Ovbiagele Bruce, Goldstein Larry B, Amarenco Pierre, Messig Michael, Sillesen Henrik, Callahan Alfred, Hennerici Michael G, Zivin Justin, Welch K Michael A
Department of Neurosciences, Medical University of South Carolina, Charleston, South Carolina.
Department of Medicine, Division of Neurology, Duke University, Durham, North Carolina.
J Stroke Cerebrovasc Dis. 2014 Apr;23(4):778-84. doi: 10.1016/j.jstrokecerebrovasdis.2013.12.001. Epub 2014 Feb 24.
Identifying patients with recent stroke or transient ischemic attack (TIA) at high risk of major vascular events (MVEs; stroke, myocardial infarction, or vascular death) may help optimize the intensity of secondary preventive interventions. We evaluated the relationships between the baseline Framingham Coronary Risk Score (FCRS) and a novel risk prediction model and with the occurrence of MVEs after stroke or TIA in subjects enrolled in the Stroke Prevention by Aggressive Reduction in Cholesterol Level (SPARCL) trial.
Data from the 4731 subjects enrolled in the SPARCL study were analyzed. Hazard ratios (HRs) from Cox regression models were used to determine the risk of subsequent MVEs based on the FCRS predicting 20% or more 10-year coronary heart disease risk. The novel risk model was derived based on multivariable modeling with backward selection. Model discrimination (c-statistics) was assessed using the areas under the receiver operating characteristic curves.
Of 3969 subjects with complete data, 27% had a baseline FCRS of 20% or more. In multivariable analysis, an FCRS of 20% or more was associated with twice the risk of subsequent MVEs (HR = 1.92, 95% confidence interval [CI]: 1.63-2.27). The novel model based on a multivariable analysis included age (HR = 1.37, 95% CI: 1.25-1.51 per 10 years), diabetes (HR = 1.82, 95% CI: 1.51-2.18), male sex (HR = 1.35, 95% CI: 1.12-1.61), and an apolipoprotein (APO)-B/APO-A1 ratio (HR = 1.56, 95% CI: 1.16-2.11). The c-statistic was .58 (95% CI: .55-.60) for the FCRS of 20% or more and .65 (95% CI: .63-.67) for the novel model.
Both a baseline FCRS of 20% or more and a novel predictive model were associated with future MVEs in SPARCL trial subjects. The novel model needs to be validated, and the benefits of using either the FCRS or the novel model in clinical practice needs to be assessed.
识别近期发生中风或短暂性脑缺血发作（TIA）且有发生重大血管事件（MVE；中风、心肌梗死或血管性死亡）高风险的患者，可能有助于优化二级预防干预的强度。我们在强化降低胆固醇水平预防中风（SPARCL）试验中，评估了基线弗明汉姆冠心病风险评分（FCRS）与一种新型风险预测模型之间的关系，以及它们与中风或TIA后MVE发生情况的关系。
对SPARCL研究中纳入的4731名受试者的数据进行分析。基于预测10年冠心病风险达20%或更高的FCRS，采用Cox回归模型的风险比（HR）来确定后续MVE的风险。新型风险模型是通过多变量建模和向后选择得出的。使用受试者工作特征曲线下的面积评估模型辨别力（c统计量）。
在3969名有完整数据的受试者中，27%的基线FCRS为20%或更高。在多变量分析中，FCRS为20%或更高与后续MVE风险增加一倍相关（HR = 1.92，95%置信区间[CI]：1.63 - 2.27）。基于多变量分析的新型模型包括年龄（每10年HR = 1.37，95% CI：1.25 - 1.51）、糖尿病（HR = 1.82，95% CI：1.51 - 2.18）、男性（HR = 1.35，95% CI：1.12 - 1.61）和载脂蛋白（APO）-B/APO-A1比值（HR = 1.56，95% CI：1.16 - 2.11）。FCRS为20%或更高时的c统计量为0.58（95% CI：0.55 - 0.60），新型模型的c统计量为0.65（95% CI：0.63 - 0.67）。
在SPARCL试验受试者中，基线FCRS为20%或更高以及新型预测模型均与未来MVE相关。新型模型需要进行验证，并且需要评估在临床实践中使用FCRS或新型模型的益处。
