Background/Purpose: We previously demonstrated that denosumab increased lumbar spine and total hip bone mineral density (BMD) significantly more than risedronate at 12 and 24 months in glucocorticoid (GC)-treated individuals (Saag ACR 2016; Saag ECTS 2018). Prespecified subgroup analyses of lumbar spine BMD at 12 months indicated that denosumab was superior to risedronate across 7 subgroups of GC-treated individuals (Saag ASBMR 2017). This analysis explored the effects of denosumab and risedronate on lumbar spine BMD in the same subgroups of GC-treated individuals at 24 months.

Methods: The phase 3, randomized, double-blind, double-dummy, active-controlled study enrolled women and men age ≥18 years receiving ≥7.5 mg prednisone or equivalent daily for <3 months (GC-initiating) or ≥3 months (GC-continuing) before screening. All subjects age <50 years were required to have a history of osteoporotic fracture. GC–continuing subjects age ≥50 years were required to have lumbar spine, total hip, or femoral neck BMD T‑score ≤‒2.0; or ≤‒1.0 with a history of osteoporotic fracture. Subjects were randomized 1:1 to denosumab 60 mg SC every 6 months or risedronate 5 mg PO daily for 24 months. All subjects were to receive daily calcium (≥1000 mg) and vitamin D (≥800 IU). The treatment difference (denosumab – risedronate) for percentage change from baseline in lumbar spine BMD at 24 months was estimated in the GC-initiating and GC-continuing subpopulations, both overall and in 7 prespecified subgroups where treatment effect might differ (sex, race, age, baseline BMD T‑score, geographic region, menopausal status, and baseline GC daily dose).

Results: The study enrolled 795 subjects (290 GC-initiating, 505 GC-continuing). Baseline characteristics were balanced between treatment groups within each subpopulation. Denosumab was superior to risedronate for gains in lumbar spine BMD at 24 months in both the GC-initiating and GC-continuing subpopulations. Within each subgroup (Table), denosumab was consistently associated with a greater increase in lumbar spine BMD at 24 months compared with risedronate. Significant quantitative interactions were observed only in the sex and race subgroups in the GC-initiating subpopulation. However, qualitative tests indicated the direction of the denosumab effect did not differ significantly by sex or race in this subpopulation.

Conclusion: Denosumab consistently increased lumbar spine BMD more than risedronate at 24 months in both GC-initiating and GC-continuing subpopulations, with no evidence of directional heterogeneity in treatment effect across 7 prespecified subgroups of GC-treated individuals. Denosumab may be a useful addition to the osteoporosis armamentarium in the common clinical setting of GC use.