As diabetes prevalence continues climbing worldwide, affecting more than 530 million adults according to the International Diabetes Federation, the conversation around long-term medication effects has expanded beyond glycemic control to include skeletal health. Recent evidence suggests that certain glucose-lowering drugs may accelerate bone density loss, raising fracture risk in vulnerable populations. This concern is particularly pressing given that diabetes itself compromises bone quality through hyperglycemia-related mechanisms, creating a compounding risk factor that demands clinical attention.

The relationship between diabetes pharmacotherapy and bone metabolism is complex and drug-specific. Some medications directly interfere with bone remodeling pathways, while others create downstream effects on calcium and phosphate homeostasis. For clinicians managing diabetes in aging populations, understanding these distinctions is essential for preserving both metabolic and skeletal health over the long term.

How Diabetes Drugs Affect Bone Density: Mechanisms and Evidence

Thiazolidinediones (TZDs): A Clear Bone Risk

Thiazolidinediones, including pioglitazone and rosiglitazone, have been the subject of longstanding concern regarding bone health. These drugs activate peroxisome proliferator-activated receptor gamma (PPARγ), a nuclear receptor that plays a central role in adipocyte differentiation. Within bone marrow stromal cells, PPARγ activation shifts the differentiation balance toward adipogenesis and away from osteoblastogenesis. The net result is a reduction in bone-forming cell production, leading to decreased bone formation and lower bone mineral density (BMD).

Clinical evidence consistently demonstrates that TZD use is associated with a 1–2 percent annual decline in BMD at both the lumbar spine and hip. Pooled analyses from randomized trials indicate a 2- to 3-fold increase in fracture risk among women using these agents, with the risk appearing dose-dependent and duration-dependent. Fractures most commonly occur at the distal forearm, humerus, and foot—sites rich in cortical bone. While men may experience similar BMD declines, the fracture signal in male TZD users has been less pronounced in available studies.

Current clinical practice guidelines from the American Diabetes Association recommend avoiding TZDs in patients with established osteoporosis or those at high fracture risk. For patients already taking a TZD who develop bone loss, transitioning to an alternative agent is warranted.

SGLT2 Inhibitors: Class-Specific Considerations

Sodium-glucose cotransporter-2 inhibitors have become cornerstone therapies in type 2 diabetes management due to their demonstrated cardiovascular and renal benefits. However, bone safety signals, particularly with canagliflozin, have prompted regulatory scrutiny and clinical caution.

Canagliflozin has been associated with a 0.5–1 percent reduction in hip BMD over one to two years of treatment in multiple studies. The CANVAS and CREDENCE trial programs reported a 23 percent increased risk of fracture with canagliflozin, predominantly affecting the upper and lower extremities. The proposed mechanism involves SGLT2 inhibition in the proximal renal tubule, which alters phosphate handling and increases serum phosphate levels. This triggers a compensatory rise in fibroblast growth factor 23 (FGF23), which suppresses renal 1α-hydroxylase activity. The resulting reduction in active vitamin D levels impairs intestinal calcium absorption, leading to secondary hyperparathyroidism and increased bone turnover.

Notably, the fracture risk appears specific to canagliflozin rather than a class-wide effect. The DECLARE-TIMI 58 trial with dapagliflozin and the EMPA-REG OUTCOME trial with empagliflozin did not demonstrate increased fracture rates. This suggests that structural differences among SGLT2 inhibitors may influence bone effects, or that the degree of phosphate elevation varies between agents. Regardless, regulatory agencies including the U.S. Food and Drug Administration advise caution when using canagliflozin in patients with fracture risk factors, particularly those with impaired kidney function or advanced age.

Insulin Therapy: Unraveling a Complex Relationship

Insulin presents a paradox in bone health. In vitro and animal models demonstrate that insulin has anabolic effects on osteoblasts, and some cross-sectional studies suggest higher BMD in insulin-treated patients. However, large epidemiological analyses tell a different story. A 2020 meta-analysis of observational studies involving over 300,000 participants found that insulin users had a 30–45 percent higher fracture risk compared to non-insulin users, even after adjusting for age, body mass index, and diabetes duration.

Understanding this discrepancy requires careful attention to confounding factors. Patients on insulin therapy typically have longer diabetes duration, poorer glycemic control, and higher rates of diabetic complications including neuropathy, retinopathy, and nephropathy. These complications independently increase fall risk and fracture susceptibility. When studies adjust for hemoglobin A1c, fall history, and comorbidity burden, the excess fracture risk attributable to insulin itself diminishes substantially.

Additionally, hypoglycemia remains a significant concern with insulin therapy, particularly in older adults using multiple daily injections. Hypoglycemic events can cause falls, dizziness, and altered mental status, directly increasing fracture risk. Basal insulin analogs with more stable pharmacokinetics, such as insulin glargine U100, insulin degludec, and insulin glargine U300, offer lower hypoglycemia rates compared to NPH insulin or premixed formulations, potentially reducing fall-related fractures.

Despite these nuances, clinicians should not dismiss the possibility of direct skeletal effects from chronic hyperinsulinemia. Some evidence suggests that sustained high insulin levels may downregulate insulin-like growth factor 1 signaling in bone or promote osteoclast activity through alternative receptor pathways. Until more definitive data emerge, a prudent approach includes bone health assessment in any patient on long-term insulin therapy, particularly those with additional fracture risk factors.

GLP-1 Receptor Agonists: Potential Bone Benefits

Glucagon-like peptide-1 receptor agonists, including semaglutide, liraglutide, dulaglutide, and exenatide, have emerged as promising agents with potentially favorable bone effects. GLP-1 receptors are expressed on osteoblasts and osteoclasts, and activation of these receptors may stimulate bone formation while inhibiting resorption in preclinical models.

Clinical data, while not derived from trials specifically designed to assess bone outcomes, are encouraging. Secondary analyses from the LEADER trial with liraglutide and the SUSTAIN-6 trial with semaglutide suggested reduced fracture rates in the active treatment groups compared to placebo. These benefits may be mediated through multiple mechanisms: direct receptor activation on bone cells, improved inflammatory profiles, weight loss that reduces skeletal loading stress, and enhanced muscle strength that improves balance and reduces falls.

However, clinicians should interpret these findings cautiously until dedicated bone studies with BMD endpoints are completed. The weight loss associated with GLP-1 agonists can transiently reduce BMD at the hip and spine, although this effect appears to plateau after 6–12 months and does not negate overall fracture reduction in available analyses.

DPP-4 Inhibitors and Metformin: Bone-Neutral Options

Dipeptidyl peptidase-4 inhibitors (sitagliptin, linagliptin, saxagliptin, alogliptin) have consistently demonstrated neutral effects on bone density and fracture risk across cardiovascular outcome trials. Their weight-neutral profile and minimal hypoglycemia risk make them suitable options for older adults concerned about skeletal health.

Metformin, the cornerstone of type 2 diabetes management, also appears bone-friendly. Observational studies suggest that metformin users have lower fracture rates compared to those using other oral agents. Proposed mechanisms include improved insulin sensitivity, reduced oxidative stress in bone tissue, and possible direct osteogenic effects through AMPK activation. As the only agent with a combined evidence base for safety, efficacy, and bone neutrality, metformin remains the foundation upon which other therapies should be built.

Clinical Evidence Quantifying Bone Density Reduction

The association between diabetes medications and bone loss is supported by a comprehensive body of clinical research. A 2021 systematic review and network meta-analysis encompassing 38 randomized controlled trials and over 70,000 participants quantified the effects of various glucose-lowering drugs on BMD and fracture outcomes.

For thiazolidinediones, the analysis demonstrated a 1.1 percent greater annual decline in lumbar spine BMD compared to placebo, with a hazard ratio for fractures of 1.56 (95% CI, 1.32–1.85). The effect was most pronounced in postmenopausal women, who showed a nearly 3-fold increase in distal forearm fractures. For canagliflozin, the pooled hazard ratio for fractures was 1.23 (95% CI, 1.06–1.44), driven primarily by upper and lower extremity events.

Beyond these medication-specific analyses, it is important to recognize that diabetes itself confers independent fracture risk. Patients with type 1 diabetes have a 3- to 6-fold increased hip fracture risk compared to the general population, while those with type 2 diabetes have a 1.3- to 2-fold increase, even after adjusting for BMD. This phenomenon—termed diabetic bone disease—results from poor bone microarchitecture, increased cortical porosity, and accumulation of advanced glycation end products that weaken collagen cross-links. The added burden of medication-induced bone loss can push patients past a critical threshold where fractures become clinically manifest.

Identifying At-Risk Populations

Not every patient using these medications will experience bone loss, but certain populations warrant heightened vigilance. Postmenopausal women are at baseline risk for osteoporosis due to estrogen withdrawal, and the addition of TZDs or canagliflozin can accelerate bone loss beyond what is expected from aging alone. Men over age 70, individuals with low body weight (body mass index less than 21 kg/m²), patients with prior fragility fractures, and those using glucocorticoids concurrently are also at elevated risk.

Duration of therapy is another critical variable. BMD loss with TZDs becomes measurable within 6–12 months of initiation and continues at a steady rate for at least 2–3 years. For SGLT2 inhibitors, the effects on bone may emerge within 12–18 months. Patients requiring these agents beyond these timeframes should be considered for bone health monitoring.

Renal function also modifies risk. Patients with chronic kidney disease (estimated glomerular filtration rate below 60 mL/min/1.73 m²) are more susceptible to the phosphate and FGF23 effects of SGLT2 inhibitors, potentially amplifying bone turnover. Similarly, patients with diabetes and coexisting hyperparathyroidism or vitamin D deficiency are at compounded risk when exposed to medications that further dysregulate mineral metabolism.

Practical Strategies for Preserving Bone Health

Baseline Assessment and Monitoring

Proactive bone health evaluation should be integrated into routine diabetes care. For patients initiating or continuing TZDs or canagliflozin, consider the following approach:

  • Dual-energy X-ray absorptiometry (DXA) scanning: Obtain a baseline DXA of the lumbar spine and hip before starting high-risk medications, with repeat scanning after 12–18 months to quantify BMD changes. For patients already established on therapy, a baseline scan is still warranted to document current bone status. The Fracture Risk Assessment Tool (FRAX), available from the University of Sheffield, can integrate BMD data with clinical risk factors to estimate 10-year fracture probability.
  • Laboratory evaluation: Measure serum calcium, phosphate, 25-hydroxyvitamin D, intact parathyroid hormone, and creatinine at baseline. For patients on canagliflozin, monitor these parameters every 6–12 months, as rising PTH may indicate emerging vitamin D insufficiency that requires supplementation.
  • Fall risk screening: Simple screening questions about recent falls, gait instability, or use of walking aids can identify patients who may benefit from physical therapy or occupational therapy evaluation. Insulin users should be specifically queried about hypoglycemia frequency and timing.

Nutritional Optimization

Adequate calcium and vitamin D intake forms the foundation of bone health preservation. The Institute of Medicine recommends 1,000–1,200 mg of total calcium daily from dietary sources and supplements. Dietary calcium from dairy products, fortified plant-based milks, leafy green vegetables, and calcium-set tofu is preferred, with supplementation limited to 500–600 mg per day for those who cannot meet needs through diet alone, due to concerns about cardiovascular safety with high-dose calcium supplements.

Vitamin D requirements may be higher in patients using SGLT2 inhibitors due to reduced renal 1α-hydroxylase activity. Target serum 25-hydroxyvitamin D levels of 30 ng/mL or higher are reasonable, often requiring 1,000–2,000 IU daily of vitamin D3. For patients with documented deficiency or those on canagliflozin, doses up to 4,000 IU daily may be needed.

Magnesium and vitamin K are also important for bone health, though routine supplementation beyond dietary intake is not recommended unless deficiency is documented.

Exercise Interventions

Weight-bearing and resistance exercises stimulate bone formation through mechanical loading and improve muscle strength and balance. A comprehensive program should include:

  • Weight-bearing aerobic activity: Brisk walking, jogging, stair climbing, dancing, or hiking for at least 30 minutes on most days of the week. For patients with gait instability, supervised walking programs or water-based exercises provide safer alternatives.
  • Resistance training: Two to three sessions per week using resistance bands, free weights, or weight machines focusing on major muscle groups. Progressive overload—gradually increasing the weight or resistance—optimizes bone stimulation.
  • Balance training: Tai chi, yoga, or specific balance exercises (single-leg stands, heel-to-toe walking) reduce fall risk and are particularly valuable for older adults.

Patients with established osteoporosis or prior vertebral fractures should avoid high-impact activities (jumping, running) or exercises involving spinal flexion (bent-over rows, sit-ups) that could increase fracture risk. Consultation with a physical therapist or exercise physiologist experienced in bone health is recommended.

Medication Management and Alternatives

When medication-induced bone loss is identified, clinicians should consider modifying the diabetes regimen. Several alternatives maintain glycemic control without compromising skeletal health:

  • Metformin: As first-line therapy, metformin offers a favorable bone profile and should be continued in all patients unless contraindicated or not tolerated. Combining metformin with other agents can reduce the need for higher-risk medications.
  • DPP-4 inhibitors: These agents are effective, well-tolerated, and bone-neutral, making them appropriate add-on therapy, particularly in older adults where hypoglycemia and bone safety are primary concerns.
  • GLP-1 receptor agonists: Beyond potential bone benefits, these drugs provide cardiovascular risk reduction, weight loss, and low hypoglycemia risk. Semaglutide, liraglutide, and dulaglutide offer potent glycemic effects that may allow discontinuation of TZDs or insulin.
  • Switching SGLT2 inhibitors: For patients requiring SGLT2 inhibition for cardiorenal protection, dapagliflozin or empagliflozin may be preferable to canagliflozin due to the absence of fracture signals in large trials.

Osteoporosis pharmacotherapy should be considered when BMD T-scores fall below -2.5 at the hip or spine, or when FRAX 10-year probabilities exceed treatment thresholds (typically 20 percent for major osteoporotic fracture or 3 percent for hip fracture in the United States). Bisphosphonates remain first-line therapy, with denosumab, teriparatide, or romosozumab reserved for higher-risk patients or those intolerant of bisphosphonates.

Emerging Research and Future Directions

The field of diabetic bone disease is evolving rapidly. Researchers are investigating novel biomarkers that could identify patients at risk for medication-induced bone loss before BMD declines are detectable. Circulating osteocyte markers such as sclerostin and Dickkopf-1 may eventually guide treatment decisions, predicting which patients will lose bone on TZDs or SGLT2 inhibitors.

Newer insulin formulations and delivery systems, including ultra-long-acting analogs and closed-loop systems, aim to reduce hypoglycemia rates and may indirectly lower fracture risk by preventing falls. Similarly, novel non-insulin agents such as tirzepatide—a dual GIP and GLP-1 receptor agonist—show promise for glycemic control and weight reduction without bone safety signals in early trials, though long-term data are awaited.

Regulatory agencies continue to monitor post-marketing bone safety data for all diabetes medications. The European Medicines Agency recently updated prescribing information for canagliflozin to include bone health warnings, and similar updates have been applied to some TZD labels. Clinicians should consult current prescribing information and remain vigilant for safety communications from regulatory bodies.

Integrating Bone Health into Diabetes Care

Preserving skeletal health alongside glycemic control requires a systematic clinical approach. The American Diabetes Association now recommends that fracture risk assessment be considered in all patients aged 65 years or older with diabetes, and in younger patients with risk factors. DXA screening is advised for women aged 65 and men aged 70, with earlier screening for those with high-risk features.

For patients using TZDs or canagliflozin, the threshold for DXA screening should be lowered by 5–10 years, and repeat scanning intervals shortened to 12–18 months rather than the standard 2–3 years. Any patient who sustains a fracture while on these medications should undergo immediate bone health evaluation and consideration of treatment modification.

Shared decision-making is essential when balancing the metabolic benefits of diabetes medications against their skeletal risks. For many patients, the cardiovascular and renal benefits of SGLT2 inhibitors outweigh the bone concerns, particularly if preventive measures are implemented. Similarly, TZDs may still have a role in selected patients with insulin resistance who cannot tolerate other agents, provided bone health is monitored proactively.

Conclusion

The potential for bone density reduction with long-term use of certain diabetes medications represents a clinically significant concern that warrants integration into routine diabetes management. Thiazolidinediones and canagliflozin carry established risks for accelerated bone loss and fracture, while insulin therapy contributes risk primarily through hypoglycemia-related falls. Metformin, DPP-4 inhibitors, and GLP-1 receptor agonists offer bone-friendly alternatives for most patients.

A proactive approach incorporating baseline bone health assessment, regular monitoring, nutritional optimization, exercise prescription, and thoughtful medication selection can mitigate these risks while maintaining glycemic control. As the diabetes pharmacopeia continues to expand, maintaining awareness of skeletal side effects will help preserve both metabolic health and bone integrity across the lifespan.

Clinicians and patients seeking additional information can consult guidelines from the American Diabetes Association, the Endocrine Society, and the Bone Health & Osteoporosis Foundation. A comprehensive scientific review on diabetes and bone disease is available in the Journal of Clinical Endocrinology & Metabolism. Additional guidance on FRAX fracture risk assessment can be found at the University of Sheffield FRAX website.