He was 51. His LDL was 62 on a moderate-intensity statin. He had never smoked, his A1c was 5.3, his blood pressure ran 118 over 74, and he had run a half marathon eight months earlier. He had an anterior MI on a Tuesday morning.

His father died at 55. His older brother had a stent at 49. Nobody in three generations of that family had ever had an Lp(a) drawn.

That is not a rare story. It is the most common story I hear from cardiology colleagues, and the test that would have explained it costs about as much as a basic metabolic panel.

This sentence has kept Lp(a) out of routine practice for thirty years, and it is wrong in two separate ways.

First, it confuses treating the number with treating the patient. You cannot lower Lp(a) with anything currently approved. You can absolutely change what you do about everything else. A 47-year-old with an Lp(a) of 180 nmol/L is not a candidate for watchful waiting on an LDL of 130. She is someone whose lifetime risk justifies aggressive LDL lowering now, a lower treatment threshold, a harder conversation about blood pressure, and a real look at whether her siblings and children should be tested.

The second is that it is about to stop being true. Four agents in late-stage development lower Lp(a) by 80% to 95%. The question of whether that translates into fewer heart attacks is being answered right now, in trials that have already finished enrolling.

Ordering the test today means that when the answer arrives, you already know which of your patients it applies to. Not ordering it means starting from zero on the day a drug gets approved.

Lp(a) is an LDL-like particle with one addition: a molecule of apolipoprotein(a) bound covalently to apoB. That single addition changes the biology substantially. Apo(a) has structural homology with plasminogen, which is the basis for the long-standing hypothesis that Lp(a) is prothrombotic as well as proatherogenic. The particle also carries oxidized phospholipids, which is the basis for the inflammatory arm of the story.

Atherogenic, inflammatory, thrombotic. Three mechanisms, one particle, none of them addressed by lowering LDL.

Clinically, the genetics matter most. Lp(a) concentration is determined overwhelmingly by variation at the LPA locus, particularly the number of kringle IV type 2 repeats. It is set at birth. It does not meaningfully drift with weight, diet, exercise, or age. A patient who has carried an Lp(a) of 200 nmol/L since childhood has had five decades of exposure by the time she reaches your office at 50, and that cumulative exposure is the thing driving her risk.

This is why the standard advice fails so badly here. Telling a high-Lp(a) patient to eat better and exercise more is not wrong as general health advice. It is simply irrelevant to the specific risk factor you just identified, and patients can tell when they are being handed a script that does not fit.

Run through the list, and the picture is bleak.

Statins. No reduction. Several analyses suggest a small increase, which is biologically interesting and clinically irrelevant next to the LDL benefit. Keep prescribing them. Just do not expect them to do anything here.

Ezetimibe. No meaningful effect.

Diet, exercise, weight loss. No meaningful effect. Worth saying out loud to the patient, because otherwise they may assume they have failed at something.

PCSK9 inhibitors. A real but partial effect, roughly 20% to 25%, observed as a secondary finding in the LDL trials. In a patient starting at 250 nmol/L, that is a reduction to something still well above any proposed treatment threshold. Useful. Not a solution.

Niacin. Lowers Lp(a) meaningfully and failed to improve outcomes in the trials that tested it. A cautionary tale about biomarker enthusiasm that is directly relevant to everything below.

Lipoprotein apheresis. Effective, approved in narrow circumstances, and about as accessible as it sounds.

That is the entire toolkit, and this is exactly why the pending trials matter more here than in almost any other area of lipidology. There is no partially adequate current option to fall back on.

All four of the agents below are investigational. None is approved for Lp(a) lowering. Every efficacy figure below is a biomarker change, not an outcome.

Pelacarsen is an antisense oligonucleotide that blocks hepatic apo(a) production, given as a monthly 80 mg subcutaneous injection. Phase 2 dose-ranging produced reductions of roughly 35% to 80% depending on regimen, and the 80 mg monthly dose was carried into phase 3.

Olpasiran is a small interfering RNA that silences LPA in the liver, dosed every 12 weeks. In OCEAN(a)-DOSE, the highest dose lowered Lp(a) by more than 95%.

Lepodisiran is also an siRNA. In the phase 2 ALPACA trial, a single 400 mg dose produced a placebo-adjusted, time-averaged reduction of 93.9% from day 60 through day 180, still above 90% at day 360. A single injection, a year of effect.

Muvalaplin is the interesting outlier: an oral small molecule that disrupts assembly of the Lp(a) particle. In the phase 2 KRAKEN trial, the highest dose produced a placebo-adjusted reduction of 85.8% by intact-particle assay.

Now the part that should govern how you read all of that. Not one of those numbers is an outcome. We have four drugs that are extremely good at moving a biomarker, and no proof that moving it helps anyone. Niacin lowered Lp(a) too. Torcetrapib raised HDL. CETP inhibition looked inevitable for a decade. Lipidology has a long and expensive history of biomarkers that moved beautifully and changed nothing, and the honest position today is that Lp(a) lowering is biologically compelling and clinically unproven.

Lp(a)HORIZON is the first cardiovascular outcomes trial for any Lp(a)-lowering therapy, which makes it the most consequential lipid readout since the PCSK9 outcome trials.

The design is clean. Researchers randomized 8,323 patients between December 2019 and July 2022 across 797 sites in 42 countries. Entry required established cardiovascular disease, defined as prior myocardial infarction, ischemic stroke, or symptomatic peripheral artery disease, plus a screening Lp(a) at or above 70 mg/dL. Patients received monthly pelacarsen 80 mg or placebo on top of optimized standard of care. The primary endpoint is time to first major adverse cardiovascular event. The trial was designed to end when 993 adjudicated primary events had accumulated. Minimum follow-up was 2.5 years. Mean age at baseline was 59.7 years, and 27.0% of participants were women.

Two things about that design deserve emphasis. It is event-driven, not calendar-driven, which means the timing of the readout depends on how fast events accrue rather than on a date someone picked. It also enrolled a secondary-prevention population with high baseline Lp(a), the group most likely to show benefit if one exists. This trial was built to succeed if the hypothesis is right.

Novartis and Ionis guided publicly to a first-half-2026 readout, with regulatory submissions to follow. The registered primary completion date has moved repeatedly across the trial's life, most recently landing on June 30, 2026. That window has now closed without a disclosure.

The rest of the field is close behind. The olpasiran outcomes trial enrolled roughly 7,000 patients with established ASCVD and Lp(a) at or above 200 nmol/L, with completion estimated for December 2026. ACCLAIM-Lp(a), testing lepodisiran, is underway. Within roughly two years we should have not one answer but three, which is an unusual position for a field to be in.

Lp(a) is the most common inherited cardiovascular risk factor in medicine; it is causal rather than merely associated, it affects roughly one person in five, and the overwhelming majority of those people have never been tested because we spent three decades telling ourselves that a number we could not lower was a number not worth knowing.

We were wrong about that even when nothing could lower it. The trials reading out over the next two years will decide how wrong.

Order the test. The drug question will resolve itself in the next two years, and you would rather find out which of your patients need the answer before it arrives than after.

The patient described at the top is a composite. No individual patient is depicted.

Data current as of August 2026. Lp(a)HORIZON had not reported topline results at the time of writing. Verify current trial status before acting on anything above.

Disclosure: The author is Chief Medical Officer of Vineyard, a telehealth obesity medicine practice.

REFERENCES

P.S. Both free one-pagers are live. The GLP-1 Evidence Cheat Sheet, twelve pivotal trials on one page. The Retatrutide Cheat Sheet, the triple agonist in plain numbers.

Following the latest trials? The live tracker is here: substance-over-noise.beehiiv.com/trials

Was this forwarded to you? Join {{active_subscriber_count}} readers and subscribe free.