- Email: [email protected]
Dual-energy X-ray absorptiometry for measuring total bone mineral content in the rat: Study of accuracy and precision
ELSEVIER
Bone and Mineral 26 (1994) 61-68
Dual-energy X-ray absorptiometry for measuring total bone mineral content in the rat: Study of accuracy and precision J-P. Casez a, R.C. Muehlbauer b, K. Lippuner a, T. Kelly c, H. Fleisch b, Ph. Jaeger *a apoliclinie of Medicine, bDepartment o[" Pathophysiology. University ~["Berne. Berne. Swit:erland CHologic lnt'.. Watthem~. MA. USA (Received 19 August 1993; revision received 4 January 1994: accepted 10 February 1994)
Abstract
Sequential studies of osteopenic bone disease in small animals require the availability of non-invasive, accurate and precise methods to assess bone mineral content (BMCJ and bone mineral density (BMD). Dual-energy X-ray absorptiometry (DXA), which is currently used in humans for this purpose, can also be applied to small animals by means of adapted software. Precision and accuracy of DXA was evaluated in 10 rats weighing 50-265 g. The rats were anesthetized with a mixture of ketamine-xylazine administrated intraperitoneally. Each rat was scanned six times consecutively in the antero-posterior incidence after repositioning using the rat whole-body software for determination of whole-body BMC and BMD (Hologic QDR 1000| software version 5.52). Scan duration was 10-20 min depending on rat size. After the last measurement, rats were sacrificed and soft tissues were removed by dermestid beetles. Skeletons were then scanned in vitro (ultra high resolution software, version 4.47). Bones were subsequently ashed and dissolved in hydrochloric acid and total body calcium directly assayed by atomic absorption spectrophotometry (TBCa[chem]). Total body calcium was also calculated from the DXA whole-body in vivo measurement (TBCa[DXA]) and from the ultra high resolution measurement (TBCa[UH]) under the assumption that calcium accounts for 40.5"/,, of the BMC expressed as hydroxyapatite. Precision error for whole-body BMC and BMD (mean • S.D.) was 1.3% and 1.5"/,,, respectively. Simple regression analysis between TBCa[DXA] or TBCa[UH] and TBCa[chem] revealed tight correlations (n = 0.991 and 0.996, respectively), with slopes and intercepts which were significantly different from 1 and 0, respectively. Mean accuracy error for TBCa[DXA] was 8.5% for all rats and 5.7% after * Corresponding author, Director of the Policlinic of Medicine, lnselspital, CH-3010, Switzerland. 0169-6009/94/$07.00 9 1994 Elsevier Science Ireland Ltd. All rights reserved SSDI 0169-6009( 94)00694-C
62
J.-P. Casez et al./ Bone Miner. 26 (1994) 61-68
excluding the 50-g rats. It was 6.4'V,,for TBCa[UH] (all rats). We conclude that whole-body DXA provides highly precise and reasonably accurate estimates of skeletal calcium content in the rat in vivo. Keywords: Rat; Skeletal calcium content; DXA, Whole body
1. Introduction
Over the last years it has become progressively apparent that development of noninvasive techniques to assess bone density in small animals in vivo was of paramount importance to further progress in the therapeutic approach to most forms of osteopenic bone disease. In particular accurate and precise sequential bone mass measurements are needed which provide information on different parts of the skeleton. Early methods for bone mass measurements in the rat were based on histomorphometry (which requires animal sacrifice) or on radiography. Bone assessment on X-ray films is possible using radiogrammetry [1] or relative optical densitometry [21. However, the level of precision reached with the former is poor and the latter can be performed only on caudal vertebrae, i.e. bones that may not be representative of the whole skeleton. Single photon absorptiometry has been used to evaluate bone mineral content (BMC) of the femur [3] or at lumbar spine [4] in the rat in vivo. Reproducibility was 3-6.6'V,,, which is not sufficient to detect small changes without a large sample size. Quantitative CT of the lumbar spine [41 and dual photon absorptiometry [5] have also been employed in vivo and in vitro. However, precision errors of 3 to 5'88 have adversely affected their ability to detect changes in skeletal status. Recently dual-energy X-ray absorptiometry (DXA) has become the method of choice to address osteopenic bone disease in humans due to high reproducibility, short investigation time, lower cost and low radiation exposure. DXA can also be applied to small animals. Amman et al. reported I.Y88 precision in measuring femoral or vertebral bone mineral density in the rat in vivo [6]. Neutron activation analysis also has been used to assess total skeletal calcium in the rat. Precision is only about 4'88 [7] and the technique is not widely available. Recently, DXA manufacturers have developed software for whole-body measurements in small animals. The present study was designed to assess precision and accuracy of DXA in measuring total skeletal mineral content in the rat using the Hologic QDR 1000| device. 2. Materials and methods
Ten Wistar rats (5 males, 5 females) aged 24-72 days and weighing - 50, 100, 150, 200 and 250 g, respectively, were anesthetized with a mixture of ketamine (50 mg/kg) and xylazine (10 mg/kg) administrated intraperitoneally. Whole body DXA measurements were carried out using a Hologic QDR 1000| densitometer equipped with a standard collimator (2.3 mm diameter). Anesthetized
63
J.-P. Casez et aL / Bone Miner. 26 (1994) 61-68
rats were placed upon the plexiglas window of the apparatus and scanned under the antero-posterior incidence, the tail being reclined under the body. Each rat was scanned six times consecutively after repositioning. Scanning resolution (pixel width x pixel length) was 0.5 x 0.5 mm for rats weighing 50 and 100 g and 0.5 x 1 mm for larger rats. Scanning duration varied between 10 and 20 min according to rat size. Analysis was performed using the rat whole-body software, version 5.52. Data were gathered as bone mineral content (BMC, g hydroxyapatite) and bone mineral density (BMD, g/cm2). BMC was then expressed as total body calcium (TBCa[DXA]) based on the assumption that calcium accounts for 40.5% o f the BMC expressed as hydroxyapatite. After the last measurement the rats were sacrificed and fed to dermestid beetles to remove the soft tissues and leave a clean skeleton [8]. The entire skeleton was then deposited onto the b o t t o m o f a plastic box and measured by D X A under a 2-cm water layer using a special collimator (0.8 m m diameter) and ultra high resolution software (version 4.47; pixel size, 0.125 x 0.254 mm). Analysis was performed using a low ( - 0 . 0 6 g/cm 2) threshold to improve the detection o f very low density bone. BMC provided by the ultra-high resolution software was also expressed as a quantity of calcium (TBCa[UH]) using the aforementioned assumption. Afterwards, skeletons were ashed and material dissolved in hydrochloric acid. Total body calcium from the ash was measured chemically (TBCa[chem]) by atomic absorption spectrophotometry (CV = 0.8%). 2.1. Statistical analysis
Precision error for BMC and B M D was calculated from the 6 whole body measurements and expressed as a mean coefficient o f variation (r~S.D./r~M). Correlations were performed using simple and multiple regression analysis. Accuracy error, expressed in percent, was calculated as the absolute difference between
Table I Weight. total body BMC and BMD measured with DXA in 10 rats in vivo (No.)
Weight (g)
BMC (M • S.D.) (g)
CVa (%)
BMD (M • S.D.) (g/cm-')
CVa (%)
1 2 3 4 5 6 7 8 9 10
50 51 106 II1 154 156 190 203 258 265
0.97 • I.I1 + 1.93 • 1.98 • 3.29 • 3.07 • 3.30 • 4.65 • 5.50 • 4.95 •
2.3 2.4 1.2 2.0 2.3 1.0 0.7 0.9 0.8 1.3
0.066 + 0.070 + 0.072 • 0.070 • 0.098 • 0.089 • 0.086 • 0.109 • 0.110 • 0.098 •
1.2 1.2 0.7 1.7 2.4 1.3 1.2 1.3 2.4 1.4
Rat
Mean CV
0.02 0.03 0.02 0.04 0.08 0.03 0.02 0.04 0.04 0.07
1.3
0.001 0.001 0.001 0.001 0.002 0.001 0.001 0.001 0.003 0.001
1.5
aCoefficients of variation calculated from 6 consecutive measurements after repositioning.
64
J.-P. Casez et al. / Bone Miner. 26 (1994) 61-68
TBCa[chem] and the value predicted from DXA measurement using the single regression equation between TBCa[DXA] and TBCa[chem]. 3. Results
3.1. Precision Table 1 shows results (mean • S.D.) obtained in each rat for BMC and BMD. The mean precision error was 1.3% for total body BMC and 1.5% for total body BMD (mean CV). 3.2. Accuracy Individual values of TBCa as assessed with DXA whole-body software, ultra high resolution software and chemical analysis are shown in Table 2. Single regression analysis between TBCa[DXA], TBCa[UH] and TBCa[chem] revealed that slopes and intercepts differed significantly from 1 and 0, respectively (Fig. 1). However, these parameters were highly correlated and linearly related and, according to regression equations, TBCa[chem] can be calculated from DXA measurements with a mean accuracy error of 6.4 -4- 3.6% for TBCa[UH] and 8.5% • 11.9% for TBCa[DXA]. These values became smaller (5.7 • 3.1% and 5.2 • 9.7%, respectively) after exclusion of the very small rats (i.e. no. I rat and no. 2 rat, 50 and 51 g, respectively). TBCa as assessed with any technique was positively correlated with body weight (Fig. 2), the tightest correlation being found for TBCa[DXA]. Multiple regression analysis revealed that TBCa[DXA] (but not TBCa[UH]) was influenced by both TBCa[chem] and body weight. The following equation could be derived:
TBCa[DXA] = 0.606 • TBCa[chem] + 0.00282 • weight + 0.130 (r = 0.995, P < 0.0001, partial F = 23.1 for TBCa[DXA] and 5.8 for body weight).
Table 2 Total body calcium in 10 rats as assessed with whole-body DXA in vivo, ultra high resolution DXA in vitro and chemical analysis of ashed bones Rat (No.)
Weight (g)
TBCa[DXA] (g)
TBCa[UH] (g)
TBCa[chem] (g)
1 2 3 4 5 6 7 8 9 10
50 51 106 111 154 156 190 203 258 265
0.394 0.448 0.778 0.801 1.330 1.244 1.337 1.874 2.226 2.004
0.259 0.239 0.570 0.570 1.391 0.788 1.411 2.064 2.446 2.062
0.278 0.282 0.630 0.621 1.221 0.867 1.214 1.853 2.265 1.955
J.-P. Casez et aL / Bone Miner. 26 (1994) 61-68
2.5
y = O,gOx - 0.239
65
.,"
r- O.W1,SEE.=0.001
...~'~
........!ii.........
2
E ,~
1.5
a
0.5
0.5
2-:t
1 1.5 2 TBCa-chem (grams)
y = 0.78 x - 0,323
r. O.Oel,SEE=O
.-"
~
2.5
..-
< x
0.5
ols
i 115
~, Z s
TBCa-UH (grams)
2.5
y : 1.14 x - O.090 r
.d~
2 A
1.5
-r"
~- 0.5
05
1 1.5 2 TBCa-chem (grams)
2.5
Fig. 1. Correlations between total body calcium as measured chemically on ashes (TBCa[chem]), with whole-body DXA in vivo (TBCa[DXA]) and with ultra high resolution DXA of the skeleton in vitro (TBCa[UHI), in 10 rats weighing 50-250 g.
66
J.-P. Casez et al. / Bone Miner. 26 (1994) 61-68
"11 y " O.O08x- 0.03, r B 0.98
~
9
1.5t
1
O.S
0
SO
100
150
200
250
Body weight 2.S
y 9 O.01x - 0.377, r 9 0.95
2
O.S1 0
so
loo 1so ~
~,o ~o
Body weight 2.5
y 80.OOOx- 0~
r m0.96
9 &
2
1 0.5 Body weight
Fig. 2. Correlations obtained in 10 rats between body weight and total body calcium as measured chemically on ashes (TBCa[chem]), with whole-body DXA in vivo (TBCa[DXA]), or with ultra high resolution DXA of the skeleton in vitro (TBCa[UH]).
J.-P. Casez et al. / Bone Miner. 26 (1994) 61-68
67
4. Discussion The present study in rats provides very strong correlations between TBCa as assessed in vivo by DXA and by chemical analysis of the whole skeleton after sacrifice. Precision error achieved with this method is in the order of 1.5% for both BMC and BMD and is independent of body weight. This precision error is similar to that obtained with DXA on single bones in anesthetized rats [6] and is superior to values previously reported with other techniques such as radiogrammetry, single photon absorptiometry, dual photon absorptiometry, quantitative CT, and neutron activation analysis [ 1,3-6]. Undoubtedly, DXA will enable investigators to carry out studies using much smaller groups of animals than previously required with other techniques. Mean accuracy error for TBCa[DXA] in the entire population of animals was 8.5%. Exclusion of the very small animals (50 g rats) reduced the accuracy error to less than 6"/0, a value similar to that obtained in vitro with the ultra high resolution software and comparable to results reported in the literature in humans with dual photon absorptiometry [9]. Multiple regression analysis showed positive correlation of TBCa[DXA] with both TBCa[chem] and body weight. Although it is obvious that bigger rats have bigger bones, a small impact of body weight on TBCa[DXA] due to differences in soft tissue thickness cannot be excluded, as suggested by Amman et al. [6]. In conclusion, whole-body DXA appears as a suitable method to assess with high precision total skeletal calcium and bone mineral density in vivo in the rat. The technique should allow prospective study of the influence of new therapeutical modalities in various animal models of osteoporosis. Acknowledgements We are grateful to Miss Joanna Fowler and to Mr Andreas Reinli for their help in performing DXA scans. References I 2 3 4
5 6
Mbuyi-Muamba JM, Dequeker J. Bone mass measured by photon absorptiometry and radiogrammetry in Wistar rats. Invest Radiol 1986;21:336-339. Del Pozo E. Bone loss in adjuvant arthritis. Triangle 1989;28:43-50. Sanchez TV, Myers RC, Bond JT, Mayor GH, Bone mineral analysis of the rat femur by direct photon absorptiometry. Calcif Tissue int 1981;33:587-590. Safadi M, Shapira D, Leichter I, Reznick A, Silbermann M. Ability of different techniques of measuring bone mass to determine vertebral bone loss in aging female rats. Calcif Tissue Int 1988;42:375-382. Kimmel DB, Wronski TJ. Nondestructive measurement of bone mineral in femurs from ovariectomized rats. Calcif Tissue lnt 1990;46:101-110. Ammann P, Rizzoli R, Slosman D, Bonjour J-Ph. Sequential and precise in vivo measurement of bone mineral density in rats using dual-energy X-ray absorptiometry. J Bone Miner Res 1992;7:311-316.
68
J.-P. Casez et al./Bone Miner. 26 (1994) 61-68
7
Schubiger PA, Jakob K, Ferreira-Marques FM, Hefti E, Trechsel U, Miihlbauer R, Fleisch H. Determination of whole-body calcium in vivo in rats by neutron activation analysis. Int J Nucl Med 1980;7:32-36. Hefti E, Trechsel U, R/ifenacht H, Fleisch H. Use of dermestid beetles for cleaning bones. Calcif Tissue I nt 1980;31:45-47. Gotfredsen A, Podenphant J, Norgaard H, Nilas L, Nielsen VH, Christiansen C. Accuracy of lumbar spine bone mineral content by dual photon absorptiometry. J Nucl Med 1988;29:248-254.
8 9
Bone and Mineral 26 (1994) 61-68
Dual-energy X-ray absorptiometry for measuring total bone mineral content in the rat: Study of accuracy and precision J-P. Casez a, R.C. Muehlbauer b, K. Lippuner a, T. Kelly c, H. Fleisch b, Ph. Jaeger *a apoliclinie of Medicine, bDepartment o[" Pathophysiology. University ~["Berne. Berne. Swit:erland CHologic lnt'.. Watthem~. MA. USA (Received 19 August 1993; revision received 4 January 1994: accepted 10 February 1994)
Abstract
Sequential studies of osteopenic bone disease in small animals require the availability of non-invasive, accurate and precise methods to assess bone mineral content (BMCJ and bone mineral density (BMD). Dual-energy X-ray absorptiometry (DXA), which is currently used in humans for this purpose, can also be applied to small animals by means of adapted software. Precision and accuracy of DXA was evaluated in 10 rats weighing 50-265 g. The rats were anesthetized with a mixture of ketamine-xylazine administrated intraperitoneally. Each rat was scanned six times consecutively in the antero-posterior incidence after repositioning using the rat whole-body software for determination of whole-body BMC and BMD (Hologic QDR 1000| software version 5.52). Scan duration was 10-20 min depending on rat size. After the last measurement, rats were sacrificed and soft tissues were removed by dermestid beetles. Skeletons were then scanned in vitro (ultra high resolution software, version 4.47). Bones were subsequently ashed and dissolved in hydrochloric acid and total body calcium directly assayed by atomic absorption spectrophotometry (TBCa[chem]). Total body calcium was also calculated from the DXA whole-body in vivo measurement (TBCa[DXA]) and from the ultra high resolution measurement (TBCa[UH]) under the assumption that calcium accounts for 40.5"/,, of the BMC expressed as hydroxyapatite. Precision error for whole-body BMC and BMD (mean • S.D.) was 1.3% and 1.5"/,,, respectively. Simple regression analysis between TBCa[DXA] or TBCa[UH] and TBCa[chem] revealed tight correlations (n = 0.991 and 0.996, respectively), with slopes and intercepts which were significantly different from 1 and 0, respectively. Mean accuracy error for TBCa[DXA] was 8.5% for all rats and 5.7% after * Corresponding author, Director of the Policlinic of Medicine, lnselspital, CH-3010, Switzerland. 0169-6009/94/$07.00 9 1994 Elsevier Science Ireland Ltd. All rights reserved SSDI 0169-6009( 94)00694-C
62
J.-P. Casez et al./ Bone Miner. 26 (1994) 61-68
excluding the 50-g rats. It was 6.4'V,,for TBCa[UH] (all rats). We conclude that whole-body DXA provides highly precise and reasonably accurate estimates of skeletal calcium content in the rat in vivo. Keywords: Rat; Skeletal calcium content; DXA, Whole body
1. Introduction
Over the last years it has become progressively apparent that development of noninvasive techniques to assess bone density in small animals in vivo was of paramount importance to further progress in the therapeutic approach to most forms of osteopenic bone disease. In particular accurate and precise sequential bone mass measurements are needed which provide information on different parts of the skeleton. Early methods for bone mass measurements in the rat were based on histomorphometry (which requires animal sacrifice) or on radiography. Bone assessment on X-ray films is possible using radiogrammetry [1] or relative optical densitometry [21. However, the level of precision reached with the former is poor and the latter can be performed only on caudal vertebrae, i.e. bones that may not be representative of the whole skeleton. Single photon absorptiometry has been used to evaluate bone mineral content (BMC) of the femur [3] or at lumbar spine [4] in the rat in vivo. Reproducibility was 3-6.6'V,,, which is not sufficient to detect small changes without a large sample size. Quantitative CT of the lumbar spine [41 and dual photon absorptiometry [5] have also been employed in vivo and in vitro. However, precision errors of 3 to 5'88 have adversely affected their ability to detect changes in skeletal status. Recently dual-energy X-ray absorptiometry (DXA) has become the method of choice to address osteopenic bone disease in humans due to high reproducibility, short investigation time, lower cost and low radiation exposure. DXA can also be applied to small animals. Amman et al. reported I.Y88 precision in measuring femoral or vertebral bone mineral density in the rat in vivo [6]. Neutron activation analysis also has been used to assess total skeletal calcium in the rat. Precision is only about 4'88 [7] and the technique is not widely available. Recently, DXA manufacturers have developed software for whole-body measurements in small animals. The present study was designed to assess precision and accuracy of DXA in measuring total skeletal mineral content in the rat using the Hologic QDR 1000| device. 2. Materials and methods
Ten Wistar rats (5 males, 5 females) aged 24-72 days and weighing - 50, 100, 150, 200 and 250 g, respectively, were anesthetized with a mixture of ketamine (50 mg/kg) and xylazine (10 mg/kg) administrated intraperitoneally. Whole body DXA measurements were carried out using a Hologic QDR 1000| densitometer equipped with a standard collimator (2.3 mm diameter). Anesthetized
63
J.-P. Casez et aL / Bone Miner. 26 (1994) 61-68
rats were placed upon the plexiglas window of the apparatus and scanned under the antero-posterior incidence, the tail being reclined under the body. Each rat was scanned six times consecutively after repositioning. Scanning resolution (pixel width x pixel length) was 0.5 x 0.5 mm for rats weighing 50 and 100 g and 0.5 x 1 mm for larger rats. Scanning duration varied between 10 and 20 min according to rat size. Analysis was performed using the rat whole-body software, version 5.52. Data were gathered as bone mineral content (BMC, g hydroxyapatite) and bone mineral density (BMD, g/cm2). BMC was then expressed as total body calcium (TBCa[DXA]) based on the assumption that calcium accounts for 40.5% o f the BMC expressed as hydroxyapatite. After the last measurement the rats were sacrificed and fed to dermestid beetles to remove the soft tissues and leave a clean skeleton [8]. The entire skeleton was then deposited onto the b o t t o m o f a plastic box and measured by D X A under a 2-cm water layer using a special collimator (0.8 m m diameter) and ultra high resolution software (version 4.47; pixel size, 0.125 x 0.254 mm). Analysis was performed using a low ( - 0 . 0 6 g/cm 2) threshold to improve the detection o f very low density bone. BMC provided by the ultra-high resolution software was also expressed as a quantity of calcium (TBCa[UH]) using the aforementioned assumption. Afterwards, skeletons were ashed and material dissolved in hydrochloric acid. Total body calcium from the ash was measured chemically (TBCa[chem]) by atomic absorption spectrophotometry (CV = 0.8%). 2.1. Statistical analysis
Precision error for BMC and B M D was calculated from the 6 whole body measurements and expressed as a mean coefficient o f variation (r~S.D./r~M). Correlations were performed using simple and multiple regression analysis. Accuracy error, expressed in percent, was calculated as the absolute difference between
Table I Weight. total body BMC and BMD measured with DXA in 10 rats in vivo (No.)
Weight (g)
BMC (M • S.D.) (g)
CVa (%)
BMD (M • S.D.) (g/cm-')
CVa (%)
1 2 3 4 5 6 7 8 9 10
50 51 106 II1 154 156 190 203 258 265
0.97 • I.I1 + 1.93 • 1.98 • 3.29 • 3.07 • 3.30 • 4.65 • 5.50 • 4.95 •
2.3 2.4 1.2 2.0 2.3 1.0 0.7 0.9 0.8 1.3
0.066 + 0.070 + 0.072 • 0.070 • 0.098 • 0.089 • 0.086 • 0.109 • 0.110 • 0.098 •
1.2 1.2 0.7 1.7 2.4 1.3 1.2 1.3 2.4 1.4
Rat
Mean CV
0.02 0.03 0.02 0.04 0.08 0.03 0.02 0.04 0.04 0.07
1.3
0.001 0.001 0.001 0.001 0.002 0.001 0.001 0.001 0.003 0.001
1.5
aCoefficients of variation calculated from 6 consecutive measurements after repositioning.
64
J.-P. Casez et al. / Bone Miner. 26 (1994) 61-68
TBCa[chem] and the value predicted from DXA measurement using the single regression equation between TBCa[DXA] and TBCa[chem]. 3. Results
3.1. Precision Table 1 shows results (mean • S.D.) obtained in each rat for BMC and BMD. The mean precision error was 1.3% for total body BMC and 1.5% for total body BMD (mean CV). 3.2. Accuracy Individual values of TBCa as assessed with DXA whole-body software, ultra high resolution software and chemical analysis are shown in Table 2. Single regression analysis between TBCa[DXA], TBCa[UH] and TBCa[chem] revealed that slopes and intercepts differed significantly from 1 and 0, respectively (Fig. 1). However, these parameters were highly correlated and linearly related and, according to regression equations, TBCa[chem] can be calculated from DXA measurements with a mean accuracy error of 6.4 -4- 3.6% for TBCa[UH] and 8.5% • 11.9% for TBCa[DXA]. These values became smaller (5.7 • 3.1% and 5.2 • 9.7%, respectively) after exclusion of the very small rats (i.e. no. I rat and no. 2 rat, 50 and 51 g, respectively). TBCa as assessed with any technique was positively correlated with body weight (Fig. 2), the tightest correlation being found for TBCa[DXA]. Multiple regression analysis revealed that TBCa[DXA] (but not TBCa[UH]) was influenced by both TBCa[chem] and body weight. The following equation could be derived:
TBCa[DXA] = 0.606 • TBCa[chem] + 0.00282 • weight + 0.130 (r = 0.995, P < 0.0001, partial F = 23.1 for TBCa[DXA] and 5.8 for body weight).
Table 2 Total body calcium in 10 rats as assessed with whole-body DXA in vivo, ultra high resolution DXA in vitro and chemical analysis of ashed bones Rat (No.)
Weight (g)
TBCa[DXA] (g)
TBCa[UH] (g)
TBCa[chem] (g)
1 2 3 4 5 6 7 8 9 10
50 51 106 111 154 156 190 203 258 265
0.394 0.448 0.778 0.801 1.330 1.244 1.337 1.874 2.226 2.004
0.259 0.239 0.570 0.570 1.391 0.788 1.411 2.064 2.446 2.062
0.278 0.282 0.630 0.621 1.221 0.867 1.214 1.853 2.265 1.955
J.-P. Casez et aL / Bone Miner. 26 (1994) 61-68
2.5
y = O,gOx - 0.239
65
.,"
r- O.W1,SEE.=0.001
...~'~
........!ii.........
2
E ,~
1.5
a
0.5
0.5
2-:t
1 1.5 2 TBCa-chem (grams)
y = 0.78 x - 0,323
r. O.Oel,SEE=O
.-"
~
2.5
..-
< x
0.5
ols
i 115
~, Z s
TBCa-UH (grams)
2.5
y : 1.14 x - O.090 r
.d~
2 A
1.5
-r"
~- 0.5
05
1 1.5 2 TBCa-chem (grams)
2.5
Fig. 1. Correlations between total body calcium as measured chemically on ashes (TBCa[chem]), with whole-body DXA in vivo (TBCa[DXA]) and with ultra high resolution DXA of the skeleton in vitro (TBCa[UHI), in 10 rats weighing 50-250 g.
66
J.-P. Casez et al. / Bone Miner. 26 (1994) 61-68
"11 y " O.O08x- 0.03, r B 0.98
~
9
1.5t
1
O.S
0
SO
100
150
200
250
Body weight 2.S
y 9 O.01x - 0.377, r 9 0.95
2
O.S1 0
so
loo 1so ~
~,o ~o
Body weight 2.5
y 80.OOOx- 0~
r m0.96
9 &
2
1 0.5 Body weight
Fig. 2. Correlations obtained in 10 rats between body weight and total body calcium as measured chemically on ashes (TBCa[chem]), with whole-body DXA in vivo (TBCa[DXA]), or with ultra high resolution DXA of the skeleton in vitro (TBCa[UH]).
J.-P. Casez et al. / Bone Miner. 26 (1994) 61-68
67
4. Discussion The present study in rats provides very strong correlations between TBCa as assessed in vivo by DXA and by chemical analysis of the whole skeleton after sacrifice. Precision error achieved with this method is in the order of 1.5% for both BMC and BMD and is independent of body weight. This precision error is similar to that obtained with DXA on single bones in anesthetized rats [6] and is superior to values previously reported with other techniques such as radiogrammetry, single photon absorptiometry, dual photon absorptiometry, quantitative CT, and neutron activation analysis [ 1,3-6]. Undoubtedly, DXA will enable investigators to carry out studies using much smaller groups of animals than previously required with other techniques. Mean accuracy error for TBCa[DXA] in the entire population of animals was 8.5%. Exclusion of the very small animals (50 g rats) reduced the accuracy error to less than 6"/0, a value similar to that obtained in vitro with the ultra high resolution software and comparable to results reported in the literature in humans with dual photon absorptiometry [9]. Multiple regression analysis showed positive correlation of TBCa[DXA] with both TBCa[chem] and body weight. Although it is obvious that bigger rats have bigger bones, a small impact of body weight on TBCa[DXA] due to differences in soft tissue thickness cannot be excluded, as suggested by Amman et al. [6]. In conclusion, whole-body DXA appears as a suitable method to assess with high precision total skeletal calcium and bone mineral density in vivo in the rat. The technique should allow prospective study of the influence of new therapeutical modalities in various animal models of osteoporosis. Acknowledgements We are grateful to Miss Joanna Fowler and to Mr Andreas Reinli for their help in performing DXA scans. References I 2 3 4
5 6
Mbuyi-Muamba JM, Dequeker J. Bone mass measured by photon absorptiometry and radiogrammetry in Wistar rats. Invest Radiol 1986;21:336-339. Del Pozo E. Bone loss in adjuvant arthritis. Triangle 1989;28:43-50. Sanchez TV, Myers RC, Bond JT, Mayor GH, Bone mineral analysis of the rat femur by direct photon absorptiometry. Calcif Tissue int 1981;33:587-590. Safadi M, Shapira D, Leichter I, Reznick A, Silbermann M. Ability of different techniques of measuring bone mass to determine vertebral bone loss in aging female rats. Calcif Tissue Int 1988;42:375-382. Kimmel DB, Wronski TJ. Nondestructive measurement of bone mineral in femurs from ovariectomized rats. Calcif Tissue lnt 1990;46:101-110. Ammann P, Rizzoli R, Slosman D, Bonjour J-Ph. Sequential and precise in vivo measurement of bone mineral density in rats using dual-energy X-ray absorptiometry. J Bone Miner Res 1992;7:311-316.
68
J.-P. Casez et al./Bone Miner. 26 (1994) 61-68
7
Schubiger PA, Jakob K, Ferreira-Marques FM, Hefti E, Trechsel U, Miihlbauer R, Fleisch H. Determination of whole-body calcium in vivo in rats by neutron activation analysis. Int J Nucl Med 1980;7:32-36. Hefti E, Trechsel U, R/ifenacht H, Fleisch H. Use of dermestid beetles for cleaning bones. Calcif Tissue I nt 1980;31:45-47. Gotfredsen A, Podenphant J, Norgaard H, Nilas L, Nielsen VH, Christiansen C. Accuracy of lumbar spine bone mineral content by dual photon absorptiometry. J Nucl Med 1988;29:248-254.
8 9