For decades, a high HDL result was the one line on a cholesterol panel a patient could feel good about. Everything else on the lipid panel came with a lower-is-better instruction. HDL was the exception: the number doctors wanted to see climb, the "good cholesterol" that seemed to clean up after the bad kind. That story held for a long time. It has not held cleanly for the last twenty years.
What HDL actually does
Cholesterol cannot travel through blood on its own, so the body packages it into lipoprotein particles that ferry it where it needs to go. Cleveland Clinic describes these particles as being like buses, with cholesterol as the passengers. LDL carries cholesterol out to tissues, where excess amounts can lodge in artery walls and start the plaque buildup behind heart attacks and strokes. HDL runs the opposite route, picking up excess cholesterol from the blood and delivering it back to the liver for breakdown and removal, a process called reverse cholesterol transport. Cleveland Clinic notes HDL particles also appear to fight inflammation and help prevent blood clots, on top of that cleanup job.
That mechanism is why a standard lipid panel treats HDL differently from every other number on it. MedlinePlus lists an optimal HDL result as 60 mg/dL or higher, and Cleveland Clinic's reference ranges call anything below 40 mg/dL in men or 50 mg/dL in women low, with 40 to 80 mg/dL (men) and 50 to 80 mg/dL (women) considered normal. Mayo Clinic adds a ceiling most people never hear about: HDL levels above 100 mg/dL are considered extreme, and some research suggests very high natural HDL may carry its own risk rather than none at all.
The theory that launched a drug
The idea that higher HDL is protective goes back to the Framingham Heart Study. A widely cited 1988 report on 2,748 Framingham participants found that, after twelve years of follow-up, people in the lowest fifth of HDL results had a higher risk of dying from coronary heart disease than people in the highest fifth, in both men and women, as summarized in a Journal of Thoracic Disease commentary on the evidence's evolution. That finding, and others like it, rested on an assumption the field did not yet know was flawed: that the relationship between HDL and death risk was a straight line, where more was simply, uniformly better.
That assumption was strong enough to launch a drug. If low HDL caused heart disease, raising it pharmacologically should prevent it. Pfizer's torcetrapib was designed to do exactly that, by blocking a protein called CETP that normally moves cholesterol out of HDL particles, and it raised HDL by 60 to 100 percent in trials, an effect no other drug had matched. Pfizer built a 15,000-patient trial to see whether that dramatic rise actually prevented heart attacks and strokes. In December 2006, the trial was stopped early: people taking torcetrapib alongside a statin were dying at a higher rate than people on the statin alone, 82 deaths versus 51 in the primary safety analysis reported by NPR, with full results later published in the New England Journal of Medicine showing significantly higher cardiovascular and non-cardiovascular mortality in the treatment group. Steven Nissen, then president of the American College of Cardiology, called it a huge setback for a strategy the field had treated as nearly settled.
Torcetrapib also raised blood pressure independent of its effect on HDL, and much of the field initially treated that as the likely explanation for the deaths, separate from the HDL-raising mechanism itself. But it was not an isolated failure: other CETP inhibitors and niacin, tested in later trials, also raised HDL substantially without producing the cardiovascular benefit the epidemiology had predicted, a pattern summarized in the Journal of Thoracic Disease's review of the field's unraveling assumptions.
What genetics revealed about function versus quantity
A 2016 study led by Daniel Rader at the University of Pennsylvania, published in Science and summarized by NIH Research Matters, offered a mechanistic clue for why raising the number alone had not worked. Researchers sequencing genes near known lipid-related regions in people with very high or very low HDL cholesterol found a rare variant in SCARB1, the gene for the main HDL receptor on liver cells. Carriers of this variant, called P376L, had unusually high HDL, exactly what the old theory would predict as protective. Instead, across nearly 50,000 people with coronary heart disease and roughly 88,000 controls, carriers had a significantly higher risk of heart disease. In cell and mouse experiments, the mutant receptor was not processed correctly and failed to reach the liver cell surface, so the liver could not take up cholesterol from HDL even though blood HDL numbers looked excellent. Rader summarized the finding directly: the protective effects of HDL depend more on how it functions than on how much of it is present.
What newer population data adds
If HDL's protective reputation rested partly on studies with limited diversity, expanding that picture mattered. A 2022 study supported by the National Institutes of Health, using data from nearly 24,000 Black and white adults in the REGARDS study and published in the Journal of the American College of Cardiology, found that low HDL predicted a higher risk of heart attack and related death in white adults, a long-accepted association, but not in Black adults, and higher HDL was not associated with reduced cardiovascular risk in either group. Senior author Nathalie Pamir, in the NIH's summary of the findings, said the results suggest the field needs to revisit how HDL is used in risk prediction, since a high number no longer guarantees the pat on the back it once did.
A separate, stranger thread comes from the extreme high end of the range. A 2017 European Heart Journal study of more than 116,000 people from two Danish cohorts, led by Christian Madsen and colleagues at the University of Copenhagen, found the relationship between HDL and all-cause mortality was U-shaped rather than a straight line: both very low and very high HDL were associated with higher death rates. Men with extremely high HDL had a mortality rate 106 percent higher than men with typical HDL, and women with extremely high HDL had a rate 68 percent higher, per the University of Copenhagen's summary. The concentration tied to the lowest mortality risk sat well within the normal range, not at its upper edge, and later reviews note this U-shaped pattern has since been reproduced in other large cohorts, including studies of United States veterans and a Korean population.
Reading your own number
None of this erases HDL from a useful lipid panel. Low HDL still tracks with real cardiovascular risk, and the behaviors known to raise it, regular exercise, not smoking, and maintaining a healthy weight, are the same behaviors that improve heart health for other reasons entirely. What has changed is the confidence with which a high number, especially a very high or medication-driven one, can be read as protection. A result comfortably inside the 40 to 80 mg/dL (men) or 50 to 80 mg/dL (women) range Cleveland Clinic describes as normal is a reasonable target. A result well above 80, and especially above 100, is worth a conversation with a clinician rather than a celebration, since current evidence cannot say that particular kind of "good" number is doing any good at all.
Common questions
Is a higher HDL always better?
Not necessarily. Within the normal range, higher has generally tracked with lower cardiovascular risk. But Mayo Clinic notes HDL above 100 mg/dL is considered extreme, and several large cohort studies have found death rates rise again at the very top of the range, not just at the bottom.
Why did the torcetrapib trial fail if HDL is supposed to be protective?
The drug raised HDL by 60 to 100 percent but also raised blood pressure and was linked to significantly more deaths, per the results published in the New England Journal of Medicine. Forcing the number up through one specific mechanism did not deliver the benefit correlated with naturally high HDL.
Can I raise my HDL with medication?
Cleveland Clinic notes some statins raise HDL somewhat as a secondary effect, but providers do not typically prescribe medication solely to raise HDL, given how the major HDL-specific drug trials have turned out.
Does smoking really lower HDL?
Yes. MedlinePlus lists tobacco use as a factor that lowers HDL, and quitting is one of the few interventions with consistent evidence for improving it.
Why do HDL targets differ between men and women?
Cleveland Clinic notes HDL is the only lipid panel number with sex-based cutoffs, with women needing a higher level (50 mg/dL) than men (40 mg/dL) to be considered acceptable, reflecting typical population differences between the sexes.
Should I ask for a more detailed HDL test?
Some researchers are exploring measures of HDL function, such as how well it removes cholesterol from cells, rather than concentration alone. These are not yet standard clinical tests, so a routine lipid panel's HDL number remains the practical starting point.
