Evaluation of vertebral volumetric vs. areal bone mineral density during growth

Bone Vol. 20, No. 6 June 1997:553-556

ELSEVIER

Evaluation of Vertebral Volumetric vs. Areal Bone Mineral Density During Growth S. M . O T T , M. O ' H A N L A N ,

E. W . L I P K I N , a n d L. N E W E L L - M O R R I S

Division of Metabolism, University of Washington, Seattle, WA, USA

projected area from DEXA scans have been used. l A recent study calculated volumetric bone density from both lateral and posterior-anterior DEXA scans and compared this with directly measured volume in adult cadavers. 18 This study includes direct measurements of the dimensions of vertebrae during growth. Here, we have studied cadavers in which we measured the areal bone density by DEXA, vertebral volume by underwater weight, mineral content by ash weight, and dimensions of lumbar vertebrae by calipers. The studies were performed in young pigtail macaques (Macaca nemestrina), because we wanted to examine primates during and after the growth period. The purpose of the study was to examine the changes in shape of vertebral bodies during growth, to derive equations to predict the volumetric bone density from the noninvasive measurements, and to determine the error associated with these measurements.

Bone mineral "density" (BMD) measured by dual-energy X-ray absorptiometry (DEXA) does not represent the volumetric density (grams per cubic centimeter), but rather the areal density (grams per square centimeter). This distinction is important during growth. The purpose of this study was to measure vertebral dimensions in cadavers of young pigtail macaques (Macaca nemestrina), and to derive equations to predict the volumetric bone density from noninvasive measurements. We measured the areal bone density by DEXA, vertebral volume by underwater weighing, mineral content by ashing, dimensions of lumbar vertebrae by calipers, and dimensions of vertebrae by radiography. Somatometric measurements of the female lumbar vertebral bodies showed that the shape changed during growth. The bone mineral content from the densitometer correlated significantly with the ash weight (r = 0.99, error 8.7%). The correlation coefficient between the volumetric bone mineral density and areal BMD measurement was significant (r = 0.68, p < 0.0001) with a 9.5% error; this improved significantly to 0.82 (7.2% error) when the BMD was divided by the vertebral depth from the radiograph. Areal BMD showed a strong correlation with age (r = 0.82,p < 0.0001), with an average increase of 7.4%/year. In contrast, volumetric mineral density showed a weak relationship with age (r = 0.43,p < 0.01), for an average increase of 1.5%/year. When studying bone mineral density during growth, the differences between volumetric and areal bone mineral density should be taken into consideration. (Bone 20: 553-556; 1997) O 1997 by Elsevier Science Inc. All rights reserved.

Materials and Methods Animals The animals were from the primate field station of the Regional Primate Research Center at the University of Washington, Medical Lake, WA. Animals who had died or been euthanized were used for this study. Causes of death included acute infection and trauma. There were 10 males and 40 females, age range 0 . 6 9 13.5 years. 14 females were feral animals that had been imported from Southeast Asia; the remaining animals were born at the field station. The ages of the feral animals were estimated from body weight at the time of admission to the colony, and could be inaccurate by up to 1 year. Unlike New World primates, macaques are not truly arboreal and do not have elongated limbs or a linear body build. They are more arboreal than are humans, and a large part of their locomotion is quadrupedal. They spend a large part of the day sitting in an upright position with weight supported on the ischial tuberosities and the feet.

Key Words: Bone mass; Bone mineral density; Vertebral volume; Vertebral growth; Primate.

Introduction Dual-energy X-ray absorptiometry (DEXA) is a commonly used method of measuring the bone mass. Resulting values are usually presented as "bone mineral density" (BMD). These values do not represent the volumetric density (grams per cubic centimeter), but rather an areal density representing the grams of bone mineral in a projected area of bone (grams per square centimeter). Several methods of estimating the vertebral volume from the

Bone Mineral Density Measurements At the time of necropsy, the lumbar spine was removed in a block section, which included the upper part of the sacrum and about 2 cm of soft tissue around the spine. The section was scanned in the supine position in both the lateral and posterior-anterior (PA) directions using a Norland dual-energy X-ray absorptiometer equipped with software for small subjects. The pixel size was 1.0 X 1.0 mm, and scan speed 60 mm/s. The second, third, and fourth lumbar vertebrae were each measured. For both the PA and lateral views, the entire vertebra was included in the range of interest, including the posterior elements and transverse processes. The spinous processes were included in the lateral view

Address for correspondence and reprints: Dr. Susan M. Ott, Division of Metabolism, University of Washington, Box 356426, Seattle, WA 98195-6426. © 1997 by Elsevier Science Inc, An rights reserved.

553

8756-3282/97/$17.00 PII $8756-3282(97)00057-4

554

S.M. Ott et al. Direct measurement of volumetric BMD

Bone Vol. 20, No. 6 June 1997:553-556 using a piece of fishing line which was strung through the vertebral arch. After 5 min, each vertebra was again dipped into the paraffin. No gas escaped, indicating that the surface of the bone was sealed with wax. Vertebra were inspected to be sure the wax layer was translucent, indicating a thin layer, and not opaque. Each vertebra was weighed three times in air on a Sartorius scale. The vertebra was then suspended in water and the weight measured three times using the scale. The weight in air minus the weight in water (in grams) equals the volume of the vertebra (in cubic centimeters).

Ashed Weight Depth Figure l. Schematic drawings of a vertebra showing locations of caliper measurements.

to avoid introducing error from an operator-placed region of interest. The densitometer results were reported as the BMC (bone mineral content, in grams), area (projected bone area as determined by the software, in square centimeters), and the BMD (bone mineral density, in grams per square centimeter, equal to the BMC/area) for each vertebra. The overall B M D was the sum of the BMCs for each vertebra divided by the total projected bone area of L 2 - 4 . To measure precision, scans were repeated two to four times (mean 2.9) in 17 spines, with repositioning between scans. The average coefficient of variation was 2.2% for the PA BMD, 4.0% for the PA BMC, 3.0% for the lateral BMD, and 4.2% for the lateral BMC.

111 vertebrae from 37 animals were placed in dried, preweighed crucibles and ashed in a muffle furnace at 550°C for 6 h. Crucibles and contents were then cooled and placed in a desiccator for at least 1 h. They were then weighed in triplicate. For 11 animals, all three vertebrae were ashed in the same crucible. Vertebrae from two animals were not ashed because they were lost.

Calculated Values The volumetric bone mineral density (grams per cubic centimeter) was the ashed weight (sum of L-2, L-3, and L-4) divided by the volume. The bone mineral apparent density (BMAD) was calculated as the BMC/scan areal5; this formula was derived by Katzman et al. s for use in human adult vertebrae, whose dimensions were assumed to be cuboidal.

Statistical Analysis Radiographs A radiograph of the spinal section was then taken in both the PA and lateral projections. The film to X-ray distance was 70 cm. The soft tissues were then removed, the vertebrae separated, and each one X-rayed in the lateral and cephalocaudal orientation. The width (transverse) and depth (PA) of each vertebra was measured with a sliding caliper. The depth included the spinous process, because this was also included in the BMD measurement.

Somatometric Measurements" After the radiographic measurements the soft tissues were removed from the spine by blunt dissection. The intervertebral disks were bisected and the vertebrae were separated. The remaining cartilage and soft tissue were removed with a fine scraper or dental pick. Using dial calipers, measurements of the vertebral bodies were taken in three dimensions (Figure 1): depth (PA), from the most anterior point in the foramen vertebra to the point directly ventral along the midline across the inferior surface of the vertebral body; height (superior-inferior), from the leftmost point on the extremitas cranialis to the leftmost point on the extremitas caudalis, and similarly on the right side; and width (transverse), along the widest transverse line across the superior and across the inferior surfaces. Each measurement was taken four times. The averages of the right and left measurements (for height) and of the superior and inferior measurements (for width) were used.

Volume Determination Vertebrae were dried for 12 h at 80°C, then cooled to room temperature. They were dipped into melted paraffin for 2 sec

Regression analysis was done using the STATVIEW program for personal computers (Abacus Concepts, Inc., Berkeley, CA, version 4.01). Regression coefficients were compared using Fisher's Z-transformation. 9 Results are reported as mean ± standard deviation unless noted. Results In 111 individual vertebrae the median ash weight was 1.45 g with a range of 0.245-3.42 g. The PA BMC from the densitometer correlated significantly with the ash weight (r = 0.99). The slope was 0.995 (95% confidence interval 0.974-1.016) and the intercept - 0 . 0 1 8 ( - 0 . 0 5 7 - 0 . 0 2 0 ) . The root-mean-square (rms) residual was 0.086 g, or 5.3%. For the lateral orientation the intercept was - 0 . 0 8 3 ( - 0 . 1 4 6 - - 0 . 1 9 0 ) , the slope 0.879 (0.849-0.910) with r = 0.98, and the rms residual was 0.138 g or 8.7%. Because these vertebrae were not independent (three from each animal) we also performed the analysis on just the L-4 vertebra with similar results. The error with the PA BMC was 5.0% and with the lateral BMC it was 9.2%. The p values for all these correlations were <0.001. The remaining analyses were done only on female vertebrae to avoid gender differences in the shape of the vertebral body. The somatometric measurements of the female vertebral bodies (average of L-2, L-3, and L-4) showed that the shape of the vertebra changed during growth (Figure 2). In the youngest animals the width was greater than the height, but by age 8 the height and width were similar. All three dimensions showed increases until approximately age 8. The mean volume of the entire irregularly shaped vertebra (sum of L-2, L-3, and L-4) was 9.37 ± 3.5 cm 3. This showed good concordance with the volume determined by multiplying the height, width, and depth of the vertebral bodies (mean 8.77 ± 3.3 cm3). The average absolute difference between these volume

Bone Vol. 20, No. 6 June 1997:553-556

S.M. Ott et al. Direct measurement of volumetric BMD .8

2.2'

::"

2 '

;"

1 . 8 '

~°

1.6'

ml•i

00

0

.7

'-o

~ .s

1 . 2 "

.3

0

o°O

~

¢

oo o°o°o

0%

.2

O0

¢

.1

o 000

•

i

•

2

1

.8

o ~i,

o

q,® O

O

o o o•

I:

.6

1 . 4 "

cm

i

•

4

i

•

6

i

,

8

J

•

10

i

•

12

14

0

0

.6

.4

555

•

-

-

,

-

.

•

,

2

-

-

-

,

4

-

.

-

8

,

-

-

8

-

i

-

-

10

•

,

12

-

-

.7

-

14

.6

AGE,years

,u,

"~

Figure 2. Dimensions of vertebral bodies in female Macaca nemestrina. Each point is the average of L-2, L-3, and L-4. Filled squares are width, open circles are height, and open diamonds are depth (PA).

.5

.4

u

measurements was only 0.69 cm 3. The regression equation was: volume = 0.182 + 1.03 × calculated volume (r = 0.97). The median volumetric bone mineral density of L - 2 - 4 in the females (n = 39) was 0.576 g/cm 3 (range 0.434-0.726). Linear regression showed that the correlation coefficient between the volumetric bone mineral density and areal PA BMD measurement was 0.68 (p < 0.0001) with a rms residual of 0.054 g (9.5% error); this improved significantly to 0.82 (7.2% error) when the PA BMD was divided by the measured depth on the radiograph (Figure 3). The BMAD showed an intermediate correlation coefficient of 0.76 (8.3% error). Equations to predict the volumetric bone mineral density were: 0.386 + 0.391 X BMD; 0.258 + 2.15 × BMAD; and 0.171 + 2.264 × BMD/X-ray depth. .75 .7 •

.65

2

•

i

4

,

i

6

•

i

•

8

u

10

,

n

12

•

14

AGE, years Figure 4. Areal vs. volumetric bone density of L-2-4 vs. age in female M a c a c a nemestrina. PA BMD was measured by DEXA; volumetric bone density (grams per cubic centimeter) was determined by ash weight and underwater weighing. Filled circles represent animals born at the field station; open circles represent feral animals. Linear regression using the areal PA BMD (L-2-4) suggests a strong correlation with age (r = 0.82, p < 0.0001), with an average increase of 7.4%/year. In contrast, the volumetric mineral density showed a weak relationship with age (r = 0.43, p < 0.01; Figure 4), for an average increase of 1.5%/year. The linear model may not be the best model in this situation; however, in our study, there were not enough animals to formulate a more detailed model to determine whether or not there was an increase in volumetric density at the time of sexual maturation (2-3 years).

Discussion

.5 .5 .4S .4

...l

.I

....

.2

t ....

.3

i ....

.4

r ....

.5

i ....

.6

l'"

.7

.8

PA BMD

.75

.7

: ..ex,4"

.65

O ~ O 0 .55

.5 .45

.4 .12

.14

.16 .18 .2 .22 BMD/Xray depth

Figure 3. Densitometry measurements vs. volumetric bone mineral density (grams per cubic centimeter) as determined by ash weight and underwater weighing of L-2-4 from female M a c a c a nemestrina. (Top) y = 0.386 + 0.391 X PA BMD (r = 0.68). (Bottom) y = 0.171 + 2.26 X BMD/X-ray depth (r = 0.82).

Our results agree with those of others that DEXA accurately measures the mineral content of bone. Our error was 5.3%. Jayo et al. 6 reported an error of 6.2% in cynomolgus macaques. Ho et al. 4 found a standard error of 8.9% using human vertebrae. An interesting observation was that the volume of the irregularly shaped vertebra was closely approximated by the product of the height, width, and depth of the vertebral body. This can be visualized by imagining a block of clay from which an oval vertebral body is carved, then the scraps used to make the rest of the vertebra. This implies that the posterior and transverse elements are proportional to the volume of the vertebral body, and thus, including spinous processes in bone density measurements will similarly affect all subjects in study cohorts. Early studies of bone mass measurements used various methods to estimate the volumetric bone mineral density from the areal density. 15'19 In a cadaver study, Ho et al. 4 measured the volume of vertebrae and found that the volumetric density was predicted from the DEXA areal BMD with a 9.2% error. However, most reports using DPA or DEXA ignored the difference between volumetric and areal density until Katzman et al. 8 focused attention on the issue by describing a calculated volumetric bone mineral density they termed "bone mineral apparent

556

S . M . Ott et al. Direct measurement of volumetric BMD

density" (BMAD). Sabin et al. is recently measured cadaveric spines from 12 elderly individuals. Volumetric B M D was calculated using measurements from the lateral scan, and volumetric density was determined by ashing and measuring the volume by displaced sand. The correlation coefficient between the volumetric bone density and the calculated D E X A volumetric bone density was 0.988, which was significantly better than the correlation with the AP B M D (r = 0.947). These results are consistent with our findings. In this study we have predicted the volumetric bone mineral density from the areal bone mineral density with an error of 9.5%. The prediction is improved to a 7% error if a measurement of the vertebral depth (from a lateral radiograph) is included. We found that the shape of the vertebral body changes during growth. This may explain our finding that an equation based on the square root of the projected area ( B M A D ) had more error than an equation which included a measurement of the depth of the vertebrae. In humans, the dimensions also change at different rates; the height increases faster than the width, z° Our study did not address the issue of whether volumetric bone mineral density could more accurately predict either bone strength or fracture risk. Epidemiologic studies in adults have suggested that smaller individuals are more likely to develop fractures, so their relatively lower areal bone density would further lower areal B M D measurement. Thus, areal B M D , although less accurate in terms of correlating with the volumetric density, might be better able to predict fractures. ~ Several studies on prediction of fractures utilizing calculated D E X A volumetric bone density have not shown enhanced ability to predict fractures. ~'~3"~6 Estimations using lateral measurements resulted in greater ability to discriminate patients with vertebral fractures. 7 Measurement o f vertebral depth by lateral X-ray improved the ability of D E X A to predict vertebral fractures. 17 When studying bone mineral density during growth, the difference between volumetric and areal bone mineral density is important. The areal density increases with growth, even without changes in the volumetric density. Our data show that most of the apparent increase in B M D can be explained by increases in size during growth; a small increase in volumetric density with age was still seen. The equations derived here are applicable to the species under study, but there is evidence that similar patterns are seen in human children. For example, in prepubertal children, studies measuring vertebral bone mineral density using QCT have indicated little change with age, 3 whereas similar studies using DEXA have shown increases with age. ILH During puberty, both size and density increase) Differences between male and female• adolescent bone density can be explained by differ.) ences in size.- Furthermore, age-related increases in BMD, as inferred from D E X A measurements, disappear when volumetric density is calculated• n).12

Acknowledgments: W e a c k n o w l e d g e the a s s i s t a n c e o f Craig A u m a n n , D i a n a G r e e n l e e , and J a n e t L o n g . T h i s w o r k w a s supported by P H S G r a n t Nos. RO1 A R 4 0 8 1 3 - 0 1 and R R 0 0 1 6 6 .

B o n e Vol. 20, No. 6 June 1997:553-556

References I. Compston, J. E. Bone density: BMC, BMD, or corrected BMD? Bone 16:57: 1995. 2. Gilsanz, V., Boechat, M. 1., Foe, T. F., Loro, M. L., Sayre, J. W., and Goodman, W. G. Gender differences in vertebral body sizes in children and adolescents. Radiology 190:673-677; 1994. 3. Gilsanz, V., Gibbens, D. T., Roe, T. F., Carlson, M., Senac, M. O., Boecha, T. M. I., Huang, H. K.. Schulz, E. E., Libanati, C. R., and Cann, C. C. Vertebral bone density in children: effect of puberty. Radiology 166:847 850; 1994. 4. Ho, C. P., Kim, R. W., Schaffler, M. B., and Sartoris, D. J. Accuracy of dual-energy radiographic absorptiometry of the lumbar spine: cadaver study. Radiology 176:171-173; 1990. 5. Hui, S. L., Slemenda, C. W. Carey, M. A.. and Johnston, C. C. Choosing between predictors of fractures. J Bone Miner Res I 0:1816-1822; 1995. 6. Jayo, M. J., Rankin, S. E., Weaver, D. S., Carlson, C. S.. and Clarkson. T. B. Accuracy and precision of lumbar bone mineral content by dual energy x ray absorptiometry in live female monkeys. Calcif Tissue Int 49:438-440; 1991. 7. Jergas, M., Breitenseher, M., Gluer, C. C., Yu, W., and Genant. H. K. Estimates of volumetric bone density from projectional measurements improve the discriminatory capability of dual X-ray absorptiometry. J Bone Miner Res 10:ll01-1110: 1995. 8. Katzman, D. K., Bachrach, L. K., Carter, D. R., and Marcus, R. Clinical and anthrometric correlates of bone mineral acquisition in healthy adolescent girls. J C l i n E n d o c r i n o l M e t a b 7 3 : 1 3 3 2 1339; 1991. 9. Kleinbaum, D. G. and Kupper, L. L. Applied Regression Analysis and Other Multivariable Methods. Boston: Duxbury: 1978. 10. Kroger, H., Kotaniemi, A., Vainio, P., and Albava, E. Bone densitometry of the spine and fumur in children by dual-energy x-ray absorptiometry. Bone Miner 17:75 85; 1992. 11. Lu, P. W., Briody, J. N. Ogle, G. D., Morley, K., Humphries. I. R. J.. Allen, J., Howman-Giles, R., Sillence, D., and Cowell, C. T. Bone mineral density of total body, spine, and femoral neck in children and young adults: A crosssectional and longitudinal study. J Bone Miner Res 9:145 I - 1458~ 1994. 12. Lu, P. W., Cowell, C. T., Lloyd-Jones. S. A., Briody, J. N.. and Howman-Giles, R. Volumetric bone mineral density in normal subjects, aged 5 27 years. J Clin Endocrinol Metab 81:1586-1590; 1996. 13. Mazess, R. B., Barden, H., Mautalen. C., and Vega, E. Normalization of spine densitometry. J Bone Miner Res 9:541 548; 1994. 14. McCormick. D. P., Ponder. S. W., Fawcett, H. D.. and Palmer, J. L. Spinal bone mineral density in 335 normal and obese children and adolescents: evidence for ethnic and sex differences. J Bone Miner Res 6:507-513; 1991. 15. Ott. S. M.. Kilcoyne. R. F., and Chesnut, C. H. Comparisons among methods of measuring bone mass and relationship to severity of vertebral fractures in osteoporosis. J Clin Endocrinol Metab 66:501 507: 1988. 16. Peel. N. F. A. and Eastell, R. Diagnostic value of estimated volumetric bone mineral density of the lumbar spine in osteoporosis. J Bone Miner Res 9:317 20; 1994. 17. Ross, P. D., Huang, C., Davis, J. W., and Wasnich, R. D. Vertebral dimension measurements improve prediction of vertebral fYacture incidence. Bone 16(Suppl.):257S 262S: 1995. 18. Sabin, M. A., Blake, G. M. MacLaughlin-Black, S. M., and Fogelman, I. The accnracy of volumetric bone density measurements in dual x ray absorptiometry. Calcif Tissue Int 56:210-214; 1995. 19. Sartoris, D. J., Sommer, F. G., Marcus, R.. and Madvig, P. Bone mineral density in the fizmoral neck: Quantitative assessment using dual-energy projection radiography. AJR 144:605-61 I; 1985. 20. Veldhuizen, A. G.. Baas, P., and Webb, P. J. Observations on the growth of the adolescent spine. J Bone Jt Surg 68B:724 728; 1986.

Date Received: June 4, 1996 Date Revised: February 4. 1997 Date Accepted: February 4, 1997