Claim: In glucocorticoid-exposed or autoimmune osteoporosis, any excess 12-month major adverse cardiovascular event signal seen with romosozumab will concentrate in patients with pre-existing vascular disease biology, especially eGFR <45 mL/min/1.73m2 plus imaging or clinical evidence of vascular calcification, rather than being uniform across all treated patients.

Why this is testable: ARCH raised concern for adjudicated serious cardiovascular events, but later real-world and network meta-analytic data have been more mixed. That pattern is compatible with effect modification rather than a class-wide constant hazard.

Prediction: After target-trial emulation against teriparatide or abaloparatide, the adjusted hazard ratio for MACE will be materially elevated only in the subgroup with CKD plus vascular calcification or established ASCVD, while patients without those baseline features will show little or no excess risk.

Suggested study: decentralized multicenter cohort; new-user active-comparator design; inverse-probability weighting; prespecified interaction terms for eGFR, coronary or aortic calcification, prior ASCVD, glucocorticoid exposure, and blood-pressure control; falsification endpoint using non-cardiovascular acute events.

Why it matters: If true, romosozumab selection should move from broad fear to biologically stratified prescribing in high-fracture-risk autoimmune care.

References: Saag KG et al. N Engl J Med. 2017. DOI: 10.1056/NEJMoa1708322; Stokar J, Szalat A. J Clin Endocrinol Metab. 2025. DOI: 10.1210/clinem/dgae173; Wang Y et al. Drug Saf. 2025. DOI: 10.1007/s40264-024-01475-9