几十年来，胆固醇风险评估一直围绕低密度脂蛋白胆固醇（LDL-C）来建立：LDL更低通常与更少心肌梗塞、中风和早死相关。 但LDL-C检测试量的是在LDL颗粒中携带的胆固醇质量，而非颗粒数量，因此两个人可以有相同的LDL值却有不同颗粒数量和不同风险。2026年3月，AHA与ACC表示apoB可能更精确；对约250,000名符合他汀治疗条件的美国成年人进行的一项JAMA建模研究发现，按apoB指导决策可预防比现行做法更多心肌梗塞和中风，并且仍然具有成本效益。尽管如此，其落地实施仍然滞后。

缓慢采用主要是行为性和历史性因素。LDL是一项50年的突破，也是非常可用的公共卫生叙事：它简单、广为理解，并与他汀的使用紧密相关。Sniderman指出，患者和医生对apoB仍知之甚少。在接受他汀治疗的人群中，Nordestgaard及其同事报告说，高apoB和非HDL胆固醇仍与更高的心肌梗塞和死亡风险相关，而LDL往往并不相关。Kausik Ray等人认为，apoB是更好的总体信号，因为它统计所有含apoB的动脉粥样硬化颗粒，但高apoB可能反映非常不同的成因，如高LDL、胰岛素抵抗、肥胖或遗传因素，这些都需要不同的治疗。

这种细分正推动更广泛的检测面板：Ray和Nordestgaard都建议加入脂蛋白(a)与残余胆固醇。数据在英国背景下尤其突出：在英国，少于5%的成年人接受了脂蛋白(a)检测，尽管一次终身测量就足够。风险在二十岁、三十岁、四十岁时就已累积，但筛查常在更晚阶段启动，因此照护仍偏向被动。新出现的代谢组与遗传模型可在标准胆固醇检测之上改进预测，并解释具有相似指标者结局差异，但也会增加成本、分析负担和不确定的治疗规则。该领域正在转向分层和连续风险评估，apoB可能居于核心，但并非最终答案。

For decades, cholesterol-risk assessment has been built around low-density lipoprotein cholesterol (LDL-C): lower LDL has been linked to fewer heart attacks, strokes, and premature deaths. But LDL-C testing measures the cholesterol mass carried in LDL particles, not particle count, so two people can share the same LDL value yet have different particle numbers and different risk. In March 2026, the AHA and ACC said apoB is likely more precise, and a JAMA modeling study of about 250,000 U.S. adults eligible for statins found apoB-guided decisions would prevent more heart attacks and strokes than current practice while remaining cost-effective. Yet implementation has lagged.

Slow adoption is largely behavioral and historical. LDL has been a 50-year breakthrough and a very usable public-health story: it is simple, widely understood, and tightly linked to statins. Sniderman notes that patients and doctors still know little about apoB. In statin-treated groups, Nordestgaard and colleagues reported that high apoB and non-HDL cholesterol still track higher heart-attack and mortality risk, while LDL often does not. Kausik Ray and others argue that apoB is a better overall signal because it counts all apoB-containing atherogenic particles, but high apoB may reflect very different causes such as high LDL, insulin resistance, obesity, or genetics, which require different therapies.

This nuance is driving broader panels: Ray and Nordestgaard recommend adding lipoprotein(a) and remnant cholesterol. The data are striking in the UK context, where fewer than 5 percent of people are tested for lipoprotein(a), despite one lifetime measurement being sufficient. Risk accumulates in twenties, thirties, and forties, yet screening often starts much later, so care stays reactive. Emerging metabolic-and-genetic models can improve prediction beyond standard cholesterol tests and help explain outcome differences among people with similar profiles, but they also add cost, analytic burden, and uncertain treatment rules. The field is moving to layered and continuous risk assessment, with apoB likely central but not the final answer.