Background: This systematic review evaluates existing risk prediction models for intradialytic hypotension (IDH) in maintenance hemodialysis (MHD) patients, aiming to inform the development of high-quality predictive tools for clinical use.
Materials and methods: Retrieve studies on the construction of predictive models for IDH risk in patients undergoing maintenance hemodialysis in CNKI and other databases. The search time frame is from the establishment of the databases to November 13, 2024. Two researchers independently screened the literature and extracted data according to the Predictive Model Study Data Extraction Form and bias risk assessment tools. The bias risk and applicability of the included literature were evaluated.
Results: A total of 21 studies were included, with 16 undergoing internal validation, and 8 reporting calibration. IDH incidence ranged from 7.3 to 51.0%. The overall applicability of the studies included in the research is good, but the overall risk of bias is high, mainly due to unreasonable sample size, lack of performance evaluation, and single-center studies.
Conclusion: The research on predictive models for IDH risk in patients undergoing maintenance hemodialysis is still in its early stages. The included studies exhibit an overall high risk of bias, and there is a lack of clinical application. In the future, it may be beneficial to utilize interpretable machine learning methods to construct predictive models with good performance and simplicity, aiming for practical clinical applications.