September 1, 2026 · Science · Biomarkers
Two Trials, One Drug, Opposite Headlines
Two randomized trials of the same cholesterol drug were reported at the European Society of Cardiology Congress on August 31, 2026. One found that lowering LDL cholesterol further reduced deaths. The other found that hitting a strict LDL target twice as often changed nothing measurable within a year. Both are correct, and read together they describe something a single lipid result on a platform dashboard cannot.
VESALIUS-CV randomized 12,257 people to evolocumab or placebo. Every participant had atherosclerosis or high-risk diabetes with elevated LDL cholesterol despite already-optimized lipid-lowering therapy, and none had a prior heart attack or stroke. Median age was 66. Over a median of 4.6 years, all-cause mortality was 20% lower in the evolocumab group. Estimated five-year mortality was 7.9% against 9.7%. The separation between the curves began at roughly 1.5 years. The prespecified mortality analysis was published in Circulation, and the work was reported through the Mass General Brigham Heart and Vascular Institute, according to MedicalXpress.
AMUNDSEN asked what happens when the same drug is started immediately. It randomized 2,161 people having a heart attack, either ST-elevation or non-ST-elevation, undergoing percutaneous coronary intervention with at least one high-risk clinical characteristic. Mean age was 67 and 21% were women. The design was open-label with blinded endpoints, and follow-up ran 12 months. Results appeared in JAMA alongside the congress Hot Line presentation.
The short version
What AMUNDSEN Recorded
AMUNDSEN's biological endpoint was LDL goal attainment, defined as a reduction of at least 50% together with a level below 55 mg/dL. Evolocumab reached that in 82% of patients against 40% on standard care. Median time to target was 9 weeks against 19. On the measure the drug acts on, the result was decisive.
The clinical endpoint was all-cause death or unplanned cardiovascular hospitalization at 12 months. That occurred in 14.6% of the evolocumab group and 15.4% of the standard-care group, a difference that did not reach significance. A per-protocol analysis, which the investigators labelled hypothesis-generating, reported 10.2% against 14.3% with a p value of 0.037. A per-protocol analysis excludes people who did not take the assigned treatment as directed, which breaks the comparability randomization creates, so it is a weaker basis for a conclusion than the primary analysis it sits beside.
The investigators' own reading was that evolocumab may lack early clinically meaningful effects beyond lipid lowering, and that the benefit of lowering lipids "may require more time". They also noted that even with immediate treatment, one in five patients had not reached target a year later.
Why the Gap Between the Marker and the Outcome Matters Here
The recurring problem in the guides we publish is that testing platforms sell measurement and readers infer outcome. A panel returns an LDL value, a dashboard colors it red or green, and the implied promise is that moving the color moves the risk. These two trials put a number on how loosely those are coupled over short periods.
In AMUNDSEN the marker moved further and faster in the treatment group by every measure reported, in a population at extremely high near-term risk, and 12 months of counted events showed no separation. That is not evidence that LDL is unimportant. VESALIUS-CV, in a lower-acuity population followed four times as long, is evidence that it is. It is evidence that the interval between changing a lipid number and observing a change in outcomes is measured in years, and that a quarterly retest showing an improved number is not itself a result.
This is the third trial in two weeks covered here where a lipid marker and a clinical outcome moved differently. The pattern is worth naming: LDL is a well-established causal risk factor with decades of trial evidence behind it, and it is still a poor short-horizon feedback signal for an individual. Those two statements are compatible, and platforms that retest quarterly tend to communicate only the first.
What This Changes for Someone Reading a Lipid Panel
Very little about the measurement, and something about the interpretation. LDL and ApoB remain among the better-evidenced things a blood panel reports. Our ApoB guide covers why the particle count separates people whose standard lipid panels look similar, and our page on normal versus optimal ranges covers how much the threshold a report chooses changes what a number appears to mean. Neither trial here disturbs either of those.
What changes is the weight a short retest interval should carry. A platform that repanels every three months can show a number falling. Nothing in that sequence tells an individual their risk has moved, because the trials that establish the link needed thousands of people and several years to detect it. Anyone whose panel returns an elevated LDL or ApoB is looking at a genuine input to a decision about treatment, and that decision belongs with a clinician who can see the rest of the record: existing diagnoses, other medications, family history, and what has already been tried.
What Would Change This Read
Longer follow-up from AMUNDSEN would change it most directly. The investigators' hypothesis is that lipid lowering needs more time than a year to produce clinical benefit in this population. Extended follow-up on the same 2,161 participants would either show curves separating after year one or show that they do not.
Full publication of the VESALIUS-CV mortality analysis would sharpen the estimate. The figures on this page are the ones reported from the congress presentation and the accompanying coverage. Confidence intervals around the 20% reduction, and the cause-of-death breakdown, would establish how precise the finding is. The Circulation paper was not retrievable at the time of writing, and no number here is taken from it.
Related Coverage
The STAREE trial in adults over 70 lowered LDL by 48 mg/dL, cut cardiovascular events, and left its primary endpoint of disability-free survival unchanged. RADICAL PC 2 raised statin use from 40% to 63% and found no difference in cardiovascular events over nearly six years. A Stanford secondary analysis tied LDL increases on low-carb diets to a genetic score. Read together with the two trials here, the through-line is that lipid markers respond quickly and reliably to treatment, and the outcomes they predict are detected on a much slower clock. The biomarkers hub collects the markers these conversations start from.
Sources
- MedicalXpress, "LDL-cholesterol lowering medication linked to lower mortality in high-risk patients," August 31, 2026, medicalxpress.com (accessed September 1, 2026). Source of the VESALIUS-CV trial name, the 12,257 randomized, the enrollment criteria and the absence of prior heart attack or stroke, the median age of 66, the median 4.6-year follow-up, the 20% lower all-cause mortality, the 7.9% and 9.7% estimated five-year mortality rates, the roughly 1.5-year onset of separation, the Circulation publication and ESC Congress 2026 presentation, the Mass General Brigham Heart and Vascular Institute attribution, and the investigators' statements that the analysis cannot definitively prove causation and that the findings were exploratory but consistently favored evolocumab.
- MedicalXpress, "Immediate evolocumab after a heart attack improves LDL-cholesterol goals but not early cardiovascular outcomes," August 31, 2026, medicalxpress.com (accessed September 1, 2026). Source of the AMUNDSEN trial name, the open-label blinded-endpoint design, the 2,161 randomized, the STEMI and NSTEMI population undergoing PCI with at least one high-risk characteristic, the mean age of 67 and 21% women, the 12-month follow-up, the 82% and 40% LDL goal attainment against a target of at least 50% reduction plus below 55 mg/dL, the 9-week and 19-week times to target, the 14.6% and 15.4% primary clinical endpoint rates, the 10.2% and 14.3% per-protocol rates with p=0.037 and their hypothesis-generating label, the JAMA publication and ESC Congress 2026 Hot Line presentation, the one-in-five figure for patients not at target at a year, and the quoted phrase on lipid-lowering benefit requiring more time.
- Circulation (VESALIUS-CV prespecified mortality analysis) and JAMA (AMUNDSEN), both dated August 31, 2026 and named as publication venues by the sources above. Neither full text was retrievable at the time of writing, and no figure on this page comes from either.
Frequently Asked Questions
What did VESALIUS-CV actually measure?
It randomized 12,257 people with atherosclerosis or high-risk diabetes and elevated LDL cholesterol, none of whom had a prior heart attack or stroke, to evolocumab or placebo on top of already-optimized lipid-lowering therapy. The mortality result reported at ESC Congress 2026 was a prespecified analysis of death from any cause over a median of 4.6 years. Estimated five-year mortality was 7.9% on evolocumab against 9.7% on placebo.
Is a 20% mortality reduction the same as the 1.8 percentage point difference?
They are the same result expressed two ways. The relative figure is 20% lower all-cause mortality. The absolute figure is 7.9% against 9.7% at five years, a gap of 1.8 percentage points. Relative reductions look larger because they scale to the underlying rate. Both belong in the same view, and the absolute number is the one that describes what changed for the group.
Why did the AMUNDSEN trial not show the same benefit?
AMUNDSEN asked a different question over a much shorter period. It randomized 2,161 people already having a heart attack to immediate in-hospital evolocumab or standard care and followed them for 12 months. Evolocumab got 82% of that group to the LDL target against 40% on standard care, and got them there in 9 weeks rather than 19. Death or unplanned cardiovascular hospitalization at 12 months was 14.6% against 15.4%, which was not a significant difference.
Does hitting an LDL target mean the outcome follows?
Not on its own, and not immediately. AMUNDSEN doubled the share of people reaching a strict LDL goal and recorded no clinical separation inside a year. VESALIUS-CV reported that the mortality curves only began to separate at roughly 1.5 years, in a trial that ran for a median of 4.6. The marker moves in weeks. The evidence that the outcome moved took years to accumulate.
Do these results apply to someone with a normal LDL and no diagnosed disease?
No. Everyone in VESALIUS-CV had established atherosclerosis or high-risk diabetes plus elevated LDL despite existing treatment, and the median age was 66. Everyone in AMUNDSEN was in hospital for a heart attack. Neither trial enrolled people without diagnosed cardiovascular disease or diabetes, so neither describes what a lipid result means for someone outside those groups.
How certain is the mortality finding?
The investigators described it as exploratory. Their stated caution is that the analysis cannot definitively establish that the mortality difference was caused by evolocumab rather than another factor, while noting the results consistently favored the drug. A prespecified analysis in a randomized trial is stronger evidence than an observational comparison, and it is still weaker than a trial designed with mortality as its primary endpoint.