This narrative review synthesizes evidence from a systematic literature search (2019–2025) on artificial intelligence (AI) applications in hemodialysis and peritoneal dialysis for managing infections and complications. Dialysis patients face infection rates 100-fold higher than the general population and sepsis mortality exceeding 35%. The AI, utilizing machine learning algorithms such as XGBoost, deep neural networks, and explainable AI tools, revolutionizes care through precise diagnostics, prognostics, and therapeutics. Key models demonstrate high accuracy in predicting bloodstream infections (area under the curve AUC 0.914), classifying peritonitis (F1 0.93), detecting SARS-CoV-2 (AUROC 0.82), and forecasting mortality (AUC 0.979). Therapeutically, AI guides antibiotic stewardship, reducing inappropriate use by 18–67%, and mitigates intradialytic hypotension via ultrafiltration adjustments, reducing incidence by 25–40%. Despite promising results, challenges include data scarcity, algorithmic bias, and the integration of these tools into clinical workflows. Future directions involve diverse datasets, explainable AI, and real-time decision support systems. The AI holds transformative potential for personalizing dialysis management and improving patient outcomes.
Waleed H. Mahallawi (2026) studied this question.