Abstract:

Objective To analyze the risk factors for interstitial lung disease （ILD）in hospitalized patients with polymyositis（PM）or dermatomyositis（DM）and to establish a clinical prediction model. Methods 379 in patients with PM or DM in Tangdu Hospital of Air Force Military Medical University from January 1，2015 to January 1，2021 were included in the study and they were divided into a control group（PM/DM）and a case group（PM/DM⁃ILD）. The clinical data on the patients were retrospectively collected. The risk factors of PM/DM⁃ ILD were summarized by binary logistic regression. A clinical prediction model was constructed by using the R software. Results The patients′ age was older；the incidences of joint pain，fever，mechanic′s hand，Raynaud′s phenomenon，chest tightness，shortness of breath，cough and expectoration were greater；and erythrocyte sedimen⁃ tation rate（ESR），serum ferritin（FER）and levels of carcinoembryonic antigen（CEA），carbohydrate antigen 153（CA153），anti⁃Jo⁃1 antibody，anti⁃Ro52 antibody and anti⁃mda5 antibody were higher in the control group than in the control group；while level of serum albumin（ALB）was lower（P < 0.05）. Binary logistic regression analysis showed that age，fever，mechanic′s hand，Reynolds phenomenon，FER，CEA，serum albumin（ALB）， anti⁃Jo⁃1 antibody，anti⁃Ro52 antibody and anti⁃mda5 antibody were the risk factors of ILD（P < 0.05）. The AUC of the constructed clinical prediction model was 0.899（95%CI：0.864 ~ 0.934）. The sensitivity was 79.5% and the specificity was 86.4%. Conclusions The risk factors of PM/DM⁃ILD are old age，fever，mechanic′s hand，Reyn⁃ olds phenomenon，increased levels of FER and CEA，a decrease in serum albumin（ALB），and positive anti⁃Jo⁃1， anti⁃Ro52 and anti⁃mda5 antibodies. The model constructed in this study can better predict the occurrence of PM/ DM⁃ILD，and it has certain reference value for early clinical prevention and treatment.

Key words:

polymyositis, dermatomyositis, interstitial lung disease, risk factors, Nomogram model