Department of Vascular Medicine, University Medical Centre Utrecht, The Netherlands.
Eur J Prev Cardiol. 2012 Dec;19(6):1486-95. doi: 10.1177/1741826711426636. Epub 2011 Oct 18.
Although the overall average 10-year cardiovascular risk for patients with manifest atherosclerosis is considered to be more than 20%, actual risk for individual patients ranges from much lower to much higher. We investigated whether information on metabolic syndrome (MetS) or its individual components improves cardiovascular risk stratification in these patients.
We conducted a prospective cohort study in 3679 patients with clinical manifest atherosclerosis from the Secondary Manifestations of ARTerial disease (SMART) study. Primary outcome was defined as any cardiovascular event (cardiovascular death, ischemic stroke or myocardial infarction). Three pre-specified prediction models were derived, all including information on established MetS components. The association between outcome and predictors was quantified using a Cox proportional hazard analysis. Model performance was assessed using global goodness-of-fit fit (χ(2)), discrimination (C-index) and ability to improve risk stratification.
A total of 417 cardiovascular events occurred among 3679 patients with 15,102 person-years of follow-up (median follow-up 3.7 years, range 1.6-6.4 years). Compared to a model with age and gender only, all MetS-based models performed slightly better in terms of global model fit (χ(2)) but not C-index. The Net Reclassification Index associated with the addition of MetS (yes/no), the dichotomous MetS-components or the continuous MetS-components on top of age and gender was 2.1% (p = 0.29), 2.3% (p = 0.31) and 7.5% (p = 0.01), respectively.
Prediction models incorporating age, gender and MetS can discriminate between patients with clinical manifest atherosclerosis at the highest vascular risk and those at lower risk. The addition of MetS components to a model with age and gender correctly reclassifies only a small proportion of patients into higher- and lower-risk categories. The clinical utility of a prediction model with MetS is therefore limited.
尽管患有明显动脉粥样硬化的患者的总体 10 年心血管风险被认为超过 20%，但个体患者的实际风险范围从低得多到高得多。我们研究了代谢综合征（MetS）或其各个组成部分是否可以改善这些患者的心血管风险分层。
我们对来自二次动脉粥样硬化表现研究（SMART）的 3679 例有临床明显动脉粥样硬化的患者进行了前瞻性队列研究。主要结局定义为任何心血管事件（心血管死亡、缺血性卒中和心肌梗死）。衍生了三个预先指定的预测模型，均包含有关既定 MetS 成分的信息。使用 Cox 比例风险分析量化了结局与预测因素之间的关联。使用整体拟合优度（χ²）、区分度（C 指数）和改善风险分层的能力来评估模型性能。
在 3679 例患者中，共发生了 417 例心血管事件，随访时间为 15102 人年（中位数随访时间为 3.7 年，范围为 1.6-6.4 年）。与仅基于年龄和性别的模型相比，所有基于 MetS 的模型在整体模型拟合度（χ²）方面表现稍好，但 C 指数无差异。与年龄和性别相比，添加 MetS（是/否）、二分类 MetS 成分或连续 MetS 成分对分类的净重新分类指数分别为 2.1%（p=0.29）、2.3%（p=0.31）和 7.5%（p=0.01）。
纳入年龄、性别和 MetS 的预测模型可以区分处于最高血管风险和处于较低风险的有临床明显动脉粥样硬化的患者。将 MetS 成分添加到年龄和性别模型中，仅正确地将一小部分患者重新分类为高风险和低风险类别。因此，具有 MetS 的预测模型的临床实用性有限。
