Background Postoperative atrial fibrillation (POAF) is a common complication after cardiac surgery. Data are lacking on the long-term prognostic implications of POAF. We hypothesized that POAF, which reflects underlying cardiovascular pathophysiologic substrate, is a predictive marker of late AF and long-term mortality. Methods We identified 603 Olmsted County, Minnesota, residents without prior documented history of AF who underwent coronary artery bypass graft and/or valve surgery from 2000 to 2005. Patients were monitored for first documentation of late AF or death at >30 days postoperatively. Multivariate Cox regression models were used to assess the independent association of POAF with late AF and long-term mortality. Results After a mean follow-up of 8.3 ± 4.2 years, freedom from late AF was less with POAF than no POAF (57.4% vs 88.9%, P < .001). The risk of late AF was highest within the first year at 18%. Univariate analysis demonstrated that POAF was associated with significantly increased risk of late AF hazard ratio (HR), 5.09; 95% CI, 3.65-7.22 and long-term mortality (HR, 1.79; 95% CI, 1.38-2.22). After adjustment for age, sex, and clinical and surgical risk factors, POAF remained independently associated with development of late AF (HR, 3.52; 95% CI, 2.42-5.13) but not long-term mortality (HR, 1.16; 95% CI, 0.87-1.55). Conversely, late AF was independently predictive of long-term mortality (HR, 3.25; 95% CI, 2.42-4.35). Diastolic dysfunction independently influenced the risk of late AF and long-term mortality. Conclusions Postoperative atrial fibrillation was an independent predictive marker of late AF, whereas late AF, but not POAF, was independently associated with long-term mortality. Patients who develop new-onset POAF should be considered for continuous anticoagulation at least during the first year following cardiac surgery.