was ∼55 g/d in a cohort of healthy North-American girls and boys 8.1 years of age with a mean body weight of 29 kg (i.e., ∼1.9 g/kg body weight/d).(42) In an Australian study, a protein intake of similar magnitude was recorded in 11.5-yr-old monozygotic twins: 75 g/d, with mean body weight of 39 kg (i.e., again ∼1.9 g/kg body weight/d).(17)
In a previous cohort also living in the Geneva district, macronutrient intakes were assessed from a 5-day diary, with all foods eaten being carefully quantified.(43) With the use of this method, animal protein intakes of prepubertal boys 9-10 yr of age were 54 g/d for a mean body weight of 32 kg.(43) This corresponds to a mean value of 1.7 g/kg body weight/d, which is an intake quite similar to that recorded in this study carried out ∼10 years later. Note that in this previous accurate dietary survey, the mean intake of vegetal proteins was 25 g/d, corresponding to about one third of the total protein intakes that amounted to 79 g/d (i.e., 2.5 g/kg body weight/d).(43)
Optimal quantity of protein for growing children
In a recent review, it was underscored that the scientific literature is essentially void of studies directed specifically at quantifying protein needs of healthy children between 8 and 12 yr of age.(44) Current recommendations for daily protein intakes in children remain speculative.(44) The RDAs(7-9) are estimates derived from interpolation of requirements determined in infants and young adults.(41,44,45) The spontaneous consumption of healthy prepubertal children 7-11 yr of age is well above the RDAs. Concern has been frequently expressed about the putative adverse effect of high-protein intakes. In adults, arguments against claims on putative detrimental effects of animal protein intakes on bone health have been developed.(36) In our cohort of prepubertal boys, a consumption of 1.7 g of animal proteins compared with the recommended allowance of 0.9-1.0 g/kg body weight/d for total proteins was not associated with any apparent detrimental effect in this cohort of healthy, quite physically active, prepubertal boys. Even protein intakes above the median were associated with greater bone mass gain. However, this observation can by no means be taken as a scientific basis to reconsider the widely accepted recommendation of protein intakes for children. On the other hand, it should not promote public health initiatives aimed at reducing the spontaneous intakes of protein to the level of the recommended allowances by national and international agencies.
Biological significance of the impact of high-protein intake in response to increased physical activity
Our results suggest that above the dietary reference value of 1.0 g/kg body weight/d, there is still a positive impact on BMC of high-protein intake on the response to increased physical activity at several sites of the skeleton. Nevertheless, the question arises of whether the observed protein intake-related difference in BMC is of biological significance. A difference in bone mineral mass by 1.0 SD is supposed to alter the risk of fragility fractures by 50%.(46-50) Increased physical activity in the presence of a mean protein intake of ∼2.0 instead of 1.5 g/kg body weight/d is associated with higher mean BMC by +0.64 instead of +0.15 Z-score (Fig. 1A). This difference of +0.49 SD may be translated in a substantial positive shift of the BMC trajectory, resulting in a higher peak bone mass and possibly in a reduction of fragility fracture risk in adulthood.
Putative mechanisms of protein intake on bone mineral mass in prepubertal children
Experimentally, it has been clearly shown by controlling for calcium and energy intake that low-protein intake exerts a severe negative impact on bone by both reducing osteoblastic formation and increasing osteoclastic resorption.(6,51,52) Clinically, correction of low-protein intake in patients with hip fracture attenuates consecutive bone loss, increases muscle strength,(53) and probably by reducing medical complications, diminishes the duration of hospitalization.(54,55) The influence of protein intake on bone and muscle mass and strength is associated in both animal and human studies by changes in the plasma level of IGF-I.(52,53) It is possible that the positive association of physical activity and protein intake on BMC is mediated in part by the stimulation of IGF-I, which in turn would impact on both skeletal muscle and bone.(56) The increase of width at femoral neck (Fig. 2) under high-protein intake and increased physical activity would be compatible with their impact on periosteal apposition.
The methods for recording dietary intakes and physical activity are not optimal in terms of accuracy. Methods using accelerometers and/or heart rate recordings can objectively assess the levels of physical activity and provide accurate data. Likewise, several diary surveys with qualitative and quantitative assessments of all foods consumed during several days and taking into account seasonality variations would certainly provide more accurate results than those collected from frequency questionnaires. Nevertheless, the questionnaire method could still provide reliable data when comparing boys of the same age category and living in the same community. Note that the frequency questionnaire for assessing the protein intake was validated against two reference methods, including diary with weighing and recording all foods and beverages over a 4-day period.(12) Furthermore, the questionnaires were filled under the direct supervision of the same certified dietitian for all boys of the investigated cohort.
Our observation with its scientific limitation of being at this stage an association without evidence for a causal relationship should also stimulate the design of randomized-controlled trials aimed at testing the impact of physical activity and protein intakes on bone and muscle development, as well as on other growth health components in children.
The authors thank Muriel Füeg-Clarisse, certified dietitian, for the assessment of food intakes and having carried out this study; François Herrmann, MD, MPH, for help with statistical analysis; Helena Francisco for administrative help; Marianne Perez for secretarial assistance; and Dr Denis Barclay, PhD, from the Nestlé Research Center, Lausanne, Switzerland. We are indebted to the Geneva Public Youth Service for the recruitment of the subjects, to the team of the bone densitometry unit, and to Prof S Suter, MD, and Prof D Belli, MD, chairpersons of the Department of Pediatrics, for constant and invaluable support in this research project. This study was supported by the Swiss National Science Foundation (Grants 32-49757-96 and 32-58962.99), Nestec, Lausanne, Switzerland, and Institute Candia, Ivry sur Seine, France.
- 72001 Apports Nutritionnels Conseillés pour la Population Française, 3rd ed. TEC&DOC, Paris, France.
- 8National Research Council 1989 Recommended Dietary Allowances, 10th ed. National Academy Press, Washington, DC, USA.
- 9World Health Organization 1985 Energy and Protein Requirements. Report of a Joint FAO/WHO/UNO Expert Consultation, vol. 724. World Health Organization, Geneva, Switzerland.
- 151972 Les Mensurations Normales de la Période de Croissance Tables Scientifiques, 7th ed. Ciba-Geigy SA, Bâle, Switzerland.,