0021-972x/96/$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright 0 1996 by The Endocrine Society

Vol. 81, No. 4 Printed in U.S.A.

Volumetric Bone Mineral Aged 5-27 Years*

Density

PEI WEN LU, CHRISTOPHER T. COWELL, SUSAN JULIE N. BRIODY, AND ROBERT HOWMAN-GILES Robert Vines Growth Research Center and the Department Alexandra Hospital for Children. Sydney, Australia

D

Medicine

(R.H.-G.),

Royal

did not change with age or height in either sex. In females, a significant inverse relationship was seen between vFBMD and weight (r2 = 0.14; P = 0.001). Male subjects had higher vFBMD than females (mean ? SD, 0.73 IT 0.11 vs. 0.70 + 0.12; P = 0.0471, but no sex difference was seen in vBMD of the femoral neck. Conversely, vBMD of Ll-L4 remained age and growth dependent, although the strength of the relationship was weaker than that for aBMD (data not shown). In conclusion, the vBMD of the femoral neck and shaft is independent of age and is less dependent on growth variables in children and young adults than is aBMD. These observations offer a different perspective from our previous concepts of aBMD. (J Clin Endocrinol Metab 81: 1586-1590, 1996)

UAL ENERGY x-ray absorptiometry (DXA), with its high accuracy and precision and low radiation, is a useful technique in assessingbone mineral density (BMD). It has been employed extensively in the field of bone mineral research, and in adults, area1bone mineral density (aBMD) is directly related to the risk of osteoporosis (1). In children and adolescents, aBMD values are closely age and growth related (2-16). Actual or true BMD is a function of the bone mineral content (BMC) per volume of bone, that is volumetric BMD (vBMD). DXA measuresBMC in the volume of bone, but due to technological limitations, it calculates a two-dimensional projected area; hence, the output is expressed as an area1 density, i.e. BMC per area. In adults, aBMD is believed to be a good substitute for vBMD, as there is little change in bone size, and the major change occurring is related to the decline in BMC (17). However, in children and adolescents,growth will inevitably lead to a much greater change in bone volume than in bone area (18). This may lead to an inappropriate substitution of aBMD for vBMD and has raised questions about the validity of the use of aBMD in pediatrics (19,201. vBMD can be assessedby quantitative computer tomography imaging (QCT) and hasbeen shown to be independent of age in spinal trabecular bone in children between 2-12 yr of age (21), in the distal radius in children between 4-11 yr Received June 27, 1995. Revision received November 20, 1995. Accepted November 27,1995. Address all correspondence and requests for reprints to: Dr. Christopher T. Cowell, Robert Vines Growth Research Center, Royal Alexandra Hospital for Children, Cm Hawkesbury Road and Hainsworth Street, Westmead, New South Wales 2145, Sydney, Australia. * The Robert Vines Growth Research Center was supported by Pharmacia during the period of the study.

Subjects,

A. LLOYD-JONES,

of Nuclear

ABSTRACT Concerns have been raised regarding the validity of using area1 bone mineral density (aBMD) as a substitute for the true volumetric bone mineral density (vBMD) in the pediatric population. We studied 209 normal subjects (109 males), aged 5-27 yr, to examine the influence of age, -gender and growth on vBMD. The femoral neck, midthird of the femoral shaft. and the four lumbar vertebral bodies (Ll-L4) were studied. Using data on bone width and height obtained by dual energy x-ray absorptiometry, bone volume was calculated with the assumption that all three sites are cylinders. In contrast to aBMD, vBMD of the femoral neck bore no relationship to age or weight in both sexes, but was significantly related to height in females (r2 = 0.07; P = 0.01). Similarly, vBMD of the femoral shaft (vFBMD)

in Normal

of age (22), and in adults (23). However, as QCT involves high radiation exposure, it is not widely used in children. Several investigators have attempted to derive bone volume mathematically based on the bone area data of DXA in normal children as well as patients with various growth problems (16,24-27). Assuming that both lumbar spine and femoral neck are cubicle, Katzman (16) found that the vBMD of the femoral neck bore no relationship to age, height, or weight in a group of 50 normal adolescents. However, in contrast to the femoral neck findings, lumbar spine remained significantly related to age and growth (16). This study was performed to investigate the influence of age and growth on vBMD using a more direct assessmentof bone volume than that employed in previous studies involving DXA (16,24,25) in lumbar spine, the femoral neck, and the midfemoral shaft, based on bone area data provided by DXA in a normal population, aged 5-27 yr. Subjects

and Methods

Subjects Two hundred and nine subjects (109 males) of a previous study (2), aged 5.6-27.0 yr, were included. The study population has been described previously (2). Individuals were selected by height and weight between the 1st and 99th percentiles (height and weight SD scores between -2.33 and 1-2.33) for age and sex and body weight above 25 kg. The latter criterion was recommended by the densitometer manufacturer because the adult software version was employed. All individuals were healthy, with no known medical illness. Written consent was obtained from parents or from subjects who were older than 16 yr of age.

Methods Anthropometry. Height scale) were measured

(Harpenden on the day

1586 Downloaded from jcem.endojournals.org on June 7, 2006

stadiometer) of the DXA

and weight scans. Weight

(electronic was mea-

VOLUMETRIC sured with minimal clothing. calculated based on published

BONE DENSITY

Height and weight SD scores (z-score) were normal data (28) according to sex and age.

Dual energy x-ray absorptiometry measurements. DXA measurements were made with a scanner (model DPX, Lunar Radiation Corp., Madison, WI) equipped with adult software for total body (version 3.4), lumbar spine (Ll-L4), and femoral neck. Detailed methodology was described previously (2). Bone volume was calculated based on the two-dimension DXA measurements on Ll-L4, femoral neck, and femoral shaft (regional analysis of the total body software). The height and width of the right femoral neck were provided by femoral neck scans, whereas in the femoral shaft, a manual analysis was needed to measure the height and width of the middle third portion (Fig. 1). All manual analyses were performed by the same investigator (S.L.J.). The estimation of Ll-L4 volume was based on a method described by KrGger et al. (24). In this method, each lumbar vertebral body was approximated as a cylinder (Fig. 1). The diameter (d,) and the height (not shown in Fig. 1) of the four vertebral bodies were obtained from DXA spine scans. The bone volume of each vertebral body is calculated as r x (d,/2)’ x height. The middle third of the femur is considered to be cylindrical in shape (29) (Fig. 1). As shown in Fig. 1, the height of the middle third of the femur is expressed as h,, and the diameter of the femur is d,. Hence, the volume of the middle third of the femur is T X h, X (d,/2)*, where d, is the area/h,. Similar to the femoral shaft, the portion of the femoral neck measured was assumed to be a cylinder (Fig. 1). The diameter and height of this cylinder are expressed as d, and h,, respectively. The bone volume of the femoral neck is m X h, X (d,/2)‘. The volumetric BMD of all three sites was expressed as BMC per volume (grams per cm3). A quality control program was instituted during the study period. The coefficient of variation (SD/mean) was 1.0% for aBMD of Ll-L4 (aLBMD) and 1.5% for aBMD of the femoral shaft (aFBMD). In vitro validation of bone volume. The accuracy of DXA in measuring the midfemoral shaft volume was validated against established methods using 3 lamb and 6 venison femora. The femoral length of these 9 specimens ranged from 15.6-28.8 cm, equivalent to the femora sizes in 5- to IO-yr-old children in the current study. Readings on the height and bone area of the middle third of each bone were obtained from DXA using manual analyses of the total body scans, and bone volume was then calculated, as described above. After the DXA scans, the middle third of each bone was obtained, and the average height and diameter were calculated based on 10 measurements of height and diameter along the shaft using a calliper. The bone marrow cavities were sealed off using Parafilm, and the bone volume was measured by a water displacement method.

IN SUBJECTS

AGED 5-27 YR

1587

Results of the validation: in vitro experiment. There was an excellent linear relationship between bone volume determined by water displacement and that determined by calculation based on calliper measurement of diameter and bone length (3 = 0.99; P = 0.0001; Fig. 2). There was also a strong relationship between bone volume measured by DXA and that measured by water displacement (3 = 0.82; P = 0.0001; Fig. 2). Statistical

analysis

The Statistical Analysis System (30) was used for data analysis. Correlations and associations between variables were examined using linear regression models. Proportions of contribution by each potential variable were examined using a multiple linear regression model similar to the method of Glauber et al. (31). The sex difference between linear models was further assessed using interaction terms. Significance was set at less than 0.05. As the aBMD of all sites increased with age until the late teens, when it plateaued, nonlinear regression procedures by least squares (Statistical Analysis System) were used with segmented polynomial models to detect the point of plateau occurring (inflection point) and to compute the best-fit curves. This method was previously described (2).

Results

The mean age was 16.3 yr for 109 male subjects and 16.4 yr for 100 female subjects. The mean height and weight SD scoresranged from 0.1-0.2 in both sexesand were not different from normal (P > 0.05). Lumbar

spine (Ll-L4)

There was a significant age-dependent increasein aLBMD in both sexes(Fig. 3). This increaseplateaued around the age of 17.4 yr in males and 15.7 yr in females. aLBMD was also height and weight dependent (Table l), and this dependence remained significant after allowing for age (P < 0.05 for all). Similarly, volumetric Ll-L4 BMD (vLBMD) was associated with age, height, and weight in both males and females (Fig. 3 and Table 1). After allowing for age, vLBMD remained significantly related to weight, but not to height, in both sexes (data not shown). We examined the relationship between the vLBMD and aLBMD z-scores computed as previously described (2). These two measurements correlated well (r* = 0.40; P = 0.0001). Midfemoral

shafr

aFBMD increasedwith age in both sexesuntil 17.2yr of age in malesand 13.5yr of agein females,when it plateaued (Fig.

gig ;pJ ~ij r/y 010 20600

‘I

t

30

40

50

70

Volume...H20

20

30

40

50

60

70

(m3)

2. In vitro validation of bone volume. The bone volume of three sheep and six venison femora (middle third) measured by DXA (volumeDXA) and calliper (volume-calliper) is plotted against that measured by the water displacement method (volume-H,O) together with lines of regression. There is a significant correlation between volumeDXA and volume-H,0 (r2 = 0.82; P = 0.0001) and between volumecalliper and volume-H,0 (r2 = 0.99; P = 0.0001). FIG.

1. Schematic illustrations of the regions studied. The shapes of both femoral shaft and neck are regarded as cylinders with diameter (d, and d,) and height (h, and h,) shown, respectively. A cross-section of a lumbar vertebrae body is illustrated with the body and the posterior and transverse processes. Only the diameter, not the height, of the vertebral body is measured cd,).

FIG.

10

Downloaded from jcem.endojournals.org on June 7, 2006

1588

LU ET AL. 1.8 ,

0

5

10

15

20

25

.6 I

.l 0

5

FIG. 3. aLBMD females together

10

15

20

and vLBMD with lines

25

0 I

.6 ,

30

0

are plotted of best-fit.

5

10

15

5

IO

15 20

against

20

25

30

25

30

age for males

and

4). Height and weight were also strongly correlated with aFBMD (Table l), and these associationsremained after allowing for age (P < 0.05 for all). In contrast, vFBMD bore no relationship with age (Fig. 4) or height (P > 0.05) in both sexes.vFBMD was constant acrossthe age range studied, and males had higher vFBMD than females (mean t SD, 0.73 + 0.11 for males and 0.70 ? 0.12 for females; P = 0.047). In females, a significant inverse relationship was seenbetween vFBMD and weight (r* = 0.14; P = 0.001). There was a significant linear relationship between vFBMD and aFBMD z-scores (r* = 0.25; P = 0.0001). Femoral

neck

Similar to the femoral shaft, aNBMD was strongly age dependent (Fig. 5). aNBMD was also related to height and weight (Table 1) even after allowing for the age factor (P < 0.05). In contrast, vNBMD had no association with age or weight in either sex (Fig. 4). vNBMD was related to height in females,but not in males(Table 1). Furthermore, there was no sex difference in vNBMD (mean t SD, 0.41 2 0.06 for males and 0.40 + 0.06 for females; P = 0.45). Similar to the lumbar spine and femoral shaft, vNBMD correlated significantly with the aNBMD z-score (r* = 0.42; P = 0.0001). Correlation

between BMD

Discussion

I

.3 -

30

values of the three sites

The aBMD values of these three regions were highly related in both sexes(r2 ranged from 0.48-0.79; P < 0.0001for all). As age and growth variables were codependent for aBMD at the three sites, the relationships among the sites were further examined using multiple linear regression. After adjusting for age, height, and weight, the proportion of aBMD that remained to explain the relationships among sites ranged from 3-18% (r* ranged from 0.03-0.18), and the relationships remained significant among the three sites for both sexes. In contrast, there was no significant correlation between vFBMD and vNBMD or between vFBMD and vLBMD in either sex (data not shown). vNBMD correlated significantly with vLBMD in both male (r* = 0.12;P = 0.0004) and female (r2 = 0.05; P = 0.03) subjects.

JCE & M . 1996 Vol81*No4

This study, performed in a large normal population using DXA, demonstrated that vBMD of femoral neck and femoral shaft is independent of age throughout childhood and early adulthood. Similar results have been obtained using QCT at different sites in small numbers (19, 21, 22). These observations offer a different perspective from our previous concepts of aBMD and are consistent with the theoretical concept that mechanical loading during growth and development should result in the same bone density at any age. The dependence between aBMD and growth parameters has been observed in several studies of normal healthy children (2, 6, 8, 12, 13), but the relevance of this relationship when assessingBMD in disorders in which growth may be affected has usually not been appreciated. Interpretation of aBMD, for example, in children with chronic renal failure will differ significantly if the values are corrected for height rather than agebecauseof their marked short stature (32).We have used predictive equations to correct for height based on our normal population data, whereas other investigators have suggesteddifferent equations (16,20). When aBMD was age adjusted using a z-score, it was not surprising that there was a positive relationship with vBMD at each of the sites studied. vBMD, becauseit is independent of age and height, does not require any adjustment, and the value of a constant reference range in the pediatric age range cannot be overemphasized. This should assistnot only in the interpretation of BMD values at the femoral neck and shaft in individuals whose growth is affected, but also in studies examining other factors that might influence both growth and bone density, such as nutrition, calcium intake, exercise, and genetic factors. vBMD measured by the current technology of DXA is restricted to sites at which bone volume can be derived from the bone area (16, 24). This requires assumptions of bone shape.The assumption of bone shapeas a solid cylinder may be true in the femoral neck region, but cannot be the casein the femoral shaft due to its apparent hollow structure. In animal femora in the same size range as that in our young children, the assumptions used to calculate bone volume as cylinders were valid, as there was an excellent linear relationship between bone volume determined by calliper and that determined by water displacement. The minor variations from real volume should not affect the mean for the study population, but may increase the range of the values observed. Alternative methods that have been employed to calculate bone volume use different assumptions for bone shapeat the femoral neck and lumbar spine (16). Comparison between our method and that of Katzman et al. (16) provides similar results at these two sites (data not shown). It is to be hoped that with advances in technology, actual volume may be measured by DXA using the low dose radiation that currently makes it a suitable investigative tool for children. In the current study there is an age and growth dependence of vBMD of Ll-L4, which is in contrast to the findings at the femoral neck and femoral shaft. It is possible that the true density of lumbar spine increaseswith age because of the increase in the number of trabeculae during growth. Alternatively, this may be due to the method that we used

Downloaded from jcem.endojournals.org on June 7, 2006

VOLUMETRIC TABLE

1. Relationship

between

BMD

BONE

and height

DENSITY

and weight

IN SUBJECTS

(values

AGED

5-27 YR

of rz and probability)

Male

Female

Ht

wt

P

I-“

aLBMD aFBMD aNBMD vLBMD vFBMD vNBMD

0.71 0.80 0.52 0.37 0.01 0.007

0

5

10

15

25

30 Age

.21 0

1 5

1 10

1 15

1 20

1 25

and vFBMD with lines

8,

P

0.75 0.75 0.59 0.40 0.01 0.006

0

5

10

15

20

25

30

5

10

15

20

25

30

.2I

1 30

0 (ye@

are plotted of best-fit.

,

against

I.2

0.0001 0.0001 0.0001 0.0001 0.27 0.44

(years)

Age

FIG. 4. aFBMD females together

Ht

l.2

0.0001 0.0001 0.0001 0.0001 0.45 0.85

20

1589

age for males

and

Wt

rz

P

0.71 0.67 0.47 0.28 0.004 0.067

0.0001 0.0001 0.0001 0.0001 0.55 0.01

0.74 0.50 0.40 0.35 0.14 0.03

P 0.0001 0.0001 0.0001 0.0001 0.001 0.07

volumetric BMD remained stable, except for some age-dependent increase between 12-15 yr of age. We found that the aBMD values of the three sites are closely related, which is similar to previous findings (8, 24). These associations remained significant after allowing for the difference in age and growth parameters. Therefore, the lack of association between vFBMD and vLBMD and between vFBMD and vNBMD was rather unexpected. The skeleton is heterogeneous in the proportion of cortical and trabecular bone at different sites (30). The femoral shaft consists predominantly of cortical bone, in contrast to trabecular bone in the femoral neck and lumbar spine regions. The significant association between vNBMD and vLBMD, but not between vFBMD and vNBMD or vLBMD suggests that the relationship of true BMD between regions may only exist in bones of similar type. In conclusion, we demonstrated a method of calculating vBMD based on the two-dimension DXA output of bone mineral content and bone area in childhood and early adulthood. In contrast to aBMD, vBMD is independent of age and less dependent on growth variables. Further studies of the clinical significance of vBMD and its relationship with the known bone density regulators, such as genetic markers, exercise, and nutritional factors, are needed. Acknowledgments

8,

I

The authors would like to thank Humphries, Kate Morley, and David recruiting subjects for the study.

Graham Sillence

Ogle, Jane Allen, for their contribuCon

Ian in

References Age (years) FIG. 5. aNBMD females together

and vNBMD with lines

are plotted of best-fit.

against

age for males

and

to approximate the bone shape of the lumbar vertebral bodies. As shown in Fig. 1, the lumbar spine vertebrae body is neither a cylinder (24) nor a cube (16). Furthermore, as the x-ray beam comes from a posterior-anterior direction, the posterior and some part of transverse processes are inevitably included in the quantitation of BMC, but not in bone area. Thus, a combination of posterior-anterior and lateral DXA scans may provide more realistic assessment of the bone dimension of the lumbar spine (33, 34). Alternatively, QCT is able to measure the bone volume of lumbar vertebral bodies (21). A study of children, aged 2-15 yr, showed that

1. Kanis JA, Melton J, Christensen C, Johnston CC, Khaltaev N. lYY4 The diagnosis of osteoporosis. J Bone Mmer Res. 9:1137-l 141. 2. Lu PW, Briody JN, Ogle GD, et al. lY94 Bone minual density of total body, spine, and femoral neck in children and young adults: a cross-sectional and longitudinal study. J Bone Miner Res. 9: 1451-1458. 3. Bonjour JP, Theintz G, Buchs B, Slosman D, Rizzoli R. 1 YYl Crltlcal years and stages of puberty for spinal and femoral bone mass accumulation durmg adolescence. J Clin Endocrinol Metab. 73:555-563. 4. Gunnes M. 1994 Bone mineral density in the cortical md trabrcular distal forearm in healthy children and adolescents. Acta I’acdiatr. 83:463-467 5. Slemenda CW, Reister TK, Hui SL, Miller JZ, Christian JC, Johnston Jr CC. 1994 Influences on skeletal mineralisation in children and ad&scents: evidence for varying effects of sexual maturation and physical actlwty. J I‘cdmtr. 125:201-207. 6. de1 Rio L, Cnrrascosa A, I’ons F, Gusinye M, Yeste D, Dotnenech FM. 1994 Bone mineral density of the lumbar spine in white Mcditermnean Spnnlsb children and adolescents: changes related to age, sex, and puberty. I’cdiatr Res 35:362-366. 7. Rubin K, Schirduan V, Gendreau P, Sarfarazi M, Mend& R, Dalsky G. 1993 Predictors of axial and peripheral bone mineral density in lwnltby children and adolescents, with special attention to the role of puberty. J I’ediatr. 123:863-870. 8. Faulkner RA, Bailey DA, Drinkwater DT, Wilkinson AA, Houston CS, McKay HA. 1993 Regional and total body bone mineral content, bone mineral

Downloaded from jcem.endojournals.org on June 7, 2006

LU ET AL.

9.

10.

11. 12.

13.

14.

15. 16.

17 18 19 20

density, and total body tissue composition in children 8-16 years of age. Calcif Tissue Int. 53:7-12. Tsukahara H, Sudo M, Umezaki M, et al. 1992 Dual-energy x-ray absorptiometry in the lumbar spine, proximal femur and distal radius in children. Pediatr Radiol. 22:560-562. Miller JZ, Slemenda CW, Meaney FJ, Reister TK, Hoi S, Johnston CC. 1991 The relationship of bone mineral density and anthropometric variables in healthy male and female children. Bone Miner. 14137-152. Southard RN, Morris JD, Mahan JD, et al. 1991 Bone mass in healthy children: measurement with quantitative DXA. Radiology. 179x735-738. Glastre C, Braillon I’, David L, Cochat P, Meunier PJ, Delmas PD. 1990 Measurement of bone mineral content of the lumbar spine by dual energy x-ray absorptiometry in normal children: correlations with growth parameters. J Clin Endocrinol Metab. 70:1330-1333. Theintz G, Buchs 8, Rizzoli R, et al. 1992 Longitudinal monitoring of bone mass accumulation in healthy adolescents: evidence for a marked reduction after 16 years of age at the levels of lumbar spine and femoral neck in female subjects. J Clin Endocrinol Metab. 75:1060-1065. Lloyd T, Rollings N, Andon MB, et al. 1992 Determinants of bone density in young women. I. Relationships among pubertal development, total body bone mass, and total body bone density in premenarchal females. J Clin Endocrinol Metab. 75~383-387. Grimston SK, Monison K, Harder JA, Hanley DA. 1992 Bone mineral density during puberty in western Canadian children. Bone Miner. 19:85-96. Katzman DK, Bachrach LK, Carter DR, Marcus R. 1991 Clinical and anthropometric correlates of bone mineral acquisition in healthy adolescent girls. J Clin Endocrinol Metab. 731332-1339. Mazess RB, Barden H, Hauialen C, Vega E. 1994 Normalisation of spine densitometry. J Bone Miner Res. 9:541-548. Carter DR, Bouxsein ML, Marcus R. 1992 New approaches for interpreting projected bone densitometry data. J Bone Miner Res. 7:137-145. Compston JE. 1995 Bone density, BMC, BMD, or corrected BMD? Bone. 165-7. Prentice A, Parsons TJ, Cole TJ. 1994 Uncritical use of bone mineral density in absorptiometry may lead to size-related artifacts in the identification of bone mineral determinants. Am J Clin Nutr. 60:837-842.

JCE & M . 1996 Vol81 . No 4

21. Gilsanz V, Gibbens DT, Roe TF, et al. 1988 Vertebral bone density in children, effect of puberty. Radiology. 166847-850. 22. SchGnaii E, Wenzlik U, Michalk D, Scheidhauer K, Klein K. 1993 Is there an increase of bone density in children? Lancet. 342:689-690. 23. Compston JE, Evans WD, Crawley EO, Evans C. 1988 Bone mineral content in normal UK subjects. Br J Radiol. 61:631-636. 24. KrSger H, Kotaniemi A, Vainio P, Alhava E. 1992 Bone densitometry of the spine and femur in children by dual-energy x-ray absorptiometry. Bone Miner. 17:75-85. 25. Bachrach LK, Loutit CW, Moss RB. 1994 Osteopenia in adults with cystic fibrosis. Am J Med. 9627-34. 26. Neely EK, Marcus R, Rosenfeld RG, Bachrach LK. 1993 Turner syndrome adolescents receiving growth hormone are not osteopenic. J Clin Endocrinol Metab. 76861-866. 27. Kohlmeier L, Gamer C, Marcus R. 1993 Bone mineral status of women with Marfan syndrome. Am J Med. 95:568-572. 28. Hamill PV, Drizd TA, Johnson CL, Reed RB, Roche AL, Moore WM. 1979 Physical growth: National Centre for Health Statistics Percentiles. Am J Clin Nutr. 32;607-629. 29. Warwick, Williams, eds. 1973 Gray’s anatomy, 35th ed. Edinburgh: Longman Group. 30. Schlotzhauer SD, Litiell RC. SAS system for elementary statistical analysis. Gary: SAS Institute. 31. Glauber HS, Vollmer WM, Nevitt MC, Ensrud KE, 01~011 ES. 1995 Body weight versus body fat distribution, adiposity, and frame size as predictors of bone density. J Clin Endocrinol Metab. 80:1118-1123. 32. Cowell CT, Lu PW, Lloyd-Jones SA, et al. 1995 Volumetric bone mineral density-a potential role in paediatrics. Acta Paediatr Stand (Suppl). 411:12-16. 33. Peel NFA, Eastell R. 1994 Diagnostic value of estimated volumetric bone mineral density of the lumbar spine in osteoporosis. J Bone Miner Res. 9:317320. 34. Jergas M, Breitenseher M, Gluber CC, Yu W, Genani HK. 1995 Estimates of volumetric bone density from projectional measurements improve the discriminatory capability of dual x-ray absorptiometry. J Bone Miner Res. 10: 1101-1110.

Downloaded from jcem.endojournals.org on June 7, 2006