Prior treatment, not vitamin D, predicts romosozumab response in 314 osteoporosis patients: A retrospective cohort study

Although calcium and vitamin D supplementation is recommended with romosozumab therapy for postmenopausal osteoporosis, real-world evidence examining the impact of vitamin D formulation, dose, and prior treatment history on the treatment response remains limited. We conducted a retrospective cohort study of 314 osteoporotic patients who received romosozumab at a general hospital in northern Japan between April 2019 and April 2025. Patients were categorized according to vitamin D co-administration (none, 0.5 μg eldecalcitol, 0.75 μg, alfacalcidol 0.5 μg, or alfacalcidol 1.0 μg) and prior treatment history. The primary outcome was the percentage change in the lumbar spine bone mineral density (BMD) at 12 months. Multivariate regression analysis was performed to identify predictors of treatment response. Overall, the lumbar spine BMD increased by 8.90 ± 6.50% and the total hip BMD by 2.90 ± 3.80% at 12 months. Patients receiving vitamin D had lower baseline 25-hydroxyvitamin D levels (15.89 vs. 17.50 ng/mL, p = 0.035), indicating a prescription bias. In the multivariate analysis, prior osteoporosis treatment was the strongest negative predictor of romosozumab response (coefficient − 4.02%, p = 0.0007), whereas vitamin D co-administration did not significantly predict outcomes ( p = 0.15). Among treatment-naïve patients, eldecalcitol 0.5 μg showed a superior lumbar spine BMD increase (10.90%) compared to alfacalcidol 1.0 μg (5.00%, p = 0.045). A significant treatment history × vitamin D interaction was observed ( p = 0.014). Prior treatment history, rather than vitamin D co-administration, was the dominant predictor of romosozumab response in patients with osteoporosis. Eldecalcitol 0.5 μg appears to be optimal for treatment-naïve patients with osteoporosis. Selection bias in vitamin D prescriptions must be considered when interpreting real-world treatment outcomes. • Prior treatment history is the dominant predictor of romosozumab response • Selection bias in vitamin D prescribing was quantified in real-world practice • Eldecalcitol showed superior BMD response in treatment-naïve patients only • Treatment history × vitamin D interaction significantly affects outcomes • STROBE-compliant study of largest single-center romosozumab cohort (n = 314)