Newly Diagnosed Atrial Fibrillation After Hospitalization: An “Irregularly Irregular” Stroke Risk to be Considered by Hospitalists

The electrocardiographic pattern of atrial fibrillation (AF) is introduced early during medical school until it is easily recognized. This “irregularly irregular” rhythm is the most frequent arrhythmia, increases in incidence with age, and implies a higher risk for adverse cardiovascular and neurologic outcomes (1). Despite growing evidence, clinicians struggle with knowledge gaps, including how to manage the cardioembolic risk for hospitalization-related, newly diagnosed AF. 

In patients with chronic, long-standing AF at high cardioembolic risk, trials have shown that anticoagulation reduces stroke by 62% (2). Consequently, guidelines strongly recommend anticoagulation in patients with an annual stroke risk above 2% and suggest considering it in those with 1% to 2% risk (3). Nevertheless, these results may not be extrapolated to patients with AF detected during hospitalization due to acute illness, which applies to 5% to 46% of critically ill patients (4). This clinical presentation of AF is poorly understood and lacks a standardized terminology (“provoked,” “secondary,” “acute,” or “induced” AF). 

Although it is the same disease, it is reasonable to approach patients with AF differently depending on whether the presentation is provoked or unprovoked. During hospitalization, factors such as inflammation, stress, pain, electrolyte imbalances, surgery, and medications can predispose to or trigger AF. Additionally, increased monitoring during hospitalization contributes to higher detection rates. After discharge, many of these variables resolve, making AF recurrence—and stroke risk—unpredictable. 

For a hospitalist orchestrating care from admission to discharge, shared decision making with patients is useful in this scenario where an unclear benefit of stroke prevention is contrasted with bleeding risk. For this purpose, knowing the natural history of the disease is crucial to communicate alternatives to patients. Recently, a paper published in Annals of Internal Medicine (5) addressed these questions and provided helpful estimates for better understanding of newly diagnosed AF during hospitalization. 

Researchers from the University of Toronto used administrative databases to conduct a retrospective cohort study (2021–2023), following over 20,000 adults without prior AF who were discharged from hospitalizations for causes other than AF. Patients developing new AF during hospitalization were classified by reason for admission (cardiac medical, noncardiac medical, cardiac surgical, or noncardiac surgical) and followed for 1 year to assess the incidence of stroke, anticoagulation initiation, bleeding, and death. 

Strikingly, 1-year mortality ranged from 3.5% to 27%, being highest in noncardiac medical patients. As a benchmark, authors reported a 15.5% mortality rate in patients hospitalized for AF. This underscores the complexity, heterogeneity, and poor prognosis of provoked AF and the need to consider competing risks. 

Although mean CHA₂DS₂-VA scores were high (3.72 to 4.89), only 1 in 4 patients was prescribed anticoagulation after discharge, highlighting the clinical equipoise to be resolved. Nevertheless, anticoagulation initiation increased with higher scores and cardiac medical admissions, suggesting that clinicians do consider the CHA₂DS₂-VA (or its variables) in this population. Overall, 1-year stroke risk was under 2%, below conventional thresholds for anticoagulation, but higher in those with CHA₂DS₂-VA scores of 5 to 8 or cardiac medical admissions, where the upper bound of CIs barely crossed the actionable threshold (1.5% [95% CI, 1.0% to 2.2%]). When censoring for anticoagulation initiation, the overall result remained similar, although differences between groups were no longer observed. For reference, stroke incidence among patients hospitalized primarily for AF was 1.4% with anticoagulation and 2.3% without. 

These results confirm that the natural history of AF differs according to the setting in which it is diagnosed, so indirect evidence does lower certainty of evidence. Even though treatment recommendations cannot be derived from this study, the presented rates are informative for shared decision making and future research. Stroke risk was generally below the actionable threshold, but modest intergroup differences could help identify patients who warrant closer follow-up and possible treatment if AF recurs. Moreover, these findings help prioritize patients for future trials: A higher event rate would allow smaller sample sizes and shorter follow-up to detect anticoagulation benefits. For instance, given the 1% annual stroke incidence in untreated patients with provoked AF, 11,000 patients would be needed to detect a 50% relative risk reduction with 80% power—the same effect size used in the NOAH-AFNET 6 trial (edoxaban in subclinical AF) (6). 

While many questions remain, and several more arise, the paper from Annals is informative and raises anticipation for what lies ahead. As we await trial-based evidence, observational data can guide nuanced decision making. Hospitalists must face uncertainty with expertise as the approach to AF transitions from a categorical problem (having or not having it) to a continuous phenomenon. After all, AF and stroke risk exist along a spectrum that runs irregularly irregular, beat after beat. 

Reference

1. Al-Khatib SM. Atrial fibrillation. Ann Intern Med. 2023;176:ITC97-ITC112. [PMID: 37429028] doi:10.7326/AITC202307180 
2. Hart RG, Benavente O, McBride R, et al. Antithrombotic therapy to prevent stroke in patients with atrial fibrillation. A meta-analysis. Ann Intern Med. 1999;131:492-501. [PMID: 10507957] 
3. Joglar JA, Chung MK, Armbruster AL, et al; Writing Committee Members. 2023 ACC/AHA/ACCP/HRS guideline for the diagnosis and management of atrial fibrillation: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. J Am Coll Cardiol. 2024;83:109-279. [PMID: 38043043] doi:10.1016/j.jacc.2023.08.017
4. Sibley S, Muscedere J. New-onset atrial fibrillation in critically ill patients. Can Respir J. 2015;22:179-82. [PMID: 26057373] 
5. Abdel-Qadir H, Gunn M, Fang J, et al. Risk for stroke after newly diagnosed atrial fibrillation during hospitalization for other primary diagnoses. A retrospective cohort study. Ann Intern Med. 2025;178:765-774. [PMID: 40258280] doi:10.7326/ANNALS-24-01967 
6. Kirchhof P, Toennis T, Goette A, et al; NOAH-AFNET 6 Investigators. Anticoagulation with edoxaban in patients with atrial high-rate episodes. N Engl J Med. 2023;389:1167-1179. [PMID: 37622677] doi:10.1056/NEJMoa2303062

José Cruz, MD

José Cruz, MD, is an internal medicine faculty physician at Fundación Cardioinfantil in Bogotá, Colombia, and an editorial fellow at Annals of Internal Medicine. He currently works as a hospitalist in general inpatient care and also serves as an intensive care physician and clinical researcher. His professional interests include hospital-based care, care transitions, evidence-based medicine implementation, and quality measurement in health care delivery.