Dietary intake of Brazilian black and white men and its relationship to the bone mineral density of the femoral neck

ABSTRACT CONTEXT AND OBJECTIVE: Osteoporosis and fragility fractures are an important public health problem. Although bone loss occurs with age universally, the incidence of bone loss fractures varies greatly between racial groups. The aim of this study was to examine the relationship between calcium, protein and energy intake and the bone mineral density of the femoral neck in Brazilian black and white men. DESIGN AND SETTING: This was a cross-sectional study, carried out in a teaching hospital in São Paulo. METHODS: The participants were 277 volunteer men, aged 50 years or older. The bone mineral density of the femoral neck (FNBMD) was measured by dual energy x-ray absorptiometry. The relationship between FNBMD and calcium, protein and energy intake, as assessed by a three-day food record, was analyzed using multiple linear regression models and was adjusted for age, height, physical activity and education level. The analysis was stratified by race (white and black). RESULTS: FNBMD presented similar means in the two racial groups (p = 0.538). Protein and energy intake did not show a significant correlation with FNBMD, either in the white or in the black population. Calcium intake showed a strong and independent correlation with FNBMD in the black men (partial r = 0.42). CONCLUSION: Calcium intake was a determinant of FNBMD for black men, aged 50 years or older, but not for the white ones.


INTRODUCTION
Osteoporosis is an important public health problem because of the significant morbidity and mortality associated with bone fractures and the high cost of treatment. 1steoporosis and fragility fractures in men constitute a considerable health care burden. 2 The risk of fracture is directly associated with bone mineral density (BMD). 3Bone loss is expected with age, 4,5 and it dramatically increases the incidence of hip and vertebral fractures in both men and women. 6The frequency of occurrence of bone fractures in South America is lower than in other regions of the world; however, it is possible that this problem is becoming intensifi ed due to the increasing life expectancy of the population on this continent. 7lthough bone loss occurs with age for everybody, the incidence of bone loss fractures varies greatly between racial groups. 8Higher bone mass has been observed in black individuals than in white individuals. 9,10alcium intake is a signifi cant determinant of BMD.In countries with high osteoporotic fracture incidence, low calcium intake among older men and women is associated with increased fracture risk. 11On the other hand, inadequate protein intake, when too high or too low, increases the risk of osteoporotic fractures. 6,11Low body weight is related to bone mass decline and to increased risk of osteoporotic fractures. 12It should be noted that the majority of studies on dietary intake, bone mass and osteoporosis have been carried out in developed countries, where the population's lifestyle and food consumption are very different from those of developing nations.Moreover, although it is known that there are important differences in bone mineral density between black and non-black individuals, the study of dietary intake differences in racial groups has received little attention.

OBJECTIVE
The aim of the present study was to determine the relationship between calcium, protein and energy intake and the bone mineral density of the femoral neck (FNBMD) in Brazilian men aged 50 years and older, comparing black and white individuals.

Population Population
This was a cross-sectional study involving volunteer men.From February to August 1997, we recruited 307 healthy male subjects aged 50 years or older, all resident in the city of São Paulo (southeastern Brazil).They were men whose wives had been referred by their primary care physicians for a bone densitometry scan, or who had responded to a newspaper advertisement.
We excluded from the study men using any kind of medication or having medical conditions that could affect bone metabolism: there was one individual with cancer of the colon, one with multiple myeloma, one with hyperthyroidism and one with rheumatoid arthritis.Twenty-two individuals were excluded due to the lack of bone densitometry data and 25 individuals were excluded because they did not answer the dietary questionnaire.We also excluded the mulattos (n = 29) in order to better separate the white and black races, thus avoiding confusions caused by the effects of racial miscegenation or mistakes in racial classifi cation.
Thus, the fi nal sample consisted of 227 men.None were taking energy, calcium or protein supplements.
The present study was approved by the Ethics Committees of Heliópolis Hospital and the School of Public Health of the University of São Paulo, Brazil.All the participants provided informed written consent.
CONTEXT AND OBJECTIVE: Osteoporosis and fragility fractures are an important public health problem.Although bone loss occurs with age universally, the incidence of bone loss fractures varies greatly between racial groups.The aim of this study was to examine the relationship between calcium, protein and energy intake and the bone mineral density of the femoral neck in Brazilian black and white men.

DESIGN AND SETTING:
This was a cross-sectional study, carried out in a teaching hospital in São Paulo.

METHODS:
The participants were 277 volunteer men, aged 50 years or older.The bone mineral density of the femoral neck (FNBMD) was measured by dual energy x-ray absorptiometry.The relationship between FNBMD and calcium, protein and energy intake, as assessed by a three-day food record, was analyzed using multiple linear regression models and was adjusted for age, height, physical activity and education level.The analysis was stratifi ed by race (white and black).
RESULTS: FNBMD presented similar means in the two racial groups (p = 0.538).Protein and energy intake did not show a signifi cant correlation with FNBMD, either in the white or in the black population.Calcium intake showed a strong and independent correlation with FNBMD in the black men (partial r = 0.42).

CONCLUSION:
Calcium intake was a determinant of FNBMD for black men, aged 50 years or older, but not for the white ones.days, including one weekend day (Saturday or Sunday).Once the data was collected, the food intake registered was converted into energy and nutrients using the Virtual Nutri program. 13

Anthropometric assessment Anthropometric assessment
Weight (kg) was measured using a standardized balance-beam scale and height (cm) using a stadiometer, and the body mass index (BMI) was calculated as the ratio between weight (kg) and the square of height (m 2 ).

Other variables Other variables
The participants answered a questionnaire about sociodemographic characteristics, including age (years), education level (years) and race (white, black, mulatto and Asian).
The information was self-reported and collected by a research assistant.
Concerning physical activity, three scores of physical activity were calculated (leisure, locomotion and occupation) using the questionnaire proposed by Baecke et al., 14 which had been adapted and validated for the adult Brazilian population. 15For more details regarding the construction of the physical activity scores, refer to Florindo et al. 16

Statistical analysis Statistical analysis
The nutrient intake (calcium and protein) was adjusted for the total energy consumption using the residual nutrient method. 17In summary, this method allows analysis of the net effect of the nutrient without the infl uence of the energy intake.Energy intake was adjusted for height.
We compared the means of FNBMD, calcium, protein and energy intake, height, weight, body mass index, age, physical activity scores and education level according to white and black race using the Student t test.The Kolmogorov-Smirnov test was conducted to assess goodness of fi t to normal distribution and only the calcium intake was transformed to log values.The relationship between nutrient (calcium and protein) and energy intakes and FNBMD was analyzed using Pearson's correlation coeffi cient and multiple linear regression models, stratifi ed by white and black race.FNBMD was considered to be the dependent variable and calcium, protein and energy intakes were the independent variables.
The control variables were: age (in years), height (in centimeters), education level (years) and physical activity (scores).The models were estimated by the stepwise forward procedure.The control variables were kept in the multiple regression model if they were statistically significant (p < 0.05) or if they adjusted the regression coeffi cients of the variables that had already been kept in the model by at least 10%.The correlation coeffi cients (r i ) and the regression coeffi cients (β i ) of the dietary variables were corrected by between and within-subject variation, in accordance with Beaton et al. 18 The effective p value for observations to be considered statistically signifi cant was 0.05.Data were analyzed using the SPSS (Statistical Package for the Social Sciences) statistical software system.

RESULTS
Most of the men (89%) were classifi ed as white (n = 246).We observed that the black individuals (n = 31) were relatively younger and presented lower education levels (2.14

Bone mineral density Bone mineral density assessment assessment
FNBMD was measured by dual energy xray absorptiometry (DEXA), using the Lunar 3.6z software (Lunar, Madison, Wisconsin, United States).We reported BMD as grams per square centimeter.The coefficient of within-subject variation for DEXA measurements was 1.5% for the femoral neck.The time taken to perform the examination was fi ve minutes and the radiation dose emitted was minimal (0.01 µΣ/h).

Dietary assessment Dietary assessment
Calcium, protein and energy intake were assessed using the food record method for three days.Food intake was recorded at all the daily meals, over three non-consecutive  *nutrient intake adjusted for the energy intake by using regression analysis; † Correlation corrected for the between and withinsubject variation in nutrient intake in the racial category; ‡ Correlation adjusted for the height (r partial) and corrected for the between and within-subject variation in energy intake in the racial category.
versus 4.41 years; p = 0.007), lower calcium intake (557 versus 720 mg; p = 0.020) and higher physical activity scores at work (2.95 versus 2.72 score units; p = 0.028) in comparison with the white individuals, as shown in Table 1.The other variables, including FNBMD (p = 0.538), presented similar means in the two racial groups.
There was no correlation between calcium, protein and energy intake and FNBMD in the white men.For the black men, calcium (r = 0.567; p < 0.001) and protein (r = 0.359; p = 0.040) intake showed a signifi cant positive correlation with FNBMD.Age showed a signifi cant negative correlation for both white and black individuals (Table 2).
For the black men (Table 4), we observed that the determiners of BMD were calcium intake and age.We found that, among the dietary intake variables, only calcium had an independent effect (partial r = 0.42; p = 0.002).Neither protein and energy consumption, nor height, physical activity scores and education level were determiners of FNBMD.Although the education level was not a determiner of BMD in the multiple model (p = 0.586), nor did it adjust the regression coeffi cient values of the dietary variables, we decided to keep this variable in the fi nal multiple model in order to control for educational differences between the racial groups.We decided to keep height in order to adjust the energy effect, and also as a control variable in relation to bone mass.However, the effect of the adjustment for the height on the energy intake regression coeffi cient was small (before the adjustment: β = 0.00008; p = 0.120; and after the adjustment: β = 0.00007; p = 0.160; data not shown in the table).
Thus, we found that, despite the small number of black individuals (n = 31), the calcium intake correlated with FNBMD only in black men, both in Pearson's correlation analysis and in the multiple regression models.

DISCUSSION
This was the fi rst study to analyze the relationship between nutrient intake and Table 3.Multiple linear regression model for bone mineral density of the femoral neck (g/cm 2 ) in Brazilian white men (n = 246) *nutrient intake adjusted for the energy intake by using regression analysis; † β corrected for the between and within-subject variation in nutrient intake in the white men; PA = physical activity.*nutrient intake adjusted for the energy intake by using regression analysis; † β corrected for the between and within-subject variation in nutrient intake in the black men.
bone mineral density in men from different races living in Latin America.We noticed that FNBMD did not differ between the racial groups and the same was true for weight, height and BMI.][10] Higher BMD in black individuals may be associated with higher obesity and muscular mass rates. 19his was not observed in the white and black men who participated in this study, since they presented similar mean weight and BMI.White race is considered to be a risk factor for low BMD, and little is known about racial differences that affect calcium intake.We found that the white individuals consumed more calcium than the blacks.Wang et al. studied young Americans of both sexes and found a lower relationship between calcium and energy in blacks than in whites and Hispanics. 10e found out that calcium intake was a determinant of FNBMD only for the black men.Other studies conducted among adult men found a relationship between current calcium intake and bone mass. 20,21However, the effect of calcium intake on bone mass is still controversial.In contrast, calcium intake was not associated with whole body, hip and lumbar spine BMD in a study carried out among adult men in Canada. 22oth in the white and black men, protein and energy intake were not correlated with FNBMD in our study.Wang et al. 10 also found no correlation between calorie, protein and calcium intake and FNBMD in Americans from different racial groups. 10ssessing another bone site, Whiting et al. 22 found that, after energy adjustment, protein intake was a signifi cant predictor of lumbar spine BMD in adult men. 22In analyzing the independent relationship between energy consumption and BMD, the action of body mass as a confounding variable needs to be considered.It is known that increased energy intake promotes increased weight and, in turn, this weight acts mechanically on the skeleton, thereby increasing BMD. 21Therefore, the

Table 1 .
Mean and standard deviation of study variables, according to white and black race *nutrient intake adjusted for the energy intake by using regression analysis; FNBMD = bone mineral density of the femoral neck; BMI = body mass index; PA = physical activity, SD = standard deviation.

Table 2 .
Correlation coeffi cient (r) between bone mineral density of the femoral neck (g/cm 2 ) and other study variables, according to white and black race

Table 4 .
Multiple linear regression model for bone mineral density of the femoral neck (g/cm 2 ) in Brazilian black men (n = 31)