High HDL Cholesterol Is Not a Free Pass: What the apoB Evidence Actually Shows

A reader writes in with a familiar story: years of HDL cholesterol above 100 mg/dL, the number every prior doctor praised, followed by a coronary artery disease diagnosis that no one saw coming. That contradiction is not a fluke. The phrase "good cholesterol" has done real damage, because it tells people that a high HDL-C reading buys safety. The evidence from the last fifteen years says otherwise, and it points to a different number worth your attention after 40: apolipoprotein B, or apoB.

This is not a claim that HDL is bad. It is a claim that HDL-C, the cholesterol carried inside HDL particles, is a weak and sometimes misleading guide to your cardiovascular risk. Here is what is established, what is still emerging, and how to act on the difference.

Established evidence: raising HDL cholesterol has failed in trial after trial

The "HDL is protective" idea came from observational data: people with higher HDL-C tend to have fewer heart attacks. The honest test of that idea is to raise HDL-C on purpose and see whether events fall. They did not.

Three drug classes reliably raise HDL-C: niacin, fibrates, and CETP inhibitors. A 2014 meta-analysis in the BMJ pooled 39 trials and 117,411 patients already on statins and found that none of these HDL-raising treatments reduced all-cause mortality, coronary death, heart attack, or stroke. The CETP inhibitor torcetrapib (the ILLUMINATE trial) raised HDL-C sharply and increased deaths. Dalcetrapib (dal-OUTCOMES) raised HDL-C and changed nothing in patients after acute coronary syndrome.

The genetics agree. When researchers compared variants in the CETP and HMGCR genes that lower LDL-C and apoB by discordant amounts, the drop in cardiovascular events tracked the drop in apoB, not the change in HDL-C. In plain terms: moving HDL-C around does not move risk. The atherogenic particles, the ones apoB counts, do.

There is a second, less comfortable finding. Very high HDL-C is not extra credit. In a Chinese cohort of 3.3 million adults, all-cause and cardiovascular mortality followed a U-shaped curve, with the lowest mortality at HDL-C of roughly 50 to 79 mg/dL and higher risk at the extremes. A Korean cohort (KoGES-HEXA) found that non-diabetic men in the highest HDL-C group carried a 31% higher risk of all-cause death. So the reader with HDL-C over 100 was not protected by it; in some datasets that range carries its own signal.

What apoB actually measures, and why it wins

Every atherogenic particle that drives plaque (LDL, VLDL, remnants, Lp(a)) carries exactly one apoB protein. Measure apoB and you have counted those particles directly, regardless of how much cholesterol each one happens to be carrying. LDL-C estimates the cholesterol mass inside LDL, which is a different thing and can disagree with particle number.

That disagreement is the whole point. Two people can share an identical LDL-C of 100 mg/dL while one has many small cholesterol-poor particles and a high apoB, and the other has fewer large particles and a lower apoB. The first person is at higher risk, and only apoB shows it. A 2025 expert consensus from the National Lipid Association concluded that apoB is a more accurate risk marker than LDL-C or non-HDL-C, echoing the physiological case laid out in the Journal of the American Heart Association in 2022. This is established enough that apoB now appears in major guidelines, not as a fringe test.

Emerging evidence: a one-time shot at lifelong low apoB

The next chapter is about driving apoB down further and more durably than a daily pill. PCSK9 is a protein that degrades the liver's LDL receptors; block it and the liver clears more apoB-containing particles. Antibody PCSK9 inhibitors already do this with injections every few weeks.

The emerging work is gene editing. In 2025, Verve Therapeutics published Phase 1b results in the New England Journal of Medicine for VERVE-102, a single-infusion base editor that switches off PCSK9 in the liver. LDL-C fell up to 62% at the highest dose, with an absolute drop of 78 mg/dL, and the trial reported no treatment-related serious adverse events in the early cohort. This is early, small, and not yet proven to prevent heart attacks. But the direction is a one-time intervention that lowers apoB for years. Treat it as a development to watch, not a decision to make.

A practical method for using these numbers after 40

1. Ask for an apoB measurement, not just a standard lipid panel. It is an inexpensive, standardized blood test, and it does not require a 12-hour fast in the way LDL-C calculation traditionally did.
2. Read your HDL-C in context, not as a score. Roughly 40 to 80 mg/dL is unremarkable. Do not interpret a high value as a credit against the rest of the panel.
3. If apoB is unavailable, use non-HDL-C (total cholesterol minus HDL-C) as the next-best proxy. It counts every atherogenic particle's cholesterol and ignores HDL entirely.
4. Track the same marker over time rather than comparing one test to another. Trend matters more than a single snapshot.
5. Bring the apoB number, plus Lp(a) measured once in your life, to a clinician and discuss whether your particle count warrants action. The decision to treat is theirs and yours together, not a website's.

Mistakes to avoid

• Treating HDL-C as a target. There is no validated way to "raise good cholesterol" into lower risk; the trials that tried failed.
• Reading HDL-C above 60 as a protective bonus. The mortality curve is U-shaped, not a straight line down.
• Leaning on the total-cholesterol-to-HDL ratio for reassurance. A flattering ratio can hide a high apoB.
• Assuming the HDL bump from alcohol or exercise adds cardiovascular protection. The HDL-C number rising does not mean risk fell.
• Ignoring apoB because LDL-C "looks fine." Discordance between the two is exactly the situation apoB exists to catch.

Observable markers to follow

• apoB in mg/dL, watched as a trend (many lipidologists treat values in the 60 to 80 range as desirable for higher-risk adults; your clinician sets your target).
• Non-HDL-C as the fallback when apoB is unavailable.
• Lp(a) measured once, since it is largely genetic and apoB-bearing.
• LDL-C still reported, but cross-checked against apoB rather than trusted alone.

Personal experimentation

If you are curious, the low-cost, low-risk experiment is informational, not pharmacological: pay for a single apoB test alongside your next standard panel and see whether the two numbers tell the same story. If your LDL-C looks reassuring but your apoB sits higher than expected, you have learned something a "good cholesterol" reading would never have told you. Bring that discordance to your physician. Nothing here is a treatment recommendation; it is a better question to walk in with.

The lesson from the reader with HDL-C over 100 is not that HDL is the enemy. It is that one comforting number was never the right one to watch. Count the particles that build plaque, follow that count over years, and make decisions with a clinician from there.

Sources