Kidney Week

Abstract: SA-PO0833

Development of a Predictive Model for Monoclonal Gammopathy of Renal Significance: A Retrospective Cohort Study from Northwest China

Session Information

• Glomerular Diseases: Management, Evolving Strategies, and Practice-Changing Advances
  October 24, 2026 | Location: Exhibit Hall A, Convention Center
  Abstract Time: 10:00 AM - 12:00 PM

Category: Glomerular Diseases

• 1402 Glomerular Diseases: Clinical, Outcomes, and Therapeutics

Authors

• He, Erhadan, The First Affiliated Hospital of Xi'an Jiaotong University, Xi'an, Shaanxi, China

Erhadan He

• Li, Huixian, The First Affiliated Hospital of Xi'an Jiaotong University, Xi'an, Shaanxi, China

Huixian Li, MD

• Yuan, Xiaohan, The First Affiliated Hospital of Xi'an Jiaotong University, Xi'an, Shaanxi, China

Xiaohan Yuan, PhD

• Xie, Xinfang, The First Affiliated Hospital of Xi'an Jiaotong University, Xi'an, Shaanxi, China

Xinfang Xie, MD, PhD

• Lu, Wanhong, The First Affiliated Hospital of Xi'an Jiaotong University, Xi'an, Shaanxi, China

Wanhong Lu, MD

Background

As chronic kidney disease (CKD) and monoclonal gammopathy (MG) increasingly coexist, distinguishing MG of renal significance (MGRS) from MG of undetermined significance (MGUS) is essential given their divergent treatments. Given the risks of renal biopsy, we developed a predictive model of MGRS in a large northwestern Chinese cohort.

Methods

In this retrospective study, MG patients who underwent renal biopsy at the First Affiliated Hospital of Xi'an Jiaotong University from 2018 to 2025 were enrolled. Multivariable logistic regression was performed to identify independent risk factors and to construct predictive models for MGRS and amyloidosis, respectively.

Results

A total of 221 patients with MG who underwent kidney biopsy were enrolled. Among them, 125 (56.6%) were diagnosed with MGRS, and 96 (43.4%) were classified as non-MGRS. In the MGRS group, immunoglobulin-related amyloidosis was the most common pathology (n=89, 71.2%), followed by monoclonal immunoglobulin deposition disease (MIDD; n=18, 14.4%). In the non-MGRS group, membranous nephropathy (MN) was the leading diagnosis (n=40, 41.7%), followed by diabetic nephropathy (n=11, 11.5%). A predictive model for MGRS was developed incorporating 11 variables: age, systolic blood pressure, diabetes, low C3, nephrotic syndrome, hematuria, lambda light chain, affected/unaffected free light chain (FLC) ratio, positive serum immunofixation (SIF), eGFR < 60 mL/min/1.73 m², and positive M protein. The model showed good discriminative ability for MGRS, with an area under the receiver operating characteristic curve (AUC) of 0.854 (95% CI: 0.804–0.904). A separate predictive model for amyloidosis was constructed using nine variables: age, systolic blood pressure, diabetes, nephrotic syndrome, hematuria, lambda light chain, affected/unaffected FLC ratio, positive SIF, and log-transformed creatinine. This model demonstrated excellent discriminative performance for identifying amyloidosis among MG patients, with an AUC of 0.925 (95% CI: 0.892–0.959).

Conclusion

In summary, this study provides real-world evidence on the epidemiological profile of MGRS in Northwest China and proposes a practical predictive model to facilitate early screening, hierarchical management, and appropriate referral to specialized centers. External validation and biomarker integration are needed to enhance its generalizability and precision.