Pencina Michael J, D'Agostino Ralph B, Zdrojewski Tomasz, Williams Ken, Thanassoulis George, Furberg Curt D, Peterson Eric D, Vasan Ramachandran S, Sniderman Allan D
Duke University, DCRI, Biostatistics and Bioinformatics, Durham, USA Framingham Heart Study, USA
Boston University, Department of Mathematics and Statistics, USA.
Eur J Prev Cardiol. 2015 Oct;22(10):1321-7. doi: 10.1177/2047487315569411. Epub 2015 Jan 29.
Analyses using conventional statistical methodologies have yielded conflicting results as to whether low-density lipoprotein cholesterol (LDL-C) or non-high-density lipoprotein cholesterol (non-HDL-C) or apolipoprotein B (apoB) is the best marker of the apoB-associated risk of coronary heart disease. The aim of this study was to determine the additional value of apoB beyond LDL-C or non-HDL-C as a predictor of coronary heart disease.
For each patient from the Framingham Offspring Cohort aged 40-75 years (n = 2966), we calculated the extent to which the observed apoB differed from the expected apoB based on their LDL-C or non-HDL-C. We added this difference to a Cox model predicting new onset coronary heart disease over a maximum of 20 years adjusting for standard risk factors plus LDL-C or non-HDL. The difference between observed and expected apoB over LDL-C or non-HDL-C was highly prognostic of future coronary heart disease events: adjusted hazard ratios 1.26 (95% confidence interval: 1.15, 1.37) and 1.20 (1.11, 1.29), respectively, for each standard deviation increase beyond expected apoB levels. When this difference between observed and expected apoB was added to standard coronary heart disease prediction models including LDL-C or non-HDL-C, prediction improved significantly (likelihood ratio test p-values <0.0001) and discrimination c-statistics increased from 0.72 to 0.73. The corresponding relative integrated discrimination improvements were 11% and 8%, respectively.
apoB improves risk assessment of future coronary heart disease events over and beyond LDL-C or non-HDL-C, which is consistent with coronary risk being more closely related to the number of atherogenic apoB particles than to the mass of cholesterol within them.
使用传统统计方法进行的分析，对于低密度脂蛋白胆固醇（LDL-C）、非高密度脂蛋白胆固醇（non-HDL-C）或载脂蛋白B（apoB）是否是与载脂蛋白B相关的冠心病风险的最佳标志物，得出了相互矛盾的结果。本研究的目的是确定apoB相对于LDL-C或non-HDL-C作为冠心病预测指标的额外价值。
对于弗雷明汉后代队列中年龄在40 - 75岁的每位患者（n = 2966），我们根据其LDL-C或non-HDL-C计算观察到的apoB与预期apoB的差异程度。我们将此差异添加到一个Cox模型中，该模型预测在最多20年时间内新发性冠心病，对标准风险因素以及LDL-C或non-HDL进行了校正。观察到的apoB与基于LDL-C或non-HDL-C的预期apoB之间的差异对未来冠心病事件具有高度预后价值：每超出预期apoB水平增加一个标准差，校正后的风险比分别为1.26（95%置信区间：1.15, 1.37）和1.20（1.11, 1.29）。当将观察到的与预期的apoB之间的这种差异添加到包括LDL-C或non-HDL-C的标准冠心病预测模型中时，预测显著改善（似然比检验p值<0.0001），鉴别c统计量从0.72增加到0.73。相应的相对综合鉴别改善分别为11%和8%。
apoB在LDL-C或non-HDL-C之外，改善了对未来冠心病事件的风险评估，这与冠心病风险与致动脉粥样硬化的apoB颗粒数量比与其中胆固醇质量更密切相关是一致的。
