Novartis and Ionis Pharmaceuticals have reported a major setback for one of the most closely watched cardiovascular drug programs, announcing that pelacarsen failed the primary endpoint of the Phase 3 Lp(a)HORIZON outcomes trial despite substantially lowering lipoprotein(a). The randomized, double-blind study enrolled 8,323 people with established cardiovascular disease and elevated Lp(a), making it the first pivotal outcomes trial specifically designed to answer whether pharmacologically lowering this largely inherited risk factor could prevent heart attacks, strokes, cardiovascular deaths and urgent coronary procedures on top of contemporary standard treatment. The answer from the topline analysis was negative: the reduction in Lp(a) did not translate into a statistically successful reduction in the trial’s four-component cardiovascular endpoint.
That result matters well beyond one experimental drug because Lp(a) has become one of cardiology’s most discussed remaining sources of residual risk. Approximately one in five people worldwide has an elevated concentration, levels are predominantly genetically determined, and they tend to change relatively little through diet or lifestyle. Pelacarsen was intended to address that risk directly using an antisense oligonucleotide that suppresses production of apolipoprotein(a), an essential component of the Lp(a) particle. Novartis has not yet disclosed the actual cardiovascular event rates, hazard ratio or exact percentage reduction in Lp(a), so the full magnitude and pattern of the failure cannot yet be assessed.
What exactly failed in the Phase 3 Lp(a)HORIZON trial?
Lp(a)HORIZON was designed around hard cardiovascular outcomes rather than cholesterol measurements or imaging biomarkers. Its primary endpoint combined cardiovascular death, nonfatal myocardial infarction, nonfatal stroke and urgent coronary revascularization requiring hospitalization. Participants had established cardiovascular disease, elevated Lp(a) and continued receiving guideline-directed therapies such as lipid-lowering and antihypertensive drugs, meaning pelacarsen had to demonstrate additional protection after many conventional risk factors were already being treated.
The study assessed the primary endpoint in an overall population with Lp(a) of at least 70 mg/dL and a prespecified subgroup with levels of at least 90 mg/dL. Novartis said the study did not meet its primary endpoint, while Ionis separately confirmed that substantial Lp(a) lowering had occurred and that the safety profile was considered acceptable. Neither company has yet released enough numerical detail to determine whether the cardiovascular curves were completely overlapping, showed a nonsignificant trend, differed by baseline Lp(a), or behaved differently for myocardial infarction, stroke and other individual components. Those questions now become critical because a narrowly missed endpoint would carry different biological implications from a trial showing essentially no separation.
Why was pelacarsen expected to work if high Lp(a) is linked to cardiovascular disease?
Lp(a) resembles an LDL-like particle attached to apolipoprotein(a), and extensive genetic and epidemiological evidence has associated elevated levels with atherosclerotic cardiovascular disease. Unlike LDL cholesterol, which is strongly influenced by diet, metabolism and several classes of medicines, Lp(a) concentration is approximately 90% genetically determined. That characteristic has made it attractive as both a risk marker and therapeutic target because people can remain at elevated risk even when LDL cholesterol and blood pressure are otherwise well controlled.
Pelacarsen uses antisense technology to interfere with RNA involved in producing apolipoprotein(a) in the liver, thereby reducing formation of circulating Lp(a). Earlier studies established that the molecule could lower the biomarker substantially, which allowed researchers to move to the much more difficult question of whether changing the laboratory value would change clinical outcomes. Lp(a)HORIZON was therefore never merely a test of whether pelacarsen could lower Lp(a); that capability had already been demonstrated. It was a test of whether the biomarker was modifiable in a way that delivered enough cardiovascular benefit to justify long-term preventive treatment.
Does this failure mean elevated Lp(a) is not actually a cause of cardiovascular disease?
No. A failed drug trial does not by itself invalidate the broader genetic and biological evidence linking elevated Lp(a) to cardiovascular risk. The result shows that the specific intervention, dose, treatment duration and study population used in Lp(a)HORIZON did not produce a statistically successful reduction in the prespecified cardiovascular endpoint. It does not yet tell researchers whether greater Lp(a) lowering, earlier treatment, longer exposure, treatment before established cardiovascular damage, or a different therapeutic modality could produce another outcome.
This distinction is particularly important because cardiovascular disease develops over decades. People in Lp(a)HORIZON already had established disease when they entered the trial, so reducing one residual risk pathway for several years may not necessarily reverse damage produced by lifelong exposure. Another possibility is that the magnitude of Lp(a) lowering achieved with pelacarsen was insufficient for the level of cardiovascular benefit required, although Novartis has not yet disclosed the numerical biomarker reduction from the Phase 3 population. The forthcoming full dataset will therefore influence how confidently the field can distinguish a molecule-specific failure from a broader challenge to Lp(a)-lowering strategies.
Why did Novartis enroll more than 8,000 patients instead of relying on a smaller biomarker study?
Cardiovascular outcomes trials require large patient populations because heart attacks, strokes and cardiovascular deaths occur much less frequently than changes in laboratory biomarkers. A drug can dramatically alter a blood test within weeks, but demonstrating that the biological change prevents enough clinical events requires thousands of participants and years of observation. ClinicalTrials.gov records 8,323 actual participants in Lp(a)HORIZON, with the trial beginning in December 2019 and continuing until its planned 2026 completion period.
The trial’s size also underscores the strength of the question Novartis and Ionis were trying to answer. Participants were not untreated high-risk patients for whom almost any intervention might produce an obvious improvement. They were people with established cardiovascular disease already receiving appropriate preventive therapies, leaving pelacarsen to address residual risk on top of statins, antihypertensive treatment and other guideline-directed care. That makes a positive result harder to achieve, but it is also the evidence required to show that adding another long-term medicine produces genuine incremental value.
What does the failure mean for other Lp(a)-lowering drugs now in development?
The result will intensify scrutiny of the entire therapeutic category, but it does not automatically determine the fate of every competing program. Different drug classes can produce different depths and durations of biomarker suppression, and some newer approaches are designed to reduce Lp(a) extremely profoundly. Their trial populations, entry thresholds and treatment durations may also differ from Lp(a)HORIZON.
The most important comparison will eventually be between achieved Lp(a) reduction and cardiovascular outcomes across different programs. If another drug produces substantially deeper suppression and succeeds in preventing events, pelacarsen may be remembered as evidence that there is a threshold of lowering required for clinical benefit. If multiple mechanistically distinct therapies dramatically suppress Lp(a) yet repeatedly fail outcomes trials, researchers would have to reconsider how directly and modifiably the biomarker contributes to residual cardiovascular risk.
For now, such conclusions would be premature. Novartis has said the complete Lp(a)HORIZON results will be presented at an upcoming medical congress, and the event curves, subgroup analyses, achieved Lp(a) concentrations and individual endpoint components will be much more informative than the topline statement alone.
What does the result mean for Ionis after licensing pelacarsen to Novartis?
Ionis discovered pelacarsen and completed its early development before Novartis exercised an option in 2019 to license rights to develop and commercialize the therapy. That arrangement shifted responsibility for the expensive global cardiovascular outcomes program to Novartis while leaving Ionis economically connected to the asset. The failure removes what could have become an important validation of antisense technology in a very large cardiovascular prevention population, although it does not undermine the mechanism of antisense medicines across unrelated diseases.
Ionis now has several independently commercialized medicines and a broader pipeline spanning neurological, cardiometabolic and other conditions. Pelacarsen was nevertheless strategically distinctive because a successful Lp(a) therapy could potentially address a vastly larger population than many rare-disease antisense programs. The companies have not announced a regulatory filing plan after the failed endpoint, and without a successful cardiovascular outcomes endpoint it would be difficult to justify assuming a conventional approval path for cardiovascular-risk reduction.
Why are the undisclosed numbers now as important as the headline failure?
The phrase “did not meet the primary endpoint” covers a broad range of possible trial outcomes. A hazard ratio very close to 1.0 would suggest little evidence of clinical benefit, while a favorable numerical trend with a confidence interval narrowly crossing the statistical threshold would generate a different scientific discussion. Researchers will also want to know whether higher baseline Lp(a) concentrations were associated with more benefit and whether particular events such as myocardial infarction responded differently from stroke or cardiovascular death.
Safety will require additional detail as well. Both companies said pelacarsen had an acceptable safety profile, but the release did not provide rates of injection reactions, laboratory abnormalities, treatment discontinuation or serious adverse events. In a preventive cardiovascular medicine that could be taken for years by otherwise stable patients, even relatively modest toxicity can materially affect the benefit-risk calculation.
The full presentation will therefore decide whether Lp(a)HORIZON becomes a clear dead end for pelacarsen or a more complicated scientific failure that informs the design of the next generation of studies. What is already clear is that biomarker credibility was not enough. The largest test so far required Lp(a) lowering to produce fewer cardiovascular events, and pelacarsen did not cross that threshold.
