High-Sensitivity Estimate of the Incidence of New-Onset Atrial Fibrillation in Critically Ill Patients

To estimate the incidence of new-onset atrial fibrillation in critically ill patients. DESIGN: Prospective cohort. SETTING: Medical-surgical ICU. SUBJECTS: Consecutive patients without a history of atrial fibrillation but with atrial fibrillation risk factors. INTERVENTIONS: Electrocardiogram patch monitor until discharge from hospital or up to 14 days. MEASUREMENTS AND MAIN RESULTS: A total of 249 participants (median age of 71 yr [interquartile range] 64–78 yr; 35% female) completed the study protocol of which 158 (64%) were admitted to ICU for medical illness, 78 (31%) following noncardiac surgery, and 13 (5%) with trauma. Median Acute Physiology and Chronic Health Evaluation II score was 16 (interquartile range, 12–22). Median duration of patch electrocardiogram monitoring, ICU, and hospital lengths of stay were 6 (interquartile range, 3–12), 4 (interquartile range, 2–8), and 11 days (interquartile range, 5–23 d), respectively. Atrial fibrillation ≥ 30 seconds was detected by the patch in 44 participants (17.7%), and three participants (1.2%) had atrial fibrillation detected clinically after patch removal, resulting in an overall atrial fibrillation incidence of 18.9% (95% CI, 14.2–24.3%). Total duration of atrial fibrillation ranged from 53 seconds to the entire monitoring time. The proportion of participants with ≥1 episode(s) of ≥6 minute, ≥1 hour, ≥12 hour and ≥24 hour duration was 14.8%, 13.2%, 7.0%, and 5.3%, respectively. The clinical team recognized only 70% of atrial fibrillation cases that were detected by the electrocardiogram patch. CONCLUSIONS: Among patients admitted to an ICU, the incidence of new-onset atrial fibrillation is approximately one in five, although approximately one-third of cases are not recognized by the clinical team.