Roxane de La Harpe, Pedro Marques-Vidal, Julien Vaucher
Division of Internal Medicine, Department of Medicine, Lausanne University Hospital and University of Lausanne, Rue du Bugnon 46, 1005 Lausanne, Switzerland.
Division of Internal Medicine, Department of Medicine and specialties, Fribourg Hospital and University of Fribourg, Fribourg, Switzerland.
Eur J Prev Cardiol. 2025 May 30. doi: 10.1093/eurjpc/zwaf213.
External validation of the new 10-year PREVENT risk score for atherosclerotic cardiovascular disease (ASCVD) is important to assess its potential clinical applicability in Switzerland and to highlight its influence in preventive treatment eligibility.
This study, which was not used in the development process of PREVENT, included 5064 individuals from a prospective Swiss cohort, aged 40 or older, without pre-existing ASCVD, and with complete data for risk score calculation. Main outcomes were adjudicated ASCVD events, including fatal and non-fatal myocardial infarction and strokes. The performances of the PREVENT score were assessed overall, and stratified by gender and age groups (<70 vs. ≥70 years), and compared with SCORE2 and the Pooled Cohort Equation (PCE) scores. Among 4356 participants followed from 2009 to 2012 over a median of 9 years, 224 experienced a first incident of ASCVD. The PREVENT cardiovascular risk prediction model demonstrated adequate discrimination performance, correctly identifying 76% of concordant pairs [C-Index, 95% Confidence Interval (CI) 0.73 to 0.79]. The model's calibration performances suggest systematic underestimation (Observed/Expected ratio 1.45, 95% CI 1.44-1.46), especially in women and those under 70 years old, yet it maintained positive clinical utility across all subgroups, particularly at the 7.5% threshold, which is the lower limit of the intermediate-risk category in clinical practice. However, PREVENT did not improve predictive performance when compared with SCORE2 and PCE.
Our study confirmed the PREVENT model demonstrated adequate discrimination and calibration capabilities, along with significant clinical utility, particularly at intermediate-risk thresholds. However, it did not outperform the established models, SCORE2 or PCE. Additionally, PREVENT may systematically underestimate risk, which could raise concerns about the underprescription of preventive treatments.
对新的10年动脉粥样硬化性心血管疾病（ASCVD）预防风险评分进行外部验证，对于评估其在瑞士的潜在临床适用性以及突出其在预防性治疗资格方面的影响具有重要意义。
本研究未用于预防风险评分的开发过程，纳入了来自瑞士一个前瞻性队列的5064名40岁及以上、无既往ASCVD且有完整风险评分计算数据的个体。主要结局为经判定的ASCVD事件，包括致命和非致命性心肌梗死及中风。对预防评分的性能进行总体评估，并按性别和年龄组（<70岁与≥70岁）分层，与SCORE2和合并队列方程（PCE）评分进行比较。在2009年至2012年随访的4356名参与者中，中位随访9年，224人发生了首次ASCVD事件。预防心血管风险预测模型显示出足够的区分性能，正确识别了76%的一致对[C指数，95%置信区间（CI）0.73至0.79]。该模型的校准性能表明存在系统性低估（观察值/预期值比率为1.45，95%CI为1.44 - 1.46），尤其是在女性和70岁以下人群中，但在所有亚组中仍保持积极的临床效用，特别是在7.5%阈值时，这是临床实践中中度风险类别下限。然而，与SCORE2和PCE相比，预防评分并未提高预测性能。
我们的研究证实，预防模型显示出足够的区分和校准能力，以及显著的临床效用，特别是在中度风险阈值时。然而，它并未优于已建立的模型SCORE2或PCE。此外，预防评分可能会系统性低估风险，这可能引发对预防性治疗处方不足的担忧。
