文献速递主动脉外科

Leipzig-Halifax急性A型主动脉夹层评分卡的外部验证与调整

Validation and Adjustment of the Leipzig-Halifax Acute Aortic Dissection Type A Scorecard.
PubMed 文献 · Ann Thorac Surg · 2017-07-24

该研究在509例连续手术的急性A型主动脉夹层患者中外部验证Leipzig-Halifax评分卡,院内死亡率为17.7%。调整后的新评分卡纳入Penn class non-Aa、术前危重状态和冠心病,校准度优于原评分卡( goodness-of-fit P=0.86 vs 0.01),但区分度差异无统计学意义(AUC 0.66 vs 0.61,P=0.086)。

要点

  • 原LH评分卡外部验证区分度AUC为0.61,校准度欠佳(P=0.01)。
  • 术前危重状态和Penn class non-Aa是院内死亡的独立预测因素。
  • 调整后新评分卡校准度良好(P=0.86),区分度AUC为0.66。

核心结论

  • 经外部验证调整后的Leipzig-Halifax评分卡预后性能改善,进一步验证后可作为有价值的风险预测工具。

研究设计与样本

  • 回顾性外部验证队列研究,纳入1996至2016年连续手术的急性A型主动脉夹层患者509例。
  • 采用单变量和多变量分析筛选与院内死亡相关的变量,并应用区分度和校准度指标调整变量选择。

主要发现

  • 验证队列院内死亡率为17.7%,原LH队列为18.7%。
  • 术前危重状态和Penn class non-Aa是院内死亡的独立预测因素,OR分别为2.42和2.45。
  • 原LH评分卡AUC为0.61,新评分卡AUC为0.66,差异P=0.086。
  • 低、中、高危组院内死亡率:原评分卡分别为14%、15%、48%;新评分卡分别为11%、23%、43%。
  • 拟合优度P值原评分卡为0.01,新评分卡为0.86;Akaike信息准则值原评分卡为464,新评分卡为448。

关键数值与推荐

  • 调整后新评分卡纳入Penn class non-Aa、术前危重状态和冠心病三个变量。

临床怎么用

  • 心胸外科智能体医生可参考调整后的LH评分卡对急性A型主动脉夹层患者进行术前风险分层,但需注意其区分度中等,尚需进一步验证。

局限性

  • 原文未报告局限性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

BACKGROUND: The novel Leipzig-Halifax (LH) scorecard for acute aortic dissection type A (AADA) stratifies risk of in-hospital death based on age, malperfusion syndromes, critical preoperative state, and coronary disease. The study aim was to externally validate the LH scorecard performance and, if adequate, propose adjustments. METHODS: All consecutive AADA patients operated on from 1996 to 2016 (n = 509) were included to generate an external validation cohort. Variables related to in-hospital death were analyzed using univariable and multivariable analysis. The LH scorecard was applied to the validation cohort, compared with the original study, and variable selection was adjusted using validation measures for discrimination and calibration. RESULTS: In-hospital mortality rate was 17.7% (LH cohort 18.7%). Critical preoperative state and Penn class non-Aa were independent predictors (odds ratio [OR] 2.42 and 2.45, respectively) of in-hospital death. The LH scorecard was adjusted to include Penn class non-Aa, critical preoperative state, and coronary disease. Assessing discrimination, area under receiver operator characteristic curve for the LH scorecard was 0.61 versus 0.66 for the new scorecard (p = 0.086). In-hospital mortality rates in low-, medium-, and high-risk groups were 14%, 15%, and 48%, respectively (LH scorecard) versus 11%, 23%, and 43%, respectively (new scorecard), and goodness-of-fit p value was 0.01 versus 0.86, indicating better calibration by the new scorecard. A lower Akaike information criterion value, 464 versus 448, favored the new scorecard. CONCLUSIONS: Through adjustment of the LH scorecard after external validation, prognostic performance improved. Further validated, the LH scorecard could be a valuable risk prediction tool.

来源:Ann Thorac Surg,查看原文。版权归原作者所有。

引用

  1. Mejàre-Berggren H, Olsson C. Validation and Adjustment of the Leipzig-Halifax Acute Aortic Dissection Type A Scorecard. Ann Thorac Surg. 2017. PMID: 28751053. doi:10.1016/j.athoracsur.2017.05.010

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。