Lipoprotein(a): the inherited risk marker hardly anyone knows

Maybe there is a lab report in front of you. One line is flagged: Lp(a), elevated. Next to it a number and a unit you have never seen before.

Or you have no report, but a question from your family. Someone had a heart attack at a young age, even though their cholesterol levels were supposedly fine. And you wonder whether that has anything to do with you.

Many people know this pattern. The standard values say everything is fine. Your gut says a piece is missing. Sometimes that piece is called lipoprotein(a), or Lp(a) for short.

A lot is changing right now: new guidelines, a dispute between specialist societies and, in September 2026, a trial result that turned out differently than hoped. I will sort out for you what is well supported today and where the limits of what is currently possible lie.

What Lp(a) is and why it sits in your genes

Imagine the LDL particle as a small delivery van. It carries cholesterol through the blood to the cells. Each of these vans carries exactly one large protein on the outside, apolipoprotein B.

Lp(a) is the same delivery van with a trailer. A second protein, apolipoprotein(a), is coupled to its apolipoprotein B. And this protein looks like a chain of repeating links.

The consensus of the European Atherosclerosis Society (EAS) summarized it in 2022 as follows: the concentration ranges from below 0.1 to above 300 mg/dL and is more than 90 percent determined by the LPA gene. The adult level is usually reached by about age five, but can still rise into adulthood. Women are on average about 5 to 10 percent higher.

That is why Lp(a) barely changes with exercise, and diet shifts it only a little. According to the EAS table, physical activity has no or only minimal effect on the level. In the EPIC-Norfolk study, Lp(a) hardly differed between people with very healthy and very unhealthy lifestyle habits.

What can still shift the value

Some conditions move Lp(a) considerably, according to the EAS on a limited data basis: an underactive thyroid can raise it, pregnancy can roughly double it, nephrotic syndrome can raise it three- to fivefold. Several inflammatory diseases are also associated with an increase. In the Copenhagen population study, Lp(a) in women after menopause was 27 percent higher.

Ancestry plays a role too. In the UK Biobank, medians ranged between 16 and 75 nmol/L, depending on the population group. The risk per 50 nmol/L, however, was similar across the groups studied, with hazard ratios between 1.07 and 1.11. Why genes and ancestry shape metabolism so differently is something I describe in Genes, ancestry and metabolism.

And now you know why this value is so stable and why you do not need to blame yourself for it.

From gene to heart attack: why this value matters

How can you tell whether a blood value helps cause a disease or merely accompanies it? People with high Lp(a) could, after all, have something else in common.

This is where Mendelian randomization comes in. Nature already drew the lots at conception. Who gets which gene variant does not depend on whether someone smokes, exercises or eats well. If people with genetically higher Lp(a) have more heart attacks, that strongly suggests a cause.

A genetic study led by Clarke confirmed this: two variants in the LPA gene were associated with more coronary heart disease (odds ratios 1.70 and 1.92), and after accounting for the Lp(a) level the association disappeared. This suggests that the genes raise risk through the level.

An analysis of the UK Biobank led by Björnson estimated Lp(a) to be about six times more atherogenic than LDL per particle (point estimate 6.6, uncertainty range 5.1 to 8.8). My assessment: because there are usually far more LDL particles than Lp(a) particles in the blood, LDL remains the bigger lever for most people.

How large is the risk?

Two numbers belong together here.

The UK Biobank with 460,506 participants showed a linear relationship without a detectable threshold: hazard ratio 1.11 per 50 nmol/L. So there is no biological switch. Cut-off values are decision aids.

For stroke the association is weaker, but present. In the Copenhagen General Population Study, people with Lp(a) above 93 mg/dL had a hazard ratio of 1.60 for ischemic stroke compared with values below 10 mg/dL. The highest absolute ten-year risk, 17 percent, was found in current smokers over 70 with high blood pressure and Lp(a) above 93 mg/dL. Smoking and blood pressure are exactly the factors that can be influenced.

The heart valve hardly anyone talks about

Lp(a) does not only affect the coronary arteries. In a genome-wide analysis led by Thanassoulis, a single gene variant reached the significance threshold for calcification of the aortic valve, and it lay in the LPA gene. Per copy of this variant, the risk of new-onset aortic valve stenosis rose with a hazard ratio of 1.68.

How many people does this affect?

The answer depends on the threshold and on the group studied. One review puts high values at 10 to 20 percent of the population. In a US laboratory population of more than 530,000 people, 24.0 percent were above 50 mg/dL, although this is not a random sample. In the UK Biobank, 12.2 percent of people without and 20.3 percent of people with pre-existing vascular disease had values of 150 nmol/L or more. The often-quoted order of magnitude of about one in five people lies within this corridor. According to the EAS, however, Lp(a) is not a risk factor for venous thrombosis.

And now you know why a single value does not have to be dramatic and can still be taken seriously.

Your value on the lab report: units, cut-offs, costs

You google your value and find three websites with three different cut-offs. One uses a factor of 2, the next 2.5. That is no coincidence, but a genuine measurement problem.

Imagine a train station. One figure gives the total weight of all trains, the other the number of trains. If every train had the same number of carriages, you could convert between them. But the chains vary in length from person to person. The EAS consensus therefore explicitly recommends against a set conversion factor, not least because measurement methods vary widely. A factor of 2 to 2.5 is considered only a rough estimate there.

Why different cut-offs circulate

The cut-offs depend on the guideline and on the unit in which they were formulated:

• EAS consensus 2022: below 30 mg/dL or 75 nmol/L unlikely to add risk, above 50 mg/dL or 125 nmol/L relevant, a grey zone in between.
• ESC/EAS update 2025: above 50 mg/dL, translated there as 105 nmol/L, to be considered clinically significant (IIa, B).
• ACC/AHA guideline 2026: values above 125 nmol/L are associated with a 1.4-fold, values above 250 nmol/L with an at least twofold increased risk.

So the same 50 mg/dL is translated once as 105 and once as 125 nmol/L. On top of that, each laboratory uses its own method. That is why there is deliberately no reference table here. Your value needs to be assessed by a physician together with its unit, the laboratory and your overall risk, and no treatment that you should work out yourself follows from a single number.

Is one measurement enough?

Usually yes, because the value stays very constant within a person over years. A second measurement can make sense if you were tested as a young woman and are now past menopause. The authors of the Copenhagen study consider repeated measurements appropriate in women over 50. The value can also be shifted in kidney disease, an underactive thyroid, during pregnancy or in inflammatory diseases.

What the test costs

Under the German fee schedule for physicians (GOÄ, item 3730), the laboratory service costs EUR 17.49 at the single rate and EUR 20.11 at the 1.15 rate. Blood draw and consultation come on top. Lp(a) is not part of the standard lipid panel. Whether a German statutory health insurer covers the cost depends on the medical justification and can only be clarified case by case.

There is also a risk calculator: the EAS consensus refers to an online algorithm from the European Atherosclerosis Society on risk and benefit in elevated Lp(a). It can help you prepare for a consultation, but it cannot replace one.

How cholesterol, apolipoprotein B and the history of blood lipid markers fit together is something I have described in detail in my article on cholesterol and blood lipid markers. Here it is only about the one value that is usually missing from that story.

And now you know why the same number can take on two different meanings on two websites.

Who should be tested and why the guidelines disagree

Maybe you have heard two statements that do not fit together. Every adult should know their Lp(a) level once. And from the family doctor's practice: we do not measure that routinely. Both statements have a guideline behind them.

The DEGAM, the German College of General Practitioners and Family Physicians, gives careful reasons for its restraint. There is no direct evidence from controlled trials that systematic measurement improves critical outcomes. And prediction in the general population barely improves.

When testing makes particular sense

Regardless of the guideline dispute, there are situations in which the value can particularly often contribute to the assessment. I deliberately do not give firm age limits, because the guidelines define differently what counts as premature.

And now you know why the same test is routine in one practice and not in another, and which situations can tip the balance for you.

High value, what now? Overall risk is the lever

You are holding the result in your hand and wondering what to do now. The honest answer begins with an image.

Imagine a river. Lp(a) is the baseline water level you were born with. You can hardly change it. But tributaries flow into this river: LDL, blood pressure, smoking, blood sugar, lack of exercise. These can often be turned down. And whether the river overflows its banks depends largely on the sum.

This is exactly what the EAS consensus recommends: early, consistent management of the other risk factors, tailored to baseline risk and to the level of Lp(a). The 2025 update explicitly mentions more intensive LDL lowering. The dissenting specialist societies emphasize that elevated Lp(a) remains a risk factor even at very low LDL. In the assessment for the German Heart Foundation (Deutsche Herzstiftung), Laufs of Leipzig University Hospital highlighted the combination with smoking.

Statins and Lp(a): honestly two-sided

According to the available studies, statins do not lower Lp(a). Whether they raise it slightly is the subject of two large analyses with different results. An individual patient data analysis led by Tsimikas found an increase under statins; the ratio of geometric means versus placebo was 1.11. A systematic meta-analysis led by de Boer with 39 studies and 24,448 participants, by contrast, found no clinically meaningful difference; the percentage difference was 0.1 percent.

The EAS consensus draws a clear conclusion from this: statin treatment should not be stopped, because the benefit for the heart and blood vessels in high Lp(a) far outweighs possible risks from moderate Lp(a) increases. When a statin may be indicated, especially at a younger age, I describe in Statins from 35? Why high LDL alone is not a diagnosis. Here I stay with the interplay with Lp(a).

The diet paradox

Now it gets counterintuitive. Anyone who cuts back on saturated fat usually lowers their LDL in studies. Lp(a), however, can rise slightly in the process.

In a small, controlled crossover study led by Tindall with 34 adults, replacing saturated fats with unsaturated fats from walnuts or vegetable oils lowered atherogenic lipoprotein classes, while Lp(a) remained unchanged under all diets. A review led by Enkhmaa concluded that diet influences Lp(a) only moderately and often in the opposite direction to LDL.

For me, this does not lead to an Lp(a) diet or to a promise that you can lower the value through food. It only leads to this: never look at one value in isolation. Lowering LDL is proven to benefit the heart and blood vessels; the small Lp(a) shifts caused by diet are not. Why the story of fats is more complicated than its reputation anyway is covered in Does fat make you fat?.

One tributary that is often underestimated is metabolism in the liver. What role sugar and fructose play in liver metabolism I describe in Sugar and fructose: the liver is where it happens. Another is fitness: VO2max and life expectancy is about why aerobic fitness is so closely linked to life expectancy. Both are tributaries you can influence, even if the baseline level stays.

And now you know why an unchanged Lp(a) level is no sign that your efforts are in vain.

Medications and procedures: the honest status in September 2026

Maybe you have read headlines saying that injections against Lp(a) are coming soon. Here is where things stand, block by block, with the level of evidence.

PCSK9 inhibitors

These medications are prescribed to lower LDL. According to the EAS consensus, they are not registered for lowering Lp(a), but can lower it as a side effect.

In ODYSSEY OUTCOMES, alirocumab lowered Lp(a) after acute coronary syndrome by a median of 5.0 mg/dL. There, Lp(a) lowering and LDL lowering independently predicted fewer events, with a hazard ratio of 0.994 per 1 mg/dL of Lp(a) lowering. That is an indication, but an observation within a trial arm and not a separate randomized proof.

Lipoprotein apheresis

In apheresis, the blood is cleaned outside the body, similar to dialysis. In Germany, a directive of the Federal Joint Committee (G-BA) strictly regulates when it can be performed in statutory outpatient care: isolated Lp(a) elevation above 60 mg/dL, LDL in the normal range, progressive vascular disease documented clinically and by imaging, a prior cardiological or angiological and lipidological assessment, and proof that no treatment alternatives exist. There, apheresis is explicitly a last resort.

In the German Pro(a)LiFe study with 170 people, the annual rate of major coronary events fell from 0.41 in the two years before to 0.09 under apheresis. Over five years of follow-up, the annual event rate was 0.11 compared with 0.58 in the two years before, and a single session lowered Lp(a) by an average of 68.1 percent. These are, however, before and after comparisons without a control group. People enrolled during a phase of clustered events often have fewer afterwards for other reasons as well. The EAS consensus, too, considers apheresis only for very high Lp(a) and progressive disease despite optimal treatment, and points to the uncontrolled data.

Niacin

Niacin, meaning high-dose nicotinic acid, can lower Lp(a). In the AIM-HIGH trial it fell by 21 percent after one year. Events still did not decrease: 16.4 percent on niacin versus 16.2 percent on placebo, and the trial was stopped early for lack of efficacy.

Targeted Lp(a)-lowering agents

The principle of the new substances in one image: the liver cell receives a building instruction for apolipoprotein(a) from the LPA gene. Antisense agents and small interfering RNA (siRNA) are designed to intercept this instruction before the protein is built. Muvalaplin is an attempt to reach the same goal with a tablet.

In terms of the lab value, the effects are large. In the phase 2 trial, pelacarsen lowered Lp(a) dose-dependently by 35 to 80 percent, compared with 6 percent on placebo. Olpasiran achieved placebo-adjusted reductions of up to 101.1 percent in phase 2, lepodisiran in the highest dose group a placebo-adjusted, time-averaged minus 93.9 percentage points between day 60 and 180.

As of September 2026, none of these medications is approved, and the Novartis announcement itself states that there is so far no approved targeted treatment for elevated Lp(a). That may change with the ongoing trials, but no one can give a credible date today. The muvalaplin figures incidentally illustrate the unit problem once more: how much the value falls depends on what it is measured with.

A genetic calculation led by Burgess from 2018 makes the open question tangible. To achieve the same effect as lowering LDL by 1 mmol/L, an absolute Lp(a) reduction of 101.5 mg/dL would be mathematically required, with an uncertainty range of 71.0 to 137.0 mg/dL. Whether the HORIZON result was due to the size of the reduction, the selection of participants or the agent itself cannot be said without the full data.

And now you know why I am not selling you hope for an Lp(a) drug any time soon, and still not telling you that there is nothing to do.

What is thinly supported and why I mention it anyway

Anyone searching for how to lower Lp(a) naturally quickly lands on websites with gram amounts for L-carnitine, coenzyme Q10 or niacin. I deliberately do not repeat these doses. But you should know what lies behind the recommendations.

Measurable in the lab value, benefit for heart and blood vessels unproven

• L-carnitine: A meta-analysis of randomized trials led by Serban found a mean Lp(a) reduction of 8.82 mg/dL, and 9.00 mg/dL with oral intake. The authors themselves call for outcome trials on benefit and safety.
• Coenzyme Q10: Seven small randomized trials with 409 people showed a reduction of 3.54 mg/dL, with no effect on other blood lipids. What Q10 means in connection with statins is covered in Coenzyme Q10 and statins.
• Aspirin: In a genetic subgroup analysis of the ASPREE trial in people aged 70 and over, there was an indication that carriers of a particular LPA variant might benefit. The interaction was only just significant (P = 0.049), Lp(a) was not measured at all, and overall aspirin increased clinically significant bleeding by 1.7 per 1,000 person-years. The EAS consensus sees no basis for guiding aspirin use by the Lp(a) level. Because of the bleeding risk, aspirin is not suitable for self-medication. Conversely: aspirin prescribed by a physician, for example after a heart attack or stroke, is not stopped on your own because of these data.
• Hormone replacement after menopause can lower Lp(a) by about 25 percent according to the EAS, on a limited data basis; in Copenhagen it was 12 percent lower under hormone therapy. That is not an indication for hormone therapy.

And now you know why a falling lab value alone says nothing yet about your heart.

Lp(a) is a family matter: parents, siblings, children

Once you know your value, the next question often comes all by itself: and my children? My siblings?

The inheritance can be explained without jargon. Every person carries two variants of the LPA gene, one from the mother and one from the father. Both contribute to the level. That is why values are very similar within families, and why a high value in you can be a signal for your relatives.

This becomes particularly clear in familial hypercholesterolemia. In the SAFEHEART cohort, systematic testing of relatives, starting from people with FH and high Lp(a), found one new case of high Lp(a) per 2.4 people tested. Over five years, the hazard ratio for a vascular event or death was 4.40 when FH and high Lp(a) occurred together, compared with people with neither.

For children, the data are more limited. In a cascade program with 103 children and adolescents from families with FH, every second child tested had Lp(a) of 30 mg/dL or more if the adult parent had FH and Lp(a) of 50 mg/dL or more. With FH and normal Lp(a), it was one in 7.5.

The EAS consensus recommends cascade testing in such families and a targeted approach for children. It also points out that values can still rise into adulthood. Whether and when a child is tested therefore belongs in a conversation with a physician.

And now you know why an inherited value is not your destiny, but an early hint of where prevention can pay off for you and your family.

Frequently asked questions about lipoprotein(a)

Where Lp(a) connects to other topics

These articles take you further from here.