| Objective:To investigate the expression and ratio of co-inhibitors and co-stimulators on lymphocytes subsets[including T lymphocytes,regulatory T cells(Treg cells)and natural killer cells(NK cells)]for predicting prognosis in sepsis patients.MethodsA prospective observational cohort study was conducted.Critically ill patients admitted to ICU and emergency ICU who met the new diagnostic criteria of sepsis 3.0 were enrolled.Observation indicators:(1)APACHE Ⅱ and SOFA scores were recorded daily within a week;(2)Cell counts and expression rate of PD-1 and CD28 of CD4 + T cells and CD8+ T cells in the peripheral blood were determined,respectively,on day 1.Along with this,cell counts and expression rate of PD-1,CD28,PD-L1 and CD86 in Treg cells and NK cells were assessed on day 1.(3)Additionally,plasma levels of IL-6,IL-8 and INF-gamma were evaluated by enzyme linked immunosorbent assay(ELISA)on day 1.Moreover,C reactive protein and procalcitonin were detected on day 1,3 and 5.Sepsis patients were divided into survival group and non-survival group according to whether they died or not on day 28 or during the hospitalized period.Nosocomial infection group and non-nosocomial infection group were based on whether the nosocomial infection had occurred during hospitalization.High expression group and low expression group were based on the best cut-off point as determined by ROC curve analysis of prognostic predictors.All data were analyzed by SPSS 20.0(SPSS,Chicago,IL,USA)software and R software(version 3.1.0).The differences of CD4+T cells,CD8+ T cells,Treg and NK cell counts,and their co-signaling molecules among the survival group and the non-survival group were compared.Logistic regression analysis was adopted to screen independent risk factors for predicting mortality and established prognostic prediction models.Following this,the predictive effect of each index on the prognosis,using ROC curve analysis,the sensitivity and specificity on the best cut-off point,the stratified analysis according to the best cut-off point etc,were evaluated.At the same time,to investigate the relationship between inflammatory mediators and expression of Co-signaling molecules in lymphocyte subsets and their correlation with APACHE Ⅱ and SOFA scores.Results1.Percentage of PD-1+ cells in CD8+T cells,and ratio of PD-1/CD28 positive cells in CD8+T cells in non-survival group were significantly higher than those in the survival group for 28-days mortality and hospital mortality.Ratio of PD-1/CD28 positive cells in CD8+T cells as well as mean fluorescence intensity of PD-1 in CD4+T cells in the nosocomial infection group were significantly higher than those in the non-nosocomial infection group.ROC curve analysis showed that percentage of PD-1+ cells in CD8+T cells,and ratio of PD-1 over CD28 positive in CD8+T cells could predict death,and ratio of PD-1 to CD28 positive percentage in CD8 T cells could predict nosocomial infection.2.Percentage of PD-1+ cells,CD86+ cells and the average fluorescence intensity of PD-1 in Treg cells in non-survival group were significantly higher than those in the survival group for 28-days mortality,while percentage of CD86+ cells and the average fluorescence intensity of PD-1 in Treg cells in non-survival group were significantly higher than those in the survival group at hospitalization.ROC curve analysis showed that percentage of PD-1+ cells,CD86 cells and the average fluorescence intensity of PD-1 in Treg cells could predict death for 28 days,and percentage of CD86+cells and the average fluorescence intensity of PD-1 in Treg cells could predict hospital mortality.3.Cell count of NK cells was significantly lower in non-survival group than that in the survival group for 28-days mortality.Percentage of PD-L1+ cells in NK cells and ratio of PD-L1 to CD86 positive percentage in NK cells in non-survival group were significantly higher than those in the survival group for 28 days.The counts of NK cells were significantly lower in non-survival group than those in the survival group for 28 days.Percentage of PD-L1+ cells in NK cells in non-survival group was significantly higher than those in the survival group for hospital mortality.ROC curve analysis showed that the counts of NK cells,PD-L1+ cells and ratio of PD-L1 to CD86 positive percentage in NK cells could predict death for 28 days,and the counts of NK cells,PD-L1+ cells in NK cells could predict hospital mortality.4.Multivariate analysis:Percentage of PD-1+ cells in CD8 T cells on day 1 was independent risk factors for predicting 28-day mortality or hospital mortality.The sensitivity and specificity of the prediction model in combination of percentage of PD-1+ cells in CD8+T cells and heart rate for predicting 28-day mortality were 90%and 87.8%,respectively;the area under the ROC curve of the model was higher than that of APACHE Ⅱ score and SOFA score[0.907(0.826-0.989)VS 0.771(0.625-0.917)VS 0.715(0.528-0.901),P = 0.132 and 0.053],but there was no statistical difference.The sensitivity and specificity of the prediction model in combination of percentage of PD-1+ cells in CD8 T cells and heart rate for predicting hospital mortality were 91.7%and 84.6%respectively;the area under the ROC curve of the model was higher than that of APACHE II score and SOFA score[0.889(0.800-0.977)VS 0.723(0.568-0.879)VS 0.670(0.486-0.854),P=0.087 and 0.025],but there was no statistical difference between the model and APACHE II score,5.Percentage of PD-1+ cells in CD8+T cells,and ratio of PD-1 to CD28 positive cells in CD8+T cells were positively correlated with CRP on day 3,while percentage of PD-1+ cells in Treg cells and the average fluorescence intensity of PD-1 in Treg cells were negatively correlated with the level of plasma INF-gamma,percentage of CD86+ cells in Treg cells was positively correlated with the level of plasma IL-8;NK cell counts were negatively correlated with the level of plasma IL-8,CRP(d3)and PCT(d5);percentage of PD-LI1 cells in NK cells was positively correlated with PCT(dl and d5);ratio of PD-L1 to CD86 positive cells were positively correlated with PCT(dl,d5)and CRP(d5).Expression of costimulatory molecules on lymphocyte subsets was positively correlated with APACHE Ⅱ score and SOFA score on day 2-7.Conclusiotn1.The counts of lymphocyte subsets and the expression of their co-signaling molecules,including PD-1+ cells in CD8 T cells,ratio of PD-1 to CD28 positive cells in CD8+T cells,percentage of PD-1+ cells,CD86+ cells and the average fluorescence intensity of PD-1 in Treg cells,the counts of NK cells,PD-L1+ cells and ratio of PD-L1 to CD86 positive cells in NK cells are risk factors for death of septic patients.2.Percentage of PD-1+ cells in CD8 T cells on day 1 was independent risk factors for predicting 28-day mortality or hospital mortality.The predictive model,which is a combination of percentage of PD-1+ cells in CD8+T cells and heart rate,may be superior to the current critical illness score in predicting prognosis.3.Ratio of PD-1 to CD28 positive cells in CD8+T cells is also a risk factor for predicting nosocomial infection.4.Expression rate of co-signaling molecules in lymphocyte subsets is associated with the severity of the disease within a week,which may affect the prognosis of septic patients by regulating the secretion of inflammatory factors. |
PDF Full Text Request