Catch up in bone acquisition in young adult men with late normal puberty

Abstract

The aim of this study was to investigate the development of bone mineral density (BMD) and bone mineral content (BMC) in relation to peak height velocity (PHV), and to investigate whether late normal puberty was associated with remaining low BMD and BMC in early adulthood in men. In total, 501 men (mean ± SD, 18.9 ± 0.5 years of age at baseline) were included in this 5‐year longitudinal study. Areal BMD (aBMD) and BMC, volumetric BMD (vBMD) and cortical bone size were measured using dual‐energy X‐ray absorptiometry (DXA) and pQCT. Detailed growth and weight charts were used to calculate age at PHV, an objective assessment of pubertal timing. Age at PHV was a strong positive predictor of the increase in aBMD and BMC of the total body (R² aBMD 11.7%; BMC 4.3%), radius (R² aBMD 23.5%; BMC 22.3%), and lumbar spine (R² aBMD 11.9%; BMC 10.5%) between 19 and 24 years (p < 0.001). Subjects were divided into three groups according to age at PHV (early, middle, and late). Men with late puberty gained markedly more in aBMD and BMC at the total body, radius, and lumbar spine, and lost less at the femoral neck (p < 0.001) than men with early puberty. At age 24 years, no significant differences in aBMD or BMC of the lumbar spine, femoral neck, or total body were observed, whereas a deficit of 4.2% in radius aBMD, but not in BMC, was seen for men with late versus early puberty (p < 0.001). pQCT measurements of the radius at follow‐up demonstrated no significant differences in bone size, whereas cortical and trabecular vBMD were 0.7% (p < 0.001) and 4.8% (p < 0.05) lower in men with late versus early puberty. In conclusion, our results demonstrate that late puberty in males was associated with a substantial catch up in aBMD and BMC in young adulthood, leaving no deficits of the lumbar spine, femoral neck, or total body at age 24 years. © 2012 American Society for Bone and Mineral Research.

BONE MINERAL DENSITY, PERIPHERAL QUANTITATIVE COMPUTED TOMOGRAPHY, MEN, LONGITUDINAL, PEAK HEIGHT VELOCITY

Introduction

The relationship between pubertal onset and adult bone mineral density (BMD) has been mainly studied in women. In postmenopausal women, late menarche has been associated with lower BMD as well as an increased risk of fractures of the forearm, vertebrae, and hip.1–5 More recent studies have shown an association between late menarche and lower BMD in young adulthood and in premenopausal women.6, 7 The influence of pubertal timing on adult BMD in men is not clear. Delayed puberty and constitutional delay of puberty in men has been reported to result in lower BMD in young adulthood.8–10 A study including boys (8–18 years of age) with constitutional delay of growth and puberty found that BMD, measured with dual‐energy X‐ray absorptiometry (DXA), adjusted for bone age increased throughout pubertal stages in a similar manner as in healthy children.11 Less is known about how adult BMD in men is affected by variations in pubertal timing within the normal range. We have earlier reported, in the Gothenburg Osteoporosis and Obesity (GOOD) study of 19‐year‐old males in Sweden, that late but normal puberty was associated with lower BMD and previous fracture in a cross‐sectional study.12 Gilsanz and colleagues13 recently reported that areal BMD (aBMD) at all sites was inversely related to the timing of puberty in teenagers of both genders, in a healthy cohort with variations in puberty within the normal range. In contrast, Jackowski and colleagues14 reported no significant differences in bone mineral content (BMC) in early, middle, and late maturing men in late adolescence. The role of pubertal timing on bone accrual in adult men has not been investigated. In the present study, the aim was to establish if pubertal timing is related to the development of aBMD, volumetric BMD (vBMD), and bone geometry in young adult men, and also to determine if late puberty was associated with remaining low bone mass in young adulthood in men. In order to investigate this, we used a well‐characterized population‐based cohort (from the GOOD study), longitudinally followed and examined by DXA and pQCT at ages 19 and 24 years.

Subjects and Methods

Subjects

The population‐based GOOD study was initiated with the aim to determine both environmental and genetic factors involved in the regulation of bone mass and fat mass. The original GOOD cohort (n = 1068) was found representative of the general young male population in Gothenburg.15, 16 For 641 subjects, a representative subset of the complete GOOD material, detailed growth and weight charts from birth until 19 years of age were obtained in order to calculate peak height velocity, as described.12 Five years later, they were contacted by letter and telephone and invited to participate in the 5‐year follow‐up study, as described.17 Out of the 641 subjects with available peak height velocity (PHV) data, 501 men, 24.1 ± 0.6 years of age, completed the measurements at the follow‐up and were included in the present study. We used a standardized self‐administered questionnaire at baseline and follow‐up, to collect information about smoking (yes/no), present physical activity (hours per week, weeks per year, and duration in years), and nutritional intake. Calcium intake was estimated from dairy product intake. No significant differences were seen between the included (n = 501) and not included (n = 40) subjects in age (included 18.9 ± 0.5 years versus not included 18.8 ± 0.6 years, p = 0.086), height (181.4 ± 6.9 cm versus 181.9 ± 67.5 cm, p = 0.501), weight (72.8 ± 10.8 kg versus 74.4 ± 12.9 kg, p = 0.174), calcium intake (1082 ± 678 mg/day versus 1095 ± 749 mg/day, p = 0.846), PHV (13.5 ± 1.0 years versus 13.7 ± 1.0 years, p = 0.086), or amount of present physical activity (4.2 ± 4.7 hours/week versus 4.9 ± 6.8 hours/week, p = 0.254) at baseline, using an independent samples t test (mean ± SD). A lower percentage of smokers was found among the included than among the not included men (7.2% [36/501] versus 13.6% [19/140], p = 0.017). The follow‐up period was 61.2 ± 2.3 months (mean ± SD); range, 56 to 70 months. The study was approved by the regional ethical review board at the University of Gothenburg. Written and oral consent was obtained from all study participants.

Anthropometrical measurements

Height was measured using a wall‐mounted stadiometer, and weight was measured to the nearest 0.1 kg. The coefficient of variation was below 1% for these measurements.

DXA

aBMD (g/cm²) of the whole body, femoral neck, total hip (of the left leg), lumbar spine, and the left and right radius were assessed using a Lunar Prodigy DXA scanner (GE Lunar Corp., Madison, WI, USA). The coefficients of variation (CVs) for the aBMD measurements ranged from 0.4% to 2.5% at baseline and 0.5% to 3% at follow‐up, depending on site. One person performed all the measurements of the baseline study, and another person performed all the measurements of the follow‐up study. The Lunar Prodigy DXA used at the follow‐up visit was not the same specimen as the one used at the baseline visit. Cross‐calibration between the two Lunar Prodigy DXA machines was performed at the time of follow‐up. Phantom measurements (phantom number 9397) were made regularly with the original Lunar Prodigy DXA, and were stable over time (1.270 g/cm² at the time of the original study and 1.268 g/cm² at the time of the follow‐up study). In the cross‐calibration, 24 young men, 24.7 ± 1.7 years of age, were measured with both Lunar Prodigy DXA devices within a period of 4 days. Regression equations using BMC derived from the baseline instrument as a dependent variable and BMC derived from the follow‐up instrument as an independent variable were calculated from BMC data from the 24 men included in the cross‐calibration. The regression coefficient and constant was 1.034 and −33.9 for BMC of the total body, 0.991 and 2.521 for the lumbar spine, 1.017 and −0.336 for the total hip, 1.008 and 0.011 for the femoral neck, 0.959 and 0.608 for the left radius, and 0.994 and 0.182 for the right radius, respectively. Regression coefficients and constants for BMD values were obtained using the same method, and have been described.17 All DXA BMD and BMC measurements at the follow‐up visit were adjusted using according to these regression equations.

pQCT

A pQCT device (XCT‐2000; Stratec Medizintechnik GmbH, Pforzheim, Germany) was used to scan the distal nondominant forearm (radius). The pQCT was calibrated every week using a standard phantom and once every 30 days using a cone phantom provided by the manufacturer. A 2‐mm‐thick single tomographic slice was scanned with a voxel size of 0.50 mm. The cortical volumetric (not including the bone marrow) BMD (vBMD; mg/cm³), cortical BMC (mg/mm), cortical cross sectional area (CSA; mm²), endosteal and periosteal circumference (EC and PC), 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) of the radius. The threshold for cortical bone was 711. Trabecular vBMD (mg/cm³) was measured using a scan through the metaphysis at 4% of the bone length in the proximal direction of the distal end of the bone. Trabecular vBMD was assessed using the inner 45% of the bone. Length of the forearm was defined as the distance from the olecranon to the ulna styloid process. All measurements were made using the same pQCT device. One person performed all the measurements of the baseline study, and another person performed all the measurements of the follow‐up study. The CVs were less than 1% for all pQCT measurements.

Estimation of age at PHV

Detailed growth and weight charts from birth until 19 years of age were used for estimation of age at PHV according to the infancy‐childhood‐puberty model.18 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 21. PHV is generally believed to be reached within 2 years after pubertal onset.18, 19

X‐ray–verified fractures

The procedure of X‐ray verification of fractures in the GOOD cohort has been described in detail.20 In summary, we searched computerized X‐ray (radiography) registers from hospitals and clinics in the greater Gothenburg area including X‐ray records from 1991 onward, and a central archive containing X‐ray records from earlier dates was searched manually. All fractures that had occurred prior to the follow‐up exam (age 24.1 ± 0.6 years, mean ± SD), were included. At the time of follow‐up, 161 subjects had experienced at least one X‐ray–verified fracture and were classified as fracture subjects. Forty‐five men had reported a fracture that could not be verified in the X‐ray registers and were excluded from the fracture analyses, leaving 295 men that were classified as nonfracture subjects. The corresponding numbers at the baseline exam were 141 subjects with at least one fracture, 335 nonfracture subjects, and 25 excluded subjects with reported fracture that could not be verified in the X‐ray registers.

Statistical analysis

To determine if age at PHV was an independent predictor of the change over 5 years in different bone variables, stepwise multiple linear regression analyses were performed. The percentage of the variation of change in each bone parameter explained (R²) by age at PHV alone and by age at PHV together with all covariates was calculated using the linear regression model. Weight was not normally distributed and was therefore log‐transformed before being entered into the regression analysis. Change in weight was calculated in kilograms. Change in physical activity was calculated in hours/week and has previously been reported to predict changes in bone variables in this cohort.21 Differences in anthropometric characteristics and bone variables between men with early, middle, and late puberty were investigated using one‐way ANOVA with Bonferroni correction for multiple comparisons. Chi square test was used to determine whether or not the distribution of smokers differed between the men with early, middle, and late puberty, and between included and not included subjects, and also to investigate whether the distribution of fractures differed between men with early, middle, and late puberty. The association between prevalent fracture and age at PHV was investigated with binary logistic regression analysis. A p value less than 0.05 was considered significant. The data were analyzed using SPSS software (version 19.0; IBM SPSS, Armonk, NY, USA).

Results

Anthropometrics and age at PHV

The mean age at PHV was 13.5 ± 1.0 (mean ± SD) (range, 10.9–16.5) years. Subjects were divided into tertiles according to age at PHV; early puberty (12.4 ± 0.5 years), middle puberty (13.6 ± 0.3 years), and late puberty (14.6 ± 0.5 years). At both baseline and follow‐up men with late and middle puberty were significantly taller than men with early puberty, and men with late puberty weighed less than men with early puberty (Table 1). At baseline, men with middle puberty spent more time on physical activity weekly than men with late or early puberty. There were no significant differences in age, calcium intake, or smoking (yes/no) (Table 1) between the groups at either the baseline or the follow‐up exam. Characteristics of the whole cohort (n = 501) and the cohort divided into three groups (each with n = 167), at baseline and follow‐up, are presented in Table 1.

1

Anthropometrics at Baseline (18–20 Years of Age) and Follow‐Up (23–25 Years of Age) of the Whole Cohort, and Divided Into Tertiles According to Age at PHV

	Mean (n = 501)	Early (n = 167)	Middle (n = 167)	Late (n = 167)	p (ANOVA)
Age at PHV (years)	13.5 ± 1.0	12.4 ± 0.5	13.6 ± 0.3	14.6 ± 0.5	N/A
Baseline
Age (years)	18.9 ± 0.5	18.9 ± 0.6	19.0 ± 0.5	18.9 ± 0.5	0.48
Height (cm)	181.4 ± 6.7	179.8 ± 6.7	182.3 ± 6.6b	182.0 ± 6.5b	<0.01
Weight (kg)	72.8 ± 10.8	74.8 ± 12.3	72.4 ± 10.1	71.1 ± 9.4b	<0.01
Smoking (%)	7.2	6.0	8.4	7.2	0.70
Physical activity (hours/week)	4.2 ± 4.7	3.5 ± 4.4	4.9 ± 5.3a	4.2 ± 4.3	0.03
Calcium intake (mg/day)	1082 ± 678	1077 ± 652	1148 ± 700	1021 ± 680	0.23
Follow‐up
Age (years)	24.1 ± 0.6	24.0 ± 0.6	24.1 ± 0.6	24.0 ± 0.6	0.49
Height (cm)	181.9 ± 6.7	180.3 ± 6.8	182.7 ± 6.5b	182.7 ± 6.5b	<0.001
Weight (kg)	77.6 ± 11.1	79.5 ± 12.9	77.0 ± 10.3	76.3 ± 9.7a	0.02
Smoking (%)	7.6	10.8	7.2	4.8	0.12
Physical activity (hours/week)	2.8 ± 3.7	2.3 ± 3.3	3.2 ± 4.1	2.9 ± 3.5	0.10
Calcium intake (mg/day)	768 ± 494	792 ± 470	762 ± 489	748 ± 522	0.71

Values are presented as mean ± SD. N = 501. Early, middle, and late puberty were defined by timing of PHV. Differences between groups were investigated by ANOVA followed by Bonferroni post hoc test. Differences in smoking were investigated with chi‐square test.

N/A = not available; PHV = peak height velocity.

^a,bStatistically significant differences between groups are indicated: ^ap < 0.05 and ^bp < 0.01 compared to men with early puberty.

Age at PHV was associated with change in aBMD and BMC in young adulthood

Changes in aBMD and BMC between 19 and 24 years of age were adjusted for follow‐up time and entered in a linear regression model (as a dependent variable) including age at PHV (as a continuous variable), change in weight, change in height, change in physical activity, change in calcium intake, change in smoking habits (started smoking yes/no, stopped smoking yes/no), and age at the baseline exam. Age at PHV was found to be an independent positive predictor of the increase in total body, lumbar spine, and radius aBMD between 19 and 24 years (p < 0.001), and could explain 11.7%, 11.9%, and 23.5% of the variation of change in these bone variables, respectively (Table 2, Fig. 1). For each year age at PHV increased, an additional gain of 0.014 g/cm², 0.022 g/cm², and 0.011 g/cm² in aBMD of the total body, lumbar spine, and radius, respectively, could be observed over the 5‐year follow‐up period. Likewise, age at PHV was found to be an independent positive predictor of the increase in total body, lumbar spine, and radius BMC between 19 and 24 years (p < 0.001), and could explain 4.3%, 10.5%, and 22.3% of the variation of change in these bone variables, respectively (Table 2). Late puberty was also associated with a lesser decrease in total hip and femoral neck aBMD and BMC between 19 and 24 years of age (p < 0.001, Table 2).

2

Age at PHV as Predictor of Change Over 5 Years in Different Bone Variables Between Baseline (18–20 Years of Age) and Follow‐Up (23–25 Years of Age)

	5‐year change	I			II
		B age at PHV	R2 age at PHV (%)	R2 whole model (%)	B age at PHV	R2 age at PHV (%)	R2 whole model (%)
DXA
Total body aBMD (g/cm2)	0.041 ± 0.04	0.014a	11.7	25.0	0.014a	11.7	14.4
Lumbar spine L2–L4 aBMD (g/cm2)	0.051 ± 0.06	0.022a	11.9	20.1	0.022a	11.9	13.7
Total hip aBMD (g/cm2)	−0.023 ± 0.06	0.013a	4.5	14.9	0.013a	4.8	4.8
Femoral neck aBMD (g/cm2)	−0.041 ± 0.07	0.016a	5.6	14.8	0.016a	5.6	5.6
Radius nondominant aBMD (g/cm2)	0.045 ± 0.02	0.011a	23.5	44.9	0.011a	23.5	36.2
Total body BMC (g)	215 ± 165	33.8a	4.3	37.8	36.1a	4.9	9.4
Lumbar spine L2–L4 BMC (g)	3.64 ± 4.0	1.28a	10.5	20.0	1.28a	10.5	14.2
Total hip BMC (g)	−0.63 ± 2.9	0.46a	2.7	12.8	0.47a	2.8	4.0
Femoral neck BMC (g)	−0.11 ± 0.4	0.12a	7.2	14.8	0.12a	7.2	7.2
Radius nondominant BMC (g)	0.87 ± 0.4	0.20a	22.3	39.0	0.20a	22.3	30.8
pQCT
Radius cortical vBMD (mg/cm3)	25 ± 16	5.8a	14.3	27.1	5.8a	14.3	22.8
Radius cortical CSA (mm2)	2.9 ± 3.6	1.0a	8.3	12.2	1.0a	8.3	14.0
Radius cortical thickness (mm)	0.11 ± 0.1	0.03a	4.4	9.6	0.03a	4.4	9.6
Radius periosteal circumference (mm)	0.13 ± 0.6	0.14a	5.7	5.7	0.14a	5.7	9.1
Radius endosteal circumference (mm)	−0.55 ± 0.9	–	–	4.4	–	–	4.4
Radius polar SSI (mm3)	15 ± 17	4.8a	8.2	12.2	4.8a	8.2	13.6
Radius trabecular vBMD (mg/cm3)	6 ± 15	3.7a	6.2	7.3	3.7a	6.2	7.3

Change over 5 years, adjusted for follow‐up time, presented as mean ± SD. N = 501. Linear regression analyses (stepwise) were performed, including (I) age at peak height velocity, change in height, change in weight, change in physical activity (hours/week), change in calcium intake, change in smoking, and age at baseline, and (II) age at peak height velocity, height, log weight, physical activity (in quartiles), calcium intake, smoking (yes/no), and age at baseline.

aBMD = areal BMD; B = unstandardized coefficient B; BMC = bone mineral content; BMD = bone mineral density; vBMD = volumetric BMD; CSA = cross‐sectional area; DXA = dual‐energy X‐ray absorptiometry; PHV = peak height velocity; SSI = strength‐strain index.

a

Statistically significant at p < 0.001.

1

Age at PHV predicts increase in radius aBMD. Higher age at PHV is associated with a larger increase in radius aBMD between 19 and 24 years of age.

Age at PHV was associated with change in vBMD and bone geometry of the radius in young adulthood

The radius was further investigated with pQCT. Changes over 5 years in vBMD and bone geometry were adjusted for follow‐up time and entered in a linear regression model in the same manner as described in the previous section, “Age at PHV was associated with change in aBMD and BMC in young adulthood.” Age at PHV was an independent positive predictor of the increase in cortical vBMD and trabecular vBMD, explaining 14.3% and 6.2% of the variation of change in these bone variables, respectively. For each year of age at PHV increased, an additional gain of 5.8 mg/cm³ in cortical vBMD and 3.7 mg/cm³ in trabecular vBMD of the radius could be observed over the 5‐year follow‐up period. Age at PHV was also an independent positive predictor of the increase in cortical CSA, cortical thickness, periosteal circumference. and polar strength‐strain index (SSI) of the radius (Table 2).

Larger increases in aBMD, BMC, vBMD, and bone geometry in men with late puberty

At the total body (0.05 ± 0.04 g/cm² versus 0.03 ± 0.04 g/cm²), lumbar spine (0.08 ± 0.07 g/cm² versus 0.03 ± 0.05 g/cm²), and radius (0.06 ± 0.02 g/cm² versus 0.03 ± 0.02 g/cm²), men with late puberty gained markedly more in aBMD than men with early puberty (p < 0.001) (Fig. 2). Men with late puberty also lost less in aBMD of the total hip and femoral neck (−0.01 ± 0.07 g/cm² versus −0.03 ± 0.05 g/cm² and −0.02 ± 0.07 g/cm² versus −0.05 ± 0.06 g/cm², respectively, p < 0.001) than men with early puberty (Fig. 2). Likewise, men with late puberty lost less in BMC of the total hip (−0.02 ± 3.7 g versus −0.95 ± 2.2 g, p < 0.01) and femoral neck than men with early puberty, as well as gained significantly more in BMC of the total body, lumbar spine, and radius (p < 0.001) (Fig. 3). As for the pQCT‐derived measurements of the radius, men with late puberty gained more in cortical vBMD (32 ± 16 mg/cm³ versus 20 ± 14 mg/cm³) and trabecular vBMD (11 ± 16 mg/cm³ versus 2 ± 13 mg/cm³) than men with early puberty (Fig. 3). The increase in bone size was also greater in the late group, with larger increases in cortical CSA (4.3 ± 4.0 mm² versus 2.0 ± 3.2 mm²), cortical thickness (0.14 ± 0.1 mm versus 0.09 ± 0.1 mm), and periosteal circumference (0.3 ± 1 mm versus 0.01 ± 1 mm) than the early group (Fig. 4). There were no significant differences between the three groups concerning change in endosteal circumference (early −0.54 ± 0.94 mm, middle −0.50 ± 0.83 mm, and late −0.59 ± 0.96 mm).

2

Change in aBMD of the total body (A), lumbar spine (B), femoral neck (C), and radius (D) in men between 19 and 24 years of age, divided into tertiles of early (n = 167), middle (n = 167), or late (n = 167) puberty according to age at peak height velocity.

3

Change in BMC of the total body (A), lumbar spine (B), femoral neck (C), and radius (D) in men between 19 and 24 years of age, divided into tertiles of early (n = 167), middle (n = 167), or late (n = 167) puberty according to age at peak height velocity.

4

Change in radius cortical vBMD (A), cortical CSA (B), periosteal circumference (C), and trabecular vBMD (D) in men between 19 and 24 years of age, divided into tertiles of early (n = 167), middle (n = 167), or late (n = 167) puberty according to age at peak height velocity.

Catch‐up effect in men with late puberty

At 24 years of age, no significant differences remained between men with early, middle, and late puberty in aBMD or BMC of the total body, lumbar spine, total hip, or femoral neck, or in BMC of the radius, whereas aBMD of the radius still was significantly lower in men with late puberty (Table 3). Of the pQCT measurements at follow‐up, only cortical and trabecular vBMD was significantly lower in men with late puberty, whereas no significant differences were seen in bone size (Table 3). All these associations remained essentially unaltered after adjustment for age, height, log weight, physical activity, calcium intake, and smoking at follow‐up, except for trabecular vBMD, for which the difference between men with early and late puberty at follow‐up was no longer significant after adjustment (Table 3).

3

Differences in Bone Variables Measured With DXA and pQCT at 23–25 Years of Age in Men With Early, Middle, or Late Normal Puberty

	Early (n = 167)	Middle (n = 167)	Late (n = 167)	p1	p2
DXA
Total body aBMD (g/cm2)	1.29 ± 0.1	1.29 ± 0.1	1.28 ± 0.1	0.731	0.869
Lumbar spine L2–L4 aBMD (g/cm2)	1.28 ± 0.2	1.27 ± 0.2	1.29 ± 0.1	0.751	0.769
Total hip aBMD (g/cm2)	1.14 ± 0.2	1.14 ± 0.2	1.15 ± 0.2	0.773	0.610
Femoral neck aBMD (g/cm2)	1.13 ± 0.2	1.12 ± 0.2	1.13 ± 0.2	0.907	0.859
Radius nondominant aBMD (g/cm2)	0.64 ± 0.05	0.62 ± 0.05b,d	0.61 ± 0.05c,f	<0.001	<0.001
Total body BMC (g)	3430 ± 514	3401 ± 455	3403 ± 479	0.832	0.734
Lumbar spine L2–L4 BMC (g)	63.8 ± 11	64.6 ± 10	64.3 ± 11	0.779	0.951
Total hip BMC (g)	42.2 ± 7	42.4 ± 6	42.7 ± 8	0.810	0.821
Femoral neck BMC (g)	6.3 ± 1	6.3 ± 1	6.4 ± 1	0.905	0.924
Radius nondominant BMC (g)	11.0 ± 2	11.0 ± 1	10.7 ± 1	0.189	0.068
pQCT
Radius cortical vBMD (mg/cm3)	1195 ± 17	1189 ± 17a,e	1187 ± 18c,f	<0.001	<0.001
Radius cortical CSA (mm2)	99 ± 12	99 ± 12	98 ± 12	0.551	0.523
Radius cortical thickness (mm)	3.1 ± 0.3	3.0 ± 0.2	3.0 ± 0.3	0.159	0.268
Radius periosteal circumference (mm)	42.0 ± 2.8	42.3 ± 2.7	41.9 ± 2.9	0.501	0.499
Radius endosteal circumference (mm)	22.8 ± 2.8	23.3 ± 2.6	22.9 ± 3.1	0.165	0.247
Radius polar SSI (mm3)	319 ± 62	323 ± 62	313 ± 61	0.341	0.273
Radius trabecular vBMD (mg/cm3)	232 ± 43	225 ± 41	221 ± 40a	0.049	0.135

Values are presented as mean ± SD. N = 501. Early, middle, and late puberty defined by timing of PHV. Differences between groups were investigated by ANOVA (p¹) followed by Bonferroni post hoc test. After adjustment for age, height, log weight, physical activity (in quartiles), calcium intake, and smoking (yes/no) at follow‐up, bone variables were again investigated by ANOVA (p²) followed by Bonferroni post hoc test.

aBMD = areal BMD; BMC = bone mineral content; BMD = bone mineral density; vBMD = volumetric BMD; CSA = cross‐sectional area; DXA = dual‐energy X‐ray absorptiometry; SSI = strength‐strain index.

^a–fStatistically significant differences between groups are indicated ^ap < 0.05, ^bp < 0.01, and ^cp < 0.001 compared to men with early puberty for unadjusted bone variables, and ^dp < 0.05, ^ep < 0.01, and ^fp < 0.001 compared to men with early puberty for adjusted bone variables.

Association between age at PHV and fractures

The 161 men with fracture did not differ from the 295 men without fracture in age, height, weight, amount of physical activity, calcium intake, or smoking status at baseline (data not shown). In total, the 161 men had experienced 215 fractures. The peak incidence of fractures in our cohort occurred at the age of 12 years (Fig. 5). The most common fracture site was the distal forearm, which constituted 29.8% (64/215) of all fractures, followed by fractures of the phalanges of the hand, 19.1% (41/215); fractures of the carpal and metacarpal region, 12.1% (26/215); and fractures of the clavicle, 8.4% (18/215). Age at PHV was associated with all prevalent X‐ray–verified fractures in a logistic regression model including age at PHV (odds ratio [OR], 1.28; 95% confidence interval [CI], 1.06–1.56), as well as age at PHV together with covariates age, smoking status, calcium intake, physical activity, height, and log weight at follow‐up (OR, 1.28; 95% CI, 1.04–1.57). Age at PHV was not significantly associated with prevalent distal forearm fracture (OR, 1.09; 95% CI, 0.83–1.43), the most common fracture in our cohort. In men with early puberty, 43 of 155 (27.7%) subjects had experienced at least one fracture during growth and early adulthood, compared to 56 of 147 (38.1%) men with middle puberty, and 62 of 154 (40.3%) men with late puberty (p = 0.049). Age at PHV was also associated with all prevalent X‐ray–verified fractures at the baseline exam (n = 476) when entered in a logistic regression model including age at PHV (OR, 1.22; 95% CI, 1.01–1.53), as well as age at PHV together with covariates age, smoking status, calcium intake, physical activity, height, and log weight at baseline (OR, 1.24; 95% CI, 1.01–1.53).

5

Fracture incidence. The number of prevalent X‐ray–verified fractures according to age (n = 456).

Discussion

In this study, we aimed to investigate the relationship between pubertal timing and bone development in young adult men. Pubertal timing affects BMD accrual during adolescence,12–14, 22 and has been reported to be associated with BMD in adulthood and old age in women in several, but not all, studies.1, 5–7, 23–26 This relationship has been mainly studied in women, partly because menarcheal age is a rather reliable and easily assessed measure of pubertal timing.27–29 No similar self‐reported measure is available in men. Age at PHV, used in the present study, is an objective assessment of pubertal timing that can be used in both genders. Age at PHV has been shown to be strongly correlated to age at menarche in females.30

In the present study, age at PHV was found to be an independent positive predictor of the increase in total body, lumbar spine, and radius aBMD and BMC between 19 and 24 years, explaining up to 23.5% and 22.5% of the variation in change in radius aBMD and BMC, respectively. Age at PHV also was an independent positive predictor of the increase in cortical vBMD, and cortical bone size increase due to increased periosteal circumference, of the radius. For each year age at PHV increased, an additional gain of 5.8 mg/cm³ in cortical vBMD could be expected, which corresponds to 23% of the mean change over 5 years in cortical vBMD. Likewise, an additional gain of 0.03 mm in cortical thickness could be expected for each year age at PHV increased, corresponding to 27% of the mean change over 5 years.

Pubertal timing and BMD in adolescents

We have reported that age at PHV was negatively associated with aBMD, vBMD, and bone size in 19‐year‐old men.12 Boot and colleagues22 observed in a study of 500 children and adolescents (4–20 years of age) that an earlier age at menarche was associated with higher BMD at the lumbar spine and total body in a substudy including girls who had experienced menarche (n = 143). The relationship between pubertal onset in boys and BMD was not investigated in that study. Jackowski and colleagues14 reported that early‐maturing males had greater BMC at the total body, lumbar spine, and femoral neck at age 13 to 15 years in a cross‐sectional analysis. In contrast, no significant differences in BMC between men with early and late puberty were observed by late adolescence and early adulthood (17–28 years of age).14 In a recently published longitudinal study (n = 78 females, and n = 85 males), Gilsanz and colleagues13 found a strong reciprocal relation between DXA values of the whole body, spine, and upper and lower extremities, and variations in the timing of puberty within the normal range in both genders, concluding that late puberty influences peak bone mass negatively. However, the mean age at follow‐up was 16.2 ± 0.9 years for boys, providing no information about the role of pubertal timing on adult BMD. Thus, the BMD differences related to pubertal timing may not remain later in life, as indicated by the present study. Nevertheless, the lower BMD observed in teenagers with late puberty has clinical relevance, because this is a period of high fracture incidence.20, 31–33 In our cohort, a 28% increased risk of having had a previous fracture during childhood, adolescence, and young adulthood (up until the age of 24 years) for each year age at PHV increased, was observed.

Pubertal timing and BMD in adults

Late menarche has been associated with low BMD and higher fracture risk at several sites in postmenopausal women,1–5 and in premenopausal women early menarche has been associated with high BMD.24, 25 In the present study, no significant differences in aBMD or BMC of the lumbar spine, femoral neck, or total body between men with early, middle, or late puberty at 24 years of age were ascertained. To our knowledge, the only previous study reporting of longitudinal bone accrual in young adult men in relation to pubertal timing was recently published by Jackowski and colleagues.14 They observed in a study with a mixed longitudinal design (age span 8–30 years) that early‐maturing female subjects (n = 121) developed significantly more BMC than middle‐ and late‐maturing females at the total body. In contrast, the amount of BMC developed at the total body, lumbar spine, and femoral neck in males (n = 109) did not differ significantly between maturational groups,14 findings in agreement with the lack of association between age at PHV and BMD of the aforementioned bone sites in our cohort. In the present study, aBMD of the radius was still 4.2% lower in men with late puberty compared to men with early puberty. This deficit was, however, markedly lower than what was observed at the baseline exam (8.4%) (data not shown). Hence, a notable catch‐up in aBMD of the radius occurred over the 5‐year study period. Chevalley and colleagues6, 7 observed an inverse relationship between aBMD of the radius and femoral neck and menarcheal age in young women (20.4 ± 0.6 years of age) and in premenopausal women (45.8 ± 3.4 years of age). When divided into two groups according to menarcheal age, the latter group had significantly lower cortical thickness, cortical density, and total density of the radius measured by high‐resolution pQCT than the early group, although no significant differences were found in trabecular microstructure of the radius.6 Thus, our study and the results provided by Chevalley and colleagues6 indicate that pubertal timing is related to remaining deficits in bone density of the radius in both adult men and adult women.

There are several strengths with the present study. It is a well‐powered longitudinal study of an in‐detail characterized cohort, in which bone traits were measured with both DXA and pQCT. Pubertal timing was assessed with an objective method, minimizing the risk of inaccurate classification. For the fracture analyses, only X‐ray–verified fractures were included. There are also some limitations with the present study. The study population was primarily white, and therefore the results cannot be directly transferred to other ethnicities. The cohort was constituted of men only and thus the conclusions are restricted to the male population. Although we performed a very careful cross‐calibration, we cannot rule out that the use of two different DXA machines at the baseline and follow‐up exams may have affected our results concerning aBMD and BMC. There may also be various difficulties in cross‐calibration at different skeletal sites, which may partly explain the disparity in results of these bone variables at different sites. However, results on longitudinal changes in cortical bone parameters obtained with a single pQCT device at both examinations support our findings from the DXA measurements concerning the results for the radius. Furthermore, the possible cross‐calibration–induced errors would theoretically affect all men, irrespective of PHV group, thus not influencing the associations seen between age at PHV and bone development. Last, even though we had a long period of follow‐up, we were unable to establish whether the difference in aBMD and vBMD of the radius between men with early and late puberty will further diminish or be maintained into old age.

In conclusion, our results demonstrate that late but normal puberty in males was associated with a substantial catch‐up in aBMD, BMC, and vBMD and bone size in young adulthood. No significant deficits in aBMD or BMC of the lumbar spine, femoral neck, or total body were seen at age 24 years in men with late puberty. aBMD of the radius was still significantly lower in men with late pubertal onset, due to lower cortical vBMD but not bone size.

Disclosures

All authors state that they have no conflicts of interest.

Acknowledgements

This study was supported by grants from the Swedish Research Council, the Swedish Foundation for Strategic Research, the European Commission, the Lundberg Foundation, the Torsten and Ragnar Söderberg Foundation, the Petrus and Augusta Hedlund Foundation, the ALF/LUA grant from the Sahlgrenska University Hospital, and the Novo Nordisk Foundation.

Authors' roles: Study design: ML and CO. Study conduct: ML and MN. Data collection: ML, MN, JK, and AD. Data analysis: AD, ML, JK, and CO. Data interpretation: AD, CO, and ML. Drafting manuscript: AD and ML. Revising manuscript content: ML, MN, JK, DM, and CO. Approving final version of manuscript: ML, MN, JK, DM, CO, and AD. AD and ML take responsibility for the integrity of the data analysis.