Objective To evaluate the efficacy and safety of endotoxin adsorption therapy added to conventional therapy in patients with sepsis or infectious shock due to Gram-negative bacterial infections in abdominal cavity while compared with only conventional therapy. Methods From 2014.1 to 2016.9，a multicenter，randomized controlled trial included 19 patients with sepsis or infectious shock due to Gram-negative bacterial infections in abdominal cavity from West China hospital of Sichuan university，The First Affiliated Hospital of Xinjiang Medical University，Zhujiang Hospital of Southern Medical University and so on. Enrolled patients randomized to either endotoxin adsorption therapy added to conventional therapy (n=8) or conventional therapy (n=11). Primary outcome mainly included the change of Sequential Organ Failure Assessment scores (SOFA) on third day after treated and the changes of SOFA scores，mean arterial pressure (MAP) and vasoactive drugs on the third、7th、14th day after treated，the incidence rate of acute kidney injury (AKI)，28-day mortality also were evaluated. Results Compared with conventional therapy only，SOFA may more obviously decrease in endotoxin adsorption therapy added to conventional therapy group. The SOFA in test group and control on the third day，7th，14th was (7.67±1.86 vs 9.88±4.73)，(7.75±4.57 vs 9.33±6.74)，and (4.67±2.08 vs 11.00±4.24) respectively. Due to the small sample size，statistical difference was not significant (P>0.05).28-day mortality also was no significant difference between the test group and control (35.7% VS.39.4%，P>0.05). Besides，the changes of MAP and vasoactive drugs also were no significant difference between the two groups. Conclusion Compared with only conventional therapy，there is a tendency to decrease SOFA for patients with sepsis or infectious shock due to Gram-negative bacterial infections in abdominal cavity treated by endotoxin adsorption therapy added to conventional therapy，however，there is no significance. More trials are needed to confirm the argument.