Benefits and harms of oral anticoagulants for atrial fibrillation in nursing home residents with advanced dementia

Abstract

Background: Approximately 20% of older persons with dementia have atrial fibrillation (AF). Nearly all have stroke risks that exceed the guideline-recommended threshold for anticoagulation. Although individuals with dementia develop profound impairments and die from the disease, little evidence exists to guide anticoagulant discontinuation, and almost one-third of nursing home residents with advanced dementia and AF remain anticoagulated in the last 6 months of life. We aimed to quantify the benefits and harms of anticoagulation in this population.

Methods: Using Minimum Data Set and Medicare claims, we conducted a retrospective cohort study with 14,877 long-stay nursing home residents aged ≥66 between 2013 and 2018 who had advanced dementia and AF. We excluded individuals with venous thromboembolism and valvular heart disease. We measured anticoagulant exposure quarterly, using Medicare Part D claims. The primary outcome was all-cause mortality; secondary outcomes were ischemic stroke and serious bleeding. We performed survival analyses with multivariable adjustment and inverse probability of treatment (IPT) weighting.

Results: In the study sample, 72.0% were female, 82.7% were aged ≥80 years, and 13.5% were nonwhite. Mean CHA₂ DS₂ VASC score was 6.19 ± 1.58. In multivariable survival analysis, anticoagulation was associated with decreased risk of death (HR 0.71, 95% CI 0.67-0.75) and increased bleeding risk (HR 1.15, 95% CI 1.02-1.29); the association with stroke risk was not significant (HR 1.08, 95% CI 0.80-1.46). Results were similar in models with IPT weighting. While >50% of patients in both groups died within a year, median weighted survival was 76 days longer for anticoagulated individuals.

Conclusion: Persons with advanced dementia and AF derive clinically modest life prolongation from anticoagulation, at the cost of elevated risk of bleeding. The relevance of this benefit is unclear in a group with high dementia-related mortality and for whom the primary goal is often comfort.

Keywords: advanced dementia; anticoagulation; atrial fibrillation.

Figure 1:

Crude outcome rates by baseline anticoagulant (table at left); adjusted hazard ratios for the associations between anticoagulant use and outcomes in multivariable adjusted models and IPT weighted models (forest plot). Abbreviations: MV = Multivariable; IPT = Inverse Probability of Treatment; AC = Anticoagulant. Covariates used in multivariable adjustment and in development of IPT weights included demographics, comorbid conditions, and nursing home assessment data, as detailed in Table 1.