The authors state that they have no conflicts of interest
Article first published online: 16 FEB 2009
Copyright © 2009 ASBMR
Journal of Bone and Mineral Research
Volume 24, Issue 7, pages 1290–1298, July 2009
How to Cite
Sheu, Y., Cauley, J. A., Wheeler, V. W., Patrick, A. L., Bunker, C. H., Kammerer, C. M. and Zmuda, J. M. (2009), Natural History and Correlates of Hip BMD Loss With Aging in Men of African Ancestry: The Tobago Bone Health Study. J Bone Miner Res, 24: 1290–1298. doi: 10.1359/jbmr.090221
Published online on February 16, 2009
- Issue published online: 4 DEC 2009
- Article first published online: 16 FEB 2009
- Manuscript Accepted: 11 FEB 2009
- Manuscript Revised: 5 JAN 2009
- Manuscript Received: 5 SEP 2008
Little is known about the magnitude, pattern, and determinants of bone loss with advancing age among men, particularly among those of African descent. We examined the rate of decline in hip BMD and identified factors associated with BMD loss among 1478 Afro-Caribbean men ≥40 yr of age. BMD was measured at baseline and after an average of 4.4 yr by DXA. The rate of decline in femoral neck BMD was 0.29 ± 0.81%/yr in the total sample (p < 0.0001). However, a U-shaped relationship between advancing age and the rate of decline in BMD was observed. The rate of decline in BMD at the femoral neck was −0.38 ± 0.77%/yr among men 40–44 yr of age, decelerated to −0.15 ± 0.81%/yr among men 50–54 yr of age, and then accelerated to −0.52 ± 0.90%/yr among those 75+ yr of age (all p < 0.003). Men who lost ≥5% of their body weight during follow-up had significantly greater BMD loss than those who remained weight stable or gained weight (p < 0.0001). The relationship between weight loss and BMD loss was more pronounced among men who were older and leaner at study entry (p < 0.03). We also observed a strong impact of advanced prostate cancer and its treatment with androgen deprivation on BMD loss. Men of African ancestry experience substantial BMD loss with advancing age that seems to be comparable to the rate of loss among white men in other studies. Additional studies are needed to better define the natural history and factors underlying bone loss with aging in men of African ancestry.
Although osteoporosis is more prevalent among women than men, men also experience substantial bone loss and an increase in fracture incidence with advancing age. However, little information exists about the natural history, magnitude, and correlates of bone loss with aging among men, especially among men of African ancestry. Osteoporosis is a global public health problem and, as the population ages, more men throughout the world will develop osteoporosis and its associated fractures, including men of African ancestry.(1) Indeed, individuals of African ancestry are expected to comprise a growing proportion of the incidence and economic burden of osteoporosis-related fractures over the next 20–50 yr in the United States(2) and worldwide.(1) These demographic trends underscore the importance of better understanding the natural history and determinants of bone loss and osteoporosis in men of all ages and racial/ethnic background.
Most longitudinal studies of BMD changes with aging in men have been conducted among white men in North America,(3–8) Europe,(9–11) and Australia.(12) These studies have identified body weight and changes in body weight as the major correlates of bone loss with aging. Smoking has also been reported as a potential risk factor for bone loss in some studies,(3,4,7,13,14) whereas other lifestyle factors such as alcohol consumption, calcium intake, and physical activity have been inconsistently related to bone loss.(3,4,7,12) To our knowledge, only a single study has characterized the magnitude and correlates of bone loss in men of African ancestry.(14) In this study, older age, lower initial body weight, and smoking were correlated with a greater decline in BMD among 119 African-American men ≥65 yr of age. The aim of this study was to further examine the magnitude, age-related patterns, and correlates of the decline in hip BMD with aging in a large cohort of middle-aged and elderly men of African ancestry.
MATERIALS AND METHODS
The population-based Tobago Bone Health Study was first conducted on the Caribbean island of Tobago in 2000.(15,16) In brief, recruitment was accomplished by word of mouth, hospital flyers, and radio broadcasting. To be eligible, men had to be ≥40 yr of age, ambulatory, and not terminally ill. Questionnaires were administered to obtain information on demographic characteristics, occupation, medical history, and lifestyle-related factors. A total of 2652 men completed an initial DXA scan for assessment of BMD and body composition at that visit. Self-reported ethnicity of the cohort is 97% African, 2% East Indian, <1% white, and <1% “other.”
Beginning in 2004, participants were recontacted for a second DXA scan to assess the rate of change in BMD. A total of 1748 men (70% of survivors) returned for the follow-up exam. We later excluded 36 men who identified themselves with ethnicity other than Afro-Caribbean and 21 men who had incomplete data. The Institutional Review Boards of the University of Pittsburgh and the Tobago Ministry of Health and Social Services approved this study, and all participants provided written informed consent before data collection.
BMD was measured at the proximal femur and subregions at both the baseline and follow-up visits using a single Hologic QDR 4500W densitometer (Hologic, Bedford, MA, USA). The left hip was scanned unless the participant had a fracture or a total hip replacement. Trained and certified technicians performed the DXA scans and followed a strict protocol for both visits. Longitudinal machine stability was assessed from plots of daily spine phantom scans and reviewed monthly. A weekly print out of quality control (QC) plots was generated to detect short-term inconsistencies and long-term drift. The scanner was stable throughout the course of the study. The long-term CV for BMD was 0.499%.
Anthropometric and body composition assessments
Body weight was measured in kilograms with participants wearing light clothing and without shoes, using a calibrated balance beam scale at both visits. Height was measured in centimeters without shoes, using a wall-mounted height board. Whole body fat and lean mass were also measured using DXA. Left and right grip strength was measured with a hand-grip dynamometer as a surrogate for upper body and overall strength (Preston Grip Dynamometer; JA Preston 136 Crop.). Average grip strength was based on two repeated measurements from the left and right hands.
Questionnaires were administered by trained interviewers and nurses to obtain information on demographic characteristics, lifestyle factors, and medical history. In this analysis, we used information from the baseline exam to assess potential factors related to the subsequent rate of decline in BMD. Mixed African ancestry was defined by self-report of one to three African-descent grandparents. Other factors that were assessed included history of cigarette smoking, alcohol consumption, time spent watching television, time spent walking, and medical history of fracture, hypertension, coronary heart disease, stroke, diabetes, chronic bronchitis, and arthritis. We also asked participants to rate their overall health compared with others of their age. Diagnosis of prostate cancer, advanced prostate cancer (prostate specific antigen > 40 or Gleason score > 7), and androgen deprivation therapy (ADT; use of leuprorelin or orchiectomy) were also recorded at baseline.
The annualized rate of change in BMD during follow-up was calculated as the percent BMD change from baseline to follow-up divided by duration in years between the two scans. Percent change in body weight or body composition was calculated as the difference between baseline and follow-up measures divided by baseline measures and multiplied by 100. We also categorized percent weight change into three groups: weight gain (>5% weight gain), weight stable (weight change between −5% and 5%), and weight loss (>5% weight loss).
We first compared the baseline characteristics between men who did (n = 1748) and did not (n = 901) return for the follow-up visit using analysis of covariance (ANCOVA) for continuous variables with age adjustment and χ2 tests for dichotomous variables. ANOVA and ANCOVA were used to evaluate the annualized percentage rate of change in BMD across different age categories (e.g., 5-yr age groups or >55 yr) and across weight change groups. The distribution of BMD change in each age group satisfied the criteria of ANOVA. We evaluated the age-adjusted and age- and weight-adjusted contribution of each individual variable to the annualized rate of change in BMD using linear regression analysis. The strength of the association is expressed as an absolute difference in units of change chosen to approximate 1 SD in the distribution for each continuous variable or null category for dichotomous variables. The formula used to calculate the absolute difference in rate of change in BMD per unit change (SD) of the independent variable was β = unstandardized β × unit change in independent variable. The corresponding 95% CIs were calculated using the following formula: β × (±1.96 × SE) × unit change. We also evaluated the interactions between weight change and age as well weight change and BMI on the rate of change in BMD using ANCOVA. We first performed the above analyses in 1691 participants with complete data. Because the effects of ADT and prostate cancer on BMD change were substantial, we excluded the 233 men on ADT or who had prostate cancer from subsequent analyses.
Multiple linear regression analysis was performed using a stepwise procedure to determine the potential independent correlates of the annualized rate of change in BMD. Variables from the age- and weight-adjusted univariate model with p < 0.10 were further entered in the multiple linear regression model. Age was forced into all models. We assessed multi-collinearity by inspecting the variance inflation factor (VIF). Because of the high correlation of body weight with lean and fat mass, we developed two different multiple linear regression models: (1) models with body weight only and (2) models substituting fat and lean mass for body weight. Statistical analyses were performed using SAS (version 9.1; SAS Institute, Cary, NC, USA).
Table 1 shows the baseline characteristics of men who did and did not attend the follow-up exam. As expected, men who did not return were older, weighed less, and had lower baseline BMI and BMD than those who participated in the follow-up exam. Nonparticipants also were more likely to smoke, report poorer health status, and have a greater prevalence of hypertension, heart disease, diabetes, and arthritis.
Rate of change in BMD and androgen deprivation
The average length of time between DXA scans was 4.4 ± 0.8 yr (range, 1.1–6.9 yr). Men with advanced prostate cancer and men who had a history of ADT for prostate cancer had a significantly greater decline in BMD compared with their counterparts. For example, the average rate of decline in BMD was 0.137 ± 0.588%/yr at the total hip and 0.333 ± 0.841%/yr at the femoral neck (data not shown) for all men. The rate of decline in BMD was 0.78%/yr greater at the total hip and 0.58%/yr greater at the femoral neck for men with advanced prostate cancer compared with men without prostate cancer (p < 0.0001 for both). Men with nonadvanced prostate cancer had a 0.08% greater (p = 0.08) rate of decline in BMD at the total hip and 0.15% greater rate (p = 0.02) at the femoral neck compared with men without prostate cancer. Approximately 6% of the study population reported that they had either taken leuprorelin or underwent orchiectomy for prostate cancer. Men on either of these treatments experienced an ∼0.4%/yr greater rate of decline in total hip and femoral neck BMD compared with men who had neither of these treatments (p < 0.0001 for both). Because prostate cancer and its treatment by androgen deprivation had a strong impact on the rate of decline in BMD, we excluded the 233 men with prostate cancer or who had undergone androgen deprivation from subsequent analyses.
Rate of change in BMD and age group
Total hip and femoral neck BMD declined by 0.10 ± 0.55%/yr and 0.29 ± 0.81%/yr among the men who did not have prostate cancer or who had not undergone ADT (p < 0.0001 for both). To examine the age-related patterns in the rate of decline in BMD among these men, we stratified the total sample by 5-yr age groups using age at study entry (Fig. 1). The rate of decline in BMD across age groups appeared to have a U-shape relationship. Men 40–44 yr of age had a significantly greater rate of decline in BMD than those 45–49 and 50–54 yr of age at both the total hip and femoral neck. Thereafter, the rate of decline in BMD accelerated with advancing age. For example, the rate of decline in total hip BMD was −0.10 ± 0.55%/yr among men 55–59 yr (p = 0.009) and increased to −0.48 ± 0.60%/yr among men 70+ yr of age (p < 0.0001). Similar results were observed at the femoral neck. Moreover, this U-shaped relationship persisted after adjusting for baseline weight and weight change (data not shown). Additional adjustment for hypertension, diabetes, arthritis, grip strength, and current smoking status yielded similar results (data not shown).
Peak BMD reference data for the Afro-Caribbean population is not available. Thus, we used the non-Hispanic black BMD reference data from NHANES III to determine the prevalence of osteoporosis in our cohort. The prevalence of proximal femur osteoporosis at study entry was 0.41% and increased to 0.75% at the follow-up exam (data not shown).
Age-adjusted correlates of the rate of change in BMD
Table 2 shows the age-adjusted association of participant characteristics with the rate of decline in BMD at the total hip and femoral neck. Age was inversely and significantly correlated with the decline in BMD. For example, the rate of decline in total hip BMD increased by 0.084%/yr every 10 yr.
A 10-kg increase in body weight was associated with a slower rate of decline in total hip BMD. Body composition measures from DXA were also significantly associated with the rate of decline in hip BMD. Whole body fat, lean mass, and percent body fat were all positively and significantly associated with the decline in hip BMD in age-adjusted analysis. Initial BMD was not related to the subsequent rate of decline in BMD.
We also evaluated the relationship between change in body weight and body composition and the rate of decline in BMD. Men gained an average of 0.2 ± 6.2% body weight during follow-up (p = 0.20). As expected, fat mass increased (p < 0.0001), whereas lean mass decreased nonsignificantly (p = 0.25) during the follow-up period. In age-adjusted analysis, men who lost weight during follow-up had a significantly greater rate of decline in BMD. For example, every 6% decrease in body weight from baseline was associated with a 0.11%/yr faster total hip and a 0.16%/yr faster femoral neck BMD loss. Similar associations were observed for the change in fat mass and lean mass.
None of the lifestyle-related characteristics examined were significantly associated with the rate of decline in BMD including current and past smoking history, time spent watching television, and alcohol intake. Diabetes was present in ∼11% of the cohort and was associated with a greater decline in total hip, but not femoral neck, BMD in age-adjusted analysis. Because there was a U-shaped relationship between age and the rate of decline in BMD, we also examined the correlates of bone loss in stratified analyses among men 40–54 and ≥55 yr of age. We also included age as a dichotomous variable in adjusted analyses. We found similar results in these analyses as in the total cohort (data not shown).
We further evaluated the association of age, BMI, weight change, and the interactions of these variables with the rate of decline in BMD in stratified analyses (Tables 3 and 4). At both the total hip and femoral neck, men who lost 5% or more of their baseline body weight had a significantly greater decline in BMD compared with men who had remained weight stable or who gained 5% or more body weight. There was also a significant loss of BMD among men who remained weight stable during follow-up. However, men who gained at least 5% body weight during follow-up did not experience a significant decline in hip BMD.
We also found a significant interaction effect of age at study entry and weight change on the rate of decline in hip BMD (Table 3). The effect of weight loss on the decline in BMD was significantly greater (total hip, p = 0.006; femoral neck, p = 0.018) among older (age ≥ 55 yr) than younger (age <55 yr) men. Finally, we also observed a significant interaction effect of initial BMI and subsequent weight change on the rate of decline in total hip BMD (Table 4). Although weight gain was associated with an overall slower rate of decline in femoral neck BMD, this effect was not apparent among the leanest men. Men with a low initial BMI (BMI < 25.0 kg/m2) experienced a more pronounced rate of decline in BMD even if they had gained weight during follow-up.
Multiple linear regression analyses
The results from multiple linear regression analyses of the independent correlates of the rate of decline in BMD are shown in Table 5. Age (−), BMI (+), weight change (+), and grip strength (+) were significant correlates of the rate of decline in BMD in multiple regression analysis. Diabetes was no longer a statistically significant correlate of the rate of decline in total hip BMD in the multivariable model. Multivariable models explained 5–6% of the variance in the rate of decline in BMD.
Compared with women, much less is known about the natural history, magnitude, and correlates of bone loss with aging in men, especially among nonwhite men. To our knowledge, only a single longitudinal study to date has evaluated age-related bone loss in men of African ancestry and that study only included 119 men ≥65 yr of age.(14) Thus, one aim of this study was to examine the pattern of bone loss at the proximal femur over 4 yr in a large population sample of middle-aged and elderly men of African ancestry. A primary finding was that the rate of decline in proximal femur BMD is substantial and may begin early in life among men of African ancestry. This early decline in BMD may reflect, in part, an early loss of trabecular bone mass from the proximal femur. Indeed, others have recently observed a loss of trabecular BMD beginning by the third decade of life among white men.(17,18) However, the pattern of BMD loss with advancing age in our study seemed to be nonlinear, with a deceleration in the rate of loss between ages 40 and 54 and an acceleration in the rate of loss thereafter that continued unabated into the seventh decade of life. A progressive acceleration of bone loss with advancing age has also been observed among white men.(3,13) Past epidemiological studies of the decline in BMD have largely focused on white men ≥65 yr of age and may have thus missed this early decline in hip BMD.
In our study of Afro-Caribbean men 40–92 yr of age, the overall unadjusted rate of decline in BMD at the femoral neck was −0.29 ± 0.81%/yr and was −0.47 ± 0.89%/yr among men ≥65 yr of age. The rate of decline in BMD was surprisingly very similar to the rate of decline reported in other longitudinal studies of white men.(3–5,7,9) For example, in the Framingham Osteoporosis Study, the rate of decline in femoral neck BMD was −0.38%/yr among 278 white American men 67–90 yr of age.(4) In the Rotterdam Study, the rate of decline in femoral neck BMD was −0.4%/yr among 1856 white European men >55 yr of age.(3) In the Rancho Bernardo study, the rate of decline in BMD at the femoral neck was −0.34%/yr among 500 white American men 45–92 yr of age.(7) In the Network in Europe for Male Osteoporosis study, the rate of decline in BMD at the femoral neck was −0.48%/yr among ∼1300 white European men 50–80 yr of age.(11) In the Baltimore Men's Osteoporosis Study, the rate of decline in femoral neck BMD was 2.1%/yr in 349 white American and 1.1%/yr in 119 black men 60–74 yr of age.(14) The higher rate of decline in BMD in this later study may reflect the use of different densitometers at the initial and follow-up exams. Comparisons across these studies are difficult because of the differences in follow-up time, densitometers used, sample sizes, geographic areas, and age distributions of the populations studied. Nonetheless, the rate of decline in femoral neck BMD with aging in our cohort of Afro-Caribbean men is very consistent with the majority of these other studies of white men.
To better understand the factors that might influence the rate of bone loss with age in men of African ancestry, we characterized a number of anthropometric, lifestyle, and medical variables and examined the relation of these variables to the rate of decline in BMD at the proximal femur. In addition to advanced age, leanness at study entry and weight loss during follow-up were independent correlates of an increased rate of BMD decline with age. A positive association between body weight or BMI and BMD among middle-aged and elderly men has been well-documented in whites and blacks.(19–24) Weight loss with aging is also a consistent predictor of bone loss in studies of white men.(4,6–8,25) In this study, men who lost 5% or more of their baseline body weight had an accelerated rate of decline in BMD compared with men who remained weight stable or gained weight. The effect of weight loss on BMD loss was more profound among men ≥55 yr of age in our study. In the Framingham, Rancho Bernardo, Osteoporotic Fractures in Men (MrOS), and EPIC studies, white men who lost 5% or more of their baseline weight also experienced significantly greater bone loss at the proximal femur than those who gained 5% or more of their baseline weight.(4,6–8) The importance of weight loss as a risk factor for accelerated loss of BMD among middle-aged and elderly men may be explained by underlying illness that results in poor health and physical inactivity,(8,25) to declines in muscle mass and strength, to decreased mechanical loading on weight-bearing skeletal sites,(26,27) to a decrease in adipose tissue mass, which is an important source of estrogens in men,(28) or to a combination of these factors. In addition, the interaction between age and weight loss on bone loss also suggests an important role of weight maintenance for bone health among older men.
Other frequently examined predictors of bone loss such as physical activity, smoking, calcium intake, and medical conditions have yielded inconsistent results across studies.(3,4,7,12) We also examined several of these variables in this study. We found that greater grip strength was associated with a slower decline in BMD, even after adjusting for age, body weight, and weight change. This association may be indirectly caused by physical activity level.
We were unable to document an association between smoking and the rate of decline in BMD. Some(22,29–31) but not all(19,32,33) studies suggest that smokers have lower BMD than nonsmokers. In some studies of white men, smokers had a greater decline in hip BMD with age compared with nonsmokers.(3,4,7,25) We were also unable to document an association between alcohol intake and the rate of decline in BMD. Moderate alcohol consumption has been associated with greater BMD in several(22,23,34) but not all(13,19) studies of middle-aged and elderly white men. Alcohol consumption was associated with a slower rate of decline in BMD in some(7,35) but not all(4,25) longitudinal studies. The absence of a significant relationship with smoking and alcohol drinking in this study may be caused by a low prevalence of these behaviors in our cohort.
Two of the strongest correlates of the rate of decline in BMD were advanced prostate cancer and ADT for prostate cancer. This association was independent of other covariates including body weight and weight change during follow-up. Prostate cancer remains the most common cancer in American men and disproportionately afflicts more men of African ancestry worldwide than other ethnic and racial groups.(15,36,37) ADT achieved surgically through orchidectomy or through gonadotropin-releasing hormone agonists, is a first-line treatment for metastatic prostate cancer and is increasingly being used in the treatment of localized, high-risk prostate cancer.(38) The prevalence of androgen deprivation therapy may be increasing(39): by 2000, ∼3% of men in the U.S. Medicare fee-for-service program alone were androgen deprived.(39) Androgens are important for the development and maintenance of BMD in men. Androgen deprivation for prostate cancer has been associated with a rapid and marked loss of BMD in numerous studies(40–47) and with an increased risk of fracture.(38,48–53) The vast majority of men in these past studies were white. To our knowledge, ours is the first to document the association of ADT with the rate of decline in BMD exclusively among men of African ancestry.
This study has several potential limitations. The small proportion of men ≥75 yr of age may have limited our ability to estimate the rate of bone loss among these men. Although walking is a common form of physical activity in this population, our questionnaire-based estimates of walking might not have accurately reflected total physical activity levels. In addition, dietary calcium and vitamin D intake was not assessed, and only selected medication use was documented at study entry. Areal BMD may be confounded by bone size changes.(54) Bone mineral apparent density (BMAD) has been suggested as a surrogate for volumetric BMD. However, we found similar results for femoral neck BMAD (data not shown). Our DXA measures of areal BMD cannot provide insight on age-related loss of trabecular and cortical BMD or bone geometry. Finally, it will be important to examine fracture rates in relation to BMD and the rate of bone loss among men of African ancestry in the future.
In conclusion, although the prevalence of osteoporosis is higher among white men than men of African ancestry,(55) men of African ancestry seem to experience a substantial loss of BMD with aging that may be comparable to the rate of loss among white men. Advancing age, lower body weight, increased weight loss, and advanced prostate cancer and its treatment by androgen deprivation were identified as potential risk factors for accelerated BMD loss among men of African ancestry. However, these factors accounted for only a small fraction of the variation in BMD loss with aging. Many additional variables, including genetic factors,(56–60) may contribute to age-related BMD loss among men of African ancestry. A more complete understanding of BMD changes with aging in men of African ancestry will require a broader examination of the potential determinants.
This study was supported in part by Grants R01-AR049747 from the National Institute of Arthritis and Musculoskeletal and Skin Diseases and by R01-CA84950 from the National Cancer Institute.
- 362002 Prostate cancer risk is three-fold higher among men, aged 50-64, of African descent compared with men of Asian-Indian descent in Trinidad and Tobago. Ethn Dis 12: S3–S30., , , , , , ,