Honest Criticisms and Limitations of the FOURIER Trial

At a glance

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

The Follow-Up Problem

The single most common criticism of FOURIER is its short observation window. A median follow-up of 2.2 years is brief for a trial designed to answer whether aggressive LDL lowering (to a median of 30 mg/dL) translates into fewer deaths and cardiovascular events long-term. Statin mega-trials like 4S and HPS ran 5+ years precisely because atherosclerosis regression takes time.

The trial's own Kaplan-Meier curves tell part of the story. Separation between evolocumab and placebo arms was modest in year one and widened through year two, suggesting that a longer study might have amplified the benefit. But it also might have introduced confounders: real-world medication switches, open-label PCSK9 inhibitor uptake in the placebo arm, and differential dropout.

Several editorialists noted that the short duration likely explains why the key secondary endpoint of CV death, MI, or stroke showed a 20% reduction while the broader primary composite (which dilutes with softer events) showed only 15%. In a longer trial, hard endpoints accumulate and the signal may have sharpened, or it may have plateaued. We do not know.

No Mortality Signal

FOURIER did not demonstrate a reduction in cardiovascular death (HR 1.05, 95% CI 0.88-1.25) or all-cause mortality (HR 1.04, 95% CI 0.91-1.19). The trial was not powered for mortality, but the complete absence of a trend troubled some observers.

Critics raised two separate concerns:

1. Statistical power: With 844 deaths in the overall cohort, a mortality benefit <10% would not reach significance. The investigators acknowledged this openly in the supplementary appendix.
2. Biological plausibility concern: Some letters to the editor questioned whether pharmacological LDL reduction to 30 mg/dL carries unforeseen risks (neurocognitive, hormonal) that offset part of the atherosclerotic benefit. The EBBINGHAUS substudy addressed cognition specifically, finding no signal over 19 months, but again, the time window was limited.

The 2018 ACC/AHA cholesterol guidelines ultimately incorporated FOURIER's data for high-risk secondary prevention despite the mortality null, judging that MACE reduction alone justified the recommendation. Whether that judgment was premature remains debated among preventive cardiologists.

Enrollment Biases and Generalizability Gaps

Age and Sex

The mean age was 62.5 years, and 75% of participants were male. Women and adults over 75, who carry the highest absolute ASCVD event rates, were underrepresented. Subgroup analyses did not show heterogeneity by sex, but the confidence intervals for women were wide enough to be uninformative.

Geographic Distribution

Over 50% of enrollees came from Europe. Roughly 5% were Black, limiting generalizability to populations with higher cardiovascular risk and different lipid metabolism profiles.

Background Therapy

Enrollment required stable statin therapy, but only 69% were on high-intensity statins. This matters because a patient already on atorvastatin 80 mg with LDL of 92 mg/dL represents a different treatment question than one on simvastatin 20 mg with LDL of 130 mg/dL. The composite result blends both scenarios without stratification in the primary analysis.

Excluded Populations

The protocol excluded patients with NYHA class III-IV heart failure, uncontrolled arrhythmia, and recent (<4 weeks) MI or stroke. Excluding the sickest patients improves internal validity but limits the clinical scenarios where the drug is most frequently considered.

The Composite Endpoint Debate

| Component | Evolocumab (n) | Placebo (n) | HR (95% CI) | |-----------|---------------|-------------|-------------| | CV death | 251 | 240 | 1.05 (0.88-1.25) | | MI | 468 | 639 | 0.73 (0.65-0.82) | | Stroke | 207 | 262 | 0.79 (0.66-0.95) | | Unstable angina hospitalization | 239 | 255 | 0.94 (0.79-1.12) | | Coronary revascularization | 759 | 965 | 0.78 (0.71-0.86) |

The primary composite bundled five events of unequal clinical weight. CV death and MI are "hard" endpoints; coronary revascularization is a physician decision influenced by practice patterns, insurance coverage, and catheterization lab availability. In FOURIER, revascularization drove the largest absolute event count in both arms.

When critics stripped the composite to CV death + MI + stroke alone (the key secondary endpoint), evolocumab showed a stronger 20% relative reduction (HR 0.80, p <0.001). This paradox, a "purer" result emerging from a narrower endpoint, suggests the five-component primary diluted the treatment effect with noise from softer outcomes.

Statistical Considerations

Multiplicity and Hierarchical Testing

The trial used a hierarchical testing procedure: the primary endpoint had to be significant before the key secondary could be formally tested. Both passed. But this sequential gatekeeping meant other subgroup analyses (by LDL quartile, by statin intensity, by time from qualifying event) were exploratory and not alpha-controlled.

Event-Driven Design Shortcomings

FOURIER was event-driven, targeting 1,630 primary events. The trial reached this threshold faster than anticipated, driven partly by the revascularization component. An event-driven design optimized for statistical power does not guarantee clinical meaningfulness of the aggregate endpoint.

Absolute Risk Reduction Context

The 15% relative reduction in the primary composite translated to an absolute risk reduction of approximately 1.5% over 2.2 years (11.3% vs. 9.8%). The number needed to treat (NNT) was roughly 67 over 26 months, or about 74 per year. For a drug costing over $14,000 annually at launch (before subsequent price reductions), the cost-effectiveness ratio raised concerns among health economists.

Industry Sponsorship and Conflict of Interest

Amgen funded FOURIER, employed the statistical team, and held the database. The trial publication disclosed that Amgen employees participated in study design, data collection, and analysis. The academic steering committee had access to the data, but the extent of independent verification was questioned in subsequent correspondence.

Specific concerns raised in published letters:

• The decision to include coronary revascularization in the primary composite (favoring a positive result) was made before unblinding, but critics noted it was also before efficacy data from OSLER-1 and OSLER-2 suggested exactly this pattern of benefit.
• Several steering committee members disclosed consulting fees from Amgen totaling six figures in the 36 months surrounding publication.
• The supplementary appendix, which contained the most detailed event adjudication data, was available only as an online-only supplement, complicating independent scrutiny.

None of these facts imply misconduct. They do, taken together, represent a standard industry-academic collaboration model that some observers believe creates structural optimism bias in trial design and interpretation.

What Post-Publication Commentary Surfaced

Letters to the NEJM (2017)

Multiple correspondence letters challenged the composite endpoint choice, the follow-up duration, and the absence of mortality benefit. The investigators' reply emphasized that the key secondary endpoint (CV death, MI, stroke) showed a consistent and stronger effect, supporting biological plausibility.

The Open-Label Extension Problem

FOURIER's open-label extension (FOURIER-OLE) allowed all participants access to evolocumab. By design, this eliminates the randomized comparison beyond 2.2 years. The OLE data published subsequently showed durable LDL lowering and suggested continued event reduction, but without a concurrent placebo arm, these data cannot rule out secular trends, survivor bias, or healthy-adherer effects.

Comparison with ODYSSEY OUTCOMES

The ODYSSEY OUTCOMES trial of alirocumab (a different PCSK9 inhibitor) enrolled post-ACS patients, ran longer (median 2.8 years), and showed a hint of mortality benefit in a pre-specified subgroup with baseline LDL above 100 mg/dL. This comparison raised questions about whether FOURIER's population selection (stable ASCVD, lower baseline LDL) handicapped the mortality signal.

Practical Clinical Translation

Despite these limitations, professional societies incorporated FOURIER into practice recommendations. The 2018 AHA/ACC Multisociety Cholesterol Guideline recommends considering PCSK9 inhibitors for patients with clinical ASCVD at very high risk whose LDL remains above 70 mg/dL on maximally tolerated statin plus ezetimibe.

The Repatha (evolocumab) prescribing information gained a cardiovascular risk reduction indication based on FOURIER. The FDA's approval letter noted the MACE benefit was driven by MI and revascularization rather than mortality, a distinction reflected in the label language.

Clinicians prescribing evolocumab should understand: the evidence supports fewer heart attacks and strokes in secondary prevention, not longer life. Whether longer treatment durations change that equation remains an open question that FOURIER, by its design, cannot answer.

Frequently asked questions

›

›

›

›

›

›

›

›

›

›

References

1. 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
2. Grundy SM, Stone NJ, Bailey AL, et al. 2018 AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA Guideline on the Management of Blood Cholesterol. J Am Coll Cardiol. 2019;73(24):e285-e350. PubMed
3. 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
4. Repatha (evolocumab) Prescribing Information. Amgen Inc. Revised 2017. FDA Label
5. Giugliano RP, Mach F, Zavitz K, et al. Cognitive Function in a Randomized Trial of Evolocumab (EBBINGHAUS). N Engl J Med. 2017;377(7):633-643. PubMed