Does BMA improve outcomes in mHSPC with bone metastases?: Landmark propensity-matched analysis and nested zoledronic acid vs denosumab.

Abstract

e17100 Background: Bone Marrow Agent (BMA) reduces skeletal-related events in bone-metastatic mCRPC, but their role in metastatic hormone-sensitive prostate cancer (mHSPC) disease remains uncertain; a LATITUDE post hoc analysis suggested longer time to skeletal-related events (SRE) with BMA use in patients with high-risk mHSPC but inconsistent OS. Methods: Retrospective cohort study using TriNetX database of mHSPC patients with bone metastases initiating first-line ADT (index). Primary analysis: early BMA (denosumab/zoledronic acid, 0–90d) vs no BMA; 90d landmark to limit immortal-time bias (exposure 0–90d; follow-up from day 90 in those alive/at risk). 1:1 PSM balanced demographics, comorbidities, tumor severity (PSA, Gleason, T/N stage), and prior therapies. Outcomes (code-defined) were OS, SREs (fracture, spinal cord compression, palliative RT), hypocalcemia, and acute kidney failure; Kaplan–Meier/Cox estimated HRs (95% CI), and risks/episode counts were summarized. Nested within early-BMA users, zoledronic acid vs denosumab was compared via 1:1 PSM using the same landmark. Results: Primary analysis: After 1:1 PSM, 505 matched pairs were analyzed. Deaths occurred in 163 (32.9%) in the BMA group vs 129 (26.0%) in the non-BMA group (p = 0.018); however, Kaplan–Meier time-to-event analysis was not significant (HR 1.06 [0.84–1.34]; p = 0.609). Pathologic fracture occurred in 23 (4.6%) vs 22 (4.4%) (p = 0.879; HR 0.91 [0.51–1.64]); spinal cord compression in 22 (4.4%) vs 15 (3.0%) (p = 0.241; HR 1.31 [0.68–2.52]); and palliative radiotherapy in 43 (8.5%) vs 46 (9.1%) (p = 0.739; HR 0.82 [0.54–1.25]). Acute kidney failure occurred in 97 (19.2%) in both groups (p = 1.000). Mean SRE episode counts were similar between groups (all p > 0.30). Nested analysis (ZOL vs DENO): After 1:1 PSM, 169 matched pairs were analyzed; only OS and AKI were reportable (other endpoints < 10 per group). Post-PSM deaths were 57 (34.3%) vs 56 (33.7%) (p = 0.908; HR 1.01 [0.70–1.46]). AKI occurred in 31 (18.3%) vs 35 (20.7%) (p = 0.583; HR 0.91 [0.56–1.48]); mean AKI episodes differed (2.45±2.10 vs 3.91±3.31, p = 0.039). Because fracture/SCC/RT/hypocalcemia were non-reportable post-PSM due to low case counts, crude estimates showed SCC 17/318 (5.35%) vs ≤10/287 (3.48%) (p = 0.268; HR 1.46 [0.67–3.19]), lower palliative radiotherapy (7.6% vs 12.5%; p = 0.040; HR 0.57 [0.34–0.96]), and lower hypocalcemia (3.5% vs 13.6%; p < 0.0001; HR 0.24 [0.12–0.46]) with zoledronic acid. P-value in negative controls > 0.1 both. Conclusions: In a landmark propensity-matched analysis, initiating BMAs early after ADT in mHSPC with bone metastases does not improve time-to-event survival or reduce skeletal-related events. In a nested comparative-effectiveness analysis, zoledronic acid and denosumab demonstrated similar survival and AKI risk, informing agent selection in routine practice.