| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Surgical and Perioperative Sciences, Sports Medicine (P.N., A.N.), Department of Community Medicine and Rehabilitation, Geriatrics (P.N.), and Department of Community Medicine and Rehabilitation, Rehabilitation Medicine (A.N.), Umeå University, 901 85 Umeå, Sweden; and Obesity Unit M73 (M.N.), Department of Medicine, Karolinska Institute, Karolinska University Hospital, 171 77 Stockholm, Sweden
Address all correspondence and requests for reprints to: Peter Nordström, M.D., Ph.D., Sports Medicine Unit, Department of Surgical and Perioperative Sciences, Umeå University, 901 85 Umeå, Sweden. E-mail: peter.nordstrom{at}idrott.umu.se.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective: The objective of the study was to investigate the changes in BMD in men between 17 and 26 yr of age.
Design: This was a longitudinal study.
Participants: Participants included 107 healthy males with a mean age of 17 yr at baseline. BMD was also measured in 81 of their fathers at a mean age of 50 yr.
Main Outcome Measures: BMD of the total body, proximal femur, and lumbar spine was measured at baseline and after mean periods of 27, 66, and 92 months in the young cohort.
Results: BMD of the total body and lumbar spine initially increased to reach a plateau during the study period. At the proximal femur, peak values were obtained at 19 yr of age, followed by significant losses of almost 0.02 g/cm2 per year (P < 0.001). At this site, the fathers’ BMD indicated a further loss of about 0.0085 g/cm2 per year up to the age of 50 yr. The BMD development at all sites was positively associated with physical activity (P < 0.05). However, changes in physical activity, weight, and height did not explain the local BMD loss seen at the proximal femur.
Conclusion: Early losses of BMD at the proximal femur were found in this male sample. The results indicate that 25% of peak BMD at this site may be lost by the age of 50 yr in men. We suggest that bone remodeling may be regulated differently at the proximal femur than at other sites.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A recent study identified bone mineral density (BMD; grams per square centimeter) as a key determinant of future risk of hip fractures in men (6). Every SD of decreased BMD was associated with a 3-fold increase in risk of fracture. The bone mineral accrual period leading up to peak BMD has been suggested to play a vital role in preventing osteoporosis (7, 8). However, there is uncertainty regarding the timing of peak BMD and the extent to which BMD is lost at different sites. The general conception is that peak BMD is attained in the end of the second decade of life, or in third decade of life, when it reaches a plateau (9, 10, 11). However, this has not been well investigated longitudinally in men.
Thus, the aim of this study was to investigate the timing of peak BMD at different anatomical sites over 8 yr in a cohort of young men as well as the influence exerted by physical activity and growth on BMD development.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Beginning in 1994, we recruited 117 healthy adolescent Caucasian males from two high schools and badminton and ice hockey clubs in Umeå in northern Sweden for this longitudinal study, which is also called the Northern Osteoporosis and Obesity Study. The recruited subjects were volunteers and not randomly selected from the general population. None of the subjects had any medication or diseases known to affect bone metabolism at baseline. The initial aim was to investigate the effects of physical activity on the development of BMD (12, 13, 14, 15) and body composition parameters such as fat mass (15, 16). After a mean period of 2 yr 3 months, a first follow-up was conducted. Of the original participants, 109 (93%) young men could be located and volunteered to participate. One subject was diagnosed with anorexia nervosa and one subject suffered from severe depression. These two subjects were excluded, leaving 107 subjects (91%) for the present study. A second follow-up was conducted after a mean period of 5 yr 6 months. Of the original sample, 104 participated in this follow-up. The third follow-up was conducted after a mean period of 7 yr 8 months. One hundred two participants (87%) were located and volunteered to participate. Those 15 subjects who did not attend to all follow-ups were not significantly different in age, weight, height, or BMD, compared with the 102 who did.
All data were collected at the Sports Medicine Unit at the University Hospital of Northern Sweden. The participants all gave informed consent, and the Ethics Committee of the Medical Faculty, Umeå University (Umeå, Sweden) approved the protocol.
Measurements
Anthropometry. Weight and height were measured in light clothing. Weight was measured to the nearest 0.1 kg by use of a digital scale, and height was measured to the nearest centimeter against a wall-mounted stadiometer.
Physical activity. Physical activity was measured by questionnaire and defined as the self-reported mean weight-bearing activity associated with sweating or breathlessness each week during the last year. At baseline, the study cohort trained for a mean of 7.3 ± 3.9 h/wk. Besides ice hockey and badminton, the group’s physical activity consisted mainly of playing soccer and floor ball, distance running, and some weight training. The questionnaire used at baseline and follow-ups also included questions on smoking habits, known illnesses, and medication intake. The amount of physical activity was validated for those subjects who were training regularly with their team coach. Pubertal stage according to Tanner was determined by self-examination of pubic hair, combined with questions on growth of beard, and height at baseline. This method has previously been shown to be accurate when determining pubertal development (17). All subjects were found to have passed the growth spurt period and were at least Tanner stage 4.
Parental information. Of a possible total number of 106 fathers, 84 could be located, volunteered to participate, and were considered for inclusion. Three were excluded because of diseases known to affect BMD, leaving 81 fathers for this study. BMD of the total body, proximal femur, and lumbar spine was measured once by the same dual-energy x-ray absorptiometry machine as in the young cohort. The fathers also completed a questionnaire, providing information on physical activity, diseases, current medication, and lifestyle characteristics.
BMD. The BMD (grams per square centimeter) of the total body and BMD and bone mineral content (BMC, grams) of the lumbar spine and right proximal femur were measured in the young cohort at all follow-ups and in their fathers using the same Lunar DPX-L (GE-Lunar, GE Healthcare, Indianapolis, IN) dual-energy x-ray absorptiometer, software version 4.6e. The bone area of the total hip and femoral neck and the bone area and height of the lumbar spine (L2-L4) were measured using the same equipment. To estimate the volumetric bone density (vBMD; grams per cubic centimeter) of the femoral neck and lumbar spine, it was assumed that these sites were cylindrical in shape. The vBMD was then estimated by dividing the BMC by the estimated volume. The coefficient of variation (SD/mean) was determined by scanning one person five times on the same day, with repositioning between each scan. Accordingly, the coefficient of variation values were 0.7% for the total body scan, approximately 1% for the femoral neck/total hip scan, 0.6% for the lumbar spine, 0.7–1.5% for the different bone areas measured, 1.9% for the lumbar spine vBMD, and 0.9% for the neck vBMD. The equipment was calibrated each day using a standardized phantom to detect drifts in the measurements and the test machine functions. The calibrations were normal. The equipment was also evaluated regularly during the study using a spine phantom. No drifts in BMD were detected. About 90% of all scans were performed by the same investigator (P.N.).
Statistical analysis. The SPSS package (version 14.0; SPSS Inc., Chicago, IL) and SAS (version 9; SAS Institute Inc., Cary, NC) were used for the statistical analyses. Data in the text are expressed as means ± SD. Differences between follow-up measurements were investigated using paired-samples t tests, and bivariate correlations were measured by Pearson’s correlation coefficient.
A mixed piecewise linear regression model with a random intercept was fitted with PROC MIXED to model the BMD trajectories over time for each site. This method uses all the repeated measurements, taking the unevenly spaced follow-up measurements and missing data points into account. This enables within-subject changes in BMD to be estimated with great precision because each individual acts as his own control (18). Hereby fixed effects that vary between subjects, but are constant within subjects over time, are accounted for. The analysis was designed to estimate how individuals’ BMD changed throughout the duration of the study and examine potential factors that influence the heterogeneity among individuals. Age, physical activity, weight and height at each time point, and baseline pubertal status were used as covariates in the initial models. A knot was inserted at age 19 yr to allow for the curvature in all the BMD variables. The choice of 19 yr was made after visual inspection of the mean BMD trajectories. A mixed-effects model was also used for assessing whether there were differences in the trajectories of total BMD, neck BMD, hip BMD, and spine BMD after age 19 yr. This was done by fitting an overall model with BMD at all sites as the response variable, age, physical activity, weight, height, pubertal status, and dummy variables for the respective sites as predictors. Interaction terms for site and age after 19 yr were used to assess differences in the development after 19 yr of age. In all models the sandwich covariance estimator was used (18).
Akaike’s information criterion (AIC) was used to compare models, in which smaller values of the AIC represent a better model. P < 0.05 was considered statistically significant.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Table 1
shows the age, anthropometric measures, hours of physical activity, BMD, and bone area measurements, together with changes during the three follow-ups. During the study period, physical activity decreased from a mean of 7.1 to 4.2 h/wk, whereas weight and height increased by 10.6 ± 7.2 kg and 2.1 ± 2.2 cm, respectively (all P < 0.0001; Table 1). At the same time, total BMD and the BMD of the lumbar spine increased and appeared to plateau around age 19 yr (Fig. 1). In parallel, the BMD in the femoral neck and hip reached a plateau around age 19 yr and thereafter started to decrease (Fig. 1). During the next approximately 6 yr of follow-up, significant losses were seen at the femoral neck and total hip (P < 0.001 for both). The estimated vBMD of the femoral neck and lumbar spine increased less before the first follow-up than the corresponding BMD at these sites, with similar patterns of loss later (Fig. 1).
|
|
The BMD at baseline was strongly associated with the corresponding BMD at study end (r = 0.69–0.76, P < 0.001). Furthermore, change in physical activity (r = 0.41–0.51, P < 0.001), weight (r = 0.21–0.50, P < 0.05), and height (r = 0.31–0.53, P < 0.01, except for spine vBMD) were also strong predictors of changes in BMD and vBMD at all sites. Change in height predicted changes in bone area at all sites (r = 0.21–0.48, P < 0.05). Change in physical activity was associated with the bone area of the hip and spine (r = 0.20–0.21, P < 0.05), whereas change in weight was associated with changes in bone area in the hip (r = 0.24, P = 0.01; data not shown).
Mixed-effects model
Physical activity, height, and weight as well as pubertal status at baseline were entered as covariates in mixed-effects models, with a spline at age 19 yr to allow for the change in curvature for the BMD trajectories (Fig. 1
). The unadjusted influence of age on the BMD changes is presented in Table 2. Age was highly significant (all P < 0.0001) in all models, even after adjustment for level and changes in physical activity, weight, and height (Table 3). Greater physical activity was associated with greater BMD, independent of age, weight, height, and pubertal status at baseline for all sites (P < 0.0001 for all but total BMD; P = 0.01; Table 3). Peak bone mass was reached already at age 19 yr in the femur neck and total hip (Tables 1–3). Thereafter BMD was lost at these sites at an annual rate of 0.016–0.019 g/cm2 (Tables 1–3). Total body BMD started to plateau after age 19 yr but did not decrease, whereas the lumbar spine BMD curve was practically flat, after adjustments for changes in physical activity, weight, height, and pubertal status at baseline.
|
|
Parental BMD
The fathers’ anthropometric data and BMD are presented in Table 4. The fathers’ BMD was related to the corresponding values for their sons at all follow-ups (r = 0.25–0.48, P < 0.05). The fathers’ BMD was compared with that of their sons and indicated an annual loss from the age of 25 to the age of 50 yr of 0.0092 (femoral neck), 0.008 (total hip), 0.0057 (lumbar spine), and 0.0036 (total body) g/cm2 (P < 0.001 for all).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Environmental factors such as physical activity and calcium intake and factors influenced by the environment, such as body weight, have previously been associated with BMD and changes in BMD in males during childhood and adolescence (11, 20, 21). In a similar cohort, we previously demonstrated an especially strong association between physical activity and the development of BMD in men (22). In the present study, changes in environmental factors such as amount of physical activity, body weight, and height were strong and independent predictors of the changes in BMD at all sites. However, the correlation pattern between changes in these factors and BMD could not explain the specific BMD loss at the proximal femur because physical activity was an equally strong predictor of spine BMD that did not decrease during the study. In the present study, the fathers’ BMD was associated with that of their sons at all follow-ups. Based on these results, we suggest a novel regulation of bone remodeling of the proximal femur, resulting in early BMD loss at this site.
We found only one previous study (23) with a longitudinal component in which BMD changes were studied in men after the age of 20 yr. Bachrach et al. (23) studied ethnic and gender differences in a large cohort consisting of 423 healthy males and females aged 9–25 yr. Using a mixed-effects model, as in our study, they concluded that BMD in men reached a plateau at 15.7 yr for the total hip and 17.6 yr for the lumbar spine. They presented BMD accumulation data only for the lumbar spine, but the cross-sectional component of their results indicated significant BMD loss at the proximal femur of about 0.008 g/cm2 annually between the ages of 20 and 25 yr.
Some cross-sectional studies have evaluated the timing of peak BMD and later losses in men. Henry et al. (24) investigated the timing of peak BMC and vBMD in 132 Caucasian children (aged 11–19 yr) and 134 adults (aged 20–50 yr). As in our study, their results indicated an early attainment of peak BMC at the hip, with declining BMC between the ages of 20 and 50 yr. In a well-performed large cross-sectional study, Ardawi et al. (25) reported reference values for BMD in 1980 randomly selected Saudis aged 20–79 yr. Regression equations showed an independent annual BMD loss at the femoral neck of 0.012 g/cm2 per year from 20 to 79 yr. The BMD loss in the present study was evaluated between ages 20 and 50 yr by comparing fathers and their sons. From this comparison, BMD at the femoral neck was lost at an annual rate of 0.009 g/cm2. Although this was a cross-sectional comparison, the corresponding figures were significantly lower for the total body (0.0036 g/cm2) and spine (0.0057 g/cm2). Together these data suggest an early and rapid BMD loss at the proximal femur after the age of 20 yr, and that continuous losses may occur of about 1% per year at the proximal femur to the age of 50 yr in men.
In contrast to our study, some previous studies suggest that peak BMD is reached earlier or later (9, 10, 23, 24, 26). This might be explained by cross-sectional vs. longitudinal study designs, attainment of peak BMD at different times at different sites, potential interactions with gender or ethnicity, or different statistical approaches. The influence of environmental factors such as physical activity might also be different and is highlighted in our sample because physical activity was strongly associated with changes in BMD and vBMD at all sites.
Bone enlargement will affect the measured BMD (grams per square centimeter) in a complex manner (27). We therefore estimated vBMD of the lumbar spine and femoral neck, assuming that these sites are cylindrical in shape. From Fig. 1 it seems that peak vBMD at these sites occurred at about the same time as peak BMD, although the gains were smaller between baseline and first follow-up. Because the bone area of the femoral neck and lumbar spine did not change after 19 yr of age, BMC or estimated vBMD changed in a manner exactly similar to BMD at the same site. Periosteal bone deposition could also increase bone size later in life and decrease the risk of fracture by increasing the moment of inertia and, consequently, the breaking strength of the bone. It has been demonstrated that estrogen-related postmenopausal bone loss in women is also related to increased periosteal diameter of the radius (28). Data from the third National Health and Nutrition Examination Survey cross-sectional material also suggested an increase in femoral neck area and diameter in men with increasing age (29). In the present material, we could, however, not demonstrate a difference in area of the proximal femur when comparing the fathers and their sons.
The present study has both strengths and weaknesses. Strengths include the longitudinal study design of a previously well-characterized rather homogenous cohort with repeated measures of BMD. However, the sample studied was not randomly drawn from the general population. Thus, we cannot exclude that selection bias would influence the BMD loss found at the proximal femur, especially because the amount of physical activity was rather high in the beginning of the study. However, the decrease in physical activity does not seem to be abnormal, compared with that in other studies (30). As discussed previously, the decreased amount of physical activity does not seem to explain the specific BMD loss as the proximal femur, which has been suggested in a previous longitudinal study and some cross-sectional studies (23, 24, 25, 29). Also, the availability of the fathers’ BMD, albeit for only one measurement, is also a strength, giving an indication of the changes in BMD between the age of 20 and 50 yr. This comparison would assume that fathers of the present study had a similar peak BMD at 20 yr of age as their sons. Given the known large influence of genetic factors on BMD and the demonstrated relationship between the fathers’ and the sons’ BMD in the present study, this comparison seems more reliable than inferences made from cross-sectional data and unrelated subjects.
In summary, we report an early attainment of peak BMD at the proximal femur in men, with immediate and substantial losses of almost 0.02 g/cm2 per year in the following 5 yr. Based on our results, we suggest that there may be a distinct regulation of bone remodeling at the proximal femur. Furthermore, it seems that the timing of peak BMD and subsequent losses at the proximal femur, as well as at other sites, are highly influenced by physical activity.
|
Footnotes |
|---|
Disclosure Statement: The authors have nothing to disclose.
First Published Online February 20, 2007
Abbreviations: AIC, Akaike’s information criterion; BMC, bone mineral content; BMD, bone mineral density; vBMD, volumetric bone density.
Received November 28, 2006.
Accepted February 9, 2007.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
