Non-HDL Cholesterol: What It Measures, What Counts as High, and What It Means for Athletes
Table of Contents
Key Takeaways
- Non-HDL cholesterol is total cholesterol minus HDL cholesterol. It captures the cholesterol carried by all lipoproteins considered atherogenic, not just LDL.
- In large prospective datasets, non-HDL cholesterol is strongly and stepwise associated with long-term cardiovascular risk, and in statin-treated patients its association with events was slightly stronger than LDL cholesterol’s.
- There is no single “normal” non-HDL cholesterol value: European guideline goals depend on a person’s overall cardiovascular risk.
- Fasting is not routinely needed. Non-HDL cholesterol changes only minimally after ordinary meals.
- Athletes are not automatically protected. Dyslipidaemia is common even in Olympic athletes, lipid profiles differ by sport, and anabolic steroid use was linked to markedly unfavourable lipid changes in a small study.
- In my clinical experience, unfavourable cholesterol profiles are already common among working-age patients, and many decisions are still made from the basic panel of total, HDL and LDL cholesterol.
Introduction: Non-HDL Cholesterol
LDL cholesterol is the number most lipid discussions focus on. Yet one of the most informative numbers on a standard lipid panel requires nothing more than subtracting HDL cholesterol from total cholesterol.
In a risk-modelling study of 398,846 people from 38 cohorts, the 30-year rate of cardiovascular events rose from 7.7% in women with non-HDL cholesterol below 2.6 mmol/L to 33.7% in women at 5.7 mmol/L or higher; in men, the corresponding rates were 12.8% and 43.6% [1].
Cholesterol is one of the most common topics at my occupational health practice. I mainly see working-age patients, and in this age group unfavourable cholesterol profiles are already common in my experience. A pattern I come across often is total cholesterol a little above 5 mmol/L, LDL cholesterol a little above 3 mmol/L and HDL cholesterol around 1 mmol/L or slightly below. Cholesterol measurements are usually part of the periodic health checks I carry out, so I order and interpret a large number of lipid panels.
This article explains what non-HDL cholesterol measures, how guidelines use it, when it tells you more than LDL, and what the research shows in athletes specifically.
What Is Non-HDL Cholesterol?
Non-HDL cholesterol is calculated as total cholesterol minus HDL cholesterol, and a joint European Atherosclerosis Society (EAS) and European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) consensus lists it alongside total cholesterol, triglycerides, HDL and LDL cholesterol as part of the primary lipid panel for estimating atherosclerotic risk [2].
Subtracting HDL leaves the cholesterol carried by the other lipoprotein classes. The authors of one meta-analysis describe non-HDL cholesterol as containing all the lipoproteins considered atherogenic: LDL, lipoprotein(a), intermediate-density lipoprotein (IDL) and very-low-density lipoprotein (VLDL) [3].
That breadth matters because non-HDL cholesterol also picks up remnant cholesterol, the cholesterol carried in triglyceride-rich remnant lipoproteins. The EAS/EFLM panel noted that non-HDL cholesterol includes remnant lipoprotein cholesterol and recommended that it be reported in all standard lipid panels [2].
In my clinical work, LDL cholesterol is usually the most important cholesterol value. Laboratory panels now also offer other values, such as apolipoprotein measurements, and non-HDL cholesterol has recently come up more often both in discussion and on lab reports. In my view it gives information closely related to LDL cholesterol, although it also includes cholesterol carried in other lipoproteins. I also see various LDL/HDL ratios on lab reports, which in my view give related information in a different form. Still, largely because of cost and availability, what I check in most patients is the basic panel of total, HDL and LDL cholesterol, and I proceed based on that.
Why Non-HDL Cholesterol Predicts Heart Risk
The evidence base for non-HDL cholesterol rests on very large datasets.
Population data. The Emerging Risk Factors Collaboration pooled individual records from 302,430 people without vascular disease in 68 prospective studies. Each 1-standard-deviation increase in non-HDL cholesterol (43 mg/dL) was associated with a hazard ratio of 1.50 for coronary heart disease after adjustment for conventional risk factors. These associations were at least as strong in participants who had not fasted as in those who had, and in the subset with direct measurements, the hazard ratio for non-HDL cholesterol (1.42) was similar to that for directly measured LDL cholesterol (1.38) [4]. The authors concluded that lipid assessment could be simplified to total and HDL cholesterol, without fasting [4].
Lifetime risk. In the Multinational Cardiovascular Risk Consortium analysis, which followed participants for a median of 13.5 years and up to 43.6 years, the adjusted hazard ratio for cardiovascular disease at non-HDL cholesterol of 5.7 mmol/L or more, compared with below 2.6 mmol/L, was 1.9 in women and 2.3 in men [1]. The authors also modelled that halving non-HDL cholesterol was associated with lower risk of a cardiovascular event by age 75, with a larger modelled benefit the earlier the reduction occurred [1]. That last point is a modelled estimate, not the result of a randomised trial.
Patients on statins. An individual-patient meta-analysis of 8 statin trials (62,154 patients) compared on-treatment markers among the 38,153 patients allocated to statin therapy. Per standard-deviation increase, adjusted hazard ratios for major cardiovascular events were 1.13 for LDL cholesterol, 1.16 for non-HDL cholesterol and 1.14 for apolipoprotein B (apoB), and the association was significantly stronger for non-HDL cholesterol than for LDL cholesterol or apoB [5].
Non-HDL Cholesterol vs LDL: When the Two Numbers Disagree
It would be an overstatement to call non-HDL cholesterol simply “better” than LDL cholesterol. An EAS/EFLM expert panel judged LDL cholesterol, non-HDL cholesterol and apoB to be comparable predictors of cardiovascular events in population studies and trials, while noting that looking at where the markers disagree improves risk prediction, partly because non-HDL cholesterol adds the remnant-cholesterol component [6]. The same panel stated plainly that there is not yet consistent evidence that treatment targeted at non-HDL cholesterol or apoB prevents more cardiovascular events than treatment targeted at LDL cholesterol [6].
Where non-HDL cholesterol earns its place is in specific situations. The panel recommended it (or apoB) as the secondary treatment target in mild-to-moderate hypertriglyceridaemia, where LDL cholesterol is measured or calculated less accurately and often predicts risk less well [6]. The 2019 ESC/EAS dyslipidaemia guidelines note that in people with elevated triglycerides, diabetes, obesity or very low LDL cholesterol, calculated or measured LDL cholesterol may underestimate risk, and they recommend measuring non-HDL cholesterol and apoB as part of routine lipid analysis in patients with diabetes, high triglycerides or very low LDL cholesterol [7].
Discordance is not a theoretical concern. In a cohort of women, when LDL cholesterol disagreed with alternative measures such as non-HDL cholesterol, apoB or LDL particle number, relying on LDL cholesterol alone under- or overestimated coronary risk by roughly 20–50% even after multivariable adjustment [8].
Normal Non-HDL Cholesterol Levels and Guideline Goals
There is no universal “normal” non-HDL cholesterol. In the 2019 ESC/EAS guidelines, non-HDL cholesterol is a secondary goal and depends on overall cardiovascular risk [7]:
| Risk category | LDL cholesterol goal | Non-HDL cholesterol goal |
|---|---|---|
| Very high | <1.4 mmol/L (55 mg/dL) | <2.2 mmol/L (85 mg/dL) |
| High | <1.8 mmol/L (70 mg/dL) | <2.6 mmol/L (100 mg/dL) |
| Moderate | <2.6 mmol/L (100 mg/dL) | <3.4 mmol/L (130 mg/dL) |
The guidelines set each non-HDL cholesterol goal 0.8 mmol/L (30 mg/dL) above the corresponding LDL cholesterol goal [7]. The 2025 focused update of these guidelines states that the LDL cholesterol goals for each risk category have not changed from 2019 [9].
For non-fasting samples, the EAS/EFLM consensus recommends that laboratory reports flag calculated non-HDL cholesterol of 3.9 mmol/L (150 mg/dL) or higher as abnormal [10]. This laboratory flag is separate from the risk-based treatment goals in the table above.
When a patient’s values are close to the reference limits and there still seems to be room to adjust their diet, what I usually discuss with them is eating less animal fat and more plant fats. In that situation, when overall cardiovascular risk is low, my typical approach has been to allow about six months for lifestyle changes and then check whether the cholesterol profile has improved meaningfully; at higher risk, I consider medication earlier according to guideline-based risk assessment. If it has not, I have often suggested a trial of cholesterol-lowering medication, and in my experience patients quite often agree to start one. Some of them develop muscle or joint pain, which in some patients is very troublesome.
Large double-blind trials suggest that many muscle symptoms reported during statin treatment are not caused by the statin itself. In an individual-participant meta-analysis of large double-blind trials, statin therapy caused a small excess of mostly mild muscle pain, and more than 90% of reports of muscle symptoms among participants allocated a statin were not due to the statin [11]. In the SAMSON trial, among 60 patients who had previously stopped a statin because of side effects, symptom scores did not differ significantly between statin and placebo months, and in their pooled analysis the authors calculated that 90% of the symptom burden elicited by the statin was also elicited by placebo [12]. Given these findings, in my view we often cannot really know why an individual patient develops muscle pain. In my experience, working out the cause is usually quite difficult, as muscle pain can have many different causes. If no explanatory factor is found, we quite often try pausing the statin for a while or switching to a different preparation. With some patients we have had to switch between several drugs, and in some cases no statin has worked, so we have stopped statins and settled for another cholesterol-lowering drug, such as ezetimibe. Biologic cholesterol-lowering drugs are not something I start in my own general practice work; in my experience they have been a specialist-level treatment.
Sometimes, however, the muscle or joint pain clearly affects a patient’s life. I had one patient who could no longer exercise because of the pain and experienced shortness of breath, and whose weight started to rise because they could not keep up exercise and healthy habits. In that case we concluded that we could not continue this way, as the patient was gaining weight and losing their joy of life. We judged it better to go without the statin, arranged strict dietary control with a clinical dietitian, and agreed on follow-up after 3 to 6 months to review the situation. That was an individual decision based on this patient’s overall situation.
With many patients I go through the FINRISK calculator, which predicts 10-year risk of coronary heart disease, stroke and their combination, based on follow-up of Finnish population cohorts [13]. In my experience it gives patients something concrete: they can see their own statistical risk of coronary heart disease or stroke. Some patients still decline medication while knowing their risk. In my view their decision should be respected, as long as they know their own values.
Do You Need to Fast Before a Non-HDL Cholesterol Test?
Usually not. The EAS/EFLM consensus recommended non-fasting lipid profiles for routine use, reporting that the maximal average change in calculated non-HDL cholesterol after habitual meals is about −0.2 mmol/L (8 mg/dL), that HDL cholesterol, apoB and lipoprotein(a) are not affected by fasting status, and that fasting and non-fasting values predict cardiovascular disease comparably [10]. The panel suggested considering a fasting sample when non-fasting triglycerides exceed 5 mmol/L (440 mg/dL) [10].
Our own laboratory offers both fasting and non-fasting lipid tests. Most of the occupational health contracts I work with, however, cover a fasting package rather than a non-fasting one, even though fasting is not necessarily required before the test. With many patients I check cholesterol values once a year and proceed accordingly. When cholesterol is clearly elevated and we are trying an intervention, I may check it more often in some patients.
Non-HDL Cholesterol in Athletes
Regular physical activity can improve the lipid profile, yet athletes can still have dyslipidaemia [14].
Dyslipidaemia is common, even at the Olympic level. In a cohort of 957 Olympic athletes not taking lipid-lowering drugs or supplements, 343 (35.8%) were classified as dyslipidaemic using the authors’ study-specific definition of LDL cholesterol of 115 mg/dL or more or a newly proposed LDL/HDL ratio cut-off of 1.90 or more [15]. The authors noted that non-HDL cholesterol has been proposed to improve the predictive value of the lipid profile, but that athlete-specific targets remain undefined [15].
Lipid profiles differ by sport type. In a study of 45 Turkish male strength athletes, 20 endurance athletes and 19 non-athletes, strength athletes had higher LDL cholesterol and apoB and lower HDL cholesterol than endurance athletes. Athletes as a whole had lower triglycerides and VLDL cholesterol but higher LDL cholesterol than non-athletes, and waist-to-height ratio and body-fat percentage consistently predicted adverse lipid results [14].
Diet pattern and lipid differences. Competitive male ultra-endurance runners habitually eating a very low-carbohydrate, high-fat diet had total, LDL and HDL cholesterol that were 65%, 83% and 60% higher, respectively, than runners on a high-carbohydrate diet (10 runners per group) [16]. The authors suggested the expanded cholesterol pool might serve a functional purpose in these athletes [16]; that is their hypothesis, and the study did not measure cardiovascular outcomes.
Anabolic steroids. In strength athletes who self-administered several anabolic-androgenic steroids for 8 or 14 weeks, HDL cholesterol fell from 1.08 to 0.43 mmol/L and apoB rose from 0.96 to 1.32 g/L, while total cholesterol and triglycerides did not change significantly; six weeks after stopping, HDL cholesterol and apoB had still not returned to baseline [17]. That study reported apoB rather than non-HDL cholesterol; the authors described the changes as an increased atherogenic lipid profile, even though total cholesterol did not change significantly [17]. In a separate randomised, placebo-controlled part of the same study, 8 weeks of nandrolone decanoate did not significantly affect total cholesterol, triglycerides, HDL cholesterol or apoB, although lipoprotein(a) decreased in the nandrolone group [17].
How Training Affects Non-HDL Cholesterol
In the trials below, exercise effects on lipids were modest and differed by training type:
- Aerobic training and HDL. A meta-analysis of 25 articles on randomised controlled trials found that aerobic training raised HDL cholesterol by a statistically significant but modest 2.53 mg/dL (0.065 mmol/L); exercise duration per session was the most important element [18].
- Aerobic training and non-HDL cholesterol. In a meta-analysis of randomised trials, in which the authors calculated non-HDL changes from reported total and HDL cholesterol, aerobic exercise alone produced no statistically significant change in non-HDL cholesterol, whereas diet combined with aerobic exercise reduced it by 11.1 mg/dL compared with controls; differences between the intervention groups themselves were not statistically significant [3].
- Resistance training. Across 29 randomised trials with 1,329 adults, progressive resistance training reduced non-HDL cholesterol by 8.7 mg/dL, equivalent to 5.6% [19].
These trials were conducted in general adult populations rather than trained athletes, so the results may not transfer directly to people who already train heavily.
Patients often ask me what else they can do besides eating less animal fat and more plant fat. In my experience, exercise has improved the lipid profile of many of my patients, so it is a natural part of those conversations. I have also found that exercise usually goes well together with lifestyle change: patients tend to be more motivated to change their habits and to exercise at all when their diet is also on track.
Practical Points
- Read the whole panel. An EAS/EFLM expert panel recommends that laboratories report non-HDL cholesterol in all standard lipid panels [6].
- Interpret against overall risk. Goals depend on cardiovascular risk category, not on a single population cut-off [7].
- Watch for disagreement with LDL. When non-HDL and LDL cholesterol disagree, the difference can reflect the remnant-cholesterol component that only non-HDL cholesterol captures [6].
- Athletes are not exempt. In the studies above, strength athletes (compared with endurance athletes) and runners on a very low-carbohydrate, high-fat diet (compared with high-carbohydrate runners) had higher LDL cholesterol, and anabolic steroid users had lower HDL cholesterol and higher apoB than at their own baseline [14][16][17].
- Cholesterol is one part of the risk. The FINRISK calculator, for example, combines total and HDL cholesterol with smoking, blood pressure, diabetes and family history [13].
Conclusion: Non-HDL Cholesterol Is the Number Worth Not Skipping
Non-HDL cholesterol is simple to calculate, needs no fasting in most cases and captures the cholesterol in all lipoproteins considered atherogenic in one number. In large studies it tracked long-term cardiovascular risk in a clear stepwise pattern [1], and in statin-treated patients its association with events was slightly stronger than LDL cholesterol’s [5]. It does not replace LDL cholesterol, but it may be particularly useful when triglycerides are high or the two numbers disagree.
For athletes, the takeaway is that fitness and lipids are related but separate questions. A favourable training history does not guarantee a favourable non-HDL cholesterol, and in the studies above athletes’ lipid profiles varied widely by sport, diet and drug use.
This matches what I see in practice: LDL cholesterol remains the main value I work with, while non-HDL cholesterol and other measures appear on lab reports more and more often. In my view, going through a patient’s overall risk, for example with the FINRISK calculator, makes the numbers concrete, and a patient’s informed decision about medication deserves respect.
References
- https://doi.org/10.1016/S0140-6736(19)32519-X
- https://doi.org/10.1016/j.atherosclerosis.2019.12.005
- https://doi.org/10.1155/2012/840935
- https://doi.org/10.1001/jama.2009.1619
- https://doi.org/10.1001/jama.2012.366
- https://doi.org/10.1373/clinchem.2018.287037
- https://doi.org/10.1093/eurheartj/ehz455
- https://pubmed.ncbi.nlm.nih.gov/24345402/
- https://doi.org/10.1093/eurheartj/ehaf190
- https://doi.org/10.1093/eurheartj/ehw152
- https://doi.org/10.1016/S0140-6736(22)01545-8
- https://doi.org/10.1056/NEJMc2031173
- https://doi.org/10.1016/j.gheart.2016.04.007
- https://doi.org/10.7717/peerj.20743
- https://doi.org/10.3390/jcm12237449
- https://doi.org/10.1136/bmjsem-2018-000429
- https://doi.org/10.1136/bjsm.2003.000199
- https://doi.org/10.1001/archinte.167.10.999
- https://doi.org/10.1016/j.ypmed.2008.10.010
