Abstract:

Peritoneal dialysis-associated peritonitis (PDAP) is a major cause of hospitalization, technique failure, and mortality in patients undergoing peritoneal dialysis. The development of effective clinical prediction models is essential for early warning and individualized intervention in PDAP. As research has shifted from analyses of individual risk factors toward multifactorial integrated prediction, PDAP risk assessment has expanded from conventional regression-based models to machine learning (ML) approaches. This narrative review summarizes the major modeling methods, key predictors, and characteristics of prediction models across different PDAP-related clinical endpoints, aiming to inform clinical decision-making and future research.

Key words: peritoneal dialysis-associated peritonitis, clinical prediction model, machine learning, nomogram, risk factors, risk assessment