1. Construction and application of fetal loss risk model in systemic lupus erythematosus patients with mild disease severity
- Author
-
Yanran Chen, Yanjuan Chen, Bo Li, Wengyi Xu, Peipei Lei, Hongyang Liu, Dongzhou Liu, and Xiaoping Hong
- Subjects
Systemic lupus erythematosus ,Mild disease severity ,Pregnancy outcome ,Fetal loss ,Prediction nomogram ,Gynecology and obstetrics ,RG1-991 - Abstract
Abstract Background This dynamic nomogram model was developed to predict the probability of fetal loss in pregnant patients with systemic lupus erythematosus (SLE) with mild disease severity before conception. Methods An analysis was conducted on 314 pregnancy records of patients with SLE who were hospitalized between January 2015 and January 2022 at Shenzhen People's Hospital, and the Longhua Branch of Shenzhen People's Hospital. Data from the Longhua Branch of the Shenzhen People's Hospital were utilized as an independent external validation cohort. The nomogram, a widely used statistical visualization tool to predict disease onset, progression, prognosis, and survival, was created after feature selection using multivariate logistic regression analysis. To evaluate the model prediction performance, we employed the receiver operating characteristic curve, calibration curve, and decision curve analysis. Results Lupus nephritis, complement 3, immunoglobulin G, serum albumin, C-reactive protein, and hydroxychloroquine were all included in the nomogram model. The model demonstrated good calibration and discriminatory power, with an area under the curve of 0.867 (95% confidence interval: 0.787–0.947). According to decision curve analysis, the nomogram model exhibited clinical importance when the probability of fetal loss in patients with SLE ranged between 10 and 70%. The predictive ability of the model was demonstrated through external validation. Conclusion The predictive nomogram approach may facilitate precise management of pregnant patients with SLE with mild disease severity before conception.
- Published
- 2024
- Full Text
- View/download PDF