Bone mineral accrual from 8 to 30 years of age: An estimation of peak bone mass

Authors


Abstract

Bone area (BA) and bone mineral content (BMC) were measured from childhood to young adulthood at the total body (TB), lumbar spine (LS), total hip (TH), and femoral neck (FN). BA and BMC values were expressed as a percentage of young-adult values to determine if and when values reached a plateau. Data were aligned on biological ages [years from peak height velocity (PHV)] to control for maturity. TB BA increased significantly from −4 to +4 years from PHV, with TB BMC reaching a plateau, on average, 2 years later at +6 years from PHV (equates to 18 and 20 years of age in girls and boys, respectively). LS BA increased significantly from −4 years from PHV to +3 years from PHV, whereas LS BMC increased until +4 from PHV. FN BA increased between −4 and +1 years from PHV, with FN BMC reaching a plateau, on average, 1 year later at +2 years from PHV. In the circumpubertal years (−2 to +2 years from PHV): 39% of the young-adult BMC was accrued at the TB in both males and females; 43% and 46% was accrued in males and females at the LS and TH, respectively; 33% (males and females) was accrued at the FN. In summary, we provide strong evidence that BA plateaus 1 to 2 years earlier than BMC. Depending on the skeletal site, peak bone mass occurs by the end of the second or early in the third decade of life. The data substantiate the importance of the circumpubertal years for accruing bone mineral. © 2011 American Society for Bone and Mineral Research

Introduction

There is now almost universal consensus that early-life experiences are important for reducing the risk of osteoporosis in later life.1 The principal cause of osteoporotic fractures is reduced bone mass, which can result from age-related bone loss and/or failure to achieve optimal peak bone mass in the growing years.2 Peak bone mass thus is a key determinant of skeletal health throughout life. The importance of maximizing bone mineral accrual during the growing years is clearly recognized.3, 4 For example, it has been reported that 60% of the risk of osteoporosis can be explained by the amount of bone mineral acquired by early adulthood.5 Childhood and adolescence is a particularly important time to maximize bone accrual because the skeleton undergoes rapid change owing to the processes of growth, modeling, and remodeling. In both males and females, it is generally believed that bone mass increases substantially during the first two decades, reaching a plateau (referred to as peak bone mass) in the late-teen or young-adult years.6 However, the timing when this occurs is still disputed. Data from some studies suggest that it is reached as early as 20 years of age,7, 8 but other data suggest that bone mineral is still being accrued into the third decade of life.9, 10 The tempo of bone accrual also has been shown to vary by site, with peak bone mass, for example, occurring earlier at the hip and spine than at the whole body.11

Knowledge of the age of attainment of peak bone mass is central to understanding its relationship to subsequent fracture risk; however, there are limited data on the tempo and timing for the accrual of bone area (BA) and bone mineral content (BMC) during the adolescent period relative to peak bone mass. The differences observed with regard to BA development and BMC accrual at different ages reflect to a certain degree the cross-sectional versus longitudinal designs of previous studies and the inability of many to control for the effects of maturation. During adolescence, there is a sudden increase in statural growth (a growth spurt), changes in body size and shape, and changes in the relative proportions of muscle, fat, and bone.12 Puberty also takes place during adolescence and is the process by which sexual maturation occurs and reproductive capacity is attained. Puberty usually starts in girls around 11 years of age and in boys closer to 12 years of age. However, there is wide variation in the timing of normal puberty, with the onset varying by as much as 4 to 5 years among normal healthy boys and girls.12 During puberty, there is a considerable increase in bone mass as a result of increased bone size, which depends on both bone length and bone width.11, 13 Furthermore, there is a dissociation in timing between peak gains in height and bone mineral accrual that are clearly related to maturational age.14 Therefore, to address the question pertaining to bone accrual in adolescence, studies have to be of a longitudinal nature, and subjects must be aligned on a biological rather than chronological time scale.

Beginning in 1991, we have monitored BA development and BMC accrual in a sample of (then) children who are now in their midtwenties. Thus we have a unique opportunity to describe BA development and BMC accrual during adolescence relative to young-adult BA and BMC values. In an earlier article, we estimated the percent of bone mineral accrued over the 2 years surrounding peak BMC velocity to represent about 26% of the total-body BMC.15 At that time, we only had data representing the adolescent growth period to 1-year after peak bone mineral accrual—not adult values. The percentage (of adult) accrual value therefore was not based on actual data but was estimated based on previous literature, by assuming that adult total-body BMC would be approximately 2200 g in females and 2800 g in males. Since publication of those data, we have continued to measure this cohort of subjects and now have up to 4 years of measurements of actual BA and BMC during young adulthood. We hypothesized that peak bone mass would be achieved in the late teens or early twenties. The specific purpose of this article is to describe the development of BA and accrual of BMC at the total body, lumbar spine, total hip, and femoral neck during adolescence, both prior to and after attainment of peak height velocity (PHV), relative to young-adult BA and BMC.

Materials and Methods

Study design

The Saskatchewan Pediatric Bone Mineral Accrual Study (PBMAS) used a mixed or multiple longitudinal cohort design, where repeated measurements were taken on more than one birth cohort or cluster. In 1991, eight chronological age clusters (entry ages 8 to 15 years) were identified, and measurements were taken for 7 successive years. Additional subjects were recruited in 1992 and 1993. Testing stopped in 1997 but was reinitiated in 2002, a break of 5 years, and measurements were taken for up to a further 4 occasions between 2002 and 2007 (Table 1). During the course of the study, the composition of these clusters remained roughly the same. The smallest numbers are found in the oldest and youngest clusters; numbers also drop between 18 and 22 years of age, reflecting the 5-year break in testing. Using this approach makes it possible to estimate a 22-year development pattern in a period of 16 years (1991–2007).

Table 1. Number of Subjects by Age Group and Year of Test
Age19911992199319941995199619972002/032003/042004/052006/07Total
  1. Male (Females)

83 (7)5 (11)0 (2)        8 (20)
910 (15)3 (9)5 (10)0 (2)       18 (36)
1019 (16)10 (18)3 (10)5 (10)0 (2)      37 (56)
1117 (12)18 (15)12 (18)3 (9)5 (10)0 (2)     55 (66)
1221 (18)17 (12)16 (15)12 (14)3 (10)5 (10)0 (2)    74 (81)
1317 (21)21 (16)17 (11)14 (16)11 (14)3 (10)5 (7)    88 (95)
1417 (16)15 (20)20 (15)14 (11)14 (16)10 (12)3 (9)    93 (99)
153 (8)17 (15)14 (20)14 (15)12 (11)10 (16)8 (8)    78 (93)
16 3 (7)17 (14)13 (12)13 (14)11 (7)8 (11)    65 (65)
17  3 (7)13 (11)13 (12)12 (12)9 (6)0 (2)   50 (50)
18   3 (6)13 (8)12 (11)8 (9)3 (3)0 (2)  39 (39)
19    3 (6)8 (6)7 (8)4 (13)3 (2)0 (2) 25 (37)
20     4 (3)5 (3)6 (13)2 (12)5 (4) 22 (35)
21      2 (2)6 (9)5 (7)2 (9)0 (2)15 (29)
22       12 (9)6 (11)6 (10)4 (7)28 (37)
23       8 (10)10 (8)7 (9)3 (7)28 (34)
24       14 (14)9 (12)10 (7)7 (10)40 (43)
25       8 (6)11 (13)8 (14)5 (12)3 2 (45)
26       3 (5)9 (6)10 (11)7 (8)29 (30)
27       1 (0)4 (5)9 (5)12 (11)26 (21)
28         3 (3)11 (14)14 (17)
29          9 (5)9 (5)
30          2 (2)2 (2)
Total107 (113)109 (123)107 (122)91 (106)87 (103)75 (89)55 (65)65 (84)59 (78)60 (74)60 (78)875 (1035)

Participants

The PBMAS has been described in detail previously.16 In brief, in 1991, of the 375 eligible students attending two elementary schools in the city of Saskatoon (population 200,000), the parents of 228 students (113 boys and 115 girls) provided written consent for their children to be involved in this study, and 220 (age range 8 to 15 years) underwent dual-energy X-ray absorptiometry (DXA) annually (Table 1). Between 1992 to 1993, an additional 31 subjects were recruited and scanned. By 1994, 197 individuals (91 males, 106 females) had been measured on two or more occasions. By 1997, the median number of scans per subject was 6. Between 2002 and 2007, 169 subjects were measured on at least one further occasion; 64% on all 4 occasions, 18% on 3 occasions, and 7% on 2 occasions. For the present analyses, subjects required measures in childhood and at least one measurement in adulthood; 75 males and 89 females fulfilled these criteria and contributed to the data for this study. Ninety-eight percent were white and of middle class socioeconomic status. Ethical approval was obtained from the University of Saskatchewan Biomedical Research Ethics Board (88-102), and all participants provided written informed consent.

Anthropometry and somatic maturation

During childhood and adolescence, anthropometric measurements were taken at 6-month intervals by trained personnel following a standard protocol.17 During young adulthood, anthropometric measurements were taken at 12-month intervals. Standing heights were recorded without shoes as stretch stature to 0.1 cm using a wall-mounted stadiometer. Body mass was measured to 0.01 kg on a calibrated electronic scale.

Controlling for maturation

To control for the well-documented maturational differences between adolescent boys and girls of the same chronological age, we determined a biological maturity age for each individual. The age of peak linear growth [age at peak height velocity (PHV)] is an indicator of somatic maturity representing the time of maximum growth in stature during adolescence. To establish age at PHV for each child, whole-year velocity values were calculated for each subject by dividing the difference between the annual distance measurements by the age increment (the mean age increment was 0.998 ± 0.048 years). A cubic spline fit then was applied to the whole-year velocity values for each child. A spline is interpolating polynomials, which uses information from neighboring points to obtain a degree of global smoothness. The cubic spline procedure was chosen over other curve-fitting protocols because it maintains the integrity of the data without transforming or modifying the underlying growth characteristics. A biological maturity age (years) was calculated by subtracting the chronological age at the time of measurement from the chronological age at PHV. Bone values were considered in terms of time before and after PHV. Thus a continuous measure of biological age (BAge) was generated. Biological age categories were constructed using 1-year intervals such that the −1 PHV age group included observations between −0.49 and −1.50 years from (ie, before) age at PHV.

Bone mineral measures

Bone area (BA, cm2) and bone mineral content (BMC, g) of the total body (TB), anteroposterior lumbar spine (LS; L1–L4), total hip (TH), and femoral neck (FN) were measured by DXA (Hologic QDR 2000; Hologic, Inc., Waltham, MA, USA). The array mode was employed for all scans, and enhanced global software Version 7.10 was used. Software Version 5.67A analyzed the TB scans, whereas the scans of the TH, FN, and LS were analyzed with software Version 4.66A. The in vivo coefficient of variation for BMC was 0.60%, 0.61%, 0.63%, and 0.91% and for BA was 0.87%, 0.69%, 0.87%, and 1.60% for the TB, LS, TH, and FN, respectively.18 The BA and BMC accrual values were aligned in biological age categories, representing biological ages surrounding the age of PHV [ie, −4 to −1, 0 (PHV), +1, +2, . . . , +15 years from PHV). The young-adult values were defined as the highest value recorded during up to four successive DXA measurements taken during the adult testing years (2002–2007) (Table 1).

Statistical analysis

Descriptive statistics were calculated for each variable: means, standard deviations (SD), and 95% confidence intervals (CI). For each individual, the percentage of adult accrual was calculated by dividing the BA or BMC at each biological age by the highest adult BA or BMC value. Paired t tests were used to compare changes between biological age categories. All data were analyzed using SPSS for Windows Version 16.0 (SPSS, Inc., Chicago, IL, USA). Significance level was set at p < .05. Alpha was adjusted using the Bonferroni technique for multiple comparisons; for example, for 16 comparisons, the α corrected for experiment-wise error rate would be 0.05/15 = 0.003.

Results

Physical characteristics of the subjects by sex and biological age category are shown in Table 2. Age and weight increased with increasing biological age (BAge), with females maturing earlier than males; at PHV (BAge = 0), the average age was 11.8 (±1.0) years in females and 13.5 (±1.0) years in males. Significant differences were found between chronological age at each biological age between males and females (p < .05), reflecting the earlier maturational pattern of females. Stature increased significantly until 3 years after PHV (p < .05) for both males and females; a similar pattern was found for weight.

Table 2. Age, Weight, and Height of Subjects by Biological Age (BAge, Years from PHV) and Sex
BAge (years from PHV)MalesFemales
NAge (years)Ht (cm)Wt (kg)NAge (years)Ht (cm)Wt (kg)
  • Values are means (SD). BAge = biological age [years from peak height velocity (PHV = 0)]; Ht = height; Wt = weight.

  • *

    p < .05 (with Bonferroni correction) between biological age groups.

−4219.8 (0.9)*140.0 (5.9)*32.0 (3.6)*148.6 (0.4)*133.2 (7.7)*28.5 (7.1)*
−33310.6 (0.9)*145.9 (5.2)*37.5 (5.8)*269.3 (0.7)*136.3 (7.4)*30.6 (6.4)*
−24111.7 (0.9)*150.3 (5.0)*41.4 (6.6)*3410.1 (0.8)*141.1 (7.5)*34.2 (8.0)*
−15412.5 (0.9)*155.4 (6.1)*45.7 (7.7)*4710.8 (1.0)*145.5 (7.8)*37.2 (8.1)*
06413.5 (1.0)*163.8 (7.2)*52.8 (8.7)*6611.8 (1.0)*153.0 (7.7)*42.3 (8.6)*
16614.3 (1.0)*172.5 (7.0)*60.2 (9.0)*7312.8 (0.9)*160.1 (7.1)*50.3 (10.2)*
26015.3 (1.0)*177.2 (7.1)*67.2 (9.7)*7713.8 (0.9)*163.4 (6.0)*55.0 (10.1)*
34716.1 (1.0)*178.9 (7.7)*71.2 (10.2)*7214.7 (1.0)*165.0 (5.8)*58.1 (9.9)*
43917.1 (1.0)*178.6 (7.5)73.5 (11.0)5915.6 (1.0)*165.4 (5.8)60.6 (10.0)
52818.0 (1.1)*179.5 (7.7)75.7 (13.0)5116.6 (1.0)*165.6 (6.1)62.2 (11.6)*
62619.2 (1.1)*178.8 (7.0)77.5 (11.7)3918.0 (0.8)*165.6 (6.2)64.1 (12.4)
72220.3 (1.1)*179.2 (5.5)80.1 (9.8)3119.0 (0.7)*167.6 (7.5)68.6 (14.6)
82421.4 (1.1)*179.5 (7.1)83.6 (12.3)2820.2 (0.8)*167.5 (6.7)69.8 (15.2)
92322.9 (1.0)*180.6 (7.1)81.0 (13.2)2521.0 (0.8)*167.0 (6.3)67.9 (13.6)
103023.2 (1.1)*179.9 (7.2)84.2 (13.2)3322.0 (1.0)*166.0 (6.7)66.0 (13.6)
113424.5 (1.1)*178.8 (7.5)84.0 (14.2)3322.7 (1.2)*165.0 (5.9)62.0 (8.9)
123625.1 (1.1)*180.4 (7.7)87.8 (14.0)4323.4 (1.1)*165.9 (6.5)69.6 (13.6)
132626.3 (1.0)*177.8 (7.8)82.5 (13.9)3724.7 (0.8)*165.6 (6.5)69.0 (15.8)
142026.8 (0.9)*178.7 (7.2)85.6 (13.6)4125.5 (0.9)*165.1 (6.0)69.8 (18.5)
15    2226.8 (0.7)*164.7 (5.1)70.1 (16.4)

The BA values at each skeletal site and the percentages of the adult values attained at each biological age category are displayed in Tables 3 and 5 for males and females, respectively. The BMC values at each skeletal site and the percentages of the adult values attained at each biological age category are displayed in Tables 4 and 6 for males and females, respectively. In males, TB BA (Table 3) increased significantly from −4 to +4 years from PHV (p < .05); in contrast, percentage accrual of adult values of TB BMC (Table 4) increased significantly from −4 years from PHV until +6 years from PHV (p < .05), showing a plateau occurring 2 years after a plateau in TB BA. No significant differences (p > .05) in accrual were found after +4 or +6 years from PHV for either BA or BMC, respectively. No further development in LS BA was seen after +3 years from PHV, and no accrual in LS BMC was seen after +4 from PHV (p > .05), a pattern repeated in TH BA development and TH BMC accrual. A plateau in FN BA development was observed at +1 year from PHV, whereas FN BMC percentage accrual increased significantly between −4 years from PHV and +2 years from PHV (p < .05). No significant increases (p > .05) were seen after +2 years from PHV in FN BMC development (Table 4). This pattern of BA development and BMC accrual was mirrored in females (Tables 3 through 6), although FN BA showed a plateau a year later in females at +2 years from PHV. 5

Table 3. Males Bone Area (BA) and Cumulative Percentage of Adult Accrual for Each Site by Biological Age (BAge, Years from PHV)
BAge (years from PHV)TBBA (cm2)NPTBBA, %LS BA (cm2)NPLSBA, %TH BA (cm2)NPTHBA, %FN BA (cm2)NPFNBA, %
  • Values are means (SD). BAge = biological age [years from peak height velocity (PHV = 0)]; TB BA = total-body bone area (cm2); PTBBA = percentage adult total-body bone area (%); LS BA = lumbar spine bone area (cm2); PLSBA = percentage adult lumbar spine bone area (%); TH BA = total-hip bone area (cm2); PTHBA = percentage adult total-hip bone area (%); FN BA = femoral neck bone area (cm2); PFNBA = percentage adult femoral neck bone area (%).

  • *

    p < .05 (with Bonferroni correction) between proceeding BAge groups.

−41173 (133)2147 (4)36.4 (5.0)1054 (7)21.4 (3.6)2149 (6)4.20 (0.38)2172 (6)
−31341 (159)*3352 (5)*40.1 (3.4)*2258(5)*23.6 (3.4)*3353 (6)*4.38 (0.36)3375 (5)*
−21461 (160)*4158 (5)*43.3 (3.8)3361 (5)*26.0 (2.8)*4159 (6)*4.49 (0.33)4177 (5)
−11594 (188)*5463 (6)*45.8 (5.0)*4065 (4)*28.7 (3.7)*5464 (5)*4.68 (0.37)*5480 (4)*
01816 (216)*6472 (6)*51.9 (5.1)*5474 (5)*33.3 (4.3)*6475 (6)*5.05 (0.42)*6486 (5)*
12076 (211)*6682 (5)*59.1 (5.7)*5884 (5)*38.1 (4.5)*6685 (5)*5.46 (0.45)*6693 (5)*
22282 (196)*6089 (4)*64.6 (5.7)*5891 (3)*40.6 (4.4)*6090 (4)*5.61 (0.45)6095 (4)*
32369 (201)*4792 (3)*67.5 (6.5)*4794 (3)*41.2 (4.2)4792 (4)5.67 (0.41)4797 (3)
42402 (209)*3995 (3)*67.7 (6.5)3996 (3)*41.0 (4.3)3992 (4)5.68 (0.42)3997 (3)
52461 (226)2895 (2)69.6 (5.9)2897 (3)41.9 (4.5)2894 (4)5.74 (0.41)2897 (3)
62476 (199)2697 (2)68.6 (5.4)2697 (3)41.6 (4.1)2695 (4)5.62 (0.44)2697 (4)
72503 (146)2298 (2)68.1 (5.7)2297 (7)41.8 (4.0)2296 (4)5.59 (0.40)2296 (4)
82524 (184)2499 (1)69.8 (7.2)2497 (7)43.1 (3.7)2497 (3)5.64 (0.62)2497 (5)
92495 (227)2399 (2)70.3 (6.9)2298 (2)43.6 (4.5)2297 (4)5.72 (0.28)2298 (2)
102545 (196)3099 (1)71.8 (7.1)3099 (2)43.4 (4.7)3096 (4)5.70 (0.39)3097 (4)
112516 (220)3498 (2)69.5 (6.3)3499 (2)44.5 (4.6)3497 (3)5.80 (0.35)3498 (2)
122598 (210)3699 (1)72.6 (6.1)3698 (3)44.4 (4.9)3697 (4)5.80 (0.39)3697 (3)
132476 (248)2699 (1)69.0 (6.8)2599 (1)41.4 (4.4)2594 (4)5.68 (0.38)2598 (2)
143540 (215)2099 (1)69.8 (6.1)2099 (2)42.9 (4.8)2096 (4)5.74 (0.45)1998 (2)
Table 4. Males' Bone Mineral Content (BMC) and Cumulative Percentage of Adult Accrual for Each Site by Biological Age (BAge, Years from PHV)
BAge (years from PHV)TB BMC (g)NPTBBMC (%)LS BMC (g)NPLSBMC (%)TH BMC (g)NPTHBMC (%)FN BMC (g)NPFNBMC (%)
  • Values are means (SD). BAge = biological age [years from peak height velocity (PHV = 0)]; TB BMC = total-body bone mineral content (g); PTBBMC = percentage adult total-body bone mineral content (%); LS BMC = lumbar spine bone mineral content (g); PLSBMC = percentage adult lumbar spine bone mineral content (%); TH BMC = total-hip bone mineral content (g); PTHBMC = percentage adult total-hip bone mineral content (%); FN BMC = femoral neck bone mineral content (g); PFNBMC = percentage adult femoral neck bone mineral content (%).

  • *

    p < .05 (with Bonferroni correction) between proceeding BAge groups.

−4977 (164)*2133 (4)*21.2 (4.0)*1029 (6)*15.4 (3.5)*2132 (5)*3.00 (0.38)*2151 (6)*
−31131 (193)*3337 (4)*24.9 (4.0)*2234 (5)*17.4 (3.8)*3335 (6)*3.15 (0.45)*3354 (7)*
−21243 (192)*4143 (5)*28.4 (5.0)*3337 (5)*19.4 (3.5)*4141 (6)*3.28 (0.42)*4159 (7)*
−11407 (234)*5448 (6)*30.7 (5.9)*4041 (5)*22.3 (4.2)*5446 (5)*3.51 (0.45)*5463 (7)*
01674 (324)*6457 (7)*38.3 (7.5)*5451 (7)*27.9 (5.8)*6458 (8)*4.02 (0.60)*6471 (8)*
12071 (339)*6670 (7)*49.8 (8.8)*5867 (8)*35.6 (6.6)*6675 (8)*4.72 (0.64)*6685 (9)*
22438 (352)*6081 (6)*59.6 (9.3)*5879 (7)*41.0 (7.3)*6085 (7)*5.16 (0.67)*6092 (8)*
32599 (351)*4787 (5)*64.9 (10.3)*4786 (7)*42.1 (7.2)*4788 (7)*5.36 (0.63)4797 (7)*
42698 (385)*3991 (4)*67.1 (10.0)*3991 (7)*42.8 (7.2)3991 (7)*5.47 (0.61)39100 (7)
52828 (433)2893 (5)70.1 (10.4)2893 (6)44.6 (7.7)2893 (7)5.60 (0.80)28101 (7)
62908 (426)*2696 (4)*71.5 (11.6)2694 (5)46.3 (7.7)2696 (5)5.72 (0.82)26100 (6)
72980 (307)2297 (3)72.6 (10.1)2294 (7)47.0 (6.1)2296 (4)5.72 (0.67)2298 (5)
83000 (239)2499 (2)76.4 (11.5)2496 (6)48.3 (5.3)2498 (3)5.69 (0.78)2497 (6)
92903 (454)2399 (2)74.4 (14.6)2298 (3)47.0 (9.3)2297 (5)5.45 (0.91)2298 (3)
102945 (399)3099 (2)75.6 (13.8)3098 (3)46.1 (8.4)3096 (5)5.50 (0.77)3098 (3)
112959 (456)3499 (2)74.0 (13.7)3498 (3)49.0 (9.9)3497 (4)5.51 (0.77)3498 (4)
123064 (435)3697 (5)77.3 (12.9)3697 (3)47.4 (9.6)3696 (5)5.50 (0.74)3697 (3)
132911 (461)2698 (8)71.9 (13.3)2597 (3)44.5 (8.9)2594 (6)5.24 (0.84)2597 (4)
143034 (416)2099 (2)74.4 (12.3)2097 (2)46.4 (7.4)2095 (5)5.45 (0.70)1997 (4)
Table 5. Females Bone Area (BA) and Cumulative Percentage of Adult Accrual for Each Site by Biological Age (BAge, Years from PHV)
BAge (years from PHV)TB BA (cm2)NPTBBA (%)LS BA (cm2)NPLSBA (%)TH BA (cm2)NPTHBA (%)FN BA (cm2)NPFNBA (%)
  • Values are means (SD). BAge = biological age [years from peak height velocity (PHV = 0)]; TB BA = total-body bone area (cm2); PTBBA = percentage adult total-body bone area (%); LS BA = lumbar spine bone area (cm2); PLSBA = percentage adult lumbar spine bone area (%);TH BA = total-hip bone area (cm2); PTHBA = percentage adult total-hip bone area (%); FN BA = femoral neck bone area (cm2), PFNBA = percentage adult femoral neck bone area (%).

  • *

    p < .05 (with Bonferroni correction) between proceeding BAge groups.

−41013 (204)1447 (5)34.2 (4.6)1054 (4)20.1 (2.3)1456 (3)3.92 (0.40)1473 (5)
−31084 (177)2651 (5)35.2 (4.7)1757 (4)20.8 (3.0)2660 (5)4.04 (0.37)2677 (6)*
−21200 (191)*3456 (5)*38.0 (4.4)*2762 (3)*23.2 (3.1)*3467 (5)*4.20 (0.37)3480 (5)*
−11314 (217)*4762 (6)*40.9 (5.4)*3567 (4)*24.7 (3.4)*4772 (6)*4.36 (0.40)4784 (5)*
01492 (217)*6671 (6)*44.0 (6.1)*4774 (6)*27.7 (3.4)*6682 (6)*4.51 (0.36)*6688 (6)*
11749 (226)*7381 (6)*50.0 (6.6)*5983 (5)*30.2 (3.1)7288 (5)*4.80 (0.37)*7293 (5)*
21880 (201)*7789 (5)*53.7 (5.6)*7090 (4)*30.8 (2.9)*7692 (6)*4.94 (0.35)*7696 (5)*
31964 (185)*7292 (4)*55.7 (5.6)*7094 (4)*31.0 (2.7)7293 (6)4.94 (0.34)7296 (4)
42015 (178)*5994 (4)*56.7 (5.3)5996 (4)30.9 (2.6)5992 (4)4.89 (0.34)5996 (5)
52024 (184)5195 (3)57.0 (4.8)5197 (4)31.0 (2.9)5193 (6)4.90 (0.37)5196 (4)
62058 (206)3997 (3)57.9 (5.1)3998 (3)31.5 (3.2)3993 (6)5.02 (0.36)3997 (4)
72116 (237)3197 (2)58.4 (5.7)3197 (5)32.5 (3.1)3195 (5)5.01 (0.37)3196 (5)
82124 (231)2898 (2)59.3 (6.8)2899 (2)33.2 (3.8)2896 (5)5.03 (0.27)2896 (4)
92104 (183)2599 (2)56.8 (5.5)2397 (3)33.0 (2.6)2397 (5)5.04 (0.34)2397 (4)
102078 (229)3399 (1)58.1 (6.0)3397 (3)32.5 (3.1)3396 (4)5.01 (0.39)3397 (4)
112037 (165)3399 (1)58.2 (5.6)3398 (2)32.4 (3.4)3396 (5)4.96 (0.32)3397 (4)
122092 (186)4399 (2)58.2 (5.2)4398 (2)32.1 (3.1)4396 (4)5.00 (0.42)4397 (6)
132092 (213)3798 (3)58.6 (5.3)3598 (3)32.3 (3.1)3596 (4)4.91 (0.41)3596 (6)
142085 (199)4198 (2)57.8 (6.0)4198 (3)32.3 (3.1)4197 (5)4.90 (0.32)4197 (4)
152108 (192)2399 (1)57.7 (6.3)2299 (2)31.8 (2.7)2296 (4)4.95 (0.41)2296 (6)

The percent of adult BMC accrued at each biological age category for the specific skeletal site is presented in Fig. 1 (males) and 2 (females). As shown, the peak accrual rates for each site occurred within the first year after PHV. In the 3 years around PHV (−1.5 to +1.5 years from PHV), 22% of TB BMC (Figs. 1A and 2A) was laid down in both males and females, and 39% was accrued in the 5 years surrounding PHV (−2.5 to +2.5 years from PHV). At the LS (Figs. 1B and 2B), 27% (males) and 25% (females) were accrued over the 3 years surrounding PHV (−1.5 to +1.5 years from PHV), and 43% (males) and 46% (females) were accrued over the 5 years surrounding PHV (−2.5 to +2.5 years from PHV). The accrual rates at the TH were 28% (males) and 25% (females) in the 3 years surrounding PHV (−1.5 to +1.5 years from PHV; Figs. 1C and 2C), and 43% (males) and 46% (females) were accrued in the 5-year period (−2.5 to +2.5 years from PHV). In the 3 years around PHV (−1.5 to +1.5 years from PHV), in both males and females, 22% of FN BMC was accrued and 33% was accrued in the 5 years surrounding PHV (−2.5 to +2.5 years from PHV; Figs. 1D and 2D).

Figure 1.

Male percentage adult accrual within 1-year biological age categories (years from peak height velocity) for four sites: (A) total-body bone mineral content (TB BMC); (B) lumbar spine bone mineral content (LS BMC); (C) total-hip bone mineral content (TH BMC); (D) femoral neck bone mineral content (FNBMC). Mean and 95% confidence intervals.

Figure 2.

Female percentage adult accrual within 1-year biological age categories (years from peak height velocity) for four sites: (A) total-body bone mineral content (TB BMC); (B) lumbar spine bone mineral content (LS BMC); (C) total-hip bone mineral content (TH BMC); (D) femoral neck bone mineral content (FNBMC). Mean and 95% confidence intervals.

Discussion

The precise age of peak bone mass has been difficult to define owing to a lack of longitudinal data, lack of control for maturational differences, different measurement modalities, and different skeletal sites assessed across previous studies. The results presented in this article are unique because they are derived from mixed longitudinal data representing an age range from 8 to 30 years. The design also allowed us to align subjects on a common maturational point. Because of the variability among subjects, an absolute bone area (BA) and bone mineral content (BMC) mean value of 100% for a specific skeletal site was not expected. This variability likely results from measurement error inherent in DXA and changing soft tissue distribution.19

Our data are restricted to an upper age range of approximately 30 years; thus we cannot demonstrate, unequivocally, that our subjects had achieved peak bone mass (PBM). However, there was no significant increase in BA development at any site after 5 years after PHV or in BMC accrual at any site after 7 years after PHV; therefore, we conclude that a plateau in PBM had been achieved. As would be expected, the bone first plateaus in area and then, roughly 1 to 2 years later, plateaus in mineralization. We estimate that the plateau in TB PBM was likely reached by 7 years after PHV in both males and females. In terms of chronological age, this equates to about 18.8 and 20.5 years of age in females and males, respectively (ie, age at PHV is 11.8 and 13.5 years for females and males, respectively). At the LS and TH, PBM was reached 5 years after PHV for females (chronological age of 16.8 years) and males (chronological age of about 18.5 years). Both males and females reached peak BMC at the FN by 3 years after PHV (chronological age of about 14.8 years for females and 16.5 years for males). As shown in these results, there was little difference between males and females in the biological ages of attaining a plateau in PBM at the various sites. This should not be surprising because it is reasonable to assume that skeletal changes are affected by the actions of sex steroids, growth hormone (GH), and insulin-like growth factors released and active during the adolescent growth period.20

Our results are consistent with the previous literature, although many previous reports are based mostly on cross-sectional data. The prevailing view, which seems to be sustained in the literature, that peak bone mass may be achieved by 30 years of age arose from the pioneering study of Recker and colleagues (1992) suggesting that PBM is achieved in the third decade of life.9 However, their objective was not to study the achievement of PBM from adolescence but the accrual of BMC in young adults, so their subjects entered the study at the mean age of 21 years (probably after the achievement of PBM based on our data). In another cross-sectional sample of females, Lin and colleagues reported that peak BMC at the hip occurred by 23 years of age, but BMC at the spine continued to increase into the early thirties.21 Nonetheless, early work by Matkovic and colleagues suggested that most of the bone mass is acquired by late adolescence.22 More recently, PBM reportedly was achieved before the age of 20 in Croatian women.23 In a study on males, PBM occurred by 18 to 20 years at the TB, FN, and LS but had not yet been reached at the distal radius or tibia.24 In another large cross-sectional sample in boys and girls, BMC and BMD at all skeletal sites had not reached a plateau by ages 16 to 17, suggesting that PBM had not yet been achieved.25 These data, however, did not include subjects beyond age 17; thus estimates of the age of PBM could not be made.

Results from other studies have suggested that PBM in the axial skeleton may be achieved by the late teens and that peak values for both LS and FN are achieved shortly after sexual maturation.11, 26 We found FN values to occur substantially earlier (3 years after PHV) than the other sites, but LS peaks were similar to the TH, with TB occurring 2 years later.

There are a paucity of data, based on longitudinal study designs, with which to compare our accretion rates. However, in a 5-year longitudinal study on Chinese girls, 951 g of BMC was accrued between the ages of 10 and 14.9 years, with a mean accretion rate of 197.4 g/year.27 Peak BMC accrual occurred about 1 year after PHV. In this study, calcium intake was only 444 mg/d—well below recommended intakes in Western societies. In a previous study, in calcium-replete participants, we reported peak bone mineral accrual values of 407 and 322 g/year for boys and girls, respectively,13 with peak accretion occurring at 14 years for boys and 12.5 years for girls. These peak accretion rates occurred approximately ½ year after the age of PHV. As noted by Zhu,27 the accretion values for their females were reasonably similar to ours (233 and 197 g/year) despite having less than 50% of the recommended calcium intake in their diets. These results suggest that these Chinese girls with comparatively low calcium intake compensated by increasing calcium absorption efficiency.

As evident in Tables 4 and 6 (and Figs. 1 and 2), BMC accretion at the FN began slightly earlier than at the other sites. This observation is consistent with the finding that the FN reached its peak BMC at an earlier age than the other sites (ie, at 3 years after PHV). Previously, we estimated that about 26% of (estimated) adult TB BMC and 32% of LS BMC were accrued over the 2 years surrounding the age of peak bone mineral accrual velocity (PBMCV).15 In the present data, we found that about 22% of the adult TB BMC and about 25% to 27% of the adult LS BMC were laid down in the 2 years surrounding PHV. The age of peak BMC accrual occurs about ½ year (0.66 year) after the age of PHV.14 Hence, if our original data were aligned on PHV rather than PBMCV, the results would be similar. Regardless of how these data are aligned, however, the main point is that a substantial amount of BMC is laid down over the adolescent growth period and over the entire 5 years of adolescent growth surrounding PHV. Depending on the skeletal site, 33% to 46% of the adult BMC is accrued. In females, this total accrual represents double the amount of bone mineral that subsequently will be lost during the postmenopausal years from age 50 to 80 years.28

Table 6. Females' Bone Mineral Content (BMC) and Cumulative Percentage of Adult Accrual for Each Site by Biological Age (BAge, years from PHV)
BAge (years from PHV)TB BMC (g)NPTBBMC (%)LS BMC (g)NPLSBMC (%)TH BMC (g)NPTHBMC (%)FN BMC (g)NPFNBMC(%)
  • Values are means (SD). BAge = biological age [years from peak height velocity (PHV = 0)]; TB BA = total-body bone area (cm2); PTBBA = percentage adult total-body bone area (%); LS BA = lumbar spine bone area (cm2); PLSBA = percentage adult lumbar spine bone area (%);TH BA = total-hip bone area (cm2); PTHBA = percentage adult total-hip bone area (%); FN BA = femoral neck bone area (cm2), PFNBA = percentage adult femoral neck bone area (%).

  • *

    p < .05 (with Bonferroni correction) between proceeding BAge groups.

−4765 (208)*1434 (4)*20.7 (6.0)*1032 (3)*12.0 (2.3)*1436 (04)*2.37 (0.47)*1452 (5)*
−3851 (182)*2638 (4)*20.7 (5.7)*1734 (3)*13.0 (2.9)*2639 (05)*2.50 (0.44)*2655 (6)*
−2950 (199)*3443 (4)*23.7 (5.7)*2738 (3)*14.8 (3.3)*3446 (06)*2.63 (0.46)*3460 (7)*
−11057 (228)*4749 (6)*26.6 (6.6)*3543 (5)*16.3 (3.6)*4752 (07)*2.81 (0.49)*4765 (8)*
01275 (259)*6659 (7)*31.1 (8.6)*4751 (6)*20.1 (5.0)*6664 (08)*3.13 (0.54)*6673 (9)*
11596 (320)*7371 (7)*41.2 (10.1)*5967 (8)*25.1 (5.3)*7277 (10)*3.72 (0.60)*7283 (8)*
21807 (280)*7782 (7)*48.3 (8.7)*7079 (8)*27.3 (4.6)*7686 (10)*4.07 (0.60)*7693 (9)*
31947 (290)*7288 (6)*52.6 (9.6)*7087 (8)*28.3 (4.9)*7289 (09)*4.21 (0.62)7297 (8)
42032 (293)*5992 (5)*55.2 (10.0)*5990 (6)*28.7 (4.8)*5991 (09)*4.31 (0.65)5999 (8)
52075 (304)*5194 (5)*56.3 (10.0)5194 (6)29.2 (4.9)5192 (08)4.36 (0.65)51100 (6)
62092 (313)*3995 (4)*57.4 (9.6)3995 (6)29.4 (5.6)3992 (08)4.34 (0.70)39100 (7)
72174 (399)3197 (3)57.9 (12.7)3196 (6)30.6 (5.5)3194 (07)4.34 (0.72)3199 (7)
82197 (394)2897 (3)59.8 (11.9)2898 (5)32.5 (7.5)2896 (06)4.35 (0.73)2895 (4)
92180 (269)2599 (2)56.3 (8.6)2396 (5)30.8 (5.4)2397 (08)4.35 (0.74)2397 (6)
102170 (392)3399 (1)58.9 (12.1)3397 (4)30.5 (7.1)3396 (05)4.21 (0.87)3397 (4)
112063 (261)3399 (2)57.3 (9.4)3397 (3)29.1 (4.5)3397 (05)4.07 (0.62)3397 (4)
122192 (320)4399 (2)58.5 (8.9)4397 (6)30.6 (5.5)4396 (04)4.25 (0.72)4397 (7)
132196 (323)3799 (2)59.5 (10.0)3597 (6)30.3 (5.2)3596 (05)4.17 (0.63)3596 (6)
142130 (317)4199 (2)57.9 (10.8)4197 (5)29.9 (5.1)4196 (07)4.12 (0.68)4197 (4)
152652 (299)2399 (2)60.4 (13.0)2297 (7)31.2 (4.9)2298 (05)4.25 (0.61)2296 (6)

Although it's believed that attaining a high peak bone mass in early life predicts a higher bone mass and hence reduced fracture risk in later life,1 there is some controversy. For example, it has been postulated that bone strength should be the primary consideration for fracture risk.29 Certainly, it is the whole bone structure, including mass, geometry, and architecture that together determine bone mechanical competence30; however, bone mass is an important component of bone strength, and there is a clear relationship between bone mass and fracture at all ages.1 It also has been argued that there is yet no clear evidence that adult bone mass acquisition is directly related to bone mass decades later.31 In contrast, there is evidence that bone mass tracks from childhood through adolescence32 and at least until early adulthood.33 Some have shown that late menarche results in lower bone mass at weight-bearing sites and that these effects (which occur at the time of peak bone accrual) are still evident 25 years later.34 These results were confirmed in this cohort, where we have shown that during adolescence, late menarche is associated with lower PBMCV35 and that from 8 to 30 years of age, late-maturing females (indexed by PHV) had compromised BMC accrual compared with their early- and average-maturing peers.36 There are also data that suggest that bone strength indices (at the proximal femur) in adolescence are related to strength indices in adulthood.37 The increasing understanding of the role of genetics in skeletal health and fracture risk also supports the conjecture that factors such as bone mass should track over time. Family and twin studies have shown that peak bone mass and bone turnover are largely regulated by genetic factors38 and that fracture risk may be programmed as early as intrauterine life,39 and control of bone phenotypes associated with fragility fractures in the elderly is expressed early in life.2, 40 Thus, although there are no definitive longitudinal studies, there is increasing evidence that bone strength indices (including bone mass) attained during childhood and adolescence as the result of both genetic and environmental factors are related to bone strength and fracture risk in later life.

Our results are limited to middle-class white males and females. Ethnic differences in bone acquisition during growth have been reported,41 and recent data from the United States have shown that BMC and bone mineral density values are greater in black children than in nonblacks.25 Another limitation to our study is that the results could be specific to our cohort and the environment to which they have been exposed. We have shown previously that our cohort was within normal ranges for height and weight compared with age-matched reference standards.42 This is important because we have shown previously that body composition plays a major role in the prediction of BMC accrual in adolescence.43 We also have shown previously in this cohort that physical activity plays a major role in predicting BMC accrual both in adolescence15 and young adulthood44 and that nutritional status plays an independent role in BMC accrual.45 Comparisons of our cohort with other cohorts is problematic, but we have no reason to believe that this cohorts' nutrition and habitual physical activity patterns are significantly different from others. Thus, although our cohort appears to be of similar height, weight, and body composition and has similar patterns of nutrition and physical activity shown in other observational studies, in this analysis, we are unable to show whether interactions have taken place and thus influenced the timing of BMC accrual.

In summary, our data provide strong evidence that, depending on site, PBM occurs by the end of the second decade or very early in the third decade. These data also confirm that the peri- and postadolescent growth periods are critical times for bone mineral acquisition in both males and females because, in total, this represents double the amount of bone mineral that will be lost subsequently during the postmenopausal years from 50 to 80 years of age.28

Disclosures

All the authors state that they have no conflicts of interest.

Acknowledgements

This work was supported in part by grants from the Canadian National Health and Research Development Program (NHRDP), the Canadian Institute of Health Research (CIHR; MOP-57671), the Saskatchewan Health Research Foundation (SHRF), and the University of Saskatchewan.

Ancillary