How many patients were enrolled in FOURIER?

27,564 patients with established atherosclerotic cardiovascular disease across 49 countries. All were on background statin therapy with LDL-C of 70 mg/dL or higher at screening.

Did evolocumab reduce cardiovascular death in FOURIER?

No. CV death was not significantly reduced during the median 2.2-year follow-up (HR 1.05; 95% CI 0.88-1.25). The open-label extension study later showed a 23% reduction in CV death with longer exposure.

How much did evolocumab lower LDL cholesterol?

LDL-C dropped 59% from baseline. The median LDL at 4 weeks was 30 mg/dL in the evolocumab arm versus 92 mg/dL in placebo.

Was the FOURIER trial double-blinded?

Yes. Patients and investigators were blinded using matching placebo injections. On-treatment lipid values were centralized and not returned to investigators, though local labs could still measure lipids for clinical care.

Why did FOURIER use a five-component rather than three-component primary endpoint?

Including revascularization and unstable-angina hospitalization increased the event rate, providing more statistical power. The three-component "hard MACE" endpoint was tested as a pre-specified key secondary endpoint and actually showed a larger effect (20% vs. 15% reduction).

How long was follow-up in FOURIER?

Median follow-up was 2.2 years (range approximately 1.8 to 2.5 years). This is shorter than many landmark statin trials, which is a recognized limitation for detecting mortality effects.

Did FOURIER lead to changes in clinical guidelines?

Yes. The 2018 ACC/AHA cholesterol guidelines cited FOURIER as the evidence base for recommending PCSK9 inhibitors (Class IIa) in patients with established ASCVD who have inadequately controlled LDL despite maximally tolerated statin therapy.

What was the cost of evolocumab at the time of FOURIER?

The original list price was approximately $14,000 per year. At that price, cost-effectiveness analyses were unfavorable. Amgen subsequently reduced the price, and insurance coverage has expanded for patients meeting FOURIER-like criteria.

Were there safety concerns with very low LDL levels in FOURIER?

The EBBINGHAUS neurocognitive sub-study found no signal of cognitive impairment at very low LDL levels achieved with evolocumab. Injection-site reactions occurred in 2.1% of evolocumab patients versus 1.6% on placebo.

Inside the FOURIER Methodology: What Most Summaries Skip

Q: What was the primary endpoint of the FOURIER trial?

A five-component composite of cardiovascular death, myocardial infarction, stroke, hospitalization for unstable angina, and coronary revascularization. The key secondary endpoint used a harder three-component composite of CV death, MI, and stroke only.

By HealthRX Medical Team

Published May 25, 2026Updated May 25, 2026Last reviewed May 25, 2026

At a glance

| Parameter | Detail | |---|---| | N | 27,564 | | Intervention | Evolocumab 140 mg Q2W or 420 mg monthly (patient choice) | | Comparator | Matching placebo injections | | Duration | Median 2.2 years (range 1.8 to 2.5) | | Primary endpoint | Composite of CV death, MI, stroke, hospitalization for unstable angina, or coronary revascularization | | Key result | 15% relative risk reduction in primary endpoint (HR 0.85; 95% CI 0.79, 0.92; p <0.001) |

The population FOURIER enrolled, and why it matters

FOURIER recruited patients aged 40 to 85 with clinically evident atherosclerotic cardiovascular disease (ASCVD) and at least one additional risk factor: prior MI, non-hemorrhagic stroke, or symptomatic peripheral artery disease. Every patient had to be on a background statin regimen with an LDL-C of 70 mg/dL or higher (or a non-HDL-C of 100 mg/dL or higher) at screening. This created a high-risk, residual-risk population where the baseline annual event rate was high enough to power the trial within a reasonable timeframe.

Two aspects of the inclusion criteria deserve closer attention. First, the LDL floor of 70 mg/dL excluded patients already at very low LDL levels on statins alone. This biased enrollment toward people with room to benefit from further LDL reduction, which is appropriate for a proof-of-concept outcomes trial but means the result does not directly address whether patients already at LDL <70 on statins gain additional protection. Second, requiring "clinically evident" ASCVD rather than just elevated risk scores ensured the trial tested secondary prevention, not primary prevention. The 2018 ACC/AHA cholesterol guidelines later cited FOURIER as evidence for PCSK9 inhibitor use specifically in this secondary-prevention population, not in broader groups.

Exclusion criteria removed patients with NYHA Class III or IV heart failure, uncontrolled arrhythmias, or planned cardiac surgery. This is standard for CV outcome trials but means the sickest heart failure patients, who account for a large share of CV deaths in real-world practice, were not represented.

Randomization and blinding architecture

Patients were randomized 1:1 to evolocumab or placebo using a central interactive voice/web response system stratified by screening LDL-C (<85 mg/dL vs. 85 mg/dL or higher) and by region. The stratification by LDL-C level is worth noting: it ensured balanced enrollment across the LDL spectrum, which allowed pre-specified subgroup analyses by baseline LDL without worrying about imbalance.

FOURIER used a double-blind, double-dummy design. Patients in the evolocumab arm self-administered subcutaneous injections using either a prefilled autoinjector every two weeks or a monthly SureClick device. Placebo patients received visually identical devices on the same schedules. Patients chose their dosing frequency at randomization, and the split was roughly 75% every-two-weeks and 25% monthly.

The blinding held despite a dramatic LDL reduction. Evolocumab lowered LDL-C by 59% from baseline (median LDL at 4 weeks: 30 mg/dL). In theory, site investigators could have unblinded themselves by checking lipid panels. The trial protocol addressed this by centralizing lipid measurements and not returning on-treatment LDL values to investigators or patients. Local labs could still measure lipids for clinical care, creating a potential source of functional unblinding, but the endpoint adjudication committee operated independently and remained fully blinded.

Defining the primary endpoint: a five-component composite

The primary efficacy endpoint was a five-component composite: cardiovascular death, myocardial infarction, stroke, hospitalization for unstable angina, and coronary revascularization. This is broader than the "hard MACE" three-component composite (CV death, MI, stroke) used in many contemporary CV trials. FOURIER pre-specified the three-component "key secondary endpoint" precisely because of this distinction.

Understanding why the investigators chose a five-component primary endpoint requires a framework for reading composite endpoints in CV trials:

HealthRX Composite-Endpoint Interpretation Framework

| Component | "Hardness" | Susceptible to practice variation? | Driven by biology alone? | |---|---|---|---| | CV death | Hard | Low | Yes | | MI (type 1) | Hard | Low-moderate (troponin thresholds) | Mostly | | Stroke | Hard | Low | Yes | | Hospitalization for unstable angina | Soft | High (admission thresholds vary) | Partially | | Coronary revascularization | Soft | High (operator/site discretion) | No |

Including revascularization and unstable-angina hospitalization increases the event rate (more statistical power, smaller sample size) but dilutes the composite with events that reflect clinical judgment, not just biology. In FOURIER, revascularization accounted for 28% of first primary-endpoint events. If revascularization decisions were influenced by any degree of functional unblinding through local lipid checks, this component is the most vulnerable.

The key secondary endpoint (CV death, MI, stroke) showed a 20% relative risk reduction (HR 0.80; 95% CI 0.73, 0.88; p <0.001). The fact that the "harder" endpoint showed a larger effect than the broader primary endpoint is reassuring. It suggests the benefit was concentrated in biologically firm outcomes rather than being driven by the softer components.

The cardiovascular-death signal: the number that drew scrutiny

The result that generated the most post-publication discussion was the lack of reduction in cardiovascular death. In the evolocumab arm, CV death occurred in 251 patients (1.8%) versus 240 (1.7%) in placebo (HR 1.05; 95% CI 0.88, 1.25). This was not statistically significant and the confidence interval comfortably crossed 1.0.

Multiple explanations exist. The most cited is follow-up duration. At a median of 2.2 years, FOURIER may not have been long enough to capture a mortality benefit. LDL lowering is believed to reduce atherosclerotic plaque burden gradually. The ODYSSEY OUTCOMES trial of alirocumab (another PCSK9 inhibitor), with slightly longer follow-up and an acute coronary syndrome population, showed a nominally significant reduction in all-cause death in a post-hoc analysis.

A second factor is statistical power. FOURIER was powered for the composite primary endpoint, not for individual components. With roughly 490 CV deaths across both arms, the trial was underpowered to detect a plausible 10-15% reduction in CV mortality as a standalone endpoint.

Third, the competing-risk structure of composite endpoints matters. When a patient has a non-fatal MI counted as the first primary event, any subsequent CV death in that patient does not contribute to the primary time-to-first-event analysis. In a trial with substantial non-fatal event reduction, this can mask a mortality signal.

Statistical design and multiplicity control

FOURIER used a hierarchical (gatekeeping) testing procedure. The primary endpoint was tested first at the two-sided alpha = 0.05 level. Only if it was significant would the key secondary endpoint be tested at the same alpha. This preserved the family-wise error rate without requiring p-value adjustment for the two endpoints.

The trial was event-driven, requiring 1,630 primary endpoint events for 90% power to detect a hazard ratio of 0.85. The actual number of primary events was 1,829, giving the trial more than its planned power. The prespecified interim analysis (one planned look at approximately 75% of events) used an O'Brien-Fleming spending function, which preserves nearly all alpha for the final analysis.

The statistical analysis plan pre-specified 29 subgroup analyses, including by age, sex, baseline LDL-C, diabetes status, statin intensity, and qualifying disease type. No formal multiplicity adjustment was applied to subgroup analyses, consistent with standard practice. The consistent direction of effect across subgroups strengthened the overall finding, but individual subgroup p-values should be treated as hypothesis-generating.

One analytical decision that shaped interpretation: FOURIER used an intention-to-treat (ITT) framework for the primary analysis. Adherence was high (about 87% at 1 year), but crossover from placebo to open-label PCSK9 inhibitor therapy was allowed after unblinding. The ITT estimate therefore represents a conservative, diluted measure of the biologic effect of sustained LDL lowering.

Duration, follow-up, and the time-varying benefit curve

The median follow-up of 2.2 years is shorter than landmark statin trials like 4S (5.4 years) and HPS (5.0 years). FOURIER's Kaplan-Meier curves for the key secondary endpoint show progressive separation over time. In the first year, the HR for the three-component endpoint was approximately 0.88. By the second year and beyond, it approached 0.75. This time-dependent pattern is consistent with the "LDL hypothesis": longer exposure to lower LDL produces larger benefit.

The open-label extension study (FOURIER-OLE) followed patients for a median of 5 additional years. Published results showed that patients originally randomized to evolocumab had a persistent 15% lower risk of the key secondary endpoint, and critically, a 23% reduction in cardiovascular death compared to those originally randomized to placebo who later started evolocumab. This longer-term data addresses the main criticism of the parent trial.

The estimand question: what FOURIER actually estimated

Regulatory science increasingly asks: what treatment effect is a trial actually estimating? Under the ICH E9(R1) addendum framework, FOURIER's primary analysis used a treatment-policy estimand. It asked: what is the effect of assigning evolocumab, regardless of whether patients actually took it, discontinued, or switched therapies? This is the ITT answer.

For clinicians making prescribing decisions, the more relevant question might be the while-on-treatment estimand: what happens to patients who actually take the drug as directed? Amgen published supplementary on-treatment analyses showing larger effect sizes (approximately 18-20% for the primary endpoint). These are hypothesis-generating, not confirmatory, because they break randomization.

The distinction matters practically. A physician prescribing evolocumab to a patient who will actually take it every two weeks can reasonably expect a benefit closer to the on-treatment estimate than the ITT estimate. But if adherence in clinical practice is lower than in the trial (injectable therapies often see adherence drop outside trial settings), the real-world effect might fall back toward or below the ITT number.

Known limitations the authors acknowledged

The FOURIER investigators listed several limitations in the primary publication and subsequent commentaries:

Follow-up duration. Too short to assess mortality benefit or long-term safety signals like neurocognitive effects (addressed separately in the EBBINGHAUS sub-study, which found no signal).
Population specificity. Only secondary prevention in established ASCVD. Not generalizable to primary prevention.
Background therapy. All patients were on statins. The trial does not address evolocumab monotherapy.
Injection-site reactions. 2.1% in evolocumab vs. 1.6% in placebo. Modest but not zero.
Cost-effectiveness not addressed. At the original list price (~$14,000/year), the incremental cost per QALY was unfavorable. Subsequent price reductions by Amgen and the evolocumab FDA label broadened access.

What FOURIER changed in practice

The ACC/AHA guidelines now include PCSK9 inhibitors as an option for patients with ASCVD and inadequately controlled LDL despite maximally tolerated statin therapy. The 2018 cholesterol guidelines and their 2022 update specifically cite FOURIER as the evidence base for this recommendation, with a Class IIa designation (reasonable to use) when LDL remains at or above 70 mg/dL.

Insurance coverage has evolved in parallel. Most US payers now cover evolocumab after prior authorization for patients meeting criteria similar to the FOURIER enrollment population: established ASCVD, on statin, LDL still elevated.

Frequently asked questions

›

References

Sabatine MS, Giugliano RP, Keech AC, et al. Evolocumab and Clinical Outcomes in Patients with Cardiovascular Disease. N Engl J Med. 2017;376(18):1713-1722. PubMed
Schwartz GG, Steg PG, Szarek M, et al. Alirocumab and Cardiovascular Outcomes after Acute Coronary Syndrome (ODYSSEY OUTCOMES). N Engl J Med. 2018;379(22):2097-2107. PubMed
Grundy SM, Stone NJ, Bailey AL, et al. 2018 AHA/ACC Cholesterol Clinical Practice Guideline. Circulation. 2019;139(25):e1082-e1143. PubMed
Repatha (evolocumab) prescribing information. Amgen Inc. FDA Label
Scandinavian Simvastatin Survival Study Group. Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the 4S study. Lancet. 1994;344(8934):1383-1389. PubMed