Geriatric Pharmacoepidemiology

2D-12 - Benzodiazepines continuation and de-intensification versus discontinuation and short-term all-cause mortality among older US adults following hospitalization for hip fracture

Sunday, August 24, 2025

4:45 PM - 4:50 PM ET

Presenting Author(s)

Meng Pan, MS

PhD Student
Department of Epidemiology, Gillings School of Global Public Health, University of North Carolina at Chapel Hill, United States

Co-Author(s)

Emilie D. Duchesneau, PhD

Assistant Professor
Department of Epidemiology and Prevention, Division of Public Health Sciences, Wake Forest University School of Medicine
Winston-Salem, North Carolina, United States
DK

Dae Hyun Kim, MD, ScD

Harvard Medical School, Massachusetts, United States
CP

Charles Poole, ScD

Associate Professor
Department of Epidemiology, Gillings School of Global Public Health, University of North Carolina at Chapel Hill
Chapel Hill, North Carolina, United States
KF

Keturah Faurot

Department of Physical Medicine and Rehabilitation, School of Medicine, University of North Carolina at Chapel Hill, United States
RG

Robert Glynn

Brigham and Women's Hospital, Massachusetts, United States
Jennifer L. Lund, PhD, MSPH

Professor
Department of Epidemiology, Gillings School of Global Public Health, University of North Carolina at Chapel Hill
Chapel Hill, North Carolina, United States
Michele Jonsson Funk, PhD

Associate Professor
Department of Epidemiology, Gillings School of Global Public Health, University of North Carolina at Chapel Hill, United States
EM

Elyse Miller

Department of Epidemiology, Gillings School of Global Public Health, University of North Carolina at Chapel Hill, United States
AE

Alan Ellis

Associate Professor
School of Social Work, North Carolina State University, United States
TS

Til Stürmer, MD, PhD

Professor
Department of Epidemiology, Gillings School of Global Public Health, University of North Carolina at Chapel Hill, North Carolina, United States

Background: The American Geriatrics Society Beers criteria recommend avoiding benzodiazepines (BZD) among older adults with history of falls or fractures. However, real-world evidence on the effect of stopping BZDs after hip fracture hospitalization on mortality is inconsistent.

Objectives: To estimate risk of one-year all-cause mortality for patients with different BZD refill patterns compared to discontinuation (as recommended by guidelines) following hospitalization for hip fracture, overall and stratified by frailty.

Methods: We assembled a cohort of Medicare beneficiaries (≥66 years) hospitalized for hip fracture (Jan 1, 2008-June 30, 2018), discharged home or to a skilled nursing facility (SNF), and survived ≥180 days post-discharge. We required patients to have ≥365 days of continuous Medicare Part A, B, and D coverage, a BZD dispensed ≤90 days pre-admission, and a hospital stay ≤10 days. BZD refill patterns were assessed by comparing the last pre-admission dispensing to the first fill ≥90 days post-discharge (with a 90-day grace period). Patients were categorized as continuers (same/increased dose), de-intensifiers (dose reduction/switch), or discontinuers (no BZD). We used inverse probability of treatment weighting (IPTW) from a generalized propensity score model (discontinuers as the common reference) to balance index-date (180 days post-discharge) covariates (demographics, comorbidities, dementia, psychiatric conditions, low-income subsidy, and hospital and SNF stay). We estimated one-year all-cause mortality risk differences (RDs, %) and 95% confidence intervals (CI) for continuers and de-intensifiers, respectively, vs. discontinuers, overall and stratified by the Faurot frailty index (FFI; low frailty: 0≤FFI < 0.1; medium: 0.1≤FFI < 0.4; high: FFI≥0.4).

Results: Among 8,395 eligible patients (52% continuers and 13% de-intensifiers; mean age 81±8 years), 84% were female, 95% were non-Hispanic White, 40% were eligible for the Medicare low-income subsidy, 44% had high frailty, and 40% had medium frailty. Crude one-year all-cause mortality risk was 16% for continuers, 21% for de-intensifiers, and 21% for discontinuers. The IPTW RD [95% CI] for continuers vs discontinuers was -2.7 [-3.9, -1.5] (overall), -1.9 [-3.7, -0.1] (low frailty), -0.6 [-2.2, 1.2] (medium), and -4.6 [-6.6, -2.6] (high). The IPTW RD [95% CI] for de-intensifiers vs discontinuers was 0.1 [-1.1, 1.4] (overall), 2.8 [0.7, 4.9] (low frailty), 2.9 [1.2, 4.7] (medium), and -2.4 [-4.5, -0.4] (high).

Conclusions: Older US adults who continued taking BZDs following hip fracture hospitalization and survived ≥180 days post-discharge had the lowest one-year mortality risk overall. Caution should be taken when discontinuing BZDs in high-frailty patients.