Analysis of risk factors and establishment of risk prediction model on interstitial lung disease complicated with respiratory failure

Abstract:

Objective：To explore the clinical characteristics of interstitial lung disease（ILD） complicated with respiratory failure （RF），compare their differences，analyze their risk factors of ILD with RF，and establish a risk prediction model. Methods：A total of 291 patients with ILD in our hospital from January 2013 to January 2018 were retrospectively analyzed and were divided into the case group（n=98） and the control group（n=193）. Their clinical data were compared and analyzed，and logistic regression analysis was used. The model was established，and Hosmer-Lemeshow goodness of fit test and receiver operating characteristic（ROC） curve were used to evaluate the model. Results：During the past 5 years，patients with ILD and RF were taken 33.7% among all hospitalized pa－tients with ILD. Compared with the control group，the mortality in the case group was significantly increased，accounting for 17.3%. Patients in the two groups had significant differences in main symptoms，the aggravated time of symptoms，smoking index，C-reactive protein（CRP） and the lesion of chest CT（P<0.05）. Logistic regression analysis revealed that acute exacerbation within short time，dyspnea，increased CRP，and CT diffusion focus were independent risk factors of patients with ILD and RF. The area under the ROC curve for the evaluation model was 0.854（95%CI=0.805-0.903），with sensitivity of 67.1% and specificity of 89.5%. Conclusion：More attention should be paid to ILD patients who have an acute exacerbation in short time，with dyspnea as main symptom，accompanying with high CRP level，diffuse lesion and ground-glass opacity，or combined with cardiovascular disease and pulmonary infection. Because those patients may be more likely to have RF，with high mortality and poor final outcome. The risk prediction model might be an effective tool for predicting whether there will be respiratory failure of ILD patients，with can be used for clinical practice.