Relationship between immune indexes and prognosis in patients with sepsis: data analysis based on MIMIC-IV database

Abstract

To investigate the relationship between the changes of immune indexes and prognosis in patients with sepsis. Based on the hospitalization information from 2008 to 2019 in the American Medical Information Mart for Intensive Care-IV v2.0 (MIMIC-IV v2.0), the hospitalized data including CD3 count, CD4 count, CD8 count, CD4/CD8 ratio, immunoglobulins (IgA, IgG, IgM) levels and lymphocyte counts, gender, age, body mass index (BMI), sequential organ failure assessment (SOFA), and prognosis of patients with coronary heart disease, hypertension, diabetes mellitus, underlying diseases, and 28-day prognosis were collected. The immunological indexes affecting the prognosis of patients with sepsis were compared and analyzed. The receiver operator characteristic curve (ROC curve) was drawn to analyze the value of immunological indexes in predicting the 28-day prognosis of patients with sepsis. A total of 33 745 patients were enrolled in the study, including 1 509 cases of sepsis and 32 236 cases without sepsis. Among the patients with sepsis, there were 1 084 cases alive and 425 cases died within 28 days. BMI (kg/m2: 33.25±28.71 vs. 28.90±15.28) and IgA levels within 3 days and 7 days after admission in sepsis group were significantly higher than those in the non-sepsis group [within 3 days after admission: (275.01±216.96) mg/L vs. (85.99±75.76) mg/L; within 7 days after admission: (275.01±216.96) mg/L vs. (85.99±75.76) mg/L, all P < 0.05]. CD3 count, CD8 count within 3 days and 7 days after admission in sepsis group were significantly lower than that in the non-sepsis group [within 3 days after admission: CD3 counts were (1 080.74±849.23)/μL vs. (1 242.91±889.24)/μL, CD8 counts were (558.07±368.77)/μL vs. (625.07±529.66)/μL; within 7 days after admission: CD3 counts were (1 079.69±850.61)/μL vs. (1 242.48±889.23)/μL, CD8 counts were (556.70±467.23)/μL vs. (624.93±429.78)/μL, all P < 0.05]. The proportion of patients with hypertension, diabetes and coronary heart disease were significantly higher than those in the non-sepsis group [hypertension: 10.07% (152/1 509) vs. 2.47% (796/32 236), diabetes: 50.10% (756/1 509) vs.15.18% (4 895/32 236), coronary heart disease: 31.21% (471/1 509) vs. 19.19% (6 186/32 236), all P < 0.05]. In the sepsis group, the CD3 count, CD4 count and CD4/CD8 ratio in the survival group were significantly higher than those in the death group within 3 days and 7 days after admission [within 3 days after admission: CD3 counts were (1 127.20±857.14)/μL vs. (938.26±810.50)/μL, CD4 counts were (559.76±507.18)/μL vs. (338.75±267.11)/μL, CD4/CD8 ratios were 1.87±0.80 vs. 1.02±0.12; within 7 days after admission: CD3 count were (1 124.01±810.53)/μL vs. (943.78±808.21)/μL, CD4 count were (559.56±507.36)/μL vs. (341.95±266.56)/μL, CD4/CD8 ratios were 1.88±0.79 vs. 1.03±0.13, all P < 0.05]. The area under the ROC curve (AUC) of CD4/CD8 ratio within 3 days after admission in predicting the prognosis of septic patients at 28 days was 0.615 [95% confidence interval (95%CI) was 0.555-0.671], the sensitivity was 70.83%, and the specificity was 52.49%. Hypertension, coronary heart disease and diabetes are more likely to induce sepsis. The changes of immune indexes in patients with sepsis will occur in the early stage. Early CD4/CD8 index can predict the prognosis of patients with sepsis to a certain extent.