Pedicle screw fixation in osteoporotic bone remains a significant challenge in instrumented spine surgery, yet most pharmacological data in the osteoporosis literature does not focus on specific biomechanical demands. Keaveny et al address this directly by retrospectively leveraging imaging substudies from two established romosozumab RCTs to investigate whether established gains in BMD translate to meaningful improvements in pedicle screw pullout resistance. Romosozumab has been shown to provide superior efficacy in improving BMD in landmark prospective trials1–3, however, its impact on surgical outcomes in spine literature remains limited, making this a timely contribution to the field.

This study identified patients from the aforementioned randomized controlled trials, who have subsequently been included in imaging substudies. A total of 79 women were included from a phase 2 trial, in which women with low BMD were randomized to receive romosozumab, teriparatide, or placebo. Patients had baseline and postbaseline CT imaging at 12 months. Another 79 women were included from a phase 3 trial, in which postmenopausal women with osteoporosis and fragility fractures were randomized to receive romosozumab versus alendronate for 12 months, before switching to alendronate-only for another 12 months. Patients had baseline and postbaseline CT imaging at 6, 12, and 24 months.

From these CT images, the L1 vertebra from each patient at each timepoint were reconstructed virtually in software, with local material properties derived from each voxel’s Hounsfield Unit to reflect local bone density, an approach that removes the variability of cadaveric testing but introduces assumptions that become more consequential when extrapolating to a postimplantation in vivo environment. Standardized pedicle screws (scaled to patients’ pedicle) were virtually implanted bilaterally. The virtual models were then loaded to failure in an axial pullout configuration, measuring the shear bone strength (in newtons) and volume of failed tissue (in cm3) when loaded to a predetermined force threshold. Timepoint measurements were compared to baseline to capture treatment-induced longitudinal changes. Secondary endpoints measured percentage change in BMD in the screws’ surrounding tissue.

Romosozumab consistently outperformed all comparators across both trials and all three endpoints. In the phase 2 patients, romosozumab improved screw shear strength by 24.7% vs teriparatide 14.8% and placebo -2.2% at 12 months. Similarly, in the phase 3 patients, romosozumab outperformed alendronate (26.3% vs 7.3%) at 12 months, with superior effects sustaining (25.2% vs 5.7%) 12 months after switching to alendronate therapy only. The persistence of these gains into the alendronate maintenance period is of particular clinical interest, suggesting a preoperative course of romosozumab may confer lasting benefit even after treatment is discontinued. The authors also employed a flexural loading model designed to simulate physiological spinal loading, demonstrating directionally consistent results. Additionally, romosozumab reduced the volume of failed tissue by 44% vs teriparatide 29.3% at 12 months, and 19.9% vs alendronate 4.5% at 12 months. Periscrew BMD showed concordant improvements, with romosozumab outperforming comparators at all timepoints across both trials.

The study presents promising results demonstrating the potential impact of romosozumab on spine surgical outcomes, however several limitations should be taken into consideration when interpreting these findings. The study’s population of women with low bone mineral density or osteoporosis is not representative of the degenerative and deformity morphology seen in a typical spine surgical population that may affect biomechanics. Furthermore, the model inadequately captures temporal bone-drug and bone-implant interaction. While the model somewhat captures the effects of preoperative optimization to initial screw fixation, it does not simulate temporal remodeling interactions after implantation, nor does it account for stress shielding introduced by an interbody implant. Moreover, although the single-level axial pullout model and flexural loading model help demonstrate physiological patterns of stress, they do not completely replicate in vivo patterns of mechanical loading, where junctional stress and cyclic fatigue play a central mechanistic role. Dedicated clinical studies will be required to verify the consistency of these findings across the domains identified above.

In summary, this study provides a timely contribution using a creative and novel methodology, demonstrating consistent biomechanical superiority of romosozumab across multiple endpoints and timepoints in both trials. These findings carry a clear clinical hypothesis: preoperative romosozumab optimization may improve initial pedicle screw fixation in osteoporotic patients undergoing instrumented fusion, a consideration of increasing relevance in today’s increasingly elderly surgical population. While promising, the non-surgical population and modeling limitations mean these findings should be interpreted cautiously, and dedicated prospective clinical studies in surgical populations will be required to confirm whether they translate to improved patient outcomes and can be incorporated into perioperative optimization pathways.