Sodium–glucose cotransporter 2 (SGLT2) inhibitors are widely used in type 2 diabetes (T2D) and have been the subject of investigation for possible effects on bone metabolism. Prior systematic reviews and meta-analyses have reported inconsistent results regarding fracture risk, bone mineral density (BMD), and biochemical bone turnover markers. The Sema3A/Neuropilin‑1 (NRP1)/Plexin‑A1 signaling axis is a molecular pathway implicated in bone homeostasis, and perturbation of this axis could plausibly affect bone formation or resorption before conventional markers change.
The report presents a propensity score–matched comparison involving postmenopausal women with T2D. The matched sample comprised 23 SGLT2 inhibitor users and 23 non-users. The publication type is a Letter in Journal of Diabetes (2026). The abstract does not detail the matching variables, duration of exposure to SGLT2 inhibitors, the specific agents used, or other protocol elements; those methodological details are not reported in the abstract.
Biomarkers highlighted in the abstract include circulating Neuropilin‑1 (NRP1) levels and the Sema3A/NRP1 ratio. Classical bone turnover markers reported were P1NP (a marker of bone formation) and CTX (a marker of bone resorption), plus a coupling index derived from these markers. Bone mineral density was assessed by DXA. The abstract does not report fracture outcomes or longitudinal changes.
Compared with matched non-users, SGLT2 inhibitor users had significantly lower circulating NRP1 concentrations (74.25 vs. 87.61 ng/mL; p = 0.015). Users also demonstrated a higher Sema3A/NRP1 ratio (0.14 vs. 0.12; p = 0.006). The authors interpret these findings as indicating functional impairment of the Sema3A/NRP1/Plexin‑A1 bone signaling axis accompanied by compensatory upregulation of Sema3A.
Classical biochemical markers of bone turnover—P1NP and CTX—together with a calculated coupling index, showed no significant differences between SGLT2 inhibitor users and non-users in the matched sample. Similarly, DXA‑derived bone mineral density did not differ significantly between groups. These negative findings suggest that alterations in the Sema3A/NRP1/Plexin‑A1 axis may occur prior to detectable changes in standard bone turnover assays or DXA measurements.
The authors propose that SGLT2 inhibitor exposure is associated with suppression of NRP1 and a shifted Sema3A/NRP1 ratio, signaling a disturbance in the Sema3A/NRP1/Plexin‑A1 axis. Because P1NP, CTX, coupling index, and BMD were unchanged, the data suggest molecular-level perturbation that may precede clinically measurable bone loss or turnover changes. This concept offers a putative mechanistic pathway through which SGLT2 inhibitors could influence bone biology without immediate changes in conventional endpoints.
The findings provide a mechanistic perspective that may help explain why prior meta‑analyses have yielded inconsistent conclusions about SGLT2 inhibitors and bone safety. If SGLT2 inhibitors alter the Sema3A/NRP1/Plexin‑A1 axis early, routine clinical measurements (DXA and common serum markers) may not detect these changes in short‑ to medium‑term studies. The report suggests a need to consider molecular biomarkers in future research evaluating bone effects of SGLT2 inhibitors.
The abstract does not report several important methodological and clinical details: specific SGLT2 agents and doses, duration of therapy, matching covariates, sample size justification, full statistical methods, and any longitudinal follow‑up or fracture outcomes. As a Letter, the report is concise and focused on biomarker differences; therefore, causality, clinical fracture risk, and the temporal sequence of molecular changes relative to conventional markers remain unanswered by the abstract. These gaps indicate areas for follow‑up in larger, fully reported studies.
Conflict of interest and publication details
The authors declared no conflicts of interest. Citation details in the abstract: J Diabetes. 2026 Sep;18(9):e70262. DOI: 10.1111/1753-0407.70262. The PubMed record includes PMID 42649130 and PMCID link for the full text. The report is indexed as a Letter and is intended to convey preliminary comparative biomarker data rather than definitive clinical outcomes.