Objective The anti-FⅩa assay can be used to monitor the blood concentration of Rivaroxaban. The aim is to evaluate the critical value and diagnostic performance of this test on bleeding risk assessment. Methods From September 2017 to June 2019, 368 patients were enrolled for retrospective cohort study, including 201 males and 167 females, aged (62.8±15.7) years old. They were divided into groups by age:≤60 years old group 105 cases,61-70 years old group 135 cases,≥71 years old group 128 cases. Anti-FⅩa was detected on ACL TOP 700 coagulation analyzer using chromogenic substrate method to quantitatively determine the plasma concentration of rivaroxaban. Anti-FⅩa data were expressed as M (P25-P75);Kruskal-Wallis H test was used for comparison among groups; Mann-Whitney U test was for data comparison between two groups; positive rate comparison was performed by χ2 test; the diagnostic performance of anti-FⅩa to assess bleeding risk was evaluated by ROC curve;Kaplan-Meier curve was used for the survival analysis;the risk ratio (HR) was obtained by Cox proportional hazard regression model. Results Both the peak and trough plasma concentrations were higher in patients aged 61-70 years old than ≤60 years old (U values were 5 618 and 5 725,respectively, P values were 0.006 and 0.011, respectively); higher in patients ≥71 years old than 61-70 years old (U values were 6 438 and 6 317, respectively, P values were 0.05).Both peak and trough blood concentrations were higher in patients with bleeding than without bleeding(U values were 1 429 and 2 185, respectively, P<0.001 and 0.001, respectively).ROC showed that the cut-off values of peak blood concentration in evaluation of the overall and the ≥61 year-old population′s bleeding risk were 200.8 ng/ml and 209.9 ng/ml,respectively, corresponding respectively with the sensitivity of 90.9% and 95.0%; the trough cut-off values were 35.1 ng/ml and 39.1 ng/ml, respectively, corresponding respectively with the sensitivity of 72.7% and 70.0%. However, all the above cut-off values gave a low diagnostic specificity. Survival analysis showed with 35.1 ng/ml as the trough cut-off value, the cumulative risk of bleeding significantly increased in patients above the cut-off value (Log-rank χ2=4.513,P=0.034). The Cox proportional regression model demonstrated that the hazard ratios for peak and trough blood concentration predictions of bleeding risk were 1.023 (95%CI: 0.834-1.256) and 0.948 (95%CI: 0.773-1.164). respectively. Conclusions Both the peak and trough values of blood concentration in bleeding patients are higher than non-bleeding patients. The peak blood concentration is highly sensitive to the risk of bleeding, and the elevated trough blood concentration levels indicate that the probability of bleeding risk increases in the short term. However, the specificity of both peak and trough values is relatively low in bleeding risk assessment. When used alone, the prediction of bleeding events does not have direct guiding significance. Dynamic monitoring and joint evaluation are recommended. Key words: Factor Ⅹa; Rivaroxaban; Plasma concentration; Blood coagulation tests; Hemorrhage; Forecasting