摘要
目的:基于Logistic回归模型构建献血相关迷走神经反应(DRVR)简易评分量表,评价其应用价值。方法:纳入2018年1月至2023年1月张家口市中心血站5 000例献血者进行回顾性研究。收集献血者基线资料,对其行Logistic回归分析构建DRVR预测模型,基于比值比(OR)构建DRVR简易评分量表。按照低危、中危、高危划分评分区间,基于Logistic回归模型预测效能评价量表的应用价值。结果:献血者DRVR发生率为4.12%(206/5 000)。Logistic回归模型确定了高龄、女性、低体质量指数、首次献血、低血压、低血红蛋白水平、长采血时间、高献血量等8项导致DRVR发生的独立危险因素(均P<0.05),基于OR值构建了DRVR简易评分量表。将量表评分划分为低危(总分值<6)、中危(总分值6-11)、高危(总分值>11且≤17)3个区间。Logistic回归模型预测低危、中危、高危献血者DRVR的AUC分别为0.784(95%CI:0.742-0.844)、0.806(95%CI:0.752-0.883)、0.842(95%CI:0.761-0.925);一致性指数分别为0.794(95%CI:0.705-0.828)、0.800(95%CI:0.745-0.852)、0.839(95%CI:0.782-0.917);校准曲线均未偏离理想拟合。外部验证结果显示,ROC曲线拟合均较为理想(P>0.05),AUC差异均无统计学意义(均P>0.05)。结论:DRVR简易评分量表对献血者的区分能力较为理想,能够为快速缩小高危群体范围提供一定的参考。DRVR简易评分量表具有一定的应用价值,但仍需不断完善和优化。
Objective:To construct a simple score scale for donation-related vasovagal reaction(DRVR)based on Logistic regression model and evaluate its application value.Methods:A retrospective study was conducted on 5000 blood donors from Zhangjiakou Central Blood Station from January 2018 to January 2023.The baseline data of the blood donors were collected.Logistic regression analysis was performed on them to construct a DRVR prediction model.A DRVR simple score scale was built based on the odds ratios.The score intervals were divided into low-risk,medium risk,and high-risk categories,the application value of the scale was evaluated based on the prediction efficacy of Logistic regression model.Results:The incidence of DRVR among the blood donors was 4.12%.The Logistic regression model identified eight independent risk factors for DRVR,including advanced age,female gender,low body mass index,first blood donation,hypotension,low hemoglobin level,long blood collection time,and high blood donation volume(all P<0.05).A simple DRVR score scale was constructed based on the odds ratio values.The scale scores were divided into three intervals:low-risk(total score<6),moderate risk(total score 6-11),and high-risk(total score>11,≤17).The area under the curve(AUC)for predicting DRVR in low-risk,medium risk,and high-risk blood donors based on Logistic regression model were 0.784(95%CI:0.742-0.844),0.806(95%CI:0.752-0.883),and 0.842(95%CI:0.761-0.925),respectively.The consistency indices were 0.794(95%CI:0.705-0.828),0.800(95%CI:0.745-0.852),and 0.839(95%CI:0.782-0.917),respectively.The calibration curves did not deviate from the ideal fit.The results of external validation showed that the receiver operating characteristic curve fittings were relatively ideal(all P>0.05),and there were no statistically significant difference in AUC(all P>0.05).Conclusion:The DRVR simple score scale has a relatively ideal ability to distinguish the blood donors and can provide some reference for quickly narrowing the scope of high-risk groups. DRVR simple score scale has certain application value, but still needs to be continuously improved and optimized.
作者
齐菲
赵建鹏
QI Fei;ZHAO Jian-Peng(Zhangjiakou Central Blood Station,Zhangjiakou 075000,Hebei Province,China)
出处
《中国实验血液学杂志》
北大核心
2025年第6期1745-1750,共6页
Journal of Experimental Hematology
基金
河北省医学科学研究课题计划项目(20211263)
河北省医学科学研究课题计划项目(20220589)。
作者简介
通信作者:赵建鹏,主任医师.E-mail:1277473912@qq.com。
