文献速递颈动脉与弓上血管

基于VQI构建并验证的无症状颈动脉狭窄CEA术后30天卒中与1年生存预测模型

Development of a validated model to predict 30-day stroke and 1-year survival after carotid endarterectomy for asymptomatic stenosis using the Vascular Quality Initiative.
PubMed 文献 · J Vasc Surg · 2017-06-02

研究利用VQI 2010-2015年无症状颈动脉狭窄首次择期CEA患者数据,构建并验证30天卒中与1年死亡风险预测模型。死亡模型纳入24,086例、10个因子(AUC 0.764,95%CI 0.72-0.80)。各中心结局差异明显,模型可辅助筛选最可能获益的患者。

要点

  • 30天卒中率0.9%、1年死亡率3.4%,各VQI中心间差异显著。
  • 1年死亡模型含10个因子,外部验证AUC 0.764(95%CI 0.72-0.80)。

核心结论

  • 无症状狭窄CEA术后30天卒中与1年死亡率在VQI各中心间存在明显差异。
  • 已开发术前风险模型,可较准确估计围术期卒中与1年死亡风险,辅助筛选最可能获益的患者。

研究设计与样本

  • 基于血管质量倡议(VQI)2010-2015年无症状颈动脉狭窄首次择期CEA患者的多变量logistic回归建模研究。
  • 卒中分析纳入31,939例(2010-2015),死亡分析纳入24,086例(2010-2014)。
  • 卒中模型采用内部验证,1年生存模型采用拆分样本验证。

主要发现

  • 30天卒中率为0.9%,1年死亡率为3.4%。
  • 265家VQI中心间卒中率范围0%-8.3%,1年死亡率范围0%-20%。
  • 1年死亡模型纳入10个显著因子,AUC 0.764,95%CI 0.72-0.80。

关键数值与推荐

  • 1年死亡风险预测模型:10个显著因子,AUC 0.764(95%CI 0.72-0.80)。

临床怎么用

  • 术前可用该模型量化无症状狭窄患者CEA的30天卒中与1年死亡风险。
  • 结合中心结局差异,优先选择风险低、预期生存足够获益的患者行CEA。

局限性

  • 卒中模型区分度一般(内部验证校正AUC 0.652)。
  • 研究基于VQI登记数据,结论外推需谨慎。

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

OBJECTIVE: Carotid endarterectomy (CEA) has been shown to be an effective treatment for patients with asymptomatic carotid artery stenosis when perioperative stroke rates are low and patients survive long enough to benefit from the intervention. Our objective was to develop and to validate a simple risk prediction model for 30-day stroke and 1-year mortality to guide optimal selection of patients for CEA. METHODS: Asymptomatic patients undergoing first-time elective CEA within the Vascular Quality Initiative (VQI) from 2010 to 2015 were selected. Outcome measures included any 30-day postoperative stroke and 1-year mortality. Patient demographics, comorbidities, carotid artery disease burden, and provider characteristics were evaluated to select a parsimonious clinical model for risk prediction using multivariable logistic regression. Internal validation was performed for stroke and split sample validation was done for 1-year survival to ensure generalizability. RESULTS: We identified 31,939 patients for inclusion in the stroke analysis (2010-2015) and 24,086 patients for the mortality analysis (2010-2014). Both the 30-day stroke rate (0.9%) and 1-year mortality rate (3.4%) varied substantially across 265 VQI centers (range, 0%-8.3% and 0%-20%, respectively). Eleven significant factors were selected for the 30-day stroke risk prediction model (area under the receiver operating characteristic curve [AUC], 0.67). Internal validation demonstrated good discrimination (bias corrected AUC = 0.652; calibration intercept and slope of 0.03 and 1.01, respectively). Similarly, 10 significant factors were selected for the 1-year mortality risk prediction model (AUC, 0.764). External validation demonstrated excellent discrimination and calibration (AUC, 0.764; 95% confidence interval, 0.72-0.80). CONCLUSIONS: Stroke and 1-year mortality rates after CEA for asymptomatic stenosis vary across VQI centers. We have developed a preoperative risk model that can be used to accurately estimate risk of perioperative stroke and 1-year mortality and to assist providers in selecting patients with asymptomatic stenosis who are most likely to benefit from CEA.

来源:J Vasc Surg,查看原文。版权归原作者所有。

引用

  1. DeMartino RR, Brooke BS, Neal D, et al. Development of a validated model to predict 30-day stroke and 1-year survival after carotid endarterectomy for asymptomatic stenosis using the Vascular Quality Initiative. J Vasc Surg. 2017. PMID: 28583737. doi:10.1016/j.jvs.2017.03.427

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。