The study objective was to determine the predictors of postoperative atrial fibrillation (POAF) in patients randomized to conventional coronary artery bypass graft (on-pump coronary artery bypass [ONCAB]) versus beating heart coronary surgery (off-pump coronary artery bypass [OPCAB]). The subgroup of 2103 patients (of 2203 enrollees) in the Randomized On Versus Off Bypass trial with no POAF was studied (1056 patients in the ONCAB group and 1047 patients in the OPCAB group). Univariate and multivariate analyses were used to identify the predictors of POAF and the impact of POAF on outcomes. Use of ONCAB versus OPCAB was not associated with increased rates of POAF. Older age (P < .0001), white race (P < .001), and hypertension (P < .002) were predictors of POAF on multivariate analysis. In general, POAF led to a higher rates of reintubation (ONCAB: 6.3% vs 0.8% no POAF, P < .001; OPCAB: 7.4% vs 1.8% no POAF, P < .0001) and prolonged ventilatory support (ONCAB: 7.1% vs 2.3% no POAF, P = .001; OPCAB: 9.2% vs 3.4% no POAF, P = .0003). The rate of any early adverse outcome was higher in patients with POAF (all patients: 10% POAF vs 4.7% no POAF, P < .0001; ONCAB: 9% POAF vs 4.3% no POAF, P = .008; OPCAB: 11% POAF vs 5.1% no POAF, P = .001). The 1-year all cause mortality was higher with POAF for both groups (ONCAB: 5.4% POAF vs 2% no POAF, P = .009; OPCAB: 5.1% POAF vs 2.6% no POAF, P = .07). POAF was independently associated with early composite end point (odds ratio [OR], 2.23; confidence interval [CI], 1.55-3.22; P < .0001), need for new mechanical support (OR, 3.25; CI, 1.39-7.61; P = .007), prolonged ventilatory support (OR, 2.93; CI, 1.89-4.55; P < .0001), renal failure (OR, 5.42; CI, 1.94-15.15; P = .001), and mortality at 12 months (OR, 1.94; CI, 1.14-3.28; P = .01). In the Randomized On Versus Off Bypass trial, the strategy of revascularization did not affect the rate of POAF. Age, race, and hypertension were predictors of POAF. POAF was independently associated with a higher short-term morbidity and higher 1-year mortality rates.