- Email: [email protected]
Do Appendicular Bone Measurements Reflect Changes in the Axial Skeleton?
Journal of Clinical Densitometry, vol. 7, no. 3, 296–301, 2004 © Copyright 2004 by Humana Press Inc. All rights of any nature whatsoever reserved. 1094-6950/04/7:296–301/$25.00
Original Article
Do Appendicular Bone Measurements Reflect Changes in the Axial Skeleton? The Use of Dual-Energy X-Ray Absorptiometry and Ultrasound Measurements During Lactation
M. Ann Laskey* DPhiL and Ann Prentice, PhD MRC Human Nutrition Research, Elsie Widdowson Laboratory, Cambridge, CB1 9NL, UK
Abstract The ability of different bone measurement techniques to monitor changes in bone mineral was studied. Lactation was used as a model because large, rapid, but reversible decreases in bone mineral content (BMC) occur in breastfeeding women. Spine and forearm dual-energy X-ray absorptiometry (DXA) and calcaneal quantitative ultrasound (QUS) measurements were made during 30 lactations. During the first 3 mo of lactation, decreases in the BMC, adjusted for area, were significant at the spine (–2.8%; standard error [SE] = 0.6; p < 0.001) but not the wrist (p = 0.40). Nonsignificant increases in normalized broadband ultrasound attenuation (nBUA) and velocity of sound (VOS) were observed at the calcaneus using QUS. From peak lactation to postlactation, the BMC increases at the spine were significant (4.1%; SE = 0.6; p < 0.001) but not those at the wrist (p = 0.17). Nonsignificant decreases were observed using QUS. Eleven breast-feeding women had longitudinal calcaneal and spine DXA measurements from peak lactation to postlactation. Significant BMC increases were observed at both sites (calcaneus: 2.4%, SE = 0.7, p < 0.01; spine: 3.3%, SE = 1.3, p < 0.03). The similarity of DXA calcaneal changes to spine changes indicates that DXA calcaneal measurements could be a useful alternative tool when it is difficult to monitor BMC at axial sites. Key Words: Dual-energy X-ray absorptiometry; ultrasound; monitoring change; spine; wrist; calcaneus.
tends to be greatest at this site. Unfortunately, there are individuals for which the spine is difficult or impossible to study. These include elderly individuals whose spines might have become compressed or contain osteophytes, subjects unable to lie on the dual-energy X-ray absorptiometer (DXA) couch, and pregnant women. In such circumstances, alternative peripheral skeletal sites rich in trabecular bone might be useful alternatives. Two peripheral skeletal sites contain significant amounts of trabecular bone: the wrist (25%) and the calcaneus (80%). The calcaneus can be studied using DXA and quantitative ultrasound (QUS). QUS is a relatively inexpensive, portable technique that does not use ionizing radiation and could be especially suitable for longitudinal studies of subjects vulnerable to radiation (e.g., children and pregnant women). QUS has been proposed as an alternative technique to DXA for evaluating osteoporosis risk (3–5). Two QUS measurements are generally reported: broadband ultrasound attenuation
Introduction Only 20% of the skeleton is composed of trabecular bone, but because of its high metabolic activity, this is where 80% of bone turnover occurs (1). Osteoporotic fractures, which result from abnormal bone turnover, occur predominantly in trabecular bone sites. The proportion of trabecular bone varies widely at different skeletal sites (spine, 70%; hip, 25–50%; wrist, 25%) (2). The spine is an ideal site for monitoring longitudinal changes in trabecular bone mineral content (BMC) because the response to physiological changes (e.g., menopause, lactation)
Received 12/09/03; Revised 02/26/04; Accepted 02/26/04. * Address correspondence to Dr. M.A. Laskey, MRC Human Nutrition Research, Elsie Widdowson Laboratory, Fulbourn Road, Cambridge, CB1 9NL, UK. E-mail: [email protected]
296
Monitoring BMC Changes in Appendicular Skeleton (BUA [dB/MHz]), which is a measure of energy lost as it passes through bone and tissue, and velocity of sound (VOS [m/s]) which represents the velocity of ultrasound transmission through bone. During lactation, there are large and rapid decreases in the BMC (6–8). These changes are greatest at skeletal regions rich in trabecular bone. When lactation declines and menstruation resumes, the BMC increases (9). The magnitude and duration of the decrease is greater in women who breast-feed for a longer time and does not occur in formula-feeding women (9). Lactation is, therefore, an ideal model to examine if changes in bone measurements (DXA and QUS) at peripheral trabecular bone sites (wrist and calcaneus) are similar to the large and rapid change known to occur at the spine. To investigate this, we have analyzed data from longitudinal DXA (spine, wrist, and calcaneus) and QUS (calcaneus) measurements in women during and after lactation and in formula-feeding women.
Materials and Methods Subjects Healthy Caucasian mothers were recruited postnatally from the local maternity hospital in Cambridge. Women with a history of bone disease or taking medications known to affect bone were excluded. Twenty-three breast-feeding and 11 formula-feeding mothers were recruited. Seven of the breast-feeding mothers volunteered to repeat the study during a subsequent lactation. Hence, measurements were made during 30 lactations. Women lactated for very variable lengths of time (median lactation length = 305 d). This study was part of a larger study of lactation, for which ethical approval had been obtained from the Ethical Committee of the MRC Dunn Nutrition Unit (of which the Nutrition and Bone Health group of MRC Human Nutrition Research was formerly a part). Informed written consent was obtained from each subject.
Methods The formula-feeding women had spine and forearm measurements at 0.5 mo (mean ± SD = 18.6 ± 6.8 d) and 3 mo after delivery (mean ± SD = 89.7 ± 13.2 d). Calcaneal DXA and ultrasound measurements in this group were incomplete and are not reported. Breast-feeding women had spine and forearm DXA and calcaneal QUS measurements at 0.5 (mean ± SD = 16.0 ± 3.7 d), 3 (mean ± SD = 90.1 ± 5.6 d), 6, and 12 mo after delivery. Additional measurements were also made at 3 mo postlactation for those mothers who lactated for more than 9 mo. To monitor recovery of bone mineral after peak lactation, the time-points corresponding to the minimum (peak lactation) and maximum (postlactation) spine bone mineral results were selected. The magnitude and timing in the nadir and recovery of spine bone mineral varied between mothers and on their length of lactation. The minimum spine bone mineral was usually observed at 3 or 6 mo postpartum (mean ± SD = 164 ± 104 d) and maximum at 12 mo postpartum or 3 mo postlactation (mean ± SD = 487 ± 215 d). Journal of Clinical Densitometry
297 Eleven of the 23 lactating women also had longitudinal DXA measurements at the calcaneus. Initial measurements were usually made at 3 or 6 mo postpartum. It was, therefore, only possible to monitor the recovery of DXA calcaneal measurements after peak lactation.
Dual-Energy X-Ray Absorptiometry Bone mineral content (BMC [g]) and bone area (BA [cm2]) of the lumbar spine (L1–L4), nondominant forearm, and calcaneus were measured by DXA (QDR-1000/W, Hologic Inc, Waltham, MA). Subsequent scans were analyzed with reference to the individual’s baseline image using the Compare Facility. The performance mode was used for the spine (software v 4.47P). Software version 5.61 was used for forearm scans. Measurements of spine and forearm were performed as described in the Hologic manual (Hologic Inc, Waltham, MA). For measurements of the left calcaneus, individuals were positioned on their left side on the scanning bed facing the operator. The left leg was bent with the left foot resting against the forearm-positioning device. The base of the foot was parallel to the scanning bed. Using the left forearm scan mode, the laser light was positioned just below and in front of the talus. The scan area included the complete calcaneus plus surrounding soft tissue. Scans were analyzed using the protocol subregion analysis. One rectangular region of interest was analyzed; it included the complete calcaneus (Fig. 1). Attempts to localize calcaneal subregions resulted in inferior precision. Quality assurance and long-term instrument stability were assessed using the Hologic spine phantom. This was scanned at the beginning of each measurement day. During the study, the coefficient of variation (%CV) of phantom measurements for both the BMC and BA was <0.4% and there was no indication of any significant drift over time. Estimates of in vivo precision of BA-adjusted BMC determined from two sets (n = 22) of scans at an interval of approx 3 mo were spine = 0.9% and radial wrist = 1.4%. Short-term precision (%CV) of the calcaneal BA-adjusted BMC from duplicate measurements on the same day with repositioning of the subject between scans (n = 25) was 1.8%.
Quantitative Ultrasound Quantitative ultrasound (QUS) measurements were made using the CUBA Clinical (McCue Ultrasonics, Winchester UK) on the same day as the DXA measurements. This instrument measured BUA (dB/MHz) and velocity of sound VOS (m/s) at the mid-calcaneus. This transmission technique used a pair of ultrasonic transducers in direct contact with the subject’s heel via silicone rubber pads. Water-soluble ultrasound gel was used to ensure acoustic coupling of the pads to the heel. Heel width was determined from the separation of the transducers and the BUA was normalized for heel width to give the normalized BUA (nBUA). The in vivo precision (%CV) of duplicate QUS measurements (n = 24) performed on the same day were nBUA = 7.2% and VOS = 0.9%. Short-term precision (%CV) was optimized by obtaining nine scans of the left foot with three positioning Volume 7, 2004
298
Laskey and Prentice
Fig. 1. Image and bone mineral results of DXA measurement of the calcaneus.
of the calcaneus. Optimized short-term precision determined from duplicate sets of nine measurements (n = 24) were nBUA = 3.6% and VOS = 0.42% (n = 24). Long-term precision (CV%), determined using the optimized protocol from two sets of scans (n = 16) at an interval of approx 6 mo, were nBUA = 3.5% and VOS = 3.4%.
Statistics The BMC was adjusted for bone size by including the BA in the analysis of covariance (ANCOVA) models (10). Investigation of changes over time within individuals was performed by ANCOVA after transformation of continuous variables to natural logarithms. Scheffe’s post hoc test was used to determine the significance of differences between specific pairs of time-points.
Results The characteristics of the women at 0.5 mo postpartum are shown in Table 1. Both breast-feeding and formula-feeding women had lost about 2 kg in weight by 3 mo postpartum. At this time, no significant changes in BMC from 0.5 mo postpartum were seen in the formula-feeding women at the spine (–0.5%; SE = 0.7; p = 0.93) or wrist (–0.2%; SE = 0.9; p = 0.99). In contrast, after 3 mo lactation, decreases in BMC were significant at the spine (p < 0.001), but not at the wrist (p = 0.4). QUS measurements had increased nonsignificantly (nBUA, p = 0.40; VOS, p = 0.61) (Fig. 2). Results showing the recovery of spine BMC after lactation (minimum to maximum spine BMC) and the changes for the other measurements at the corresponding time-points, are shown in Fig. 3. As lactation declined, BMC increased at both the spine and wrist, but changes were only significant at the spine (p < 0.001 and p = 0.17 respectively). Nonsignificant decreases were observed at the calcaneus using QUS (nBUA, Journal of Clinical Densitometry
Table 1 Subject Characteristics at 0.5 mo Postpartum Breast-feeders (n = 30)
Age (yr) Height (m) Weight (kg) Spine BMD (g/cm2) Wrist BMD (g/cm2) VOS (m/s) nBUA (dB/MHz)
Formula-feeders (n = 11)
Mean
SD
Mean
SD
33.6 1.65 68.4 1.052
3.5 0.05 12.6 0.131
29.5 1.67 75.8 1.073
3.4 0.06 24.5 0.171
0.432
0.047
0.454
0.057
1695 90.5
58 20.0
NM NM
NM NM
Abbr: NM, not measured.
p = 0.49; VOS, p = 0.80). The corresponding DXA results for the 11 women with both spine and calcaneal DXA measurements are shown in Fig. 4. This shows that significant increases in BMC, similar to those observed at the spine, were observed in these women.
Discussion This is the first comparison of longitudinal changes in calcaneal bone using both DXA and QUS during lactation. The decrease in BMC during lactation followed by recovery as lactation declines, known to occur at the spine (8,9), were not Volume 7, 2004
Monitoring BMC Changes in Appendicular Skeleton
Fig. 2. Mean (SE) percentage changes during lactation from 0.5 to 3 mo postpartum (pp). DXA spine and wrist measurements were BA-adjusted BMC; calcaneal quantitative ultrasound (QUS) measurements were in decibels per megahertz for BUA and in meters per second for VOS (n = 30). NS = nonsignificant.
Fig. 3. Mean (SE) percentage changes, from time of minimal to maximum spine bone mineral. DXA spine and wrist measurements were BA-adjusted BMC; calcaneal QUS measurements were in decibels per megahertz for BUA and in meters per second for VOS (n = 30). NS = nonsignificant. Journal of Clinical Densitometry
299
Fig. 4. Mean (SE) percentage changes, from time of minimal to maximum spine bone mineral, in BA-adjusted BMC at the spine and calcaneus using DXA (n = 11).
detected by calcaneal QUS. The initial decrease in BMC during the first 3 mo of lactation could not be measured, as initial calcaneal DXA measurements were obtained 3–6 mo postpartum. However, the recovery in BMC, as lactation declined, was detected by calcaneal DXA measurements. These DXA changes at the calcaneus were significant despite the small size of the study. No significant postpartum changes were detected by DXA at either the spine or wrist in the formula-feeding women or at the wrist in breast-feeding women. Previous studies have measured BMC changes at the forearm during pregnancy and lactation (7,9,11). These studies reported no changes at the radial shaft, which is predominantly cortical bone. Reported changes in wrist BMC were variable. Kalkwarf et al. reported no changes (7), whereas Koltoff et al. reported a 2% decrease at the wrist during the second half of pregnancy but no change during lactation (11). Another study only observed significant changes in BMC at the wrist after 6 mo or more of lactation (9). Our analyses indicate that the direction of BMC changes at the wrist and spine are identical, but the magnitude of the change is much smaller at the wrist, probably reflecting its lower trabecular bone content (25%; cf. spine 70%). The ability of DXA, but not QUS, calcaneal measurements to monitor bone mineral changes might be the result of the inferior precision of QUS compared to DXA measurements. Short-term precision for the CUBA Clinical was similar to previous reports (12–14). The precision of nBUA and VOS was greatly improved by taking many repeat measurements. Long-term precision of VOS was inferior to short-term precision. This has been reported previously (15) and might be the result of nonspecific drift, temperature dependence and inadequate quality control procedures of QUS systems (12,16). Volume 7, 2004
300 At present, it is unclear whether QUS can reliably measure longitudinal changes in bone over a short time-span (17). One QUS study was unable to detect the longitudinal changes in bone mineral detected by DXA during hormone replacement therapy (HRT) (18). Other studies have reported a positive response to HRT in studies over 4 yr (19). In agreement with our data, one study detected no change in VOS or BUA during lactation (20). However, other studies have reported decreases in QUS measurements during lactation and pregnancy (21–23). The comparison of QUS studies is complicated by the increasing diversity of commercial QUS devices that measure different peripheral sites. Commercial QUS devices use either the transmission of ultrasound through accessible limb bones (calcaneus and phalanges) or the reflectance of ultrasound waves along the bone surface. Results for apparently similar QUS measurements obtained on one QUS system cannot necessarily be directly translated into performance statements of other technologically different QUS systems. Correlation coefficients vary between 0.4 and 0.8 (13). These are similar to the correlation coefficients for DXA measurements at different skeletal sites (between 0.4 and 0.7) (24) and between the QUS measurements and DXA measurements at different skeletal sites found by us (0.3–0.7, unpublished data) and others (14,17). These moderate correlations indicate that different QUS systems measure different properties of bone (4) and these are not closely related to BMC measured by DXA. During lactation, the decrease in BMC at the spine is associated with higher bone turnover and increase in bone remodeling sites. However, during aging, trabecular bone could thin and then break, resulting in the bone fragility associated with osteoporosis. BUA measurements depend on the trabecular architecture of cancellous bone and might relate to the separation and connectivity of trabeculi. This could explain why QUS measurements are predictive of fracture risk in elderly subjects (25–27) but are poor at monitoring changes in bone mineral in young subjects. The calcaneus has potential advantages as a DXA measurement site. It is unlikely to be affected by artifacts such as deformity and osteophytes commonly seen at the spine of elderly subjects. The effective dose from radiation is very low (<0.1 µSv for pencil-beam DXAs), making it an especially suitable site for studies of children, pregnant women, and other vulnerable subjects. The calcaneus has a high percentage of trabecular bone (>80%), similar to that found in the spine (70%), and is, therefore, likely to show similar high metabolic activity. There are relatively few longitudinal studies comparing changes in BMC at the calcaneus to changes at the spine and other clinically important sites. However, accelerated bone loss has been detected by DXA at both the spine and calcaneus after the menopause (28). DXA measurements of the calcaneus can be made using specially designed equipment (GE Lunar PIXI, Madison, WI; HeelScan DX-2000, KDK, Kyoto, Japan) or conventional DXA equipment can be adapted using modified software as described in this article and other studies (29–31). In conclusion, using lactation as a model of BMC change, DXA but not QUS measurements of the calcaneus demon-
Journal of Clinical Densitometry
Laskey and Prentice strated a significant increase in bone mineral, similar to that observed at the spine. These findings indicate that DXA measurements of the calcaneus might be a superior tool to calcaneal QUS when it is difficult or impossible to monitor changes in the BMC at conventional axial skeletal sites. Further larger studies are required to confirm this observation.
References 1. Fleisch H. 1996 Mechanism of Bone Loss and Modes of Action of Antiresorptive Therapies. In: Osteoporosis 1996. Papapoulos SE, Lips P, Pols HAP, Johnson CC, Delmas PD, eds. Elseveir, Amsterdam, pp. 7–15. 2. Mundy GR, ed. 1999 Bone Remodelling. Philadelphia: Lippincott Williams and Wilkins. 3. Funke M, Kopka L, Vosshenrich R, et al. 1995 Broadband ultrasound attenuation in the diagnosis of osteoporosis: correlation with osteodensitometry and fracture. Radiology 194:77–81. 4. Gluer CC. 1997 Quantitative ultrasound techniques for the assessment of osteoporosis: expert agreement on current status. The International Quantitative Ultrasound Consensus Group. J Bone Miner Res 12:1280–1288. 5. Hadji P, Hars O, Gorke K, Emons G, Schulz KD. 2000 Quantitative ultrasound of the os calcis in postmenopausal women with spine and hip fracture. J Clin Densitom 3:233–239. 6. Laskey MA, Prentice A, Hanratty LA, et al. 1998 Bone changes after 3 mo of lactation: influence of calcium intake, breast-milk output, and vitamin D-receptor genotype. Am J Clin Nutr 67:685–692. 7. Kalkwarf HJ, Specker BL, Bianchi DC, Ranz J, Ho M. 1997 The effect of calcium supplementation on bone density during lactation and after weaning. N Engl J Med 337:523–528. 8. Sowers M, Corton G, Shapiro B, et al. 1993 Changes in bone density with lactation. JAMA 269:3130–3135. 9. Laskey MA, Prentice A. 1999 Bone mineral changes during and after lactation. Obstet Gynecol 94:608–615. 10. 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. 11. Kolthoff N, Eiken P, Kristensen B, Nielsen SP. 1998 Bone mineral changes during pregnancy and lactation: a longitudinal cohort study. Clin Sci (Lond) 94:405–412. 12. Iki M, Kajita E, Mitamura S, Nishino H, Yamagami T, Nagahama N. 1999 Precision of quantitative ultrasound measurement of the heel bone and effects of ambient temperature on the parameters. Osteoporos Int 10:462–467. 13. Stewart A, Reid DM. 2000 Precision of quantitative ultrasound: comparison of three commercial scanners. Bone 27:139–143. 14. Tromp AM, Smit JH, Deeg DJ, Lips P. 1999 Quantitative ultrasound measurements of the tibia and calcaneus in comparison with DXA measurements at various skeletal sites. Osteoporos Int 9:230–235. 15. Ingle BM, Eastell R. 1996 Qualitative ultrasound measurements: short and long-term precision. J Bone Miner Res 11:1830. 16. Krieg MA, Cornuz J, Hartl F, et al. 2002 Quality controls for two heel bone ultrasounds used in the Swiss Evaluation of the Methods of Measurement of Osteoporotic Fracture Risk Study. J Clin Densitom 5:335–341. 17. Greenspan SL, Bouxsein ML, Melton ME, et al. 1997 Precision and discriminatory ability of calcaneal bone assessment technologies. J Bone Miner Res 12:1303–1313. 18. Lees B, Garland SW, Stevenson JC. 1996 Comparison of changes in ultrasound parameters and bone mineral density in
Volume 7, 2004
Monitoring BMC Changes in Appendicular Skeleton
19.
20.
21. 22. 23. 24.
women on hormone replacement therapy. J Bone Miner Res 11:1817. Sahota O, San P, Cawte SA, Pearson D, Hosking DJ. 2000 A comparison of the longitudinal changes in quantitative ultrasound with dual-energy X-ray absorptiometry: the four-year effects of hormone replacement therapy. Osteoporos Int 11:52–58. Yamaga A, Taga M, Minaguchi H, Sato K. 1996 Changes in bone mass as determined by ultrasound and biochemical markers of bone turnover during pregnancy and puerperium: a longitudinal study. J Clin Endocrinol Metab 81:752–756. Sowers MF, Scholl T, Harris L, Jannausch M. 2000 Bone loss in adolescent and adult pregnant women. Obstet Gynecol 96:189–193. Javaid MK, Taylor P, Shore SR, Inskip HM, Godfrey KM, Cooper C. 2001 Longitudinal assessment of bone status during pregnancy using calcaneal ultrasound. Osteoporos Int 12:S28. Giorgino R, Lorusso D, Paparella P. 1996 Pregnancy and lactation: longitudinal evaluation of bone status by ultrasound densitometry. Osteoporos Int 6:187. Sundberg M, Gardsell P, Johnell O, Ornstein E, Sernbo I. 1998 Comparison of quantitative ultrasound measurements in calcaneus with DXA and SXA at other skeletal sites: a populationbased study on 280 children aged 11–16 years. Osteoporos Int 8:410–417.
Journal of Clinical Densitometry
301 25. Thompson P, Taylor J, Fisher A, Oliver R. 1998 Quantitative heel ultrasound in 3180 women between 45 and 75 years of age: compliance, normal ranges and relationship to fracture history. Osteoporos Int 8:211–214. 26. Hans D, Dargent-Molina P, Schott AM, et al. 1996 Ultrasonographic heel measurements to predict hip fracture in elderly women: the EPIDOS prospective study. Lancet 348: 511–514. 27. Njeh CF, Kuo CW, Langton CM, Atrah HI, Boivin CM. 1997 Prediction of human femoral bone strength using ultrasound velocity and BMD: an in vitro study. Osteoporos Int 7:471–477. 28. Ito M, Nakamura T, Tsurusaki K, Uetani M, Hayashi K. 1999 Effects of menopause on age-dependent bone loss in the axial and appendicular skeletons in healthy Japanese women. Osteoporos Int 10:377–383. 29. Kang C, Speller R. 1999 The effect of region of interest selection on dual energy X-ray absorptiometry measurements of the calcaneus in 55 post-menopausal women. Br J Radiol 72:864–871. 30. Yamada M, Ito M, Hayashi K, Nakamura T. 1993 Calcaneus as a site for assessment of bone mineral density: evaluation in cadavers and healthy volunteers. Am J Roentgenol 161:621–627. 31. Rees S, McCoy K, Allen J, Taggart H. 2001 Measurement of bone density on the foot using dual energy x-ray absorptiometry. Osteoporosis Int 12:S27.
Volume 7, 2004
Original Article
Do Appendicular Bone Measurements Reflect Changes in the Axial Skeleton? The Use of Dual-Energy X-Ray Absorptiometry and Ultrasound Measurements During Lactation
M. Ann Laskey* DPhiL and Ann Prentice, PhD MRC Human Nutrition Research, Elsie Widdowson Laboratory, Cambridge, CB1 9NL, UK
Abstract The ability of different bone measurement techniques to monitor changes in bone mineral was studied. Lactation was used as a model because large, rapid, but reversible decreases in bone mineral content (BMC) occur in breastfeeding women. Spine and forearm dual-energy X-ray absorptiometry (DXA) and calcaneal quantitative ultrasound (QUS) measurements were made during 30 lactations. During the first 3 mo of lactation, decreases in the BMC, adjusted for area, were significant at the spine (–2.8%; standard error [SE] = 0.6; p < 0.001) but not the wrist (p = 0.40). Nonsignificant increases in normalized broadband ultrasound attenuation (nBUA) and velocity of sound (VOS) were observed at the calcaneus using QUS. From peak lactation to postlactation, the BMC increases at the spine were significant (4.1%; SE = 0.6; p < 0.001) but not those at the wrist (p = 0.17). Nonsignificant decreases were observed using QUS. Eleven breast-feeding women had longitudinal calcaneal and spine DXA measurements from peak lactation to postlactation. Significant BMC increases were observed at both sites (calcaneus: 2.4%, SE = 0.7, p < 0.01; spine: 3.3%, SE = 1.3, p < 0.03). The similarity of DXA calcaneal changes to spine changes indicates that DXA calcaneal measurements could be a useful alternative tool when it is difficult to monitor BMC at axial sites. Key Words: Dual-energy X-ray absorptiometry; ultrasound; monitoring change; spine; wrist; calcaneus.
tends to be greatest at this site. Unfortunately, there are individuals for which the spine is difficult or impossible to study. These include elderly individuals whose spines might have become compressed or contain osteophytes, subjects unable to lie on the dual-energy X-ray absorptiometer (DXA) couch, and pregnant women. In such circumstances, alternative peripheral skeletal sites rich in trabecular bone might be useful alternatives. Two peripheral skeletal sites contain significant amounts of trabecular bone: the wrist (25%) and the calcaneus (80%). The calcaneus can be studied using DXA and quantitative ultrasound (QUS). QUS is a relatively inexpensive, portable technique that does not use ionizing radiation and could be especially suitable for longitudinal studies of subjects vulnerable to radiation (e.g., children and pregnant women). QUS has been proposed as an alternative technique to DXA for evaluating osteoporosis risk (3–5). Two QUS measurements are generally reported: broadband ultrasound attenuation
Introduction Only 20% of the skeleton is composed of trabecular bone, but because of its high metabolic activity, this is where 80% of bone turnover occurs (1). Osteoporotic fractures, which result from abnormal bone turnover, occur predominantly in trabecular bone sites. The proportion of trabecular bone varies widely at different skeletal sites (spine, 70%; hip, 25–50%; wrist, 25%) (2). The spine is an ideal site for monitoring longitudinal changes in trabecular bone mineral content (BMC) because the response to physiological changes (e.g., menopause, lactation)
Received 12/09/03; Revised 02/26/04; Accepted 02/26/04. * Address correspondence to Dr. M.A. Laskey, MRC Human Nutrition Research, Elsie Widdowson Laboratory, Fulbourn Road, Cambridge, CB1 9NL, UK. E-mail: [email protected]
296
Monitoring BMC Changes in Appendicular Skeleton (BUA [dB/MHz]), which is a measure of energy lost as it passes through bone and tissue, and velocity of sound (VOS [m/s]) which represents the velocity of ultrasound transmission through bone. During lactation, there are large and rapid decreases in the BMC (6–8). These changes are greatest at skeletal regions rich in trabecular bone. When lactation declines and menstruation resumes, the BMC increases (9). The magnitude and duration of the decrease is greater in women who breast-feed for a longer time and does not occur in formula-feeding women (9). Lactation is, therefore, an ideal model to examine if changes in bone measurements (DXA and QUS) at peripheral trabecular bone sites (wrist and calcaneus) are similar to the large and rapid change known to occur at the spine. To investigate this, we have analyzed data from longitudinal DXA (spine, wrist, and calcaneus) and QUS (calcaneus) measurements in women during and after lactation and in formula-feeding women.
Materials and Methods Subjects Healthy Caucasian mothers were recruited postnatally from the local maternity hospital in Cambridge. Women with a history of bone disease or taking medications known to affect bone were excluded. Twenty-three breast-feeding and 11 formula-feeding mothers were recruited. Seven of the breast-feeding mothers volunteered to repeat the study during a subsequent lactation. Hence, measurements were made during 30 lactations. Women lactated for very variable lengths of time (median lactation length = 305 d). This study was part of a larger study of lactation, for which ethical approval had been obtained from the Ethical Committee of the MRC Dunn Nutrition Unit (of which the Nutrition and Bone Health group of MRC Human Nutrition Research was formerly a part). Informed written consent was obtained from each subject.
Methods The formula-feeding women had spine and forearm measurements at 0.5 mo (mean ± SD = 18.6 ± 6.8 d) and 3 mo after delivery (mean ± SD = 89.7 ± 13.2 d). Calcaneal DXA and ultrasound measurements in this group were incomplete and are not reported. Breast-feeding women had spine and forearm DXA and calcaneal QUS measurements at 0.5 (mean ± SD = 16.0 ± 3.7 d), 3 (mean ± SD = 90.1 ± 5.6 d), 6, and 12 mo after delivery. Additional measurements were also made at 3 mo postlactation for those mothers who lactated for more than 9 mo. To monitor recovery of bone mineral after peak lactation, the time-points corresponding to the minimum (peak lactation) and maximum (postlactation) spine bone mineral results were selected. The magnitude and timing in the nadir and recovery of spine bone mineral varied between mothers and on their length of lactation. The minimum spine bone mineral was usually observed at 3 or 6 mo postpartum (mean ± SD = 164 ± 104 d) and maximum at 12 mo postpartum or 3 mo postlactation (mean ± SD = 487 ± 215 d). Journal of Clinical Densitometry
297 Eleven of the 23 lactating women also had longitudinal DXA measurements at the calcaneus. Initial measurements were usually made at 3 or 6 mo postpartum. It was, therefore, only possible to monitor the recovery of DXA calcaneal measurements after peak lactation.
Dual-Energy X-Ray Absorptiometry Bone mineral content (BMC [g]) and bone area (BA [cm2]) of the lumbar spine (L1–L4), nondominant forearm, and calcaneus were measured by DXA (QDR-1000/W, Hologic Inc, Waltham, MA). Subsequent scans were analyzed with reference to the individual’s baseline image using the Compare Facility. The performance mode was used for the spine (software v 4.47P). Software version 5.61 was used for forearm scans. Measurements of spine and forearm were performed as described in the Hologic manual (Hologic Inc, Waltham, MA). For measurements of the left calcaneus, individuals were positioned on their left side on the scanning bed facing the operator. The left leg was bent with the left foot resting against the forearm-positioning device. The base of the foot was parallel to the scanning bed. Using the left forearm scan mode, the laser light was positioned just below and in front of the talus. The scan area included the complete calcaneus plus surrounding soft tissue. Scans were analyzed using the protocol subregion analysis. One rectangular region of interest was analyzed; it included the complete calcaneus (Fig. 1). Attempts to localize calcaneal subregions resulted in inferior precision. Quality assurance and long-term instrument stability were assessed using the Hologic spine phantom. This was scanned at the beginning of each measurement day. During the study, the coefficient of variation (%CV) of phantom measurements for both the BMC and BA was <0.4% and there was no indication of any significant drift over time. Estimates of in vivo precision of BA-adjusted BMC determined from two sets (n = 22) of scans at an interval of approx 3 mo were spine = 0.9% and radial wrist = 1.4%. Short-term precision (%CV) of the calcaneal BA-adjusted BMC from duplicate measurements on the same day with repositioning of the subject between scans (n = 25) was 1.8%.
Quantitative Ultrasound Quantitative ultrasound (QUS) measurements were made using the CUBA Clinical (McCue Ultrasonics, Winchester UK) on the same day as the DXA measurements. This instrument measured BUA (dB/MHz) and velocity of sound VOS (m/s) at the mid-calcaneus. This transmission technique used a pair of ultrasonic transducers in direct contact with the subject’s heel via silicone rubber pads. Water-soluble ultrasound gel was used to ensure acoustic coupling of the pads to the heel. Heel width was determined from the separation of the transducers and the BUA was normalized for heel width to give the normalized BUA (nBUA). The in vivo precision (%CV) of duplicate QUS measurements (n = 24) performed on the same day were nBUA = 7.2% and VOS = 0.9%. Short-term precision (%CV) was optimized by obtaining nine scans of the left foot with three positioning Volume 7, 2004
298
Laskey and Prentice
Fig. 1. Image and bone mineral results of DXA measurement of the calcaneus.
of the calcaneus. Optimized short-term precision determined from duplicate sets of nine measurements (n = 24) were nBUA = 3.6% and VOS = 0.42% (n = 24). Long-term precision (CV%), determined using the optimized protocol from two sets of scans (n = 16) at an interval of approx 6 mo, were nBUA = 3.5% and VOS = 3.4%.
Statistics The BMC was adjusted for bone size by including the BA in the analysis of covariance (ANCOVA) models (10). Investigation of changes over time within individuals was performed by ANCOVA after transformation of continuous variables to natural logarithms. Scheffe’s post hoc test was used to determine the significance of differences between specific pairs of time-points.
Results The characteristics of the women at 0.5 mo postpartum are shown in Table 1. Both breast-feeding and formula-feeding women had lost about 2 kg in weight by 3 mo postpartum. At this time, no significant changes in BMC from 0.5 mo postpartum were seen in the formula-feeding women at the spine (–0.5%; SE = 0.7; p = 0.93) or wrist (–0.2%; SE = 0.9; p = 0.99). In contrast, after 3 mo lactation, decreases in BMC were significant at the spine (p < 0.001), but not at the wrist (p = 0.4). QUS measurements had increased nonsignificantly (nBUA, p = 0.40; VOS, p = 0.61) (Fig. 2). Results showing the recovery of spine BMC after lactation (minimum to maximum spine BMC) and the changes for the other measurements at the corresponding time-points, are shown in Fig. 3. As lactation declined, BMC increased at both the spine and wrist, but changes were only significant at the spine (p < 0.001 and p = 0.17 respectively). Nonsignificant decreases were observed at the calcaneus using QUS (nBUA, Journal of Clinical Densitometry
Table 1 Subject Characteristics at 0.5 mo Postpartum Breast-feeders (n = 30)
Age (yr) Height (m) Weight (kg) Spine BMD (g/cm2) Wrist BMD (g/cm2) VOS (m/s) nBUA (dB/MHz)
Formula-feeders (n = 11)
Mean
SD
Mean
SD
33.6 1.65 68.4 1.052
3.5 0.05 12.6 0.131
29.5 1.67 75.8 1.073
3.4 0.06 24.5 0.171
0.432
0.047
0.454
0.057
1695 90.5
58 20.0
NM NM
NM NM
Abbr: NM, not measured.
p = 0.49; VOS, p = 0.80). The corresponding DXA results for the 11 women with both spine and calcaneal DXA measurements are shown in Fig. 4. This shows that significant increases in BMC, similar to those observed at the spine, were observed in these women.
Discussion This is the first comparison of longitudinal changes in calcaneal bone using both DXA and QUS during lactation. The decrease in BMC during lactation followed by recovery as lactation declines, known to occur at the spine (8,9), were not Volume 7, 2004
Monitoring BMC Changes in Appendicular Skeleton
Fig. 2. Mean (SE) percentage changes during lactation from 0.5 to 3 mo postpartum (pp). DXA spine and wrist measurements were BA-adjusted BMC; calcaneal quantitative ultrasound (QUS) measurements were in decibels per megahertz for BUA and in meters per second for VOS (n = 30). NS = nonsignificant.
Fig. 3. Mean (SE) percentage changes, from time of minimal to maximum spine bone mineral. DXA spine and wrist measurements were BA-adjusted BMC; calcaneal QUS measurements were in decibels per megahertz for BUA and in meters per second for VOS (n = 30). NS = nonsignificant. Journal of Clinical Densitometry
299
Fig. 4. Mean (SE) percentage changes, from time of minimal to maximum spine bone mineral, in BA-adjusted BMC at the spine and calcaneus using DXA (n = 11).
detected by calcaneal QUS. The initial decrease in BMC during the first 3 mo of lactation could not be measured, as initial calcaneal DXA measurements were obtained 3–6 mo postpartum. However, the recovery in BMC, as lactation declined, was detected by calcaneal DXA measurements. These DXA changes at the calcaneus were significant despite the small size of the study. No significant postpartum changes were detected by DXA at either the spine or wrist in the formula-feeding women or at the wrist in breast-feeding women. Previous studies have measured BMC changes at the forearm during pregnancy and lactation (7,9,11). These studies reported no changes at the radial shaft, which is predominantly cortical bone. Reported changes in wrist BMC were variable. Kalkwarf et al. reported no changes (7), whereas Koltoff et al. reported a 2% decrease at the wrist during the second half of pregnancy but no change during lactation (11). Another study only observed significant changes in BMC at the wrist after 6 mo or more of lactation (9). Our analyses indicate that the direction of BMC changes at the wrist and spine are identical, but the magnitude of the change is much smaller at the wrist, probably reflecting its lower trabecular bone content (25%; cf. spine 70%). The ability of DXA, but not QUS, calcaneal measurements to monitor bone mineral changes might be the result of the inferior precision of QUS compared to DXA measurements. Short-term precision for the CUBA Clinical was similar to previous reports (12–14). The precision of nBUA and VOS was greatly improved by taking many repeat measurements. Long-term precision of VOS was inferior to short-term precision. This has been reported previously (15) and might be the result of nonspecific drift, temperature dependence and inadequate quality control procedures of QUS systems (12,16). Volume 7, 2004
300 At present, it is unclear whether QUS can reliably measure longitudinal changes in bone over a short time-span (17). One QUS study was unable to detect the longitudinal changes in bone mineral detected by DXA during hormone replacement therapy (HRT) (18). Other studies have reported a positive response to HRT in studies over 4 yr (19). In agreement with our data, one study detected no change in VOS or BUA during lactation (20). However, other studies have reported decreases in QUS measurements during lactation and pregnancy (21–23). The comparison of QUS studies is complicated by the increasing diversity of commercial QUS devices that measure different peripheral sites. Commercial QUS devices use either the transmission of ultrasound through accessible limb bones (calcaneus and phalanges) or the reflectance of ultrasound waves along the bone surface. Results for apparently similar QUS measurements obtained on one QUS system cannot necessarily be directly translated into performance statements of other technologically different QUS systems. Correlation coefficients vary between 0.4 and 0.8 (13). These are similar to the correlation coefficients for DXA measurements at different skeletal sites (between 0.4 and 0.7) (24) and between the QUS measurements and DXA measurements at different skeletal sites found by us (0.3–0.7, unpublished data) and others (14,17). These moderate correlations indicate that different QUS systems measure different properties of bone (4) and these are not closely related to BMC measured by DXA. During lactation, the decrease in BMC at the spine is associated with higher bone turnover and increase in bone remodeling sites. However, during aging, trabecular bone could thin and then break, resulting in the bone fragility associated with osteoporosis. BUA measurements depend on the trabecular architecture of cancellous bone and might relate to the separation and connectivity of trabeculi. This could explain why QUS measurements are predictive of fracture risk in elderly subjects (25–27) but are poor at monitoring changes in bone mineral in young subjects. The calcaneus has potential advantages as a DXA measurement site. It is unlikely to be affected by artifacts such as deformity and osteophytes commonly seen at the spine of elderly subjects. The effective dose from radiation is very low (<0.1 µSv for pencil-beam DXAs), making it an especially suitable site for studies of children, pregnant women, and other vulnerable subjects. The calcaneus has a high percentage of trabecular bone (>80%), similar to that found in the spine (70%), and is, therefore, likely to show similar high metabolic activity. There are relatively few longitudinal studies comparing changes in BMC at the calcaneus to changes at the spine and other clinically important sites. However, accelerated bone loss has been detected by DXA at both the spine and calcaneus after the menopause (28). DXA measurements of the calcaneus can be made using specially designed equipment (GE Lunar PIXI, Madison, WI; HeelScan DX-2000, KDK, Kyoto, Japan) or conventional DXA equipment can be adapted using modified software as described in this article and other studies (29–31). In conclusion, using lactation as a model of BMC change, DXA but not QUS measurements of the calcaneus demon-
Journal of Clinical Densitometry
Laskey and Prentice strated a significant increase in bone mineral, similar to that observed at the spine. These findings indicate that DXA measurements of the calcaneus might be a superior tool to calcaneal QUS when it is difficult or impossible to monitor changes in the BMC at conventional axial skeletal sites. Further larger studies are required to confirm this observation.
References 1. Fleisch H. 1996 Mechanism of Bone Loss and Modes of Action of Antiresorptive Therapies. In: Osteoporosis 1996. Papapoulos SE, Lips P, Pols HAP, Johnson CC, Delmas PD, eds. Elseveir, Amsterdam, pp. 7–15. 2. Mundy GR, ed. 1999 Bone Remodelling. Philadelphia: Lippincott Williams and Wilkins. 3. Funke M, Kopka L, Vosshenrich R, et al. 1995 Broadband ultrasound attenuation in the diagnosis of osteoporosis: correlation with osteodensitometry and fracture. Radiology 194:77–81. 4. Gluer CC. 1997 Quantitative ultrasound techniques for the assessment of osteoporosis: expert agreement on current status. The International Quantitative Ultrasound Consensus Group. J Bone Miner Res 12:1280–1288. 5. Hadji P, Hars O, Gorke K, Emons G, Schulz KD. 2000 Quantitative ultrasound of the os calcis in postmenopausal women with spine and hip fracture. J Clin Densitom 3:233–239. 6. Laskey MA, Prentice A, Hanratty LA, et al. 1998 Bone changes after 3 mo of lactation: influence of calcium intake, breast-milk output, and vitamin D-receptor genotype. Am J Clin Nutr 67:685–692. 7. Kalkwarf HJ, Specker BL, Bianchi DC, Ranz J, Ho M. 1997 The effect of calcium supplementation on bone density during lactation and after weaning. N Engl J Med 337:523–528. 8. Sowers M, Corton G, Shapiro B, et al. 1993 Changes in bone density with lactation. JAMA 269:3130–3135. 9. Laskey MA, Prentice A. 1999 Bone mineral changes during and after lactation. Obstet Gynecol 94:608–615. 10. 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. 11. Kolthoff N, Eiken P, Kristensen B, Nielsen SP. 1998 Bone mineral changes during pregnancy and lactation: a longitudinal cohort study. Clin Sci (Lond) 94:405–412. 12. Iki M, Kajita E, Mitamura S, Nishino H, Yamagami T, Nagahama N. 1999 Precision of quantitative ultrasound measurement of the heel bone and effects of ambient temperature on the parameters. Osteoporos Int 10:462–467. 13. Stewart A, Reid DM. 2000 Precision of quantitative ultrasound: comparison of three commercial scanners. Bone 27:139–143. 14. Tromp AM, Smit JH, Deeg DJ, Lips P. 1999 Quantitative ultrasound measurements of the tibia and calcaneus in comparison with DXA measurements at various skeletal sites. Osteoporos Int 9:230–235. 15. Ingle BM, Eastell R. 1996 Qualitative ultrasound measurements: short and long-term precision. J Bone Miner Res 11:1830. 16. Krieg MA, Cornuz J, Hartl F, et al. 2002 Quality controls for two heel bone ultrasounds used in the Swiss Evaluation of the Methods of Measurement of Osteoporotic Fracture Risk Study. J Clin Densitom 5:335–341. 17. Greenspan SL, Bouxsein ML, Melton ME, et al. 1997 Precision and discriminatory ability of calcaneal bone assessment technologies. J Bone Miner Res 12:1303–1313. 18. Lees B, Garland SW, Stevenson JC. 1996 Comparison of changes in ultrasound parameters and bone mineral density in
Volume 7, 2004
Monitoring BMC Changes in Appendicular Skeleton
19.
20.
21. 22. 23. 24.
women on hormone replacement therapy. J Bone Miner Res 11:1817. Sahota O, San P, Cawte SA, Pearson D, Hosking DJ. 2000 A comparison of the longitudinal changes in quantitative ultrasound with dual-energy X-ray absorptiometry: the four-year effects of hormone replacement therapy. Osteoporos Int 11:52–58. Yamaga A, Taga M, Minaguchi H, Sato K. 1996 Changes in bone mass as determined by ultrasound and biochemical markers of bone turnover during pregnancy and puerperium: a longitudinal study. J Clin Endocrinol Metab 81:752–756. Sowers MF, Scholl T, Harris L, Jannausch M. 2000 Bone loss in adolescent and adult pregnant women. Obstet Gynecol 96:189–193. Javaid MK, Taylor P, Shore SR, Inskip HM, Godfrey KM, Cooper C. 2001 Longitudinal assessment of bone status during pregnancy using calcaneal ultrasound. Osteoporos Int 12:S28. Giorgino R, Lorusso D, Paparella P. 1996 Pregnancy and lactation: longitudinal evaluation of bone status by ultrasound densitometry. Osteoporos Int 6:187. Sundberg M, Gardsell P, Johnell O, Ornstein E, Sernbo I. 1998 Comparison of quantitative ultrasound measurements in calcaneus with DXA and SXA at other skeletal sites: a populationbased study on 280 children aged 11–16 years. Osteoporos Int 8:410–417.
Journal of Clinical Densitometry
301 25. Thompson P, Taylor J, Fisher A, Oliver R. 1998 Quantitative heel ultrasound in 3180 women between 45 and 75 years of age: compliance, normal ranges and relationship to fracture history. Osteoporos Int 8:211–214. 26. Hans D, Dargent-Molina P, Schott AM, et al. 1996 Ultrasonographic heel measurements to predict hip fracture in elderly women: the EPIDOS prospective study. Lancet 348: 511–514. 27. Njeh CF, Kuo CW, Langton CM, Atrah HI, Boivin CM. 1997 Prediction of human femoral bone strength using ultrasound velocity and BMD: an in vitro study. Osteoporos Int 7:471–477. 28. Ito M, Nakamura T, Tsurusaki K, Uetani M, Hayashi K. 1999 Effects of menopause on age-dependent bone loss in the axial and appendicular skeletons in healthy Japanese women. Osteoporos Int 10:377–383. 29. Kang C, Speller R. 1999 The effect of region of interest selection on dual energy X-ray absorptiometry measurements of the calcaneus in 55 post-menopausal women. Br J Radiol 72:864–871. 30. Yamada M, Ito M, Hayashi K, Nakamura T. 1993 Calcaneus as a site for assessment of bone mineral density: evaluation in cadavers and healthy volunteers. Am J Roentgenol 161:621–627. 31. Rees S, McCoy K, Allen J, Taggart H. 2001 Measurement of bone density on the foot using dual energy x-ray absorptiometry. Osteoporosis Int 12:S27.
Volume 7, 2004