The Drug Lowered Lp(a). So Why Didn't the Trial Prove a Heart Benefit?
In a Nutshell
The Horizon trial showed pelacarsen lowered Lp(a) but failed to reduce cardiovascular events in 8,323 patients with existing disease, highlighting that lowering a marker without proven clinical benefit may not justify approval. Lp(a) is genetically determined and causally linked to heart disease, but decades of arterial damage mean that treating after plaque has formed may be too late for measurable benefit in short trials. Future trials like OCEAN(a) are testing earlier intervention in at-risk patients before first events, with results expected 2028-2031.
These notes were generated by AI and may contain inaccuracies.
A patient who hiked, lifted weights, and maintained a careful diet still required a heart stent after a severe blockage was discovered in the left anterior descending artery. A heart scan revealed this critical finding. The patient carried another risk factor beyond lifestyle: elevated Lp(a). Genes largely determine levels of this cholesterol-carrying particle, and neither diet nor exercise reliably lowers it to any significant degree.
The physician closely monitored Lp(a) drug trials hoping for a treatment that would prevent heart attacks and strokes. Pelacarsen's Horizon trial lowered Lp(a) levels but failed to prove cardiovascular event reduction. The trial did not meet its primary endpoint. The question emerged: can a drug be approved that lowers only the Lp(a) marker without improving the heart attacks and strokes that matter clinically?
Atherosclerosis develops over decades, beginning in childhood. The timing of treatment initiation matters significantly. Starting a trial in adulthood after disease has already established may limit observable benefits. This principle applies broadly to longevity and preventive medicine.
The Horizon trial enrolled 8,323 participants with established cardiovascular disease and Lp(a) levels of at least 70 mg/dL. Participants received monthly pelacarsen or placebo alongside standard care. The primary composite endpoint included cardiovascular death, non-fatal heart attack, non-fatal stroke, and urgent coronary revascularization requiring hospitalization.
Trial failure on the primary endpoint does not prove the drug had zero beneficial effect or caused harm. Two scenarios can produce the same headline: one where groups showed nearly identical outcomes with precise measurement, and another where wide confidence intervals leave room for meaningful effects. The Horizon results fell into the latter category, showing uncertainty rather than definitive lack of benefit.
Approximately one in five people has elevated Lp(a), varying by population and cutoff definitions. Standard risk factors like weight, fitness, or diet do not reliably identify affected individuals. A routine lipid panel does not detect Lp(a); a separate test is required.
Lp(a) resembles LDL but contains an additional protein called apolipoprotein(a). This particle carries oxidized phospholipids that promote inflammation. Lp(a) particles cause vascular damage independent of other lifestyle factors. Classic lipid-lowering therapies, including PCSK9 inhibitors, do not lower Lp(a) sufficiently. Lp(a) also links to aortic valve calcification and stenosis.
Population studies demonstrate that inherited genetic variants associated with higher Lp(a) levels correlate with increased coronary risk. Genetic, population, and biological evidence together support Lp(a) as a causal disease factor.
Pelacarsen acts in the liver where apolipoprotein(a) is synthesized. The drug binds to RNA instructions, directing the liver to produce less apolipoprotein(a). It does not alter DNA. Effects cease upon drug discontinuation. Earlier trials demonstrated significant Lp(a) reduction, though Horizon did not specify the magnitude or consistency achieved.
An 80% reduction from different baseline levels produces different absolute remaining amounts. Lowering Lp(a) from 100 mg/dL to 20 mg/dL versus from 300 mg/dL to 60 mg/dL both represent 80% reduction, but the remaining particle burden differs substantially. Whether benefit depends on reaching a specific safe threshold, the magnitude of change, or duration at lower levels remains unknown.
Monthly injectable administration raises adherence concerns. Real-world factors including concurrent treatments, adherence rates, and background care may narrow observable differences between groups.
Decades of risk accumulation followed by testing treatments after plaque has formed may not reverse established disease trajectories, particularly with calcified plaque. Scientists enroll participants with existing disease because they are more motivated, recruitment occurs faster, and events accumulate within feasible timeframes. Primary prevention trials without disease would require decades of follow-up.
Horizon tested secondary prevention: preventing additional events in participants who already experienced cardiovascular disease. Primary prevention—preventing first events—remains the goal. A 35-year-old with high Lp(a) and no disease differs fundamentally from a 60-year-old with identical Lp(a) levels but 25 additional years of arterial exposure.
Treatment benefit depends on three factors: age at treatment initiation, duration of Lp(a) lowering achieved, and time required for clinical benefit to manifest. A drug may rapidly change blood levels while disease modification requires substantially longer periods.
Horizon participants had average Lp(a) levels in the two-digit range. Many patients present with levels exceeding 100 mg/dL. Higher baseline levels may show more dramatic absolute changes within limited trial durations.
Evolocumab (Repatha) reduced cardiovascular events in participants with established disease on statins within 2.2 years in the FOURIER trial. This drug lowered both LDL and Lp(a), demonstrating that medications can produce measurable benefits in secondary prevention populations despite existing disease.
Horizon participants received guideline-directed care in both treatment arms. Pelacarsen needed to demonstrate incremental benefit beyond existing therapies, which may reduce observable effect sizes over short timeframes.
Following low-risk individuals for decades presents logistical challenges. Participants relocate, initiate other medications, discontinue study drugs, and experience changing medical care. Very few events may occur within study periods, preventing statistical divergence between groups. The National Institute on Aging notes that aging mechanisms may require decades to produce clinical effects.
The HOPE-3 trial enrolled participants without cardiovascular disease but at intermediate risk. Over 5.6 years median follow-up, cardiovascular death, heart attack, or stroke occurred in 3.7% of the rosuvastatin group versus 4.8% of placebo, yielding a 1.1 percentage point absolute difference. This represents the magnitude of benefit achievable in primary prevention over relatively short durations.
Horizon planned to conclude after 993 confirmed primary events with minimum 2.5 years follow-up. Event-driven designs wait for predetermined information accumulation rather than fixed calendar time. The study completed in July 2026, with participants enrolled at staggered times. Later enrollees contributed less follow-up time.
Grouping outcomes (death, stroke, urgent procedures) facilitates statistical analysis but obscures differential effects on individual outcomes. A drug might reduce deaths while increasing procedures, or vice versa, with meaningful clinical implications despite overall composite non-significance.
Horizon planned analysis of participants with Lp(a) ≥90 mg/dL alongside the overall population. Results from this higher-risk subgroup warrant careful examination, as greater baseline levels may demonstrate larger treatment effects.
The FDA approved inclisiran (Leqvio) based on LDL reduction at 24 weeks without completed cardiovascular outcomes data, citing extensive prior evidence that LDL lowering prevents events. Lp(a) lacks equivalent accumulated evidence for surrogate endpoint qualification. Inclisiran's cardiovascular outcomes trial completion is estimated for 2029.
Biological age tests predict lifespan but lack evidence that interventions lowering these markers extend life. Similar evidentiary gaps exist for Lp(a) as a surrogate endpoint.
Lp(a) should be measured once per US guidance recommendations. Identifying inherited risk informs intensity of modifiable risk factor management. Testing once suffices without available interventions requiring serial monitoring.
When Lp(a) is elevated, evaluate LDL, ApoB, blood pressure, smoking status, diabetes, kidney disease, family history, and existing cardiovascular disease together. Units matter: 160 nmol/L differs from 160 mg/dL, representing particle concentration versus mass without reliable conversion factors.
Heart disease prevalence varies across populations, with elevated rates in South Asian and East Asian groups. However, high Lp(a) occurs across ancestry groups. Testing is preferred over assumptions based on background. Familial clustering warrants testing parents, siblings, and children when an index case is identified.
A patient with Lp(a) 160 nmol/L, elevated ApoB, blood pressure >130/80 mmHg, and parental early heart attack warrants aggressive prevention. The same patient with Lp(a) 116 nmol/L, well-controlled ApoB, normal blood pressure, never-smoker status, and favorable calcium score or CT angiography has substantially lower concern. European consensus estimates indicate 50 mg/dL Lp(a) carries approximately 40% higher cardiovascular risk versus 7 mg/dL; 100 mg/dL approximately doubles risk.
Suppose a drug truly reduces your relative risk of cardiovascular disease by 20% over 5 years. In a group with a 20% starting risk, that would mean 40 fewer events per 1,000 people. With a 2% starting risk, that would mean four fewer per thousand. Same relative reduction, but 10 times the absolute benefit in that first group. Then even if a drug has only partial effect, it can really matter for your outcomes. If you're starting younger or you plan on living another 100 years, no matter what your current age is, then all of this just expands into a much bigger effect.
Niacin can make the Lp(a) result look better, roughly 20 to 30% lower at doses used as a cholesterol drug. Yet large trials adding niacin-based treatment to effective statin therapy did not show any extra cardiovascular benefit. HPS2-Thrive tested niacin with laropiprant and found more serious adverse effects including diabetes problems and bleeding, which is why niacin is not used. The treatment that barely helps this number can help the person. The treatment that improves it may not add benefit. That is why one shouldn't just take a particular medication just to improve the number. It's a lot more complicated than it seems.
If you today have a high Lp(a) and even though there are no drugs to directly target it, take a look at it. It lowers your LDL. Even if it's not a treatment for Lp(a), the injectable PCSK9 antibodies, evolocumab and alirocumab, they lower the LDL substantially and will lower Lp(a) roughly 20 to 30% and they do reduce actual heart attacks and strokes and other cardiovascular outcomes.
Lipoprotein apheresis equipment actually goes through your blood. Your blood is getting pumped into a machine and removes lipoproteins from your blood then returns the blood to you cleansed in a sense. So Lp(a) can fall from 60 to 80% in a single session. But it will rise between treatments. So you have to do this pretty frequently, at least once every few months. Currently the US Lp(a) related indication for doing this is quite narrow. An inherited cholesterol disorder called heterozygous familial hypercholesterolemia, high Lp(a) and documented coronary or peripheral artery disease and also an inadequate response to treatment or inability to tolerate treatment otherwise.
Hormone therapy can lower Lp(a), but one wouldn't start it just for that number. It seems to be this added benefit. Aspirin is not related at all. It addresses the clotting. Whether that's got an added benefit outweighing the bleeding risk is a totally separate question.
Pelacarsen failed. But the next trials ask about other drugs, earlier treatment or earlier dosing. We have olpasiran and zerlasiran that are still in trials today. They use a different mechanism, a small interfering RNA to reduce that apolipoprotein(a) production. In addition, we've got muvalaplin, which is an experimental pill that interferes with the assembly of the Lp(a) particle. We're still early in the trials for all of these, so we don't have those outcomes yet.
Look out for the OCEAN(a) pre-event trials results. It's studying selected people 50 years or over with a high Lp(a) and other risks or evidence of plaque before a first major cardiovascular event or procedure to restore blood flow such as a stent. Earlier prevention is getting a trial. It won't answer all of our questions, especially starting really early. It will answer some really good questions. The trials on this screen list are estimating our primary completion dates from 2028 through 2031. For sure, these dates certainly can move, but these are the time periods that we're looking at for the results that we're hoping for.
There is one failed trial, but it points to a lot of problems with studying interventions for long-term disease processes. There are still other drugs that are out there that are still being studied. The results are still fairly close in the future compared to the length of time it takes for atherosclerosis to cause damage. Even if you've got a high Lp(a), don't lose hope. Keep an eye out and work on your other risks.
Keep Hillary Lin, MD in your library
Save the videos and channels worth coming back to, and find them again in one place.
