What Are Thiazolidinediones?

Thiazolidinediones (TZDs) are oral antihyperglycemic agents that improve insulin sensitivity by activating peroxisome proliferator‑activated receptor gamma (PPAR‑γ). This nuclear receptor modulates gene expression involved in adipogenesis, glucose uptake, and lipid metabolism. The two main drugs in this class are pioglitazone and rosiglitazone. Approved by the U.S. Food and Drug Administration (FDA) in the late 1990s and early 2000s, they were widely used as second‑ or third‑line therapy for type 2 diabetes, often combined with metformin, sulfonylureas, or insulin. Despite their ability to lower hemoglobin A1c and sustain glycemic control, their use has declined because of cardiovascular safety concerns and the emergence of a substantial skeleton‑related adverse effect: an elevated risk of fractures.

The chemical structure of TZDs includes a thiazolidine‑2,4‑dione ring, which is essential for PPAR‑γ binding. Rosiglitazone (brand name Avandia) and pioglitazone (Actos) differ in their binding affinity and downstream effects—pioglitazone has a slightly weaker PPAR‑γ activation but may also interact with PPAR‑α, giving it a more favorable lipid profile. Because of these differences, the fracture risk appears consistent across both agents, though some observational studies suggest a marginally higher hazard with rosiglitazone.

Historically, TZDs were considered a breakthrough for managing insulin resistance. However, as safety data accumulated, regulatory agencies placed restrictions on their use. The FDA issued a safety communication in 2011 regarding rosiglitazone’s cardiovascular risks, and more recent warnings have highlighted the fracture danger. Today, pioglitazone is more commonly prescribed than rosiglitazone, but both require careful consideration of bone health before initiation.

The Evidence Linking TZDs to Bone Fractures

Clinical Trial Signals

The first clear evidence of an adverse bone effect came from the ADOPT trial (A Diabetes Outcome Progression Trial, 2006), which randomized 4,360 patients with newly diagnosed type 2 diabetes to rosiglitazone, metformin, or glyburide. Women receiving rosiglitazone had significantly more fractures—particularly of the upper arm, hand, and foot—than women in the comparator groups. In the RECORD trial, which evaluated rosiglitazone added to metformin or sulfonylurea, a similar pattern emerged: women experienced a 40–70% increase in fracture risk, while men showed a smaller, often non‑significant increase. The DREAM trial, assessing rosiglitazone in individuals with impaired glucose tolerance, also observed increased fractures, mostly in women.

Subsequent analyses from the ACCORD trial and from the Veterans Affairs Diabetes Trial (VADT) confirmed these findings. ACCORD reported a 30% higher fracture rate among women using rosiglitazone compared with those on other glucose‑lowering therapies. A pooled analysis of five large randomized controlled trials, published in Diabetologia in 2014, found an overall fracture odds ratio of 1.42 for women (95% CI 1.18–1.71) and 1.13 for men (95% CI 0.90–1.41).

Meta‑Analyses and Observational Data

Multiple meta‑analyses have confirmed and quantified the hazard. A 2016 Diabetes Care meta‑analysis (over 60,000 participants) reported a relative risk of 1.45 for fractures in women and 1.20 for men (the latter not reaching statistical significance). A more recent meta‑analysis from 2020, encompassing 22 studies with more than 2.5 million participants, reported a pooled relative risk of 1.35 for any fracture in women (95% CI 1.25–1.46) and 1.17 for men (95% CI 1.02–1.34), now reaching statistical significance in men as well. Observational studies have provided real‑world evidence: a Canadian cohort of more than 200,000 adults found that TZD use was associated with a 35% higher hip fracture risk in women and a 20% higher risk in men (FDA Safety Communication).

The risk appears within the first year of use and persists with continued exposure. Fracture sites are predominantly non‑vertebral: wrist, humerus, hip, and foot. Hip fractures are especially concerning because of their high morbidity and mortality in older populations. Importantly, vertebral fractures—which are characteristic of postmenopausal osteoporosis—are not significantly elevated, suggesting a distinct pattern of bone fragility induced by TZDs.

Fracture Risk Versus Bone Density Changes

Bone mineral density (BMD) declines of about 1–2% per year have been documented in TZD users, a rate comparable to early postmenopausal loss. This loss particularly affects cortical bone at the hip and forearm, consistent with the observed fracture distribution. Dual‑energy X‑ray absorptiometry (DXA) studies show that TZD‑induced BMD loss is independent of baseline BMD, meaning even patients with normal bone mass are at risk, although those with preexisting osteopenia or osteoporosis experience the greatest absolute risk increase. A 2022 systematic review in Osteoporosis International found that the average BMD loss at the total hip after two years of TZD therapy is about 2.5%—a magnitude that doubles the risk of hip fracture in elderly populations.

Notably, the fracture risk may be underestimated by BMD changes alone. Some studies suggest that TZDs also impair bone quality—specifically, they reduce bone strength indepently of density through alterations in collagen cross‑linking and microarchitecture. High‑resolution peripheral quantitative computed tomography (HR‑pQCT) studies have shown that TZD users have thinner cortices and less trabecular bone volume than non‑users matched for BMD.

Mechanisms of TZD‑Induced Bone Loss

PPAR‑γ and Mesenchymal Stem Cell Fate

PPAR‑γ is expressed in bone marrow mesenchymal stem cells, osteoblasts, and osteoclasts. When TZDs activate PPAR‑γ, they tip the balance of mesenchymal stem cell differentiation from osteoblastogenesis toward adipogenesis. This reduces the number of bone‑forming osteoblasts and increases marrow adiposity. Histomorphometry studies in animals and humans confirm decreased bone formation rates and reduced trabecular thickness. The effect is mediated primarily through the PPAR‑γ2 isoform, which is preferentially expressed in adipocytes and osteoblast precursors. In knockout mouse models, deletion of PPAR‑γ in mesenchymal stem cells completely prevents TZD‑induced bone loss, confirming the central role of this pathway.

Osteoclast Activity and Bone Resorption

In addition to suppressing formation, TZDs may increase osteoclast activity via upregulation of receptor activator of nuclear factor‑κB ligand (RANKL) and reduced osteoprotegerin levels. The net effect is a negative bone balance: bone resorption either increases or remains unchanged while bone formation declines. Markers such as procollagen type 1 N‑terminal propeptide (P1NP) drop, while C‑telopeptide (CTX) often rises or stays stable. Some studies have also shown that TZDs increase the production of sclerostin, an inhibitor of Wnt signaling that further suppresses osteoblast activity. This dual mechanism—reducing formation and increasing resorption—creates a rapid and sustained bone deficit.

Other Skeletal Effects

Beyond the PPAR‑γ pathway, TZDs may impair local production of insulin‑like growth factor 1 (IGF‑1), which normally supports bone formation. They can also alter calcium and phosphate homeostasis, although these mechanisms are less well established. Rodent studies demonstrate that TZD‑treated animals develop thinner cortices, reduced trabecular number, and compromised bone material properties, correlating with increased fracture rates in controlled loading tests. In human bone biopsies, TZD exposure is associated with decreased osteoid thickness and mineralization lag time, indicating a state of low bone turnover that may impair microdamage repair.

Impact on Bone Quality Beyond Density

Emerging evidence suggests that TZDs affect bone collagen and mineralization. Animal models show that TZD treatment increases the ratio of immature to mature collagen cross‑links, leading to reduced bone toughness. Additionally, TZD‑induced changes in bone marrow fat composition may interfere with the normal mechanosensing of osteocytes. These non‑density effects may explain why fracture risk rises more than expected from BMD decline alone.

Patient Populations at Greatest Risk

Postmenopausal Women

Sexual dimorphism is a robust finding: women, particularly those past menopause, are at substantially higher fracture risk than men. Estrogen deficiency already accelerates bone loss, and TZDs compound this effect. In ADOPT, the fracture incidence in women on rosiglitazone was 9.3 per 1,000 patient‑years versus 3.5 per 1,000 for metformin. Postmenopausal women using TZDs have roughly double the fracture risk of non‑users. The mechanism may involve estrogen’s modulation of PPAR‑γ activity; in postmenopausal women, the absence of estrogen amplifies TZD‑induced osteoblast suppression.

Older Adults

Age is an independent risk factor. Patients older than 65 years have higher absolute fracture rates, even if the relative risk increase is similar across age groups. Frailty, sarcopenia, and impaired balance increase the likelihood of falls, which often precipitate fractures. The combination of TZD‑induced bone fragility and age‑related fall risk is especially dangerous. A study from the Kaiser Permanente Northern California database found that TZD users aged 70 and older had a hip fracture incidence of 14.5 per 1,000 person‑years, compared with 8.2 per 1,000 in non‑users.

Patients with Preexisting Osteoporosis or Low BMD

Individuals with low bone mass or a prior fragility fracture are most vulnerable. TZDs can accelerate bone loss, quickly pushing patients with osteopenia into the osteoporotic range. The National Osteoporosis Foundation recommends avoiding TZDs in those with a prior fragility fracture or a T‑score below –2.5 (Bone Health & Osteoporosis Foundation). In patients with a history of wrist or hip fracture, the absolute risk increase with TZD therapy may be as high as 2–3 additional fractures per 100 patient‑years.

Duration of Therapy and Cumulative Exposure

The fracture risk is dose‑ and duration‑dependent. A meta‑analysis of six trials reported relative risks of 1.3 for ≤12 months of therapy, 1.5 for 12–24 months, and 1.7 for >24 months. Long‑term users face the greatest danger, and the risk persists even after discontinuation, though it may diminish over time. A 2018 cohort study from Denmark found that the fracture risk remained elevated for at least two years after stopping TZDs, suggesting that the bone deficit is not quickly reversible. Therefore, limiting TZD exposure is a key preventive strategy.

Diabetes itself impairs bone microarchitecture through hyperglycemia‑induced oxidative stress and accumulation of advanced glycation end‑products (AGEs). AGEs cross‑link collagen, reducing bone toughness and increasing fragility. Neuropathy and retinopathy increase fall risk. Concurrent medications such as loop diuretics, glucocorticoids, proton pump inhibitors, and selective serotonin reuptake inhibitors can further compromise bone health. A thorough medication review is essential before starting a TZD; special attention should be paid to drugs that cause orthostatic hypotension or sedation.

Regulatory Actions and Clinical Guidelines

In 2011, the FDA restricted rosiglitazone dispensing because of cardiovascular concerns, but the fracture risk was already noted on the label. Pioglitazone’s label was updated in 2016 to include a warning about bone fractures in women. The American Diabetes Association (ADA) Standards of Care now recommend that TZDs be used with caution in patients at high risk for fractures, especially postmenopausal women. The use of FRAX (Fracture Risk Assessment Tool) is encouraged before initiating therapy (WHO FRAX). Many guidelines now list TZDs as an option only after metformin and when other agents have been considered, reflecting the accumulating skeletal evidence.

Clinical Implications and Management

Pre‑Treatment Assessment

Before prescribing a TZD, clinicians should evaluate fracture risk using validated tools such as FRAX. For patients with a 10‑year major osteoporotic fracture probability exceeding 20% or with a history of fragility fracture, TZDs should be avoided. Baseline DXA is recommended for postmenopausal women, men ≥50 years, and anyone with additional risk factors. Check serum calcium, 25‑hydroxyvitamin D, and parathyroid hormone to identify and correct bone‑health deficits before initiating therapy.

Additionally, review the patient’s fall history. Those who have had two or more falls in the past year are at very high risk for fracture and should not be prescribed TZDs unless absolutely necessary and accompanied by aggressive fall prevention measures. Electromyography or nerve conduction studies may be warranted if peripheral neuropathy is suspected.

Monitoring During Therapy

For patients already on TZDs, repeat DXA every 1–2 years. Vitamin D supplementation (800–2,000 IU/day) and adequate calcium intake (1,000–1,200 mg/day from diet or supplements) are essential. Although routine bone turnover markers are not universally recommended, they can be useful in specific cases—for example, if BMD loss is rapid or if antiresorptive therapy is added. A low fasting CTX (below 100 ng/mL) may indicate suppressed bone formation and help guide decisions about discontinuing the TZD. Fall prevention strategies—including medication review to avoid sedation or hypotension, physical therapy for balance and strength, and home safety assessments—should be implemented proactively.

When to Discontinue or Add Bone‑Protective Agents

If a patient develops a low‑trauma fracture or shows significant BMD loss (e.g., >5% at the hip over 1–2 years), discontinue the TZD and transition to an alternative diabetes medication. For patients who cannot discontinue (e.g., due to failure of all other agents), consider adding an antiresorptive agent such as a bisphosphonate (alendronate, risedronate) or denosumab. A 2015 randomized trial demonstrated that alendronate 70 mg weekly prevented TZD‑associated hip BMD loss over 18 months. Limited evidence suggests that raloxifene may attenuate TZD‑induced bone loss in postmenopausal women, but it is not FDA‑approved for this purpose. Consultation with a bone health specialist is advisable in complex cases.

Fall Prevention as a Core Strategy

Because many TZD‑associated fractures result from falls, a multifaceted fall‑prevention program can substantially reduce risk. Screen for orthostatic hypotension, wean or avoid medications that cause dizziness or sedation, recommend vitamin D to improve muscle strength, and prescribe supervised exercise programs that include balance and resistance training. The American Diabetes Association (ADA) emphasizes fall risk screening as part of comprehensive diabetes care (ADA Standards of Medical Care in Diabetes). A 2019 cohort study found that TZD users who participated in fall prevention programs had a 40% lower fracture incidence than those who did not.

Alternative Diabetes Medications with Skeletal Safety

A wide array of glucose‑lowering agents are now available that have neutral or favorable bone effects. Choosing an alternative is often the simplest way to avoid TZD‑related fracture risk.

  • Metformin: First‑line therapy; multiple observational studies show no adverse fracture risk and possibly a slight protective effect against hip fractures. It remains the cornerstone of diabetes management.
  • Sulfonylureas: Neutral on bone metabolism, but carry risks of hypoglycemia and weight gain that may increase fall risk. Use with caution in older adults.
  • SGLT2 inhibitors (canagliflozin, dapagliflozin, empagliflozin): Initial concerns from the CANVAS trial about fracture risk with canagliflozin have not been replicated in larger meta‑analyses. Current evidence suggests overall neutral or even beneficial effects on BMD, possibly mediated through weight loss and improved physical function. Use with caution in patients with advanced chronic kidney disease or high fracture risk (PubMed review).
  • GLP‑1 receptor agonists (liraglutide, semaglutide, dulaglutide): These agents promote weight loss and may improve bone formation markers. Fracture risk data are mixed but largely reassuring. They are reasonable candidates for TZD replacement, especially in patients who need weight reduction.
  • DPP‑4 inhibitors (sitagliptin, saxagliptin, linagliptin): Animal studies suggest a possible reduction in osteoclast activity, and human data show no increased fracture risk. They are a neutral alternative.
  • Insulin: Exogenous insulin has no direct negative effect on bone, though hypoglycemia can increase falls. It remains a safe option for patients who require intensive glucose control and have high fracture risk.

When transitioning from a TZD, consider individual patient factors: those with obesity may benefit from GLP‑1 agonists; those with heart failure or chronic kidney disease may benefit from SGLT2 inhibitors; and those intolerant to other agents can use DPP‑4 inhibitors or sulfonylureas. In patients with established osteoporosis, the best alternatives are metformin, DPP‑4 inhibitors, or insulin, as they have the most reassuring bone data.

Practical Recommendations for Clinicians

  1. Perform a structured fracture risk assessment using FRAX and clinical risk factors before initiating a TZD. Document the risk‑benefit discussion in the medical record.
  2. Order baseline DXA for all women aged ≥50 and men aged ≥60, and for any patient with additional risk factors (prior fracture, glucocorticoid use, low body weight, family history of hip fracture).
  3. Limit TZD duration to the shortest period necessary to achieve glycemic goals. Aim for ≤12 months of therapy when possible, especially in high‑risk patients.
  4. Counsel patients explicitly about the increased fracture risk. Encourage them to report falls, fractures, and any new bone pain. Provide written educational materials.
  5. Monitor bone health annually: repeat DXA if abnormal or if other risk factors develop; check vitamin D and calcium status; supplement as needed.
  6. Prioritize fall prevention with exercise prescriptions, medication reconciliation, and home safety evaluations.
  7. Deprescribe promptly if a fragility fracture occurs or if BMD declines substantially. Switch to an alternative agent with better skeletal safety.
  8. Consult a specialist (endocrinologist, bone health specialist) for patients with established osteoporosis or those requiring continued TZD therapy with concurrent antiresorptive treatment.
  9. Stay current with evolving evidence. The FDA continues to update labeling for TZDs (FDA TZD labeling change). Review ADA Standards of Care annually.

Emerging Research and Future Directions

Recent studies are exploring whether certain TZD analogues or selective PPAR‑γ modulators can retain glycemic benefits without the bone toxicity. For example, balaglitazone and other partial agonists have shown less adipogenic effect in preclinical models. Clinical trials are needed to determine whether these agents can divorce metabolic efficacy from skeletal harm. Additionally, research into the role of PPAR‑γ in osteocyte mechanosensing may lead to co‑therapies that protect bone while allowing TZD use. In the meantime, the safest approach remains to use TZDs only in patients with low fracture risk and to monitor bone health diligently.

Conclusion

Thiazolidinediones remain a pharmacologically distinct class that can effectively improve insulin sensitivity and glycemic control in type 2 diabetes. However, the well‑established elevation of fracture risk—driven by PPAR‑γ‑mediated suppression of bone formation, increased bone resorption, and accelerated BMD loss—requires careful patient selection and systematic monitoring. Women (especially postmenopausal), older adults, and those with preexisting bone deficits face the highest danger. With a rich armamentarium of alternative diabetes medications now available—most of which have neutral or beneficial skeletal effects—TZDs should be reserved for carefully chosen patients in whom the glycemic benefits clearly outweigh the skeletal risks, and even then used at the lowest effective dose for the shortest necessary duration. By integrating bone health into routine diabetes care, clinicians can reduce the burden of fragility fractures and improve overall outcomes for patients with type 2 diabetes.