Recently, I got an email from a 70-year-old reader of Outlive that I haven’t been able to stop thinking about. It read, in part:
“I don’t think it is an understatement to say ‘you saved my life.’ Everything in my lipid panel was fine except my high density lipoprotein cholesterol (HDL-C), which has been over 100 mg/dL for the past five years. No one seems to really know what that means, but it kept me pushing for answers and for the next thing I perceived as critical on your list—the CT angiogram. I think the cardiologist was convinced my request was nonsense, since I had no chest pain or shortness of breath or high blood pressure or any of the other classic symptoms of heart disease. In the end, I think he agreed to give me a CT angiogram just to shut me up… I think you can guess where this is going. Both the doctor and I were surprised to see that my CTA showed 50% stenosis of the left anterior descending coronary artery. I’m scheduled for a stress echocardiogram next month and am now on lipid lowering medications.”
Stories like this are part of why I do what I do. She pushed for answers when something didn’t add up, and it saved her life. But this email also surfaced something I hear constantly, from patients and clinicians alike: the belief that high HDL-C is some kind of free pass from cardiovascular disease.
For decades, HDL-C has been labeled “good cholesterol,” and that label has created a deeply rooted—but misleading—assumption that having a lot of it means you’re protected from atherosclerosis. In this reader’s case, her very high HDL-C was incorrectly interpreted as reassurance. And had she not pushed for a deeper look, the consequences could have been serious.
Why we started calling it “good cholesterol”
To understand the “good cholesterol” trap, it helps to go back to the beginning. The label didn’t come out of nowhere. Early epidemiological data consistently showed that people with higher HDL-C had lower rates of cardiovascular disease, while people with lower HDL-C tended to have higher rates. The association was real, it was reproducible, and it showed up across large populations.
The biology seemed to back it up: HDL particles perform genuinely important protective functions. The classic example is reverse cholesterol transport (RCT)—the process by which HDL particles shuttle cholesterol out of peripheral tissues and back to the liver, where it can be excreted. (We’ve covered RCT in more depth in this article and on The Drive.) If HDL is part of the system removing cholesterol from places it shouldn’t accumulate, then more HDL—and by extension, more HDL-C—should mean less atherosclerosis. That was the logic.
So when researchers in the 1970s and 1980s needed a way to communicate these findings, “good cholesterol” was an easy shorthand. At the population level, the directional association still holds: low HDL-C tracks with higher cardiovascular risk, high HDL-C with lower risk. The problem is that this shorthand was such a gross oversimplification of the underlying biology that it has done more harm than good.
Our research team spends hundreds of hours each month vetting studies and distilling dense literature to deliver evidence-informed insights on health and longevity. If you find value in our work, consider becoming a premium member and supporting our mission.
Where the label falls apart
There are several problems with calling HDL-C “good cholesterol,” and they compound on each other.
Start with the phrase itself. The cholesterol molecule is essential for life: it’s a structural component of every cell membrane, a precursor to vitamin D, steroid hormones, and bile acids. The cholesterol inside an HDL particle is chemically identical to the cholesterol inside an LDL particle. Calling one “good” and the other “bad” implies there are different kinds of cholesterol in the body, when in fact what matters is where cholesterol ends up and how much of it gets there.
The deeper problem is that the HDL-C reported on your blood tests doesn’t tell you anything about HDL function. HDL-C is a measurement of the total mass of cholesterol being carried inside all the HDL particles in a given volume of plasma, typically reported in mg/dL. It tells you little about how many HDL particles there are, what those particles are made of, or how well they’re doing their job. HDL biology is genuinely important—but the number on a standard lipid panel is a crude snapshot that misses most of what actually matters. HDL particles carry out dozens of distinct functions, and while researchers can assess some of these in the lab, no clinical-grade test for HDL functionality is currently available to practicing physicians. We’ve blurred the line between HDL function and HDL-C, treating an easily measured number as a stand-in for the biology we actually care about.
Then there’s the correlation-causation problem. Low HDL-C often clusters with insulin resistance, elevated triglycerides, and other markers of metabolic dysfunction—factors that are very well-known to increase cardiovascular risk. High HDL-C, in contrast, tends to travel with a more favorable metabolic profile. HDL-C may simply be reflecting the metabolic context that drives risk, rather than driving the risk itself.
For decades, the assumption ran the other way: if higher HDL-C meant lower cardiovascular risk, then raising HDL-C should lower it further. That assumption has not held up. Niacin, fibrates, and multiple CETP inhibitors—including torcetrapib and others—have all raised HDL-C substantially in large outcome trials without producing the expected reduction in major adverse cardiovascular events. (I’ve discussed the CETP story in more depth in the obicetrapib episode of The Drive.) These failures are a strong signal that HDL-C is not the variable we should be trying to manipulate.
The institutional response has followed the evidence. No major guideline—not the American Heart Association, not the American College of Cardiology, not the National Lipid Association—recommends raising HDL-C as a treatment target or sets a therapeutic goal for it. That’s a striking silence given how often patients are left believing that their high HDL-C is protecting them.
When very high HDL-C is a warning sign
So far, the case has been that high HDL-C may not be protective. There’s a stronger version of this argument: in some patients, very high HDL-C may be a sign that something has gone wrong.
The cleanest evidence comes from human genetics. People with certain variants in SCARB1 have high HDL-C and paradoxically elevated cardiovascular risk.1 Normally, the SCARB1 receptor allows the liver to pull cholesterol out of HDL particles for excretion or utilization. But when genetic variants impair the receptor’s function, HDL can’t offload its cargo. Cholesterol stays stuck in circulation on HDL particles, driving HDL-C up even though the reverse cholesterol transport process—the very function HDL is supposed to perform—has stalled.
The same pattern—high HDL-C tracking with elevated risk—shows up in large observational studies. Several major analyses, including data from the UK Biobank and a meta-analysis of more than a million individuals, have found a U-shaped relationship between HDL-C and mortality.2,3 At low HDL-C, risk is elevated—no surprise there. But risk also rises at very high HDL-C, generally above 80 mg/dL, and the effect appears more pronounced in men. Some of this signal may reflect confounding by factors like alcohol intake, which both raises HDL-C and independently increases mortality, but the U-shape persists in adjusted analyses. Regardless, this isn’t what you’d see if HDL-C were a straightforward “more is better” marker.
In a cohort of patients with very high HDL-C—median 98 mg/dL—roughly half had detectable coronary artery calcium, and its presence predicted all-cause mortality more strongly than traditional risk factors did.4 A population that on paper looked maximally protected was carrying a substantial burden of subclinical atherosclerosis.
The bottom line
None of this means HDL-C is useless. Low HDL-C, especially when it travels with high triglycerides and apoB, is a reasonable biomarker for metabolic dysfunction and elevated cardiovascular risk. It’s just not the whole story, and a high HDL-C does not mean the rest of the picture is fine.
That doesn’t mean every person with high HDL-C should panic or rush out for advanced imaging. It means HDL-C should be interpreted in context—alongside apoB, metabolic markers, blood pressure, family history, and where appropriate, imaging. Risk lives in the full picture, not in any single number.
Maybe the most useful thing we can do is retire the phrase “good cholesterol.” It was a well-intentioned shorthand that captured a real population-level association, but it has outlived its usefulness. It implies a level of certainty the biology doesn’t support, encourages patients and clinicians to treat a single number as a verdict, and can quietly steer people away from the questions they should be asking. High HDL-C is not a free pass. The reader whose story opened this piece may have had a reassuring number on paper, but what changed her trajectory was refusing to let that number end the conversation.
For a list of all previous weekly emails, click here.
References
1. Zanoni P, Khetarpal SA, Larach DB, et al. Rare variant in scavenger receptor BI raises HDL cholesterol and increases risk of coronary heart disease. Science. 2016;351(6278):1166-1171. doi:10.1126/science.aad3517
2. Liu C, Dhindsa D, Almuwaqqat Z, et al. Association between high-density lipoprotein cholesterol levels and adverse cardiovascular outcomes in high-risk populations. JAMA Cardiol. 2022;7(7):672-680. doi:10.1001/jamacardio.2022.0912
3. Chen L, Zhao Y, Wang Z, et al. Very high HDL-C (high-density lipoprotein cholesterol) is associated with increased cardiovascular risk in patients with NSTEMI (non-ST-segment elevation myocardial infarction) undergoing PCI (percutaneous coronary intervention). BMC Cardiovasc Disord. 2023;23(1):357. doi:10.1186/s12872-023-03383-9
4. Razavi AC, Mehta A, Wong ND, et al. Coronary artery calcium for risk stratification among persons with very high HDL cholesterol. JACC Adv. 2024;3(10):101217. doi:10.1016/j.jacadv.2024.101217
