Racial Differences in Bone Mineral Density in Older Men

Abstract

Studies have examined factors related to BMD in older white, but not black, men. We measured BMD in older white and black men and examined factors related to racial differences in BMD. Black men had significantly higher adjusted BMD at all sites. These results may explain, in part, the lower incidence of fractures in older black men.

Introduction: Several studies have examined factors associated with bone mineral density (BMD)in older men. None, however, have had sufficient numbers of black men to allow for meaningful comparisons by race.

Materials and Methods: A total of 503 white and 191 black men aged 65 and older(75.1 ± 5.8 and 72.2 ± 5.7 years, respectively) were recruited from the Baltimore metropolitan area. All men completed a battery of self-administered questionnaires, underwent a standardized examination, and had BMD measured at the femoral neck, lumbar spine, and total body. Data were analyzed using multiple variable linear regression models, adjusted for potential confounding variables; two-way interactions with main effects were included in models where appropriate.

Results: Black men had significantly higher adjusted BMD at the femoral neck (difference 0.09 [95% CI: 0.07, 0.12] mg/cm²), lumbar spine (0.07 [0.04, 0.10] mg/cm²), and total body (0.06 [0.03, 0.08] mg/cm²) than white men.

Conclusions: Older black men have significantly higher BMD than older white men, even after adjustment for factors associated with BMD. These differences, especially at the femoral neck, may explain the reduced incidence of hip fracture in black compared with white men.

INTRODUCTION

OSTEOPOROSIS IN MEN is recognized to be a significant public health issue,^(1–4) and low bone mineral density (BMD) has been shown to be a strong risk factor for hip fracture in men.^(1–5) Men accounted for 28% of the 1.7 million hip fractures that occurred worldwide in 1990.⁽⁶⁾ With the improving longevity of men and the increasing size of the older population, the number of men with hip fracture worldwide is estimated to reach 6.8 million in 2050.⁽⁶⁾ Of all the fractures, hip fractures account for the greatest morbidity and mortality, especially in men.^(7–10) The 1-year mortality rate after hip fracture is estimated to be 31% (95% CI, 27–35%) for men compared with 17% (95% CI, 15–19%) for women.⁽¹⁰⁾ Data from Europe suggest that the prevalence of vertebral deformity, possibly secondary to vertebral compression fractures, may be higher in men than in women.⁽¹¹⁾

The prevalence of osteoporosis and low BMD among men and women in the United States was estimated based on data from the Third National Health and Nutrition Survey (NHANES III; 1988–1994).⁽¹²⁾ Using non-Hispanic white men aged 20–29 years as the reference group and the World Health Organization's recommended criteria for the definition of osteoporosis in white women,^(13,14) white men 50 years and older had a prevalence of osteoporosis of 7% and black men in the same age group had a prevalence of 5%.⁽¹⁵⁾ The prevalence of osteopenia using the same reference group was almost 50% in white men and 30% in black men.^(12,15) These studies, however, did not adjust for other risk factors shown to be associated with osteoporosis in men and women. Consistent with their lower prevalence of osteoporosis, black men have a lower rate of hip fractures than white men.⁽¹⁶⁾

To investigate differences in bone mass and fracture risk between white and black men, we established a cohort of older men who will be followed longitudinally. Reported herein are analyses of baseline data comparing BMD at the femoral neck, lumbar spine, and total body by race after adjusting for factors associated with these outcomes.

MATERIALS AND METHODS

Participants

From July 2000 through July 2001, 694 men (503 white and 191 black) aged 65 and older were enrolled in the Baltimore Male Osteoporosis Study. Volunteers were recruited from population-based listings of age-eligible male drivers in Baltimore and surrounding counties. In addition, age-eligible husbands, brothers, and sons of postmenopausal women participating in one or more ongoing longitudinal cohort studies or placebo-controlled randomized trials conducted at the University of Maryland Osteoporosis Research Clinic were also invited to participate. Men with bilateral total hip replacements, weight over 300 lb or who were unable to give informed consent were excluded from the study.

This study was approved by the Institutional Review Boards of the University of Maryland Baltimore and the Maryland Veterans Affairs Health Care System at Baltimore. All participants gave written informed consent for participation in the study.

DXA

BMD was measured by DXA using a QDR-2000 (Hologic, Inc., Waltham, MA, USA). Measurements obtained by DXA included bone mineral content (BMC; g), bone area (cm²), and BMD (g/cm²) for the femoral neck, lumbar spine, and total body. The DXA system was calibrated daily to provide accurate BMC measurements in vivo using an anthropomorphic phantom. The DXA precision error rate ranged from 0.5–1%. All DXA measurements were performed by Hologic-certified technicians.

Self-administered questionnaire and interview

Data were obtained from completion of a self-administered questionnaire modified from that used in the Study of Osteoporotic Fractures⁽¹⁷⁾ as well as an interview by a trained examiner at the time of clinic examination. Information included (1) demographic characteristics, (2) lifestyle and habits, (3) dietary history, (4) current and previous medical conditions, (5) current medications, (6) occupational history, (7) participation in recreational physical activities, (8) personal fracture history, (9) family history of fracture, and (10) health-related quality of life measures.

Examination data

Anthropometric data obtained at the time of clinic examination included (1) measurement of weight without shoes and outer clothing using a standard balance beam scale and recorded in kilograms to the nearest 0.1 kg, and (2) height without shoes using a Harpenden stadiometer (Veeder-Root, Elyabethtown, NC, USA) recorded to the nearest centimeter. Quadriceps strength was determined by performance on the Bodymaster MD 110 leg extension chair (Lafayette Instruments, Lafayette, IN, USA).

Statistical analysis

Data were obtained from the baseline visit for all participants enrolled in the study. Univariate analyses were performed to evaluate racial differences in demographic and clinical variables. Differences in categorical variables were examined using Pearson's χ² test; differences in continuous variables were evaluated by Students' t-test.

Multiple linear regression analyses were performed to identify risk factors predictive of BMD, separately for femoral neck BMD, lumbar spine BMD, and whole body BMD. Our primary interest was the predictive power of race beyond the predictive effects of other risk factors. Risk factors that were found to be associated with these outcomes (see Table 2) or which had been reported to be associated with these outcomes in the literature^(1–4) were examined in all models, along with the two-way interactions of race with these risk factors. Using a backward-elimination approach, main effects were retained in the models when they themselves were significant or when their associated interaction terms were significant. All statistical analyses were performed using Stata 6 software (Stata Corp., College Station, TX, USA). Statistical significance for all analyses was defined as p < 0.05.

Table Table 2. Regression Coefficients for Correlates of BMD Measures for the Full Sample and Stratified by Race

RESULTS

The descriptive and clinical characteristics of the participants studied are shown in Table 1. White men were older, taller, had more years of formal education, and were more likely to be current drinkers, to engage in regular physical activity during the past year, and to rate their overall health as excellent. Black men were heavier, more likely to be current smokers, and more likely to report a history of hypertension. The vast majority of participants had served in the United States armed forces (i.e., Army, Navy, Air Force, or Marines).

Table Table 1. Descriptive and Clinical Characteristics by Race

Black men had significantly higher BMD at the femoral neck, lumbar spine, and total body than white men (Table 1; Figs. 1A–1C).

Figure FIG. 1.. (A) Unadjusted (raw) and adjusted means for white and black men on femoral neck BMD. (B) Unadjusted (raw) and adjusted means for white and black men on lumbar spine BMD. (C) Unadjusted (raw) and adjusted means for white and black men on whole body BMD.

Correlates of femoral neck BMD

Demographic, clinical, and lifestyle variables were examined as potential correlates of femoral neck BMD based on a review of the extant literature and results of the descriptive analyses. The list of potential correlates of BMD that were evaluated included age, height, weight, years of education, mean quadriceps strength, current alcohol use, current smoking, race, self-rated health status, comorbid diseases, and regular physical activity. Age, height, weight, years of education, and mean quadriceps strength were analyzed as continuous variables; race, smoking status, self-rated health status, comorbid diseases, and regular physical activity were analyzed as categorical variables because of the manner in which the data were collected from participants. Pairwise correlation coefficients for selected risk factors with BMD measures are presented in Table 2 for the total sample and for white and black men separately.

Racial differences in correlates of BMD

Age was found to be inversely correlated with femoral neck BMD for white but not black men. Height was associated with femoral neck BMD for both groups, although the association was stronger in whites. Weight and quadriceps strength were associated with femoral neck BMD in both groups. Current smoking was associated with decreased femoral neck BMD in blacks but not in whites. Physical activity was associated with femoral neck BMD only in black men.

Increased height and quadriceps strength were associated with increased lumbar spine BMD for white men but not for black men. Current smoking status was associated with decreased lumbar spine BMD for black men but not for white men. Increased weight was associated with increased lumbar spine BMD for men in both race groups.

Physical activity was associated with increased whole body BMD for blacks only. Current smoking was associated with decreased whole body BMD for whites only. Height, weight, and quadriceps strength were all associated with increased whole body BMD for both black and white men.

Multiple regression analyses

Risk factors and race by risk factor interactions were entered into multiple linear regression analyses using a backward elimination procedure to determine the most parsimonious models for femoral neck, lumbar spine, and whole body BMD measurements (Table 3). The main focus of these analyses was examination of racial differences in the presence of other predictive risk factors, with emphasis on any different predictive strength of risk factors across the two races.

Table Table 3. Regression Coefficients for Association Between Predictor Variables and BMD Criterion Variables for Full Sample and Stratified by Race

Femoral neck BMD:

Black race, current smoking, weight, mean quadriceps strength, and status as a former drinker were all significant predictors of femoral neck BMD. Former drinkers had lower femoral neck BMD than current drinkers and nondrinkers combined. Current smokers had lower femoral neck BMD than former and nonsmokers combined. Higher mean quadriceps strength and weight were both associated with increased femoral neck BMD. Regular physical activity, while not a significant predictor by itself, did modify the effect of race on femoral neck BMD, such that regular physical activity was associated with higher femoral neck BMD for blacks but not for whites (Figs. 1A). Even after adjusting for these factors independently associated with BMD in older men, blacks had higher femoral neck BMD than whites (Fig. 1A). The final model accounted for 34% of the variance in femoral neck BMD.

Lumbar spine BMD:

Older age and higher weight were both associated with increased lumbar spine BMD (Table 3), while the use of seizure medication was associated with lower lumbar spine BMD. This model accounted for almost 10% of the variance in lumbar spine BMD. After adjusting for these factors, black men showed higher lumbar spine BMD than did white men (Fig. 1B).

Whole body BMD:

Higher weight was also related to increased whole body BMD. Current smokers had lower whole body BMD than former and nonsmokers combined. While not by itself significantly related to whole body BMD, physical activity modified the effect of race on whole body BMD such that physical activity was associated with increased whole body BMD for blacks but not for whites (Fig. 1C). This model accounted for almost 20% of the variance in total body BMD (Table 3).

Exploratory analyses

Given that racial differences were found in BMD at multiple sites, it was of interest to explore whether or not the higher BMD in black men was related to higher BMC or smaller bone area in the bone regions of interest. In univariate analyses, black men had significantly higher BMC at the femoral neck (t = 6.60, p < 0.0001) and whole body (t = 5.00, p < 0.0001) than white men; no significant difference in BMC was found for the lumbar spine (Table 1). On the other hand, white men had significantly larger bone area than black men at the femoral neck (t = 3.10, p < 0.002) and lumbar spine (t = 5.16, p < 0.0001), but smaller total body bone area (t = 2.30, p < 0.022; Table 1).

In multiple variable linear regression models, racial differences in BMD persisted even when BMC or bone area also were included in the final models (Table 3, middle and right hand columns). In models controlling for bone area, there was slight attenuation of the coefficient for race in the model for femoral neck BMD while there was a 50% increase in the coefficient for race in the model for lumbar spine BMD; no change in the coefficient for race was noted for total body BMD. In models controlling for BMC, there was substantial attenuation of the coefficient for race in both the models for femoral neck and total body BMD and only a slight attenuation for the coefficient for race in the model for lumbar spine BMD. Furthermore, the proportion of the variance explained by the models that included BMC was greater than 60% for femoral neck BMD and greater than 80% for lumbar spine and total body BMD.

DISCUSSION

This study found that older black men had higher BMD than older white men, even after adjusting for factors known to be associated with bone mass in this and other populations. The higher BMD was related to both higher BMC and/or smaller bone area, depending on the region of interest. These findings extend results from previous studies by Bell et al. and others^(18–22) who demonstrated racial differences in BMD between American blacks and whites as well as between West Africans and whites living in the United Kingdom.⁽²¹⁾ Whether these differences are caused by the acquisition of a higher peak bone mass, a slower rate of bone loss, or both is not known at the present time; longitudinal observations from this study will provide data on the rate of bone loss in older men of both racial groups.

Previous studies have shown that black men achieve a higher peak bone mass than white men.⁽¹⁹⁾ The intestinal absorption of calcium is the same in both races, but blacks had lower urinary calcium excretion because of enhanced renal tubular calcium reabsorption.⁽²³⁾ This may be related to higher serum parathyroid hormone levels with increased circulating levels of 1,25-dihydroxyvitamin D in blacks than whites.⁽²⁴⁾ Harris et al.^(25,26) have shown that parathyroid hormone levels were higher in blacks than in whites, and this difference was only partially explained by the racial differences in 25-hydroxyvitamin D levels. Wright et al.⁽²⁷⁾ noted greater secretion of growth hormone and higher serum 17-β estradiol levels in black than in white men, which also may account for greater BMD. Serum samples from the older black and white men in the present study have not yet been analyzed to determine 1,25-dihydroxyvitamin D, 17-β estradiol, and parathyroid and growth hormone levels.

Indirect evidence also exists to support a slower rate of bone loss in black than white men. In studies in which bone histomorphometry of the iliac crest was performed, a reduced rate of bone formation was seen in normal black men and women.⁽²⁸⁾ This may account for lower rate of bone turnover in blacks than in whites because bone resorption and bone formation are closely coupled in the steady state. Consistent with a reduction in bone resorption is the finding of decreased mean serum osteocalcin levels in young and older black men and women.^(29–31)

BMD is calculated from the ratio of BMC and bone area; therefore, differences in BMD may be caused by differences in BMC and/or bone area. We found that older black men had higher BMC in the femoral neck and total body as well as smaller bone area in the femoral neck and lumbar spine than white men (Table 1). These results are similar to those reported by Looker et al.⁽³²⁾ for the femoral neck for non-Hispanic white and black male participants in NHANES III (their Tables 8 and 12). In the present study, racial differences in adjusted BMD persisted even after the inclusion of BMC or bone area in multiple linear regression models, suggesting that neither measure by itself explains the observed differences in BMD. These findings may account, in part, for probable differences in bone strength that underlie the differences in hip fracture rate observed between older black and white men. Future plans include performing hip structural analysis on the scans acquired in this study.⁽³³⁾

Several potential limitations of the present study should be acknowledged. First, selection of the sample from the pool of licensed automobile drivers in the state of Maryland limits the overall generalizability of these findings. It is possible that volunteers recruited by this particular method may have higher socioeconomic status than older men without licenses. In addition, it has been noted that driving cessation is related to decreased physical activity in U.S. samples⁽³⁴⁾; therefore, the study sample may represent a healthier segment of the older male population and results may not be applicable to less healthy older men. Although this is a potential limitation, it should be noted that this recruitment strategy was used successfully for the recruitment of women at the Baltimore site for the Study of Osteoporotic Fractures (SOF) as well as for participants in the Fracture Intervention Trial^(17,35) and that results from these studies have been confirmed in other cohorts. An additional limitation is that data for this study were collected from a single geographic area and results may not apply to older men living in other areas of the United States.

In summary, older black men seem to have consistently higher BMD than older white men at all sites even after adjusting for factors known to be associated with BMD. These differences, approaching 1 SD in size, may account for the almost 50% reduction in rates of osteoporotic fractures among black compared with white men in the United States.

Acknowledgements

This study was supported by grants from the Department of Veterans Affairs and the Arthritis Foundation, Maryland Chapter.