Clinical Aspects of Diabetic Bone Disease: An Update
- 268 Downloads
Older adults with type 2 diabetes have a higher risk of fracture but do not have decrements in bone density. The reasons for greater bone fragility in diabetes are still not clearly understood, but progress has been made in identifying potential contributors. With new imaging techniques, increased cortical porosity has been identified as a possible contributing factor to bone fragility. Cortical porosity may be especially detrimental to bone strength in the presence of higher levels of advanced glycation end products. Initial results have reported higher levels of marrow adiposity in diabetic men, suggesting that diabetes may contribute to marrow stem-cell lineage allocation toward adipocytes rather than osteoblasts. Higher levels of sclerostin have also been observed in diabetic patients, indicating that osteocyte function may play a role in diabetic bone. An important clinical question, “How to best predict fracture in diabetic patients?” has been addressed with studies demonstrating that BMD T-score and FRAX score predict fractures but underestimate absolute risk in diabetic patients. For prevention of fractures, there is now evidence from a clinical trial that intensive glycemic control does not increase fracture or fall risk but also does not reduce these outcomes. The pursuit of these new findings promises to provide insights into the reasons for greater fragility of diabetic bone and the best methods for fracture prevention in this population.
KeywordsDiabetes Fracture Bone mineral density Hyperglycemia Advanced glycation end products Marrow adiposity Cortical porosity
Diabetes is associated with an increased risk of fracture. For both type 1 and type 2 diabetes, higher risk has been reported for hip fractures [1, 2, 3, 4, 5, 6, 7, 8, 9]. In a meta-analysis, Vestergaard estimated a risk ratio for diabetes and hip fracture of 1.38 (95% CI, 1.25–1.53) for type 2 and 6.94 (95% CI, 3.25–14.78) for type 1 diabetes . Increased risk of non-spine fractures in general has also been reported for type 2 diabetes [3, 9, 11, 12, 13]. Less information is available on other fracture sites in type 1 diabetes although risks appear somewhat elevated for non-spine fractures as a group [6, 14]. Surprisingly, the higher fracture risks in those with type 2 diabetes are observed in spite of bone density by dual X-ray absorptiometry (DXA) that is similar to, or even higher than, BMD in those without diabetes . In other words, those with type 2 diabetes are more likely to fracture at a given BMD than non-diabetic patients of a similar age. Type 1 diabetes is associated with modest decrements in BMD, apparently consistent with their increased fracture risk, but upon closer examination, the increase in hip fracture risk is much larger than would be predicted from the lower BMD (1.42 vs. 6.94) . Thus, in both types of diabetes, patients have an increased risk of fracture at a given BMD, compared with non-diabetic patients.
Type 2 diabetes (T2DM) in particular affects an increasing proportion of older adults, and there has been growing concern regarding our ability to prevent and treat osteoporosis in this population. Previous reviews have considered the relationship between type 2 diabetes and fracture risk [15, 16]. This review will provide an update on several areas with recent findings of interest, including our understanding of the factors that account for increased fracture risk in diabetes, our ability to predict who will fracture in this population, and the clinical risk factors for fracture in older diabetic adults.
Increased Fracture Risk in Type 2 Diabetes
BMD and Falls
Older adults with type 2 diabetes are more likely to fall [17, 18], and this likely contributes to their higher fracture risk. However, in studies of diabetes and fracture that have measured falls, the increased frequency of falls does not account for the higher fracture risk associated with diabetes [3, 12]. Combined with the evidence that BMD tends to be higher in those with type 2 diabetes, this suggests that there is a deficit in bone strength that is not identified with DXA measurement. The concept of reduced bone strength for a given BMD is not unique to diabetes. A similar phenomenon is observed with older age. On average, for a given BMD by DXA, those who are older have an increased risk of fracture. Similarly, diabetes is associated with decreased bone strength at a given BMD.
Medications prescribed for control of diabetes may also have effects on bone. Thiazolidinedione (TZD) use in particular is associated with increased fracture risk, particularly in women . Use of TZDs may account in part for increased fracture risk with diabetes. However, rosiglitazone and pioglitazone were not available in the United States until 1999. Most of the follow-up time in studies identifying increased fracture risk with diabetes occurred prior to widespread use of this class of medication . The effects of TZDs on bone and the potential effects of other anti-diabetic medications are reviewed by Lecka–Czernik in this issue.
Over the last several years, studies have employed different imaging techniques to assess bone structure and its relationship to diabetes. Based on quantitative computed tomography (QCT) scans of the hip and spine, Melton et al.  reported that, although aBMD was higher in T2DM, strength-to-load ratios at the femoral neck and spine were not improved in T2DM compared with non-diabetic patients. Thus, the higher BMD in the T2DM patients did not appear to confer added protection against fracture. In a study of older men, pQCT at the radius and tibia revealed higher vBMD with smaller bone area at the distal sites . In the mid-shaft area, men with T2DM had similar cortical vBMD but smaller bone area, resulting in lower estimated bending strength, adjusted for body weight. The response of bone to loading appears to be shifted in T2DM. Smaller cross-sectional area suggests deficits in the expansion of bone by periosteal apposition, normally observed with greater loading .
Although DXA scans do not provide the 3D assessment possible with QCT, it is possible to identify 2D bone geometry at the hip and compute strength indices that are associated with hip fracture risk . In a study of women during the menopause transition, an index of femoral strength developed by Karlamangla  was lower in women with diabetes .
Cortical porosity decreases bone strength [35, 36]. Earlier studies, using bone biopsies or cadaver specimens, showed that higher cortical porosity is associated with hip and vertebral fracture [37, 38]. The pattern of increased cortical porosity with maintenance of trabecular bone, seen in the diabetic patients with fracture history, increases the vulnerability of the bone to bending, rather than compression, forces. It is notable that cortical porosity seems to vary widely among those with diabetes, with high levels in those with prevalent fractures and modest elevation in those without fracture, compared to non-diabetic subjects . This variability suggests that factors related to the development and progression of diabetes could confer different risks of increased porosity. Identification of increased cortical porosity in T2DM patients with a history of fracture provides the first clear indication of a deficit in bone associated with fractures in this population. If this association is confirmed with further research, assessment of the factors that contribute to cortical porosity could provide insight into the causes of increased fracture risk in diabetic patients. Cortical porosity may also prove useful as a clinical marker of increased fracture risk in this population.
Advanced Glycation End products (AGEs)
The material properties of bone may be compromised with diabetes, especially with longer duration. Advanced glycation end products (AGEs) accumulate with diabetes, including in bone collagen, and have been proposed as a factor contributing to bone fragility independent of BMD. AGEs are formed through a series of non-enzymatic reactions between glucose and proteins, resulting in a highly stable cross-linked product. Hyperglycemia, oxidative stress, and reduced renal function contribute to AGE levels . AGEs can form with any protein, but accumulation occurs in tissues with low turnover, including collagen. At higher levels, AGEs cause subtle changes in collagen structure, increasing stiffness in arteries, skin, cartilage, and bone .
In vitro studies that have induced higher levels of AGEs in specimens of human and bovine bone have demonstrated negative effects on bone strength [41, 42, 43]. Cadaver studies have reported reduced bone strength with higher AGE levels in bone [44, 45, 46]. And, in studies using surgical specimens, patients with femoral neck fracture had higher levels of AGEs than those without fracture .
Given the difficulties in obtaining measurements of AGEs in bone collagen, clinical studies have relied on measurements of circulating AGEs. Pentosidine is a well-characterized AGE with established assay procedures that can be measured in serum and urine. However, bone is only one of many tissues contributing to the circulating level of pentosidine. Not surprisingly, a study comparing pentosidine levels in serum and cortical bone demonstrated a low level of correlation (r = 0.25) . Several studies have reported an increased risk of fracture with higher levels of pentosidine in the serum or urine in those with diabetes. Serum pentosidine was associated with prevalent vertebral fracture in diabetic women, but not men . An association between higher urine pentosidine levels and increased non-spine fracture risk was found in older diabetic adults . In those without diabetes, one study found increased risk of vertebral fracture with higher pentosidine , while others have reported no increased risk of non-spine fracture in men and women  and of clinical fracture in women .
An intriguing possibility is a synergistic effect of microstructure and material properties on bone strength. Yang et al.  recently explored this hypothesis with finite element models assessing fracture resistance with different combinations of cortical porosity and AGE levels. Their results suggest that the effects of higher AGEs on bone are nonlinear and are worsened by increasing porosity.
Diabetes is thought to reduce osteoblast function and bone formation. Such effects have been widely demonstrated in rodent models, but the clinical evidence for an effect of diabetes on bone metabolism is sparse. A bone histomorphometry study published in 1995 demonstrated reduced bone formation in 6 patients with type 2 diabetes and 2 patients with type 1 diabetes compared with healthy premenopausal controls . However, a recent study of bone histomorphometry in women with type 1 diabetes (N = 18) found no differences in apposition rate compared with healthy controls . In most studies of bone turnover markers, osteocalcin, a marker of formation, is decreased with type 2 diabetes [56, 57, 58, 59, 60]. However, other formation markers are not consistently different in diabetic patients [56, 57]. Resorption markers have been reported as increased , decreased [57, 58], or not different [57, 59] in those with diabetes.
There is a growing appreciation of the role of osteocytes in bone metabolism although the tools for clinical studies remain limited. Sclerostin, a product of osteocytes, antagonizes the Wnt signaling pathway, resulting in inhibition of osteoblasts . The use of anti-sclerostin antibodies as an anabolic treatment for osteoporosis is being actively explored . In a recent study among older women, higher serum sclerostin levels at baseline were associated with an increased risk of hip fracture . At the same time, sclerostin was positively associated with bone mineral density. In a separate study, higher sclerostin levels have been reported in type 2 diabetes . These results suggest that diabetes may be associated with changes in osteocyte function and the Wnt pathway, possibly contributing to bone fragility. Further research is needed to identify the effects of diabetes on sclerostin levels and on osteocytes.
Studies in older adults have established that a higher proportion of fat in bone marrow is associated with osteoporosis [65, 66, 67, 68]. In rodent models, diabetes is associated with increased marrow fat , and TZD use also produces higher levels of marrow fat . Recently, a study using magnetic resonance spectroscopy to measure vertebral marrow fat in older men reported higher levels in men with diabetes (59%) compared to those without diabetes (55%) . The historic conception of marrow fat as merely a passive filler has been superseded by an appreciation for this fat depot as a dynamic player in bone health [72, 73]. The clinically observed associations between increased marrow adiposity, older age, and decreased bone density may result from the increasing commitment of bone marrow-derived mesenchymal stem cells (MSCs) to the adipocyte rather than the osteoblast lineage with advancing age [74, 75, 76, 77]. Factors that may influence the lineage commitment of MSCs include loading of the skeleton through weight or vibration . Immobilization results in increased marrow adipogenesis and decreased osteoblastogenesis while loading has the opposite effect [79, 80]. Oxidative stress may promote increased marrow adipogenesis and decreased osteoblastogenesis via the Wnt pathway . In rodent models, ovariectomy results in reduced bone density and increased marrow fat . Diabetes may be another factor that influences this relationship. Marrow fat also produces factors that may directly affect osteoblasts and osteoclasts. In co-cultures with marrow adipocytes, osteoblast activity was inhibited , possibly due to release of free fatty acids . Fat in the marrow, like other fat depots, produces inflammatory cytokines that can promote osteoclast recruitment and bone loss . Age and osteoporosis may alter marrow fat and its products. In vitro studies comparing MSCs obtained from osteoporotic versus control donors have reported changes favoring increased adipogenesis [86, 87]. Further research is needed to clarify the relationship between diabetes, accumulation of marrow fat, and effects on bone metabolism.
Risk Factors for Fracture and Prediction of Fractures in Type 2 Diabetes
BMD as a Risk Factor for Fracture
Other Traditional Risk Factors for Fracture
Traditional risk factors, other than BMD, that are known to be associated with fracture risk among older adults, also appear to be associated with fracture risk in those with diabetes. In a study of risk factors for fracture in those with diabetes, Melton et al.  identified 1061 cases of moderate trauma fracture among 1964 residents of Rochester, MN, followed during 1970–1994. Fractures were predicted by many traditional risk factors including older age, female gender, prior fracture, reduced physical activity, lower BMI, prevalence of factors related to falling, use of corticosteroids, and use of osteoporosis drugs.
FRAX Score: BMD Combined with Other Risk Factors
The relationship between hyperglycemia and fracture risk does not appear to be linear. Studies have reported no increase in risk , or even decreased risk [94, 95], comparing those with impaired glucose tolerance to those with normoglycemia. It is possible that positive effects of overweight and hyperinsulinemia are the main influence on bone during impaired glucose tolerance, but that, with the development of frank diabetes, the presence of complications and higher levels of advanced glycation end products result in an overall negative effect on bone.
Among those with diabetes, there is not an established relationship between A1C and fracture risk. Most observational studies have found no effect [9, 11, 14, 96, 97]; one study reported increased vertebral fracture risk with higher A1C in obese men . A 1-year clinical trial reported improved bone density at the femoral neck with improved glycemic control in older adults who presented with poor control (mean A1C, 11.6%) .
Recently, the effects of glycemic control on the risk of fractures and falls were tested in the Action to Control Cardiovascular Risk in Diabetes (ACCORD), a randomized trial of intensive versus standard glycemia therapy in a population with long-standing type 2 diabetes and a history of CVD or significant cardiovascular risk factors. Over an average follow-up of 3.8 years, those in the intensive group maintained an average A1C of 6.4%, while those in the standard group had an average A1C of 7.5%. There were no differences in non-spine fracture rates between the two ACCORD treatment groups (HR = 1.04; 95% CI, 0.86–1.27) . A substantial proportion of participants in the standard therapy group (58%) were prescribed a TZD, primarily rosiglitazone, during the trial, and use was even more frequent in the intensive therapy arm (92%). TZD use is associated with increased fracture risk, particularly in women, and this high prevalence of use might have obscured an effect of glycemic control on fracture risk. When results were examined separately in men, who are not as susceptible to the negative skeletal effects of TZDs, fracture risk was slightly reduced in the intensive group, but the differences were not statistically significant (hazard ratio = 0.93; 95% CI, 0.70–1.25). The rate of falls also did not differ between the two glycemia therapy groups (rate ratio = 1.10; 95% CI, 0.84–1.43) . These results indicate that improving glycemic control beyond an average A1C of 7.5% does not have a clinically important effect on fracture risk, at least over a period of several years. The ACCORD trial was stopped early due to higher mortality in the intensive glycemic control group . Given this higher mortality risk, the intensive glycemia therapy goals and regimen employed in ACCORD are not recommended for the treatment of diabetes. However, improved glycemic control with maintenance of A1C levels below 7% continues to be an important goal of diabetes treatment . In this context, the ACCORD results provide reassurance that tighter control is not likely to increase the risk of fractures and falls, a possibility due to the increased frequency of hypoglycemic episodes with lower A1C levels. At the same time, the lack of improvement in fracture risk indicates that glycemic control by itself is not sufficient to reduce fracture risk in diabetic patients, at least over a period of 3–4 years. The longer-term effect of improved glycemic control on fracture risk is not known. Improved control is associated with reduced microvascular complications, and it is possible that such reductions might translate into prevention of fractures over more extended time periods.
A few studies have reported on diabetes-related complications as risk factors for fracture in those with type 2 diabetes, but results have not been consistent. In the Health Aging and Body Composition study in older adults, unadjusted analyses comparing the prevalence of complications in diabetic participants with (N = 30) and without fracture found higher prevalence of neuropathy and history of stroke/TIA in those with fracture but no difference in cardiovascular disease . However, because of small numbers, this study was not able to develop multivariable models for fracture risk in diabetic participants. In Rochester, MN residents with type 2 diabetes, neuropathy was a risk factor for fracture (HR = 1.3; 95% CI, 1.1–1.6) in multivariable models, but renal failure and retinopathy were not . A study in Korean patients with diabetes also reported peripheral neuropathy as a risk factor for fracture . In a large case–control study among patients with type 2 diabetes, conducted using the Danish National Hospital Discharge Register, investigators did not find increased fracture risk associated with macrovascular complications, diabetic eye disease, or neuropathy considered separately . However, there was a modest increase in fracture risk for multiple complications.
Preventing Fractures in Type 2 Diabetes
Since BMD is a risk factor for fracture in older adults with type 2 diabetes, it is likely that treatments to preserve bone density will reduce fractures in this population. However, there are concerns that anti-resorptive therapies, the most common available treatments for osteoporosis, might lower bone turnover too far in diabetic patients. As discussed above, there is evidence that diabetes is characterized by lower bone turnover. The anti-resorptive therapies reduce turnover even further, potentially resulting in the accumulation of microcracks in bone and ultimately increasing fracture risk. Some reports have suggested that atypical subtrochanteric fractures, potentially linked to bisphosphonate use, may be increased with diabetes . However, atypical fractures are rare. They are a small portion of subtrochanteric fractures, which constitute only about 3% of hip fractures . Limited results are available on the efficacy of osteoporosis therapies in those with diabetes. In a subgroup analysis of data from the Fracture Intervention Trial (FIT), alendronate was found to improve bone density at the total hip, femoral neck, and spine, compared with placebo, in postmenopausal women with diabetes . Interestingly, comparing the diabetic and non-diabetic women assigned to alendronate, those without diabetes had a greater improvement in BMD (Fig. 1). In a study of postmenopausal women in Turkey, a similar pattern of better BMD gains with alendronate at the hip and forearm, but not the spine, was reported among non-diabetic compared with diabetic women, but the study did not have a placebo group for comparison . A subgroup analysis of results from the Multiple Outcomes of Raloxifene (MORE) trial found that treatment, compared with placebo, was effective in preventing vertebral fractures in those with type 2 diabetes. In a small study of alendronate in postmenopausal Japanese women, without a placebo group, Iwamoto et al.  reported that women with and without diabetes had similar gains in spine BMD. However, 3 (19%) of 16 diabetic women experienced a non-vertebral fracture compared with 4 of 135 (3%) (P < 0.05). The authors noted that 2 of the 3 diabetic women with fracture had a history of thiazolidinedione use. In a study using the Danish registries for hospital discharges and pharmacy sales, Vestergaard et al. analyzed the fracture experience of patients exposed to anti-resorptive therapy (bisphosphonates or raloxifene). In general, those using an anti-resorptive therapy had an increased rate of fractures, consistent with prescription of these therapies to those at highest risk of fracture. However, when the effect of therapy on fracture rate was compared in patients with and without diabetes, there were no statistically significant differences in the relative rate of fractures associated with any of the considered therapies. For example, the relative rate of hip fracture associated with alendronate therapy was 1.85 (1.68–2.04) in those without diabetes and 1.97 (1.31–2.95) in those without diabetes (p for homogeneity = 0.77). To date, the available studies indicate that fracture efficacy of the anti-resorptive therapies is similar in those with and without diabetes. However, information is not available comparing the efficacy of anabolic and anti-resorptive therapies. With glucocorticoid-induced osteoporosis, also characterized by reduced bone formation, alendronate is effective in preventing fractures, but anabolic therapy using PTH was more effective than alendronate in preserving BMD and preventing vertebral fractures . Additional research is warranted to understand the most effective therapies for use in type 2 diabetes.
Over the past several years, studies have more firmly established the association between fracture risk and type 2 diabetes and have increased our appreciation of the contribution of bone fragility, independent of BMD, to this risk. Efforts to identify the key elements contributing to lower bone strength in type 2 diabetes have progressed. Better imaging techniques have revealed deficits in the microstructure of cortical bone in diabetic patients with fracture and have identified increased marrow adiposity in diabetes. In vitro studies suggest that cortical porosity may be particularly detrimental in the presence of higher levels of advanced glycation end products that accumulate in diabetes. Gains have also been made in our understanding of fracture prediction in diabetic patients, with studies demonstrating the importance of lower BMD as a risk factor but also identifying the tendency of BMD T-score and FRAX to underestimate absolute fracture risk in diabetes. For prevention of fractures, maintenance of intensive glycemic control over several years does not increase fracture or fall risk but also does not reduce these outcomes. Thus, a better understanding of the efficacy of standard osteoporosis therapies for fracture prevention in this population is needed.
- 22.Strotmeyer ES, Boudreau RM, Marshall LM, Schwartz AV, Bauer DC, Barrett-Connor E, et al. Higher bone mineral denisty loss in older men with diabetes: The Osteoporotic Fractures in Men Study. In ASBMR 30th annual meeting, 2008. Montreal, ASBMR; 2008.Google Scholar
- 31.Burghardt AJ, Issever AS, Schwartz AV, Davis KA, Masharani U, Majumdar S, et al. High-resolution peripheral quantitative computed tomographic imaging of cortical and trabecular bone microarchitecture in patients with type 2 diabetes mellitus. J Clin Endocrinol Metab. 2010;95:5045–55.CrossRefPubMedGoogle Scholar
- 32.Yap SP, Baum T, Burghardt AJ, Link TM. Cortical porosity identifies fragility fractures in type-2 diabetic postmenopausal women. In European congress of radiology 2011, Book of Abstracts. 2011;S193.Google Scholar
- 33.Shu A, Yin MT, Stein E, Cremers S, Dworakowski E, Ives R, et al. Bone structure and turnover in type 2 diabetes mellitus. Osteoporos Int. 2011 [Epub ahead of print].Google Scholar
- 34.Nicks KM, Amin S, Atkinson EJ, Riggs BL, Melton LJ III, Khosla S. Relationship of age to bone microstructure independent of areal bone mineral density. J Bone Miner Res. 2011 [Epub ahead of print].Google Scholar
- 56.Suzuki K, Kurose T, Takizawa M, Maruyama M, Ushikawa K, Kikuyama M, et al. Osteoclastic function is accelerated in male patients with type 2 diabetes mellitus: the preventive role of osteoclastogenesis inhibitory factor/osteoprotegerin (OCIF/OPG) on the decrease of bone mineral density. Diabetes Res Clin Pract. 2005;68:117–25.CrossRefPubMedGoogle Scholar
- 71.Sheu Y, Amati F, Schwartz A, Li X, Lee C, Gordon CL, et al. The relationship of bone marrow fat with bone geometry and strength differs by diabetic status. J Bone Miner Res. 2011;26. Available at http://www.asbmr.org/Meetings/AnnualMeeting/Abstract2011.aspx.
- 80.David V, Martin A, Lafage-Proust MH, Malaval L, Peyroche S, Jones DB, et al. Mechanical loading down-regulates peroxisome proliferator-activated receptor gamma in bone marrow stromal cells and favors osteoblastogenesis at the expense of adipogenesis. Endocrinology. 2007;148:2553–62.CrossRefPubMedGoogle Scholar
- 91.World Health Organization. FRAX WHO fracture risk assessment tool. Access date: 2008; http://www.sheffield.ac.uk/FRAX/.
- 93.Giangregorio L, Leslie W, Lix L, Johansson H, Oden A, McCloskey E, et al. FRAX underestimates fracture risk in patients with diabetes. J Bone Miner Res. 2011 [Epub ahead of print].Google Scholar
- 97.Viegas M, Costa C, Lopes A, Griz L, Medeiro MA, Bandeira F. Prevalence of osteoporosis and vertebral fractures in postmenopausal women with type 2 diabetes mellitus and their relationship with duration of the disease and chronic complications. J Diabetes Complicat. 2011;25:216–21.CrossRefPubMedGoogle Scholar
- 100.Schwartz AV, Vittinghoff E, Margolis KL, Ambrosius WT, Bonds DE, Josse RG, et al. Intensive glycemic control and fracture risk. Diabetes. 2011;60:A372.Google Scholar
- 101.Schwartz AV, Vittinghoff E, Bonds DE, Hamilton BP, Ambrosius WT, Palermo L, et al. Intensive glycemic control not linked to falls in ACCORD. Diabetes. 2011;60:A372.Google Scholar
- 103.Standards of medical care in diabetes—2009. Diabetes Care. 2009;32 Suppl 1:S13–S61.Google Scholar
- 104.Kim JH, Jung MH, Lee JM, Son HS, Cha BY, Chang SA. Diabetic peripheral neuropathy is highly associated with non-traumatic fractures in Korean patients with type 2 diabetes mellitus. Clin Endocrinol (Oxf). 2011 [Epub ahead of print].Google Scholar
- 110.Saag KG, Zanchetta JR, Devogelaer JP, Adler RA, Eastell R, See K, et al. Effects of teriparatide versus alendronate for treating glucocorticoid-induced osteoporosis: thirty-six-month results of a randomized, double-blind, controlled trial. Arthritis Rheum. 2009;60:3346–55.CrossRefPubMedGoogle Scholar