Low HDL Cholesterol: Causes, Is It Dangerous and What to Do

HDL is reported most often in mg/dL, and in scientific papers in mmol/L; the conversion factor is roughly 38.7 (so 40 mg/dL is 1.0 mmol/L, and 50 mg/dL is about 1.3 mmol/L). The most commonly used cut-offs for 'low' HDL, below 40 mg/dL in men and below 50 mg/dL in women, come from the criteria for metabolic syndrome and serve to pick out people at risk, not to make a diagnosis. Why do women have a higher cut-off? Estrogens raise HDL, so average values are higher in women.

A single result is not decisive. HDL concentration changes over time and depends on recent illness, weight change, exercise and the laboratory method, so with a worrying result it is worth repeating the test after a few weeks under standard conditions. A large pooled analysis (the Emerging Risk Factors Collaboration) also showed that a lipid panel taken without fasting has at least as strong an association with risk as one taken fasting, so strict fasting requirements matter less for HDL itself than for triglycerides and calculated LDL.

Remember that HDL-C measures the amount of cholesterol in HDL particles, not their number or function. This is an important distinction we return to: two people with the same HDL-C can have very different particles. We describe in detail how to read the whole lipid panel and how HDL relates to LDL and triglycerides in our article on reading a lipid panel and in the knowledge-base entry.

The most common cause of low HDL in the general population is not a disorder of HDL itself but a metabolic state in which HDL is the 'victim'. In insulin resistance, abdominal obesity and type 2 diabetes, the liver produces an excess of triglyceride-rich lipoproteins (VLDL). These particles exchange their fat with HDL particles in a process dependent on the protein CETP: HDL loses cholesterol and gains triglycerides. Such 'triglyceride-enriched' HDL is cleared from the blood faster, so its concentration falls.

The practical consequence: low HDL and high triglycerides are one 'family' of disorders, and improving one usually goes hand in hand with improving the other. When you see low HDL together with raised triglycerides, raised fasting glucose, a wide waist or hypertension, that is the classic picture of metabolic syndrome and insulin resistance (we describe them in separate entries). Then metabolism, not HDL, is the right subject for the conversation with your doctor.

Besides insulin resistance, several well-known factors are listed. Tobacco smoking lowers HDL and quitting raises it: a review of 45 studies (Forey et al.) found that after quitting smoking HDL-C rises quickly, by a weighted mean of about 0.06 mmol/L (about 2 mg/dL), with a significant increase seen in all analyzed subgroups. Low physical activity and excess body weight also lower HDL, and a diet very rich in refined carbohydrates promotes high triglycerides and, through them, low HDL.

Genetic disorders of HDL metabolism are rare, but worth knowing about. A particularly low HDL (in the range of a few to a dozen mg/dL) together with other unusual features, such as corneal clouding, an enlarged liver or spleen or sensory disturbances, is a reason for evaluation at a lipid center. With a typical HDL of around 30-40 mg/dL and a clear metabolic background, such causes are unlikely.

This question concerns many men on testosterone replacement therapy and people who use steroids to build muscle. The answer depends on the dose. In a controlled study using single doses of testosterone enanthate in 25 healthy men, doses of 250 mg and 500 mg lowered apolipoprotein A1 and HDL, while the lowest dose studied (125 mg) did not change the lipid profile (Gårevik et al. 2014). This indicates that high, supraphysiological doses, typical of steroid abuse, can clearly lower HDL.

A reasonable conclusion: in men with a real testosterone deficiency treated with replacement doses, the effect on HDL is usually small, whereas high anabolic doses can lower it clearly. We write in detail about the relationship between testosterone and the lipid panel in the knowledge-base entry on testosterone and cholesterol. If you notice a marked fall in HDL on therapy, more important than HDL itself is checking ApoB, LDL and hematocrit and discussing the dose with your doctor.

This is strong observational evidence that low HDL is a marker of higher risk. Observational, however, means it does not settle causality: people with low HDL have on average more insulin resistance and higher triglycerides, are more likely to smoke and move less, and each of those factors on its own increases risk. The key question is therefore not 'does low HDL go together with heart disease' (it does) but 'does raising HDL reduce risk'.

This result is consistent with what large clinical trials showed. In AIM-HIGH (3414 patients with atherosclerotic disease and low HDL, treated intensively with a statin), adding extended-release niacin raised median HDL from 35 to 42 mg/dL and lowered triglycerides, but did not reduce cardiovascular events (HR 1.02; 95% CI 0.87-1.21), and the trial was stopped after a mean of 3 years because of lack of efficacy.

Not all drugs that raise HDL failed in the same way, which shows that the matter is not black and white. In REVEAL (30,449 patients on intensive atorvastatin), the CETP inhibitor anacetrapib raised HDL-C by 43 mg/dL (by 104%) and at the same time lowered non-HDL-C by 17 mg/dL (by 18%), and the risk of major coronary events fell by 9% (rate ratio 0.91; 95% CI 0.85-0.97). The result does not, however, allow the conclusion that raising HDL was what helped: the drug simultaneously reduced the number of ApoB-carrying particles, and the trial was not designed to separate these two effects. Longer follow-up showed a 12% reduction in coronary events over the whole period and no significant adverse effects on non-vascular mortality.

This result suggests that the function of HDL particles (their ability to carry cholesterol away from tissues) may be a better indicator of risk than the HDL-C concentration alone. Cholesterol efflux is not measured routinely, however, and the study does not prove that improving HDL function lowers risk. The second surprise is the U-shaped curve: in large Danish cohorts (more than 116,000 people) both very low and extremely high HDL-C were associated with higher mortality. The lowest mortality was seen at about 73 mg/dL (1.9 mmol/L) in men and 93 mg/dL in women, and at HDL-C of 116 mg/dL or more in men the adjusted hazard of death from any cause was 2.06 (Madsen et al. 2017). The authors cautioned that the results need confirmation.

A practical lesson follows from these data: there is no point striving for the highest possible HDL-C. The goal is healthier metabolism, not a record number. And when HDL is low, remember that it is the result of a single parameter, best interpreted together with LDL, ApoB, triglycerides, glycemia and blood pressure.

The modest size of these effects (a few mg/dL) is worth treating as an honest point of reference: no lifestyle change will raise HDL from 30 to 70 mg/dL. That does not mean it is not worth changing. Exercise, quitting smoking and weight reduction simultaneously improve insulin sensitivity, triglycerides, blood pressure and overall risk, and their benefit far exceeds the small change in HDL-C itself.

Likewise, there is no evidence that supplements marketed as 'raising good cholesterol' (for example some herbal blends) improve prognosis. Even if a preparation does increase HDL-C by a few mg/dL, we know from the trials above that a change in the number alone is not equivalent to benefit. Instead of looking for a supplement, it pays more to address the cause: metabolism, exercise and smoking.

A reasonable set of tests for low HDL is: a full lipid panel (LDL, triglycerides), ApoB or non-HDL, fasting glucose and HbA1c (to assess insulin resistance), blood pressure and waist circumference, and where needed TSH (hypothyroidism can cause lipid disturbances) and liver and kidney tests. For a person without symptoms, what matters is establishing what is behind the low HDL, not the result itself.

Low HDL is best treated like a warning light on the dashboard, not like engine failure. It signals that it is worth checking metabolism, exercise, smoking and the rest of the lipid panel, but raising the HDL number itself is neither a proven goal nor a proven route to lower risk. The best thing you can do is address the causes and control what has been shown to be causal: LDL and ApoB, blood pressure, glycemia and smoking.