Predicting the risk of lupus nephritis based on a nomogram in patients with childhood-onset systemic lupus erythematosus
Shuolan Jing, Xianglin Yang, Shihao Li, Liqun DongAbstract
Background
Lupus nephritis (LN) is one of the leading causes of death and a major cause of morbidity in childhood-onset systemic lupus erythematosus (SLE). However, currently there is no tool to predict the risk of LN in patients with childhood-onset SLE.
Methods
This study collected and retrospectively analysed data of paediatric patients who were first diagnosed with SLE at the West China Second Hospital of Sichuan University from January 2013 to December 2022. A nomogram prediction model for predicting the risk of LN was developed using LASSO-logistic regression analysis. Receiver operating characteristic (ROC) curve, calibration curves, and decision curve analysis were used to assess the performance of the prediction model.
Results
A total of 264 paediatric patients with childhood-onset SLE were included; 171 with and 93 without LN. LASSO regression and multivariate logistic regression analyses showed that serositis, anti-dsDNA positivity, low IgG, and low albumin were independent risk factors for identifying patients at high risk for LN and were thus included in predictive models to identify such high-risk groups. The final results show that the calibration curve is close to the ideal prediction situation, and the decision curve analysis and ROC curve (AUC = 0.874) suggest that the model has good predictive power.
Conclusions
Four indicators, serositis, anti-dsDNA positivity, low IgG, and low albumin, were independent risk factors predictive of high-risk LN in patients with childhood-onset SLE. The model has been validated internally and performs well.