Pubertal Timing Predicts Previous Fractures and BMD in Young Adult Men: The GOOD Study†
The authors state that they have no conflicts of interest.
The importance of pubertal timing for adult BMD in males was studied through association of pubertal timing with young adult bone phenotype. Pubertal timing was found to predict both cortical and trabecular volumetric BMD and previous fractures in young adult men. Thus, late puberty is a risk factor for low BMD and previous fractures in young adult men.
Introduction: Peak bone mass (PBM), achieved during puberty, is a determinant of the risk for osteoporosis and future fractures. The role of variations within the normal range in pubertal timing for fractures during pubertal development and for adult bone mass in men is unknown.
Materials and Methods: The aim of this study was to investigate the importance of pubertal timing for adult BMD and for fractures before achievement of PBM in men. The population-based Gothenburg Osteoporosis and Obesity Determinants (GOOD) study is a well-characterized cohort of young adult Swedish males 18–20 years of age. Detailed growth charts from birth to 18–20 years of age were retrieved for 642 men participating in the GOOD study. Age at peak height velocity (PHV) was estimated and used as an assessment of pubertal timing. The skeletal phenotype was analyzed at young adult age using DXA and pQCT and previous fractures were assessed by questionnaires.
Results: Age at PHV was a negative independent predictor of both adult cortical and trabecular volumetric BMD and of total body and radius areal BMD. Moreover, age at PHV was associated with previous fractures in a logistic regression analysis. The OR for cortical osteopenia was 2.49 (95% CI, 1.91–3.24; p < 0.001) and for previous upper limb fractures was 1.35 (95% CI, 1.04–1.75; p < 0.05) per year increment in age at PHV.
Conclusions: Age at PHV is a negative independent predictor of BMD and a positive predictor of previous fractures in young adult men. Longitudinal studies to determine if pubertal timing also predicts BMD and fractures in elderly men are required.
Osteoporosis, characterized by low bone mass and increased susceptibility to fractures, is an economic burden in society and a cause of great suffering to the affected individual. Peak bone mass (PBM), attained mainly during adolescence, is an important determinant of the risk for development of osteoporosis and fractures later in life.(1–4)
The role of pubertal timing for adult BMD in men has recently been widely discussed. Treatments aiming at prolonging the growth period are used to augment height in short adolescents,(5) but whether these treatments affect adult BMD is unclear. It is therefore important to determine if pubertal timing affects adult BMD in men.
The objective assessments of pubertal timing include age at peak height velocity (PHV) and assessment of sexual maturation by Tanner stages. Subjective methods include self-reported recalled age at menarche in women, which correlates rather well with the age at PHV,(6,7) but the method is less certain if there is a long time interval between menarche and recall.(8,9) In men, no reliable method for the recall of self-reported pubertal timing exists and consequently less is known about male puberty.
Previous studies, investigating the role of pubertal timing for the skeleton, have been limited by the use of a 2D measurement (areal BMD [aBMD]) of the bone using the DXA.(10–15) aBMD provided by DXA is dependent on size,(2,16) and therefore, a large bone will be falsely perceived as having higher aBMD. Moreover, DXA does not discriminate between trabecular and cortical bone compartments, which is necessary for measurements of cortical and trabecular volumetric BMD (vBMD) and bone geometry, an important factor for bone strength.(17)
In females, some studies report a role for timing of puberty (subjectively assessed) for adult aBMD and osteoporosis.(13,14) In contrast, less is known about the role of variations within the normal range of pubertal timing for the adult skeleton in men. However, Finkelstein et al.(11,12) reported permanently reduced adult aBMD in males with constitutionally delayed puberty (CDP; defined in the Finkelstein study as onset of puberty after 15 years of age). Bertelloni et al.(10) and Yap et al.(15) showed reduced aBMD in men with CDP, but if aBMD was converted to apparent vBMD, there was no difference between subjects with normal puberty and subjects with CDP. These unclear data may be caused by small sample sizes (n ≤ 32) and by the use of a 2D technique (DXA) instead of a 3D technique such as QCT.
Puberty is not only a period of rapid longitudinal growth and accrual of bone mass, but also a period of high incidence of fractures.(18) The upper limb is the most frequent site of fractures during childhood and adolescence, and an increasing incidence for these fractures has been reported.(19,20) Thus, fracture incidence is bimodal, with peaks later in life caused by osteoporosis and during adolescence for reasons not completely understood.(18) The role of pubertal timing for fractures before the achievement of PBM is unknown.
The aim of this study was to investigate the importance of variations within the normal range of pubertal timing for adult BMD and bone dimensions as well as for self-reported fractures before achievement of PBM in a population-based cohort of 642 young adult Swedish men. We show that pubertal timing predicts previous fractures and BMD in young adult men.
MATERIALS AND METHODS
The Gothenburg Osteoporosis and Obesity Determinants (GOOD) study was initiated with the aim to determine both environmental and genetic factors involved in the regulation of bone and fat mass as previously described.(21) Study subjects were randomly identified using national population registers, contacted by telephone, and asked to participate in this study. One thousand sixty-eight men, 18.9 ± 0.6 (SD) years of age, representative of the greater Gothenburg area, were included. Of the 1068 subjects, complete growth charts for determination of PHV were available for 642 subjects. Age and BMI did not differ between the subset with PHV and the complete cohort, indicating that the subset is representative for the complete GOOD study (data not shown). A standardized questionnaire was used to collect information about amount of present physical activity (hours/week, duration in years), nutritional intake, (dairy products, vegetables and vitamin intake), and smoking. Calcium intake was estimated from dairy product intake. Questionnaires were used to collect data on previous fractures. The information collected on fractures was restricted to prevalence of previous fracture and did not include information on age at fracture.
The GOOD study was approved by the local ethics committee at Gothenburg University. Written and oral informed consent was obtained from all the study participants.
Height and weight were measured using standardized equipment. The CV values were <1% for these measurements.
Bone area (cm2), BMC (g), and aBMD (g/cm2) of the whole body, femoral neck (of the left leg), lumbar spine, and left radius were assessed using the Lunar Prodigy DXA (GE Lunar Corp., Madison, WI, USA). All measurements were performed using a single machine. The CVs for the aBMD measurements were 0.4% for total body, 0.8% for spine, 0.6% for femoral neck, and 2.5% for radius.
pQCT scans were performed using a pQCT device (XCT-2000; Stratec Medizintechnik, Pforzheim, Germany) as previously described.(22) The cortical vBMD (mg/cm3), cortical BMC (mg/mm), cortical cross-sectional area (CSA; mm2), and cortical thickness (mm) were measured using a scan through the diaphysis (at 25% of the bone length in the proximal direction of the distal end of the bone), and trabecular vBMD (mg/cm3) was measured using a scan through the metaphysis of the radius and tibia (at 4% of the bone length in the proximal direction of the distal end of the bone). All measurements were performed using a single machine. The CVs were <1% for all pQCT measurements. It should be noted that the pQCT measures the true volumetric BMD, as opposed to the 2D measurement (aBMD) of the bone using DXA, which is dependent on bone size. Hence, a large bone will be falsely perceived as having higher aBMD.
Estimation of PHV
For 642 subjects, a representative subset of the complete GOOD material, detailed growth and weight charts from birth until 19 years of age, have been used for estimation of PHV according to the infancy-childhood-puberty (ICP) model.(23) Age at PHV was defined as the age at maximum growth velocity during puberty and was estimated by the algorithm. The average number of measurements between birth and 19 years of age was 23. PHV is generally believed to be reached within 2 years after pubertal onset.(23,24)
The independent predictors of the various bone parameters were tested using linear regression analyses nonadjusted (crude) or with height, age at bone analyses, weight, physical activity, calcium intake, and smoking included as covariates (adjusted) together with age at PHV in the regression analyses. Standardized β values were used. To clearly show the differences between early, average, and late puberty, a one-way ANOVA followed by Tukey's posthoc test was used to compare tertiles, and p < 0.05 was considered significant. Binary logistic regression analyses were performed for osteopenia of the radius bone parameters or previous fractures according to PHV (1-year increments), and given as ORs with 95% CIs. The same covariates were used for the logistic regression analyses as described above for the linear regression analyses. For all the statistical analyses, SPSS software (version 13.0) was used.
Descriptive characteristics of age at bone analyses, anthropometrics, age at PHV, and bone variables as measured by DXA and pQCT are given in Table 1. The average age at PHV was 13.6 ± 1.0 years of age, ranging from 10.9 to 16.9 years of age.
Table Table 1.. Anthropometrics, Age at Peak Height Velocity, and Bone Variables
2D DXA measurements
The impact of age at PHV for aBMD in young adult age was first examined using linear regression analyses. Age at PHV was a negative predictor of aBMD at all bone sites studied (Table 2). Linear regression analyses including age at DXA analysis, height, weight, calcium intake, smoking, and present physical activity as covariates revealed that age at PHV was an independent negative predictor of total body and left radius aBMD (Table 2). Subjects were thereafter divided into tertiles based on age at PHV (early, average, and late PHV; Table 3) and aBMD was compared. A one-way ANOVA analysis revealed that age at PHV predicted aBMD at all bone sites measured (Table 3). Posthoc analyses comparing the different PHV tertiles revealed that total body aBMD and radius aBMD were low in the late tertile and high in the early tertile compared with the average tertile. Radius aBMD in the early PHV tertile was increased by 7.9% compared with the late tertile (p < 0.001; Table 3). Measurements of BMC showed similar results as seen for aBMD (data not shown), with high levels in the early PHV tertile and low levels in the late PHV tertile, whereas the respective bone areas were mainly unaffected by age at PHV (Table 3).
Table Table 2.. Age at Peak Height Velocity as a Predictor of aBMD. 2D DXA Analyses
Table Table 3.. Anthropometric and Bone Variables in Young Adult Men According to Early, Average, and Late Peak Height Velocity
3D pQCT measurements
Age at PHV was a strong negative independent predictor of both cortical and trabecular vBMD in both the radius and in the tibia (Table 4). Age at PHV was also a negative independent predictor of cortical bone size, reflected by affected cortical cross-sectional area in the radius and cortical thickness in the radius and tibia (Table 4). However, the impact of age at PHV for cortical and trabecular vBMD was clearly larger than its impact for cortical bone dimensions (Table 4), supported by the finding that PHV explained 13.9% of the variation in cortical vBMD, 4.3% of the variation in trabecular vBMD, but only 1.2% of the variation in cortical CSA of the radius. One-way ANOVA analysis of tertiles based on age at PHV revealed that age at PHV predicted both cortical and trabecular vBMD and cortical thickness in both radius and tibia (Table 3). Posthoc analyses comparing the different PHV tertiles revealed that cortical and trabecular vBMD was lower in the late tertile, whereas cortical vBMD was higher in the early tertile compared with the average tertile (Table 3).
Table Table 4.. Age at Peak Height Velocity as a Predictor of Cortical and Trabecular Bone Variables in the Radius: 3D pQCT Analyses
Age at PHV predicts adult BMD
To quantify the impact of age at PHV for osteopenia (< −1 SD), ORs were computed by logistic regression analyses. These analyses showed that age at PHV was an independent predictor of osteopenia, measured as both aBMD and trabecular and cortical vBMD in the radius. These findings implicate a large impact of age at PHV for adult cortical vBMD but also for trabecular vBMD and aBMD in young adult men (Table 5).
Age at PHV is associated with previous fractures
The total number of subjects with at least one previous fracture was 175 (27.3%), and 71 (11.1%) of these had suffered upper limb fractures, the major bone site for pubertal fractures. Age at PHV was associated with all previous fractures (p < 0.05) and of previous upper limb fractures (p < 0.01; Table 5) even after adjustment in a logistic regression analysis. A possible involvement of BMD for the impact of age at PHV for upper limb fractures was tested after adjustment for radius aBMD. As a result of adjustment for radius aBMD, age at PHV was no longer associated with previous fractures nor previous fractures of the upper limb fractures in the logistic regression analysis (p = 0.439, CI = 1.08 [0.89–1.31] for all fractures and p = 0.084, CI = 0.35 [0.00–8.77] for fractures of the upper limb).
Table Table 5.. Crude and adjusted OR for Osteopenia (Defined as −1 SD) of the Radius and Previous Fractures According to Peak Height Velocity (1-Year Increment)
We studied the role of variations within the normal range in pubertal timing for fractures before achievement of PBM and for adult bone phenotype in a well-characterized population-based cohort of 642 men. Age at PHV was objectively determined using detailed growth charts and was associated with previous fractures and adult bone phenotypes as measured by 2D DXA and by 3D pQCT techniques. The novel main finding is that variations within the normal range in pubertal timing predict previous fractures and adult aBMD and cortical and trabecular vBMD in young adult men.
Previous studies have implicated a significant but weak association between self-reported recalled age at menarche and adult aBMD in women.(13,14) Although the DXA measurements in these studies do not allow for the differentiation of trabecular and cortical bone, the results indicate that pubertal timing, similarly as seen in this study in men, predicts adult aBMD in women. In males, the role of variations within the normal range in pubertal timing has not been pre- viously investigated, but subjects with CDP, a condition defined as pathologically late pubertal onset followed by normal sexual maturation, displayed reduced aBMD.(10–12,15) Finkelstein et al.(11,12) reported permanently reduced aBMD in the radius, spine, and femur in adult men with CDP. Bertelloni et al.(10) and Yap et al.(15) confirmed the finding that subjects with CDP had reduced aBMD, but the difference disappeared after conversion of aBMD to calculated vBMD. The studies on CDP in men were restricted by small numbers of study subjects (n ≤ 32), and by the fact that only 2D measurement of aBMD (DXA) was performed. We present a unique setting with a large (n = 642) population-based cohort where the association between variations within the normal range in pubertal timing, objectively assessed by determination of PHV, and BMD as assessed both by 2D DXA and 3D pQCT techniques and previous self-reported fractures have been studied. This experimental setting allows us to show that the timing of pubertal growth spurt is a negative independent predictor of aBMD as well as of both cortical and trabecular vBMD. Interestingly, the pubertal timing explained as much as 13.9% of the variation in cortical vBMD, indicating that it might be a major, but previously unknown, determinant of adult cortical vBMD.
In addition to trabecular and cortical vBMD, cortical bone geometry is important for the biomechanical properties of the skeleton.(2,17,25) In this study, we showed that pubertal timing, besides strongly predicting vBMD, slightly predicts cortical bone geometry variables such as cortical CSA and cortical thickness in the long bones. We believe that the effect of pubertal timing on cortical bone dimensions is of less physiological importance than its effects on BMD, supported by the finding that pubertal timing only explained 1.2% of the variation in cortical CSA.
Fracture incidence is rather high during childhood and adolescence, supported by the finding that 27.3% of the study subjects in this study reported at least one previous fracture. Pubertal timing predicted previous self-reported fractures in this study. The upper limb is the most frequent site for fractures during puberty,(18) and in this study, 11.1% of the subjects reported at least one previous upper limb fracture. Interestingly, age at PHV was in this study a clear predictor of previous upper limb fractures. After adjustment for radius aBMD, age at PHV no longer predicted previous fractures, indicating that reduced BMD in boys with late puberty might be of importance for the mechanical strength of the skeleton already before PBM is achieved.
It was recently shown that treatment of adolescents with short stature and normally timed puberty with an luteinizing hormone–releasing hormone (LHRH) agonist for 3.5 years prolonged the growth period and increased final height but substantially decreased adult BMD.(5) The present finding in the GOOD study indicates that reduced BMD after treatment with LHRH agonist might be the consequence of delayed cessation of longitudinal growth.(5) Thus, our findings support the notion that treatments aiming at prolonging the growth period should not be routinely recommended to augment height in short adolescents with normally timed puberty because it might result in reduced adult BMD(5) and increased risk for fractures in the upper limb before PBM is achieved.
The main limitations of this study are that we could not determine if the observed predictive role of age at puberty for BMD in young adult men will be preserved and predict fractures in elderly men. The ages at bone analyses in the cohort vary too little to be able to differentiate between a primary permanent effect of age at PHV and a secondary sampling effect where the young men with late age at PHV are still increasing their BMD at the time of study. However, the previously reported finding of permanently reduced BMD in men with CDP(11,12) indicates that the strong predictive value of age at puberty for adult BMD, observed in this study, also will be permanent. Furthermore, when interpreting the associations with cortical thickness and cortical CSA, one should be aware of that the estimation of these parameters are dependent on the algorithm used, the pixel size, and the different density thresholds used. However, the whole procedure for the estimation of cortical thickness, used in this study, has previously been validated(26) and used(21,27) for the analyses of cortical thickness of the diaphysis of the radius. The strengths of this study includes (1) the large number of subjects (n = 642) in a population-based cohort, (2) objective determination of age at PHV, and (3) the use of the 3D pQCT technique for the adult bone analyses.
In conclusion, we showed that variation within the normal range of pubertal timing is a strong determinant of both adult cortical and trabecular vBMD. Our data show that late puberty is a risk factor for low BMD and previous fractures in young adult men and suggest that a minor delay in the pubertal growth spurt, either caused by variations within the normal range or disease or treatment, might result in adult osteoporosis. To study whether late puberty confers increased risk of fractures later in life, longitudinal studies are needed. Thus, these findings support the notion that treatments aimed at prolonging the growth period should be avoided because they might result in reduced adult BMD and an increased risk for fractures before PBM is achieved.
The authors thank Anette Hansevi and Maud Petterson for excellent technical assistance. This study was supported by the Swedish Research Council, the Swedish Foundation for Strategic Research, European Commission Grant QLK4-CT-2002-02528, The ALF/LUA research grant in Gothenburg, the Lundberg Foundation, the Torsten and Ragnar Söderberg's Foundation, Petrus and Augusta Hedlunds Foundation, The Göteborg Medical Society, and the Novo Nordisk Foundation.