Objective:To explore the risk factors influencing the short-term mortality of patients with sepsis in intensive care unit (ICU) and the combined value of predicting prognosis.Methods:A retrospective analysis was performed on 104 patients with sepsis admitted to emergency ICU of Jiangsu Provincial Hospital of Traditional Chinese Medicine from January 2018 to August 2019. Multiple general information containing gender, age, past history as well as complications and sequential organ failure assessment (SOFA) score, mean arterial pressure (MAP), blood routine examination, hepatic and renal function, coagulation indicators and procalcitonin (PCT) were collected within 24 hours of admission. Patients were divided into death group and survival group according to the 28-day outcome. Univariate and multivariate Logistic regression analysis were used to find the effective factors influencing the prognosis of sepsis. Receiver operating characteristic (ROC) curve was drawn to evaluate the value of related indexes in predicting the prognosis of sepsis. Correlation between parameters that might be relevant to disease severity and SOFA score was evaluated by Pearson or Spearman correlation analysis.Results:104 patients were enrolled for final analysis, of whom 60 patients survived, while the others died with a 28-day mortality of 42.3%. ① Univariate analysis results: the incidence of acute kidney injury (AKI), SOFA score, serum creatinine (SCr), D-dimer, activated partial thromboplastin time (APTT), international normalized ratio (INR) and PCT in the death group were significantly higher than those in the survival group [incidence of AKI: 70.5% (31/44) vs. 36.7% (22/60), SOFA score: 11.0 (8.0, 13.0) vs. 8.0 (6.2, 10.0), SCr (μmol/L): 108.8 (65.5, 235.6) vs. 75.1 (55.1, 109.5), D-dimer (mg/L): 4.1 (1.6, 11.6) vs. 2.1 (1.2, 4.3), APTT (s): 42.6 (37.7, 55.7) vs. 40.3 (35.9, 44.7), INR: 1.3 (1.2, 1.5) vs. 1.2 (1.1, 1.4), PCT (μg/L): 3.1 (0.4, 39.9) vs. 0.3 (0.1, 3.4), all P < 0.05]. ② Multivariate Logistic regression analysis results: all indicators of univariate analysis were included in the multivariate Logistic regression model considering interaction between each variable. Multivariate Logistic regression analysis was repeated based on conditional backward method. Age, SOFA score, MAP, neutrophil (NEU), lymphocyte (LYM) and APTT were automatically selected by SPSS software to build the predicting model. Analysis results showed that SOFA score, NEU and LYM were independent risk factors for the short-term prognosis of sepsis [SOFA score: odds ratio ( OR) = 1.22, 95% confidence interval (95% CI) was 1.04-1.44, P = 0.02; NEU: OR = 1.14, 95% CI was 1.03-1.26, P = 0.01; LYM: OR = 0.79, 95% CI was 0.66-0.95, P = 0.01]. ③ ROC curve analysis results: the above six-variable prediction model had the optimal fitting degree defaulted by SPSS. ROC curve showed that the combination of age [area under ROC curve (AUC) = 0.60], SOFA score (AUC = 0.71), MAP (AUC = 0.53), NEU (AUC = 0.59), LYM (AUC = 0.54) and APTT (AUC = 0.61) had better sensitivity (79.5%) and specificity (65.0%) as well as the maximal AUC (AUC = 0.75), which suggested that combined prediction had higher diagnostic value in predicting the short-term prognosis of sepsis.④ Correlation analysis showed that NEU, D-dimer, prothrombin time (PT), APTT, INR and PCT were positively correlated with SOFA score ( r values were 0.26, 0.28, 0.21, 0.22, 0.10, 0.38, respectively, all P < 0.05). Conclusions:SOFA score, NEU and LYM were independent risk factors for the short-term prognosis of sepsis. The combination of age, SOFA score, MAP, NEU, LYM and APTT were more accurate than any single factor in predicting the short-term prognosis of sepsis and had higher diagnostic value. NEU, D-dimer, PT, APTT, INR and PCT were correlated with SOFA score.