The authors state that they have no conflicts of interest.
Genetic and Environmental Determinants of Volumetric and Areal BMD in Multi-Generational Families of African Ancestry: The Tobago Family Health Study†
Article first published online: 16 JAN 2007
Copyright © 2007 ASBMR
Journal of Bone and Mineral Research
Volume 22, Issue 4, pages 527–536, April 2007
How to Cite
Wang, X., Wheeler, V. W., Patrick, A. L., Bunker, C. H. and Zmuda, J. M. (2007), Genetic and Environmental Determinants of Volumetric and Areal BMD in Multi-Generational Families of African Ancestry: The Tobago Family Health Study. J Bone Miner Res, 22: 527–536. doi: 10.1359/jbmr.070106
- Issue published online: 4 DEC 2009
- Article first published online: 16 JAN 2007
- Manuscript Accepted: 10 JAN 2007
- Manuscript Revised: 20 DEC 2006
- Manuscript Received: 24 OCT 2006
BMD is higher and fracture risk is lower among individuals of African versus European descent, but little is known about the genetic architecture of BMD in the former group. Heritabilities of areal and volumetric BMD were moderate in our large families of African descent but differed for trabecular and cortical BMD.
Introduction: Populations of African ancestry have lower osteoporotic fracture risk and higher BMD than other ethnic groups. However, there is a paucity of information regarding the genetic and environmental influences on bone health among populations of African heritage.
Materials and Methods: We dissected the genetic architecture of areal BMD measured by DXA at the proximal femur, lumbar spine, and whole body and volumetric BMD measured by pQCT at the distal and proximal radius and tibia in 283 women and 188 men ≤18 years of age (mean, 43 years) from eight multigenerational Afro-Caribbean families (mean family size > 50). Using quantitative genetic methods, we estimated the residual heritability and the effects of anthropometric, demographic, lifestyle, and medical variables on areal and volumetric BMD.
Results: Compared with U.S. non-Hispanic blacks and whites, areal BMD at the femoral neck was highest in the Afro-Caribbean men and women at all ages. Trabecular volumetric BMD decreased linearly with increasing age, whereas cortical volumetric BMD did not decrease until age 40–49, especially in women. Anthropometric, lifestyle, and medical factors accounted for 12–32% of the variation in areal and volumetric BMD, and residual heritabilities (range, 0.23–0.52) were similar to those reported in other ethnic groups. Heritability of cortical BMD was substantially lower than that of areal or trabecular volumetric BMD, although the measured covariates accounted for a similar proportion of the total phenotypic variation.
Conclusions: Our study is the first comprehensive genetic epidemiologic analysis of volumetric BMD measured by QCT and the first analysis of these traits in extended families of African descent. Genes account for as much or more of the total variation in areal and volumetric BMD than do environmental factors, but these effects seem to differ for trabecular and cortical bone.
Osteoporosis is a systemic skeletal disorder characterized by low bone mass and microarchitectural deterioration of bone with a consequent increase in bone fragility and susceptibility to fractures.(1) BMD is one of the most important indicators of the mechanical strength of the skeleton and the risk of osteoporosis-related fractures. BMD can be assessed by 2D DXA or by 3D CT. QCT may provide several advantages over DXA for the assessment of BMD. First, QCT is a 3D technique that measures BMD volumetrically (g/cm3) and is independent of the potential confounding effects of bone size. Bone volume estimates derived from DXA images have been proposed(2) but are thought to be of limited use. Second, QCT directly examines the trabecular and cortical bone compartments separately, whereas DXA yields a measure of integral BMD (i.e., cortical and trabecular) that may be confounded by the high prevalence of extravertebral mineralization among older adults (e.g., osteophytes).(3–6) Furthermore, measures of trabecular BMD may confer the highest risk for vertebral fracture(7,8) and are more sensitive to changes in bone metabolism.
Many intrinsic and extrinsic factors have been correlated with areal BMD, including heredity, age, sex, body mass index, sex steroid hormones, physical activity, calcium intake, and diabetes.(9–13) However, most heritability studies have relied solely on DXA measures of integral BMD; none to our knowledge have comprehensively dissected the relative contribution of genetics and environmental factors to the determination of trabecular and cortical volumetric BMD as measured with QCT. Furthermore, although several studies report that populations of African ancestry have substantially higher BMD and lower osteoporotic fracture risk than other ethnic groups,(14–17) there is a paucity of information about the genetic and environmental factors that contribute to skeletal health in populations of African origin. Thus, dissection of the genetic architecture of QCT measures of volumetric BMD, especially in populations of African heritage, is important to advance our understanding of the etiology of osteoporosis.
In this study, we investigated the genetic and environmental contributions to areal BMD measured by traditional DXA and volumetric BMD measured by pQCT in 471 individuals from eight large, multi-generational Afro-Caribbean families. We also compared our results to those obtained in populations of European descent and blacks.
MATERIALS AND METHODS
As part of a large population-based prostate cancer surveillance project of all age-eligible men on the Caribbean island of Tobago,(18) we recorded the number of living siblings for each participant and the vital status and residence of their parents. These data revealed that the sibship size in Tobago is large, with a median living sibship size of eight.
Thus, the men in our population-based study served as potential probands for the family study and were selected without regard to medical history or BMD. To be eligible, a proband had to be a healthy Afro-Caribbean resident, have had a spouse who was willing to participate in the study, and have at least six living offspring and/or siblings ≤18 years of age who were residing in Tobago. These potential probands were sorted by family size, and individuals with the largest family sizes were recruited first. All eight probands for the Tobago Family Health Study were between 52 and 103 years of age, and none of them had previously been diagnosed with prostate or any other cancer. In addition, all first-, second-, and third-degree relatives of these probands and their spouses were invited to participate regardless of their medical history or BMD. To date, we recruited 471 individuals ≤18 years of age in eight multigenerational families of the following sizes: 102, 26, 49, 28, 113, 21, 38, and 94, with a mean family size of >50 individuals. These 283 women and 188 men with phenotype data ranged in age from 18 to 103 years (mean age, 43 years). Among these 471 individuals in eight pedigrees, we have the following relationships: 361 parent-offspring, 495 full siblings, 101 grandparent-grandchildren, 1137 avuncular, 61 half-sibs, and 1380 cousins. Written informed consent was obtained from every participant, using forms and procedures approved by the Tobago Ministry of Health and Social Services and University of Pittsburgh Institutional Review Boards.
Single axial tomographic slices of the nondominant forearm and left tibia were scanned using a Stratec XCT 2000 scanner (Stratec Medizintechnik, Pforzheim, Germany) according to standardized measurement and analysis procedures. Each scan was acquired with a 0.5-mm voxel size, slice thickness of 2.5 mm, and at a speed of 20 mm/s. The precise position of the measurement sites were determined in a 30-mm planar scout view using the medial endplate of the radius and tibia as standard anatomic landmarks and automatically set by the software at 4% (i.e., distal) or 33% (i.e., shaft) of the length of the radius and tibia proximal to the distal endplate. These anatomical sites were chosen to assess primarily trabecular and cortical bone, respectively. Tibia length was measured from the medial malleolus to the medial condyle of the tibia, and forearm length was measured from the olecranon to the ulna styloid process.
Image processing was performed using the Stratec software package (Version 5.5E). All 4% ultra distal radius and tibia scans were analyzed using identical parameters for contour finding and separation of trabecular and cortical bone (contour mode 2, T = 169 mg/cm3; peel mode 1, area = 45%) to determine the volumetric BMD of the total (mg/cm3) and trabecular (mg/cm3) bone compartments. All 33% proximal radius and tibia shaft scans were analyzed using identical parameters for contour finding and separation of total and cortical bone (contour mode 2, T = 169 mg/cm3; cortmode 1, T = 710 mg/cm3) to determine the volumetric BMD of the total (mg/cm3) and cortical (mg/cm3) bone compartments.
The short-term in vivo precision of the separate pQCT measurements for 15 subjects ranged from 0.43% (for total density at the distal tibia) to 6.21% (for total density at distal radius). The test–retest correlations for all six pQCT traits ranged between 0.85 (for total density at the distal radius) and 0.996 (total density at the distal tibia). A phantom was scanned daily to maintain quality assurance.
Areal BMD (g/cm2) of the lumbar spine (L1–L4) and proximal femur (total hip, femoral neck) was measured by DXA using the array beam mode on a Hologic QDR 4500W scanner (Hologic, Bedford, MA, USA). Standardized procedures for participant positioning and scan analysis were followed according to the manufacturer's recommended protocol. Scans were analyzed with QDR software version 8.26a.
The short-term in vivo precision of the DXA measurements was assessed in 12 subjects. All CVs were ≤1.16%, and all test–retest correlations were >0.99. A phantom was scanned daily and reviewed by DXA Resource Group (Worcester, MA, USA) to maintain longitudinal quality assurance of the scanner during the course of the study.
Body weight was measured to the nearest 0.1 kg with participants wearing indoor clothing but without shoes using a balance beam scale. Standing height was measured to the nearest 0.1 cm without participants wearing shoes using a wall-mounted stadiometer. The average of two measurements was used. Body mass index was calculated by dividing body weight (kg) by height (m2). An inelastic tape measure was used to determine the waist circumference (cm) at the umbilicus.
Information on demographic characteristics, medical history, and lifestyle habits was obtained by questionnaire and interview by trained and certified clinical staff. Race/ethnicity was based on self-declaration, and participants provided detailed information on the ethnic origin of their parents and grandparents. Respondents were assigned to an ethnic group if they reported that all four grandparents belonged to that group. The Tobago population is predominantly of West African origin (97% of the island according to the most recent census data), with low non-African admixture. Previous studies using molecular markers have confirmed a low admixture (6% non-African) in this population.(19)
Subjects were classified as current smokers (yes/no). Participants who had smoked <100 cigarettes in their lifetime were considered nonsmokers. Information on alcohol consumption was obtained by questionnaire and expressed as drinks per week. Information on calcium and vitamin D supplement use was also obtained (yes/no). Milk consumption at several ages was assessed by questionnaire.(20) We assessed caffeine consumption by assuming the following: one cup of coffee, 95 mg of caffeine; one cup of tea, 55 mg; one cup of cola, 45 mg.(21,22) Subjects were asked if they walk for exercise (yes/no). Physical activity was also assessed as a continuous variable by the number of minutes walked and hours spent watching television per week.
Reproductive characteristics including ever being pregnant, menopausal status, use of hormone therapy, and current oral contraceptive use were recorded (all coded as yes/no). Because only 5/283 (1.8%) women reported ever using postmenopausal hormone therapy, we did not consider this variable as a potential correlate of BMD in subsequent analyses. Women were defined as postmenopausal if they declared that they had no menses for at least 12 months and were >40 years old (n = 76) or they had experienced a hysterectomy or ovariectomy (n = 9).
Participants were asked to bring prescription and nonprescription medications to the clinic for verification. Current use was defined as use within the preceding 30 days. A study-specific medication dictionary was used to categorize the type of medication from product brand and generic names obtained from the medication containers. Dose or duration of use or specific indication was not queried. Subjects were asked whether a doctor or health care provider had ever told them they had certain medical conditions including arthritis, cancer, or cardiovascular disease. Hypertension was defined as a diastolic blood pressure >90 mmHg, systolic blood pressure >140 mmHg (n = 91), or currently taking blood pressure medication (n = 41). Diabetes was defined as fasting glucose level >126 mg/dl (n = 45) or currently taking diabetes medication (n = 26).
Before statistical analysis, the distributions of all BMD traits were assessed for non-normality. Subsequently, all outliers (±4 SD) were removed for each BMD trait, and no more than four values were removed for a single variable.
The major goal of our analysis was to determine the extent to which genetic and environmental factors contribute to the total phenotypic variation of the areal and volumetric BMD measures. Initially, to identify possible covariates influencing each BMD trait, we first performed combined forward and backward stepwise linear regression analysis, ignoring the nonindependence of the subjects, using the R statistical package (Version 2.2.1). We required each variable remaining in the model to be significant at p ≤ 0.10. We subsequently evaluated each of the potentially significant covariates using a variance component framework that enabled us to take into account the correlations among family members.(13) Briefly, the variance components approach involves partitioning the variance of a quantitative trait into components attributable to individual-specific covariates (e.g., age, BMI, diabetes status), an additive genetic (polygenic) component, and a residual nonmeasured environmental component. The additive genetic component is modeled as a random effect from the covariance matrix, which is a function of the coefficient of relatedness between all pairs of individuals. For example, two full sibs share one half of their genes in common on average, and thus have a coefficient of relatedness equal to 0.5. Thus, the effects of all independent variables on BMD are estimated conditional on the correlations among related individuals. The significance of a particular independent variable (e.g., diabetes status) was assessed by the likelihood ratio test, which compares the likelihood of a full model (e.g., age, BMI, and diabetes status) to that of a nested model (e.g., age and BMI only, with the diabetes status effect constrained to be zero). Similarly, the significance of the residual heritability (h2r), was determined by comparing the likelihood of a model in which h2r was estimated to a nested model in which h2r was constrained to 0. These analyses were performed using the program SOLAR.(23) In addition to h2r, which represents the proportion of the residual phenotypic variation caused by additive genetic effects, we also estimated h2, which represents the proportion of total phenotypic variation (including variation caused by measured covariates) that is attributable to additive genetic effects. The relationship between these two terms can be expressed by: h2r = h2/(1 − r2), where r2 is the proportion of total phenotypic variation explained by the measured covariates.
To compare the effects of covariates across all traits, we report the strength of association between covariates and the BMD traits as a percent difference in the BMD trait per unit of the covariate instead of the nonstandardized β coefficients. Percent differences in each BMD trait for each covariate were calculated as (β coefficient >unit range)/mean BMD. For continuous covariates, the unit range was every 5 years for age, every 10 kg for body weight, and every 8.5 cm (1 SD) for height. The unit range for dichotomous covariates equaled 1.(3) The percentage difference between younger and older age groups was calculated as [(Trait Mean 60+ age group − Trait Mean 18–29 age group)/Trait Mean 18–29 age group] >100% for each sex.
All p values that tested for sex differences were computed using SOLAR univariate regression analysis, which accounts for the nonindependence of the family data.
Among the 283 women and 188 men, BMI was higher (p < 0.001), but waist circumference was similar in women than men (Table 1). Although the prevalence of cigarette smoking and alcohol consumption was higher in men than women, the prevalence of hypertension, diabetes, cardiovascular disease, and arthritis was comparable in both sexes. Mean number of minutes walked or time watching TV per week was comparable in men and women. Current use of supplemental vitamin D and calcium and milk consumption were all significantly higher in women than in men (p < 0.05).
Sex differences in BMD
Mean areal and volumetric BMD in men and women with and without adjustment for height and weight are shown in Table 2. All measures of areal BMD were significantly greater in men than in women. Mean differences in unadjusted areal BMD between men and women ranged from 0.07 g/cm2 or 7% at the lumbar spine and femoral neck to 0.11 g/cm2 or 10% at the total hip and whole body. The magnitude of this sex difference represents ∼0.5 (lumbar spine, femoral neck) to 1 (whole body) SD. These sex differences in areal BMD were only slightly attenuated and remained statistically significant after controlling for height and body weight.
Sex differences in volumetric BMD were generally greater at the distal than proximal (shaft) regions of the radius and tibia and were greater for trabecular than cortical BMD (Table 2). Trabecular volumetric BMD was 14% greater at the distal radius and 7% greater at the distal tibia among men than women, differences that represent ∼0.75 and 0.5 SD. Adjusting for height and body weight had little effect on these results. Cortical volumetric BMD, on the other hand, was significantly greater in women than men at the radius and tibia. Although these BMD differences were statistically significant and persisted after adjustment for height and body weight, they were small (∼1%).
Age-related patterns in BMD by sex and comparisons with other populations
The sex-specific mean BMD values for young participants (18–29 years) and the absolute and percentage difference compared with older participants (60+ years) are shown in Table 3. Consistent with observations for the entire sample (Table 2), the young adult Afro-Caribbean men generally had greater areal BMD values than women. These relative differences were largely maintained throughout life (Fig. 1 shows femoral neck BMD; other skeletal sites not shown).
The largest areal BMD difference between young and old Afro-Caribbean men and women was observed at the femoral neck: −11% and −14% over life in men and women, respectively. We also compared the apparent age- and sex-related patterns in BMD at the femoral neck in our Afro-Caribbean family members with published values for non-Hispanic blacks and non-Hispanic whites in the U.S. NHANES III study (Fig. 1).(24) As expected, mean areal BMD at the femoral neck decreased in all age and sex groups (Fig. 1) and was highest among the Afro-Caribbean men and women throughout life. However, comparison of the slopes of the curves across age groups within each sex indicate that Afro-Caribbean men and women may lose bone at a similar tempo to U.S. white and black men and women. A similar pattern was observed for total hip BMD (data not shown). At the lumbar spine in Afro-Caribbeans, areal BMD increased over life by 14% in men but decreased by 8% in women. Lumbar spine BMD data were not available in the U.S. NHANES III survey for comparison.
The apparent decline with age in total volumetric BMD was similar in magnitude at the distal radius and distal tibia and somewhat greater in women (−17% to −18%) than men (−11% to −12%; Table 3). The apparent age-related decline in total volumetric BMD at the radius shaft and tibia shaft was also similar in magnitude. However, the decline with age was much greater in women (−10% to −11%) than in men (0.9% to −1%). Separate measures of trabecular and cortical volumetric BMD at these skeletal sites also revealed different patterns by sex and anatomic region. For instance, age-related declines in volumetric BMD were greater for trabecular than cortical bone, particularly among men. Moreover, the age-related decline in trabecular volumetric BMD tended to be greater in men than in women, whereas the decline in cortical volumetric BMD was greater in women than in men.
We also plotted the age- and sex-specific mean volumetric BMD at the radius and tibia across life for the Afro-Caribbean men and women (Figs. 2A and 2B). Mean trabecular volumetric BMD (Fig. 2A) was higher in men than women at both skeletal sites over life and decreased with increasing age in both groups and at both sites. In contrast, cortical volumetric BMD at both skeletal sites increased until age 40–49 and was higher in women than men. In subsequent age groups, mean cortical BMD decreased markedly in women, whereas it remained fairly constant with advancing age in men, at least until the oldest age group (70+).
Anthropometric, lifestyle, reproductive, and medical correlates of BMD
We next studied the potential correlates of the DXA and pQCT BMD traits. Table 4 describes the percent difference in BMD for the most important correlates. With few exceptions, areal and volumetric BMD were generally higher in men than in women, decreased with increasing age and height and increased with increasing body weight. In contrast to the other BMD traits, volumetric cortical BMD at the radius and tibia decreased with increasing body weight. With the exception of total spine and cortical volumetric BMD of the radius and tibia, these four factors (age, sex, height, body weight) accounted for the majority (73–99%) of the total variation caused by significant covariates (Table 5).
In addition to these four covariates, postmenopausal women had lower measures of BMD than premenopausal women. This difference tended to be more prominent for areal BMD at the lumbar spine versus proximal femur and for trabecular versus cortical volumetric BMD. Parity was associated with greater cortical but not trabecular volumetric BMD at both the radius and tibia. Diabetes was associated with an increased BMD for most DXA and pQCT traits.
History of hypertension and cardiovascular diseases showed weaker and inconsistent associations with BMD phenotypes. We were unable to detect strong associations between BMD phenotypes and walking time, TV watching time, or supplemental calcium intake.
Heritability of BMD
After incorporating all significant covariates, additive genetic effects (residual heritability), accounted for 55–65% of the remaining variation in areal BMD (Table 4). In contrast, residual heritability of trabecular volumetric BMD at the radius and tibia was higher and accounted for ∼70% of the residual variation at these skeletal sites. Residual heritability of cortical volumetric BMD was lower than for trabecular volumetric BMD and ranged from 29% to 42% of the total residual variation. Residual heritability of total volumetric BMD at the radius and tibia ranged from 46% to 73% of the total residual variation, with sites comprised predominantly of trabecular bone (distal radius and tibia) having higher residual heritability than sites comprised predominantly of cortical bone (radial and tibial shaft).
Altogether, significant measured covariates explained from 12% to 38% of the total phenotypic variation in areal and volumetric BMD (Table 5). Furthermore, the proportion of total variation caused by additive genetic factors accounted for as much as or more of the total phenotypic variation than did measured covariates. Heritability of trabecular BMD was higher than heritability of cortical BMD (e.g., 0.52 versus 0.23 for the radius), although measured covariates accounted for a similar proportion (21–26%) of the total phenotypic variation for each of these traits.
To our knowledge, the Tobago Family Health Study is the first comprehensive analysis of the genetic and environmental factors influencing BMD measured by both DXA and QCT in families, and it is the largest family study of individuals of African descent to date. Clinical data, including BMD phenotypes and intrinsic and extrinsic environmental covariates, were collected on 471 members of eight multi-generational families with a mean family size of >50 individuals and with nearly 3500 relative pairs. Each family had at least three to four generations with phenotypic data. These large families combined with QCT measurements enabled us to begin to disentangle the relative contributions of genes (heredity) and environment to the determination of integral volumetric BMD and volumetric BMD in the cortical and trabecular compartments and to compare our results with those from other populations. Our analyses revealed substantial overall heritability of areal and volumetric BMD with heterogeneity by skeletal site and bone compartment.
Sex-specific areal BMD at the femoral neck was higher at all ages among Afro-Caribbeans than in non-Hispanic blacks or whites in the U.S. NHANES III cohort (Fig. 1). Except for the youngest and oldest age groups, mean areal BMD at the femoral neck among Afro-Caribbean women was almost as high as or higher than BMD of U.S. black men. The higher BMD in Afro-Caribbeans than in non-Hispanic blacks may reflect lower admixture in the Afro-Caribbeans compared with U.S. blacks, 6% versus 20%, respectively.(19,25) Although areal BMD at the femoral neck in U.S. black men is almost as high as that in our Afro-Caribbean family members at 18–29 years of age, it seems to decrease with age to a larger extent among U.S. black men. Between the 18–29 and 60+ age groups, mean femoral neck BMD decreased 11.1% in Afro-Caribbean men, 15.6% in U.S. white men, and 20.5% in U.S. black men. Although the total percent difference in femoral neck areal BMD also was lowest in Afro-Caribbean women (14.3%), the percent decline for U.S. black women (18.9%) was slightly less than that for U.S. white women (20.6% femoral neck).(24)
Lumbar spine areal BMD increased across the lifespan among Afro-Caribbean men. Similar age-related increases in lumbar spine areal BMD have been noted in other studies.(3) Manifestations of spinal degenerative disease such as disc space narrowing, vertebral endplate sclerosis, and osteophytes at the spinous processes and facet joints are prevalent among the elderly and may be more common among older men than women.(4,26) Such degenerative changes are likely to increase the apparent BMD measured in the posterior-anterior position.(8,27,28)
Skeletal site–specific differences in the decline in volumetric BMD over life may be attributable to differences in bone composition (e.g., the relative composition of trabecular and cortical bone), mechanical loading (e.g., weight-bearing versus non–weight-bearing), or other factors.(29) In men and women of European ancestry, volumetric trabecular BMD at both the central and peripheral skeleton decreases with increasing age, and this decrease begins before midlife in both sexes. In contrast, volumetric cortical BMD at both the central and peripheral skeleton does not seem to change in either sex before midlife, after which there is a dramatic decrease, and this decrease is greater in women than in men.(30) The overall age patterns in volumetric BMD measured at the radius and tibia in our Afro-Caribbean families are similar to those reported by Riggs et al.(30) in U.S. whites: men had higher mean trabecular BMD than women at both bone sites, and mean trabecular BMD decreased earlier and across all age groups. Furthermore, similar to a previous report,(30) men had lower peak cortical BMD at both skeletal sites, and mean cortical BMD did not decrease until after midlife in either men or women. The percent difference between the youngest and oldest age groups was almost twice as high among U.S. whites as Afro-Caribbeans, respectively, for trabecular BMD at the radius (∼29% versus 16%) and the tibia (∼24% versus 12%). Similarly, for cortical BMD, the overall loss between the youngest and oldest age groups was <4% for both Afro-Caribbean men and women at the radius and tibia, whereas U.S. men lost ∼17% and U.S. women lost ∼27% at these sites.(30) The reasons for these apparent differences in loss of areal and volumetric BMD across the lifespan among these ethnic groups are unknown, but they could be caused in part by differences in environmental factors (e.g., prevalence of smoking and alcohol use are lower among these Afro-Caribbeans than among U.S. whites and blacks), genetic variation, or genotype by environment interactions.
We also studied the heritability of areal and volumetric BMD in our Afro-Caribbean families. Unlike several previous family studies that considered only age, sex, and BMI when estimating heritability, but similar to a study of areal BMD in Mexican Americans,(13) we examined a large number of potential covariates for BMD. In general, four covariates, age, sex, height, and weight, accounted for most (73–99%) of the variation in areal and volumetric BMD caused by measured covariates. For three traits, areal BMD at the lumbar spine and cortical BMD at the radius and tibia, inclusion of other covariates, especially menopausal status, dramatically increased the proportion of total phenotypic variation caused by measured covariates. Although menopausal status is significantly associated with almost all of the areal and volumetric BMD traits, the strongest and most significant associations were with lumbar spine BMD and cortical BMD at the radius and tibia.
We found that smoking was associated with decreased trabecular, but not areal, BMD or total or cortical volumetric BMD. Several cross-sectional studies(31–33) have found lower BMD among current compared with never smokers. Male smokers had 0.3 SD lower femoral neck BMD compared with never smokers in a meta-analysis of five published studies, an effect size that was similar to that observed among women.(31) Compared with never smokers, past smokers had 5% or 0.3 SD lower calcaneal BMD independent of important covariates.(34) Men 20–29 years old who smoked had 10% lower hip BMD compared with nonsmokers, suggesting that smoking may also reduce peak BMD.(35) The lower BMD among current smokers persisted after adjusting for important covariates including age, body weight, alcohol intake, and physical activity. However, our sample size was too small to look for sex by smoking effects as reported by Deng et al.(36)
Similar to some investigators,(37,38) but not others,(39) we observed the slightly counterintuitive result that ever-pregnant status (i.e., parity) was associated with increased areal BMD. Although pregnancy is known to be associated with transient decreases in BMD, the observation that increased parity in associated with increased BMD is not well understood, although it may be caused by increases in body weight, intestinal calcium absorption, and later age at menopause.(26) Further investigation in our study indicated that ever-pregnant status was associated with increased cortical, but not trabecular, BMD at both skeletal sites.
As reported by others,(40,41) we found that diabetes was associated with increased areal and volumetric BMD. The diabetes association was stronger for trabecular than cortical BMD at both the radius and tibia. Furthermore, as described above, we observed that smoking was associated with decreased trabecular but not cortical BMD, and parity was associated with increased cortical, but not trabecular, BMD at both skeletal sites. Thus, unlike analyses of areal BMD, in which conflicting results are often obtained, analyses of volumetric BMD may reveal that different bone compartments are being influenced by specific environmental factors, and this knowledge may lead to a better understanding of the possible mechanisms involved.
Except for cortical volumetric BMD, residual heritabilities of areal and volumetric BMD in our Afro-Caribbean families ranged from 0.55 (for femoral neck) to 0.70 (for radius trabecular BMD). Similar to family studies in other population groups,(13,42) genes accounted for a larger proportion of the total variation than did measured covariates in our Afro-Caribbean families. We do not know why heritability of cortical volumetric BMD at the radius and tibia was considerably lower than that for trabecular volumetric BMD or areal BMD. The proportion of variation caused by measured covariates was similar across all volumetric and areal BMD traits, indicating that perhaps unmeasured covariates may have a larger effect—or perhaps that we are not adequately powered to model the effects of some covariates. However, we did not observe any difference in the magnitude of the effect of measured covariates on loaded (tibia) versus unloaded (radius) bones, indicating that the potential effects of unmeasured covariates may not be large.
In conclusion, our study provides the first comprehensive genetic epidemiologic analysis of volumetric BMD measured by QCT, and the first analysis of these traits in extended families of African descent. Our findings reveal that genes account for as much or more of the total variation in areal and volumetric BMD than do environmental factors, but also that the magnitude of the effect of genetic and environmental factors differs between trabecular and cortical bone. Identification of the genetic and environmental determinants of both trabecular and cortical bone mass could reveal novel insights into the etiology of osteoporosis.
This study was supported by Grants R03-AR050107 and R01-AR049747 from the National Institute of Arthritis and Musculoskeletal Diseases.
- 302004 Population-based study of age and sex differences in bone volumetric density, size, geometry, and structure at different skeletal sites. J Bone Miner Res 19: 1945–1954., , , , , , , , ,Direct Link: