人工智能在维持性血液透析糖尿病患者血糖管理中的应用研究进展

Research Progress on the Application of Artificial Intelligence in Glycemic Management of Diabetic Patients Undergoing Maintenance Hemodialysis

摘要: 人工智能(artificial intelligence, AI)技术的发展为糖尿病管理提供了新的手段。维持性血液透析(maintenance hemodialysis, MHD)糖尿病患者由于血糖波动大、胰岛素代谢异常及多因素干扰，其血糖管理复杂，传统经验性调整模式存在局限。本文综述人工智能在该人群血糖管理中的应用进展，重点分析其在血糖预测、胰岛素剂量优化及风险评估等方面的研究现状。现有研究表明，基于机器学习、深度学习及强化学习的模型能够整合多维临床数据，提高血糖预测准确性并优化治疗策略。同时，人工智能在临床应用中仍面临数据质量、模型可解释性及临床整合等问题。未来需通过多中心研究及前瞻性验证进一步提升其可靠性与可推广性。人工智能有望成为透析患者血糖管理的重要辅助工具。

Abstract: The rapid development of artificial intelligence (AI) has introduced novel approaches for diabetes management. Glycemic control in diabetic patients undergoing maintenance hemodialysis (MHD) remains particularly challenging due to pronounced glucose fluctuations, altered insulin metabolism, and multiple dialysis-related interfering factors, rendering traditional experience-based adjustment strategies insufficient. This review summarizes recent advances in the application of AI for glycemic management in this population, with particular emphasis on its roles in glucose prediction, insulin dose optimization, and risk assessment. Current evidence suggests that machine learning-, deep learning-, and reinforcement learning-based models can integrate multidimensional clinical data to improve the accuracy of glucose prediction and optimize therapeutic strategies. Nevertheless, challenges remain regarding data quality, model interpretability, and integration into clinical workflows. Future multicenter studies and prospective validations are warranted to further enhance the reliability and generalizability of these technologies. AI is expected to become an important adjunctive tool in glycemic management for patients undergoing hemodialysis.