Evaluation Of Predictive Models For The Impact Of Immune Function Status On The Severity Of Community Acquired Pneumonia Patients

Objective:Using the existing clinical evaluation methods of immune function,the nomogram method was used to establish a prediction model of immune function in patients with community acquired pneumonia(CAP)to predict the severity of the disease,so as to identify the abnormal immune function early.In CAP patients with possible disease progression,the prognosis of the disease should be indicated as early as possible,and the clinical immunotherapy and comprehensive treatment intervention should be carried out early,so as to improve the prognosis of patients.Methods:A total of 113 CAP patients admitted to Wuhan Central Hospital from October 2019 to January 2022 were selected as the research objects.The clinical data and immune-related examination data of the patients were collected.The subjects were randomly divided into training set and validation set according to the ratio of 7:3.The patients were divided into severe group and mild-moderate group.The immune-related factors in the two sets were analyzed by univariate analysis,and the severity-related immune factors were screened for multivariate Logistic regression analysis.R4.2.2(Microsoft open R4.2.2)software was used to construct a nomogram model for predicting the severity of CAP patients’immune function.The receiver operating characteristic(ROC)curve was drawn and the Hosmer-Lemeshow goodness of fit test was performed.According to the relevant results,the discrimination and calibration of the nomogram model for predicting the severity of CAP were evaluated.The performance of the prediction model was evaluated.Results:The training set included 79 patients,and the age of patients was 67.0(57.0,75.0)years old.There were 30 male patients(62.0%)and 49female patients(38.0%).The validation set included 34 patients,with an age of 71.0(59.5,81.3)years,including 16 males(47.1%)and 18 females(52.9%).Univariate analysis of immune-related factors in the training set showed that the platelet count,lymphocyte count,and the percentages of CD3+,CD8+,and CD19+cells in lymphocytes in the severe group were lower than those in the mild-to-moderate group,and the differences were statistically significant(P=0.020,P=0.004,P=0.002,P=0.000,P=0.047).The white blood cell count,neutrophil count,neutrophil-to-lymphocyte ratio(NLR)and CD4/CD8 in the severe group were higher than those in the mild-to-moderate group,and the differences were statistically significant(P=0.024,P=0.007;P=0.000,P=0.011).There were no statistically significant differences in gender,platelet-lymphocyte ratio(PLR),CD3-CD16+CD56+cells and CD4+cells in lymphocytes between the severe group and the mild and moderate group(P=0.670,P=0.108,P=0.766,P=0.936).Multivariate regression Logistic analysis of patients in the training set showed that age(OR=1.149,95%CI=1.021～1.292;P=0.021),platelet count(OR=0.983,95%CI=0.971-0.995,P=0.005),NLR(OR=1.149,95%CI=1.02～1.292,P=0.021),P=0.021)and the percentage of CD8+cells in lymphocytes(OR=0.853,95%CI=0.766-0.961,P=0.009)were independent influencing factors for the severity of CAP patients(P<0.05).A nomogram prediction model for the severity of CAP was established based on age,platelet count,NLR,and CD8+cell percentage in lymphocytes in the training set and validated in the validation set.the area under the curve(AUC)of ROC curve was0.961(95%CI:92.16%-100.12%).In the validation set,the AUC value of the prediction model was 0.885(95%CI:76.76%-100.16%),suggesting that the prediction model established by the above four factors had a good discrimination ability for the severity of CAP patients.In the training set,Hosmer-Lemeshowχ~2=2.970,P=0.936>0.05;In the validation set,Hosmer-Lemeshowχ~2=3.870,P=0.869>0.05,suggesting that the prediction model established by the above four immune-related factors has a good calibration ability for the poor severity of CAP patients.Conclusion:Age,platelet count,NLR and the percentage of CD8+cells in lymphocytes are immune-related factors for the severity of CAP.The prediction model for the severity of CAP based on these four indicators has a good early identification ability for the severity of CAP,which can help doctors to identify patients with poor prognosis early and intervene as soon as possible.