Impact of immune marker-based microinflammation on dialysis efficacy in diabetic nephropathy patients: Development of an early prediction model.

Abstract

Background: Peritoneal dialysis (PD) is widely used for treating end-stage renal disease (ESRD) in diabetic nephropathy patients. Suboptimal dialysis outcomes are often linked to chronic microinflammation, yet predictive models integrating immune markers remain limited.

Materials and methods: In this prospective cohort study, 236 diabetic nephropathy patients undergoing PD were categorized into high-inflammation (HI) and microinflammation (MI) groups based on serum levels of CRP, IL-6, and TNF-α. Clinical and biochemical data - including Kt/V, infection rates, and survival - were collected. Logistic regression was applied to identify predictors of poor dialysis outcomes and construct a predictive model.

Results: The HI group showed significantly lower dialysis efficiency (Kt/V < 1.5), increased peritonitis and catheter infections, and reduced survival. Elevated CRP, IL-6, and TNF-α levels were observed alongside higher creatinine, urea nitrogen, fasting glucose, lipids, and notable electrolyte and acid-base imbalances. Logistic regression identified CRP, IL-6, TNF-α, Kt/V, infection rates, and selected biochemical markers as independent predictors of poor outcomes. The resulting predictive model yielded an AUC of 0.87, demonstrating strong discriminative power.

Conclusion: Microinflammation is a key determinant of dialysis efficacy in diabetic nephropathy patients receiving PD. The proposed model, based on inflammatory and dialysis-related parameters, offers a promising approach for early risk stratification and personalized treatment. Further validation may enhance long-term management and survival outcomes in this population.