A New Perspective on Predicting PLA2R-Associated Membranous Nephropathy Relapse: The Value of a Genetic Risk Score

Purpose: Membranous nephropathy (MN) is a common glomerular disease characterized by high relapse rates and heterogeneous outcomes. This study aimed to develop a genetic risk score (GRS) based on five MN-associated single nucleotide polymorphisms (SNPs) and assess its predictive value for disease relapse. Methods: In this prospective study, we analyzed data from 234 patients with phospholipase A2 receptor (PLA2R)-associated MN between January 2020 and December 2023. Genotyping for five SNPs associated with MN risk (rs28383345, rs2187668, rs35771982, rs3749117, and rs4664308) was performed. A GRS was constructed and Cox regression models were used to assess risk factors for remission and relapse. Predictive performance was evaluated using Cox regression, time-dependent receiver operating characteristic (tROC) curves, net reclassification improvement (NRI), integrated discrimination improvement (IDI), Akaike information criterion (AIC), Bayesian information criterion (BIC), likelihood ratio test (LRT), and cross-validation. Results: Over a median follow-up duration of 28.0 (IQR 20.0, 39.0) months, the cumulative remission rate was 85.5%, with 47% relapsing. A high GRS was significantly associated with the risk of relapse (HR = 1.885, 95% CI: 1.331– 2.585; P < 0.001). Adding GRS to the base model consistently increased the time-dependent AUC at years 2, 4, and 5 (all P < 0.05). Notably, assessments using risk reclassification metrics (IDI/NRI) and model fit metrics (LRT/AIC/BIC) also verified significant improvements in model performance across multiple years. Critically, rigorous repeated cross-validation demonstrated that the overall C-index gain provided by the GRS was both stable and significant (P < 0.05), and further year-by-year cross-validation confirmed that this advantage persisted across all evaluated years (all P < 0.05). Furthermore, sensitivity analysis further confirmed the robustness of the GRS. Conclusion: This study is the first to apply a GRS in predicting relapse in PLA2R-associated MN. GRS significantly enhances predictive accuracy, offering a valuable tool for personalized risk assessment.