Chronic co-prescription of opioids and benzodiazepines is associated with higher frailty prevalence in older US adults: A nationally representative cross-sectional analysis of NHANES 2005-2018.

内容摘要

Opioid-benzodiazepine co-prescription carries a US Food and Drug Administration black-box warning for respiratory depression, yet its association with frailty in community-dwelling older adults has not been quantified at population scale. We analysed 18,434 adults aged &#x2265;50&#xa0;years from seven cycles of the National Health and Nutrition Examination Survey (NHANES 2005-2018). Current opioid and benzodiazepine use was ascertained from prescription-medication interview data, and participants were classified as neither, opioid-only, benzodiazepine-only, or co-prescription under two exposure definitions (any use in the past 30&#xa0;days; chronic use &#x2265;90&#xa0;days). A 36-item questionnaire-based frailty index (FI) was constructed following Rockwood's standard procedure (&#x2265;30 items meets the canonical threshold), and frailty was defined as FI &#x2265;0.25. Survey-weighted logistic regression with three progressively adjusted models estimated odds ratios, complemented by sub-group stratification across seven effect modifiers, multiple imputation by chained equations, cycle-stratified sensitivity analysis, polypharmacy adjustment, triple-therapy gradient modelling (including gabapentinoids), and E-value computation. Chronic co-prescription prevalence was 1.4% (n&#xa0;=&#xa0;259). In the fully adjusted model, chronic co-prescription was associated with an odds ratio for frailty of 2.88 (95% CI 2.47-3.35, P&#xa0;=&#xa0;3.1&#xa0;&#xd7;&#xa0;10-21) relative to neither drug, with a monotonic gradient (benzodiazepine-only 1.60; opioid-only 2.07; co-prescription 2.88; P-trend&#xa0;=&#xa0;5.1&#xa0;&#xd7;&#xa0;10-29). Adding a gabapentinoid (triple therapy, n&#xa0;=&#xa0;67) raised the odds ratio to 3.71 (3.07-4.48). The association remained in all 22 pre-specified sub-groups (all P&#xa0;<&#xa0;0.01), after multiple imputation (OR 2.68), after adjustment for the number of non-study drugs (OR 1.40), and across NHANES cycle halves (OR 2.72 in 2005-2010; 2.96 in 2011-2018). The E-value of 5.20 indicated robustness to moderate unmeasured confounding. Chronic opioid-benzodiazepine co-prescription was consistently associated with a near three-fold higher burden of frailty in older US adults. These findings support prospective evaluation of this high-risk combination in deprescribing and frailty-screening studies.