Association of Early-Onset Atrial Fibrillation and Estimated Atrial Fibrillation Burden with Mortality in Critically Ill Adults: A Landmark Cohort Study

Background New-onset atrial fibrillation (AF) occurs in 5%–15% of non-cardiac-surgery intensive care unit (ICU) patients and is associated with increased mortality. Most studies classify AF over the entire ICU stay, introducing immortal time bias, and treat AF as binary, ignoring dose-response relationships. Aims To estimate associations of early-onset AF and estimated AF burden with mortality using a 48-h landmark design. Methods We retrospectively analyzed Medical Information Mart for Intensive Care IV adults with at least 48 h of ICU observation whose first informative rhythm record indicated sinus rhythm, excluding cardiac-surgery patients. Early-onset AF and cumulative AF duration were estimated from intermittently charted rhythm records during the first 48 h; follow-up began at the landmark. Cox and logistic regression estimated adjusted associations with 28-day, 1-year, and in-hospital mortality. Results Among 20,925 landmark survivors, 1650 (7.9%) had early-onset AF. After adjustment, early-onset AF was associated with 28-day (HR 1.22, 95% CI 1.10-1.36), 1-year (HR 1.12, 95% CI 1.03-1.22), and in-hospital mortality (OR 1.50, 95% CI 1.31-1.72). Proportional hazards were not met for 28-day mortality: the HR was 1.32 (95% CI 1.17-1.48) during days 0–14 after the landmark and 0.90 (95% CI 0.70-1.16) during days 15–28. Among AF-positive patients, each approximate doubling of estimated AF duration was associated with an 8% higher hazard of 28-day mortality (HR 1.08, 95% CI 1.01-1.16). Conclusions Among ICU patients who remained under observation and survived to a 48-h landmark, early-onset AF was associated with higher subsequent mortality, and longer estimated AF duration was modestly associated with mortality among AF-positive patients. These observational findings do not establish benefit from AF-directed treatment or reducing AF duration.