Previous studies proposed the predictive value of baseline serum uric acid (SUA) in the prognosis of coronary artery bypass grafting (CABG) patients. The association of perioperative SUA variation with in-hospital adverse outcomes in CABG patients is unknown. A total of 2,453 patients were included in the study and were divided into four groups (G1-G4) according to perioperative SUA variation (ΔSUA) (G1, ΔSUA ≤ -90 μmol/L; G2, -90 μmol/L < ΔSUA < 0; G3, 0 ≤ ΔSUA < 30 μmol/L; G4, 30 μmol/L ≤ ΔSUA.) The basic characteristics and incidence of adverse outcomes were compared between the groups in the overall population and the subgroups. Multivariate logistic regression was performed to explore the association between perioperative SUA increases and adverse outcomes, and receiver operating characteristic analysis was used to obtain the cutoff value of SUA increases. The patients had a mean age of 60.9 years and the majority were males (76.7%). In the group with the most significant increase in SUA (G4), incidences of in-hospital all-cause death and fatal arrhythmia were higher than in other groups in the overall population and the subgroups. Multivariate logistic regression showed that an increase in the SUA level of ≥30 µmol/L was significantly associated with in-hospital all-cause death and fatal arrhythmia, independent of the baseline SUA level and renal function. This association was significant in most subgroups for in-hospital fatal arrhythmia and in the ≥60 years, myocardial infarction, and female subgroups for in-hospital all-cause death. The cutoff values of SUA increases in the overall population were 54.5 µmol/L for in-hospital all-cause death and 42.6 µmol/L for in-hospital fatal arrhythmia. The perioperative SUA increase significantly correlated with a higher incidence of in-hospital all-cause death and fatal arrhythmia in CABG patients, independent of the baseline SUA level and renal function. Perioperative SUA variation may provide complementary information in the identification of patients potentially at risk.