ТЯЖЕЛЫЕ ПОРАЖЕНИЯ ЛЕГКИХ И ПРОГНОЗИРОВАНИЕ РИСКА ИХ РАЗВИТИЯ У ПАЦИЕНТОВ СО СПОНДИЛОАРТРИТАМИ И РЕВМАТОИДНЫМ АРТРИТОМ

Abstract

The severity of pulmonary involvement in patients with spondylarthritеs and rheumatoid arthritis is determined by the nature of lung affection and functional disorders. Objective. To single out severe affection of the lungs among the varieties of pulmonary involvement in patients with spondylarthritеs and rheumatoid arthritis and to devise a model predicting the risk of development of severe lung affection in patients with spondylarthritеs and rheumatoid arthritis. Materials and methods. The study involved 172 patients with spondylarthritеs and rheumatoid arthritis. The patients were interviewed. The presence of antibodies to atypical lung infections in blood was determined. The multispiral computer tomography of thorax was performed. The diffusion lung capacity was tested. In order to process the data Statistica 10 software package was used. With aim to predict the risk of development of severe lung affection logistic regression was carried out, also a logistic model was built. Results. Varieties of lung involvement leading to a decrease in the diffusion lung capacity were qualified as severe lung involvement. In view of the revealed risk factors a logistic model was built which predicted the risk of development of severe lung affection in patients with spondylarthritеs and rheumatoid arthritis. The predictive value of the model was 80,7%. According to ROC-analysis the cut-off point was 0,47. AUC=0,81±0,05; р<0,001. Model sensitivity was 83,9%, its specificity was 70,0%. Its working capacity was confirmed on an independent sample of patients. In accordance with the check results the predictive value of the model was 78,1%. Conclusion. Severe lung affection in patients with spondylarthritеs and rheumatoid arthritis leads to a decrease in the lung gas exchange. The logistic regression model makes it possible to predict the risk of development of severe lung involvement in this group of patients.