Study suggests SGLT2 inhibitors may subtly disrupt bone signalling pathways before detectable changes occur

Research indicates that sodium-glucose cotransporter-2 inhibitors could influence bone biology earlier than standard tests reveal, raising questions about their long-term skeletal safety.

A new study in the journal Diabetes, Obesity and Metabolism adds to the debate over whether sodium-glucose cotransporter-2 inhibitors affect bone health, suggesting that the drugs may alter a key signalling pathway before standard bone tests show any change.

In a propensity score-matched analysis of 23 users and 23 non-users, researchers found that women taking SGLT2 inhibitors had lower levels of neuropilin-1, or NRP1, and a higher semaphorin 3A to NRP1 ratio. That pattern points to reduced signalling through the Sema3A/NRP1/Plexin-A1 axis, which helps regulate bone formation and breakdown. By contrast, conventional measures such as bone turnover markers and dual-energy X-ray absorptiometry, or DXA, did not differ significantly between the two groups.

The finding is notable because previous meta-analyses have not agreed on how SGLT2 inhibitors affect the skeleton. A 2024 review in BMC Endocrine Disorders reported higher parathyroid hormone and CTX, a marker of bone resorption, but no clear effect on bone mineral density. Another review in Osteoporosis International found no meaningful change in fracture risk, bone density or most bone markers. Together, those mixed results have left clinicians without a clear mechanistic explanation.

The new paper offers one possible answer: SGLT2 inhibitors may disturb bone signalling in a way that is too early or too subtle for routine testing to catch. The authors say the compensatory rise in the Sema3A to NRP1 ratio may reflect an attempt to preserve balance, but they caution that the study was cross-sectional and small, and that it did not separate effects by individual drug or treatment duration. Even so, the work suggests that bone safety questions around this drug class may extend beyond fracture counts and into the biology of bone signalling itself.