Allan G Michael, Garrison Scott, McCormack James
aEvidence-Based Medicine, Department of Family Medicine, University of Alberta, Alberta bFaculty of Pharmaceutical Sciences, University of British Columbia, Vancouver, British Columbia, Canada.
Curr Opin Lipidol. 2014 Aug;25(4):254-65. doi: 10.1097/MOL.0000000000000095.
The cardiovascular benefit of many preventive interventions (like statins) is strongly dependent on the baseline cardiovascular risk of the patient. Many lipid and vascular primary prevention guidelines advocate for the use of cardiovascular risk calculators.
There are over 100 cardiovascular risk prediction models, and some of these models have spawned scores of calculators. Only about 25 of these models/calculators have been externally validated. The ability to identify who will have events frequently varies little (<5%) between models. However, disagreement between risk calculators is common with one in three paired comparisons disagreeing on risk category. In part, this disagreement is because calculators vary according to the database they are derived from, choice of clinical endpoints and risk interval duration upon which the estimate is based. Additional risk factors do little to improve the basic risk predictions performance, except perhaps coronary artery calcium which still requires further study before regular use.
The estimates provided by cardiovascular risk calculators are ballpark approximations and have a margin of error. Physicians should use models derived from, or calibrated for, populations similar to theirs and understand the endpoints, duration, and special features of their selected calculator.
许多预防性干预措施（如他汀类药物）对心血管系统的益处很大程度上取决于患者的基线心血管风险。许多血脂和血管疾病一级预防指南提倡使用心血管风险计算器。
有超过100种心血管风险预测模型，其中一些模型衍生出了大量的计算器。这些模型/计算器中只有约25种经过了外部验证。不同模型识别出谁会发生心血管事件的能力差异通常很小（<5%）。然而，风险计算器之间存在分歧很常见，三分之一的配对比较在风险类别上存在分歧。部分原因在于，不同的计算器所依据的数据库、临床终点的选择以及估计所基于的风险间隔时长各不相同。除了冠状动脉钙化可能还需要进一步研究才能常规使用外，其他额外的风险因素对改善基本风险预测性能作用不大。
心血管风险计算器提供的估计值只是大致近似值，存在误差范围。医生应使用源自与其患者群体相似的人群或针对该人群进行校准的模型，并了解所选计算器的终点、时长和特殊特征。
