Risk of atypical femoral fractures among osteoporotic patients on denosumab treatment in real-life setting: a population-based study

Uri Goldwaser; Uri Yoel; Liana Tripto-Shkolink; Merav Fraenkel

doi:10.1530/endoabs.110.P216

P216

Prev Next

Section Contents Cite

Endocrine Abstracts (2025) 110 P216 | DOI: 10.1530/endoabs.110.P216

ECEESPE2025 Poster Presentations Bone and Mineral Metabolism (112 abstracts)

Risk of atypical femoral fractures among osteoporotic patients on denosumab treatment in real-life setting: a population-based study

Uri Goldwaser ¹ , Uri Yoel ² , Liana Tripto-Shkolink ³ & Merav Fraenkel ²

Author affiliations

¹Faculty of Health Scince Ben-Gurion University of the Negev, Goldman School of Medicine, Beer-Sheva, Israel; ²Soroka University Medical Center, Faculty of Health Science Ben-Gurion University of the Negev, Endocrinology, Beer-Sheva, Israel; ³Sheba Medical Center, Mineral and Bone Diseases Service, Division of Endocrinology, Diabetes and Metabolism, Tel-Hshomer, Israel

JOINT3319

Background: Atypical femoral fractures (AFF) are rare but serious adverse events with distinct radiographic criteria but lack specific ICD-9 diagnostic codes. While the association between bisphosphonates (BP) treatment and AFF is well established, denosumab’s (DMAB) risk profile remains unclear, particularly when used after BP.

Aim: To evaluate AFF risk in patients treated with DMAB after prior BP use.

Methods: This population-based study included Clalit HMO patients aged ≥50 years who received osteoporosis treatment between 2002 and 2024. Patients were classified as BPs only (at least one purchase of zoledronic acid or more than one oral BP purchase), or DMAB after BP (at least 2 purchases of DMAB following no more than 3 years from last BP purchase). AFF cases were identified using a validated algorithm combining specific femoral fracture diagnostic and surgical procedure codes, achieving 96. 88% sensitivity, 95. 10% specificity, and 38. 75% positive predictive value when validated against radiographically adjudicated cases meeting ASBMR 2014 criteria. Cases with concurrent pathological fractures, motor vehicle accidents, or peri-prosthetic fractures within 14 days were excluded. We used Cox regression with treatment status (e. g., BP, DMAB as second-line) being a time-varying covariate, to compare AFF risk between the groups. Hazard ratios (HR) were adjusted for age.

Results: 262, 482 patients were included in the analysis: 242, 117 patients in the BP-only group (85% female, mean age 68. 8±10. 0 years), and 16, 629 patients in the DMAB-after-BP group (93% female, mean age 65. 9±8. 8 years). 2, 622 AFF cases occurred during 1, 866, 704 BP patient-years and 98 AFF cases occurred during 76, 743 DMAB patient-years. AFF risk was significantly lower during DMAB treatment after BP compared to BP treatment (HR: 0. 25, 95% CI: 0. 20-0. 31). Patients who developed AFF during DMAB treatment after BP had previously received BP therapy for a mean duration of 10. 8 ± 4. 6 years before switching to DMAB and had a mean DMAB exposure of 2. 9 ± 2. 4 years.

Conclusion: In our analysis, DMAB treatment following BP was associated with a lower risk of AFF compared to BP therapy. This finding is particularly relevant given the common clinical practice of transitioning patients from BP to DMAB therapy. A prospective study is needed to further support the safety of DMAB therapy after BP.