Marrugat Jaume, Vila Joan, Baena-Díez José Miguel, Grau María, Sala Joan, Ramos Rafel, Subirana Isaac, Fitó Montserrat, Elosua Roberto
Grupo de Epidemiología y Genética Cardiovascular, Instituto Municipal de Investigación Médica, Barcelona, España.
Rev Esp Cardiol. 2011 May;64(5):385-94. doi: 10.1016/j.recesp.2010.12.011. Epub 2011 Apr 8.
Cardiovascular risk screening requires accurate risk functions. The relative validity of the Framingham-based REGICOR adapted function is analyzed and the population distribution of cardiovascular 10-year cardiovascular events is described by risk group.
A population cohort of 3856 participants recruited between 1995 and 2000, aged 35 to 74 years from Girona without symptoms of cardiovascular diseases, was followed between 2006 and 2009. Standardized laboratory and blood pressure measurements, questionnaires, and case definitions were used. The follow-up combined cross-linkage of our databases with our regional mortality registry, reexamination, and telephone contact with participants. Coronary disease endpoints alone were considered.
A total of 27 487 person-years were obtained (mean follow-up 7.1 years), and the follow-up was achieved in 97% of participants (120 coronary disease events). Validity was good: the regression coefficients estimated with the cohort data did not differ from those obtained in the original Framingham function. Function calibration was good: the observed incidence of cardiovascular events in the decile groups of risk did not differ from the function prediction (P=.127 in women, and P=.054 in men). The C statistic (discrimination) was 0.82 (95% confidence interval, 0.76-0.88) in women, and 0.78 (95% confidence interval, 0.73-0.83) in men. More than 50% of cardiovascular events occurred in participants whose 10-year risk was 5% to 14.9%.
The studied function accurately predicts coronary disease events at 10 years. Risk stratification could be simplified in 4 groups: low (<5%), moderate (5%-9.9%), high (10%-14.9%) and very high (≥15%).
心血管风险筛查需要准确的风险函数。分析了基于弗雷明汉的REGICOR适配函数的相对有效性，并按风险组描述了心血管10年心血管事件的人群分布。
对1995年至2000年间招募的3856名年龄在35至74岁之间、来自赫罗纳且无心血管疾病症状的参与者组成的人群队列，在2006年至2009年期间进行随访。使用了标准化的实验室和血压测量、问卷以及病例定义。随访结合了我们数据库与区域死亡率登记处的交叉链接、重新检查以及与参与者的电话联系。仅考虑冠心病终点。
共获得27487人年（平均随访7.1年），97%的参与者完成了随访（120例冠心病事件）。有效性良好：用队列数据估计的回归系数与原始弗雷明汉函数获得的系数无差异。函数校准良好：风险十分位数组中心血管事件的观察发病率与函数预测无差异（女性P = 0.127，男性P = 0.054）。女性的C统计量（区分度）为0.82（95%置信区间，0.76 - 0.88），男性为0.78（95%置信区间，0.73 - 0.83）。超过50%的心血管事件发生在10年风险为5%至14.9%的参与者中。
所研究的函数能准确预测10年冠心病事件。风险分层可简化为4组：低（<5%）、中（5% - 9.9%）、高（10% - 14.9%）和非常高（≥15%）。
