Soluble urokinase plasminogen activator receptor (suPAR) is a biomarker of chronic low-grade inflammation and a potent predictor of cardiovascular events. We hypothesized that plasma suPAR levels would predict new-onset atrial fibrillation (AF) in a large cohort of con-secutively admitted acute medical patients during long-term follow-up. In 14,764 acutely ad-mitted patients without prior or current AF, median suPAR measured upon admission was 2.7 ng/ml (interquartile range (IQR) 1.9-4.0). During a median follow-up of 392 days (IQR 218-577), 349 patients (2.4%) were diagnosed with incident AF. suPAR levels at admission significantly predicted subsequent incident AF (HR per doubling of suPAR: 1.21, 95% CI 1.05-1.41, adjusted for age and sex). After further adjustment for Charlson score, plasma C-reactive protein (CRP), plasma creatinine and blood hemoglobin-levels, the result remained essentially unaltered (HR per doubling of suPAR: 1.20, 95% CI: 1.01-1.42). In multivariate ROC curve analysis, combining age, sex, Charlson score, CRP, creatinine, and hemoglobin (AUC 0.77, 95% CI 0.75-0.79), the addition of suPAR did not improve the prediction of incident AF (AUC 0.77, 95% CI 0.75-0.79, P=0.89). Plasma suPAR is independently associated with subsequent new-onset AF in a population of recently hospitalized patients, but the addition of suPAR to baseline risk markers appears not to improve the prediction of AF.