Fractures present an under-recognized complication of diabetes
Diabetes is a disease whose seriousness is attributable as much to its complications as to the immediate effects of uncontrolled glucose levels, and patients rightly focus on reducing their risks for heart disease and stroke. Although bone fractures contribute substantially to mortality and decreased quality of life, this complication is often at the bottom of the list of conditions addressed by health care providers.
“When you think about diabetic complications, you think about angiopathy, neuropathy, retinopathy, accelerated cardiovascular risk — nobody talks about fractures or low bone density, and I think it’s a major issue,” Clifford J. Rosen, MD, senior scientist and director for the Center for Clinical and Translational Research at Maine Medical Center Research Institute in Scarborough, told Endocrine Today. “There’s certainly a lot of morbidity associated with fractures, especially hip fractures.”
According to Sundeep Khosla, MD, a past president of the American Society for Bone and Mineral Research, that organization is partnering with the American Diabetes Association to raise awareness of increased fracture risk in diabetes.
“We would like for the diabetes community to think about bone as yet another diabetic complication,” Khosla, professor of medicine and physiology at Mayo Clinic in Rochester, Minnesota, told Endocrine Today. “Poor glycemic control likely alters bone material properties and bone quality so that — particularly in type 2 diabetes, but likely also in type 1 diabetes — the relationship between the mass of bone and the strength of bone is altered and you fracture despite having a relatively normal bone density.”
Endocrine Today interviewed several experts about the risks for fracture among patients with diabetes, the quality of diabetic bone and the challenges of managing diabetes in patients with bone fragility.
Increased fracture risk
Several studies attest to elevated fracture risk among patients with type 2 diabetes: 20% higher risk at any site among postmenopausal women with type 2 diabetes vs. those without the disease in the Women’s Health Initiative observational study; 64% higher in men and women with type 2 diabetes in the Health, Aging and Body Composition Study. Meta-analyses have shown double to triple the risk for hip fracture among adults with type 2 diabetes.
Diabetes severity and duration have been shown to affect risk. In the Singapore Chinese Health Study, researchers found a 40% increased risk for hip fracture among participants with diabetes for less than 5 years, rising to 266% among those with diabetes for at least 15 years vs. those without the disease.
Results of a recent study from Canada confirmed these results among women with type 2 diabetes. Using data from a clinical DXA registry linked to Manitoba administration databases on 8,840 women with type 2 diabetes and 49,048 women without diabetes, Sumit R. Majumdar, MD, MPH, associate professor in the department of medicine at the University of Alberta, and colleagues found increased risk for a major osteoporotic fracture among women with diabetes, rising from 7% for disease duration less than 5 years to 34% for duration more than 10 years. These women also had increased risk for hip fracture: more than 50% for diabetes duration less than 5 years and more than 90% for longer than 10 years’ duration.
Increased risk for fracture, beginning in childhood and extending throughout life, is also associated with type 1 diabetes. In a recent study, David R. Weber, MD, MS, of Golisano Children’s Hospital, University of Rochester School of Medicine and Dentistry, and colleagues evaluated data from The Health Improvement Network on 30,394 participants aged 0 to 89 years with type 1 diabetes and 303,872 controls without diabetes.
An increased risk for fracture was found for participants with type 1 diabetes in all age categories, rising from 14% among men younger than 20 years to 218% in those aged 60 to 69 years vs. controls. In women, risk was elevated 35% in the group younger than 20 years to more than double among those aged 40 to 49 years.
Bone fragility–diabetes connection
Although experts have little doubt that people with diabetes are at increased risk for fracture, demonstrating a causal relationship has been difficult.
To some extent, fractures can be explained by other diabetes complications that limit mobility in these patients, such as peripheral neuropathy, poor proprioception, vasculopathy and retinopathy, which might also explain fracture placement, according to Mone Zaidi, MD, PhD, professor of medicine, endocrinology, diabetes and bone disease at Mount Sinai Hospital.
“They seem to trip and fall and hurt themselves in the wrong regions,” Zaidi told Endocrine Today. “They fracture at the ankle, the toes, the fingers and at other unusual sites that are not normally sites of fractures, for example, in postmenopausal osteoporosis. That is an interesting phenomenon, which to some extent has been attributed to [peripheral neuropathy], but I don’t think it is the sole cause.”
Rosen agrees. “Is that a function of just falls and neuropathy or vasculopathy, or is it truly something that is related to how the distal skeleton may be different from the spine, for example?” he asked.
Bone disease resulting from diabetes is multifactorial, Rosen said. Besides the toxic effects on the skeleton of prolonged high glucose levels and excess lipids, the atherosclerosis present in patients with diabetes may lead to vascular changes that affect the skeleton; and neuropathic changes can also cause damage. His research group is also examining the role of inflammation in fat mass and possible effects of an individual’s microbiome on bone.
“How these different components come together to produce diabetic bone disease is very difficult to reconcile,” Rosen said.
Furthermore, the disease process appears to be different between type 1 and type 2 diabetes. According to Zaidi, type 1 diabetes can impair bone formation through reduced insulin and insulin-like growth factor. Those whose diabetes develops during childhood or adolescence may not accrue bone; their bone density remains low, and their fracture risk may remain low. Those who develop type 1 diabetes at an older age may lose bone when the formation stimulus is lost; their fracture risk is higher.
“There are physiologic differences in bone quantity primarily in diabetics who do not accrue as much bone vs. in diabetics who have lost bone after they begin to have type 1 diabetes,” Zaidi said. “Fracture risk in both circumstances, though, is not as high as in type 2 diabetics.”
Paradoxically, patients with type 2 diabetes have an increased risk for fragility fractures despite often excess body weight and normal or higher bone mineral density, factors that tend to be protective for fracture.
“BMD is just one of the key factors that determine facture risk, and there are many others that determine the high risk for fracture,” Zaidi said. “So you can get an increase in bone density, but still, bone becomes more brittle.
“That is a clinical paradox but an adequately explainable clinical paradox because of changes in material properties of bone in diabetics that are not accounted for by bone density changes,” Zaidi said.
This suggests that there are changes in bone quality in these patients, likely the result of greater exposure to the hyperglycemia and reduced insulin that characterize diabetes, according to Mario Skugor, MD, FACE, of the department of endocrinology, diabetes and metabolism at Cleveland Clinic and co-director of the endocrinology and reproductive organ block at Cleveland Clinic Lerner College of Medicine of Case Western Reserve University. Changes in bone formation and resorption can decrease bone turnover and result in denser but more brittle bones.
Similar to patients with diabetes, those with prediabetes tend to have greater body weight and higher BMD, but their fracture risk is not elevated compared with the general population. “But then they develop diabetes, and as time goes on they lose their protection,” Skugor told Endocrine Today.
Thus, fracture risk in patients with diabetes is not associated with BMD in the same way as in those without the disease. In a 2011 meta-analysis, Ann V. Schwartz, PhD, professor in the department epidemiology and biostatistics at the University of California, San Francisco School of Medicine, and colleagues found a higher fracture risk for a given T-score or WHO fracture risk algorithm score among older adults with type 2 diabetes compared with patients without diabetes.
“This has led to the hypothesis that it’s not just the mass of bone, but the quality of the bone that’s abnormal in diabetes,” Khosla said. “And to test this hypothesis, several years ago we used a novel tool to assess bone material properties in patients with diabetes.”
Khosla and colleagues used a reference point indentation technique that employs a hand-held instrument called an OsteoProbe (ActiveLife Scientific). The device is not FDA approved, but is being used for commercial research applications; it works like a center punch for setting nails. After numbing the skin on the tibia, the user inserts the probe through the periosteum and makes a series of microscopic cracks, which are measured and compared against a standardized reference to give a bone material strength index (BMSi) value.
“Essentially, the deeper you can poke the bone, the worse the material properties of the bone,” Khosla said.
In the proof-of-concept study, Khosla and colleagues compared BMD and BMSi of 30 postmenopausal women with diabetes for at least 10 years with the values for 30 postmenopausal women without diabetes.
“As with other studies, there was no difference in bone density between the diabetics and the nondiabetics — if anything at some sites the diabetics had a higher bone density — but the diabetics clearly had a worse bone material property, as assessed by the instrument. ... This really suggests that the bone quality is impaired in diabetics.”
A recent study conducted by Jessica R. Furst, MD, an endocrinology fellow at Columbia University College of Physicians and Surgeons, and colleagues confirmed these results and correlated them with advanced glycation end-products (AGEs), determined by skin autofluorescence, and HbA1c.
“Admittedly, they didn’t measure AGEs in bone,” Khosla said. “But they were able to measure it in another related tissue — in skin, which also has collagen. ... [It is] very plausible that chronic poor [glycemic] control leads to AGE accumulation in bone, which impairs the material properties of bone and also alters cellular remodeling, and the net effect of that is, despite having relatively normal bone mass, your bone is more brittle and you fracture more.”
Effects of diabetes therapies on bone quality
Not only is the disease process likely to impair bone quality, but also implicated are some of the medications used to treat diabetes.
“This is particularly true for the thiazolidinediones, which probably work by shifting precursor cells that might want to become bone-forming osteoblasts to becoming, instead, adipocytes,” Nelson Watts, MD, FACP, MACE, director of Mercy Health Osteoporosis and Bone Health Services in Cincinnati, told Endocrine Today.
“Some recent research with the SGLT2 inhibitor canagliflozin suggested an early increase in fractures — peripheral fractures, not the ones we usually associate with osteoporosis, and fractures that may have little or no clinical significance,” Watts said.
The FDA warned last fall that the SGLT2 inhibitors canagliflozin (Invokana, Janssen) and canagliflozin/metformin hydrochloride (Invokamet, Janssen) may cause an increased risk for bone fractures and that the agency is continuing to examine other drugs in the class to determine whether they also increase the risk for fractures.
“There was some evidence that the new SGLT-2 inhibitors had increased glucose in the urine, and may actually have a detrimental effect on bone, possibly by changing the electrolyte or sodium concentration around the osteoblast of the bone-forming cells,” Rosen said. “But those are all theoretical.”
Watts and colleagues evaluated fracture risk presented by canagliflozin in a meta-analysis of nine yearlong or longer randomized, placebo- or active-controlled, double blind phase 3 trials of the agent (n = 10,194), including the pivotal Canagliflozin Cardiovascular Assessment Study (CANVAS; n = 4,327). The researchers found an increased fracture risk with canagliflozin (2.7%) compared with placebo and other comparator agents (1.9%), but the elevated risk disappeared when CANVAS was removed from the analysis (risk, 1.7% for canagliflozin vs. 1.5% for non-canagliflozin). CANVAS participants were older, with a history of or higher risk for cardiovascular disease and decreased kidney function, two conditions treated with medications that can increase risk for falls and, therefore, fractures, the researchers noted.
“Although the cause of the increased fracture risk with canagliflozin is unknown, the small, inconsistent changes in total hip BMD ... and the fact that an early increase in fractures was observed in only a subgroup of patients treated with canagliflozin suggest that extrinsic factors related to canagliflozin, possibly related to falls or other indirect effects of canagliflozin on bone strength, may be a more likely explanation for this observed imbalance,” the researchers wrote.
Other drug classes present neutral effects on bone or may even prove to be protective. For example, in an analysis of fracture data from the TECOS cardiovascular outcomes trial of the DPP-IV inhibitor sitagliptin (Januvia, Merck), Robert G. Josse, BSc, MBBS, FRCP, FRCPC, FACP, FACE, associate scientist at Li Ka Shing Knowledge Institute and professor of medicine at the University of Toronto, and colleagues found that the agent conferred no added risk or benefit for incident fractures, major osteoporotic fractures or hip fractures in adults with type 2 diabetes.
“Another large DPP-IV inhibitor study (SAVOR-TIMI) also did not find an increase in fractures compared with placebo,” Josse told Endocrine Today.
“Original observational database meta-analyses suggest that the DPP-IV inhibitors might be associated with a decrease in fractures, and there is some biological plausibility for that suggestion,” Josse said. “More fracture studies with incretins, like [glucagon-like peptide-1] and GLP-2 analogues, are required to determine whether these incretins can decrease fracture risk in patients with diabetes with or without prior fractures. More mechanistic studies are also necessary to explore pathophysiology.”
Good evidence for the effects on bone of the most commonly used agents for diabetes management — insulin and metformin — are not available.
“With insulin treatment, patients who achieve low HbA1c [levels] with type 2 diabetes may be more likely to fall, and part of the increased fracture risk in patients with diabetes is due to falling,” Watts said. “Clinicians with large volumes of patients with diabetes should be aware of this and consider fracture risk assessment early on in the management, but certainly periodically thereafter.”
Rosen’s research group has returned to studying metformin recently. “All unpublished stuff, but we have evidence that metformin is probably actually good for the skeleton and that it may have effects on both the bone-forming and the bone-reabsorbing cells in positive ways,” he said.
Surgical treatment of diabetes through bariatric surgery also appears to affect bone metabolism. “When you treat with gastric surgery or gastric bypass, people do lose bone,” Rosen said. “They lose a lot of fat mass, their diabetes is cured, transiently, maybe permanently, but in that you’re also losing some bone. We need to understand what that mechanism is as well.”
According to Skugor, the available evidence is not yet enough to recommend against prescribing particular diabetes therapies. “It would not stop me from using pioglitazone because the increased risk may be mitigated by benefits,” he said.
Treatments for osteoporosis have not been evaluated in patients with diabetes, and in fact, clinical trials for these therapies have most often excluded patients with severe diabetes. With only limited data from retrospective studies available, Rosen said he treats people with diabetes with conventional osteoporosis medications.
Similarly, Khosla recommends using these agents and maintaining a lower threshold for prescribing because patients with diabetes have increased fracture risk for any given bone density. The WHO fracture risk assessment tool (FRAX) does not adjust for diabetes, and so it tends to underestimate the risk it confers and to underestimate the further risk from longer-duration disease.
“One way to correct FRAX for diabetes is to click the rheumatoid arthritis button,” Khosla said. “I’ve been talking to my clinical colleagues about that and doing that myself.”
Khosla attributes the practice to the population studies conducted by William Leslie, MD, MSc, FRCPC, professor of medicine and radiology at the University of Manitoba in Winnipeg.
Obtaining a trabecular bone score is another option to adjust for diabetes risk, Khosla said. “There is evidence showing that the trabecular bone score, which is a very specific measurement of bone texture, is worse in diabetics than nondiabetics, so you could either use the rheumatoid arthritis button or you can go into FRAX and do the trabecular bone score adjustment for your patient. That will more appropriately reflect the fracture risk in the diabetic patients,” he said.
For patients with diabetes, Skugor attempts to maximize bone health in general by optimizing calcium and vitamin D levels, improving physical fitness and treating other conditions that may affect bone health, such as celiac disease or hypogonadism. “There are a lot of dominoes to line up,” he said.
Directions for research
Researchers are now attempting to sort out cellular, molecular and biomechanical factors involved in diabetic bone disease.
Whether better glycemic control translates to better bone quality has yet to be determined, according to Rosen. His research group is conducting studies of cells exposed to extremely high glucose levels and, at least in the preliminary data, seeing dysfunctional glucose utilization.
Rosen is also interested in effects of diet and microbiome on bone quality. “Before the onset of insulin, when an individual in the late 19th century or early 20th century got diabetes, they died. But before they died, and before insulin, they used to starve them, in hopes that by reducing their sugar load that would be enough,” he said.
“But what we’re seeing is that if you feed mice that have no insulin, you actually can protect their skeletons. ... We’re trying to understand how fat and, particularly, dietary content can impact the skeleton both through direct impact, but also maybe through the microbiome or through other mechanisms,” Rosen said.
Research has also begun to show that bone can affect glucose homeostasis — that not only does diabetes affect bone, but bone affects diabetes. Gerard Karsenty, MD, PhD, professor and chair in the genetics and development department at Columbia University, and colleagues have shown that, at least in mice, bone regulates insulin action through osteocalcin.
“The evidence from the mouse models is fascinating,” Khosla said. “It’s up to many of us as clinical investigators to begin to test those concepts in humans.” – by Jill Rollet with additional reporting by Regina Schaffer
- Bilik D, et al. J Clin Endocrinol Metab. 2010;doi:10.1210/jc.2009-2638.
- Bonds DE, et al. J Clin Endocrinol Metab. 2006;91:3404-3410.
- Farr JN, Khosla S. Bone. 2016;doi:10.1016/j.bone.2015.07.027.
- Furst JR, et al. J Clin Endocrinol Metab. 2016;doi:10.1210/jc.2016-1437.
- Gennari L, et al. J Clin Endocrinol Metab. 2012;doi:10.1210/jc.2011-2958.
- Josse RG, et al. Diabetes Obes Metab. 2016;doi:10.1111/dom.12786.
- Koh WP, et al. Diabetes Care. 2010;doi:10.2337/dc10-0067.
- Majumdar SR, et al. J Clin Endocrinol Metab. 2016;doi:10.1210/jc.2016-2569.
- Rico H, et al. Calcif Tissue Int. 1989;45:71-73.
- Schwartz AV, et al. J Bone Miner Res. 2005;20:596-603.
- Schwartz AV, et al. JAMA. 2011;doi:10.1001/jama.2011.715.
- Weber DR, et al. Diabetes Care. 2015;doi:10.2337/dc15-0783.
- Winkler DG, et al. EMBO J. 2003;22:6267-6276.
- For more information:
- Robert G. Josse, BSc, MBBS, FRCP, FRCPC, FACP, FACE, can be reached at St. Michael’s Hospital, 61 Queen St. E., 6th Floor, Toronto, Ontario, M5C 2T2; email: firstname.lastname@example.org.
- Sundeep Khosla, MD, can be reached at Mayo Clinic, 200 First St. SW, Rochester, MN 55902; email: email@example.com.
- Clifford J. Rosen, MD, can be reached at Maine Medical Center, 81 Research Drive, Scarborough, ME 04074; email: firstname.lastname@example.org.
- Mario Skugor, MD, FACE, can be reached at Cleveland Clinic Main Campus, 9500 Euclid Ave., Cleveland, OH 44195.
- Nelson Watts, MD, FACP, MACE, can be reached at 4760 E. Galbraith Road #212, Cincinnati, OH 45236; email: email@example.com.
- Mone Zaidi, MD, PhD, can be reached at Icahn School of Medicine at Mount Sinai, 5 E. 98th St., 3rd Floor, New York, NY 10029; email: firstname.lastname@example.org.
Disclosure: Khosla, Rosen and Skugor report no relevant financial disclosures. Josse reports receiving grants or personal fees from Amgen, AstraZeneca, Boehringer Ingelheim, Eli Lilly, Janssen and Merck. Watts reports serving formerly as an unpaid consultant to Janssen.