(080) Opioid Deprescribing Rates in Older Adults with Frailty and Risk Factors for Prolonged Use after an Acute Fracture

Background: Opioids are commonly used to control pain after an acute fracture in older adults with frailty or dementia. Little is known about the duration and deprescribing rates of opioids in this high-risk population.

Objectives: Our objective was to describe the trends in opioid deprescribing following an acute fracture within a population of older adults. We compared 1-year opioid deprescribing rates based on frailty and dementia status.

Methods: We conducted a retrospective cohort study using a U.S. national sample of fee-for-service Medicare beneficiaries from January 2013 to December 2018. The study cohort included individuals aged 65 years or older with an opioid dispensing (index date = opioid dispensing date) within 30 days after experiencing an acute inpatient fracture. Exclusions encompassed individuals with prior opioid use, chronic pain, or contraindications to opioid use in the 6 months before the index date. We employed Nelson-Aalen cumulative hazard functions to estimate opioid discontinuation rates, defined as a gap in prescription refills >15 days. We assessed the association between 22 covariates measured in the 180 days before the index date and opioid discontinuation in Cox proportional hazards regression. We used Fine and Grey’s method to address competing risks, and inverse probability of censoring weights (IPCW) to handle informative censoring.

Results: In our sample (N=21,728, mean age [SD]=80 [9] years, female=77%, White=91%), 89% discontinued opioids and 2% died within 1-year. The unadjusted 30-day discontinuation rate was 85% (95% CI, 83-86%). The IPCW-adjusted discontinuation rate at 30 days was 85% (95% CI, 83-86%) in those with dementia versus 87% (95% CI, 86-87%) for those without. The 30-day discontinuation rate was 85% (95% CI, 83-87%) in the moderate/severe frail group and 90% (95% CI, 89-91%) in the non-frail group. The corresponding Fine and Grey estimates were very similar. After multivariable adjustment, significant predictors of opioid discontinuation included older age (hazard ratio [HR]: 1.05, 95% CI, 1.03-1.08), recent years of cohort entry (HR comparing 2018 to 2013: 1.25, 95% CI, 1.18-1.32), and less frailty (HR comparing non-frail to moderate/severe: 1.16, 95% CI, 1.07-1.26). Dementia was not associated with opioid discontinuation rates.

Conclusions: Our findings suggest that older age and greater frailty—but not dementia status—were associated with longer durations of opioid use.