Analysis Of Related Influencing Factors Of Interstitial Pulmonary Disease Complicated With Pulmonary Hypertension And Construction Of Prediction Model

Objective:Taking all patients with ILD as the research object,including their various laboratory examinations and imaging features,the risk factors of pulmonary hypertension were screened through statistical analysis,and the risk prediction model was established based on the above results,so as to detect the progress of the disease early and timely,and predict the possibility of pulmonary hypertension.Methods:From November 2019 to November 2022,244 patients with ILD were selected from the Department of Respiratory and Critical Care Medicine of the Affiliated Hospital of Chengde Medical College.They were divided into non-pulmonary hypertension group and pulmonary hypertension group according to the presence of pulmonary hypertension.Compare the baseline data,laboratory examination and imaging characteristics of the two groups,carry out logistic regression analysis on the influencing factors(p<0.05 is statistically significant),use the AIC information criteria to select the optimal model,and use the receiver operating characteristic curve(ROC)and Hosmer-Lemeshow goodness of fit test to evaluate the model,And draw the model calibration diagram and decision curve analysis(DCA)to understand the accuracy of the model.Results:1.Comparison of basic data between the two groupsThe patients were divided into non-PH groupand PH groupaccording to the results of pulmonary artery pressure in echocardiography(with 40mm Hg as the boundary).In the PH group,there were 107 patients with an average age of 65 years,50.5%(n=54)were women,and 49.5%(n=53)were men.In the non-PH group,there were 137 patients with an average age of 63 years.48.9%(n=67)were women and 51.1%(n=70)were men.After comparison,NE,RDW,NT-pro BNP,D-D,NLR,ALB,CA125,CYFRA21.1,P/F and RVEDD were significantly different between the two groups(p<0.05).2.Regression analysisThe above variables were included in the single-factor logistic regression analysis.The results showed that NE(OR=1.155,95%CI:1.021-1.305,P=0.021),NLR(OR=1.131,95%CI:1.038-1.234,P=0.005),RDW(OR=1.282,95%CI:1.181-1.391,P<0.001),NT-pro BNP(OR=1.005,95%CI:1.013-1.061,P<0.001),D-D(OR=1.506,95%CI:1.091-2.078,P=0.013),ALB(OR=0.919,95%CI:0.872-0.968,P=0.002),CA125(OR=1.010,95%CI:1.004-1.017,P=0.003),CYFRA21.1(OR=1.269,95%CI:1.118-1.439,P<0.001),P/F(OR=0.992,95%CI:0.989-0.995,P<0.001),RVEDD(OR=1.559,95%CI:1.371-1.774,P<0.001),the above variables have statistical significance,and they are analyzed by multivariate logistic regression.The above variables were further included in the multivariate logistic regression analysis.The results showed that RDW(OR=1.178,95%CI:1.063-1.305,P=0.002),NT-pro BNP(OR=1.013,95%CI:1.023-1.118,P=0.004),CYFRA21.1(OR=1.267,95%CI:1.053-1.523,P=0.120),RVEDD(OR=1.604,95%CI:1.338-1.922,P<0.001)were independent risk factors for PH.3.Model establishment and model visualizationAccording to the multi-factor logistic regression analysis,the results show that the four variables RDW,NT-pro BNP,CYFRA21.1 and RVEDD are the final predictors,and the AIC information criterion is used to select the optimal model.The Nomogram diagram is drawn to visualize the model.4.Model evaluationThe area under ROC(AUC)drawn according to the model is 0.894(95%CI:0.858-0.939),which proves that the model has the ability to identify the onset of PH.The goodness of fit test result of Hosmer-Lemeshow x~2=3.513,P=0.898,indicating that the fitting degree is good.The calibration curve shows the curve between the model and the predicted value,which proves the high feasibility of the model.The decision curve shows that this model can bring maximum benefits to patients.Conclusions:1.For patients with definite diagnosis of ILD,when one or more of RDW,NT-pro BNP,RVEDD and CYFRA21.1 are elevated in the laboratory,they should be alert to the subsequent occurrence of pulmonary hypertension.It is recommended to actively screen for possible primary disease and select the best treatment method.2.The diagnostic prediction model established in this study based on risk factors can guide the clinical prediction of the occurrence of pulmonary hypertension.