Withholding Oral Anticoagulation at Discharge and Post‐Discharge Stroke and Bleeding Vulnerability in a Predominantly Older Cohort Hospitalized With Acute Heart Failure and Atrial Fibrillation

Vollständiger Abstract

Worum geht es in dieser Arbeit?

ABSTRACT Aim In older adults hospitalized with acute heart failure and atrial fibrillation, bleeding concern often drives withholding of oral anticoagulation at discharge. Whether withholding marks a genuinely lower bleeding‐risk phenotype or residual clinical vulnerability is uncertain. We examined its association with post‐discharge ischemic stroke, major bleeding, and death in a predominantly older cohort. Methods We analyzed a registry of patients hospitalized with acute heart failure and atrial fibrillation. In a discharge‐based analysis of patients discharged alive and free of stroke or major bleeding at discharge, we used stabilized inverse‐probability‐of‐treatment weighting for discharge anticoagulation. Ischemic stroke and major bleeding were modeled with weighted Fine–Gray competing‐risk regression (death as competing event) and death with weighted Cox regression, contrasting no anticoagulant versus anticoagulant. Age‐restricted analyses (≥ 75, ≥ 80 years) and absolute risks were prespecified. Results Among 560 patients (no anticoagulant 72, anticoagulant 488; mean age 78 years), withholding was associated with higher ischemic stroke (subdistribution hazard ratio 3.14, 95% confidence interval [CI] 1.26–7.81) and did not identify a lower bleeding‐risk phenotype, with numerically higher major bleeding overall (1.98, 0.95–4.14) that reached significance among patients aged ≥ 80 years (2.41, 1.05–5.54). No significant association with death was observed (hazard ratio 1.34, 0.70–2.55). Three‐year cumulative incidences of bleeding were 25.6% (no anticoagulant) versus 13.4% (anticoagulant). Conclusion Discharge anticoagulation withholding did not identify a lower post‐discharge bleeding‐risk phenotype but marked residual thromboembolic vulnerability, with a higher bleeding‐risk trajectory, particularly among patients aged ≥ 80 years, supporting structured post‐discharge anticoagulation reassessment in older adults.