This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
FOURIER-OLE was a phase 3, single-group, open-label extension study of evolocumab in participants with dyslipidemia. The study enrolled 5035 participants and evaluated adverse events during the OLE study period, with the primary endpoint defined over up to 5 years.
| Feature | FOURIER-OLE |
|---|---|
| Study title | Further Cardiovascular Outcomes Research With PCSK9 Inhibition in Subjects With Elevated Risk Open-label Extension |
| Phase | Phase 3 |
| Condition | Dyslipidemia |
| Design model | Single-group |
| Allocation | NA |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 5035 |
| Number of arms | 1 |
| Intervention | Evolocumab (biological) |
| Status | Completed |
| Start date | 2016-09-02 |
| Primary completion date | 2022-03-21 |
| Lead sponsor | Amgen |
| Sponsor type | Industry |
| ClinicalTrials.gov | NCT02867813 |
2. Clinical Question
The central clinical question represented by this open-label extension is how adverse events are characterized during long-term exposure to evolocumab in participants with dyslipidemia. Unlike a parallel-group randomized efficacy trial, the FOURIER-OLE registry record describes a single-group extension design, so the primary endpoint is an adverse-event outcome rather than a between-group efficacy comparison.
Population
Participants enrolled in the FOURIER-OLE phase 3 open-label extension with the registry condition of dyslipidemia.
Intervention
Evolocumab, classified in the registry as a biological intervention.
Comparator
The study is registered as a single-group design with no masking and no separate randomized comparator arm.
Primary question
How many participants experience an adverse event during the OLE study period, with the registry specifying a time frame of up to 5 years?
3. Trial Design
Evolocumab
- Evolocumab
- Biological intervention
- Open-label study
- Primary purpose: treatment
Interpretive consequence
- The registry does not define a randomized control arm for the OLE.
- The primary endpoint is an adverse-event count rather than a randomized treatment contrast.
- Between-group causal comparisons are therefore not the principal statistical structure of this study.
4. Primary Endpoint
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Number of Participants Who Experienced an Adverse Event | All adverse event summaries for the primary analysis of the primary endpoint (OLE study period only) included all treatment-emergent events reported on the Event electronic case report form (eCRF), including CEC positively reviewed events and disease-related events. | Up to 5 years |
The endpoint is explicitly participant-based: it asks for the number of participants who experienced an adverse event. The registry definition also specifies that the primary analysis concerns the OLE study period and includes treatment-emergent events reported on the Event electronic case report form.
The inclusion of CEC positively reviewed events and disease-related events is important statistically because the endpoint is broader than a single narrowly defined adverse-event category. The event collection framework determines which observations enter the numerator of an adverse-event summary.
5. Statistical Methodology
Descriptive adverse-event analysis
For a participant-level adverse-event endpoint in a single-group extension study, the natural primary summary is descriptive: the number and proportion of participants experiencing at least one qualifying adverse event during the specified observation period. Event-level summaries can provide additional information about the number and type of events, but the registry's primary endpoint is explicitly defined at the participant level.
The numerator represents participants with at least one qualifying treatment-emergent event during the specified OLE period. The denominator must correspond to the analysis population used for the registry's primary endpoint.
Why participant counts matter
Counting participants rather than simply counting adverse-event records avoids treating repeated events in the same participant as if they represented independent people. One participant may have multiple adverse-event records while still contributing only once to a binary endpoint defined as "experienced an adverse event."
Event-level versus participant-level summaries
Clinical safety databases commonly contain multiple events per participant. An event-level table answers questions about the volume or categories of reported events, whereas a participant-level analysis answers whether each participant experienced at least one event. These are different estimands and should not be substituted for one another.
Time frame and exposure
The registry specifies a time frame of up to 5 years. A long observation window creates an important distinction between the probability of experiencing an event during a defined period and an instantaneous event rate. Participants followed for different lengths of time may have different opportunities to experience an event.
6. Statistical Methods Explained
Why is the primary endpoint participant-based?
The endpoint is the number of participants who experienced an adverse event. This makes the basic outcome binary at the participant level: a participant either experienced at least one qualifying event during the specified period or did not. This avoids giving a participant with many repeated events disproportionate weight in the primary participant-count endpoint.
Why is a single-group design important?
FOURIER-OLE is registered as a single-group study. In a randomized two-arm trial, an adverse-event rate can be compared directly between treatment groups while preserving the benefits of randomization. In a single-group extension, the primary descriptive question is instead how frequently events occur among participants receiving the extension intervention.
What does "treatment-emergent" mean statistically?
The registry specifies treatment-emergent events for the primary endpoint. Conceptually, this identifies events according to their relationship to the treatment period defined by the study rather than simply counting every medical event in a participant's lifetime. The precise operational definition is determined by the study's event collection procedures.
Why does the denominator matter?
An adverse-event proportion depends on both the number of participants with events and the population included in the denominator. Using the enrolled population, treated population, or another safety-analysis population can produce different percentages. A participant-level safety estimate is therefore meaningful only when its analysis population is clearly specified.
Why can a five-year safety summary differ from an incidence rate?
A five-year participant proportion describes whether an event occurred during a period extending up to 5 years. An incidence rate incorporates time at risk into the denominator and is therefore sensitive to differences in follow-up duration. The two measures answer related but distinct questions.
Why are repeated adverse events not the same as repeated participants?
If one participant experiences several adverse events, an event-level database may contain several records while a participant-level endpoint still counts that person once. This distinction is essential when interpreting safety tables and prevents the number of event records from being mistaken for the number of affected participants.
7. Reported Results
The registry indicates that results were posted. The primary endpoint is defined as the number of participants who experienced an adverse event during the OLE study period, but the registry information available for this analysis does not provide a numerical estimate, confidence interval, or p-value for that primary endpoint.
Serious Adverse Events by Parent-Study Treatment
The registry reports serious adverse-event counts by treatment received in the parent study. These figures describe parent-study treatment groups and should not be interpreted as randomized treatment groups within the single-group FOURIER-OLE design.
| Parent-study treatment | Participants with serious adverse events | Participants at risk |
|---|---|---|
| Placebo in Parent Study | 1077 | 2532 |
| Evolocumab in Parent Study | 1045 | 2499 |
The reported counts show that serious adverse events occurred among participants entering the extension from both parent-study treatment groups. The figures are presented as affected participants over participants at risk, rather than as an effect estimate comparing randomized FOURIER-OLE treatment groups.
The difference between 1077/2532 and 1045/2499 should not be treated as a randomized treatment effect for FOURIER-OLE. The parent-study treatment categories describe participants' prior treatment assignment, while the OLE itself is registered as a single-group study.
These counts also do not by themselves provide a measure of statistical precision. A confidence interval would quantify uncertainty around a specified parameter, while a p-value would address a prespecified hypothesis under a defined statistical model. Neither is reported here for these safety counts.
8. Safety Analysis
Safety is the central statistical domain represented by the registry's primary endpoint. The endpoint covers all treatment-emergent adverse events reported on the Event electronic case report form during the OLE study period, including CEC positively reviewed events and disease-related events.
| Safety component | Registry information |
|---|---|
| Primary safety endpoint | Number of Participants Who Experienced an Adverse Event |
| Observation period | Up to 5 years |
| Event source | Event electronic case report form (eCRF) |
| Event scope | All treatment-emergent events for the primary analysis |
| CEC-reviewed events | Included when positively reviewed by the CEC |
| Disease-related events | Included |
| Serious adverse events by parent-study treatment | 1077/2532 for Placebo in Parent Study; 1045/2499 for Evolocumab in Parent Study |
The serious adverse-event figures illustrate another important safety-analysis principle: the analysis population and treatment history determine the interpretation of a count. Because the registry labels these groups according to parent-study treatment, they should not be recast as concurrent FOURIER-OLE comparator arms.
9. What the Safety Endpoint Does — and Does Not — Mean
The primary endpoint asks whether a participant experienced at least one qualifying adverse event during the OLE study period. It therefore summarizes the occurrence of participants with events rather than the severity, frequency, duration, or clinical importance of every individual event.
A participant who experiences multiple adverse events can still contribute only once to a participant-level "experienced an adverse event" endpoint.
An adverse-event count alone does not establish causality. It also does not establish comparative treatment risk because FOURIER-OLE is registered as a single-group study. The endpoint does not by itself distinguish mild events from serious events unless the analysis separately defines and reports those categories.
The endpoint has a time frame of up to 5 years. Longer observation generally provides more opportunity for an event to occur, so the duration of follow-up is an essential part of interpreting a participant-level safety proportion.
10. Analysis Population and Denominators
The registry reports enrollment of 5035 participants. That number establishes the overall study enrollment, but a primary safety analysis can use a more specific analysis population depending on treatment exposure and evaluability.
Enrollment
5035 participants were enrolled in the study.
Study structure
The registry identifies 1 arm and a single-group design.
Safety denominator
The appropriate denominator for a participant-level adverse-event proportion is the population defined for that safety analysis.
Parent-study counts
The serious adverse-event denominators are 2532 and 2499 for the two parent-study treatment-history categories.
These distinctions are important because denominators answer the question "out of whom?" A safety percentage without its corresponding analysis population can be difficult to interpret, particularly in a long-term extension where participants may have differing exposure and follow-up.
11. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The registry defines a participant-level adverse-event endpoint over up to 5 years in a single-group extension. The principal statistical description is therefore the frequency of participants experiencing qualifying treatment-emergent events within the specified analysis population.
Clinical interpretation
The safety record is intended to characterize adverse events during long-term exposure in the extension population. Clinical meaning depends on event definitions, seriousness, severity, timing, exposure, and the population represented by each summary.
12. Why a Single-Group Extension Changes the Statistical Question
A randomized controlled trial and an open-label extension can address different statistical questions even when they involve the same therapeutic intervention.
| Feature | Randomized parallel-group comparison | FOURIER-OLE registry structure |
|---|---|---|
| Allocation | Typically assigns participants between treatment groups | NA |
| Design model | Between-group comparison | Single-group |
| Masking | May be blinded | None |
| Primary statistical contrast | Treatment versus comparator | Participant-level safety occurrence |
| Interpretive emphasis | Relative or absolute treatment effect | Descriptive safety experience during the extension |
The absence of a concurrent randomized comparator is not merely a formatting detail. It changes the estimand. A single-group extension is suited to describing outcomes among participants exposed in the extension, while a randomized trial is structured to estimate a causal contrast between randomized strategies.
13. Serious Adverse Events and Parent-Study Treatment History
The registry provides serious adverse-event counts separately for participants who received placebo or evolocumab in the parent study.
The chart is a visual display of the reported participant counts, not an adjusted treatment-effect analysis. The corresponding denominators are different: 2532 participants for the placebo parent-study group and 2499 for the evolocumab parent-study group.
14. Long-Term Safety and Censoring
A five-year extension raises a fundamental time-to-event issue: not every participant necessarily contributes the same amount of observation time. Participants may enter or leave the risk set at different times, and a participant who remains event-free through the end of available follow-up contributes information through that point.
If the scientific question were the time until a first adverse event, a time-to-event analysis could use Kaplan-Meier estimation or an appropriate incidence-rate framework. Such an analysis would require event times, censoring times, and a prespecified definition of the event and risk period.
The registry's primary endpoint is stated as a participant count rather than a posted time-to-event estimate. A time-to-event analysis would therefore represent a different statistical summary from the primary participant-count endpoint.
15. Why the Event Definition Matters
The registry specifies that all adverse-event summaries for the primary analysis include all treatment-emergent events reported on the Event eCRF. This definition determines the population of observations considered eligible for the primary endpoint.
Treatment-emergent events
The endpoint is restricted to treatment-emergent events within the OLE study period.
Event eCRF
The registry identifies the Event electronic case report form as the reporting source for the primary analysis.
CEC review
CEC positively reviewed events are included in the adverse-event summaries.
Disease-related events
Disease-related events are included in the endpoint's specified event scope.
For statistical interpretation, endpoint definitions are part of the estimand. Changing the event definition can change the numerator, the denominator, or both. Consequently, two safety percentages cannot be assumed comparable unless their event definitions and analysis populations are aligned.
16. Trial Timeline
Study start
The registry lists 2016-09-02 as the FOURIER-OLE start date.
Open-label extension
The study is registered as a phase 3, single-group, unmasked treatment study of evolocumab in dyslipidemia.
Primary endpoint time frame
The primary endpoint is the number of participants who experienced an adverse event over a time frame of up to 5 years.
Primary completion
The registry lists 2022-03-21 as the primary completion date.
Registry status
The current study record identifies FOURIER-OLE as completed.
17. What the Serious Adverse-Event Counts Do — and Do Not — Mean
The values 1077/2532 and 1045/2499 indicate the numbers of participants affected by serious adverse events and the corresponding numbers at risk in the two parent-study treatment-history groups reported by the registry.
They are not hazard ratios, risk ratios, odds ratios, or p-values. They also do not establish that the difference between the two parent-study treatment-history groups is caused by prior treatment assignment.
Comparing raw counts without considering the numbers at risk can be misleading. Even when two groups have similar numbers of affected participants, their underlying proportions can differ if their denominators differ.
18. Limitations and Interpretation Issues
- Single-group design: the OLE is registered as a single-group study, so the primary design does not provide a concurrent randomized comparator.
- Open-label structure: masking is listed as none. Knowledge of treatment can affect reporting and assessment of some safety outcomes.
- Long observation period: the primary endpoint extends up to 5 years, making duration of follow-up an important part of safety interpretation.
- Participant versus event counts: the primary endpoint counts participants with adverse events, whereas individual participants may contribute multiple event records.
- Parent-study treatment history: serious adverse-event counts are categorized by treatment received in the parent study and should not be interpreted as concurrent FOURIER-OLE randomized arms.
- Denominator definition: enrollment of 5035 does not automatically establish the denominator for every safety summary.
- No posted formal statistical analyses: the registry provides the endpoint and selected safety counts but no formal statistical analysis with an effect estimate, confidence interval, or p-value for the primary endpoint.
- Generalizability: participants entering an open-label extension represent a selected population and may differ from the broader population with dyslipidemia.
- Causality: the occurrence of an adverse event during treatment does not by itself establish that the intervention caused the event.
19. Why This Trial Matters Statistically
FOURIER-OLE is useful as a teaching case because it illustrates how the statistical interpretation of a clinical study depends heavily on its design. The study has a large enrollment of 5035 participants and a long primary endpoint time frame of up to 5 years, yet it is registered as a single-group, unmasked extension rather than a concurrent randomized comparison.
| Concept | How it appears in FOURIER-OLE |
|---|---|
| Single-group design | The registry identifies the design model as single-group with 1 arm. |
| Open-label treatment | Masking is listed as none. |
| Participant-level safety endpoint | The primary endpoint counts participants who experienced an adverse event. |
| Treatment-emergent events | The primary endpoint includes treatment-emergent events during the OLE study period. |
| Long-term follow-up | The primary endpoint has a time frame of up to 5 years. |
| Event-source definition | The Event electronic case report form is specified for the primary analysis. |
| CEC review | CEC positively reviewed events are included. |
| Parent-study treatment history | Serious adverse events are reported separately for placebo and evolocumab in the parent study. |
| Denominator interpretation | Serious adverse-event counts are paired with distinct numbers at risk. |
| Descriptive safety analysis | The primary endpoint naturally supports participant counts and proportions rather than a randomized treatment contrast. |
20. Planned Analysis
The registry defines the primary outcome as the Number of Participants Who Experienced an Adverse Event, with a time frame of up to 5 years. The primary analysis is specified to include all treatment-emergent events reported on the Event electronic case report form during the OLE study period, including CEC positively reviewed events and disease-related events.
For an endpoint of this type, the principal statistical summary would normally be the number and proportion of participants experiencing at least one qualifying adverse event in the prespecified analysis population. Additional safety summaries can distinguish event categories, seriousness, severity, timing, and repeated events when those data are available.
Because this is a single-group extension, the principal interpretation is descriptive rather than a randomized comparison between concurrent treatment arms. A time-to-first-event analysis could be appropriate for a separately defined time-to-event safety question, using methods such as Kaplan-Meier estimation, provided that event and censoring times are available.
21. Key Statistical Takeaways
Design drives interpretation
A large sample does not convert a single-group extension into a randomized comparative trial. The estimand remains tied to the study design.
Safety is participant-based
The primary endpoint asks how many participants experienced an adverse event, not how many individual adverse-event records occurred.
Time frame matters
"Up to 5 years" means that exposure and follow-up duration are important when interpreting the occurrence of safety events.
Counts need context
The serious adverse-event counts must be read with their denominators and their parent-study treatment-history labels.
22. Sources
- ClinicalTrials.gov: FOURIER-OLE, NCT02867813.
Continue through the Clinical Biostats knowledge graph
Explore the broader collection of clinical-trial results, statistical methods, and analysis resources from Clinical Biostats.
23. Record Summary
FOURIER-OLE is a phase 3, completed, single-group, open-label extension study of evolocumab in dyslipidemia with 5035 participants enrolled. Its registry-defined primary endpoint is the Number of Participants Who Experienced an Adverse Event over a time frame of up to 5 years, with treatment-emergent events identified from the Event electronic case report form and including CEC positively reviewed events and disease-related events.
The most important statistical distinction is that FOURIER-OLE is not registered as a concurrent randomized comparison. Consequently, its primary safety endpoint is naturally interpreted as a participant-level descriptive outcome rather than as a treatment-effect estimate between randomized arms. The registry also reports serious adverse-event counts according to treatment received in the parent study: 1077/2532 for Placebo in Parent Study and 1045/2499 for Evolocumab in Parent Study. Those categories describe parent-study treatment history rather than separate FOURIER-OLE comparator groups.