- Email: [email protected]
Accuracy and precision of bone mineral density and bone mineral content in excised rat humeri using fan beam dual-energy X-ray absorptiometry
Bone Vol. 30, No. 1 January 2002:243–246
Accuracy and Precision of Bone Mineral Density and Bone Mineral Content in Excised Rat Humeri Using Fan Beam Dual-energy X-ray Absorptiometry S. KASTL, T. SOMMER, P. KLEIN, W. HOHENBERGER, and K. ENGELKE Department of Surgery and Institute of Medical Physics, University of Erlangen, Erlangen, Germany
opathies and fracture risk in small laboratory animals a number of different osteodensitometric methods, such as peripheral quantitative computed tomography (pQCT)10,11 and dual-energy X-ray absorptiometry (DXA) have been used. BMC and BMD precision and accuracy measurements in rats have been reported by some investigators1,3,5,8,9,12; however, to our knowledge, an in vitro analysis of excised rat bones scanned with a fan beam DXA scanner has not yet been published. In this study we compare ash weight and ash mineral content with DXA BMC results of excised rat humeri and determine the short-term reproducibility of BMD and BMC. We also investigate the effect of the salinity of the water bath in which the samples were scanned. Excised bones were typically scanned in tap water, whereas, in in vivo scans, a physiological concentration of approximately 0.9% saline was present.
The aim of this study was the evaluation of fan beam dualenergy X-ray absorptiometry (DXA) for measuring bone mineral density (BMD) and bone mineral content (BMC) of isolated rat humeri. Defleshed rat humeri from male Lewis rats were examined with a Hologic QDR 4500 A (Hologic, Inc., Bedford, MA) high-resolution densitometer both in water and 0.9% saline solution. The small animal scan protocol with the regional high-resolution mode was used. BMC measured by DXA was compared with bone dry weight, ash weight, and bone calcium content. Furthermore, DXA BMD and BMC precision were determined. We also evaluated the effect of salinity of the water bath in which the bones were measured. Correlations (r2) of BMC, as determined by DXA with dry weight, ash weight, and bone calcium content, were 0.978, 0.988, and 0.890, respectively. DXA overestimated ash weight by 5%–9%. Precision errors for BMC (BMD) were 0.90% (0.76%) without and 1.3 (0.86) with repositioning. Changes in the salinity of the water bath had a significant influence on the DXA results: At the 0.9% physiological level, BMC (ⴚ4.4%) and area (ⴚ4.1%), but not BMD, values were significantly lower (p < 0.005) compared with measurements in tap water. Fan beam DXA is a highly accurate and precise technique for measuring BMC and BMD in excised small animal bones. A physiological saline concentration in the water bath had a significant impact on BMC and area, but not on BMD, and should therefore be strictly controlled to avoid an underestimation of BMC. (Bone 30:243–246; 2002) © 2002 by Elsevier Science Inc. All rights reserved.
Materials and Methods Specimens and DXA Scans In this study we used excised humeri of 11-week-old healthy male Lewis rats. The rats had served as a control group in another study, which was approved by the regional animal ethics committee (Regierung von Mittelfranken, Az. 621-2531.3-10/95). The bones were freed of all soft tissue and then frozen. Before the DXA examination they were first defrosted for 30 min and then weighed. DXA scans were performed with a fan beam QDR 4500 A densitometer (Hologic, Inc., Bedford, MA) calibrated daily in accordance with the manufacturer’s recommendations. The regional high-resolution mode of the small animal scan protocol (scan field 5.0 [width] ⫻ 6.0 [height] cm2, pixel size 0.31 mm2, scan time 2 min) was used. The specimens were positioned centrally at the bottom of a square, thin-walled plastic container using the manufacturer-recommended water bath at a level of 2.5 cm. The outline of each humerus was marked on the container bottom to assure a high positioning reproducibility during all measurements. The deltoid tuberosity was faced upward to avoid an irregular projectional shape. BMC and BMD of the entire humeri were analyzed (Figure 1). All DXA measurements and analyses were performed by the same investigator (S. K.). After manually positioning the bone and the laser beam that marks the start of the scan the measurements were performed in a fully automated manner. Following DXA the humeri were first dried at room temperature and then at 100°C, and were finally incinerated at 800°C. The three steps lasted 24 h each. The ash was weighed and dissolved in 6 mol/L HCl and the calcium content was measured
Key Words: Fan beam dual-energy X-ray absorptiometry (DXA); Rat; Ash weight; Accuracy; Precision; Bone mineral density (BMD). Introduction The determination of bone mineral density (BMD) and content (BMC) in rats is helpful in experimental surgery to examine the influence of different operation techniques on calcium metabolism. For example, skeletal effects of small bowel resection, ovariectomy, orchiectomy, and organ transplantation can be described and quantified precisely. For the assessment of osteAddress for correspondence and reprints: Dr. Sigrid Kastl, Department of Surgery, University of Erlangen, Krankenhausstrasse 12, 91054 Erlangen, Germany. E-mail: [email protected] © 2002 by Elsevier Science Inc. All rights reserved.
243
8756-3282/02/$22.00 PII S8756-3282(01)00641-X
244
S. Kastl et al. DXA accuracy and precision in rat humeri
Bone Vol. 30, No. 1 January 2002:243–246
Figure 2. Bland and Altman plot BMC and ash weight from Table 1. Mean and ⫾ 2 SD (0.0137 ⫾ 0.0094 g) of the difference between BMC and ash weight are shown as dotted horizontal lines.
Figure 1. Typical DXA scan of an excised rat humerus examined with a Hologic QDR 4500 A. The bone outline is determined automatically without user interaction. The R1 subregion, also determined automatically, was not used in this study.
linear regression as well as Bland–Altman plots.4 The values measured in water and in 0.9% saline solution were compared by within-subject analysis of variance (ANOVA) using the effect of salinity and of the six repetitions of the measurements as independent factors. All statistical analyses were performed with SPSS software (v 10.0). Results
by complexiometry (Model 940 Calcium Analyzer, Corning, Halstead, UK). Accuracy and Precision DXA accuracy was evaluated by comparing BMC of six rat humeri with dry weight, ash weight, and calcium content of the incinerated bone. DXA measurements were repeated six times each and averaged. Precision was determined with and without repositioning by measuring six different bones six times each. Salinity of the Solution Used for DXA Measurements Six separate humeri were measured six times each without repositioning in water and in a physiological saline solution of 0.9% NaCl. Statistical Methods Results are given as mean, standard deviation (SD), and coefficient of variation (c.v.). Mean c.v.s and SDs were calculated as root-mean-square (RMS) values.2 Confidence limits have also been calculated for precision results elsewhere.2 The differences between DXA BMC and ash weight are given in percent of ash weight (⌬ mineral ⫽ [BMC ⫺ ash weight]/ash weight ⴱ 100). The accuracy analysis also included a calculation of correlation coefficients and the standard error of the estimate (SEE) of a
Accuracy results are given in Table 1. Figure 2 shows the corresponding Bland–Altman plot. The mean difference between BMC and ash weight was 0.0138 g (95% Cl 0.0091– 0.0185 g). Correlation coefficients (r2) between BMC and dry weight, ash weight, and calcium content were 0.978, 0.988, and 0.890, respectively. Performing a linear regression of BMC against the same parameters resulted in standard errors of the estimates (SEEs) of 0.0036 g, 0.0010 g, and 0.0074 g. SEEs for ash and dry weight regressions were comparable with the standard deviations of the repeated BMC measurements (see Table 1); the SEE of 0.0074 g for calcium content was two to three times higher. DXA precision results are shown in Table 2. Without repositioning, the mean coefficient of variation (95% CI interval) was 0.76% (0.62–1.01) for area, 0.90% (0.74 –1.19) for BMC, and 0.76% (0.63–1.01) for BMD. With repositioning, we obtained 2.77% (2.28 –3.67) for area, 1.32% (1.08 –1.74) for BMC, and 0.86% (0.71–1.14) for BMD. In the 0.9% saline solution, DXA BMC (F ⫽ 42.58) and DXA area (F ⫽ 33.05), but not DXA BMD (F ⫽ 0.00051), were significantly lower (p ⬍ 0.005) than in tap water. The critical F value (␣ ⫽ 0.05) was 6.61. The second factor included in the ANOVA (six repeated measurements) did not explain the difference in the data (F ⬍ 0.7 for DXA BMC and DXA area and F ⫽ 3.4 for DXA BMD; p ⬎ 0.3). DXA-BMC decreased by 4.1% and DXA area by 4.4%. DXA-BMD did not change significantly (see Table 3).
Table 1. Accuracy results of bone mineral content (BMC) compared with ash weight as determined by dual-energy X-ray absorptiometry (DXA) Humerus
BMC (g)
Dry weight (g)
Ash weight (g)
⌬ Mineral (%)
Ca (mmol/L)
1 2 3 4 5 6
0.249 ⫾ 0.0017 0.211 ⫾ 0.0034 0.236 ⫾ 0.0021 0.221 ⫾ 0.0034 0.265 ⫾ 0.0028 0.223 ⫾ 0.0039
0.328 0.290 0.320 0.306 0.353 0.307
0.229 0.201 0.223 0.211 0.246 0.212
8.7 5.0 5.8 4.7 7.7 5.2
235 201 232 203 243 204
BMC values expressed as mean ⫾ standard deviation. ⌬ Mineral (%) indicates the difference between BMC and ash weight.
Bone Vol. 30, No. 1 January 2002:243–246
S. Kastl et al. DXA accuracy and precision in rat humeri
245
Table 2. Precision results for area, bone mineral content (BMC), and bone mineral density BMD Without repositioning 2
With repositioning 2
2
Humerus
Area (cm )
BMC (g)
BMD (g/cm )
Area (cm )
BMC (g)
BMD (g/cm2)
1 2 3 4 5 6
1.235 ⫾ 0.0139 1.073 ⫾ 0.0076 1.174 ⫾ 0.0077 1.142 ⫾ 0.0046 1.294 ⫾ 0.0116 1.153 ⫾ 0.0065
0.247 ⫾ 0.0019 0.209 ⫾ 0.0018 0.234 ⫾ 0.0020 0.217 ⫾ 0.0019 0.267 ⫾ 0.0024 0.222 ⫾ 0.0024
0.200 ⫾ 0.0020 0.194 ⫾ 0.0015 0.199 ⫾ 0.0012 0.190 ⫾ 0.0018 0.206 ⫾ 0.0009 0.193 ⫾ 0.0014
1.234 ⫾ 0.0114 1.140 ⫾ 0.0701 1.198 ⫾ 0.0095 1.164 ⫾ 0.0194 1.284 ⫾ 0.0134 1.168 ⫾ 0.0208
0.249 ⫾ 0.0017 0.211 ⫾ 0.0034 0.236 ⫾ 0.0021 0.221 ⫾ 0.0034 0.265 ⫾ 0.0028 0.223 ⫾ 0.0039
0.202 ⫾ 0.0008 0.189 ⫾ 0.0014 0.197 ⫾ 0.0007 0.190 ⫾ 0.0025 0.206 ⫾ 0.0027 0.193 ⫾ 0.0009
Values expressed as mean ⫾ standard deviation. Each humerus was measured six times. Table 3. Effect of salinity on area, bone mineral content (BMC), and bone mineral density (BMD) Tap water 2
0.9% saline solution 2
2
Humerus
Area (cm )
BMC (g)
BMD (g/cm )
Area (cm )
BMC (g)
BMD (g/cm2)
1 2 3 4 5 6
1.142 ⫾ 0.0142 1.132 ⫾ 0.0100 1.233 ⫾ 0.0084 1.224 ⫾ 0.0107 1.084 ⫾ 0.0078 1.110 ⫾ 0.0081
0.218 ⫾ 0.0036 0.219 ⫾ 0.0025 0.236 ⫾ 0.0017 0.239 ⫾ 0.0016 0.193 ⫾ 0.0017 0.199 ⫾ 0.0025
0.191 ⫾ 0.0021 0.193 ⫾ 0.0009 0.191 ⫾ 0.0026 0.195 ⫾ 0.0025 0.178 ⫾ 0.0014 0.179 ⫾ 0.0019
1.114 ⫾ 0.0096 1.092 ⫾ 0.0156 1.149 ⫾ 0.0271 1.177 ⫾ 0.0099 1.053 ⫾ 0.0105 1.053 ⫾ 0.0126
0.212 ⫾ 0.0026 0.206 ⫾ 0.0030 0.224 ⫾ 0.0073 0.226 ⫾ 0.0085 0.189 ⫾ 0.0017 0.189 ⫾ 0.0018
0.190 ⫾ 0.0010 0.189 ⫾ 0.0011 0.196 ⫾ 0.0020 0.195 ⫾ 0.0018 0.180 ⫾ 0.0009 0.180 ⫾ 0.0019
Values expressed as mean ⫾ standard deviation. Each humerus measured six times in tap water and in 0.9% saline solution.
Discussion We determined DXA accuracy and precision in excised rat humeri using a fan beam scanner. Similar to other studies, the accuracy results showed excellent correlation with ash weight (r ⫽ 0.986) and, as expected, slightly lower correlations with dry weight and calcium content.11 Like other investigators, the observed overestimation of BMC by DXA was ⬍10%.9 The Bland–Altman diagram shows an offset between BMC measured by DXA and by ash of 0.137 g. The 95% confidence interval (0.0091– 0.0185 g), was on the order of 1 SD, which is smaller than 5% of the mean of DXA BMC and ash weight. According to Figure, 2 differences tended to be higher for higher BMC values, indicating a relative discrepancy rather than a constant offset between DXA BMC and ash weight. This is also reflected by the 1 SD values (⫾0.0047 g) of the Bland–Altman plot, which were fourfold higher than the SEE of the regression of BMC vs. ash weight. However, our samples are too homogeneous to draw final conclusions. Taken together, our data show a high comparability of the two methods. Furthermore, we observed few short-term precision errors. Errors without repositioning were ⬍1%. With repositioning, the precision error for area increased but the effect on BMD was modest as effects of BMC and area canceled out. Our values are in general agreement with those reported in the literature in vivo using intact animals.1,6,7,13 A BMD precision of ⬍1%, even with repositioning, is an excellent result in these very small bones. In the background of Figure 1 window and level were adjusted to visualize a vertical pattern, indicating that sensitivity differences in the individual detectors were not fully eliminated. This pattern was constant in the images analyzed—thus precision errors may have been larger if, compared with our experiment, the specimen was less carefully repositioned. A limitation of our study is the relative homogeneity of the bones examined: BMC varied by approximately only 25% (we examined inbred male sibling rats). Larger variations may have caused slightly larger errors as reported in other publications.7 Also, we did not investigate intact animals. Furthermore, we did not vary the depth of the water bath but rather
used the recommended value of 2.5 cm only. We did, however, change the salinity. Compared with tap water, DXA BMC and DXA area decreased significantly by roughly 4% each when the 0.9% physiological saline concentration was used. There was no effect on DXA BMD. In the regional high-resolution scan mode the threshold settings used to detect the bone edges were probably optimized for water, the medium recommended by the manufacturer. Thus, bone edges may vary if the bone-soft tissue contrast is altered by a compositional change of this medium. It has been reported that older scanners are also sensitive to the height of the water bath (kept constant in this study).5 DXA theory assumes that soft tissue composition in the path containing no bone and that containing bone are identical. However, this is not the case when using tap water as the marrow composition inside the bone is probably closer to the 0.9% salinity level. We conclude that DXA is a simple, accurate, and precise technique for measuring BMC and BMD in isolated small animal bones, and is therefore an excellent tool for assessing osteopathies. However, only water should be used for determination of BMC and BMD by DXA.
Acknowledgments: This work was supported by a grant from the Erlangen University Research Fund “ELAN.”
References 1. Casez, J. P., Muehlbauer, R. C., Lippuner, K., Kelly, T., Fleisch, H., and Jaeger, P. Dual-energy X-ray absorptiometry for measuring total bone mineral content in the rat: Study of accuracy and precision. Bone Miner 26(1):61– 68; 1994. 2. Glu¨ er, C.-C., Blake, G., Lu, Y., Blunt, B. A., Jergas, M., and Genant, H. K. Accurate assessment of precision errors: How to measure the reproducibility of bone densitometry techniques. Osteopor Int 5:262–270; 1995. 3. Griffin, M. G., Kimble, R., Hopfer, W., and Pacifici, R. Dual-energy X-ray absorptiometry of the rat: Accuracy, precision and measurement of bone loss. J Bone Miner Res 8(7):795– 800; 1993. 4. Bland, J. M. and Altman, D. G. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet i307–310; 1986
246
S. Kastl et al. DXA accuracy and precision in rat humeri
5. Hagiwara, S., Lane, N., Engelke, K., Sebastian, A., Kimmel, D. B., and Genant, H. K. Precision and accuracy for rat whole body and femur bone mineral determination with dual X-ray absorptiometry. Bone Miner 22:57– 68; 1993. 6. Lu, P. W., Briody, J. N., Howman-Giles, R., Trube, A., and Cowell, C. T. DXA for bone density measurement in small rats weighing 150 –250 grams. Bone 15(2):199 –202; 1994. 7. Makan, S., Bayley, H. S., and Webber, C. E. Precision and accuracy of total body bone mass and body composition measurements in the rat using x-raybased dual photon absorptiometry. Can J Physiol Pharmacol 75(10):1257– 1261; 1997. 8. Mitlak, B. H., Schoenfeld, D., and Neer, R. M. Accuracy, precision, and utility of spine and whole-skeleton mineral measurements by DXA in rats. J Bone Miner Res 9(1):119 –126; 1994. 9. Pastoureau, P., Chomel, A., and Bonnet, J. Specific evaluation of localized bone mass and bone loss in the rat using dual-energy X-ray absorptiometry subregional analysis. Osteopor Int 5(3):143–149; 1995. 10. Rosen, H. N., Tollin, S., Balena, R., Middlebrooks, V. L., Beamer, W. G., Donohue, L. R., Rosen, C., Turner, A., Holick, M., and Greenspan, S. L. Differentiating between orchiectomized rats and controls using measurements
Bone Vol. 30, No. 1 January 2002:243–246 of trabecular bone density: A comparison among DXA, histomorphometry, and peripheral quantitative computerized tomography. Calcif Tissue Int 57(1):35– 39; 1995. 11. Rozenberg, S., Peretz, A., Praet, J. P., Vandromme, J., and Ham, H. Bone mineral density measured by dual photon and dual energy X-ray absorptiometry: A problem of accuracy. Nucl Med Commun 14(3):189 –191; 1993. 12. Rozenberg, S., Vandromme, J., Neve, J., Aguilera, A., Muregancuro, A., Peretz, A., Kinthaert, J, and Ham, H. Precision and accuracy of in vivo bone mineral measurement in rats using dual-energy X-ray absorptiometry. Osteopor Int 5(1):47–53; 1995. 13. Yamauchi, H., Kushida, K., Yamazaki, K., and Inoue, T. Assessment of spine bone mineral density in ovariectomized rats using DXA. J Bone Miner Res 10(7):1033–1039; 1995.
Date Received: January 4, 2001 Date Revised: July 20, 2001 Date Accepted: July 20, 2001
Accuracy and Precision of Bone Mineral Density and Bone Mineral Content in Excised Rat Humeri Using Fan Beam Dual-energy X-ray Absorptiometry S. KASTL, T. SOMMER, P. KLEIN, W. HOHENBERGER, and K. ENGELKE Department of Surgery and Institute of Medical Physics, University of Erlangen, Erlangen, Germany
opathies and fracture risk in small laboratory animals a number of different osteodensitometric methods, such as peripheral quantitative computed tomography (pQCT)10,11 and dual-energy X-ray absorptiometry (DXA) have been used. BMC and BMD precision and accuracy measurements in rats have been reported by some investigators1,3,5,8,9,12; however, to our knowledge, an in vitro analysis of excised rat bones scanned with a fan beam DXA scanner has not yet been published. In this study we compare ash weight and ash mineral content with DXA BMC results of excised rat humeri and determine the short-term reproducibility of BMD and BMC. We also investigate the effect of the salinity of the water bath in which the samples were scanned. Excised bones were typically scanned in tap water, whereas, in in vivo scans, a physiological concentration of approximately 0.9% saline was present.
The aim of this study was the evaluation of fan beam dualenergy X-ray absorptiometry (DXA) for measuring bone mineral density (BMD) and bone mineral content (BMC) of isolated rat humeri. Defleshed rat humeri from male Lewis rats were examined with a Hologic QDR 4500 A (Hologic, Inc., Bedford, MA) high-resolution densitometer both in water and 0.9% saline solution. The small animal scan protocol with the regional high-resolution mode was used. BMC measured by DXA was compared with bone dry weight, ash weight, and bone calcium content. Furthermore, DXA BMD and BMC precision were determined. We also evaluated the effect of salinity of the water bath in which the bones were measured. Correlations (r2) of BMC, as determined by DXA with dry weight, ash weight, and bone calcium content, were 0.978, 0.988, and 0.890, respectively. DXA overestimated ash weight by 5%–9%. Precision errors for BMC (BMD) were 0.90% (0.76%) without and 1.3 (0.86) with repositioning. Changes in the salinity of the water bath had a significant influence on the DXA results: At the 0.9% physiological level, BMC (ⴚ4.4%) and area (ⴚ4.1%), but not BMD, values were significantly lower (p < 0.005) compared with measurements in tap water. Fan beam DXA is a highly accurate and precise technique for measuring BMC and BMD in excised small animal bones. A physiological saline concentration in the water bath had a significant impact on BMC and area, but not on BMD, and should therefore be strictly controlled to avoid an underestimation of BMC. (Bone 30:243–246; 2002) © 2002 by Elsevier Science Inc. All rights reserved.
Materials and Methods Specimens and DXA Scans In this study we used excised humeri of 11-week-old healthy male Lewis rats. The rats had served as a control group in another study, which was approved by the regional animal ethics committee (Regierung von Mittelfranken, Az. 621-2531.3-10/95). The bones were freed of all soft tissue and then frozen. Before the DXA examination they were first defrosted for 30 min and then weighed. DXA scans were performed with a fan beam QDR 4500 A densitometer (Hologic, Inc., Bedford, MA) calibrated daily in accordance with the manufacturer’s recommendations. The regional high-resolution mode of the small animal scan protocol (scan field 5.0 [width] ⫻ 6.0 [height] cm2, pixel size 0.31 mm2, scan time 2 min) was used. The specimens were positioned centrally at the bottom of a square, thin-walled plastic container using the manufacturer-recommended water bath at a level of 2.5 cm. The outline of each humerus was marked on the container bottom to assure a high positioning reproducibility during all measurements. The deltoid tuberosity was faced upward to avoid an irregular projectional shape. BMC and BMD of the entire humeri were analyzed (Figure 1). All DXA measurements and analyses were performed by the same investigator (S. K.). After manually positioning the bone and the laser beam that marks the start of the scan the measurements were performed in a fully automated manner. Following DXA the humeri were first dried at room temperature and then at 100°C, and were finally incinerated at 800°C. The three steps lasted 24 h each. The ash was weighed and dissolved in 6 mol/L HCl and the calcium content was measured
Key Words: Fan beam dual-energy X-ray absorptiometry (DXA); Rat; Ash weight; Accuracy; Precision; Bone mineral density (BMD). Introduction The determination of bone mineral density (BMD) and content (BMC) in rats is helpful in experimental surgery to examine the influence of different operation techniques on calcium metabolism. For example, skeletal effects of small bowel resection, ovariectomy, orchiectomy, and organ transplantation can be described and quantified precisely. For the assessment of osteAddress for correspondence and reprints: Dr. Sigrid Kastl, Department of Surgery, University of Erlangen, Krankenhausstrasse 12, 91054 Erlangen, Germany. E-mail: [email protected] © 2002 by Elsevier Science Inc. All rights reserved.
243
8756-3282/02/$22.00 PII S8756-3282(01)00641-X
244
S. Kastl et al. DXA accuracy and precision in rat humeri
Bone Vol. 30, No. 1 January 2002:243–246
Figure 2. Bland and Altman plot BMC and ash weight from Table 1. Mean and ⫾ 2 SD (0.0137 ⫾ 0.0094 g) of the difference between BMC and ash weight are shown as dotted horizontal lines.
Figure 1. Typical DXA scan of an excised rat humerus examined with a Hologic QDR 4500 A. The bone outline is determined automatically without user interaction. The R1 subregion, also determined automatically, was not used in this study.
linear regression as well as Bland–Altman plots.4 The values measured in water and in 0.9% saline solution were compared by within-subject analysis of variance (ANOVA) using the effect of salinity and of the six repetitions of the measurements as independent factors. All statistical analyses were performed with SPSS software (v 10.0). Results
by complexiometry (Model 940 Calcium Analyzer, Corning, Halstead, UK). Accuracy and Precision DXA accuracy was evaluated by comparing BMC of six rat humeri with dry weight, ash weight, and calcium content of the incinerated bone. DXA measurements were repeated six times each and averaged. Precision was determined with and without repositioning by measuring six different bones six times each. Salinity of the Solution Used for DXA Measurements Six separate humeri were measured six times each without repositioning in water and in a physiological saline solution of 0.9% NaCl. Statistical Methods Results are given as mean, standard deviation (SD), and coefficient of variation (c.v.). Mean c.v.s and SDs were calculated as root-mean-square (RMS) values.2 Confidence limits have also been calculated for precision results elsewhere.2 The differences between DXA BMC and ash weight are given in percent of ash weight (⌬ mineral ⫽ [BMC ⫺ ash weight]/ash weight ⴱ 100). The accuracy analysis also included a calculation of correlation coefficients and the standard error of the estimate (SEE) of a
Accuracy results are given in Table 1. Figure 2 shows the corresponding Bland–Altman plot. The mean difference between BMC and ash weight was 0.0138 g (95% Cl 0.0091– 0.0185 g). Correlation coefficients (r2) between BMC and dry weight, ash weight, and calcium content were 0.978, 0.988, and 0.890, respectively. Performing a linear regression of BMC against the same parameters resulted in standard errors of the estimates (SEEs) of 0.0036 g, 0.0010 g, and 0.0074 g. SEEs for ash and dry weight regressions were comparable with the standard deviations of the repeated BMC measurements (see Table 1); the SEE of 0.0074 g for calcium content was two to three times higher. DXA precision results are shown in Table 2. Without repositioning, the mean coefficient of variation (95% CI interval) was 0.76% (0.62–1.01) for area, 0.90% (0.74 –1.19) for BMC, and 0.76% (0.63–1.01) for BMD. With repositioning, we obtained 2.77% (2.28 –3.67) for area, 1.32% (1.08 –1.74) for BMC, and 0.86% (0.71–1.14) for BMD. In the 0.9% saline solution, DXA BMC (F ⫽ 42.58) and DXA area (F ⫽ 33.05), but not DXA BMD (F ⫽ 0.00051), were significantly lower (p ⬍ 0.005) than in tap water. The critical F value (␣ ⫽ 0.05) was 6.61. The second factor included in the ANOVA (six repeated measurements) did not explain the difference in the data (F ⬍ 0.7 for DXA BMC and DXA area and F ⫽ 3.4 for DXA BMD; p ⬎ 0.3). DXA-BMC decreased by 4.1% and DXA area by 4.4%. DXA-BMD did not change significantly (see Table 3).
Table 1. Accuracy results of bone mineral content (BMC) compared with ash weight as determined by dual-energy X-ray absorptiometry (DXA) Humerus
BMC (g)
Dry weight (g)
Ash weight (g)
⌬ Mineral (%)
Ca (mmol/L)
1 2 3 4 5 6
0.249 ⫾ 0.0017 0.211 ⫾ 0.0034 0.236 ⫾ 0.0021 0.221 ⫾ 0.0034 0.265 ⫾ 0.0028 0.223 ⫾ 0.0039
0.328 0.290 0.320 0.306 0.353 0.307
0.229 0.201 0.223 0.211 0.246 0.212
8.7 5.0 5.8 4.7 7.7 5.2
235 201 232 203 243 204
BMC values expressed as mean ⫾ standard deviation. ⌬ Mineral (%) indicates the difference between BMC and ash weight.
Bone Vol. 30, No. 1 January 2002:243–246
S. Kastl et al. DXA accuracy and precision in rat humeri
245
Table 2. Precision results for area, bone mineral content (BMC), and bone mineral density BMD Without repositioning 2
With repositioning 2
2
Humerus
Area (cm )
BMC (g)
BMD (g/cm )
Area (cm )
BMC (g)
BMD (g/cm2)
1 2 3 4 5 6
1.235 ⫾ 0.0139 1.073 ⫾ 0.0076 1.174 ⫾ 0.0077 1.142 ⫾ 0.0046 1.294 ⫾ 0.0116 1.153 ⫾ 0.0065
0.247 ⫾ 0.0019 0.209 ⫾ 0.0018 0.234 ⫾ 0.0020 0.217 ⫾ 0.0019 0.267 ⫾ 0.0024 0.222 ⫾ 0.0024
0.200 ⫾ 0.0020 0.194 ⫾ 0.0015 0.199 ⫾ 0.0012 0.190 ⫾ 0.0018 0.206 ⫾ 0.0009 0.193 ⫾ 0.0014
1.234 ⫾ 0.0114 1.140 ⫾ 0.0701 1.198 ⫾ 0.0095 1.164 ⫾ 0.0194 1.284 ⫾ 0.0134 1.168 ⫾ 0.0208
0.249 ⫾ 0.0017 0.211 ⫾ 0.0034 0.236 ⫾ 0.0021 0.221 ⫾ 0.0034 0.265 ⫾ 0.0028 0.223 ⫾ 0.0039
0.202 ⫾ 0.0008 0.189 ⫾ 0.0014 0.197 ⫾ 0.0007 0.190 ⫾ 0.0025 0.206 ⫾ 0.0027 0.193 ⫾ 0.0009
Values expressed as mean ⫾ standard deviation. Each humerus was measured six times. Table 3. Effect of salinity on area, bone mineral content (BMC), and bone mineral density (BMD) Tap water 2
0.9% saline solution 2
2
Humerus
Area (cm )
BMC (g)
BMD (g/cm )
Area (cm )
BMC (g)
BMD (g/cm2)
1 2 3 4 5 6
1.142 ⫾ 0.0142 1.132 ⫾ 0.0100 1.233 ⫾ 0.0084 1.224 ⫾ 0.0107 1.084 ⫾ 0.0078 1.110 ⫾ 0.0081
0.218 ⫾ 0.0036 0.219 ⫾ 0.0025 0.236 ⫾ 0.0017 0.239 ⫾ 0.0016 0.193 ⫾ 0.0017 0.199 ⫾ 0.0025
0.191 ⫾ 0.0021 0.193 ⫾ 0.0009 0.191 ⫾ 0.0026 0.195 ⫾ 0.0025 0.178 ⫾ 0.0014 0.179 ⫾ 0.0019
1.114 ⫾ 0.0096 1.092 ⫾ 0.0156 1.149 ⫾ 0.0271 1.177 ⫾ 0.0099 1.053 ⫾ 0.0105 1.053 ⫾ 0.0126
0.212 ⫾ 0.0026 0.206 ⫾ 0.0030 0.224 ⫾ 0.0073 0.226 ⫾ 0.0085 0.189 ⫾ 0.0017 0.189 ⫾ 0.0018
0.190 ⫾ 0.0010 0.189 ⫾ 0.0011 0.196 ⫾ 0.0020 0.195 ⫾ 0.0018 0.180 ⫾ 0.0009 0.180 ⫾ 0.0019
Values expressed as mean ⫾ standard deviation. Each humerus measured six times in tap water and in 0.9% saline solution.
Discussion We determined DXA accuracy and precision in excised rat humeri using a fan beam scanner. Similar to other studies, the accuracy results showed excellent correlation with ash weight (r ⫽ 0.986) and, as expected, slightly lower correlations with dry weight and calcium content.11 Like other investigators, the observed overestimation of BMC by DXA was ⬍10%.9 The Bland–Altman diagram shows an offset between BMC measured by DXA and by ash of 0.137 g. The 95% confidence interval (0.0091– 0.0185 g), was on the order of 1 SD, which is smaller than 5% of the mean of DXA BMC and ash weight. According to Figure, 2 differences tended to be higher for higher BMC values, indicating a relative discrepancy rather than a constant offset between DXA BMC and ash weight. This is also reflected by the 1 SD values (⫾0.0047 g) of the Bland–Altman plot, which were fourfold higher than the SEE of the regression of BMC vs. ash weight. However, our samples are too homogeneous to draw final conclusions. Taken together, our data show a high comparability of the two methods. Furthermore, we observed few short-term precision errors. Errors without repositioning were ⬍1%. With repositioning, the precision error for area increased but the effect on BMD was modest as effects of BMC and area canceled out. Our values are in general agreement with those reported in the literature in vivo using intact animals.1,6,7,13 A BMD precision of ⬍1%, even with repositioning, is an excellent result in these very small bones. In the background of Figure 1 window and level were adjusted to visualize a vertical pattern, indicating that sensitivity differences in the individual detectors were not fully eliminated. This pattern was constant in the images analyzed—thus precision errors may have been larger if, compared with our experiment, the specimen was less carefully repositioned. A limitation of our study is the relative homogeneity of the bones examined: BMC varied by approximately only 25% (we examined inbred male sibling rats). Larger variations may have caused slightly larger errors as reported in other publications.7 Also, we did not investigate intact animals. Furthermore, we did not vary the depth of the water bath but rather
used the recommended value of 2.5 cm only. We did, however, change the salinity. Compared with tap water, DXA BMC and DXA area decreased significantly by roughly 4% each when the 0.9% physiological saline concentration was used. There was no effect on DXA BMD. In the regional high-resolution scan mode the threshold settings used to detect the bone edges were probably optimized for water, the medium recommended by the manufacturer. Thus, bone edges may vary if the bone-soft tissue contrast is altered by a compositional change of this medium. It has been reported that older scanners are also sensitive to the height of the water bath (kept constant in this study).5 DXA theory assumes that soft tissue composition in the path containing no bone and that containing bone are identical. However, this is not the case when using tap water as the marrow composition inside the bone is probably closer to the 0.9% salinity level. We conclude that DXA is a simple, accurate, and precise technique for measuring BMC and BMD in isolated small animal bones, and is therefore an excellent tool for assessing osteopathies. However, only water should be used for determination of BMC and BMD by DXA.
Acknowledgments: This work was supported by a grant from the Erlangen University Research Fund “ELAN.”
References 1. Casez, J. P., Muehlbauer, R. C., Lippuner, K., Kelly, T., Fleisch, H., and Jaeger, P. Dual-energy X-ray absorptiometry for measuring total bone mineral content in the rat: Study of accuracy and precision. Bone Miner 26(1):61– 68; 1994. 2. Glu¨ er, C.-C., Blake, G., Lu, Y., Blunt, B. A., Jergas, M., and Genant, H. K. Accurate assessment of precision errors: How to measure the reproducibility of bone densitometry techniques. Osteopor Int 5:262–270; 1995. 3. Griffin, M. G., Kimble, R., Hopfer, W., and Pacifici, R. Dual-energy X-ray absorptiometry of the rat: Accuracy, precision and measurement of bone loss. J Bone Miner Res 8(7):795– 800; 1993. 4. Bland, J. M. and Altman, D. G. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet i307–310; 1986
246
S. Kastl et al. DXA accuracy and precision in rat humeri
5. Hagiwara, S., Lane, N., Engelke, K., Sebastian, A., Kimmel, D. B., and Genant, H. K. Precision and accuracy for rat whole body and femur bone mineral determination with dual X-ray absorptiometry. Bone Miner 22:57– 68; 1993. 6. Lu, P. W., Briody, J. N., Howman-Giles, R., Trube, A., and Cowell, C. T. DXA for bone density measurement in small rats weighing 150 –250 grams. Bone 15(2):199 –202; 1994. 7. Makan, S., Bayley, H. S., and Webber, C. E. Precision and accuracy of total body bone mass and body composition measurements in the rat using x-raybased dual photon absorptiometry. Can J Physiol Pharmacol 75(10):1257– 1261; 1997. 8. Mitlak, B. H., Schoenfeld, D., and Neer, R. M. Accuracy, precision, and utility of spine and whole-skeleton mineral measurements by DXA in rats. J Bone Miner Res 9(1):119 –126; 1994. 9. Pastoureau, P., Chomel, A., and Bonnet, J. Specific evaluation of localized bone mass and bone loss in the rat using dual-energy X-ray absorptiometry subregional analysis. Osteopor Int 5(3):143–149; 1995. 10. Rosen, H. N., Tollin, S., Balena, R., Middlebrooks, V. L., Beamer, W. G., Donohue, L. R., Rosen, C., Turner, A., Holick, M., and Greenspan, S. L. Differentiating between orchiectomized rats and controls using measurements
Bone Vol. 30, No. 1 January 2002:243–246 of trabecular bone density: A comparison among DXA, histomorphometry, and peripheral quantitative computerized tomography. Calcif Tissue Int 57(1):35– 39; 1995. 11. Rozenberg, S., Peretz, A., Praet, J. P., Vandromme, J., and Ham, H. Bone mineral density measured by dual photon and dual energy X-ray absorptiometry: A problem of accuracy. Nucl Med Commun 14(3):189 –191; 1993. 12. Rozenberg, S., Vandromme, J., Neve, J., Aguilera, A., Muregancuro, A., Peretz, A., Kinthaert, J, and Ham, H. Precision and accuracy of in vivo bone mineral measurement in rats using dual-energy X-ray absorptiometry. Osteopor Int 5(1):47–53; 1995. 13. Yamauchi, H., Kushida, K., Yamazaki, K., and Inoue, T. Assessment of spine bone mineral density in ovariectomized rats using DXA. J Bone Miner Res 10(7):1033–1039; 1995.
Date Received: January 4, 2001 Date Revised: July 20, 2001 Date Accepted: July 20, 2001