ORIGINAL ARTICLE. Abstract. Introduction

Transcription

1 ORIGINAL ARTICLE Response of Serum Carboxylated and Undercarboxylated Osteocalcin to Risedronate Monotherapy and Combined Therapy with Vitamin K2 in Corticosteroid-Treated Patients: A Pilot Study Takayoshi Hozuki, Tomihiro Imai, Emiko Tsuda, Akihiro Matsumura, Daisuke Yamamoto, Takanobu Toyoshima, Syuuichiro Suzuki, Rika Yamauchi, Takashi Hayashi, Shin Hisahara and Shun Shimohama Abstract Objective The aim of this study was to investigate the responses of serum osteocalcin (OC), undercarboxylated osteocalcin (ucoc) and N-terminal telopeptide of type I collagen (NTx) to corticosteroids, and to examine the effects of risedronate therapy with or without vitamin K2 supplementation on bone metabolic markers in corticosteroid-treated patients. Methods Sixteen patients on corticosteroid therapy for neuromuscular disorders were assigned randomly to 2 groups (A: risedronate monotherapy, n=8; B: combined risedronate and vitamin K2 therapy, n=8) and treated for 1 year. Another 6 patients who received intravenous steroid pulse therapy were assigned to group C for investigation of the effects of corticosteroids on OC and ucoc 1 month after pulse therapy. Results Serial measurements revealed that significant decreases of OC, ucoc and NTx persisted with a similar time course profile during 1 year of treatment in groups A and B, and between-group analysis failed to demonstrate any additional effects of vitamin K2 on risedronate therapy. Intravenous steroid pulse therapy induced a transient depression of OC and ucoc within 1 week in group C. Conclusion These results indicate that serum concentrations of OC and ucoc become consistently low during corticosteroid administration despite risedronate therapy with or without vitamin K2 supplementation, and the serum ucoc level may not be a reliable indicator of vitamin K status under corticosteroid administration. Key words: vitamin K2, risedronate, corticosteroid, osteocalcin, undercarboxylated osteocalcin, N-terminal telopeptide of type I collagen (Inter Med 49: , 2010) () Introduction Vitamin K is a cofactor of γ-carboxylase that mediates the conversion of undercarboxylated osteocalcin (ucoc) to osteocalcin (OC) by transforming glutamyl (Glu) residues of OC to γ-carboxyglutamic acid (Gla) (1, 2). Recent in vivo and in vitro studies showed that vitamin K and its analogues have a therapeutic effect on bone metabolism. In particular, menatetrenone, vitamin K2 with four isoprene units, has been reported to be effective in promoting bone formation in postmenopausal women (3, 4), improving remodeling in hemodialysis patients (5), and preventing the occurrence of new fractures in osteoporotic patients (6). These reports led us to expect that vitamin K2 may also prevent corticosteroidinduced osteoporosis (7, 8). In fact, in vivo experiments using rats have suggested that vitamin K2 may prevent prednisolone-induced bone loss (9, 10). In 2004, the Japanese Society for Bone and Mineral Research established guidelines on the management and treat- Department of Neurology, Sapporo Medical University School of Medicine, Sapporo Received for publication June 4, 2009; Accepted for publication November 4, 2009 Correspondence to Dr. Tomihiro Imai, toimai@sapmed.ac.jp 371

2 ment of corticosteroid-induced osteoporosis, and recommended vitamin K2 as a second-line therapy to be used in combination with bisphosphonates as the first-line therapy (11). However, there have been few attempts to evaluate the effect of vitamin K2 on bone metabolism in corticosteroid-treated patients. In the present study, we performed serial measurements of serum bone metabolic markers such as OC, ucoc and N- terminal telopeptide of type I collagen (NTx) to investigate the additive effects of vitamin K2 on bone metabolism, when used in combination with risedronate in corticosteroidtreated patients with neuromuscular disorders. Patients Patients and Methods Between March 2004 and September 2007, we enrolled 23 patients for a long-term observation study and 6 patients for a short-term observation study at Sapporo Medical University Hospital. This study was approved by the local ethical committee, and written informed consent was obtained from all participants. Long-term observation Twenty-three patients were enrolled in the long-term observation. The inclusion criteria were: 1) patients with corticosteroid-treatable neuromuscular diseases; 2) patients who had never received corticosteroid therapy; 3) patients who would receive corticosteroid treatment for at least one year after initiation of the observation; and 4) patients who agreed to receive prophylactic therapy for corticosteroidinduced osteoporosis according to the Japanese guidelines (11). During the study period, 7 of 23 patients were excluded from observation because of difficulty in oral intake (n=3), gastrointestinal problems (n=3) or liver dysfunction (n=1). Eventually the data of 16 patients who completed the long-term observation were analyzed. Their neuromuscular disorders included polymyositis (n=2), polymyalgia rheumatica (n=1), neuralgic amyotrophy (n=4), chronic inflammatory demyelinating polyneuropathy (n=4), myasthenia gravis (n=3), and multiple sclerosis (n=2). Of16patients,8(3malesand5females)aged29to77 years (mean, 54.9 years) were assigned randomly to receive risedronate monotherapy (group A) and 8 (2 males and 6 females) aged 34 to 73 years (mean, 53.6 years) were assigned to receive combined risedronate (2.5 mg/day) and menatetrenone (45 mg/day) therapy (group B). Treatment with risedronate and/or menatetrenone was started within 8 days after corticosteroid therapy was initiated, and continued for 1 year. Serial measurements of serum OC, ucoc and NTx were performed before, at 1 week, 1 month, 3 months and 1 year after the start of treatment with corticosteroid. The ucoc/ OC ratio was calculated from the OC and ucoc levels. The percent changes of these markers from the baseline at each measurement were calculated and compared between groups AandB. Short-term observation This study was performed to evaluate the short-term effects of corticosteroids on serum OC and ucoc levels, because it is known that serum OC concentrations are significantly decreased within one week after the start of oral steroid intake (12). Six patients (3 males and 3 females) aged 20 to 80 years (mean, 44.1 years) were assigned to group C, and received intravenous steroid pulse therapy (methylprednisolone 1,000 mg/day for 3 days) to treat demyelinating diseases of the central nervous system, such as multiple sclerosis, without any prophylactic therapy for osteoporosis. No tapering corticosteroid therapy was given after the steroid pulse therapy. Serial measurements of serum OC and ucoc were performed before, at 1 week, and 1 month after the completion of one cycle of steroid pulse therapy. Percent changes of OC and ucoc from the baseline were calculated at each time point. Measurement of bone metabolic markers Blood samples were centrifuged at 3,000 rpm for 10 minutes at 4. The separated serum samples were stored frozen at -80 or lower until analysis. Measurements of bone metabolic markers were performed using an immunoradiometric assay (BGP; Mitsubishi Chemical Medience, Tokyo, Japan) for OC, an electrochemiluminescence immunoassay (Picolumi; Sanko Junyaku, Co., Ltd., Tokyo, Japan) for ucoc, and an enzyme-linked immunosorbent assay (Osteomark; Inverness Medical, Tokyo, Japan) for NTx. Statistical analysis Results are expressed as mean ± SE. Wilcoxon matchedpairs signed ranks test was used to determine whether there was a significant change compared to the baseline occurred at each observation time within a group. Between-group differences were assessed by the Mann-Whitney U test. A probability <0.05 was recognized as statistical significance. The JMP statistical program (SAS Institute Inc., Cary, NC) was used for data analysis. Baseline characteristics Results Table 1 lists the baseline characteristics and biochemical parameters of the patients. There were no significant differences in baseline age, dose of prednisolone, bone turnover markers and bone mineral density (BMD) between groups A and B in the long-term study. In the long-term observation, the mean time of initiation of risedronate and/or menatetrenone was 3.5 days in group A and 2.3 days in group B. We tried to start the prophylactic therapy as early as possible to assess the effect of risedronate and/or menatetrenone on changes in bone metabolic 372

3 Table1. CharacteristicsofthePatientsatBaseline Long-term observation Short-term observation Group A Group B P value Group C Age (years) Prednisolone (mg/day) Serum OC (ng/ml) Serum ucoc (ng/ml) Serum NTx (nmol BCE/L) ucoc/oc Lumbar (L2-L4) BMD (g/cm 2 ) Total femur BMD (g/cm 2 ) All the data are expressed mean SE. There were no significant differences between groups A and B on the basis of their background data. OC, osteocalcin; ucoc, undercarboxylated osteocalcin; NTx, N-terminal telopeptide of type I collagen ; BMD, bone mineral density Figure1. Oralprednisolonedoseafterstartingrisedronate monotherapy(group A;closed circles)orcombined risedronateandvitamink2therapy(groupb;opencircles).data aremean±se.between-groupanalysisshowednosignificant diferencesataltimepointsalthoughthemeanvalueswere relativelyhighingroupaupto3months. Figure2. Percentchangesofserum osteocalcin(oc)level afterstartingrisedronatemonotherapy(groupa;closedcir cles)orcombinedrisedronateandvitamink2therapy(group B;opencircles).Dataaremean±SE.Theasterisksindicate significantdiferencescomparedwithbaseline(p<0.05). markers, because corticosteroids, even prednisolone at the low-dose of 5 mg/day, could suppress bone metabolic markers as rapidly as within 2 weeks (13). Changes of prednisolone dose in the long-term observation The prednisolone doses (mg/day) at 1 week, 1 month, 3 months and 12 months were 36.3±5.0, 28.6±4.2, 16.6±3.7 and 9.4±3.2, respectively, in group A, and 32.2±3.8, 21.3± 3.4, 16.6±2.8 and 11.0±4.5 in group B (Fig. 1). There were no significant differences between the two groups at all time points throughout the entire observation period. Changes of BMD in the long-term observation The lumbar and femur BMD (g/cm 2 ) at 12 months after treatment were 1.068±0.075 and 1.075±0.063, respectively, in group A, and 0.935±0.061 and 0.901±0.060 in group B. In both groups, there were no significant changes in lumbar and femur BMD compared to baseline values (Table 1). Changes of OC and ucoc in the long-term observation The percent changes of OC from the baseline at 1 week, 1 month, 3 months and 12 months were -68.2±3.9, -70.0± 4.5, -48.2±10.9 and -36.9±10.2%, respectively, in group A, and -59.1±6.4, -48.4±12.0, -39.7±14.5 and -32.3±12.5% in group B (Fig. 2), while the corresponding values of ucoc were -71.1±10.1, -74.9±6.3, -63.8±9.8 and -38.9±16.6% in group A, and -65.2±8.2, -72.8±6.3, -64.3±14.3 and -64.2± 12.5% in group B (Fig. 3). A significant decrease of OC persisted until 12 months in both groups (Fig. 2). On the other hand, the significant change of ucoc disappeared at 12 months in group A, although a significant decrease below -60% persisted throughout the entire course of study in group B (Fig. 3). The percent changes of ucoc/oc ratio at 1 week, 1 373

4 Figure3. Percentchangesofserum undercarboxylatedos teocalcin(ucoc)levelafterstartingrisedronatemonother apy(groupa;closedcircles)orcombinedrisedronateandvi tamink2 therapy(groupb;opencircles).dataaremean± SE.Theasterisksindicatesignificantdiferencescompared withthebaseline(p<0.05). Figure5. Percentchangesofserum typeicolagencroslinkedn-telopeptide(ntx)levelafterstartingrisedronate monotherapy(group A;closed circles)orcombined risedronateandvitamink2therapy(groupb;opencircles).data aremean±se.theasterisksindicatesignificantdiferences comparedwithbaseline(p<0.05). Figure4. PercentchangesoftheucOC/OCratioafterstart ingrisedronatemonotherapy(group A;closed circles)or combined risedronateand vitamin K2 therapy(group B; opencircles).dataaremean±se. month, 3 months and 12 months were -16.9±25.0, -21.8± 12.7, -18.8±21.1 and -13.8±15.7%, respectively in group A, and 4.5±29.5, -10.8±31.9, -21.9±35.8 and -41.4±19.4% in group B (Fig. 4). Analyses of the time course of mean changes of the two groups detected no significant difference between the groups at all time points throughout the entire observation period. Changes of NTx in the long-term observation The percent changes from baseline at 1 week, 1 month, 3 months and 12 months were -24.6±7.1, -28.2±9.1, -29.6±9.3 and -15.9±21.2%, respectively, in group A, and -15.1±7.6, ±4.3, -38.1±4.8 and -30.1±8.3% in group B. The decreases compared to baseline were statistically significant (p <0.05) at all time points in groups A and B, except at 12 months in group A and at 1 week in group B. Betweengroup analysis showed no significant differences at all time points (Fig. 5). Figure6. Theefectofintravenoussteroidpulsetherapy onpercentchangesinserum osteocalcin(oc;closedcircle), undercarboxylatedosteocalcin(ucoc;opencircle)andtype I colagen cros-linked N-telopeptide (NTx;gray circle) levels.dataaremean±se.theasterisksindicatesignificant diferencescomparedwithbaseline(p<0.05). Changes of bone metabolic markers in the shortterm observation Steroid pulse therapy led to significant changes in OC (-47.4±14.0%) and ucoc (-56.6±15.0%) from the baseline values after 1 week (p<0.05). After 1 month, the percent changes of OC and ucoc increased to -18.2±11.2% and 28.5±23.8%, respectively, from the baseline (Fig. 6). On the other hand, there were no significant changes in NTx from the baseline after 1 week (-2.9±17.8%) and 1 month (-9.7± 13.2%). Discussion Osteoclacin is one of the most abundant noncollagenous proteins produced by osteoblasts in bone matrix (14, 15). Since a part of the OC enters the bloodstream, the serum OC concentration is used as a marker of bone forma- 374

5 tion (16, 17). On the other hand, the serum ucoc concentration is used as an indicator of the status of vitamin K2 supplementation to promote bone formation in osteoporotic patients, because vitamin K mediates the conversion of ucoc to OC by transforming Glu residues of OC to Gla. Several studies have reported that the circulating level of vitamin K is decreased in osteoporotic patients with bone fractures (18-21) and that a high serum concentration of ucoc is related to the risk of fracture in elderly women or osteoporotic patients (22-26). However, ucoc is not a reliable indicator to estimate the status of vitamin K2 in corticosteroidtreated patients because a significant decrease of ucoc concomitant with OC reduction is induced by corticosteroids early after administration regardless of vitamin K2 supplementation, as shown in the present study. Another striking finding in this study is that serum concentrations of OC and ucoc follow a very similar time course (27). The precise turnover of OC and ucoc in human serum has not been reported in the literature, although a study in rat models reported that one-half of the administered bone Gla protein may be cleared from serum in less than 5 minutes (17). Various studies have demonstrated that corticosteroids depress bone formation with a decrease in osteoblastic activity that is generally attributed to the direct inhibitory effect of corticosteroids on bone formation (28-30). This would explain the significant decrease in OC during corticosteroid therapy as was also reported previously (12). However, this is the first report in the literature of significant decreases in ucoc and OC immediately after corticosteroid administration. One possible explanation of why ucoc decreased significantly with a similar time course profile as OC is the facilitation of γ-carboxylation of serum ucoc following a steroid-induced reduction of serum OC. Another possibility is an as-yet unrecognized effect of steroid that reduces serum OC and ucoc simultaneously in the early stages of corticosteroid intake. The present results indicate the latter possibility, and also suggest that the reduction of ucoc by corticosteroids may be stronger than that by vitamin K2 alone according to previous studies (4, 31). In the long-term observation, the mean ucoc/oc ratios up to 3 months were relatively lower in group A than in group B in spite of combined therapy with vitamin K2 in group B (Fig. 4). This paradoxical result may be explained by the relatively high dose of prednisolone used in group A compared with group B, although the dose difference did not reach a statistical significance (Fig. 1). Further measurements of serum OC and ucoc concentrations in the earlier periods immediately after steroid administration (eg. within several hours) are required to resolve whether or not steroids have a direct effect on serum ucoc. Consistent with a previous report (32), the present study also showed that risedronate significantly reduced NTx in corticosteroid-treated patients, and the magnitude of reduction observed was similar to that previously reported in osteoporotic patients (33, 34). Since NTx has been used as a bone resorption marker in several reports (35, 36), our result suggests that risedronate can induce a significant reduction of bone turnover against the increased osteoclast-mediated bone resorption induced by corticosteroids (37). Although past studies have shown that vitamin K2 exhibits two contradictory functions; activation (38) and inhibition (39, 40) of osteoclasts, our results indicate that vitamin K2 may not influence NTx when used concomitantly with risedronate in corticosteroid-treated patients. In summary, vitamin K2 when used in combination with risedronate as prophylactic therapy for corticosteroidinduced osteoporosis had no additive effects on bone metabolic markers compared to risedronate alone (41). Since OC and ucoc were equally depressed shortly after corticosteroid treatment, ucoc is not a reliable indicator to estimate the status of vitamin K2 under corticosteroid administration. Although the bone metabolic markers were changed according to the administration of corticosteroids, treatment with risedronate and/or menatetrenone appears to be effective to maintain the BMD level for at least 12 months. This study examined a small number of subjects and did not assess the incidence of osteoporotic fracture. These are the limitations of this study and further studies are needed to clarify whether vitamin K2 supplementation in risedronate therapy provides additional benefits in the prevention of corticosteroid-induced osteoporosis. Acknowledgement This study was supported in part by Smoking Research Foundation. We would like to thank Eisai Pharmaceutical Co., Ltd. for comments on our manuscript. References 1. Akedo Y, Hosoi T, Inoue S, et al. Vitamin K2 modulates proliferation and function of osteoblastic cells in vitro. Biochem Biophys Res Commun 187: , Price PA. Vitamin K-dependent formation of bone Gla protein (osteocalcin) and its function. Vitam Horm 42: , Hirao M, Hashimoto J, Ando W, Ono T, Yoshikawa H. Response of serum carboxylated and undercarboxylated osteocalcin to alendronate monotherapy and combined therapy with vitamin K2 in postmenopausal women. J Bone Miner Metab 26: , Ozuru R, Sugimoto T, Yamaguchi T, Chihara K. Time-dependent effects of vitamin K2 (menatetrenone) on bone metabolism in postmenopausal women. Endocrine J 49: , Nakashima A, Yorioka N, Doi S, Masaki T, Ito T, Harada S. Effects of vitamin K2 in hemodialysis patients with low serum parathyroid hormone levels. Bone 34: , Shiraki M, Shiraki Y, Aoki C, Miura M. Vitamin K2 (menatetrenone) effectively prevents fractures and sustains lumbar bone mineral density in osteoporosis. J Bone Miner Res 15: , Gennari C. Glucocorticoids and bone. In: Bone and Mineral Research. Vol. 3. Peck WA, Ed. Excepta Medica, Amsterdam, 1998: Lukert BP, Raisz LG. Glucocorticoid-induced osteoporosis: Patho- 375

6 genesis and management. Ann Intern Med 112: , Hara K, Akiyama Y, Ohkawa I, Tajima T. Effects of menatetrenone on prednisolone-induced bone loss in rats. Bone 14: , Hara K, Kobayashi M, Akiyama Y. Vitamin K2 (menatetrenone) inhibits bone loss induced by prednisolone partly through enhancement of bone formation in rats. Bone 31: , Nawata H, Soen S, Takayanagi R, et al. Guidelines on the management and treatment of glucocorticoid-induced osteoporosis of the Japanese society for bone and mineral research (2004). J Bone Miner Metab 23: , Godshalk MF, Downs RW. Effect of short-term glucocorticoids on serum osteocalcin in healthy young men. J Bone Miner Res 3: , Ton FN, Gunawardene SC, Lee H, Neer RM. Effects of low-dose prednisolone on bone metabolism. J Bone Miner Res 20: , Hauschka PV, Lian JB, Gallop PM. Direct identification of the calcium-binding amino acid, gamma-carboxyglutamate, in mineralized tissue. Proc Natl Acad Sci USA 72: , Price PA, Pose JW, Raman N. Primary structure of γ- carboxyglutamic acid-containing protein from bovine bone. Proc Natl Acad Sci USA 73: , Delmas PD. Biochemical markers of bone turnover. I: Theoretical considerations and clinical use in osteoporosis. Am J Med 95: 11 S-16S, Price PA, Williamson MK, Lothringer JW. Origin of the vitamin K-dependent bone protein found in plasma and its clearance by kidney and bone. J Biol Chem 256: , Bitensky L, Hart JP, Catterall A, Hodges SJ, Pilkington MJ, Chayen J. Circulating vitamin K levels in patients with fractures. J Bone Jt Surg Br 70: , Hart JP, Shearer MJ, Klenerman L, et al. Electrochemical detection of depressed circulating levels of vitamin K1 in osteoporosis. J Clin Endocrinol Metab 60: , Hodges SJ, Pilkington MJ, Stamp TCB, et al. Depressed levels of circulating menaquinones in patients with osteoporotic fractures of the spine and femoral neck. Bone 1: , Hodges SJ, Akesson K, Vergnaud P, Obrant K, Delmas PD. Circulating levels of vitamin K1 and K2 decreased in elderly women with hip fracture. J Bone Miner Res 8: , Liu G, Peacock M. Age-related changes in serum undercarboxylated osteocalcin and its relationships with bone density, bone quality, and hip fracture. Calicif Tissue Int 62: , Luukinen H, Kakonen SM, Pettersson K, et al. Strong prediction of fractures among older adults by the ratio of carboxylated to total serum osteocalcin. J Bone Miner Res 15: , Plantalech L, Guillaumont M, Vergnaud P, Leclerocq M, Delmas PD. Impairment of gamma carboxylation of circulating osteocalcin (bone Gla protein) in elderly women. J Bone Miner Res 6: , Szulc P, Chapuy M, Meunier PJ, Delmas PD. Serum undercarboxylated osterocalcin is a marker of the risk of hip fracture in elderly women. J Clin Invest 91: , Vergnaud P, Garnero P, Meunier PJ, Breart G, Kamihagi K, Delmas PD. Undercarboxylated osteocalcin measured with a specific immunoassay predicts hip fracture in elderly women. J Clin Endocrinol Metab 82: , Obrant KJ, Kakonen SM, Astermark J, et al. The proportion of carboxylated to total or intact osteocalcin in serum discriminates warfarin-treated patients from control subjects. J Bone Miner Res 14: , Bressot C, Meunier PJ, Chapuy MC, LeJeune E, Edouard C, Darby AJ. Histomorphometric profile, pathophysiology and reversibility of corticosteroid-induced osteoporosis. Metab Bone Dis Rel Res 1: , Jowsey J, Riggs BL. Bone formation in hypercortisonism. Acta Endocrinol 63: 21-28, Meunier PJ, Dempster D, Edouard C, Chapuy MC, Arlot M, Charhon P. Bone histomorphometry in corticosteroid-induced osteoporosis and Cushing s syndrome. Adv Exp Med Biol 171: , Miki T, Nakatsuka K, Naka H, et al. Vitamin K2 (menaquinone 4) reduces serum undercarboxylated osteocalcin level as early as 2 weeks in elderly women with established osteoporosis. J Bone Miner Metab 21: , Wallach S, Cohen S, Reid DM, et al. Effects of risedronate treatment on bone density and vertebral fracture in patients on corticosteroid therapy. Calcif Tissue Int 67: , Majima T, Shimatsu A, Komatsu Y, et al. Effects of risedronate or alfacalcidol on bone mineral density, bone turnover, back pain, and fractures in Japanese men with primary osteoporosis: results of two-year strict observational study. J Bone Miner Metab 27: , Takada J, Iba K, Yamashita T. Changes in bone resorption markers among Japanese patients with postmenopausal osteoporosis treated with alendronate and risedronate. J Bone Miner Metab 25: , Eastell R, Mallinak K, Weiss S, et al. Biological variability of serum and urinary N-telopeptides of type I collagen in postmenopausal women. J Bone Miner Res 15: , Montagnani A, Gonnelli S, Cepollaro C, et al. A new serum assay to measure N-terminal fragment of telopeptide of type I collagen in patients with renal osteodystrophy. Eur J Intern Med 14: , Khosla S, Melton III LJ. Corticosteroid-induced osteoporosis. In: Osteoporosis, Etiology, Diagnosis and Management. 2nd ed. Riggs BL, Melton III LJ, Eds. Lippincott-Raven, Philadelphia, 1995: Chenu C, Colucci S, Grano M, et al. Osteocalcin induces chemotaxis, secretion of matrix proteins, and calcium-mediated intracellular signaling in human osteoclast-like cells. J Cell Biol 127: , Hara K, Akiyama Y, Nakamura T, Murota S, Morita I. The inhibitory effects of vitamin K2 (menatetrenone) on bone resorption may be related to its side chain. Bone 16: , Kameda T, Miyazawa K, Mori Y, et al. Vitamin K2 inhibits osteoclastic bone resorption by inducing osteoclast apoptosis. Biochem Biophys Res Commun 220: , Booth SL, Dallal G, Shea MK, Gundberg C, Peterson JW, Dawson-Hughes B. Effects of vitamin K supplementation on bone loss in elderly men and women. J Clin Endochrinol Metab 93: , The Japanese Society of Internal Medicine 376