PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of diacareAmerican Diabetes AssociationSubscribeSearchDiabetes Care Journal
 
Diabetes Care. 2009 December; 32(12): 2263–2268.
Published online 2009 September 14. doi:  10.2337/dc09-0901
PMCID: PMC2782988

Low Serum Level of the Endogenous Secretory Receptor for Advanced Glycation End Products (esRAGE) Is a Risk Factor for Prevalent Vertebral Fractures Independent of Bone Mineral Density in Patients With Type 2 Diabetes

Masahiro Yamamoto, MD, PHD, Toru Yamaguchi, MD, PHD, Mika Yamauchi, MD, PHD, and Toshitsugu Sugimoto, MD, PHD

Abstract

OBJECTIVE

Patients with type 2 diabetes are known to have an increased risk for fracture compared with non–type 2 diabetic control subjects, despite having higher bone mineral density (BMD). We previously showed that serum pentosidine, one of the advanced glycation end products (AGEs), was associated with prevalent vertebral fractures (VFs) in those with type 2 diabetes. The involvement of the endogenous secretory receptor for AGEs (esRAGE) in VFs in those with type 2 diabetes, however, is still unknown.

RESEARCH DESIGN AND METHODS

We compared parameters including esRAGE, pentosidine, and BMD in Japanese type 2 diabetic patients (137 men >50 years old and 140 postmenopausal women) with and without VFs.

RESULTS

The esRAGE-to-pentosidine ratio in type 2 diabetic patients with VFs was significantly lower than in those without VFs (men: 7.1 ± 2.8 vs. 9.4 ± 6.2, P = 0.013, respectively; women: 4.7 ± 2.7 vs. 8.2 ± 5.4, P < 0.001, respectively). Multivariate logistic regression analysis adjusted for age, BMI, A1C, serum creatinine, duration of diabetes, therapeutic agents, diabetes complications, osteoporotic risk factors, and lumbar BMD identified the serum esRAGE level and esRAGE-to-pentosidine ratio as factors associated with the presence of VFs, independent of BMD in men (odds ratio [OR] 0.46 [95% CI 0.25–0.84], P = 0.012; and OR 0.34 [0.15–0.76], P = 0.009, respectively) and in women (OR 0.32 [0.16–0.67], P = 0.002; and OR 0.14 [0.04–0.43], P = 0.001, respectively).

CONCLUSIONS

These results show that serum esRAGE level and esRAGE-to-pentosidine ratio are more useful than BMD for assessing the risk of VFs in type 2 diabetic patients.

The association between diabetes and osteoporosis has been investigated in many studies because these two disorders affect a large proportion of the elderly population. Recent meta-analyses of accumulating studies have shown that patients with type 2 diabetes have an increased risk for hip fracture compared with non–type 2 diabetic control subjects, despite their higher bone mineral density (BMD) (1,2). We have also shown that patients with type 2 diabetes have an increased risk for vertebral fractures (VFs) and that BMD at any site fails to assess the risk of VF (3). Because bone strength reflects integration of bone density and bone quality (4), these findings suggest that bone quality may be more important than bone density in defining bone strength in type 2 diabetic patients.

Bone quality is known to be determined by bone architecture, turnover, accumulation of microdamage, mineralization, and properties of bone matrix proteins such as collagen (4). In diabetic patients, advanced glycation end products (AGEs) are generated by sequential nonenzymatic glycosylation of protein amino groups (5). Pentosidine is one of the well-known AGEs, and its bone content in spontaneous diabetic rats has been shown to increase concurrently with the onset of diabetes, resulting in impaired mechanical properties of the bone despite normal BMD (6). We have shown clinically that the serum pentosidine level is associated with the presence of VFs in postmenopausal diabetic women independent of BMD (7). These findings suggest that AGEs, including pentosidine, may act as causative factors for poor bone quality in type 2 diabetic patients.

The receptor for AGEs (RAGE) belongs to the immunoglobulin superfamily of cell surface receptors and is capable of interacting with multiple ligands, including AGEs (8). When transgenic mice overexpressing human RAGE in vascular cells were crossbred with a transgenic line that develops insulin-dependent diabetes shortly after birth, a more progressive histological change of diabetic nephropathy was observed compared with controls (9), confirming that RAGE is associated with the development of diabetes complications. Endogenous secretory RAGE (esRAGE), a splice variant of one of the naturally occurring secretory forms, is known to carry all the extracellular domains but lacks the transmembrane and cytoplasmic domains (10). Secreted esRAGE in the extracellular space is thought to act as a decoy receptor that binds AGEs and results in reducing the activity of intercellular signal pathways via RAGE (10). Indeed, administration of a genetically engineered murine-soluble RAGE suppressed the development of diabetic atherosclerosis in a dose-dependent manner in streptozotocin-induced apoE-null diabetic mice (11). Recently, RAGE-knockout mice have been shown to increase BMD and biomechanical bone strength by decreasing osteoclast formation as well as serum levels of interleukin-6 and pyridinoline (12). We have also shown that the combination of high glucose with AGEs inhibits osteoblastic mineralization through glucose-induced increases in the expression of RAGE in vitro (13). These experimental findings suggest that enhanced RAGE activity may also be clinically linked to reduced bone strength in diabetic patients. Given the neutralizing nature of esRAGE, it is possible that the ratio of serum esRAGE-to-AGE levels could be linked to clinical bone problems, such as fractures, more prominently than either parameter alone.

To examine this issue, we compared serum levels of esRAGE and pentosidine as well as the esRAGE-to-pentosidine ratio between type 2 diabetic patients with and without VFs and evaluated the usefulness of these markers for assessing the risk of VFs in the population.

RESEARCH DESIGN AND METHODS

We consecutively enrolled 277 Japanese patients with type 2 diabetes (137 men [age range 50–82 years] and 140 postmenopausal women [age range 46–87 years]) who underwent BMD measurements at the outpatient clinic of Shimane University Hospital. The patients were referred to our hospital from community clinics for the treatment of diabetes. We excluded patients who had higher-than-normal range of serum creatinine (normal range for women, 0.44–0.83 mg/dl; men, 0.56–1.23 mg/dl) or >300 mg albumin/g urine creatinine for urinary albumin excretion because serum esRAGE levels are known to be influenced by decreased renal function (14). We also excluded patients with an abnormality of calcium metabolism such as primary hyperparathyroidism or a history of falls or traffic accidents to eliminate the possibility of injury-associated fractures. We defined the onset of type 2 diabetes as the first time when glucosuria or hyperglycemia was noticed. None of the patients were taking any drugs or hormones that affected bone metabolism, including sex steroids, warfarin, and bisphosphonates. Baseline characteristics of the subjects are shown in Table 1. Serum esRAGE levels in men (0.294 ± 0.102 ng/ml) and women (0.257 ± 0.161 ng/ml) were equivalent to those previously reported in subjects with metabolic syndrome (0.253 ± 0.111 ng/ml) (15) and lower than those in normal control subjects (0.436 ± 0.121 ng/ml) (16). Serum pentosidine levels in men (0.0413 ± 0.0194 μg/ml) and women (0.0400 ± 0.0159 μg/ml) were higher than those previously reported in normal control subjects (0.0261 ± 0.0007 μg/ml) (17). There were 48 men (35%) and 61 women (44%) with diabetic retinopathy, whereas 83 men (61%) and 93 women (66%) had diabetic neuropathy. There were 95 men (69%) and 6 women (4%) who smoked >20 cigarettes/day and 86 (63%) men and 12 (9%) women with habitual alcohol drinking. This study was cross-sectional and was approved by the ethics review board of our institution, in compliance with the Declaration of Helsinki. All subjects agreed to participate in the study and provided written informed consent.

Table 1
Background data of men and postmenopausal women with type 2 diabetes

Biochemical measurements

Fasting blood was obtained, and fasting plasma glucose (FPG), A1C, and serum creatinine were measured by automated techniques at the central laboratory of our hospital. Serum bone-specific alkaline phosphatase (BAP) and urinary N-telopeptide (uNTX) were commercially measured using enzyme-linked immunosorbent assay (ELISA).

Serum concentrations of human esRAGE were measured using an esRAGE ELISA kit (Daiichi Fine Chemicals, Takaoka, Japan) as reported previously (10,14). The interassay coefficient of variation (CV) for repeated measurements ranged from 2.5 to 5.1%. Serum pentosidine levels were detected using a competitive ELISA kit (FSK pentosidine ELISA kit; Fushimi Pharmaceutical, Kagawa, Japan) as described previously (7,18). The inter- and intra-assay CVs of absorbance were 6.6 and 8.0%, respectively. This ELISA was highly correlated with the conventional high-performance liquid chromatography method (r = 0.9356) (18).

Assessment of fractures

In all subjects VFs were identified on lateral X-ray films of the thoracic and lumbar spine according to the semiquantitative method (19) by two investigators who were blinded to the other's readings. VFs were classified as follows: mild (grade 1), a reduction of 20–25%; moderate (grade 2), 25–40%; severe (grade 3), >40%. If judgment of VFs was not in agreement, the film was independently reassessed. If the reevaluated findings were still different, we regarded the case as a milder grade.

Densitometry

BMD values of the lumbar spine (L), the femoral neck (FN), and one-third the radius (R) were measured by dual-energy X-ray absorptiometry using the QDR-4500 system (Hologic, Waltham, MA). Values were expressed relative to the SD of age- and sex-matched normal Japanese mean values of BMD provided by the manufacturer (Z score). The coefficients of variation of measurement of the L-, FN-, and R-BMD were 1.0, 1.0, and <1.0%, respectively.

Statistical analysis

All data are expressed as the means ± SD for each index. An unpaired t test was used to compare parameters between subjects with and without VFs. Comparisons of categorical variables were made using a χ2 test. Statistical analyses were performed using the computer program StatView for Windows, version 5.0 (SAS Institute, Cary, NC).

RESULTS

Simple regression analysis revealed that serum esRAGE level in diabetic men was significantly and positively correlated with creatinine (r = 0.377; P < 0.001) and significantly and negatively correlated with t score of R-BMD (r = −0.186; P = 0.029). There were no significant correlations between serum esRAGE levels and age, BMI, other BMD values, A1C, duration of diabetes, serum pentosidine level, or bone metabolic markers in either sex (Table 2).

Table 2
Simple regression analysis between esRAGE level and various parameters

We compared biochemical parameters, including serum esRAGE level and esRAGE-to-pentosidine ratio, between type 2 diabetic patients with and without VFs for each sex (Table 3). Diabetic men with VFs were significantly older (P < 0.001) and shorter in height (P < 0.001), and they had a significantly lower esRAGE-to-pentosidine ratio (P = 0.013) than those without VFs. Diabetic women with VFs were significantly older (P < 0.001) and shorter in height (P = 0.021), and they had a significantly higher serum pentosidine level (P = 0.006) and significantly lower values for the R-BMD t score (P = 0.047), serum esRAGE (P = 0.007), and esRAGE-to-pentosidine ratio (P < 0.001) than those without VFs (Table 4). There were no significant differences in other BMD values, bone metabolic markers, or demographic confounders such as therapeutic agents, diabetes complications, smoking status, or alcohol consumption between those with and without VFs for either sex.

Table 3
Comparison of various parameters between type 2 diabetic patients with and without VFs
Table 4
Associations between serum esRAGE level, serum pentosidine level, esRAGE-to-pentosidine ratio, and BMD versus the presence of VFs in type 2 diabetic patients

To determine the association between the presence of VFs and serum esRAGE, pentosidine, and esRAGE/pentosidine ratio, logistic analyses were performed (Table 4). When no adjustment was made (model 1), esRAGE-to-pentosidine ratio in men and serum esRAGE, pentosidine, and esRAGE-to-pentosidine ratio in women were significantly associated with VFs (men: esRAGE-to-pentosidine OR per SD increase 0.53 [95% CI 0.31–0.88], P = 0.015; and women: esRAGE OR 0.53 [0.33–0.85], P = 0.009; pentosidine OR 1.65 [1.13–2.41], P = 0.010; esRAGE-to-pentosidine OR 0.34 [0.19–0.62], P < 0.001). In contrast, no associations between VFs and BMD at any site were found in either sex. esRAGE level in men became significantly associated with VFs after adjustments for age, BMI, A1C, and creatinine level (model 2). These observations remained significant when multivariate logistic regression analysis was performed after the addition of therapeutic agents, the presence of diabetes complications, and risk factors for osteoporosis (e.g., smoking and habitual alcohol drinking) (model 3). However, esRAGE, pentosidine, or esRAGE-to-pentosidine ratio was not associated with severe VFs, including grade 2/3 or multiple ones (data not shown).

CONCLUSIONS

This is the first clinical study to show that serum esRAGE level is significantly and negatively associated with the presence of prevalent VFs in patients with type 2 diabetes. This association was independent of BMD, bone metabolic markers, therapeutic agents, diabetes complications, and risk factors for osteoporosis. esRAGE is known to bind various ligands, including AGEs, in the extracellular space and to inhibit the connection between cell surface RAGE and ligands (10). Thus, an insufficient amount of esRAGE to counteract AGEs could intensify the binding of AGEs to RAGE and exert harmful effects on organs via RAGE-transmitted signals in diabetic patients. We found that esRAGE was negatively associated with VFs, whereas pentosidine showed a positive association. esRAGE-to-pentosidine ratio was thought to be more suitable to assess the risk of VFs than serum esRAGE level alone because of its more significant P values. These findings support the concept that esRAGE could be beneficial to bones in those with type 2 diabetes by neutralizing the harmful effects of pentosidine or other AGEs rather than by direct activity.

We found that esRAGE was more significantly associated with VFs than pentosidine in diabetic patients. The association between esRAGE and VFs was significant in both sexes after adjustments for multivariate, whereas the association between pentosidine and VFs was only significant in women. Thus, esRAGE seems to be more useful than pentosidine to assess the risk of VFs irrespective of sex in type 2 diabetes. Several studies have shown that not only pentosidine but also other AGEs are harmful to bone. It has been documented that RAGE is expressed in human bone-derived cells (20) and that AGE-BSA inhibits the synthesis of type I collagen and osteocalcin in primary human osteoblasts and the secretion of parathyroid hormones from human parathyroid cells (21). We have also shown that combination of high glucose with either AGE2 or AGE3 inhibits osteocalcin expression and mineralization through glucose-induced increases in RAGE expression in cultured osteoblasts (13). Thus, various kinds of AGEs seem to impair bone matrix production and mineralization of osteoblasts, which may lead to the bone fragility seen in diabetic patients. Furthermore, esRAGE is known to interact with nonglycated proteins such as proinflammatory calcium-binding S100/calgranulins proteins and nuclear high-mobility group protein box-1 (HMGB1), which are released by cellular stress through RAGE (22). Given that esRAGE has extensive neutralizing effects against various AGEs that are harmful to bone as well as pentosidine, it seems reasonable that esRAGE was more highly associated with the presence of VFs than pentosidine in type 2 diabetic patients in this study.

This study has some limitations. First, it was not population-based, and the sample size was not large enough to make definitive conclusions. Second, the patients enrolled in this study might have had relatively severe cases of type 2 diabetes and might not be representative of standard Japanese diabetic patients, given that the diabetic conditions of the patients who attended Shimane University Hospital, a tertiary care center, were considered to be more serious than those of other diabetic patients. Third, we did not exclude patients with confounders known to affect bone strength to avoid a considerable reduction in the study population. Diabetic complications such as diabetic retinopathy, longer diabetes duration, and a history of insulin treatment are known to be associated with non-VFs in diabetic patients (23). A recent meta-analysis showed that long-term use of thiazolidinediones was associated with an increased risk of fractures and reduced L- and FN-BMD in female type 2 diabetic patients (24). Multiple logistic regression analysis in this study, however, revealed that both serum esRAGE levels and the esRAGE-to-pentosidine ratio were associated with the presence of VFs independent of diabetes complications, duration of diabetes, or therapeutic agents including insulin and pioglitazone, suggesting that none of these factors affected our observations. Finally, BMI of our patients was 23.3 in men and 24.5 in women, showing that they were less obese by Western standards. It is well known that low BMI increases risk for all kind of fractures (25). Although the significant results were still obtained after adjustment for BMI in the present study, our findings need to be confirmed in other ethnic groups.

In conclusion, we found that serum esRAGE level as well as esRAGE-to-pentosidine ratio were significantly and inversely associated with the presence of prevalent VFs in type 2 diabetic patients. These associations were independent of BMD and diabetes-related confounders and were stronger than serum pentosidine level alone. These findings suggest that the AGE-RAGE system as a whole may affect bone quality, which is not determined by BMD. The present study as well as our previous studies (3) suggest that BMD may not be sensitive enough to assess the risk of VFs, given that BMD was not associated with the presence of VFs in type 2 diabetic patients. Serum esRAGE level and esRAGE-to-pentosidine ratio seem to be useful surrogate markers, which could compensate for the ineffectiveness of BMD in evaluating the risk of VFs in type 2 diabetic patients. However, our study was cross-sectional, and we did not investigate non-VFs in the subjects. Longitudinal studies on both VFs and non-VFs are needed to prove the importance of the AGE-RAGE system in assessing the fracture risk of type 2 diabetic patients.

Acknowledgments

This work was supported in part by Grants-in-Aid 20791039 (to M.Y.) and 20590699 (to T.Y.) from the Ministry of Science, Education and Culture of Japan and by the 2008 Japan Osteoporosis Society (JOS) Young Scientist Award (to M.Y.).

No potential conflicts of interest relevant to this article were reported.

Footnotes

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

References

1. 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– 444 [PubMed]
2. Janghorbani M, Van Dam RM, Willett WC, Hu FB.: Systematic review of type 1 and type 2 diabetes mellitus and risk of fracture. Am J Epidemiol 2007; 166: 495– 505 [PubMed]
3. Yamamoto M, Yamaguchi T, Yamauchi M, Kaji H, Sugimoto T.: Diabetic patients have an increased risk of vertebral fractures independent of BMD or diabetic complications. J Bone Miner Res 2009; 24: 702– 709 [PubMed]
4. NIH Consensus Development Panel on Osteoporosis Prevention, Diagnosis, and Therapy March 7–29, 2000: highlights of the conference.South Med J 2001; 94: 569– 573 [PubMed]
5. Brownlee M.: Advanced protein glycosylation in diabetes and aging. Annu Rev Med 1995; 46: 223– 234 [PubMed]
6. Saito M, Fujii K, Mori Y, Marumo K.: Role of collagen enzymatic and glycation induced cross-links as a determinant of bone quality in spontaneously diabetic WBN/Kob rats. Osteoporos Int 2006; 17: 1514– 1523 [PubMed]
7. Yamamoto M, Yamaguchi T, Yamauchi M, Yano S, Sugimoto T.: Serum pentosidine levels are positively associated with the presence of vertebral fractures in postmenopausal women with type 2 diabetes. J Clin Endocrinol Metab 2008; 93: 1013– 1019 [PubMed]
8. Kislinger T, Fu C, Huber B, Qu W, Taguchi A, Du Yan S, Hofmann M, Yan SF, Pischetsrieder M, Stern D, Schmidt AM.: N(epsilon)-(carboxymethyl)lysine adducts of proteins are ligands for receptor for advanced glycation end products that activate cell signaling pathways and modulate gene expression. J Biol Chem 1999; 274: 31740– 31749 [PubMed]
9. Yamamoto Y, Kato I, Doi T, Yonekura H, Ohashi S, Takeuchi M, Watanabe T, Yamagishi S, Sakurai S, Takasawa S, Okamoto H, Yamamoto H.: Development and prevention of advanced diabetic nephropathy in RAGE-overexpressing mice. J Clin Invest 2001; 108: 261– 268 [PMC free article] [PubMed]
10. Yonekura H, Yamamoto Y, Sakurai S, Petrova RG, Abedin MJ, Li H, Yasui K, Takeuchi M, Makita Z, Takasawa S, Okamoto H, Watanabe T, Yamamoto H.: Novel splice variants of the receptor for advanced glycation end-products expressed in human vascular endothelial cells and pericytes, and their putative roles in diabetes-induced vascular injury. Biochem J 2003; 370: 1097– 1109 [PubMed]
11. Park L, Raman KG, Lee KJ, Lu Y, Ferran LJ, Jr, Chow WS, Stern D, Schmidt AM.: Suppression of accelerated diabetic atherosclerosis by the soluble receptor for advanced glycation endproducts. Nat Med 1998; 4: 1025– 1031 [PubMed]
12. Ding KH, Wang ZZ, Hamrick MW, Deng ZB, Zhou L, Kang B, Yan SL, She JX, Stern DM, Isales CM, Mi QS.: Disordered osteoclast formation in RAGE-deficient mouse establishes an essential role for RAGE in diabetes related bone loss. Biochem Biophys Res Commun 2006; 340: 1091– 1097 [PubMed]
13. Ogawa N, Yamaguchi T, Yano S, Yamauchi M, Yamamoto M, Sugimoto T.: The combination of high glucose and advanced glycation end-products (AGEs) inhibits the mineralization of osteoblastic MC3T3–E1 cells through glucose-induced increase in the receptor for AGEs. Horm Metab Res 2007; 39: 871– 875 [PubMed]
14. Koyama H, Yamamoto H, Nishizawa Y.: RAGE and soluble RAGE: potential therapeutic targets for cardiovascular diseases. Mol Med 2007; 13: 625– 635 [PMC free article] [PubMed]
15. Koyama H, Shoji T, Yokoyama H, Motoyama K, Mori K, Fukumoto S, Emoto M, Shoji T, Tamei H, Matsuki H, Sakurai S, Yamamoto Y, Yonekura H, Watanabe T, Yamamoto H, Nishizawa Y.: Plasma level of endogenous secretory RAGE is associated with components of the metabolic syndrome and atherosclerosis. Arterioscler Thromb Vasc Biol 2005; 25: 2587– 2593 [PubMed]
16. Katakami N, Matsuhisa M, Kaneto H, Matsuoka TA, Sakamoto K, Nakatani Y, Ohtoshi K, Hayaishi-Okano R, Kosugi K, Hori M, Yamasaki Y.: Decreased endogenous secretory advanced glycation end product receptor in type 1 diabetic patients: its possible association with diabetic vascular complications. Diabetes Care 2005; 28: 2716– 2721 [PubMed]
17. Yoshida N, Okumura K, Aso Y.: High serum pentosidine concentrations are associated with increased arterial stiffness and thickness in patients with type 2 diabetes. Metabolism 2005; 54: 345– 350 [PubMed]
18. Sanaka T, Funaki T, Tanaka T, Hoshi S, Niwayama J, Taitoh T, Nishimura H, Higuchi C.: Plasma pentosidine levels measured by a newly developed method using ELISA in patients with chronic renal failure. Nephron 2002; 91: 64– 73 [PubMed]
19. Genant HK, Wu CY, van Kuijk C, Nevitt MC.: Vertebral fracture assessment using a semiquantitative technique. J Bone Miner Res 1993; 8: 1137– 1148 [PubMed]
20. Takagi M, Kasayama S, Yamamoto T, Motomura T, Hashimoto K, Yamamoto H, Sato B, Okada S, Kishimoto T.: Advanced glycation endproducts stimulate interleukin-6 production by human bone-derived cells. J Bone Miner Res 1997; 12: 439– 446 [PubMed]
21. Yamamoto T, Ozono K, Miyauchi A, Kasayama S, Kojima Y, Shima M, Okada S.: Role of advanced glycation end products in adynamic bone disease in patients with diabetic nephropathy. Am J Kidney Dis 2001; 38: S161– S164 [PubMed]
22. Yan SF, Ramasamy R, Schmidt AM.: Mechanisms of disease: advanced glycation end-products and their receptor in inflammation and diabetes complications. Nat Clin Pract Endocrinol Metab 2008; 4: 285– 293 [PubMed]
23. Ivers RQ, Cumming RG, Mitchell P, Peduto AJ. Blue Mountains Eye Study. Diabetes and risk of fracture: The Blue Mountains Eye Study. Diabetes Care 2001; 24: 1198– 1203 [PubMed]
24. Loke YK, Singh S, Furberg CD.: Long-term use of thiazolidinediones and fractures in type 2 diabetes: a meta-analysis. CMaj 2009; 180: 32– 39 [PMC free article] [PubMed]
25. De Laet C, Kanis JA, Odén A, Johanson H, Johnell O, Delmas P, Eisman JA, Kroger H, Fujiwara S, Garnero P, McCloskey EV, Mellstrom D, Melton LJ, 3rd, Meunier PJ, Pols HA, Reeve J, Silman A, Tenenhouse A.: Body mass index as a predictor of fracture risk: a meta-analysis. Osteoporos Int 2005; 16: 1330– 1338 [PubMed]

Articles from Diabetes Care are provided here courtesy of American Diabetes Association