The study by Philips and Davey Smith¹ from 1991 addressed the subject of inferring causality from observational epidemiology in situations where exposures are highly correlated and measurement variability is not similar for the variables. The authors exemplified this problem with triglycerides and high-density lipoprotein HDL cholesterol in the risk for cardiovascular disease, with inversely associated concentrations of the two biomarkers and with a higher variability in concentrations of triglycerides than in HDL cholesterol. The association of high concentrations of triglycerides with risk for cardiovascular disease is attenuated when adjusted for HDL cholesterol; however, when adjusting the association of low HDL cholesterol with risk for cardiovascular disease for triglyceride concentrations, the association is not attenuated to the same extent.¹ This has previously often led to the interpretation that HDL cholesterol is more important than triglycerides for causing cardiovascular disease. In consequence, huge efforts went into developing and testing HDL cholesterol-raising drugs, which surprisingly did not prevent cardiovascular disease.²

中文翻译：

---

评论：甘油三酸酯或HDL胆固醇是心血管疾病的真正元凶？

飞利浦和戴维·史密斯的研究¹1991年的研究解决了在暴露高度相关且变量的测量变异性不相似的情况下从观察流行病学推断因果关系的问题。作者举例说明了甘油三酸酯和高密度脂蛋白HDL胆固醇在罹患心血管疾病中的风险，这两种生物标志物的浓度成反比，并且甘油三酸酯的浓度比HDL胆固醇具有更高的可变性。调整高密度脂蛋白胆固醇后，高浓度甘油三酸酯与心血管疾病风险的相关性减弱。但是，当调整低HDL胆固醇与甘油三酯浓度引起的心血管疾病风险的相关性时，该相关性不会减弱到相同的程度。^1个以前，这常常导致人们认为HDL胆固醇在引起心血管疾病方面比甘油三酸酯更为重要。结果，在开发和测试HDL胆固醇升高药物上付出了巨大的努力，令人惊讶的是，这种药物并不能预防心血管疾病。^2个