Can a Normal LDL Hide Higher Heart Attack Risk?

A normal LDL cholesterol is not always the all-clear it feels like. ApoB measures the number of artery-clogging particles in your blood, while LDL cholesterol measures the cholesterol packed inside LDL particles. Usually those two things move together, so LDL-C does a decent job. But in a meaningful share of people they part ways, and when normal or low LDL-C comes with high ApoB, the higher observed risk tends to track ApoB more closely. So if your LDL-C looks fine but you have reason to think your risk is higher, ApoB can add information the standard lipid panel may miss.

Why two people with the same LDL can differ

Every LDL, IDL, VLDL, and Lp(a) particle carries one ApoB molecule. Count ApoB and you have a close count of those particles. LDL cholesterol instead weighs cholesterol cargo, and cargo per particle varies. Some people carry many cholesterol-poor particles, often small dense LDL particles, each holding less cholesterol. Their LDL-C can come back normal because there is not much cholesterol to measure, while the particle count remains high. In discordance studies, the observed risk tends to follow the higher particle count.

The two markers disagree more often than a single number implies, and how often depends heavily on how discordance is defined and how LDL-C is calculated. Martin/Hopkins and Sampson estimates of LDL-C show less ApoB discordance than the older Friedewald equation. The disagreement is more common in a recognizable group: insulin resistance, type 2 diabetes, metabolic syndrome, extra weight around the middle, and high triglycerides can all shift people toward many small particles, though a high ApoB cannot be reliably predicted from those traits alone. In these patients the LDL-C number can understate particle burden. In discordant cases, particle number often maps risk better than LDL-C.

What happens when the numbers disagree

The cleanest way to see this is to compare people in the same LDL-C range and split them by ApoB. In a UK Biobank analysis of 293,876 adults with no cardiovascular disease at baseline, people with LDL-C around 130 mg/dL had about 7 cardiovascular events per 100 people over ten years when ApoB sat at the high end of its range for that LDL-C, versus about 4 per 100 when ApoB sat at the low end. Same LDL-C range, materially different risk. That gap is invisible on LDL-C alone.

The pattern holds in younger people and earlier disease. In the CARDIA study, young adults, average age 25, whose ApoB ran high while their LDL-C ran low had about 55 percent higher odds of coronary calcium in midlife, independent of the usual risk factors. Coronary calcium is subclinical atherosclerosis, not a heart attack, but it is a marker of future event risk. In a separate statin-treated Copenhagen cohort, people with low LDL-C but high ApoB had about 50 percent higher myocardial infarction risk than people low on both. And in a clinical chemistry analysis of people with very low calculated LDL-C, 55 to 70 mg/dL, 40 percent had discordantly higher ApoB by that study's definition. The point is not that LDL-C is useless. It is that particle burden can remain after LDL-C looks reassuring.

Why the big studies look underwhelming

There is a real counterweight. In the largest pooled analyses, the Emerging Risk Factors Collaboration and a large UK Biobank study led by Welsh found that adding apolipoproteins to models already containing standard cholesterol measures produced very little extra prediction. That sounds like it kills the case. It does not. It explains why Sniderman's older meta-analysis, Marston's separate MI analysis, and the discordance cohorts can favor ApoB while average prediction metrics still look small. In the UK Biobank, ApoB and LDL-C were tightly correlated, so once you know one, the other adds little on average. Non-HDL-C, which comes free on a standard panel, carries much of the same particle-linked information, and a 2026 Copenhagen analysis found discordantly high non-HDL-C and discordantly high ApoB carried similar myocardial infarction risk. Average is the operative word. Those studies answer a population question. The discordance studies answer the individual question: what if your LDL-C and particle count do not match? A test that changes little for most people can still matter for a person with discordance.

One caveat cuts the other way: because ApoB counts every particle equally, it can understate risk when Lp(a) is high. A 2024 genetic analysis estimated Lp(a) particles were roughly six times more atherogenic than LDL particles on a per-particle basis, though estimates ranged from about four to twelve times across datasets. This came from a Mendelian randomization analysis, not a trial. A normal ApoB is not a complete all-clear if you have never checked Lp(a). The 2026 ACC/AHA dyslipidemia guideline recommends at least one Lp(a) measurement in adulthood.

Who this actually changes something for

No major U.S. guideline makes ApoB mandatory for every adult with a normal LDL-C. Where guidance has converged is on the targeted case. The 2024 National Lipid Association consensus and the 2026 ACC/AHA dyslipidemia guideline support ApoB measurement when the standard panel is likely to miss risk: elevated triglycerides, diabetes, metabolic syndrome or CKM syndrome, obesity, discordant lipid profiles, or low achieved LDL-C on treatment. When ApoB and LDL-C disagree, the NLA says ASCVD risk generally aligns better with ApoB or non-HDL-C. It also says ApoB can be used as an additional treatment target, while noting that thresholds are less well established than LDL-C thresholds and that judging therapy by ApoB still requires clinical judgment.

The thresholds are worth knowing because guidelines do not agree. The older ACC/AHA ApoB cutoff of 130 mg/dL was a high-risk flag near the 90th percentile of the untreated population, not the point at which risk first begins. The NLA's 2024 suggested treatment thresholds are lower: under 90 mg/dL for borderline-to-intermediate risk, under 70 mg/dL for high risk, and under 60 mg/dL for very high risk. These were derived by regression from LDL-C targets rather than from ApoB-targeted trials, but if you are using ApoB to optimize prevention rather than merely clear a high-risk flag, those lower thresholds are more relevant than 130. ApoB is an available blood test that can usually be measured without fasting. It is most worth checking if your LDL-C looks fine but you carry metabolic risk factors, high triglycerides, a strong family history, or other signs your risk is higher than your cholesterol suggests.

A high ApoB with a normal LDL-C is not a diagnosis and not an automatic prescription. It is a reason to take overall cardiovascular risk more seriously than the LDL-C number alone invited, and to use a fuller prevention discussion, especially if medication decisions are on the table. What would sharpen this further is a completed trial showing that treating to ApoB in people with normal LDL-C prevents more heart attacks than treating to LDL-C or non-HDL-C. No completed head-to-head outcomes trial has answered that yet, and the discordance evidence is mainly observational or post hoc. Even so, when LDL-C and ApoB split, risk has more often aligned with particle count.