Longitudinal Tracking of DXA Bone Measures Over 6 Years in Children and Adolescents: Persistence of Low Bone Mass to Maturity

Abstract

Objectives

Early assessment of bone mass may be useful for predicting future osteoporosis risk if bone measures “track” during growth. This prospective longitudinal multicenter study examined tracking of bone measures in children and adolescents over 6 years to sexual and skeletal maturity.

Study design

240 healthy males and 293 healthy females, ages 6–17 years, underwent yearly evaluations of height, weight, body mass index, skeletal age, Tanner stage, and dual-energy x-ray absorptiometry (DXA) bone measurements of the whole body, spine, hip, and forearm for 6 years. All subjects were sexually and skeletally mature at final follow-up. Correlation was performed between baseline and 6-year follow-up measures, and change in DXA Z-scores was examined for subjects who had baseline Z < −1.5.

Results

DXA Z-scores (r = 0.66–0.87) had similar tracking to anthropometric measures (r = 0.64–0.74). Tracking was stronger for bone mineral density (BMD) compared with bone mineral content (BMC) and for girls compared with boys. Tracking was weakest during mid to late puberty, but improved when Z-scores were adjusted for height. Almost all subjects with baseline Z < −1.5 had final Z-scores below average, with the majority remaining below −1.0.

Conclusions

Bone status during childhood is a strong predictor of bone status in young adulthood, when peak bone mass is achieved. This suggests that bone mass measurements in children and adolescents may be useful for early identification of individuals at risk for osteoporosis later in life.

Due to the difficulties in longitudinally studying subjects from childhood to an elderly age, the contention that senile osteoporosis is the result of inadequate bone acquisition during growth lacks supporting data. This notion is supported, however, by knowledge that the prevalence of osteoporosis is substantially lower in men than women, and in subjects of African American than European and Asian descent, due to sex and racial differences in bone mass that are already present in childhood (1–5). Additional support for this concept comes from studies showing a strong resemblance between mother-daughter bone traits and that this resemblance is present even before the daughters have begun puberty (6, 7). If bone loss was indeed the exclusive determinant of late life bone mass, then one would not expect such a strong resemblance in bone traits between girls and mothers.

Peak bone mass is nearly achieved by the end of sexual and skeletal development (8, 9). Early assessment of bone mass is only useful if bone measurements “track” during growth, with children who have low bone mass continuing to have low bone mass as they become young adults. We have previously shown that dual-energy x-ray absorptiometry (DXA) bone measures track longitudinally over 3 years during childhood and adolescence (10). The purpose of this prospective longitudinal multicenter study was to examine tracking of DXA bone measures in subjects from the same large pediatric cohort who were followed for 6 years and had reached sexual and skeletal maturity by the final follow-up. We hypothesized that tracking would be maintained over this longer period.

METHODS

The study sample derived from the Bone Mineral Density in Childhood Study (BMDCS), a multi-center longitudinal study examining bone accretion in a racially diverse cohort of 1,554 healthy boys and girls in the United States (U.S.) (11). Subjects were recruited from July 2002 to November 2003 at five medical centers across the U.S. Consent and assent were obtained from parents/guardians and participants following IRB approved protocols. Detailed information about the study participants and procedures has been published previously (11). Briefly, the BMDCS cohort represents healthy children with height, weight, and body mass index (BMI) between the 3^rd and 97^th percentile and no previous or current conditions that might affect bone acquisition. The subjects were evaluated annually for up to 6 years.

At each visit, height, weight, and BMI were measured, and Z-scores for sex and age were determined using normative data from the Centers for Disease Control and Prevention 2000 growth charts. A pediatrician or pediatric endocrinologist assessed sexual maturation based on breast stage in girls and testicular volume in boys using the Tanner criteria (12). A pediatric radiologist assessed skeletal maturity based on roentgenograms of the left hand and wrist according to the method of Greulich and Pyle (13). DXA measures were obtained on the same day.

This paper examines the 533 subjects who completed the full 6 years of follow-up and were sexually and skeletally mature at final follow-up. Skeletal maturity was defined as closure of the long bone physes, which occurs at a skeletal age of 17 years in boys and 16 years in girls. Of the 1,554 subjects in the original BMDCS cohort, 977 completed the year 6 visit with 533 of these being mature at final follow-up. The mean ± standard deviation (SD) chronological age of the 533 subjects at baseline was 13.2 ± 1.8 years with a comparable skeletal age (13.5 ± 2.2 years) (Table I). At final follow-up, average chronological and skeletal age were 19.2 ± 1.8 and 17.7 ± 0.9 years, respectively.

Table 1.

Baseline characteristics of the study subjects

	Male (N=240)	Female (N=293)	All Subjects (N=533)
Developmental Characteristics
Chronological Age (yr)	13.7 (1.8)	12.8 (1.7)	13.2 (1.8)
Skeletal age (yr)	13.8 (2.3)	13.3 (2.0)	13.5 (2.2)
Sexual maturation
Tanner 1	24 (10%)	26 (9%)	50 (9%)
Tanner 2	47 (20%)	45 (15%)	92 (17%)
Tanner 3	31 (13%)	58 (20%)	89 (17%)
Tanner 4	48 (20%)	74 (25%)	122 (23%)
Tanner 5	90 (38%)	90 (31%)	180 (34%)
Anthropometric Characteristics
Height Z-score	0.2 (0.8)	0.2 (0.8)	0.2 (0.8)
Weight Z-score	0.3 (0.8)	0.5 (0.8)	0.4 (0.8)
BMI Z-score	0.3 (0.9)	0.5 (0.8)	0.4 (0.8)
DXA Z-score
Whole body less head BMC	0.0 (1.0)	0.2 (0.9)	0.1 (1.0)
Whole body less head BMD	0.1 (0.9)	0.2 (0.9)	0.1 (0.9)
AP spine BMC	−0.1 (1.0)	0.1 (1.0)	0.0 (1.0)
AP spine BMD	−0.1 (1.0)	0.0 (1.0)	0.0 (1.0)
Hip total BMC	0.0 (1.0)	0.1 (1.0)	0.1 (1.0)
Hip total BMD	0.0 (1.0)	0.1 (1.0)	0.1 (1.0)
Hip neck BMC	0.0 (0.9)	0.1 (0.9)	0.1 (0.9)
Hip neck BMD	0.0 (1.0)	0.1 (1.0)	0.0 (1.0)
1/3 forearm BMC	−0.1 (0.9)	0.1 (0.9)	0.0 (0.9)
1/3 forearm BMD	−0.1 (1.0)	0.1 (1.1)	0.0 (1.1)

Continuous variables are shown as mean (standard deviation).

Categorical variables are N (%).

Whole body, anterior-posterior (AP) lumbar spine, non-dominant forearm, and left proximal femur DXA scans were performed using Hologic, Inc. (Bedford, MA) bone densitometers (QDR4500A, QDR4500W, and Delphi A models) and analyzed using pediatric software (Hologic version 12.3). Scanner calibration was described previously (11). The precision errors were coefficients of variation <1.4% for BMC and <1% for BMD of the spine and whole body less head, and <1.7% for BMD of all other sites (14). Standard and height-adjusted DXA Z-scores were calculated using the appropriate black or non-black normative data for sex and age as reported previously by the BMDCS (11, 15, 16).

Statistical Analyses

Statistical analyses were conducted using Stata (version 12.0; StataCorp LP, College Station, TX). Pearson correlations were performed between baseline anthropometric and DXA measures and the same measures 6 years later. Correlations were calculated by sex and by Tanner stage at baseline. Pearson correlations were also used to compare the change in BMC and BMD Z-scores over 6 years to the change in Z-scores for anthropometric measures.

To assess further changes over time, we categorized children by their DXA Z-scores at baseline (<−1.5, −1.5 to 1.5, or > 1.5) and calculated the mean Z-score by baseline category for each study year. Cutoffs of −1.5 and +1.5 were used rather than -2.0 and +2.0 because the subjects were required to be between the 3^rd and 97^th percentiles for height, weight, and BMI, resulting in only a small number with Z-scores above 2.0 or below −2.0. Differences in mean Z-scores at the final 6-year follow-up were compared among subjects in the different baseline groups using analysis of variance with Bonferroni post-hoc tests.

RESULTS

Baseline Z-scores for anthropometric and DXA bone measures were close to zero with standard deviations close to one (Table I). The tracking of DXA Z-scores over 6 years (r = 0.66 to 0.87) was similar in magnitude, or slightly greater, than the tracking of anthropometric measures (Table II). In general, tracking was stronger for BMD than for BMC and for height-adjusted DXA Z-scores compared with unadjusted Z-scores. Tracking also tended to be better for girls compared with boys.

Table 2.

Correlation of Z-scores between baseline and 6 years later

		BMC	BMD	Height adjusted BMC	Height adjusted BMD
Boys
Height	.71	---	---	---	---
Weight	.64	---	---	---	---
BMI	.68	---	---	---	---
DXA whole body	---	.66	.69	.71	.76
DXA spine	---	.69	.72	.78	.78
DXA hip total	---	.70	.75	.78	.80
DXA hip neck	---	.68	.74	.75	.77
DXA forearm	---	.73	.73	.80	.78
Girls
Height	.72	---	---	---	---
Weight	.69	---	---	---	---
BMI	.74	---	---	---	---
DXA whole body	---	.68	.74	.70	.81
DXA spine	---	.73	.78	.84	.87
DXA hip total	---	.73	.79	.75	.85
DXA hip neck	---	.70	.79	.75	.82
DXA forearm	---	.79	.79	.85	.83
All Subjects
Height	.71	---	---	---	---
Weight	.66	---	---	---	---
BMI	.71	---	---	---	---
DXA whole body	---	.67	.72	.71	.78
DXA spine	---	.71	.75	.81	.83
DXA hip total	---	.71	.78	.76	.83
DXA hip neck	---	.69	.76	.75	.80
DXA forearm	---	.75	.76	.83	.81

When examined as a function of sexual maturity at baseline, tracking of unadjusted Z-scores was strong before puberty (Tanner 1; boys: r = 0.73 to 0.89, N = 24, girls: r = 0.83 to 0.93, N = 26) and after sexual maturity had been achieved (Tanner 5; boys: r = 0.74 to 0.85, N = 90, girls: r = 0.75 to 0.89, N = 90). However, tracking was weaker during puberty (Tanner 2–4; boys: r = −0.03 to 0.78, N = 126, girls: r = 0.56 to 0.88, N = 177), particularly in Tanner stage 3 (boys: r = −0.03 to 0.65, N = 31; girls: r = 0.60 to 0.75, N = 58) (Figure 1, A; available at www.jpeds.com). Tracking generally improved with height adjustment, especially in Tanner stage 3 (Figure 1, B).

Figure 1.

Correlation between baseline and final (6-year) Z-scores for whole body less head BMC as a function of baseline Tanner stage. Similar trends were observed for other DXA bone measures.

There was a significant relationship between the change in BMC Z-scores and the change in Z-scores for height (r = 0.50 to 0.68) and, to a lesser extent, weight (r = 0.38 to 0.48). A similar, though weaker, relationship was observed for BMD Z-scores (r = 0.31 to 0.53 for height; r = 0.27 to 0.48 for weight). Almost all subjects with poor tracking of BMC Z-scores (change exceeding ±1.5) experienced a change in height in the same direction with a median magnitude of change of 0.3 to 1.7. Similar but weaker trends were observed for BMD Z-scores.

We examined the persistence of values at the tails of the distribution by categorizing children according to their baseline Z-score as LOW (<−1.5), HIGH (> 1.5), or INTERMEDIATE (−1.5 to 1.5). Although regression towards the mean was observed over time, final Z-scores remained significantly different among subjects in the different baseline groups (Figure 2). Notably, among boys and girls who had initial Z-scores below −1.5, almost all remained below the mean of normal (Z = 0) and the majority stayed below −1.0 at final follow-up (Figure 3). This was true for both BMC and BMD, regardless of measurement site and whether Z-scores were adjusted for height (Figure 4; available at www.jpeds.com). Likewise, final values for subjects who had initial Z-scores >1.5 tended to remain above the mean (Figure 3). Only 3.8% of the children crossed from INTERMEDIATE to LOW; 12 of 256 girls and 6 of 209 boys. We were not able to identify any characteristics distinguishing the children who crossed Z-score categories.

Figure 2.

Mean ± SD whole body less head BMC Z-scores over time for subjects subdivided by baseline Z-score (<−1.5, −1.5 to 1.5, >1.5). Final Z-scores differed significantly among the different baseline groups (p < 0.001).

Figure 3.

Change between baseline and final (6-year) Z-scores for whole body less head BMC subdivided by baseline Z-score: > 1.5 (top row), −1.5 to 1.5 (middle row), < −1.5 (bottom row). Similar trends were observed for other DXA bone measures.

Figure 4.

Final height-adjusted Z-scores for subjects starting with adjusted Z < −1.5 at baseline.

DISCUSSION

This study indicates that DXA measures of bone strength track through childhood until the time of skeletal and sexual maturity. This was true in both the axial and appendicular skeleton, for both males and females. On average, DXA measures of BMC and BMD during childhood accounted for approximately 50–70% of the variation observed 6 years later when subjects were mature. Almost all boys and girls with low bone mass (Z-score < −1.5) became young men and women with values for BMC and BMD below the mean (Z-score < 0). These results corroborate previous findings indicating that the susceptibility for osteoporosis is present early in life.

Ample data suggest that the amount of bone accrued by skeletal maturity is the main contributor to peak bone mass, which, in turn, is a major determinant of osteoporosis and fractures in the elderly (17). We have also recently shown in the same cohort that bone mass during childhood is a predictor of pediatric fractures (18). Although the time of life when DXA values reach their peak has been of considerable controversy (19, 20), most estimates indicate that bone mass does not significantly increase after the third decade (5). Because the degree of tracking during young adulthood should be comparable or greater than during growth, the findings of our study underscore the potential of these measures to predict bone mineral status in older adulthood. Support for this notion is that the degree of tracking for DXA measures of bone mineral accrual in our large, nationally representative cohort of healthy children was comparable with that of height and weight.

We were not able to identify any characteristics that predicted which subjects would cross Z-score categories, such as changing from low to intermediate or intermediate to low. Changes in bone status may have been related to growth, nutrition, changes in activity, or many other factors. It is important to recognize that although there is strong overall tracking of DXA Z-scores, it is still possible for the bone status of a given individual to worsen or improve.

The large sample of well characterized healthy children representative of the current U.S. population, the relatively long follow-up period until skeletal maturity, and the highly standardized assessments of bone density measured by DXA are strengths of this study. The findings of this 6-year longitudinal study corroborate previous studies with shorter follow-ups suggesting that children retain their bone phenotypes throughout growth (9, 21). Most studies to date, however, have been constrained by small sample sizes (22–24) and/or BMD determinations using techniques not widely used in clinical practice (23, 25). Our results are also consistent with smaller studies of pre- and early pubertal children that were followed for 7 to 8 years (26–29). However, the current study is limited to the natural history of bone mineral accrual in healthy children. Studies are needed for children who have chronic medical conditions that may alter bone mineral accrual and cause fluctuations in tracking.

Bone mineral status during childhood is a strong predictor of bone mineral status in the axial and appendicular skeletons in young adulthood. These results suggest that bone mass measurements may be useful for early identification of children at risk for osteoporosis later in life. They also provide a distinct example of how phenotypic traits, like genetic information, could assume a major role in personalized medicine and the customization of healthcare. Because current treatment for osteoporosis in the elderly does not significantly restore loss of bone, efforts should be directed toward developing preventive measures that increase bone accrual before the completion of skeletal maturity. Studies are needed to establish whether DXA bone measures associated with weaker bones in childhood can be altered as a result of simple nutritional, mechanical, or pharmacological intervention.

ABBREVIATIONS

BMC

bone mineral content

BMD

bone mineral density (areal)

BMDCS

Bone Mineral Density in Childhood Study

BMI

body mass index

DXA

dual-energy x-ray absorptiometry

Appendix

Collaborators in the pediatric endocrine divisions of each Bone Mineral Density in Childhood Study Group clinical centers include: Children’s Hospital of Philadelphia: Andrea Kelly, David Langdon, Thomas Moshang, Steve Willi, Lorraine Katz, Charles Stanley, and Craig Alter; Children’s Hospital Los Angeles: Lynda Fisher, Mitchell Geffner, Debra Jeandron, Steven Mittelman, Pisit Pitukcheewanont, and Francine Kaufman; Cincinnati Children’s Hospital Medical Center: Susan Rose, Frank Biro, Peggy Stenger, Debbie Elder, and James Heubi; Columbia University Medical Center-St. Luke’s Hospital: Mary Horlick, Natasha Leibel, and Abeer Hassoun; and Creighton University: Jean-Claude Desmangles.

Data Safety and Monitoring Board members: Clifford Rosen, Ralph D’Agostino, Ingrid Holm, James Reynolds, and Reginald Tsang.