Chapter 4. Discussion
4.4 Conclusions
This study, for the first time, identified metabolome for predicting the risk of
low BMD in postmenopausal women. Metabolomics can quantify a large-scale of
metabolites to characterize response to drugs, diet, lifestyle, environmental stimuli and
genetic modulations.15 Future studies exploring gene-metabolite interactions using
liquid chromatograph/mass spectrometry for low abundant metabolites and animal
studies are warranted to shed light on the association between metabolome and BMD
in human.
20
References
1. Prentice A. Diet, nutrition and the prevention of osteoporosis. Public Health Nutr
2004;7:227-43.
2. Dawson-Hughes B, Looker AC, Tosteson AN, Johansson H, Kanis JA, Melton
LJ, 3rd. The potential impact of the National Osteoporosis Foundation guidance
on treatment eligibility in the USA: an update in NHANES 2005-2008.
Osteoporosis international : a journal established as result of cooperation
between the European Foundation for Osteoporosis and the National
Osteoporosis Foundation of the USA 2012;23:811-20.
3. Lin YC, Pan WH. Bone mineral density in adults in Taiwan: results of the
Nutrition and Health Survey in Taiwan 2005-2008 (NAHSIT 2005-2008). Asia
Pac J Clin Nutr 2011;20:283-91.
4. Raisz LG. Clinical practice. Screening for osteoporosis. N Engl J Med
2005;353:164-71.
5. Kanis JA. Diagnosis of osteoporosis and assessment of fracture risk. Lancet
2002;359:1929-36.
6. Garnero P, Hausherr E, Chapuy MC, Marcelli C, Grandjean H, Muller C,
Cormier C, Breart G, Meunier PJ, Delmas PD. Markers of bone resorption
predict hip fracture in elderly women: the EPIDOS Prospective Study. J Bone
21
Miner Res 1996;11:1531-8.
7. Garnero P, Mulleman D, Munoz F, Sornay-Rendu E, Delmas PD. Long-term
variability of markers of bone turnover in postmenopausal women and
implications for their clinical use: the OFELY study. J Bone Miner Res
2003;18:1789-94.
8. McLean RR, Jacques PF, Selhub J, Tucker KL, Samelson EJ, Broe KE, Hannan
MT, Cupples LA, Kiel DP. Homocysteine as a predictive factor for hip fracture
in older persons. N Engl J Med 2004;350:2042-9.
9. van Meurs JB, Dhonukshe-Rutten RA, Pluijm SM, van der Klift M, de Jonge R,
Lindemans J, de Groot LC, Hofman A, Witteman JC, van Leeuwen JP, Breteler
MM, Lips P, Pols HA, Uitterlinden AG. Homocysteine levels and the risk of
osteoporotic fracture. N Engl J Med 2004;350:2033-41.
10. Leboff MS, Narweker R, LaCroix A, Wu L, Jackson R, Lee J, Bauer DC, Cauley
J, Kooperberg C, Lewis C, Thomas AM, Cummings S. Homocysteine levels and
risk of hip fracture in postmenopausal women. J Clin Endocrinol Metab
2009;94:1207-13.
11. Enneman AW, van der Velde N, de Jonge R, Heil SG, Stolk L, Hofman A,
Rivadeneira F, Zillikens MC, Uitterlinden AG, van Meurs JB. The association
between plasma homocysteine levels, methylation capacity and incident
22
osteoporotic fractures. Bone 2012.
12. Gjesdal CG, Vollset SE, Ueland PM, Refsum H, Drevon CA, Gjessing HK, Tell
GS. Plasma total homocysteine level and bone mineral density: the Hordaland
Homocysteine Study. Arch Intern Med 2006;166:88-94.
13. Jimi E, Aoki K, Saito H, D'Acquisto F, May MJ, Nakamura I, Sudo T, Kojima T,
Okamoto F, Fukushima H, Okabe K, Ohya K, Ghosh S. Selective inhibition of
NF-kappa B blocks osteoclastogenesis and prevents inflammatory bone
destruction in vivo. Nat Med 2004;10:617-24.
14. Yuan J, Akiyama M, Nakahama K, Sato T, Uematsu H, Morita I. The effects of
polyunsaturated fatty acids and their metabolites on osteoclastogenesis in vitro.
Prostaglandins Other Lipid Mediat 2010;92:85-90.
15. Beckonert O, Keun HC, Ebbels TM, Bundy J, Holmes E, Lindon JC, Nicholson
JK. Metabolic profiling, metabolomic and metabonomic procedures for NMR
spectroscopy of urine, plasma, serum and tissue extracts. Nat Protoc
2007;2:2692-703.
16. Nicholson JK, Foxall PJ, Spraul M, Farrant RD, Lindon JC. 750 MHz 1H and
1H-13C NMR spectroscopy of human blood plasma. Anal Chem
1995;67:793-811.
17. Nicholson G, Rantalainen M, Maher AD, Li JV, Malmodin D, Ahmadi KR,
23
Faber JH, Hallgrimsdottir IB, Barrett A, Toft H, Krestyaninova M, Viksna J,
Neogi SG, Dumas ME, Sarkans U, The Molpage C, Silverman BW, Donnelly P,
Nicholson JK, Allen M, Zondervan KT, Lindon JC, Spector TD, McCarthy MI,
Holmes E, Baunsgaard D, Holmes CC. Human metabolic profiles are stably
controlled by genetic and environmental variation. Mol Syst Biol 2011;7:525.
18. Keun HC, Ebbels TM, Antti H, Bollard ME, Beckonert O, Schlotterbeck G,
Senn H, Niederhauser U, Holmes E, Lindon JC, Nicholson JK. Analytical
reproducibility in 1H NMR-based metabonomic urinalysis. Chem Res Toxicol
2002;15:1380-6.
19. Dumas ME, Maibaum EC, Teague C, Ueshima H, Zhou B, Lindon JC,
Nicholson JK, Stamler J, Elliott P, Chan Q, Holmes E. Assessment of analytical
reproducibility of 1H NMR spectroscopy based metabonomics for large-scale
epidemiological research: the INTERMAP Study. Analytical chemistry
2006;78:2199-208.
20. Nicholson JK, Lindon JC, Holmes E. 'Metabonomics': understanding the
metabolic responses of living systems to pathophysiological stimuli via
multivariate statistical analysis of biological NMR spectroscopic data.
Xenobiotica 1999;29:1181-9.
21. Fiehn O. Metabolomics--the link between genotypes and phenotypes. Plant Mol
24
Biol 2002;48:155-71.
22. Gibney MJ, Walsh M, Brennan L, Roche HM, German B, van Ommen B.
Metabolomics in human nutrition: opportunities and challenges. Am J Clin Nutr
2005;82:497-503.
23. Long WF, Li L, Chen HQ, Tang Y, He XL, Jing RZ. 1H-NMR-based
metabonomics analysis of plasma from osteoporotic rats induced by ovariectomy.
Sichuan Da Xue Xue Bao Yi Xue Ban 2009;40:843-7.
24. Xue L, Wang Y, Liu L, Zhao L, Han T, Zhang Q, Qin L. A 1HNMR-Based
Metabonomics Study of Postmenopausal Osteoporosis and Intervention Effects
of Er-Xian Decoction in Ovariectomized Rats. Int J Mol Sci 2011;12:7635-51.
25. Lerner UH. Bone remodeling in post-menopausal osteoporosis. J Dent Res
2006;85:584-95.
26. Viant MR. Improved methods for the acquisition and interpretation of NMR
metabolomic data. Biochem Biophys Res Commun 2003;310:943-8.
27. Barton RH, Waterman D, Bonner FW, Holmes E, Clarke R, Nicholson JK,
Lindon JC. The influence of EDTA and citrate anticoagulant addition to human
plasma on information recovery from NMR-based metabolic profiling studies.
Mol Biosyst 2010;6:215-24.
28. Purohit PV, Rocke DM, Viant MR, Woodruff DL. Discrimination models using
25
variance-stabilizing transformation of metabolomic NMR data. OMICS
2004;8:118-30.
29. Eriksson L, Johansson E, Kettaneh-Wold N, Wold S. Introduction to multi-
andmegavariate dta analysis using projection methods (PCA & PLS): Umetrics;
1999.
30. van den Berg RA, Hoefsloot HC, Westerhuis JA, Smilde AK, van der Werf MJ.
Centering, scaling, and transformations: improving the biological information
content of metabolomics data. BMC Genomics 2006;7:142.
31. Xia J, Psychogios N, Young N, Wishart DS. MetaboAnalyst: a web server for
metabolomic data analysis and interpretation. Nucleic Acids Res
2009;37:W652-60.
32. Manolagas SC. From estrogen-centric to aging and oxidative stress: a revised
perspective of the pathogenesis of osteoporosis. Endocr Rev 2010;31:266-300.
33. Cohen SA. Amino acid analysis using precolumn derivatization with
6-aminoquinolyl-N-hydroxysuccinimidyl carbamate. Methods Mol Biol
2000;159:39-47.
34. Sengupta S, Park SH, Patel A, Carn J, Lee K, Kaplan DL. Hypoxia and amino
acid supplementation synergistically promote the osteogenesis of human
mesenchymal stem cells on silk protein scaffolds. Tissue Eng Part A
26
2010;16:3623-34.
35. Frayn KN. Calculation of substrate oxidation rates in vivo from gaseous
exchange. J Appl Physiol 1983;55:628-34.
36. Owen OE, Trapp VE, Skutches CL, Mozzoli MA, Hoeldtke RD, Boden G,
Reichard GA, Jr. Acetone metabolism during diabetic ketoacidosis. Diabetes
1982;31:242-8.
37. Vestergaard P. Discrepancies in bone mineral density and fracture risk in patients
with type 1 and type 2 diabetes--a meta-analysis. Osteoporos Int
2007;18:427-44.
38. Shan PF, Wu XP, Zhang H, Cao XZ, Gu W, Deng XG, Gu C, Liao EY. Bone
mineral density and its relationship with body mass index in postmenopausal
women with type 2 diabetes mellitus in mainland China. J Bone Miner Metab
2009;27:190-7.
39. Ma L, Oei L, Jiang L, Estrada K, Chen H, Wang Z, Yu Q, Zillikens MC, Gao X,
Rivadeneira F. Association between bone mineral density and type 2 diabetes
mellitus: a meta-analysis of observational studies. Eur J Epidemiol
2012;27:319-32.
40. Adami S. Bone health in diabetes: considerations for clinical management. Curr
Med Res Opin 2009;25:1057-72.
27
41. Felson DT, Zhang Y, Hannan MT, Anderson JJ. Effects of weight and body mass
index on bone mineral density in men and women: the Framingham study. J
Bone Miner Res 1993;8:567-73.
42. Frederich RC, Hamann A, Anderson S, Lollmann B, Lowell BB, Flier JS. Leptin
levels reflect body lipid content in mice: evidence for diet-induced resistance to
leptin action. Nat Med 1995;1:1311-4.
43. Thomas T, Gori F, Khosla S, Jensen MD, Burguera B, Riggs BL. Leptin acts on
human marrow stromal cells to enhance differentiation to osteoblasts and to
inhibit differentiation to adipocytes. Endocrinology 1999;140:1630-8.
44. Martin A, de Vittoris R, David V, Moraes R, Begeot M, Lafage-Proust MH,
Alexandre C, Vico L, Thomas T. Leptin modulates both resorption and formation
while preventing disuse-induced bone loss in tail-suspended female rats.
Endocrinology 2005;146:3652-9.
45. Adami S, Braga V, Zamboni M, Gatti D, Rossini M, Bakri J, Battaglia E.
Relationship between lipids and bone mass in 2 cohorts of healthy women and
men. Calcif Tissue Int 2004;74:136-42.
46. Brownbill RA, Ilich JZ. Lipid profile and bone paradox: higher serum lipids are
associated with higher bone mineral density in postmenopausal women. J
Womens Health (Larchmt) 2006;15:261-70.
28
47. Yamaguchi T, Sugimoto T, Yano S, Yamauchi M, Sowa H, Chen Q, Chihara K.
Plasma lipids and osteoporosis in postmenopausal women. Endocr J
2002;49:211-7.
48. Cui LH, Shin MH, Chung EK, Lee YH, Kweon SS, Park KS, Choi JS.
Association between bone mineral densities and serum lipid profiles of pre- and
post-menopausal rural women in South Korea. Osteoporos Int 2005;16:1975-81.
49. Orozco P. Atherogenic lipid profile and elevated lipoprotein (a) are associated
with lower bone mineral density in early postmenopausal overweight women.
Eur J Epidemiol 2004;19:1105-12.
50. Serre CM, Farlay D, Delmas PD, Chenu C. Evidence for a dense and intimate
innervation of the bone tissue, including glutamate-containing fibers. Bone
1999;25:623-9.
51. Skerry TM, Genever PG. Glutamate signalling in non-neuronal tissues. Trends
Pharmacol Sci 2001;22:174-81.
52. Chenu C, Serre CM, Raynal C, Burt-Pichat B, Delmas PD. Glutamate receptors
are expressed by bone cells and are involved in bone resorption. Bone
1998;22:295-9.
53. Behl C, Skutella T, Lezoualc'h F, Post A, Widmann M, Newton CJ, Holsboer F.
Neuroprotection against oxidative stress by estrogens: structure-activity
29
relationship. Mol Pharmacol 1997;51:535-41.
54. Singer CA, Figueroa-Masot XA, Batchelor RH, Dorsa DM. The
mitogen-activated protein kinase pathway mediates estrogen neuroprotection
after glutamate toxicity in primary cortical neurons. J Neurosci
1999;19:2455-63.
55. Zhu X, Liu X, He P, Cao B, Lv Y, Zhang W, Ni X. Metabolomics in serum of
ovariectomised rats and those exposed to 17beta-oestradiol and genistein.
Gynecol Endocrinol 2010;26:760-7.
56. Ma B, Zhang Q, Wang GJ, A JY, Wu D, Liu Y, Cao B, Liu LS, Hu YY, Wang YL,
Zheng YY. GC-TOF/MS-based metabolomic profiling of estrogen
deficiency-induced obesity in ovariectomized rats. Acta Pharmacol Sin
2011;32:270-8.
57. Rapuri PB, Kinyamu HK, Gallagher JC, Haynatzka V. Seasonal changes in
calciotropic hormones, bone markers, and bone mineral density in elderly
women. J Clin Endocrinol Metab 2002;87:2024-32.
58. Frankenfeld CL, McTiernan A, Thomas WK, LaCroix K, McVarish L, Holt VL,
Schwartz SM, Lampe JW. Postmenopausal bone mineral density in relation to
soy isoflavone-metabolizing phenotypes. Maturitas 2006;53:315-24.
30
Figure 1. Flowchart of participant recruitment Participants for statistical
analyses (610)
Lack of blood sample (45) or NMR data (30)
or NMR (30)
Lack of BMD data at spine (51)
Take steroid (3) or have hormone replacement therapy (34)
Women aged 40-55 (773)
31
s o.138
Figure 2. The tertiles of bone mineral density
T1: BMD < 1.138 g/cm2; T2: 1.138 g/cm2 ≤ BMD <1.258 g/cm2; T3: BMD ≥ 1.258 g/cm2.
1.138 1.258
0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
0 5 10 15 20 25 30
Percent
Bone mineral density
Bone Mineral Density (g/cm2)
32
Figure 3. PCA score plots from the analysis of CPMG NMR spectra using women plasma samples
(a) High BMD group: n=399; low BMD group: n=211
(b) Premenopausal women. High BMD group: n=349; low BMD group: n=134 (c) Postmenopausal women. High BMD group: n=45; low BMD group: n=77
High BMD indicates the 2nd and 3rd tertiles of BMD; low BMD indicates the 1st tertile of BMD.
Abbreviations: PCA, principle components analysis; CPMG, Carr-Purcell-Meiboom-Gill;
BMD, bone mineral density.
33
Figure 4. PLS-DA score plots from the analysis of CPMG NMR spectra using women plasma samples
(a) High BMD group: n=399; low BMD group: n=211
(b) Premenopausal women. High BMD group: n=349; low BMD group: n=134 (c) Postmenopausal women. High BMD group: n=45; low BMD group: n=77
High BMD indicates the 2nd and 3rd tertiles of BMD; low BMD indicates the 1st tertile of BMD.
Abbreviations: PCA, principle components analysis; CPMG, Carr-Purcell-Meiboom-Gill;
BMD, bone mineral density.
34
Figure 5. Receiver operating characteristic curves of comparing models for classification of high and low BMD
(a) All participants. Model 1: AUC= 0.59 (95% CI=0.55-0.64), Model 2: AUC=0.60 (95%
CI=0.55-0.65). Glutamine alone: AUC=0.54 (95% CI=0.50-0.59).
(b) Premenopausal women. Model 1: AUC=0.57 (95% CI= 0.52-0.63), Model 2: AUC=0.60 (95% CI=0.55-0.66). Glutamine alone: AUC=0.53 (95% CI: 0.48-0.58).
(c) Postmenopausal women. Model 1: AUC=0.69 (95% CI: 0.67-0.72), Model 2: AUC=0.70 (95% CI: 0.67-0.73). Glutamine alone: AUC=0.65 (95% CI: 0.63-0.68).
Model 1: lactate, acetone, lipid, VLDL, and glutamine.
Model 2: Model 1 plus acetate and glucose.
High BMD indicates the 2nd and 3rd tertiles of BMD; low BMD indicates the 1st tertile of BMD.
Abbreviations: AUC, area under the curve; BMD, bone mineral density; CI, confidence interval;
VLDL, very low density lipoprotein.
35
Figure 6. Acetone distribution by fasting glucose level
(High:glucose> 100 mg/dL; Low: glucose ≤ 100 mg/dL)
≤
≤
36
Figure 7. Postulated mechanism relates important metabolites with bone
mineral density among postmenopausal women
Acetone is referred to ketone bodies which may be generated during diabetic ketosis36, and then promote to release hormonal factors that suppresses estrogen deficiency and the subsequent bone loss.37-41 Lactate may against bone loss via regulating collagen biosynthesis during osteogenesis.34 Lipid and very low density lipoprotein as well as lipid profiles can be reflected by leptin level which is related to differentiation of stromal cells to osteoblasts and thus prevent further loss of bone mass.42-44 Bone resorption may occur when glutamine interconverts to excess amount of glutamate and then cause bone loss as the composition of osteoclast expresses glutamate receptors.5237
Table 1. Characteristics of the study population
Variables BMD Waist circumference (cm) 73.6 (0.35) 71.6 (0.44) 0.0008 Diastolic blood pressure (mmHg) 64.9 (0.51) 63.7 (0.64) 0.13
Fasting glucose (mg/dL) 101.4 (1.16) 97.2 (0.61) 0.01
Triglycerides (mg/dL) 94.1 (2.70) 87.3 (3.10) 0.10
High-density lipoprotein cholesterol (mg/dL)
69.8 (0.77) 72.7 (1.25) 0.05
Total cholesterol (mg/dL) 198.4 (1.66) 202.7 (2.27) 0.13 Alkaline phosphatase (IU) 60.2 (0.84) 67.3 (1.20) <0.0001
Creatinine (mg/dL) 0.80 (0.005) 0.78 (0.006) 0.03
Abbreviation: BMD, bone mineral density.
*P-values were obtained from Student’s t tests (normally distributed continuous variables) and Chi-square tests (categorical variables) by comparing participants with high and low BMD.
Number in bold indicates statistically significant finding.
38
Table 2. The PLS-DA parameters and permutation test for differentiating high and low BMD levels
Comparison Groups
PLS-DA parameters Ppermutation testc
No.b R2 Q2†
Abbreviations: PLS-DA, partial least squares-discriminant analysis; BMD, bone mineral density.
a High BMD indicates the 2nd and 3rd tertiles of BMD; low BMD indicates the 1st tertile of BMD.
b The number of components based on Q2 indicates the best classifier of PLS-DA using 10-fold cross-validation method.
c Distribute 1000 permutations.
†Predictive capability.
Number in bold indicates statistically significant finding.
39
Table 3. The change of plasma metabolites in postmenopausal women to distinguish high and low BMD
Metabolites Multiplicitya Signal assignmentb δH (ppm) Abbreviations: BMD, bone mineral density; VLDL, very low density lipoprotein; δH, chemical shift; ppm, parts per million.
a s,singlet; d, doublet; t, triplet; m, complex multiplet; ddd, doublet of doublets of doublets.
b Signal assignment provides precise bioanalytical and dynamic information.
c All metabolites with variable importance in projection (VIP) score >1.5.
d High BMD indicates the 2nd and 3rd tertiles of BMD; low BMD indicates the 1st tertile of BMD.
*Pathway information was obtained from KEGG PATHWAY Database (http://www.genome.jp/kegg/).
40
Table 4. Association between plasma metabolites and bone mineral density (T1 vs. T2 + T3)
Abbreviations: AOR, adjusted odds ratio; BMD, bone mineral density; VLDL, very low density lipoprotein; CI, confidence intervals.
a All metabolites with variable importance in projection (VIP) score >1.5.
b High BMD indicates the 2nd and 3rd tertiles of BMD; low BMD indicates the 1st tertile of BMD.
c All models were adjusted for age (continuous), body weight (continuous), height (continuous), waist circumference (continuous), menopausal status (yes/no), creatinine (> median, 79 mg/dL, yes/no), regular exercise (≥2 times/week, yes/no), serum alkaline phosphatase (≥ median, 60 IU, yes/no).
Number in bold indicates statistically significant finding.
Metabolitesa
41
Table 5. Association between plasma metabolites and bone mineral density stratified by menopausal status (T1 vs. T2+T3)
Abbreviations: AOR, adjusted odds ratio; BMD, bone mineral density; VLDL, very low density lipoprotein; CI, confidence intervals.
a All metabolites with variable importance in projection (VIP) score >1.5.
b High BMD indicates the 2nd and 3rd tertiles of BMD; low BMD indicates the 1st tertile of BMD.
c All models were adjusted for age (continuous), body weight (continuous), height (continuous), waist circumference (continuous), menopausal status (yes/no), creatinine (> median, 79 mg/dL, yes/no), regular exercise ≥2 times/week (yes/no), serum alkaline phosphatase (≥ median, 60 UI, yes/no).
Number in bold indicates statistically significant finding.
Metabolitesa
Premenopause Postmenopause p interaction
Low BMD/High BMDb Glutamine 1.32 (0.90-1.95) 1.24 (0.78-1.96) 2.88 (1.45-5.72) 6.04 (1.57-23.21) 0.04 Glucose 0.96 (0.90-1.03) 0.97 (0.90-1.04) 0.95 (0.85-1.06) 1.14 (0.92-1.40) 0.38
42
Table 6. Receiver operating characteristic contrast tests of pairwise comparison between different models to classify high and low BMD
a Model 1 included lactate, acetone, lipid, VLDL, and glutamine; model 2 included variables in model 1 plus acetate and glucose.
*P-values were obtained from Chi-square tests.
Number in bold indicates statistically significant finding.
Contrast models All participants Premenopause Postmenopause P value*
Model 2 vs. Mode 1a 0.38 0.16 0.74
Glutamine alone vs. Model 1 0.05 0.12 0.03
Glutamine alone vs. Model 2 0.03 0.02 0.01
43
Table 7. Model comparisons for the association between plasma metabolites and bone mineral density (T1 vs. T2+T3)
Lactate Acetone Acetate Lipid VLDL Glutamine Glucose
Low BMDa (399) / High BMDa (211) AOR (95% CI)
Model 1b 0.92 (0.81-1.04) 0.90 (0.79-1.02) 0.42 (0.97-1.03) 1.08 (0.44-2.66) 0.88 (0.75-1.04) 1.47 (1.07-2.03) 0.96 (0.91-1.02) Model 2c 0.86 (0.76-0.99) 0.82 (0.71-0.95) 1.55 (1.05-2.30) 0.56 (0.21-1.48) 0.80 (0.67-0.95) 1.67 (1.19-2.35) 0.96 (0.91-1.02) Model 3d 0.99 (0.86-1.15) 0.95 (0.82-1.10) 1.07 (0.70-1.63) 1.04 (0.37-2.90) 0.93 (0.77-1.11) 1.22 (0.84-1.76) 0.99 (0.93-1.06) Model 4e 0.94 (0.80-1.09) 0.92 (0.78-1.08) 1.17 (0.74-1.83) 0.85 (0.29-2.51) 0.89 (0.74-1.09) 1.40 (0.94-2.09) 1.00 (0.93-1.07) Model 5f 0.90 (0.76-1.05) 0.87 (0.74-1.03) 1.32 (0.82-2.10) 0.72 (0.24-2.12) 0.85 (0.69-1.04) 1.55 (1.03-2.33) 0.99 (0.93-1.06) Abbreviations: AOR, adjusted odds ratio; BMD, bone mineral density; VLDL, very low density lipoprotein; CI, confidence intervals.
a High BMD indicates the 2nd and 3rd tertiles of BMD; low BMD indicates the 1st tertile of BMD.
b Model 1: Unadjusted model.
c Model 2: Adjusted for age (continuous), menopausal status (yes/no).
d Model 3: Adjusted variables in model 2 plus body weight (continuous) and height (continuous)
e Model 4: Adjusted variables in model 3 plus serum alkaline phosphatase (≥ median, 60 IU, yes/no) and regular exercise (≥2 times/week, yes/no)
f Model 5: Adjusted variables in model 4 plus waist circumference (continuous) and creatinine (> median, 79 mg/dL, yes/no).