Clinical characteristics and risk model of lower gastrointestinal involvement in systemic sclerosis.

Abstract

Objectives: Systemic sclerosis (SSc) is a systemic autoimmune disease that could involve multiple organs. Lower gastrointestinal involvement (LGIT) in systemic sclerosis (SSc-LGIT) is a serious manifestation and has a poor prognosis. This study aims to explore the related risk factors of SSc-LGIT that require hospitalisation and build a clinical risk model.

Methods: SSc patients with LGIT admitted to Peking Union Medical College Hospital (PUMCH) were enrolled between December 2003 and December 2023. The controls were selected in SSc patients without LGIT in the same period after matching age and gender at a ratio of 1:3. Clinical data of both groups was collected to build the SSc-LGIT clinical prediction model by machine learning using R software.

Results: A total of 42 SSc patients with LGIT and 126 matched SSc patients without LGIT were enrolled. Compared to the control group, SSc-LGIT patients had lower level of body mass index (BMI), haemoglobin (HB), albumin (ALB). Cardiomyopathy and puffy finger were more common, but arthritis/arthralgia was less. Higher hs-CRP and a higher rate of anti-Ro-52 antibody positivity were found in these patients. A multivariate analysis revealed BMI, cardiomyopathy, HB, ALB, hs-CRP as independent related factors for SSc-LGIT, and a clinical risk model containing these five items was built.

Conclusions: A clinical risk model of SSc-LGIT was established and it has demonstrated capability in predicting the risk of severe lower gastrointestinal involvement in systemic sclerosis.