A clinical model for identifying proliferative lupus nephritis: a multicenter study

Objectives Proliferative lupus nephritis (PLN) is associated with poorer renal outcomes than non-proliferative lupus nephritis. Although kidney biopsy remains the gold standard for histologic classification, its use may be limited by invasiveness, contraindications, and limited healthcare access. We aimed to develop and externally validate a noninvasive model for identifying PLN. Methods In this multicenter retrospective study, 443 patients with biopsy-confirmed lupus nephritis from five centers were included. The development cohort comprised 337 patients from center A, and the external validation cohort included 106 patients from centers B to E. Candidate predictors were screened using univariable analysis and least absolute shrinkage and selection operator regression, followed by multivariable logistic regression to construct the final nomogram. Internal validation was performed using 5-fold cross-validation and bootstrap resampling. Results The final model incorporated seven predictors, namely lymphocyte percentage, hematuria, leukocyturia, estimated glomerular filtration rate, uric acid, anti-dsDNA positivity, and C3. The model showed good discrimination, with an area under the receiver operating characteristic curve of 0.852 in the development cohort and 0.895 in the external validation cohort. Calibration was acceptable in both cohorts, and decision curve analysis suggested potential clinical utility. Conclusion We developed and externally validated a noninvasive clinical model based on routinely available clinical variables for identifying PLN. This model may provide a practical tool for risk stratification and support clinical evaluation in patients with lupus nephritis.