SEARCH

SEARCH BY CITATION

Keywords:

  • BONE MASS;
  • BMD;
  • GROWTH;
  • CHILDREN;
  • FRACTURES

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

A fracture in childhood is associated with low bone mineral density (BMD), but it is debated whether a fracture at growth also predicts low BMD in young adulthood. The purpose of this work was to gender-specifically evaluate whether children with a fracture are at increased risk of low BMD in young adulthood. Distal forearm BMD (g/cm2) was measured with single-photon absorptiometry (SPA) in 47 boys and 26 girls (mean age 10 years, range 3–16 years) with an index fracture and in 41 boys and 43 girls (mean age 10 years, range 4–16 years) with no fracture. BMD was re-measured mean 27 years later with the same SPA apparatus and with dual-energy absorptiometry (DXA), quantitative ultrasound (QUS), and peripheral computed tomography (pQCT). Individual Z-scores were calculated using the control cohort as reference population. Data are presented as means with 95% confidence intervals (95% CI) within brackets and correlation with Pearson's correlation coefficient. Boys with an index fracture had at fracture event a distal forearm BMD Z-score of −0.4 (95% CI, −0.7 to −0.1) and at follow-up −0.4 (95% CI, −0.7 to −0.1). Corresponding values in girls were −0.2 (95% CI, −0.5 to 0.1) and −0.3 (95% CI, −0.7 to 0.1). The deficit in absolute bone mass was driven by men with index fractures in childhood due to low energy rather than moderate or high energy. There were no changes in BMD Z-score during the follow-up period. The BMD deficit at follow-up was in boys with an index fracture verified with all advocated techniques. A childhood fracture in men was associated with low BMD and smaller bone size in young adulthood whereas the deficit in women did not reach statistical significance. © 2013 American Society for Bone and Mineral Research.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

Fractures are a general health problem because around one-half of all women and 25% of all men will sustain a fracture after the age of 50 years.1 But because close to one-half of all children will sustain a fracture before the age of 18 years,2, 3 fractures are also a huge pediatric problem, associated with large health care costs and significant individual suffering.3, 4 Therefore, it is imperative to identify risk factors for fractures in all ages, enabling identification of high-risk individuals. One such risk factor is low bone mineral density (BMD), found to be associated with increased fracture risk in both adults5, 6 and children,7–16 and 1 SD lower BMD is usually reported to be associated with doubled fracture risk.5 Research has therefore focused on factors that influence BMD, both the loss during aging17 and the accrual during growth.18 Osteoporosis has long since been attributed to predominantly high bone loss in adult life, but because 50% of BMD at age 65 years has been estimated to be predicted by peak bone mass,19, 20 the accrual of BMD during growth has gradually attracted interest. This especially accounts for the peripubertal period, because 36% of the total amount of adult BMD is acquired during the 4 peripubertal years, similar to the total amount of loss in adult life.21 Recent data have also inferred that both benefits22, 23 and deficits24, 25 in BMD acquired during growth may be retained into adulthood. So, if children with low BMD could be identified, this would open possibilities for targeted interventions.22, 23

A childhood fracture is one such risk for low BMD,7–8, 13, 26 predominantly fractures following a low-energy trauma,13 but possibly also moderate- to high-energy trauma.26 Because there is a peak in fracture incidence in childhood after a period with increased skeletal size but without an accompanying similar increase in mineralization, one hypothesis suggests that a childhood fracture is associated with a delayed maturational pattern that creates transient reduced BMD.8, 21, 27, 28 This period has, however, been shown to be followed by an extended period of mineralization that will lead to normal peak bone mass.21

Another hypothesis infers that maturational delayed children are overrepresented among children with fractures, but that an extended growth period after the fracture event would lead to normal peak bone mass.29, 30 A third hypothesis infers that BMD tracks from childhood to adulthood so that any BMD deficit in childhood would also be reflected by low peak bone mass.16 But up to now no study has prospectively been able to shown that children with a fracture actually reach low peak bone mass.13, 26 The current knowledge is based on cross-sectional studies19, 20 and short-term prospective observational studies.16, 28, 31 But because peak bone mass is reported to be reached at the end of the second or even third decade in life,30 it is debated whether peak bone mass was actually reached in the published studies.16, 28, 31 Therefore, there is a need for long-term prospective controlled data. That is the reason why this study was designed as a prospective controlled study with the aim of following BMD in children with an index fracture for close to three decades. We hypothesized that a childhood fracture would be associated with low BMD in young adulthood.

Subjects and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

Children at baseline

A skeletal evaluation was performed by single-photon absorptiometry (SPA) in 90 children with an index fracture between 1979 and 1981,3 57 boys and 33 girls with a mean age of 10 (range, 4–16) years. The scans were performed 40 ± 25 days (mean ± SD) after they had sustained the fracture. All types of fractures except hand, finger, skull, tooth, and rib fractures were included. Fifty-five children were reported with a fracture due to low-energy trauma, 31 due to moderate energy trauma, and 4 due to high-energy trauma.3 A control cohort that included 131 volunteers within the same ages with no index fracture, 65 boys and 66 girls with a mean age of 10 (range, 4–16) years, was measured during the same period. All participants were white without diseases or medications known to affect the bone metabolism, none were malnourished or had impairment in growth. In the original report, children with an index fracture were reported with significantly lower BMD than the children with no index fracture.13

Follow-up evaluation

The follow-up evaluation was performed a mean 27 (range, 25–29) years after the fracture event. At follow-up, 2 men and 1 woman had died, 13 men and 7 women had relocated, 12 men and 14 women could not be found, 7 men and 6 women declined further measurements due to unwillingness to participate, 1 man was unable to attend due to illness, and 1 additional fracture case was excluded because we were not able to verify the type of trauma severity. Thus 157 of the original 221 were finally included in this report when they were at a mean age of 37 (range, 30–44) years. This corresponded to a 71% participation rate in total, 47 of 57 (82%) boys and 28 of 33 (85%) girls in the fracture cohort, and 41 of 65 (63%) boys and 43 of 66 (65%) girls in the control cohort. In the re-measured fracture cohort, 28 boys and 19 girls had experienced the index fracture due to a low-energy trauma, 16 boys and 7 girls due to a moderate-energy trauma, and 3 boys and 0 girls due to a high-energy trauma. The dropout analysis revealed that there were no statistically significant group differences regarding, age height, weight, or body mass index (BMI) registered between participants and nonparticipants.

Bone mass measurements

Bone mineral content (BMC; g/cm) and BMD (g/cm2) were measured both at baseline and at follow-up on the forearm 6 cm proximal to the ulnar styloid process by the same SPA apparatus; the scanning technique is described in detail in previous reports.13, 17 Both arms were scanned, after which the mean value was used. In individuals with a history of forearm fracture, the nonfractured arm was used. Twenty-eight children had a fractured upper extremity, 11 on the right side and 17 on the left side. The coefficient of variation (CV) was 2% with a standardized phantom and 4% determined by double measurements after the subject was repositioned. The long-term drift was 0.1%/year (95% confidence interval [CI], −0.2 to 0.4), evaluated by a standardized phantom every second week during the entire study.17 One technician performed all baseline measurements and one performed all follow-up measurements, and one of the authors analyzed all the plots.

At follow-up, BMC and BMD were also measured by dual X-ray absorptiometry (DXA) (Lunar DPX-L scanner, software version 1.3z; Lunar, Madison, WI, USA) in total body by a total body scan, in the first to fourth lumbar vertebra (L1–L4) by a lumbar spine scan and in the femoral neck and total hip by a hip scan. Daily calibration of the apparatus was done with the Lunar phantom. The CV evaluated in 14 individuals after repositioning was 0.4% to 3.0% for BMD depending on the measured region. Qualitative ultrasound (QUS) evaluated broadband ultrasonic attenuation (BUA; db/MHz) and speed of sound (SOS; m/s) in both calcanei, after which the mean value was used. The CV evaluated in 14 individuals after repositioning was 2.2% for BUA and 0.3% for SOS. Peripheral quantitative computed tomography (pQCT) (XCT 2000; Stratec, Pforzheim, Germany) measured BMD, cross-sectional area (CSA; mm2) and stress-strain index (SSI, mm3) in the left radius and left tibia. We measured at the 4% and 38% level from the ankle joint and at 6% and 66% level from the wrist. Daily calibration of the apparatus was done with a standard phantom. The CV evaluated in 14 individuals after repositioning was 1.1% to 4.6% for CSA depending on the measured region. Three research technicians performed all the DXA, QUS, and pQCT measurements and analyzed all the scans.

Anthropometric measurements and registration of lifestyle factors and incident fractures

We measured body weight to the nearest 0.1 kg with an electric scale and body height to the nearest 0.5 cm by a wall-tapered height meter. Questionnaires registered lifestyle factors, diseases, and medications both at baseline13 and at follow-up.32

Statistical evaluation

Statistical calculations were performed with PASW Statistic software SPSS (version 18.0, SPSS, Inc., Cicago, IL, USA). Data are presented as numbers (n), means with 95% CI, and as proportions (%). Group differences were evaluated by chi-square test and ANCOVA with adjustment for age. Individual Z-scores, the number of SDs above or below the age-predicted mean, were derived by linear regression using the control cohort as a reference population.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

BMC and BMD in individuals with an index fracture, independent of trauma type

There were no differences in age or anthropometrics between boys and girls with or without an index fracture, neither at baseline nor at follow-up (Table 1). Boys with an index fracture had at fracture event a distal forearm BMC Z-score of −0.4 (95% CI, −0.6 to −0.1) and BMD Z-score of −0.4 (95% CI, −0.7 to −0.1) and at follow-up of −0.5 (95% CI, −0.8 to −0.2) and −0.4 (95% CI, −0.7 to −0.1), respectively. Thus there were no changes in the BMC or BMD deficit from growth into adulthood (Table 2). The BMD deficit in adult men with a former index fracture was statistically captured by all scanning techniques with the largest Z-score deficit registered by DXA (total hip Z-score −1.0; 95% CI, −1.3 to −0.7) (Tables 3 and 4). Men with an index fracture also had smaller CSA with the largest deficit in tibia (Z-score −0.5; 95% CI, −0.7 to −0.3), resulting in a lower SSI (Z-score −0.5; 95% CI, −0.7 to −0.3) (Tables 3 and 4). Girls with an index fracture had at fracture event a distal forearm BMC Z-score of −0.3 (95% CI, −0.7 to 0.1) and BMD Z-score of −0.2 (95% CI, −0.5 to 0.1) and at follow-up of −0.3 (95% CI, −0.6 to 0.1) and −0.3 (95% CI, −0.7 to 0.1), respectively. Thus there were no changes in the BMC and BMD deficit from childhood into adulthood. The deficits in adulthood when measured by the other scanning techniques did not reach statistical significance (Table 3), even though pQCT measured a cortical radius BMD Z-score of −0.4 (95% CI, −0.9 to −0.0) (Table 4).

Table 1. Age, Height, Weight, BMI, BMC, and BMD Were Measured in 47 Boys and 26 Girls When They Sustained a Fracture in Childhood and at a Mean 27 Years Later
 Baseline Follow-up 
 CasesControlspCasesControlsp
  1. Comparisons were made with 41 boys and 43 girls with no index fracture at baseline. Bone mass measurements were done by SPA in distal forearm. Data are shown as unadjusted means ± SD. Comparisons of the two groups are adjusted for age, and statistically significant differences (p < 0.05) are in bold. No group comparison was made in girls with moderate/high-energy trauma versus controls due to the small sample size.

  2. BMI = body mass index; BMC = bone mineral content; BMD = bone mineral density; SPA = single-photon absorptiometry.

Fractures due to all types of trauma      
 Menn = 47n = 41 n = 47n = 41 
  Age (years)9.7 ± 4.09.8 ± 3.50.9336.4 ± 4.137.4 ± 3.60.25
  Height (cm)140,5 ± 25.6140.8 ± 22.80.95180.1 ± 7.0181.4 ± 8.20.42
  Weight (kg)35.2 ± 18.036.7 ± 19.00.7185.7 ± 12.988.7 ± 15.00.31
  BMI (kg/m2)17.1 ± 3.217,3 ± 3.20.8226.4 ± 3.226.9 ± 3.80.50
  BMC (g/cm)0.44 ± 0.190.48 ± 0.180.051.00 ± 0.141.07 ± 0.130.04
  BMD (g/cm2)0.41 ± 0.110.43 ± 0.090.090.65 ± 0.070.68 ± 0.070.08
 Womenn = 26n = 43 n = 26n = 43 
  Age (years)9.8 ± 3.410.6 ± 3.70.3736.3 ± 3.638.2 ± 3.70.04
  Height (cm)139.5 ± 17.9142.9 ± 20.40.49167.9 ± 5.5166.7 ± 6.60.46
  Weight (kg)33.6 ± 11.038.5 ± 14.60.1474.4 ± 15.972.6 ± 16.20.65
  BMI (kg/m2)16.7 ± 2.118.0 ± 2.40.0326.4 ± 5.326.1 ± 5.50.82
  BMC (g/cm)0.40 ± 0.130.45 ± 0.160.170.70 ± 0.090.73 ± 0.090.30
  BMD (g/cm2)0.39 ± 0.090.42 ± 0.090.380.53 ± 0.060.55 ± 0.060.34
Fractures due to low-energy trauma      
 Menn = 28n = 41 n = 28n = 41 
  Age (years)9.8 ± 4.39.8 ± 3.50.9936.6 ± 4.537.4 ± 3.60.41
  Height (cm)141.2 ± 29.1140.8 ± 22.80.95179.5 ± 6.6181.4 ± 8.20.30
  Weight (kg)35.2 ± 19.636.7 ± 19.00.7183.8 ± 13.288.7 ± 15.00.17
  BMI (kg/m2)17.1 ± 2.917.3 ± 3.20.8326.0 ± 3.326.9 ± 3.80.31
  BMC (g/cm)0.45 ± 0.210.48 ± 0.180.121.00 ± 0.131.07 ± 0.130.02
  BMD (g/cm2)0.40 ± 0.110.43 ± 0.090.050.65 ± 0.070.68 ± 0.070.04
 Womenn = 19n = 43 n = 19n = 43 
  Age (years)10.1 ± 3.710.6 ± 3.70.3736.6 ± 3.838.2 ± 3.70.12
  Height (cm)141 ± 18.2142.9 ± 20.40.49168.0 ± 4.9166.7 ± 6.60.48
  Weight (kg)34.1 ± 11.538.5 ± 14.60.1474.4 ± 16.172.6 ± 16.20.68
  BMI (kg/m2)16.6 ± 2.318.0 ± 2.40.0326.3 ± 5.326.1 ± 5.50.86
  BMC (g/cm)0.41 ± 0.140.45 ± 0.160.190.69 ± 0.090.73 ± 0.090.13
  BMD (g/cm2)0.40 ± 0.090.42 ± 0.090.360.53 ± 0.050.55 ± 0.060.47
Fractures due to moderate/high-energy trauma      
 Menn = 19n = 41 n = 19n = 41 
  Age (years)9.4 ± 3.69.8 ± 3.50.6936.0 ± 3.537.4 ± 3.60.16
  Height (cm)137.9 ± 20.1140.8 ± 22.80.65181.0 ± 7.9181.4 ± 8.20.85
  Weight (kg)34.6 ± 16.136.7 ± 19.00.6788.6 ± 12.588.7 ± 15.00.99
  BMI (kg/m2)17.2 ± 3.717.3 ± 3.20.8827.0 ± 3.226.9 ± 3.80.89
  BMC (g/cm)0.43 ± 0.150.48 ± 0.180.181.03 ± 0.151.07 ± 0.130.35
  BMD (g/cm2)0.41 ± 0.10.43 ± 0.090.060.66 ± 0.080.68 ± 0.070.56
 Womenn = 7n = 43 n = 7n = 43 
  Age (years)9.0 ± 2.710.6 ± 3.735,6 ± 2.938.2 ± 3.7
  Height (cm)135.5 ± 17.8142.9 ± 20.4167.7 ± 7.1166.7 ± 6.6
  Weight (kg)32.2 ± 10.038.5 ± 14.674.4 ± 16.872.6 ± 16.2
  BMI (kg/m2)17.1 ± 1.818.0 ± 2.426.4 ± 5.826.1 ± 5.5
  BMC (g/cm)0.38 ± 0.120.45 ± 0.160.72 ± 0.10.73 ± 0.09
  BMD (g/cm2)0.38 ± 0.080.42 ± 0.090.55 ± 0.080.55 ± 0.06
Table 2. BMC and BMD Measured by SPA in 47 Boys and 26 Girls When They Sustained a Fracture in Childhood and a Mean 27 Years Later
 Individuals with fractures due to all types of traumaIndividuals with fractures due to low-energy traumaIndividuals with fractures due to moderate/high-energy trauma
 MenWomenMenWomenMenWomen
 Cases (n = 47)Cases (n = 26)Cases (n = 28)Cases (n = 19)Cases (n = 19)Cases (n = 7)
  1. Data are shown as mean Z-scores with 95% confidence interval within brackets. The sample mean in the controls (zero) was calculated based on data in 41 boys and 43 girls who were followed during the same period but had no index fracture at baseline. Individuals' Z-scores, the number of SDs above or below the age predicted mean, were derived by linear regression using our control cohort (with a mean Z-score of zero) as a reference population.

  2. BMC = bone mineral content; BMD = bone mineral density; SPA = single-photon absorptiometry.

Baseline      
 BMC Z-score−0.38 (−0.62 to −0.14)−0.31 (−0.66 to 0.05)−0.36 (−0.67 to −0.04)−0.34 (−0.78 to 0.10)−0.34 (−0.73 to 0.04)−0.22 (−1.0 to 0.56)
 BMD Z-score−0.37 (−0.67 to −0.07)−0.21 (−0.53 to 0.10)−0.50 (−0.90 to −0.10)−0.25 (−0.60 to 0.11)−0.17 (−0.68 to 0.34)−0.13 (−1.03 to 0.78)
Follow-up      
 BMC Z-score−0.45 (−0.76,−0.15)−0.31 (−0.73 to 0.12)−0.54 (−0.92 to −0.16)−0.39 (−0.88 to 0.10)−0.26 (−0.81 to 0.30)−0.08 (−1.17 to 1.00)
 BMD Z-score−0.40 (−0.73,−0.07)−0.31 (−0.72 to 0.10)−0.49 (−0.89 to −0.09)−0.42 (−0.83 to −0.01)−0.19 (−0.80 to 0.42)−0.03 (−1.13 to 1.26)
Delta Z-score      
 BMC Z-score0.07 (−0.22 to 0.37)−0.00 (−0.34 to 0.34)0.18 (−0.18 to 0.54)0.05 (−0.33 to 0.43)−0.08 (−0.64 to 0.47)−0.14 (1.08 to 0.80)
 BMD Z-score0.03 (−0.36 to 0.41)0.10 (−0.30 to 0.49)−0.01 (−0.44 to 0.43)0.17 (−0.28 to 0.63)0.02 (−0.07 to 0.74)−0.10 (−1.13 to 0.92)
Table 3. BMD Measured by DXA, SOS and BUA Measured by QUS, and BMD Measured by pQCT in 47 Men and 26 Women a Mean 27 Years After They Sustained an Index Fracture in Childhood
ModalityFractures due to all types of traumaFractures due to low-energy traumaFractures due to moderate/high-energy trauma
MenWomenMenWomenMenWomen
Cases (n = 47)Controls (n = 41)Cases (n = 26)Controls (n = 43)Cases (n = 28)Controls (n = 41)Cases (n = 19)Controls (n = 43)Cases (n = 19)Controls (n = 41)Cases (n = 7)Controls (n = 43)
  1. Comparisons are made with 41 men and 43 women with no index fracture at baseline. Adjustments for differences in age are advocated in the group comparison and statistically significant differences (p < 0.05) are in bold. Due to small sample size no group comparison was made between girls with moderate/high-energy trauma and controls. Data are shown as unadjusted means with 95% CI in parentheses.

  2. BMD = bone mineral density; DXA = dual-energy X-ray absorptiometry; SOS = speed of sound; BUA = broadband attenuation; QUS = quantitative ultrasound; pQCT = peripheral computed tomography; FN = femoral neck; L1–L4 = lumbar spine vertebrae 1 to 4; SSI = stress strain index; CI = confidence interval.

DXA            
 Total body BMD (g/cm2)1.24 (1.21–1.26)1.30 (1.27–1.33)1.18 (1.15–1.21)1.20 (1.18–1.22)1.23 (1.20–1.26)1.30 (1.27–1.33)1.17 (1.13–1.20)1.20 (1.18–1.22)1.24 (1.20–1.28)1.30 (1.27–1.33)1.22 (1.14–1.30)1.20 (1.18–1.22)
 Total hip BMD (g/cm2)1.04 (1.01–1.08)1.16 (1.12–1.19)1.01 (0.96–1.06)1.04 (1.01–1.08)1.03 (0.98–1.07)1.16 (1.12–1.19)1.00 (0.95–1.05)1.04 (1.01–1.08)1.07 (1.01–1.13)1.16 (1.12–1.19)1.05 (0.91–1.19)1.04 (1.01–1.08)
 FN BMD (g/cm2)1.03 (0.99–1.06)1.13 (1.10–1.17)1.00 (0.95–1.06)1.03 (0.99–1.06)1.01 (0.97–1.05)1.13 (1.10–1.17)0.99 (0.93–1.05)1.03 (0.99–1.06)1.06 (0.99–1.12)1.13 (1.10–1.17)1.04 (0.89–1.19)1.03 (0.99–1.06)
 L1–L4 BMD (g/cm2)1.20 (1.16–1.24)1.30 (1.26–1.34)1.22 (1.17–1.27)1.24 (1.20–1.28)1.19 (1.14–1.23)1.30 (1.26–1.34)1.19 (1.13–1.25)1.24 (1.20–1.28)1.22 (1.16–1.29)1.30 (1.26–1.34)1.31 (1.21–1.41)1.24 (1.20–1.28)
QUS            
 SOS (m/s)1568 (1553–1583)1592 (1579–1604)1586 (1575,1599)1591 (1580–1601)1561 (1535–1586)1592 (1579,1604)1576 (1565–1587)1591 (1580,1601)1578 (1562–1593)1592 (1579,1604)1609 (1583–1633)1591 (1580,1601)
 BUA (dB/MHz)117 (113–121)124 (119–129)123 (113–133)120 (115–125)115 (109–120)124 (119–129)115 (109–121)120 (115–125)120 (113–128)124 (119–129)146 (114–178)120 (115–125)
pQCT            
 Trabecular tibia (4%)            
  BMD (g/cm2)1.45 (1.40–1.50)1.56 (1.49–1.62)1.23 (1.15–1.31)1.24 (1.20–1.27)1.45 (1.37–1.53)1.56 (1.49–1.62)1.20 (1.11 1.28)1.24 (1.20–1.27)1.47 (1.40–1.53)1.56 (1.49–1.62)1.32 (1.13–1.52)1.24 (1.20–1.27)
 Cortical tibia (38%)            
  BMD (g/cm2)1.97 (1.94–2.00)2.04 (2.00–2.09)1.81 (1.77–1.85)1.83 (1.80–1.87)1.97 (1.92–2.01)2.04 (2.00–2.09)1.80 (1.75–1.84)1.83 (1.80–1.87)1.99 (1.94–2.03)2.04 (2.00–2.09)1.84 (1.75–1.93)1.83 (1.80–1.87)
  Cross sectional area (mm2)464 (446–482)499 (476–522)385 (367–403)387 (372–402)460 (441–478)499 (476–522)382 (363–400)387 (372–402)470 (445–495)499 (476–522)385 (338–432)387 (372–402)
  SSI (mm3)1942 (1853–2031)2166 (2017–2315)1464 (1368–1560)1504 (1423–1585)1909 (1790–2028)2166 (2017–2315)1440 (1331–1549)1504 (1423–1585)1993 (1837–2149)2166 (2017–2315)1523 (1273–1774)1504 (1423–1585)
 Trabecular radius (6%)            
  BMD (g/cm2)0.95 (0.92–0.98)1.00 (0.97–1.05)0.75 (0.70–0.79)0.77 (0.75–0.80)0.94 (0.89–0.99)1.00 (0.97–1.05)0.74 (0.71–0.78)0.77 (0.75–0.80)0.97 (0.91–1.03)1.00 (0.97–1.05)0.78 (0.69–0.88)0.77 (0.75 0.80)
 Cortical radius (66%)            
  BMD (g/cm2)1.12 (1.10–1.15)1.16 (1.13–1.19)0.99 (0.96–1.02)1.02 (0.99–1.04)1.11 (1.08–1.14)1.16 (1.13–1.19)0.98 (0.95–1.02)1.02 (0.99–1.04)1.15 (1.11–1.19)1.16 (1.13–1.19)1.00 (0.96–1.06)1.02 (0.99–1.04)
  Cross sectional area (mm2)180 (170–189)188 (177–199)130 (122–139)138 (131–145)174 (161–186)188 (177–199)131 (120–143)138 (131–145)190 (174–205)188 (177–199)127 (111–144)138 (131–145)
  SSI (mm3)413 (382–443)440 (407–473)273 (250–296)293 (274–311)406 (363–449)440 (407–473)275 (246–304)293 (274–311)429 (380–477)440 (407–473)268 (229–307)293 (274–311)
Table 4. Z-Scores for Bone Traits Measured by DXA, QUS, and pQCT in 47 Men and 26 Women a Mean 27 Years After They Sustained an Index Fracture in Childhood
 Individuals with fractures due to all types of traumaIndividuals with fractures due to low-energy traumaIndividuals with fractures due to moderate- or high-energy trauma
 Men (n = 47)Women (n =26)Men (n = 28)Women (n = 19)Men (n = 19)Women (n = 7)
  1. Data are shown as means with 95% confidence interval in parentheses. The sample mean in the controls (zero) are calculated based on data in 41 boys and 43 girls with no index fracture at baseline. Individuals' Z-scores, the number of SDs above or below the age predicted mean, were derived by linear regression using our control cohort (with a mean Z-score of zero) as a reference population.

  2. DXA = dual-energy X ray absorptiometry; QUS = quantitative ultrasound; pQCT = peripheral computed tomography; BMD = bone mineral density; L1–L4 = lumbar spine vertebrae 1 to 4; SOS = speed of sound; BUA = broadband attenuation; SSI = stress strain index.

DXA      
 Total body BMD Z-score−0.70 (−0.96 to −0.45)−0.31 (−0.78 to 0.16)−0.72 (−1.05 to −0.38)−0.53 (−1.02 to −0.03)−0.72 (−1.25 to −0.20)0.28 (−0.97 to 1.54)
 Total hip BMD Z-score−1.01 (−1.33 to −0.70)−0.30 (−0.76 to 0.17)−1.16 (−1.57 to −0.75)−0.42 (−0.93 to 0.09)−0.62 (−1.03 to −0.20)0.04 (−1.24 to 1.32)
 Femoral neck BMD Z-score−0.92 (−1.23 to −0.62)−0.19 (−0.70 to 0.32)−1.04 (−1.41 to −0.68)−0.32 (−0.89 to 0.25)−0.67 (−1.23,−0.11)0.15 (−1.25,1.56)
 L1−L4 BMD Z-score−0.77 (−1.05 to −0.48)−0.16 (−0.58 to 0.26)−0.85 (−1.20 to −0.50)−0.43 (−0.90 to 0.05)−0.56 (−1.08 to −0.04)0.56 (−0.23 to 1.35)
QUS      
 SOS Z-score−0.64 (−1.04 to −0.24)−0.11 (−0.47 to 0.25)−0.83 (−1.49 to −0.17)−0.42 (−0.73 to −0.11)−0.37 (−0.78 to 0.04)0.73 (−0.14 to 1.60)
 BUA Z-score−0.48 (−0.76 to −0.19)0.21 (−0.44 to 0.87)−0.64 (−1.02 to −0.26)−0.33 (−0.74 to 0.09)−0.26 (−0.74 to 0.22)1.69 (−0.43 to 3.80)
pQCT      
 Trabecular tibia (4%)      
  BMD Z-score−0.49 (−0.74 to −0.24)−0.03 (−0.74 to 0.69)−0.49 (−0.88 to −0.11)−0.35 (−1.13 to 0.42)−0.42 (−0.72 to 0.12)0.82 (−1.01 to 2.66)
 Cortical tibia (38%)      
  BMD Z-score−0.50 (−0.73 to −0.27)−0.20 (−0.54 to 0.14)−0.54 (−0.87 to −0.21)−0.31 (−0.71 to 0.09)−0.42 (−0.77 to 0.07)0.09 (−0.72 to 0.89)
  Cross-sectional area Z-score−0.50 (−0.70 to −0.30)−0.09 (−0.43 to 0.26)−0.56 (−0.82 to −0.30)−0.11 (−0.49 to 0.28)−0.42 (−0.77 to −0.06)−0.04 (−1.02 to 0.95)
  SSI Z-score−0.49 (−0.68 to −0.29)−0.15 (−0.53 to 0.21)−0.56 (−0.82 to −0.30)−0.25 (−0.68 to 0.18)−0.38 (−0.71 to −0.04)0.07 (−0.90 to 1.06)
 Trabecular radius (6%)      
  BMD Z-score−0.50 (−0.78 to −0.23)−0.27 (−0.80 to 0.27)−0.58 (−0.95 to −0.21)−0.42 (−1.02 to 0.18)−0.34 (−0.80 to 0.13)0.15 (−1.26 to 1.57)
 Cortical radius (66%)      
  BMD Z-score−0.42 (−0.68 to −0.16)−0.44 (−0.86,−0.01)−0.58 (−0.92 to −0.23)−0.52 (−1.03 to −0.01)−0.14 (−0.55 to 0.26)−0.21 (−1.14 to 0.73)
  Cross-sectional area Z-score−0.25 (−0.56 to 0.05)−0.35 (−0.76 to 0.05)−0.42 (−0.82 to −0.03)−0.31 (−0.84 to 0.22)0.07 (−0.44 to 0.57)−0.47 (−1.20 to 0.26)
  SSI Z-score−0.28 (−0.60 to 0.04)−0.34 (−0.73 to 0.05)−0.36 (−0.79 to 0.08)−0.31 (−0.83 to 0.21)−0.10 (−0.60 to 0.41)−0.41 (−1.06 to 0.24)

All group differences in Tables 1 and 3 remained after adjusting for height, weight, and age (data not shown).

BMC and BMD in children with an index fractures due to low-energy trauma

Boys with an index fracture due to a low-energy–related trauma had a distal forearm BMC Z-score of −0.4 (95% CI, −0.7 to −0.0) and BMD Z-score of −0.5 (95% CI, −0.9 to −0.1) and at follow-up of −0.5 (95% CI, −0.9 to −0.2) and −0.5 (95% CI, −0.9 to −0.1), respectively. Thus, there were no changes in the BMC or BMD deficit from growth into adulthood (Table 2). The BMD deficit in adult men with a former low-energy–related index fracture was statistically captured by all scanning techniques, with the largest Z-score deficit registered by DXA (total hip Z-score −1.2; 95% CI, −1.6 to −0.8) (Tables 3 and 4). Men with an index fracture also had smaller CSA with the largest deficit in tibia (Z-score of −0.6; 95% CI, −0.8 to −0.3), resulting in a lower SSI (Z-score −0.6; 95% CI, −0.8 to −0.3) (Tables 3 and 4). Girls with an index fracture due to a low-energy–related trauma had a distal forearm BMC Z-score of −0.3 (95% CI, −0.8 to 0.1) and BMD Z-score of −0.3 (95% CI, −0.6 to 0.1) and at follow-up of −0.4 (95% CI, −0.9 to 0.1) and −0.4 (95% CI, −0.8 to −0.0), respectively. Thus, there were no changes in the BMC or BMD deficit from growth into adulthood (Table 2). The deficits in adulthood when measured by the other scanning techniques did not reach statistical significance (Table 3), even though DXA measured a total body BMD Z-score of −0.5 (95% CI, −1.0 to 0.0) (Table 4). All group differences in Tables 1 and 3 remained after adjusting for height, weight, and age (data not shown).

BMC and BMD in children with an index fracture due to moderate- or high-energy trauma

Boys with a moderate- or high-energy–related index fracture had a distal forearm BMC Z-score of −0.3 (95% CI, −0.7 to 0.0) and BMD Z-score of −0.2 (95% CI, −0.7 to 0.3) and at follow-up of −0.3 (95% CI, −0.8 to 0.3) and −0.2 (95% CI, −0.8 to 0.4). Thus, there were no changes in the BMC or BMD deficit from growth into adulthood (Table 2). The BMD deficit in adult men with a former moderate- or high-energy–related index fracture was statistically captured by DXA with the largest Z-score deficit found with a total body Z-score of −0.7 (95% CI, −1.3 to −0.2) (Tables 3 and 4). The deficits in adulthood when measured by the other scanning techniques did not reach statistical significance (Table 3), even though pQCT measured a tibial CSA Z-score of −0.4 (95% CI, −0.8 to −0.1)), resulting in a SSI Z-score of −0.4 (95%CI, −0.8 to −0.0) (Table 4).

There were only 7 girls with an index fracture due to a moderate- or high-energy–related trauma; therefore, no further statistical evaluation was done in this group.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

A childhood fracture in men was associated with low BMD and smaller bone size whereas the deficit in women did not reach statistical significance. It is widely accepted that low BMD in adults is associated with an increased fracture risk.10 Recently, a systematic review and meta-analysis that included all relevant articles published in 1965–2005 found that this also accounts for children,7 concluding that children with a fracture have a mean BMD deficit of −0.3 SD compared to children with no fracture.7 This is in close accordance with the −0.4 SD deficit in the boys and the −0.2 SD deficit in the girls at fracture event in the current study, a finding which indicates that our cohort includes a representative study population so that our inferences can be generalized. However, the studies included in the meta-analysis were predominantly retrospective case-control studies7 and the authors of the review summarized their publication by concluding that there is a need for well-conducted prospective studies that evaluate whether the deficit in BMD at fracture event is transient or retained into adulthood. Our study has this design, with measurement performed with the same scanner and in the same skeletal region both at fracture event and at follow-up close to three decades later, indicating that the participants actually had reached peak bone mass, something that has been discussed in previous prospective reports with shorter follow-up periods.16, 28, 31

To our knowledge there are to date only five case-control studies that have prospectively followed children with fractures into older ages, four published articles2, 16, 28, 31, 33 and one abstract.33 These reports infer that the BMD deficit that was found at fracture event was retained years after the injury. However, the end-point was in these studies set at ages when it could be questioned whether peak bone mass really had been reached, as other studies have suggested that peak bone mass may be reached as late as the third or even fourth decade of life.17, 30 In contrast, our study chose an end-point at older ages, to increase the probability that the participants had reached peak bone mass. With this study design and with the possibility to present longitudinal data with the same scanner, we could actually state that there were no significant changes in the BMD deficits in the children with an index fracture during the follow-up period. This implies that a childhood fracture ought to be regarded as a factor that is associated with low BMD in young adulthood and that low BMD at childhood ought to be regarded as a risk factor for also reaching low peak bone mass.

According to published prospective observational studies, a deficit of −0.4 SD would be associated with a 40% higher fracture risk than expected by age.34 However, because bone size also influence who will sustain further fractures, we speculate that the fracture risk in adult boys with an index fracture possibly could be even larger than 40% given the smaller than expected bone size in young adulthood.

Opposed to our findings, The European Prospective Osteoporosis Study (EPOS) group presented results from a large (over 12,000 subjects) and well-conducted study in 2009, in which they could not find a statistical difference in BMD in adults (>50 years old) with or without self-reported fractures in childhood (8–18 years old).35 A possible explanation could be that information on fractures was self-reported, thus introducing recall bias. Another difference was that the cited study included children with an index fracture up to an age of 18 years, thus actually fully grown individuals. There is a need for more longitudinal studies with larger cohorts included and with fractures prospectively registered.

The association between low BMD and fractures in the elderly has generally been related to low-energy trauma, i.e., fragility fractures,10 as has the association generally described in children.13 A recent large prospective study from the UK has, however, challenged this view, reporting that there is also an association between low BMD and childhood fractures due to high-energy trauma.26 In our study we found that the BMD deficit in boys with a low-energy–related index fracture, at fracture event as well as at follow-up, was statistically significant for boys but not for girls.

In addition, the deficit in boys with a high-energy–related index fracture did not reach statistical significance. Comparison between girls with a high-energy fracture and controls was not tested due to small sample size. All subgroup analysis is, however, difficult to interpret due to the power problem; there is therefore a need for more longitudinal studies including larger cohorts.

Study strengths include the prospective controlled study design and gender-specific measurements performed in close conjunction with the fracture event, which avoids a major influence of posttraumatic osteopenia, as it is well known that posttraumatic osteopenia may have influences on bone mass both locally and generally.26, 31, 36 The long duration of the follow-up period makes it probable that peak bone mass was reached, even though we have not performed serial measurements in adulthood and therefore could not pinpoint the exact peak value for each person. Our cohort is, however, at follow-up, in the same age range as the reference population usually used when calculating T-scores in the definition of osteoporosis. The T-score is usually regarded as an estimation of peak bone mass. The use of the same apparatus with measurements done at the same region, and with available phantom data during the entire follow-up period, made it possible to exclude bias introduced by long-term drift. The use of modern scanners using different techniques, which all verified the remaining adult bone mass deficit also strengthens the view that there actually is a remaining deficit at young adulthood. An attendance rate of 71%, 27 years after fracture event must also be regarded as advantageous compared with earlier cited studies.21, 28, 31, 36 A low participation rate will increase the risk of making a type II error when evaluating the outcome and also increase the risk of achieving a nonrepresentative cohort as a result of bias in those who denied further participation Finally, the similarity between the fracture and the control cohort and participants and dropouts with respect to anthropometry and lifestyle reduce the risk of selection bias and increase the possibility of generalizing our inferences.

Study limitations include the sample size, which creates the risk of committing type II errors especially in girls and in individuals with high-energy–related fractures. The different participation rate among the index group and the control cohort is also a weakness. A low participation rate will increase the risk of making a type II error when evaluating the outcome and also increase the risk of achieving a nonrepresentative cohort due to bias in those who denied further participation. It would also have been advantageous to have prospective data with the modern scanning techniques. However, these techniques were not available at study start. It would also have been advantageous to have a registration of Tanner stage at baseline, to be able to correlate prepubertal BMD and bone mass in young adulthood based on maturational stage and not chronological age. It would have been advantageous to register whether any women at follow-up had reached menopause, and then had possibly also experienced a period of postmenopausal bone loss. If so, peak bone mass would not have been captured. But, as none of the women were above age 45 years, this ought to be a minor confounding factor. Finally, it would have been advantageous to have performed serial measurements of bone mass in adulthood to be able to predict actual peak bone mass.

In summary, a childhood fracture in men was associated with low BMD and smaller bone size in young adulthood whereas the deficit in women did not reach statistical significance.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References

This work was supported by grants from the Swedish Society of Medicine, Skåne University Hospital, the Österlund Foundation, the Pahlsson Foundation, and the Kocks Foundation.

Authors' roles: Study design: CB, BR, LL, MT, and MK. Data collection, analysis, and interpretation: CB, BR, JÅ, LL, MT, and MK. Drafting manuscript: CB, BR, JÅ, LL, MT, and MK. Approved the final version of the submitted manuscript: CB, BR, JÅ, LL, MT, and MK.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References