6A-01 - Follow-on therapy after denosumab discontinuation among United States Medicare beneficiaries

Background: Denosumab (DMAB) is an injectable osteoporosis (OP) therapy prescribed twice annually to older adults. Discontinuation of DMAB may result in rebound fractures and rapid bone loss. Guidelines recommend starting another OP medication six months after the last DMAB dose, but implementation of these guidelines in routine clinical practice is not well-described.

Objectives: To describe the incidence and predictors of follow-on OP medication use for US older adults who discontinued DMAB.

Methods: We used Medicare data for a 20% random sample of beneficiaries. The first claim for DMAB between 1/2010-12/2019 among adults age ≥66 years was identified using Part B and D claims. DMAB discontinuation was defined as a 60-day gap in supply. During follow-up (through 12/2020), we identified those who initiated a follow-on OP therapy (i.e., oral and intravenous bisphosphonates, teriparatide, and abaloparatide) within 1 year after DMAB discontinuation. Patients were censored at the earliest of follow-on medication claim, DMAB reinitiation (restart after therapy gap), death, long-term care entry, Medicare disenrollment, or 1 year (no therapy). A multinomial logistic regression model with marginal standardization to total population covariate distributions was used to estimate risk ratios (RR) with delta method 95% confidence limits (CL) for associations between covariates and follow-on medication use after DMAB discontinuation. The outcome variable had four levels: no follow-on medication within one year (reference category), follow-on medication initiation, DMAB reinitiation, and loss to follow-up.

Results: Of 97,595 DMAB initiators, we identified 65,818 individuals (mean age=79.2 [SD=7.2] years; 90.7% female) who discontinued DMAB. Only 4,010 (6.1%) individuals initiated a follow-on medication after discontinuation (mean time to therapy=96.1 [SD=101.3] days). Alendronate (n=2,293; 3.5%), ibandronate (n=536; 0.8%), and zoledronic acid (n=503; 0.8%) were most common. Younger age, female sex, prior OP therapy, and having 1-2 DMAB claims (vs. more) before discontinuation were strong predictors of initiating follow-on medication (vs. not). There was no evidence of association between prior fractures and starting follow-on therapy: RR=0.96 (95% CL: 0.87-1.04). Most people did not initiate therapy (44.6%), reinitiated DMAB after a gap (40.7%), or were lost to follow-up (8.6%).

Conclusions: Our study suggests that many patients stopping DMAB may be at excess risk of fractures because they do not initiate follow-on therapies. Future work should compare the benefits of initiating different follow-on therapies after DMAB discontinuation.