← Clinical Trials
Cardiovascular Phase 3 Completed NCT03872401

VESALIUS-CV: Complete Statistical Analysis of Evolocumab in High Cardiovascular Risk

An independent statistical review of the randomized phase 3 VESALIUS-CV trial evaluating evolocumab versus placebo in patients at high cardiovascular risk without prior myocardial infarction or stroke.

Trial period: 2019-06-11 to 2025-07-25  ·  Enrollment: 12301  ·  Lead sponsor: Amgen
Scope of this record

This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. The numerical results presented here are restricted to the VESALIUS-CV ClinicalTrials.gov record.

1. Trial at a Glance

VESALIUS-CV was a randomized, quadruple-masked, parallel-group phase 3 trial evaluating evolocumab versus placebo in patients at high cardiovascular risk without prior myocardial infarction or stroke and with coronary heart disease. The registry reports 12301 enrolled participants and two primary time-to-event endpoints.

12301
Enrolled
Two treatment groups
2
Arms
Parallel randomized design
0.75
Primary HR
95% CI 0.65–0.86
< 0.001
Primary P-value
Two-sided log-rank
FeatureVESALIUS-CV
Trial nameVESALIUS-CV
PhasePhase 3
ConditionCoronary Heart Disease (CHD)
PopulationPatients at high cardiovascular risk without prior myocardial infarction or stroke
DesignRandomized, parallel, quadruple-masked
AllocationRandomized
Primary purposeTreatment
Enrollment12301
InterventionsEvolocumab; placebo
Primary endpointsTwo binary registry endpoints analyzed as time-to-event outcomes
ResultsPosted on ClinicalTrials.gov
ClinicalTrials.govNCT03872401

2. Clinical Question

The trial addresses whether evolocumab changes the time to first occurrence of major cardiovascular events in patients at high cardiovascular risk without prior myocardial infarction or stroke who were receiving optimized lipid lowering therapy.

Population

Patients at high cardiovascular risk without prior myocardial infarction or stroke and with coronary heart disease.

Intervention

Evolocumab 140 mg Q2W.

Comparator

Placebo Q2W.

Primary question

Does evolocumab reduce the hazard of experiencing the prespecified first cardiovascular event compared with placebo?

3. Trial Design

01
Enroll12301 participants
02
RandomizeTwo parallel groups
03
TreatEvolocumab or placebo Q2W
04
FollowTime-to-event outcomes
05
AnalyzeStratified survival analysis
Allocation
Randomized allocation in a parallel-group design.
Masking
Quadruple masking.
Purpose
Treatment.
Trial status
Completed.
PLACEBO GROUP · 6122 AT RISK FOR REPORTED SERIOUS AE SUMMARY

Placebo Q2W

  • Placebo administered every two weeks.
  • Comparator group for the randomized efficacy analyses.
EVOLOCUMAB GROUP · 6125 AT RISK FOR REPORTED SERIOUS AE SUMMARY

Evolocumab 140 mg Q2W

  • Evolocumab 140 mg administered every two weeks.
  • Intervention group for the randomized efficacy analyses.
Analysis population: The posted primary analyses used the full analysis set (FAS), defined as all randomized participants. Participants were analyzed according to their randomized treatment assignment.

4. Endpoints

The registry lists two primary endpoints. Although the registered endpoint wording is expressed as the number of participants who experienced an event, the posted statistical analyses treat these outcomes as time-to-event endpoints using a log-rank test and a hazard ratio.

Primary endpointRegistered time frameStatistical treatment
Number of Participants Who Experienced Coronary Heart Disease (CHD) Death, Myocardial Infarction (MI), or Ischemic Stroke, Whichever Occurred First From enrollment to last confirmed survival status date; median (min, max) time on trial was 55.2 (0.0, 72.7) months Time-to-event; stratified log-rank test; hazard ratio from a stratified Cox model
Number of Participants Who Experienced CHD Death, MI, Ischemic Stroke, or Any Ischemia-driven Arterial Revascularization, Whichever Occurred First From enrollment to last confirmed survival status date; median (min, max) time on trial was 55.2 (0.0, 72.7) months Time-to-event; stratified log-rank test; hazard ratio from a stratified Cox model

Endpoint adjudication

The registry states that all deaths and individual components were adjudicated by an independent external clinical events committee (CEC), using standardized definitions. This is important because composite cardiovascular endpoints depend on consistent classification of each component event.

5. Statistical Methodology

Full analysis set

The primary efficacy analyses used the full analysis set (FAS), which included all randomized participants. Participants were analyzed according to their randomized treatment assignment. This preserves the treatment comparison created by randomization rather than redefining treatment groups according to what participants subsequently received.

Time-to-event analysis

The registry identifies the primary and posted secondary analyses as time-to-event analyses. The analysis therefore considers not only whether an event occurred, but also when the first qualifying event occurred during follow-up.

Core time-to-event quantity
Time origin → first qualifying event or censoring

Participants who do not experience the event during observable follow-up can contribute information until their last confirmed survival status date. This is one reason survival-analysis methods are appropriate for long follow-up with incomplete event observation.

Stratified log-rank test

The primary analyses used a 2-sided log-rank test stratified by randomization factors collected via interactive voice and web response systems (IVRS/IWRS). Stratification allows the treatment comparison to account for the randomization strata rather than treating the entire randomized population as an unstratified sample.

Stratified Cox model

The hazard ratio was based on a Cox model stratified by the randomization stratification factors collected via IVRS/IWRS. The Cox model provides a relative measure of the instantaneous event rate between randomized treatment groups while accounting for the specified stratification structure.

Hazard-ratio interpretation
HR < 1  →  lower estimated hazard in the evolocumab group

A hazard ratio compares estimated instantaneous event rates under the fitted survival model. It is not the same quantity as a relative risk, an absolute risk reduction, or the proportion of participants who benefit.

Superiority hypothesis

Both primary analyses were identified as superiority analyses. The posted tests were two-sided, so the statistical question was whether the treatment groups differed in either direction under the prespecified survival-analysis framework, with the observed estimates indicating the direction of the difference.

6. Primary Results

Primary Endpoint 1: CHD Death, MI, or Ischemic Stroke

The first primary endpoint counted the first occurrence of CHD death, myocardial infarction, or ischemic stroke. The posted analysis used the FAS, a stratified two-sided log-rank test, and a stratified Cox model.

Hazard ratio for the first primary endpoint

0.75

95% CI: 0.65–0.86   ·   P < 0.001

Comparison: Placebo Q2W vs Evolocumab 140 mg Q2W

FeatureReported result
EndpointCHD death, MI, or ischemic stroke, whichever occurred first
Analysis populationFAS; all randomized participants analyzed according to randomized treatment assignment
Method2-sided stratified log-rank test
Effect measureHazard ratio
Hazard ratio0.75
95% confidence interval0.65–0.86
P-value< 0.001
HypothesisSuperiority
Clinical Biostats interpretation

The reported HR of 0.75 means that, under the stratified Cox model, the estimated instantaneous rate of experiencing the first qualifying event was approximately 25% lower with evolocumab than with placebo. The 25% figure is the direct relative interpretation of 1 − 0.75; it is not an absolute reduction in the probability of an event.

The estimate does not mean that exactly 25% fewer participants experienced an event, nor does it mean that every participant had a 25% reduction in individual risk. A hazard ratio is a relative time-to-event measure.

The 95% CI of 0.65–0.86 describes uncertainty around the estimated hazard ratio under the analysis framework. It does not give a range in which individual treatment effects must fall. The interval is entirely below 1, consistent with the direction of the reported superiority analysis.

The P < 0.001 result addresses evidence against the null hypothesis in the reported statistical test. It does not measure the magnitude of the treatment effect; the hazard ratio and its confidence interval provide the effect estimate and its statistical precision.

Because the result is based on a Cox model, interpretation of a single HR also depends on the model's proportional-hazards framework. The ClinicalTrials.gov record does not report a separate assessment of that assumption, so the HR should be understood as the reported model-based summary rather than as a claim that hazards were constant or proportional at every time point.

Primary Endpoint 2: CHD Death, MI, Ischemic Stroke, or Ischemia-driven Revascularization

The second primary endpoint broadened the composite to include any ischemia-driven arterial revascularization as an additional component, with the first qualifying event determining the endpoint time.

Hazard ratio for the second primary endpoint

0.81

95% CI: 0.73–0.89   ·   P < 0.001

Comparison: Placebo Q2W vs Evolocumab 140 mg Q2W

FeatureReported result
EndpointCHD death, MI, ischemic stroke, or any ischemia-driven arterial revascularization, whichever occurred first
Analysis populationFAS; all randomized participants analyzed according to randomized treatment assignment
Method2-sided stratified log-rank test
Effect measureHazard ratio
Hazard ratio0.81
95% confidence interval0.73–0.89
P-value< 0.001
HypothesisSuperiority
Clinical Biostats interpretation

The reported HR of 0.81 corresponds to an estimated 19% lower instantaneous rate of the composite event in the evolocumab group relative to placebo under the fitted Cox model. As with the first endpoint, this is a relative hazard interpretation, not a statement that 19% of participants avoided an event.

The 95% CI of 0.73–0.89 quantifies uncertainty around the estimated HR. It remains below 1 throughout the interval, while the width of the interval indicates that the estimate is not an exact measure of the underlying treatment effect.

The P < 0.001 value indicates strong evidence against the null hypothesis under the reported two-sided log-rank test. It should not be read as a 0.1% probability that the treatment effect is real, nor as a measure of clinical magnitude.

The broader composite includes arterial revascularization. Composite endpoints should therefore be interpreted as the endpoint as defined, rather than assuming that the hazard ratio applies identically to every individual component.

Educational note: the ClinicalTrials.gov record provides hazard ratios, confidence intervals, and P-values but do not provide Kaplan-Meier coordinates or sufficient event/censoring data to reconstruct a valid curve. No fabricated Kaplan-Meier curve is presented here.

7. Secondary Endpoint Results

The registry also posts ten secondary statistical analyses. They use the same general analytical framework: the FAS, randomized treatment assignment, a two-sided stratified log-rank test, and a stratified Cox model for the hazard ratio.

Secondary endpointHR95% CIP-value
MI, ischemic stroke, or any ischemia-driven arterial revascularization0.790.72–0.88< 0.001
CHD death, MI, or any ischemia-driven arterial revascularization0.790.72–0.88< 0.001
Cardiovascular death, MI, or ischemic stroke0.730.64–0.84< 0.001
CHD death or MI0.730.62–0.87< 0.001
MI0.640.52–0.79< 0.001
Any ischemia-driven arterial revascularization0.790.70–0.88< 0.001
CHD death0.890.68–1.16= 0.39
Cardiovascular death0.790.64–0.98= 0.031
Death due to any cause0.800.70–0.91< 0.001
Ischemic stroke0.790.62–1.01= 0.062

How the secondary estimates should be read

The secondary estimates are not interchangeable. Each hazard ratio applies to a different endpoint definition, so the numerical value must always be interpreted together with the event definition and follow-up period.

For example, the HR of 0.64 for MI represents the reported relative hazard for myocardial infarction specifically, whereas the HR of 0.80 for death due to any cause represents a different time-to-event outcome. A smaller HR does not automatically mean that the endpoint is more clinically important or that the underlying absolute difference is larger.

Multiplicity matters: the ClinicalTrials.gov record identifies the analyses as superiority analyses but do not provide a multiplicity-adjustment hierarchy or alpha allocation for the complete set of secondary endpoints. The individual P-values should therefore be interpreted as the reported statistical results for their respective analyses, without assuming that every secondary comparison represents an independently multiplicity-controlled confirmatory test.

8. Individual Secondary Outcomes in More Detail

Myocardial infarction

MI hazard ratio

0.64

95% CI: 0.52–0.79   ·   P < 0.001

The MI analysis has an estimated hazard ratio of 0.64. In relative terms, this corresponds to an estimated 36% lower instantaneous hazard of myocardial infarction in the evolocumab group under the reported model. The confidence interval, 0.52–0.79, represents the statistical uncertainty around that estimate.

Cardiovascular death

Cardiovascular-death hazard ratio

0.79

95% CI: 0.64–0.98   ·   P = 0.031

The reported HR of 0.79 corresponds to an estimated 21% lower instantaneous hazard of cardiovascular death in the evolocumab group under the fitted model. The 95% CI extends from 0.64 to 0.98, showing greater uncertainty than some of the composite endpoint estimates.

CHD death

CHD-death hazard ratio

0.89

95% CI: 0.68–1.16   ·   P = 0.39

The estimated HR of 0.89 is below 1, but the 95% CI of 0.68–1.16 includes 1. The reported P-value is 0.39. This illustrates why an estimated HR below 1 should not automatically be described as statistically demonstrated evidence of a treatment effect: the confidence interval and hypothesis test must be considered together.

Ischemic stroke

Ischemic-stroke hazard ratio

0.79

95% CI: 0.62–1.01   ·   P = 0.062

The estimated HR of 0.79 corresponds to a 21% lower estimated instantaneous hazard under the fitted model, but the 95% CI of 0.62–1.01 includes 1 and the reported P-value is 0.062. The appropriate statistical description is therefore the reported estimate together with its uncertainty and P-value, rather than converting the point estimate into a definitive conclusion about the endpoint.

9. Time-to-Event Statistics Explained

Why use time-to-event analysis for these endpoints?

Each primary endpoint is defined by the first occurrence of one of several cardiovascular events and is followed from enrollment to the last confirmed survival status date. Participants can therefore have different lengths of observable follow-up. Time-to-event methods are designed for precisely this structure because they incorporate both event timing and censored follow-up.

Kaplan-Meier concept
S(t) = estimated probability of remaining event-free beyond time t

The Kaplan-Meier estimator updates the estimated event-free survival whenever an event occurs while retaining information from participants who remain under observation without experiencing the event.

Why does the analysis use the first event?

The primary endpoints specify that the qualifying event occurring whichever occurred first determines the endpoint. This converts several possible clinical events into a single time-to-first-event outcome. Once the first qualifying event occurs, that participant has experienced the endpoint regardless of whether additional qualifying events occur later.

Why is the Cox model used with the log-rank test?

The log-rank test provides a statistical comparison of the time-to-event distributions between randomized groups. The Cox model supplies a hazard ratio that summarizes the relative event rate between groups. In VESALIUS-CV, both are stratified according to the randomization factors collected through IVRS/IWRS.

10. Statistical Methods Explained

What does a hazard ratio of 0.75 mean?

An HR of 0.75 means that the estimated instantaneous rate of the endpoint is 0.75 times the corresponding rate in the comparator group under the fitted Cox model. Equivalently, 1 − 0.75 = 0.25, so the model-based relative hazard is estimated to be 25% lower. This is not the same as saying that 25% of participants avoided the event.

Why is the confidence interval important?

The point estimate alone does not communicate statistical precision. The 95% CI of 0.65–0.86 for the first primary endpoint describes the uncertainty surrounding the estimated HR. A relatively narrow interval provides more precision than a very wide interval, although precision should always be considered alongside the clinical context and endpoint definition.

Why does the P-value not measure effect size?

A P-value measures the evidence against a specified null hypothesis under the statistical test. It depends on both the magnitude of the observed difference and the amount of information in the analysis. Consequently, a very small P-value can occur with a modest effect when the sample contains substantial information, while a larger P-value can occur for an effect estimate with considerable uncertainty.

Why was a stratified log-rank test used?

The registry states that the log-rank test was stratified by randomization factors collected via IVRS/IWRS. Stratification preserves the structure used when assigning participants to treatment groups and allows the comparison of event-time distributions within the specified strata rather than ignoring those factors.

What does the Cox model add?

The stratified Cox model converts the treatment comparison into a hazard-ratio estimate with a confidence interval. The model is particularly useful because it expresses the relative event rate on a common scale while retaining the time-to-event structure and the prespecified stratification.

Why can a composite endpoint differ from one of its components?

A composite endpoint combines several possible first events. Its hazard ratio reflects the combined endpoint as defined. A treatment can therefore have one hazard ratio for a composite of CHD death, MI, and ischemic stroke and a different hazard ratio for MI alone. The component and composite analyses answer different questions.

What does a confidence interval crossing 1 mean for a hazard ratio?

For a hazard ratio, 1 represents equal estimated hazard between groups. A confidence interval that includes 1 indicates that equality remains within the interval of values compatible with the statistical uncertainty at that confidence level. In VESALIUS-CV, the CHD-death interval of 0.68–1.16 and ischemic-stroke interval of 0.62–1.01 illustrate this distinction.

11. Follow-up and Trial Timing

2019-06-11

Trial start

The registry lists June 11, 2019 as the study start date.

2025-07-25

Primary completion

The registry lists July 25, 2025 as the primary completion date.

Completed

Registry status

The ClinicalTrials.gov record identifies the trial as completed and reports statistical results for 12 outcome measures.

The endpoint time frame is from enrollment to the last confirmed survival status date. The registry reports a median time on trial of 55.2 months, with a minimum of 0.0 months and a maximum of 72.7 months.

Important distinction: the median time on trial is a follow-up descriptor. It is not a median time to an endpoint and should not be interpreted as such.

12. Safety Results

The ClinicalTrials.gov record reports serious adverse events by randomized treatment arm using affected participants divided by participants at risk.

Safety measurePlacebo Q2WEvolocumab 140 mg Q2W
Serious adverse events, affected / at risk1772 / 61221726 / 6125
Participants affected by reported serious adverse events
Placebo Q2W
1772
Evolocumab 140 mg Q2W
1726

The serious-adverse-event figures should be read as affected participants divided by participants at risk, exactly as reported in the ClinicalTrials.gov record. They are not the same statistical quantity as the time-to-event efficacy hazard ratios and should not be compared by simply subtracting the two affected-participant counts.

Analysis-set distinction: the registry-reported efficacy analyses use the FAS and randomized treatment assignment, whereas the serious-adverse-event summary is presented by treatment arm with its own affected/at-risk denominators. The ClinicalTrials.gov record does not provide a formal hypothesis test or hazard ratio for serious adverse events.

13. Randomization and Stratified Analysis

Randomization is the central design feature supporting the treatment comparison. Once participants are randomized, the treatment groups can be compared without requiring the statistical analysis to explain every baseline characteristic individually. The registry-reported primary analyses retain that randomized assignment through the FAS.

The survival analyses additionally use the randomization stratification factors collected through IVRS/IWRS. The purpose of stratification is not to change the treatment groups but to incorporate the design's stratification structure into the analysis.

Randomization

Creates the basis for comparing treatment groups under the randomized design.

FAS analysis

Includes all randomized participants and analyzes them according to randomized treatment assignment.

Stratified log-rank

Compares time-to-event outcomes while accounting for the randomization stratification factors.

Stratified Cox model

Provides the hazard ratio and confidence interval using the same stratification structure.

14. Understanding the Two Primary Endpoints Together

The two primary endpoints are related but not identical. The first includes CHD death, MI, or ischemic stroke. The second adds any ischemia-driven arterial revascularization to those events.

Primary endpointHR95% CIP-value
CHD death, MI, or ischemic stroke0.750.65–0.86< 0.001
CHD death, MI, ischemic stroke, or any ischemia-driven arterial revascularization0.810.73–0.89< 0.001

The HRs are not expected to be identical because the endpoints contain different event definitions. Adding revascularization creates a broader composite and changes the population's observed time-to-first-event distribution.

Both reported estimates are below 1 and both have two-sided P-values below 0.001. The statistical interpretation should nevertheless remain endpoint-specific: the first result concerns the narrower composite, while the second concerns the broader composite.

15. What the Hazard Ratio Does — and Does Not — Mean

Relative effect

An HR of 0.75 means the estimated instantaneous hazard under the model is 0.75 times the comparator hazard. This corresponds to a 25% lower estimated hazard relative to the comparator.

Not an absolute risk difference

The hazard ratio does not tell us how many percentage points the cumulative probability of an event changed. Absolute risk requires the underlying event probabilities or survival estimates at a specified time point.

Not a probability of benefit

An HR does not mean that a particular percentage of individual participants benefited. Treatment effects vary across individuals, while the HR is a population-level model-based summary.

Confidence interval

The 95% confidence interval describes statistical uncertainty around the estimated HR. For the first primary endpoint, 0.65–0.86 indicates that the point estimate of 0.75 should not be treated as exact.

P-value

The P-value addresses the hypothesis test under the reported statistical framework. It does not quantify the magnitude of the observed treatment effect and should be interpreted alongside the HR and confidence interval.

16. Limitations

Important registry caveat: the exclusion of 44 enrolled participants from all efficacy and safety analyses means that the enrollment total of 12301 should not automatically be treated as the denominator for every reported analysis. The primary analyses instead specify the FAS of randomized participants, while the safety summary supplies its own at-risk denominators.

17. Why This Trial Matters Statistically

VESALIUS-CV is a useful teaching case because it illustrates how a large randomized cardiovascular trial can combine binary registry endpoint labels with formal time-to-event methodology. The statistical story is not contained in the event definition alone: the analysis depends on follow-up time, censoring, stratification, the log-rank comparison, and the Cox hazard-ratio model.

ConceptHow it appears in VESALIUS-CV
RandomizationRandomized parallel-group phase 3 design
BlindingQuadruple masking
FAS analysisAll randomized participants analyzed according to randomized treatment assignment
Time-to-event endpointsPrimary and secondary cardiovascular outcomes analyzed according to time from enrollment to event or censoring
Log-rank testTwo-sided comparison stratified by IVRS/IWRS randomization factors
Hazard ratioPrimary effect measure for treatment comparisons
Stratified Cox modelUsed to obtain the reported hazard ratios and confidence intervals
Confidence intervalsQuantify uncertainty around the reported HR estimates
Composite endpointsPrimary outcomes combine several cardiovascular event types
MultiplicityMultiple primary and secondary analyses require careful interpretation of P-values
Safety analysisSerious adverse events reported as affected participants over participants at risk by arm

18. Related Tutorials

Learn more about the methods used in this trial:

19. Related Calculators

20. Sources

Continue through the Clinical Biostats statistical pathway

Explore the underlying survival-analysis concepts and statistical tools used to understand randomized time-to-event trials.

21. Record Summary

VESALIUS-CV provides a detailed example of randomized time-to-event analysis in a large cardiovascular trial. The ClinicalTrials.gov record uses the full analysis set, randomized treatment assignment, two-sided stratified log-rank testing, and stratified Cox models to estimate hazard ratios with 95% confidence intervals. Both primary endpoints produced reported HRs below 1 with P-values below 0.001, while the secondary analyses illustrate why endpoint definition, confidence intervals, and multiplicity considerations matter when interpreting a large set of cardiovascular outcomes.

The most useful statistical reading of the record combines the hazard ratio, confidence interval, P-value, endpoint definition, analysis population, follow-up structure, and stratification strategy. The reported serious-adverse-event counts provide a separate safety perspective and should not be treated as another form of the efficacy hazard-ratio analysis.

Clinical Biostats methodology: A trial-results page should distinguish the registry's reported evidence from statistical interpretation. For VESALIUS-CV, that means preserving the registered endpoint definitions and reported estimates while explaining what the survival-analysis methods do, what the hazard ratios mean, and what cannot be concluded from the registry-level information.