摘要
目的 系统评价全身麻醉患者围手术期低体温风险预测模型.方法 检索中国知网、中国生物医学文献数据库、万方数据库、维普期刊库、PubMed、Web of Science、Embase、Cochrane Library、Medline中与全身麻醉患者围手术期低体温风险预测模型相关的研究,检索时限为建库至2024年3月12日.研究者根据纳入、排除标准筛选文献,2名研究员依据预测模型研究数据提取表和偏倚风险评估工具独立进行资料提取和质量评价.结果 共纳入22项研究,22个模型.21项研究受试者操作特征曲线下面积(AUC)/一致性指数(C-index)为0.709～0.898,其中1项研究未报告AUC.年龄、体重指数、输液量/补液量、手术时间、手术室温度、麻醉时间、基础体温为预测模型重复报告的独立预测因子.22篇文献整体呈现高偏倚风险和好的适用性.结论 现有的全身麻醉患者低体温风险预测模型预测性能较好,但存在方法学缺陷和高偏倚风险.未来建议针对现有模型进行验证和更新,参考不同方法学研究构建规范、针对不同人群的风险预测模型.
Abstract
Objective To systematically evaluate the model for predicting the risk of perioperative hypothermia in patients undergoing general anesthesia.Methods The studies related to perioperative hypothermia risk prediction models for patients taking general anaesthesia were searched from China National Knowledge Infrastructure,China Biomedical Literature Database,Wanfang Database,Wikipedia Journal Library,PubMed,Web of Science,Embase,Cochrane Library,and Medline from their establishment to 12 March 2024.The investigators screened the literatures according to the inclusion and exclusion criteria,and 2 researchers independently carried out data extraction and quality assessment based on the Data Extraction Form for Predictive Modelling Studies and the Risk of Bias Assessment Tool.Results A total of 22 studies containing 22 models were included;the area under the receiver operating characteristic curve(AUC)/C-index of 21 studies was 0.709-0.898;1 study did not report the AUC.Age,body mass index,infusion volume/rehydration volume,operation time,operation room temperature,anesthesia time,and basal body temperature were the independent predictors reported repeatedly by the predictive model.22 articles overall present a high risk of bias and good applicability.Conclusions The existing prediction models for hypothermia risk in patients taking general anesthesia have good predictive performance,but there are methodological deficiencies and high bias risks.In the future,it is recommended to validate and update the existing models,refer to different legal studies to construct standardized risk prediction models for different populations.
维普
万方数据