# Unveiling risk factors: a prognostic model of frequent peritonitis in peritoneal dialysis patients

## TL;DR

This study develops a predictive model to identify peritoneal dialysis patients at high risk of frequent peritonitis episodes.

## Contribution

The novel contribution is a risk prediction model for frequent peritonitis episodes in peritoneal dialysis patients using clinical biomarkers.

## Key findings

- The model includes risk factors like diabetes mellitus, hemoglobin, and serum albumin.
- The model achieved an AUC of 0.75 in training and 0.76 in testing, showing good predictive performance.
- A nomogram was developed to help clinicians assess patient risk more intuitively.

## Abstract

Peritoneal dialysis-associated peritonitis (PDAP) is a serious complication of peritoneal dialysis (PD) patients. The aim of this study was to construct a risk prediction model for frequent episodes in PDAP patients.

This retrospective cohort study included PDAP patients in our center from January 1, 2010 to December 31, 2021. The risk prediction model for frequent episodes in PDAP patients was constructed by the binary logistic regression.

We included 371 PDAP patients, of which 235 patients had single episode and 136 had frequent episodes. We randomly allocated the patients into training set (296 patients) and test set (75 patients) in the ratio of 8:2. In the training set, we found several independent risk factors significantly associated with frequent episodes in PDAP patients, including diabetes mellitus (DM), hemoglobin (HB), serum albumin (ALB), lactatic dehydrogenase (LDH), serum potassium (K), N-terminal pro-brain natriuretic peptide (NT-proBNP) and peritoneal dialysate white cell counts on day 1. And we constructed a prediction model with an area under curve (AUC) values of 0.75 in the training set and 0.76 in the test set, which showed excellent predictive performance.

We constructed a predictive model that demonstrated excellent predictive performance for identifying high-risk frequent episodes in PDAP patients and developed a more intuitive nomogram for evaluating the risk. However, multicenter studies with a larger sample size are warranted to validate the model in the future.