# Risk factors and predictive model for renal outcomes in autoimmune membranous nephropathy with and without acute kidney injury: a retrospective cohort study

## TL;DR

This study identifies risk factors and builds a predictive model for kidney outcomes in autoimmune membranous nephropathy patients, with and without acute kidney injury.

## Contribution

A novel predictive model for renal outcomes in autoimmune membranous nephropathy patients is developed and validated.

## Key findings

- AKI, triglycerides, serum creatinine, hematuria, and anti-M-type PLA2R staining are risk factors for poor renal outcomes.
- Serum C3 is a protective factor against renal endpoints.
- The nomogram model shows strong predictive performance with a C-index of 0.845 and high AUC values.

## Abstract

This study aimed to delineate the risk factors affecting renal outcomes in autoimmune membranous nephropathy (aMN) with or without acute kidney injury (AKI) and develop a predictive model.

This retrospective cohort study included 441 patients with biopsy-confirmed aMN from the First Affiliated Hospital of Fujian Medical University (January 2010 to March 2023). Patients were grouped based on the presence of AKI and followed up until a renal endpoint event (progression to end-stage renal disease, initiation of dialysis, or either a >40% decline in estimated glomerular filtration rate from baseline or a doubling of serum creatinine levels from baseline, both sustained for ≥3 months) or study endpoint (March 2024). Clinicopathological and renal outcomes were collected and analyzed. Risk factors for renal endpoints were identified via Cox regression analyses, and a nomogram was constructed. Model performance was evaluated using the C-index, time-dependent receiver operating characteristic (Time-ROC) curves, calibration curves, and decision curve analysis (DCA). Kaplan–Meier survival curves compared renal survival between AKI subgroups.

Among 441 patients, 109 (24.72%) experienced AKI. Renal endpoint events occurred in 40.4% of the AKI group and 4.5% of the non-AKI group. Multivariate Cox regression identified AKI (HR = 7.298, P < 0.001), triglycerides (HR = 1.140, P = 0.002), serum creatinine (HR = 1.008, P = 0.012), hematuria (HR = 2.246, P = 0.040), and kidney anti-M-type phospholipase A2 receptor staining 4+ (HR = 2.473, P = 0.003) as independent risk factors, while serum C3 (HR = 0.082, P < 0.001) was an independent protective factor. The nomogram had a C-index of 0.845 (P < 0.001), with Time-ROC AUCs of 0.92, 0.81, 0.83, and 0.87 for 3 to 6 years, respectively. Calibration plots revealed good consistency between the predicted and actual probabilities. DCA indicated that the nomogram had potential clinical utility. Kaplan–Meier analysis showed lower cumulative renal survival in patients with AKI (P < 0.001).

The risk factor model suggests that renal outcomes in patients with aMN can be predicted. Early assessment and management targeting these identified risk factors could help delay renal function decline in these patients.