Raw data.

Objective
Accurate risk stratification is crucial for personalized management in peritoneal dialysis (PD). This study aimed to develop and validate a novel biomarker, the SII-PNI score, integrating systemic immunoinflammatory index (SII) and prognostic nutritional index (PNI), for personalized mortality risk prediction in PD patients.
Methods
A retrospective cohort study analyzed data from 330 patients initiating PD between December 2005 and June 2023 at a single center. Patients were stratified into four risk groups based on median cut-off values for SII and PNI. The association between SII-PNI risk groups and mortality (all-cause, cardiovascular [CVD], infection-related) was assessed using Kaplan-Meier survival analysis and multivariable Cox proportional hazard models. The predictive performance of the SII-PNI score was evaluated using receiver-operating characteristic (ROC) curves and compared to individual components (SII, PNI) and CRP. Random survival forests assessed variable importance.
Results
The high-risk group (G4: high SII + low PNI) had the shortest PD duration, highest mortality, and worst survival outcomes. Compared to the low-risk group (G3: low SII + high PNI), G4 had significantly increased risks of all-cause mortality (adjusted HR 3.36, 95% CI 1.93–8.67), CVD mortality (adjusted HR 3.74, 95% CI 2.41–19.60), and infection mortality (adjusted HR 4.32, 95% CI 2.58–20.4) in fully adjusted models. The SII-PNI score demonstrated superior predictive ability (AUC: all-cause 0.80, CVD 0.80, infection 0.81) compared to SII, PNI, or CRP alone. Random survival forests confirmed the critical importance of the individual components (platelets, neutrophils, lymphocytes, albumin) for outcomes.
Conclusions
The SII-PNI score, derived from readily available blood parameters, is a powerful and convenient tool for personalized risk stratification in PD. Patients identified as high-risk warrant intensified monitoring and early interventions. This composite biomarker represents a significant step towards personalized and precision medicine in PD care, with potential for implementation in routine clinical practice using standard laboratory data.