Can you lower Lp(a)? The drugs in trials, and why none is approved yet

By Marcus Renshaw · · updated

For years the only honest advice on a high Lp(a) was to drive every other risk factor down, because nothing touched the number itself. Diet did not move it, exercise did not move it, and statins nudged it the wrong way. Five drugs in trials lower Lp(a) by 80 to 95 percent, and for two years the reasonable assumption was that one of them would close the gap. In September 2026 the first of them finished the trial that mattered and did not close it. That result is the most useful thing to have happened in this field, and not in the way anyone wanted.

As of September 2026 no drug is approved to lower Lp(a), and the first outcomes trial has now failed. Pelacarsen lowered Lp(a) in 8,323 patients with established heart disease and did not reduce cardiovascular events. Four other drugs remain in trials. Measure your Lp(a) once anyway, because it still changes how hard you push everything else.

Why has Lp(a) had no treatment, when cholesterol has so many?

Because Lp(a) is not really a cholesterol problem, it is a genetics problem. Lipoprotein(a) is an LDL-like particle with an extra protein bolted on, and your level is set by the gene you inherited. It is roughly ninety percent heritable and stable across your adult life, which is exactly why the standard cholesterol toolkit does nothing useful to it. Statins lower LDL and can raise Lp(a) slightly. Diet and exercise, which move so much else, leave it flat.

Compare that to how fast the cholesterol side keeps advancing. On July 16, 2026, the FDA approved the first oral PCSK9 inhibitor, a once daily pill that cuts ApoB and LDL substantially. That is a real step, and it is worth naming precisely because it is not an Lp(a) drug. Oral PCSK9 inhibition lowers Lp(a) only modestly, on the order of 20 to 25 percent, and it is not indicated for it. So the same month that cholesterol got easier to treat, the Lp(a) gap stayed open. The drugs built to close it are a different class aimed at this one job, and none is approved.

What are the drugs that actually lower Lp(a)?

Five, at different stages, using three mechanisms. Four are injections that switch off Lp(a) production in the liver, either by degrading the messenger RNA for apo(a) (an antisense oligonucleotide) or by silencing it (a small interfering RNA). The fifth is an oral pill that blocks the particle from assembling in the first place.

DrugTypeHow it is givenLp(a) reduction (Phase 2)
PelacarsenAntisense oligonucleotideMonthly injectionAbout 80 percent
OlpasiransiRNAInjection every 12 weeksOver 95 percent
LepodisiransiRNA (long acting)Infrequent injectionAbout 94 percent
ZerlasiransiRNAInjectionAround 80 to 90 percent
MuvalaplinOral small moleculeDaily pillUp to 85.8 percent

One caveat on that table, and it is the same lesson as asking for your own result in the right unit. Muvalaplin’s headline figure depends on which assay measured it: in the KRAKEN trial the top dose cut Lp(a) by 85.8 percent on an assay for intact Lp(a) but by 68.9 percent on the traditional apo(a)-based assay, because the drug blocks particle assembly and leaves fragments the older assay still counts. Both numbers are from the same participants.

The reductions are real and large, and they were always the reason to slow down rather than speed up. Every number in that column is a Phase 2 result, and every one of them measures Lp(a) itself, the marker. None of them is a result about heart attacks. That gap between lowering a number and preventing an event is the whole question, and in September 2026 it got its first answer.

Does lowering Lp(a) actually prevent heart attacks?

On the only outcomes trial that has reported, no.

What happened in Lp(a)HORIZON?

On 4 September 2026 Novartis announced that Lp(a)HORIZON did not meet its primary endpoint. It was the first cardiovascular outcomes trial of any Lp(a) lowering therapy, and the result is the one nobody in the field had planned for.

The trial design was not the problem. Lp(a)HORIZON randomized 8,323 patients with established cardiovascular disease, defined as a prior heart attack, ischemic stroke or symptomatic peripheral artery disease, and an Lp(a) of 70 mg/dL or higher (about 149 nmol/L), to monthly injections of pelacarsen 80 mg or placebo on top of optimized standard care. The primary endpoint was a composite of cardiovascular death, nonfatal heart attack, nonfatal stroke, or urgent coronary revascularization requiring hospitalization, and it was designed to be tested both in that whole group and in a prespecified subgroup with Lp(a) at or above 90 mg/dL (about 192 nmol/L). Minimum follow up was two and a half years, and the trial was built to run until 993 confirmed primary events had accumulated.

Novartis’s development chief put the result plainly: “Although lower Lp(a) levels were observed with pelacarsen, the findings did not demonstrate that this translated into reduced cardiovascular risk in the overall study population.” So the drug did its pharmacological job. The patients did not get the benefit.

Three things are not yet known, and it is worth being precise about them rather than filling the gaps. The topline release did not report the size of the Lp(a) reduction achieved, it did not report the result in the high Lp(a) subgroup, and it did not publish the event counts. Novartis says the full data will be presented at an upcoming medical congress. Until that happens, the defensible statement is that the primary endpoint was missed in the overall population, and nothing more specific than that.

Does that kill the Lp(a) hypothesis?

Not by itself, and the reasons are worth understanding rather than treating as consolation.

The history of lipid drugs is littered with markers that moved beautifully while patients did not benefit. Torcetrapib raised HDL and increased deaths. Niacin lowered several markers and failed to prevent events in the trials that mattered. Lp(a) has now joined that list, at least for this drug in this population, and the caution that section of this article carried in July was the correct one.

It also landed in a fortnight that made the general point better than any single trial could. Six days earlier, on 29 August, STAREE reported that a statin cut cardiovascular events by 30% in healthy adults over 70 without improving disability free survival, so the marker moved, the events moved, and function did not. The next day, a trial in 33,995 people reported that controlling uncontrolled hypertension cut dementia by 15% and held the effect for three years after the program stopped. Three trials inside a week, three markers, and only the cheapest one to measure delivered the outcome people actually care about. That is not a coincidence about biology. It is what happens when a field finally runs the trials instead of arguing from the markers.

But the genetic case that made Lp(a) a target has not gone anywhere. People who inherit naturally low Lp(a) have less heart disease across their whole lives. The obvious tension is that inheriting a low level means sixty years of low exposure, while a drug trial in people who already have established heart disease offers perhaps five years of it, starting after the arteries are already damaged. A lifetime risk factor treated late for a short time is a genuinely different experiment from a lifetime of low exposure, and that argument was made in advance rather than invented after the result, which is what makes it worth taking seriously.

The counterargument is just as real. Everyone in Lp(a)HORIZON was on optimized background therapy, which in a modern secondary prevention trial means aggressively treated LDL. If the remaining risk attributable to Lp(a) in a well treated patient is small, no amount of Lp(a) lowering will show up as fewer events. That is an explanation for the miss and also a limit on how much the drug could ever have been worth to that patient.

Two other outcomes trials now carry the question, and both are larger or longer than the one that failed. Amgen’s OCEAN(a) Outcomes trial of olpasiran has enrolled 7,297 patients and is no longer recruiting, with primary completion listed as March 2028. Lilly’s ACCLAIM-Lp(a) trial of lepodisiran has enrolled 17,300 and lists primary completion in March 2029, and it includes high risk primary prevention patients rather than only people who have already had an event. If the late and short explanation is right, ACCLAIM is the trial shaped to show it. Both dates are later than the ones widely quoted a year ago, including in the earlier version of this article, so treat any timeline you read on this subject as provisional.

What does the 2026 guideline tell you to do right now?

Measure it once. The 2026 ACC/AHA multisociety dyslipidemia guideline made measuring Lp(a) at least once in every adult a Class I recommendation, its strongest tier, and the first time universal Lp(a) screening has been recommended at that level. A single test captures your lifetime exposure, because the number does not drift.

The thresholds tell you what the result means. Risk climbs continuously, with roughly a 1.4 times increase in cardiovascular risk above 125 nmol/L and about double above 250 nmol/L. Ask for the result in nmol/L, the current particle based unit, rather than the older mg/dL mass units. If your number is high, you are in the roughly one in five adults who carry this, and the guideline’s logic is the same as it has always been: you cannot lower this one, so drive every other risk factor lower to compensate, ApoB first.

How should a self-experimenter play this?

The clinical floor here is do the one time test. That has not changed, and the case for it is now simpler than it was in July, because it no longer leans on a drug arriving.

Here is the reasoning. Lp(a) is one of the highest value tests you will ever buy per krona: a single measurement, a few hundred SEK, never repeated. The argument for it is knowing whether to push your other risk factors harder, and a high result is a standing instruction to be more aggressive about ApoB, blood pressure and everything else you can actually move. In July this article carried a second argument, that having the number on file would make you ready when a drug was approved. That argument is weaker now, and I would rather retract it than let it stand: pelacarsen’s miss pushes any approval further out and makes the size of the eventual benefit an open question. Test because the number changes what you do this year, not because of what might be prescribable in 2029.

Two honest cautions for this crowd specifically. First, do not let the pipeline talk you into chasing anything early, and note that the pipeline just supplied the clearest possible reason why. These drugs are investigational, the one that finished its outcomes trial failed it, and the reductions in the table above are marker reductions. Lipoprotein apheresis, the one existing way to physically strip Lp(a) from blood, is reserved for extreme cases and is not a biohack. Second, be clear about what your current stack does not do. Nattokinase and omega-3 work on clotting and triglycerides, not on Lp(a). A statin lowers ApoB and does nothing good for Lp(a). None of what you may already be running touches this number, which is the entire reason a dedicated drug class had to be built, and the reason a high result changes what you do everywhere else rather than here. As always, these markers inform a decision you make with a clinician, not instead of one.

If you have never measured your Lp(a), order it on your next draw and record it as the one time value it is. The pelacarsen result changes what a high number means for your prospects of ever treating it directly. It does not change what a high number means for how hard you work on everything else, which is the part that was always inside your control.

FAQ

Can you lower Lp(a) with medication?

No, as of September 2026 there is no approved drug. Five drugs that lower Lp(a) by 80 to 95 percent have been in clinical trials, using antisense, siRNA, or an oral small molecule. Approval depends on cardiovascular outcomes data, and the first such trial reported in September 2026: pelacarsen lowered Lp(a) but missed its primary cardiovascular endpoint.

Can you lower Lp(a) naturally with diet or exercise?

No. Lp(a) is roughly ninety percent genetic and stable across adult life. Diet, exercise, and weight loss do not meaningfully move it, and statins can raise it slightly. This is why a single measurement captures your lifetime exposure, and why a high result is managed by lowering every other risk factor rather than by targeting Lp(a) itself.

What Lp(a)-lowering drugs are in trials?

Pelacarsen (antisense), olpasiran, lepodisiran, and zerlasiran (all siRNA injections), and muvalaplin (an oral pill). In Phase 2, injectable agents lowered Lp(a) by roughly 80 to over 95 percent, and the oral agent by up to about 85 percent. These are reductions in the marker, not proven reductions in heart attacks.

When will an Lp(a) drug be approved?

Later than expected, and possibly not at all for the leading candidate. Pelacarsen’s Lp(a)HORIZON trial, the first to test whether lowering Lp(a) prevents cardiovascular events, missed its primary endpoint in September 2026. Approval would follow a positive outcomes result, so attention moves to olpasiran, with primary completion listed for March 2028, and lepodisiran, listed for March 2029.

Does lowering Lp(a) actually reduce heart attack risk?

On the only trial that has tested it, no. Lp(a)HORIZON gave 8,323 patients with established cardiovascular disease monthly pelacarsen, lowered their Lp(a), and did not reduce cardiovascular events. The genetic case remains that people born with low Lp(a) have less heart disease, but that reflects a lifetime of low exposure rather than a few years of treatment after disease is established. Two larger trials are still running.

Is any Lp(a) drug the same as the new oral cholesterol pill?

No. The oral PCSK9 inhibitor the FDA approved in July 2026 lowers LDL and ApoB and is a cholesterol drug, not an Lp(a) drug. It lowers Lp(a) only modestly, around 20 to 25 percent, and is not indicated for it. The dedicated Lp(a)-lowering drugs are a separate class and remain investigational.

Did the pelacarsen trial fail?

Yes, on its primary endpoint. Novartis announced on 4 September 2026 that Lp(a)HORIZON did not meet its primary endpoint of reducing cardiovascular death, nonfatal heart attack, nonfatal stroke, or urgent coronary revascularization, in 8,323 patients with established cardiovascular disease and an Lp(a) of 70 mg/dL or higher. Lp(a) levels did fall. The company has not yet released the event counts, the size of the Lp(a) reduction, or the result in the higher risk subgroup, and says full data will come at a medical congress.

Should I get my Lp(a) tested now if drugs are not approved yet?

Yes. The 2026 guideline recommends measuring it once in every adult, and that recommendation is unaffected by the pelacarsen result. Knowing your number tells you how hard to push the risk factors you can move, which matters more now that lowering Lp(a) itself looks further away. It is a one time test that never needs repeating.

Sources