节点文献
TBI术后高渗治疗对ICP和短期死亡的影响及短期死亡预测模型的研究
Effect of Hyperosmotic Therapy on Intracranial Pressure and Short-term Mortality after Decompressive Craniectomy and Derivation of Predictive Models
【作者】 程峰；
【导师】 王之敏；
【作者基本信息】 苏州大学 ， 神经外科（专业学位）， 2019， 博士
【摘要】 第一部分:颅脑创伤患者行去骨瓣减压术后高渗治疗对颅内压和短期死亡的影响目的:探讨高渗盐水(Hypertonic Saline,HTS)和甘露醇对去骨瓣减压(Decompressive Craniectomy,DC)术后颅脑创伤(Traumatic Brain Injury,TBI)患者中颅内压(Intracranial Pressure,ICP)升高的控制作用及短期(两周内)死亡的影响。方法:回顾性分析2012年1月1日至2017年8月31日期间入住江苏大学附属昆山医院行DC术后TBI患者的临床资料。其中只接受一种高渗剂(3%HTS或20%甘露醇)治疗的患者被录入研究。观察指标包括:(1)每日ICP负担(小时/天);(2)对单次大剂量高渗剂的反应;(3)两组间重症监护室(Intensive Care Unit,ICU)和总住院天数;(4)两周内死亡率。结果:30例仅接受3%HTS治疗的患者和30例仅接受20%甘露醇治疗的患者进行近似匹配和数据分析。两组患者的人口统计学特征具有可比性,HTS组每日ICP负担明显低于甘露醇组(分别为0.89±1.02小时/天:2.11±2.95小时/天;P=0.038)。HTS组单次给药后ICP降低的程度较甘露醇组更高(P=0.001)。而短期死亡率在两组间无显著性差异(2,66.67%[HTS组]:1,33.33%[甘露醇组];P=0.554)。结论:DC术后TBI患者使用等渗剂量高渗剂治疗时,3%HTS的降颅压效果优于20%甘露醇。但这种优势对此亚组患者的短期死亡率没有影响。第二部分:颅脑创伤去骨瓣减压术后短期内死亡发生的危险因素分析及预测模型的建立目的:分析DC术后TBI患者短期死亡的相关危险因素,并建立短期死亡预测模型。方法:回顾性分析2012年1月1日至2018年12月31日期间入住江苏大学附属昆山医院行DC术后TBI患者的临床资料,包括年龄、性别、受伤机制、GCS评分、GCS动作评分、瞳孔反应、术前低氧经历、头部简略损伤评分(the Abbreviated Injury Score of the Head region,HAIS)、收缩压水平、头颅CT特征及实验室检查资料。经单因素分析、错误发现率(False Discovery Rate,FDR)检验和多因素分析筛选出相关独立危险因素,用于建立DC术后TBI患者短期死亡的预测模型。通过验证集数据检验模型并观察模型性能。最后,借助编程技术开发预测工具。结果:337例DC术后TBI患者中,112(33.2%)例发生短期死亡。经单因素分析、FDR检验和多因素分析后,发现年龄、入院GCS评分、瞳孔反应、Marshall CT分级、HAIS和Fg与短期死亡的发生独立相关。基于独立危险因素建立的预测模型的C统计值为0.930,阈值P为0.359,当P≥0.359时认为发生短期死亡的可能性大。测试集的验证结果显示MMC值为0.799,F-measure值为0.865。结论:建立的预测模型可以早期、方便、准确地预测DC术后TBI患者短期死亡的发生,开发的预测工具可辅助临床决策的制定。
【Abstract】 Part I:Effect of hypertonic therapy on intracranial pressure and short-term mortality in patients with traumatic brain injury after decompressive craniectomyObjectives: The current retrospective study was carried out to compare the effect of hypertonic saline(HTS)versus mannitol on lowering increased intracranial pressure(ICP)burdens and decreasing short-term mortality in head-injured patients undergoing decompressive craniectomy(DC).Methods: We reviewed data on TBI patients after DC admitted between January 1,2012,and August 31,2017.Patients who received only 1 type of hyperosmotic agent,3% HTS or 20% mannitol,were included.Daily ICP burden(hours/day)and response to hyperosmolar agent were used as primary outcome measures.The numbers of intensive care unit days,numbers of hospital days and 2-week mortality rates were also compared between the groups.Results: 30 patients who received 3% HTS only and 30 who received 20% mannitol only were identified for the approximate matching and additional data analyses.The demographic data of the patients were comparable between the 2 groups.The daily ICP burden was significantly lower in the HTS group than in the mannitol group(0.89±1.02 hours/day vs 2.11±2.95 hours/day,respectively;P = 0.038).The slope of fall in ICP in response to a bolus dose at a baseline value of ICP was higher with HTS compared to mannitol(p = 0.001).However,the 2-week mortality was not statistically significant(2,66.67% [HTS] vs 1,33.33% [mannitol];P = 0.554).Conclusion: When used in equiosmolar doses,reduction in ICP with 3% HTS is favorable to that with 20% mannitol in TBI patients after DC.However,these superiorities do not seem to confer any additional benefit terms of short-term mortality.Part II:Analysis of risk factors for short-term death after decompressive craniectomy in patients with traumatic brain injury and establishment of prediction modelObjectives: The objective of this part was to analyze the risk factors of short-term mortality in traumatic brain injury(TBI)patients after DC and to establish a short-term mortality prediction model.Methods: We reviewed data on TBI patients after DC admitted between January 1,2012,and December 31,2018,including age,sex,injury mechanism,GCS score,GCS motor score,pupil reaction,preoperative hypoxia experience,the abbreviated injury score of the head region(HAIS),systolic blood pressure,head CT features and laboratory examination data.We used univariate analysis,false discovery rate(FDR)test and multivariate analysis were used to screen out the independent risk factors for the establishment of short-term mortality prediction model for TBI patients after DC surgery.Test the model by validation set data and observe the performance of the model.Finally,the prediction tool is developed with the help of programming technology.Results: Of 337 TBI patients after DC,112(33.2%)had short-term death.After univariate analysis,FDR test and multivariate analysis,age,admission GCS score,pupil response,Marshall CT grade,HAIS and Fg were found to be independently associated with short-term mortality.The C statistic value of the predictive model based on independent risk factors was 0.930,and the threshold P was 0.359.When P≥0.359,it was considered that the possibility of short-term mortality was high.The validation results of the test set showed that the MMC value was 0.799 and the F-measure value was 0.865.Conclusion: The established predictive model can predict the short-term mortality of TBI patients after DC early,conveniently and accurately.The developed predictive tool can assist clinical decision-making.