Abstract: Objective To conduct a systematic evaluation of prediction models for the risk of acquired weakness in mechanically ventilated patients. Methods We retrieved literature on prediction models for the risk of acquired weakness in mechanically ventilated patients in PubMed, Web of Science, Embase, Cochrane Library, CNKI, Wanfang, and VIP. The retrieval period spanned from the inception of the databases to December 2023. Two researchers independently screened the literature, extracted data, and assessed the bias risk and applicability of the included studies. Results Sixteen articles were included. The area under the receiver operating characteristic curve for the included models ranged from 0.710 to 0.926. Bias risk assessment showed high bias risk in all models, but good applicability. The top 5 predictive factors in terms of frequency were mechanical ventilation time, age, Acute Physiology and Chronic Health Evaluation II score, blood lactate level, and multi-organ dysfunction. The combined AUC value of the six validation models was 0.800 (95%CI: 0.740-0.850), indicating good discrimination. Conclusion Prediction models for the risk of acquired weakness in mechanically ventilated patients demonstrate overall good predictive performance, but further optimization is needed in terms of data sources, design, and statistical analysis. Future efforts should focus on external validation of existing models or the development of high-quality predictive models.

Key words: mechanical ventilation, acquired weakness, intensive care unit, risk prediction model, systematic review