HDL's U-Shape: Why 'More Good Cholesterol' May Not Mean a Sharper Brain
A new analysis of more than 7,500 older adults suggests very high HDL — long sold as the 'good' cholesterol — may track with lower cognitive scores. Here's what tired parents need to know about their own labs (and their parents' labs).
If you've been running on three hours of sleep and a half-cold coffee, the last thing you want is another headline rewriting what counts as healthy. Fair. But this one is worth a minute, because it nudges a story most of us absorbed years ago and never updated: that HDL — the so-called "good" cholesterol — is something you simply want more of. A new longitudinal analysis from China suggests the picture is more nuanced, particularly for the grandparents helping you survive the toddler years.
The short version: in a study of 7,509 older adults, the relationship between HDL cholesterol and cognitive function looked like a U-shape, not a straight line up. Middle-range HDL tracked with the best cognitive scores. Very low and very high HDL both tracked with worse ones.
Key takeaways
- In a 7,509-person Chinese cohort of older adults, HDL cholesterol and cognitive scores showed a U-shaped relationship, with a turning point around 67.43 mg/dL.
- Below that threshold, higher HDL was modestly associated with better cognitive scores; above it, higher HDL was associated with lower scores.
- This is an observational finding — it shows a pattern, not a cause — and the effect sizes are small.
- The takeaway isn't to chase a number. It's to treat lipid panels as one signal among many, and to talk with a clinician rather than reorganize anyone's diet around HDL alone.
- The "higher is always better" framing for HDL is increasingly out of step with the research, including emerging cardiovascular signals around ultra-high HDL.
What the study actually found
The analysis drew on the China Health and Retirement Longitudinal Study (CHARLS), using a 2011 baseline of 7,509 older adults. Researchers measured HDL cholesterol and scored cognition using a tool that captured episodic memory and mental status, then ran the kind of statistics designed to catch curves rather than straight lines — multiple linear regression, restricted cubic splines, and threshold effect analysis. The published results in PLOS One describe a clearly nonlinear pattern.
Compared with the lowest HDL group (under 35 mg/dL), the two highest groups had lower cognitive scores — Q4 with a β of −0.622 and Q5 with a β of −0.322. The spline analysis pinpointed a turning point at 67.43 mg/dL. Below that, a one-standard-deviation increase in HDL was linked to roughly a 0.08-SD bump in cognitive score. Above it, the same increase was linked to a 0.07-SD decline.
Those are small numbers. They're statistically meaningful in a sample this size, but no one is going to notice a 0.07-SD swing in how Grandma does the crossword. What's notable is the shape: the curve bends back down at the high end, instead of continuing to rise the way the old consumer-health story would predict.
Why "more good cholesterol" got complicated
For decades, HDL was cast as the cleanup crew of the bloodstream — ferrying cholesterol back to the liver, associated in observational studies with lower heart-disease risk. From there it was a short leap to "more is better," a leap that powered a generation of food marketing and well-meaning advice from in-laws.
The trouble is that biology rarely cooperates with tidy stories. Drugs designed to push HDL up did not deliver the cardiovascular benefit researchers expected. And in recent years, large cohort analyses have flagged that very high HDL — well above population norms — appears to travel with its own risks, not just protection. This new cognitive-health analysis is part of that broader rethink: HDL isn't a score you want to maximize; it's a marker that behaves differently at the extremes than in the middle.
Mechanism-wise, researchers are still working it out. One line of thinking is that at very high concentrations, HDL particles may become dysfunctional — less efficient at the jobs that made them look protective in the first place. Another is that very high HDL can reflect underlying conditions, genetics, or medications that themselves influence brain and vascular health. The honest answer for now is: we see the curve, we're not sure why it bends.
What this means for tired parents (and the grandparents in the group chat)
A few grounded points, since the temptation with any new study is either to ignore it or to overhaul dinner.
It's about older adults, not you in the baby years
This was a cohort of older Chinese adults, and the outcome was cognitive function, not toddler-era brain fog. If you're the one trying to remember whether you already started the dishwasher, this study isn't really about your HDL. It's much more relevant when you're looking at a parent's lab results.
Numbers are signals, not scoreboards
If a relative's HDL is on the low side, or unusually high, that's information to bring to a clinician — alongside blood pressure, blood sugar, triglycerides, family history, and what's actually going on in daily life. The U-shape is a reminder that a single lipid number is a poor proxy for "health."
Don't reverse-engineer your diet around HDL
There is no good evidence that nudging HDL up or down with food choices alone will protect cognition, and certainly none in this study. The smallest useful step is usually the dull one: regular movement, sleep when you can claw it back, blood-pressure checks, and the standard cardiometabolic basics that benefit the brain whether or not the HDL needle moves.
Be skeptical of supplements promising HDL boosts
The "raise your good cholesterol" pitch is everywhere. Given how the science is moving, that framing is increasingly outdated — and there's no reason to assume a higher HDL produced by a supplement carries any cognitive benefit at all.
The honest caveats
A few things worth holding in mind. This is observational data: the analysis can show that very high HDL and lower cognitive scores cluster together, but it can't prove that HDL is doing anything to the brain. People with very high HDL may differ in other ways — genetics, medications, alcohol intake, underlying illness — that themselves shape cognition.
It's also a single population at a single time point in one country, with cognition measured by a screening tool rather than a clinical diagnosis of dementia. The effect sizes are modest. And nothing here changes the standing advice for people with genuinely low HDL or established cardiovascular risk — that conversation still belongs with a clinician.
What the study does do, quietly and well, is chip away at a slogan. "Good cholesterol, the more the better" was always more catchphrase than science, and the curve in this data is one more reason to retire it. For most of us, the practical move isn't to track HDL more obsessively. It's to track it less reverently — and to let the people who interpret labs for a living tell us what a given number actually means.