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
OPTIMAS (OPtimal TIMing of Anticoagulation After Acute Ischaemic Stroke) is a randomized, parallel-group trial evaluating direct oral anticoagulant (DOAC) treatment in patients with acute ischaemic stroke and atrial fibrillation. The registered primary endpoint is a composite of recurrent symptomatic ischaemic stroke, symptomatic intracranial haemorrhage, and systemic embolism, assessed at 90 days from randomisation.
| Feature | OPTIMAS |
|---|---|
| Trial acronym | OPTIMAS |
| Full title | OPTIMAS: OPtimal TIMing of Anticoagulation After Acute Ischaemic Stroke : a Randomised Controlled Trial |
| Population | Patients with acute ischaemic stroke and atrial fibrillation |
| Intervention | Direct oral anticoagulant (DOAC) |
| Allocation | Randomized |
| Design | Parallel |
| Masking | Single |
| Primary purpose | Treatment |
| Enrollment | 3648 |
| Number of arms | 2 |
| Lead sponsor | University College, London |
| Sponsor type | OTHER |
| Trial status | UNKNOWN |
| Start date | 2019-06-18 |
| Primary completion date | 2024-05-30 |
| ClinicalTrials.gov | NCT03759938 |
2. Clinical Question
The central clinical question is whether early initiation of DOAC treatment in patients with acute ischaemic stroke and atrial fibrillation is as effective as, or better than, standard initiation of DOAC treatment in preventing the registered composite outcome of recurrent symptomatic ischaemic stroke, symptomatic intracranial haemorrhage, and systemic embolism at 90 days from randomisation.
Population
Patients with acute ischaemic stroke and atrial fibrillation.
Intervention
Direct oral anticoagulant (DOAC) treatment.
Comparator
Standard initiation of DOAC treatment.
Primary question
Is early initiation of DOAC treatment as effective as, or better than, standard initiation in preventing the registered 90-day composite outcome?
3. Trial Design
The design is fundamentally a randomized parallel-group comparison. Randomization is important because it creates the basis for comparing outcomes between treatment strategies while reducing systematic differences in measured and unmeasured baseline characteristics between the groups.
The single-masked designation also matters statistically. Masking can reduce the influence of treatment knowledge on aspects of outcome assessment or participant and investigator behavior, depending on which parties are masked. The registry identifies the study as single masked but does not provide further masking detail in the information reported here.
4. Endpoints
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Primary endpoint | Composite outcome of the combined incidence of recurrent symptomatic ischaemic stroke, symptomatic intracranial haemorrhage and systemic embolism | At 90 days from randomisation |
The primary endpoint is a composite outcome
The registered primary endpoint combines three clinically distinct event types: recurrent symptomatic ischaemic stroke, symptomatic intracranial haemorrhage, and systemic embolism. The statistical analysis therefore concerns the occurrence of any component of the composite within the specified 90-day period.
Composite endpoints can increase the number of observed events relative to an individual component, which can be useful when the components are clinically related. At the same time, interpretation requires attention to the contribution of each component. A composite treatment effect can be driven disproportionately by one component, particularly when components differ substantially in frequency, severity, or clinical importance.
5. Statistical Methodology
The registry identifies a randomized, parallel-group treatment comparison with a binary composite outcome assessed at a defined 90-day time point. For an endpoint of this type, the primary statistical analysis would ordinarily compare the proportion of participants experiencing the composite outcome in the randomized groups.
Binary outcome analysis
At 90 days, each participant can be classified according to whether the registered composite outcome occurred. The basic data structure can therefore be represented by a two-by-two table:
| Composite outcome by 90 days | No composite outcome by 90 days | |
|---|---|---|
| Early DOAC initiation | Events | Non-events |
| Standard DOAC initiation | Events | Non-events |
Several effect measures can describe such a comparison. The risk ratio compares the probability of the composite outcome between groups. The risk difference subtracts the two probabilities and expresses the absolute difference. An odds ratio compares the odds of the outcome between groups.
Each measure answers a different question. The risk difference is expressed in absolute probability units, whereas the risk ratio and odds ratio are relative measures.
Confidence intervals
A confidence interval would normally accompany the estimated treatment effect. For example, a confidence interval around a risk difference describes the statistical uncertainty associated with the estimated difference in event probabilities under the chosen analysis framework.
Confidence intervals are especially useful for a trial such as OPTIMAS because the clinical question concerns whether two treatment-timing strategies have sufficiently similar or different event risks. The interval provides information about both the direction and precision of the estimated treatment effect.
Hypothesis testing
A hypothesis test for the primary endpoint would normally evaluate a prespecified null hypothesis concerning the difference between the treatment strategies. The resulting p-value would quantify the compatibility of the observed data with that null hypothesis under the specified statistical model; it would not quantify the magnitude or clinical importance of the treatment effect.
For a 90-day binary composite endpoint, the most informative presentation would generally include the event proportion in each randomized group together with an absolute effect measure, a relative effect measure, a confidence interval, and the prespecified hypothesis-test result. The effect estimate and its confidence interval describe the size and precision of the comparison; the p-value addresses a separate hypothesis-testing question.
6. Statistical Methods Explained
Why is the primary endpoint a composite?
A composite endpoint combines multiple clinically relevant events into one outcome. In OPTIMAS, the registered composite combines recurrent symptomatic ischaemic stroke, symptomatic intracranial haemorrhage, and systemic embolism. Statistically, this allows a participant to count as having experienced the primary endpoint if any one of these events occurs within the 90-day assessment period.
Why does the 90-day time frame matter?
The time frame defines the period during which an event contributes to the registered primary endpoint. It prevents an indefinite follow-up period from being implicitly substituted for the prespecified outcome window and makes the treatment comparison explicitly about outcomes by 90 days from randomisation.
What is the difference between a risk ratio and a risk difference?
A risk ratio describes the relative probability of an event between the two treatment groups. A risk difference describes the absolute separation in probabilities. For example, two treatment strategies could have a similar relative effect in different populations while producing different absolute differences if their baseline risks differ.
Why is an absolute effect important?
Relative measures can describe the proportional change in outcome risk but do not directly indicate how many additional or fewer events occur per population. An absolute risk difference puts the comparison into the original probability scale and is therefore complementary to a relative measure.
Why does a composite endpoint require component-level interpretation?
The composite treats all qualifying components as contributing to the same primary outcome. However, recurrent symptomatic ischaemic stroke, symptomatic intracranial haemorrhage, and systemic embolism are not interchangeable clinical events. A complete interpretation should therefore examine the overall composite together with the components when component-specific results are available.
Why is randomization important?
Randomization creates the design-based foundation for a causal comparison of the treatment strategies. It means that, apart from random variation, systematic differences between groups should not determine treatment assignment. The resulting comparison is therefore not equivalent to an observational comparison in which treatment timing is chosen according to clinical characteristics.
What does single masking contribute?
Masking can reduce the influence of knowledge of treatment assignment on behavior, assessment, or reporting. The precise implications depend on which participants, investigators, or outcome assessors were masked. The registry identifies the trial as single masked but does not provide the additional detail needed to specify those roles here.
7. Planned Analysis
The registered primary endpoint is the combined incidence of recurrent symptomatic ischaemic stroke, symptomatic intracranial haemorrhage, and systemic embolism at 90 days from randomisation. For this type of fixed-time binary composite endpoint, a typical analysis would compare the 90-day event proportions between the randomized treatment groups and report an appropriate measure of treatment effect with its confidence interval and hypothesis-test result.
Primary comparison
Compare the incidence of the registered composite outcome between the randomized treatment strategies at 90 days.
Effect measures
Risk difference and relative measures such as a risk ratio or odds ratio can describe the treatment contrast from complementary perspectives.
Uncertainty
A confidence interval should accompany the treatment-effect estimate to show its statistical precision.
Component interpretation
When component-specific results are available, the three components should be examined alongside the overall composite.
8. Understanding the Primary Endpoint Statistically
The primary endpoint is different from a continuous outcome such as a laboratory measurement or a score. Each participant contributes a binary outcome at the prespecified 90-day time point: the composite either occurred or did not occur during the relevant assessment period.
| Statistical feature | Implication for OPTIMAS |
|---|---|
| Outcome type | Binary composite outcome assessed at 90 days |
| Unit of analysis | Randomized participant |
| Primary comparison | Early versus standard initiation of DOAC treatment |
| Composite components | Recurrent symptomatic ischaemic stroke; symptomatic intracranial haemorrhage; systemic embolism |
| Effect scale | Typically expressed using absolute and/or relative measures for binary outcomes |
| Uncertainty | Typically expressed using a confidence interval around the treatment-effect estimate |
| Assessment time | 90 days from randomisation |
The statistical interpretation should remain anchored to the randomized comparison. Because the endpoint is assessed at a fixed time point, the central question is whether the probability of experiencing at least one qualifying event by 90 days differs between the two randomized strategies.
9. Composite Endpoint Interpretation
Composite endpoints are often statistically efficient because they combine several clinically relevant outcomes into one primary analysis. Their interpretation, however, depends on how the component events behave.
Recurrent symptomatic ischaemic stroke
This is one of the three events that causes a participant to meet the registered primary endpoint.
Symptomatic intracranial haemorrhage
This is a second qualifying event type included in the registered composite.
Systemic embolism
This is the third qualifying event type contributing to the composite outcome.
Overall composite
A participant contributes to the primary endpoint if the combined outcome occurs within the 90-day assessment period.
A treatment effect on the composite should not automatically be interpreted as an identical treatment effect on every component. If component-level estimates become available, they provide essential context for understanding which events contribute to the overall result.
10. Analysis Population and Randomization
The trial is registered as randomized with 3648 participants and two parallel arms. Randomization establishes the principal comparison between early and standard initiation strategies.
For a randomized efficacy analysis, the intention-to-treat principle is generally important because participants remain associated with their randomized treatment assignment. This preserves the comparison created by randomization even if treatment exposure subsequently differs between participants.
An intention-to-treat analysis estimates the effect of being assigned to a treatment strategy rather than restricting the analysis only to participants who fully adhered to the assigned strategy.
The registry information reported here does not provide additional analysis-population definitions or exclusions. Those details would determine the exact population used for any formal statistical analysis.
11. Missing Data and Follow-Up
A fixed 90-day endpoint requires an ascertainment rule for determining whether a participant experienced the composite outcome during the specified period. In a complete statistical analysis, participants without complete endpoint information would need to be handled according to prespecified rules.
Possible statistical approaches to incomplete binary outcomes include prespecified missing-data assumptions, sensitivity analyses, or other methods appropriate to the source and mechanism of missingness. The choice should be determined before examining treatment-group outcome differences whenever possible.
12. Non-Inferiority and Superiority Logic
The registered endpoint description states that OPTIMAS will investigate whether early initiation of DOAC treatment is as effective as, or better than, standard initiation in preventing the specified 90-day composite outcome.
This wording raises an important statistical distinction. A conventional superiority test asks whether the treatment groups differ. A non-inferiority analysis instead asks whether the new strategy is not unacceptably worse than the comparator by more than a prespecified margin.
| Question | Statistical interpretation |
|---|---|
| Superiority | Is there evidence that the treatment strategies differ in the prespecified direction? |
| Non-inferiority | Is the early strategy sufficiently close to standard initiation that any disadvantage does not exceed a prespecified acceptable margin? |
| Clinical meaning | The conclusion depends on the prespecified margin, effect measure, confidence interval, and analysis population. |
The ClinicalTrials.gov record does not state a numerical non-inferiority margin. Consequently, a numerical non-inferiority conclusion cannot be derived from the registered information alone.
13. Time Point Versus Time-to-Event Analysis
The registered primary endpoint is specified as the combined incidence of the three events at 90 days from randomisation. This is conceptually different from an endpoint defined as time from randomisation until the first occurrence of an event, with participants followed until an event or censoring.
| Feature | 90-day composite incidence | Time-to-event analysis |
|---|---|---|
| Primary scale | Whether the event occurred by a fixed time | Time until the event occurs |
| Typical effect measures | Risk difference, risk ratio, odds ratio | Hazard ratio and survival probabilities |
| Time frame | 90 days from randomisation | Depends on follow-up and censoring rules |
| Interpretation | Probability of the composite by 90 days | Relative event rate over follow-up under the specified model |
The distinction matters because a hazard ratio should not be substituted automatically for a fixed-time risk comparison. The appropriate effect measure depends on the prespecified statistical analysis and endpoint definition.
14. Statistical Interpretation of Effect Measures
A risk ratio compares the probability of the 90-day composite outcome between early and standard DOAC initiation. A value below 1 would indicate a lower estimated probability in the numerator group, while a value above 1 would indicate a higher estimated probability.
A risk difference expresses the absolute separation between the two 90-day event probabilities. It is directly interpretable in percentage-point terms and can be translated into an approximate number of events per population when expressed on an appropriate scale.
An odds ratio compares the odds of experiencing the composite outcome between the randomized groups. It is not identical to a risk ratio, particularly when the outcome is not rare.
A confidence interval describes the statistical uncertainty around the estimated treatment effect under the specified analysis. It does not describe the range of outcomes experienced by individual participants.
A p-value addresses the compatibility of the observed data with a specified null hypothesis. It does not measure the size of the treatment effect, the probability that the treatment is beneficial, or the clinical importance of the result.
15. Limitations
- No posted statistical analyses: the ClinicalTrials.gov record does not provide a formal statistical analysis result for the registered primary endpoint.
- No posted treatment-effect estimate: the registry information reported here does not provide a 90-day event estimate, confidence interval, or p-value for the primary composite.
- Composite endpoint: the primary outcome combines recurrent symptomatic ischaemic stroke, symptomatic intracranial haemorrhage, and systemic embolism. Component-specific behavior is important for interpretation.
- Non-inferiority margin: the registry information reported here does not state a numerical margin, so the non-inferiority criterion cannot be evaluated numerically.
- Analysis details: the registry information reported here does not specify the exact statistical model, hypothesis-testing procedure, confidence level, missing-data strategy, or analysis populations for the primary endpoint.
- Masking detail: the study is identified as single masked, but the registry information reported here does not specify which trial roles were masked.
- Fixed-time endpoint: a 90-day incidence comparison and a time-to-event analysis answer different statistical questions and should not be treated as interchangeable.
16. Why This Trial Matters Statistically
OPTIMAS provides a useful example of how trial design and endpoint definition determine the appropriate statistical framework. The key statistical issue is not simply whether two groups are compared, but what is being compared, when it is measured, and what inferential question the comparison is intended to answer.
| Concept | How it appears in OPTIMAS |
|---|---|
| Randomization | Participants are assigned using a randomized allocation. |
| Parallel-group design | The trial has 2 arms evaluated in parallel. |
| Binary endpoint | The primary outcome is the incidence of a defined composite event by 90 days. |
| Composite endpoint | Three clinically distinct events contribute to the primary outcome. |
| Fixed follow-up | The primary endpoint is assessed at 90 days from randomisation. |
| Absolute effect | A risk difference describes the absolute separation in event probabilities. |
| Relative effect | A risk ratio or odds ratio can describe the relative treatment contrast. |
| Confidence interval | Quantifies uncertainty around the estimated treatment effect. |
| Hypothesis testing | A formal test can assess the prespecified null hypothesis. |
| Non-inferiority logic | The registry describes the question as whether early treatment is as effective as, or better than, standard initiation. |
17. Trial Timeline
Trial start
The registered start date for OPTIMAS is June 18, 2019.
Randomized parallel-group trial
The registry reports enrollment of 3648 participants across 2 arms.
Primary endpoint assessment
The registered primary endpoint is the combined incidence of recurrent symptomatic ischaemic stroke, symptomatic intracranial haemorrhage, and systemic embolism at 90 days from randomisation.
Primary completion
The registered primary completion date is May 30, 2024.
18. What a Complete Statistical Result Would Show
For this trial's primary endpoint, a statistically complete results presentation would place the two randomized treatment groups side by side and report the number or proportion of participants who experienced the composite outcome by 90 days.
| Result element | Purpose |
|---|---|
| Number of participants analyzed | Defines the population contributing to the primary analysis. |
| Composite events by arm | Shows the observed frequency of the primary outcome. |
| Event proportion by arm | Provides the absolute 90-day risk in each randomized group. |
| Risk difference | Shows the absolute treatment contrast. |
| Relative effect | Shows the proportional treatment contrast. |
| Confidence interval | Quantifies statistical precision. |
| P-value | Addresses the prespecified hypothesis test. |
| Component outcomes | Shows how each constituent event contributes to the composite. |
These quantities should be interpreted together rather than relying on a single statistic. In particular, a relative effect without the underlying event proportions can obscure the absolute clinical frequency of the outcome, while an event percentage without an uncertainty interval does not fully describe statistical precision.
19. Planned Analysis: Non-Inferiority Perspective
The wording of the registered primary objective makes non-inferiority reasoning especially relevant. In a non-inferiority framework, the critical quantity is the confidence interval for the treatment contrast relative to a prespecified margin that represents the largest clinically acceptable disadvantage of early initiation.
The direction of the comparison and the definition of the margin depend on the selected effect measure. A non-inferiority conclusion cannot be determined from a p-value alone.
The registry information reported here does not provide the numerical non-inferiority margin or the formal testing procedure. Those quantities are essential for a numerical evaluation of the non-inferiority hypothesis.
20. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The trial is structured to compare the 90-day incidence of a prespecified composite outcome between randomized strategies for DOAC initiation. The appropriate analysis must account for the binary fixed-time endpoint and the stated objective of determining whether early initiation is as effective as, or better than, standard initiation.
Clinical interpretation
The clinical question concerns the timing of DOAC treatment after acute ischaemic stroke in patients with atrial fibrillation, with recurrent symptomatic ischaemic stroke, symptomatic intracranial haemorrhage, and systemic embolism forming the registered primary composite outcome.
Keeping these two levels separate is important. Statistical evidence describes the estimated difference and its uncertainty. Clinical interpretation considers what that difference means in the context of the treatment-timing question and the individual components of the composite endpoint.
21. Registry Record Summary
OPTIMAS is registered as a randomized, single-masked, parallel-group treatment trial involving 3648 participants and 2 arms. The intervention is direct oral anticoagulant treatment, and the clinical question concerns early versus standard initiation after acute ischaemic stroke in patients with atrial fibrillation.
The primary endpoint is a composite of recurrent symptomatic ischaemic stroke, symptomatic intracranial haemorrhage, and systemic embolism, assessed at 90 days from randomisation. Statistically, this is a fixed-time binary composite endpoint for which treatment effects can ordinarily be expressed using absolute and relative measures accompanied by confidence intervals and an appropriate hypothesis test.
22. Sources
- ClinicalTrials.gov: OPTIMAS — NCT03759938.
- PubMed: PubMed record 41919368.
- PubMed: PubMed record 40402086.
- PubMed: PubMed record 39966982.
- PubMed: PubMed record 39491870.
- PubMed: PubMed record 35018878.
Continue through the Clinical Biostats clinical trial library
Explore additional randomized-trial analyses and statistical methods across clinical research.