Lipoprotein-associated phospholipase A2 (Lp-PLA2) probably plays an important role in the development of acute coronary syndrome (ACS). However, alterations of Lp-PLA2 levels during ACS and its association with cardiovascular outcome are unclear. Our aim was to investigate the change in Lp-PLA2 and its association with cardiovascular outcome in patients with ACS. A total of 79 patients with ACS came from the coronary care unit (CCU) between June 1, 2015 and August 31, 2016 in this longitudinal study. Serum levels of Lp-PLA2, troponin I, and creatine kinase isoenzymes MB (CK-MB) were measured at admission, on the first morning (D1), on the second morning of hospitalization (D2), and on the last second morning before discharge (D4). The patients were followed up till November 30, 2016. The primary outcomes were cardiovascular death and cardiovascular rehospitalization. Kaplan-Meier analysis and Cox proportional hazard models were used to identify risk factors for poor outcome in patients with ACS. All patients were followed up for 10.6 +/- 4.7 months. The patients were divided into 2 groups according to the median of Lp-PLA2: lower Lp-PLA2 group and higher Lp-PLA2 group. Elevated levels of Lp-PLA2 significantly decreased during the early phases of ACS in higher Lp-PLA2 group. And Lp-PLA2 level increased at first and then decreased in lower Lp-PLA2 group. Kaplan-Meier analysis showed that patients with elevated Lp-PLA2 had a lower cardiovascular event-free survival (log-rank chi(2)=4.736, P=.030) than those with lower Lp-PLA2. Cox regression analysis indicated that high Lp-PLA2 level (hazard ratio [HR] = 1.005, 95% confidence interval [CI] = 1.002-1.008, P=.003), time delay from symptom onset to admission (HR=1.088, 95% CI=1.038-1.139, P<.001) independently predicted cardiovascular event in patients with ACS after adjusting for potential confounders. Serum level of Lp-PLA2 altered considerably during the early phase of ACS and increased Lp-PLA2 independently predicted cardiovascular outcome in patients with ACS after adjustment for potential confounders.