Elevated Serum Chloride Levels Contribute to a Poor Prognosis in Patients with IgA Nephropathy

Abstract

Introduction. The identification of reliable prognostic factors is a crucial requirement for patients with IgA nephropathy (IgAN). Here, we explored the relationship between serum chloride levels and prognosis in patients with IgAN. Methods. We recruited all patients with primary IgAN, as diagnosed by renal biopsy, between 1st January 2015 and 1st April 2019. Patients were divided two groups (high chloride group and low chloride group) based on the best cut-off values from survival receiver operating characteristic (ROC) curves. The baseline clinicopathological characteristics of two groups were then compared. Cox proportional hazard models were used to determine the prognostic value of serum chloride levels in patients with IgAN. Finally, we screened reliable prognostic indicators and built a clinical prediction model and validated the performance of the model. Results. Compared with patients in the high chloride group, patients in the low chloride group had significantly lower levels of 24-hour urinary total protein (24 h-UTP), serum creatinine (sCr), and higher levels of hemoglobin (Hb), albumin (all $p<0.05$ ), and less proportion of Oxford classification grade E1 (endothelial cell proliferation) and T2 (renal tubule atrophy or renal interstitial fibrosis). Cox analysis revealed that $\text{serum}\text{chloride}\text{level}\ge 105.4\text{mmol}/\text{L}$ was a significant and independent risk factor for prognosis in patients with IgAN ( $p<0.05$ ). Serum chloride, sCr, T, hypertension, and Hb were used to generate a predictive model for prognosis. The $c$ -indices of our predictive model were 0.80, 0.86, and 0.78, for 1, 2, and 3 years, respectively; Brier scores were 0.06, 0.09, and 0.16, respectively. Conclusions. A $\text{serum}\text{chloride}\text{level}\ge 105.4\text{mmol}/\text{l}$ was identified as a significant and independent risk factor for the prognosis of patients with IgAN. A predictive prognosis model was generated using serum chloride, sCr, T, hypertension, and Hb; this model exhibited a good predictive effect.