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
OASIS-6 was a randomized, parallel, quadruple-masked, phase 3 treatment trial with 12,092 enrolled participants. It evaluated fondaparinux versus control therapy across four intervention arms, with primary endpoints defined as death or recurrent myocardial infarction through day 30 and severe hemorrhage through Day 9.
| Feature | OASIS-6 |
|---|---|
| Trial name | OASIS-6 |
| ClinicalTrials.gov identifier | NCT00064428 |
| Phase | Phase 3 |
| Status | Completed |
| Condition | Thromboembolism |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Quadruple |
| Primary purpose | Treatment |
| Enrollment | 12,092 |
| Lead sponsor | GlaxoSmithKline |
| Sponsor type | Industry |
2. Clinical Question
The registry title frames the study question as an evaluation of the safety and efficacy of fondaparinux versus control therapy in patients with ST-segment elevation acute myocardial infarction. The registered primary endpoints operationalize that question through an efficacy composite and a bleeding endpoint.
Population
Patients with ST-segment elevation acute myocardial infarction, as described in the trial title.
Intervention
Fondaparinux, with separate registered intervention groups for settings in which UFH was not indicated and in which UFH was indicated.
Comparator
Control therapy, including a control group in which UFH was not indicated and a control group in which UFH was used.
Primary questions
How does treatment compare with respect to death or recurrent myocardial infarction through day 30 and severe hemorrhage through Day 9?
3. Trial Design
The study used randomized allocation and a parallel design. The registry identifies four intervention groups, allowing the trial to distinguish fondaparinux from control therapy in two different UFH-indication settings.
Fondaparinux — UFH not indicated
- Fondaparinux
- UFH not indicated
- Registered as a drug intervention
Control — UFH not indicated
- Control
- UFH not indicated
- Registered as an other intervention
Fondaparinux — UFH indicated
- Fondaparinux
- UFH indicated
- Registered as a drug intervention
Control — UFH
- Control
- UFH used
- Registered as a drug intervention
4. Trial Timeline
Trial start
The registry lists August 2003 as the study start date.
Primary completion
The registry lists February 2006 as the primary completion date.
Registry status
The ClinicalTrials.gov record identifies OASIS-6 as completed.
5. Primary Endpoints
The registry specifies two primary endpoints with different clinical meanings and different time frames. The first is a composite time-limited efficacy endpoint. The second is a bleeding endpoint defined using modified TIMI criteria.
| Endpoint | Time frame | Registry definition | Statistical structure |
|---|---|---|---|
| Death or recurrent myocardial infarction | up to day 30 | the first occurrence of any component of death (all-cause mortality) or recurrent myocardial infarction | Time-to-first-event composite endpoint |
| Severe hemorrhage | up to Day 9 | Severe hemorrhage (modified TIMI criteria) | Binary safety endpoint over a defined follow-up period |
Why the composite endpoint deserves attention
The first primary endpoint combines two clinically distinct events: death from any cause and recurrent myocardial infarction. The endpoint is reached at the first occurrence of either component. Consequently, a statistical comparison of the composite is not the same as a comparison of mortality alone or recurrent myocardial infarction alone.
For a participant who experiences recurrent myocardial infarction before death, the composite endpoint occurs at the time of the recurrent myocardial infarction. For a participant who dies without a preceding recorded recurrent myocardial infarction, the endpoint occurs at death. This time-to-first-event structure is central to interpreting the endpoint.
Why the time frames are different
The efficacy endpoint is followed through day 30, whereas severe hemorrhage is evaluated through Day 9. These are therefore not two measurements taken over an identical observation window. The differing windows should remain explicit when comparing or interpreting the endpoints.
6. Statistical Methodology
Time-to-event analysis for death or recurrent myocardial infarction
The first primary endpoint is naturally structured as a time-to-first-event outcome. A typical analysis would begin with Kaplan-Meier estimation of the probability of remaining free from the composite endpoint through time. Treatment groups can then be compared with a log-rank test, with a Cox proportional-hazards model providing a relative hazard estimate when its assumptions are appropriate.
Here, di represents the number of first composite events at event time ti, while ni represents the number at risk immediately before that time.
The registry does not post a formal statistical analysis for this endpoint. Accordingly, Kaplan-Meier, log-rank, and Cox methods are best understood here as the standard analytical framework for this type of endpoint rather than as a statement that a particular model was used in OASIS-6.
Binary analysis for severe hemorrhage
Severe hemorrhage is defined over a fixed period ending at Day 9. Unlike the composite endpoint, its registry definition does not specify a time-to-first-event analysis. A conventional analysis could summarize the proportion of participants experiencing severe hemorrhage in each treatment group and compare those proportions using an absolute risk difference, relative risk, odds ratio, or an appropriate regression model.
The risk difference describes the absolute difference in event probability, while the risk ratio expresses the treatment-group probability relative to the comparator probability.
The registry does not post a specific statistical procedure for the severe-hemorrhage endpoint. The choice among these methods would normally depend on the prespecified statistical analysis plan, event frequency, treatment contrasts, and handling of the four registered arms.
7. Planned Analysis
No formal statistical analyses are posted on ClinicalTrials.gov for the registered primary endpoints. The registry nevertheless defines exactly what is to be measured, which provides a useful framework for understanding how the primary analysis would normally be structured.
| Primary endpoint | Planned statistical perspective | What the endpoint measures |
|---|---|---|
| Death or recurrent myocardial infarction | Typically analyzed as a time-to-first-event outcome using survival methods such as Kaplan-Meier estimation and a treatment-effect model such as Cox regression. | Time until the first occurrence of all-cause death or recurrent myocardial infarction through day 30. |
| Severe hemorrhage | Typically analyzed as a fixed-window binary safety outcome using event proportions and an appropriate treatment-group comparison. | Occurrence of severe hemorrhage according to modified TIMI criteria through Day 9. |
8. Statistical Methods Explained
Why is death or recurrent myocardial infarction analyzed as a composite?
A composite endpoint can capture the first occurrence of more than one clinically relevant event. Here, the registry explicitly defines the endpoint as the first occurrence of either all-cause death or recurrent myocardial infarction. This means every participant contributes to one common time-to-first-event endpoint rather than producing separate primary observations for the two components.
What does "first occurrence" mean?
It means the endpoint clock stops at whichever component occurs first. If recurrent myocardial infarction occurs before death, the recurrent myocardial infarction is the event defining the composite. If death occurs first, death defines the event. This distinction matters because the composite cannot be interpreted simply as the sum of the two event types.
Why would Kaplan-Meier estimation be appropriate?
Kaplan-Meier methods are designed for time-to-event data in which participants can have different follow-up times. They allow participants who have not yet experienced the event by their last observed time to contribute information without being treated as though the event occurred.
What would a Cox hazard ratio mean?
If a Cox model were used, its hazard ratio would describe the relative instantaneous rate of experiencing the first composite event between treatment groups under the fitted model. A hazard ratio below 1 would indicate a lower estimated event hazard for the numerator group; it would not mean that the same percentage of individual participants avoided the event.
Why is severe hemorrhage analyzed differently?
The registered severe-hemorrhage endpoint is defined simply as severe hemorrhage according to modified TIMI criteria through Day 9. The registry does not describe a time-to-event formulation for this endpoint. A fixed-window binary analysis therefore provides a natural framework: count participants who meet the definition by Day 9 and compare the resulting event probabilities across treatment groups.
Why does the four-arm structure matter?
The intervention structure contains two fondaparinux groups and two control groups, separated by whether UFH was indicated or used. A statistical analysis must therefore define the estimand before calculating a treatment effect. For example, combining both fondaparinux groups against both control groups would answer a broader fondaparinux-versus-control question than comparing only one matched pair of arms.
9. Interpreting the Primary Efficacy Endpoint
The registered endpoint is death or recurrent myocardial infarction, measured through day 30 and defined by the first occurrence of either component. The endpoint therefore represents an early composite clinical outcome rather than mortality alone.
Any reported hazard ratio or risk difference for this endpoint would need to be interpreted in light of the component events. A treatment effect on the composite could arise from differences in death, differences in recurrent myocardial infarction, or both.
A relative effect measure can describe how treatment groups compare proportionally, but it does not directly state the absolute number of additional or fewer events. For a 30-day composite endpoint, absolute event probabilities and absolute risk differences can therefore add important clinical context.
If a time-to-event treatment effect were reported, its confidence interval would quantify statistical uncertainty around the estimated effect under the specified analysis. A narrow interval indicates greater statistical precision than a wide interval, but neither interval width nor a point estimate describes the experience of every individual participant.
10. Interpreting Severe Hemorrhage
The second primary endpoint is severe hemorrhage (modified TIMI criteria) through Day 9. Statistically, this endpoint has a different structure from the death-or-recurrent-myocardial-infarction composite.
Fixed observation window
The registry defines the endpoint through Day 9. The analysis therefore concerns hemorrhage occurring within that specified period.
Safety interpretation
Severe hemorrhage is a safety outcome. Its interpretation should remain separate from the efficacy endpoint rather than being folded into a single numerical measure of overall treatment value.
Binary event
For a conventional analysis, each participant can be classified according to whether the registered severe-hemorrhage definition was met during the specified period.
Modified TIMI criteria
The registry identifies modified TIMI criteria as the definition for severe hemorrhage, providing a prespecified clinical classification framework.
A useful statistical presentation would report the severe-hemorrhage event proportion in each relevant treatment group together with an absolute risk difference and a relative measure, each accompanied by an appropriate confidence interval.
11. Analysis Populations and Randomization
The registry identifies randomized allocation but does not provide posted definitions of specific efficacy or safety analysis populations. In a randomized treatment trial, the primary efficacy analysis would ordinarily preserve randomized treatment assignment through an intention-to-treat framework unless the prespecified statistical analysis plan specifies otherwise.
| Concept | Statistical role |
|---|---|
| Randomization | Creates the basis for comparing treatment groups while reducing systematic differences attributable to measured and unmeasured baseline factors. |
| Intention-to-treat principle | Typically maintains participants in their randomized groups for efficacy analysis, preserving the principal benefit of randomization. |
| Safety population | Often defined around actual exposure to study treatment, although the registry excerpt does not provide the OASIS-6 safety-population definition. |
| Four-arm structure | Requires explicit specification of the treatment contrast before estimating an overall fondaparinux effect or a setting-specific effect. |
12. Masking and Bias Control
The OASIS-6 registry record identifies the study as quadruple masked. Masking is a design feature intended to reduce the influence of knowledge of treatment assignment on trial conduct and assessment.
For an efficacy endpoint involving death or recurrent myocardial infarction, the objective nature of death is different from the clinical adjudication and identification of recurrent myocardial infarction. For a bleeding endpoint defined by modified TIMI criteria, consistent endpoint ascertainment is likewise important. Masking can therefore be relevant to reducing assessment-related sources of bias.
Masking, however, is not itself a statistical adjustment. It is a trial-design mechanism that operates before and during outcome assessment rather than a mathematical correction applied after data collection.
13. Multiplicity and the Two Primary Endpoints
OASIS-6 has two registered primary endpoints: death or recurrent myocardial infarction through day 30, and severe hemorrhage through Day 9. When a study has more than one primary endpoint, the statistical interpretation depends on how the endpoints are incorporated into the prespecified testing strategy.
| Issue | Why it matters |
|---|---|
| Two primary endpoints | Multiple primary hypotheses can create a multiplicity question if each is tested formally. |
| Different clinical domains | The endpoints address efficacy and safety rather than measuring the same underlying construct. |
| Different time frames | The efficacy composite is followed through day 30, whereas severe hemorrhage is evaluated through Day 9. |
| Testing strategy | The interpretation of p-values depends on whether the protocol prespecified a joint, hierarchical, co-primary, or other multiplicity-control strategy. |
The registry excerpt does not report a multiplicity strategy. Therefore, a formal claim about familywise type I error control cannot be reconstructed from the endpoint names alone.
14. Missing Data and Censoring
For the death-or-recurrent-myocardial-infarction endpoint, incomplete follow-up creates a censoring issue rather than simply a missing binary value. A participant who has not experienced the composite event by the last observed time may contribute information up to that point, provided the analysis assumptions governing censoring are appropriate.
For severe hemorrhage through Day 9, incomplete ascertainment before the end of the defined observation window creates a different problem. The statistical analysis would need prespecified rules for determining how participants with incomplete safety follow-up are handled.
Censoring
Relevant to time-to-event analysis when a participant's event-free follow-up ends before the event is observed.
Missing endpoint status
Relevant to fixed-window safety analysis when complete ascertainment through Day 9 is unavailable.
Prespecification
Rules for censoring and missingness should be established before unblinding to reduce analysis flexibility.
Interpretation
The handling of incomplete observations can affect both the estimated treatment effect and its uncertainty.
15. Why This Trial Matters Statistically
OASIS-6 is a useful statistical teaching case because its registry design combines randomized treatment allocation, four intervention groups, masking, a composite time-to-event endpoint, and a separately defined fixed-window bleeding endpoint.
| Concept | How it appears in OASIS-6 |
|---|---|
| Randomization | The registry identifies randomized allocation. |
| Parallel design | The design model is registered as parallel. |
| Quadruple masking | The registry identifies the study as quadruple masked. |
| Four treatment groups | Two fondaparinux groups and two control groups are registered according to UFH indication/use. |
| Composite endpoint | Death or recurrent myocardial infarction is defined as the first occurrence of either component. |
| Time-to-event analysis | The 30-day composite has a time-to-first-event structure. |
| Safety endpoint | Severe hemorrhage is defined using modified TIMI criteria through Day 9. |
| Different follow-up windows | The two primary endpoints use day 30 and Day 9 time frames. |
| Multiplicity | Two primary endpoints create a need to understand the prespecified testing strategy. |
| Estimand definition | The four-arm structure makes the treatment contrast an essential part of the statistical question. |
16. Important Limitations and Interpretation Issues
- No posted statistical analyses: the ClinicalTrials.gov record does not post statistical estimates, confidence intervals, or p-values for the registered primary endpoints.
- Four-arm structure: the registered interventions distinguish two fondaparinux groups and two control groups. An overall treatment comparison requires an explicit definition of how those groups are combined.
- Composite endpoint: death and recurrent myocardial infarction are clinically different outcomes. A treatment effect on the composite does not automatically imply the same effect on each component.
- Different time frames: the efficacy composite is defined through day 30, while severe hemorrhage is defined through Day 9. Their event probabilities therefore describe different observation periods.
- Analysis specification: the registry does not report the statistical model, censoring rules, multiplicity strategy, or analysis population definitions needed to reproduce a formal primary analysis.
- Safety versus efficacy: severe hemorrhage is a safety endpoint and should be interpreted independently from the efficacy composite rather than combined into an informal overall score.
- Hazard-ratio assumptions: if a Cox proportional-hazards model were used for the composite endpoint, interpretation would depend on the proportional-hazards framework being appropriate for the data.
17. What a Complete Statistical Report Would Need to Show
A complete statistical presentation of OASIS-6 would need more than the endpoint definitions. It would need to identify the prespecified treatment contrast across the four registered arms and then report the corresponding event counts, treatment-effect estimates, confidence intervals, and hypothesis-test results.
| Component | What a complete analysis would report |
|---|---|
| Death or recurrent myocardial infarction | Number of first composite events, time-to-event estimates, treatment-effect estimate, confidence interval, and inferential result. |
| Composite components | Event counts for death and recurrent myocardial infarction separately to show which components contribute to the composite. |
| Severe hemorrhage | Number and proportion meeting modified TIMI severe-hemorrhage criteria through Day 9 in each relevant treatment group. |
| Treatment contrast | Explicit identification of which of the four registered arms are being compared. |
| Multiplicity | Prespecified method for controlling or interpreting type I error across the two primary endpoints, if applicable. |
| Analysis populations | Definitions for efficacy and safety populations and any exclusions or deviations from randomized assignment. |
This structure illustrates an important statistical principle: an endpoint definition tells the analyst what to measure, while the statistical analysis plan specifies how the measurement becomes an estimand and an inferential result.
18. Statistical Interpretation of a Future Hazard Ratio
If the primary composite were analyzed with a Cox model, a hazard ratio below 1 would indicate a lower estimated instantaneous rate of the first composite event in the numerator treatment group relative to the denominator group.
It would not mean that the same percentage of participants avoided death or myocardial infarction, nor would it mean that every participant experienced a proportional reduction in individual risk.
A 95% confidence interval would describe statistical uncertainty around the estimated treatment effect under the model and sampling framework. It would not describe the range of treatment effects experienced by individual participants.
A p-value would quantify the compatibility of the observed data with a specified null hypothesis under the statistical model. It would not measure the size of the treatment effect, the clinical importance of the result, or the probability that the treatment hypothesis is true.
For a Cox analysis, the treatment effect is interpreted through a hazard model and therefore requires attention to censoring assumptions and the proportional-hazards framework. A single hazard ratio should not automatically be treated as a complete description of the treatment-group survival experience over time.
19. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The primary efficacy endpoint is a time-to-first-event composite, while severe hemorrhage is a fixed-window safety endpoint. Each requires an analysis aligned with its endpoint structure and with the registered four-arm design.
Clinical interpretation
The trial was designed to evaluate both clinical efficacy and serious bleeding risk. These outcomes address different dimensions of treatment effect and should be examined separately before considering their combined clinical context.
20. Sources
- ClinicalTrials.gov: NCT00064428 — OASIS-6.
- PubMed: PubMed record for the OASIS-6 publication.
- PubMed: PubMed record for the OASIS-6 publication.
Continue through the Clinical Biostats knowledge graph
Explore clinical-trial design, survival analysis, composite endpoints, and safety-event methodology through the broader Clinical Biostats library.
21. Record Summary
OASIS-6 is a randomized phase 3, parallel, quadruple-masked treatment trial with 12,092 enrolled participants and four registered intervention groups. Its two primary endpoints have deliberately different statistical structures: death or recurrent myocardial infarction is a first-occurrence composite measured through day 30, while severe hemorrhage is defined using modified TIMI criteria through Day 9.
The most important statistical issue is therefore not simply selecting a familiar test. A rigorous analysis must first identify the treatment contrast across the four registered arms, preserve the endpoint definitions, distinguish the time-to-event efficacy analysis from the fixed-window safety analysis, and account for the presence of two primary endpoints. The ClinicalTrials.gov record does not post formal statistical analyses or outcome estimates for these endpoints, so the appropriate interpretation remains focused on the registered design and the statistical framework required to analyze it.