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
STOPDAPT-2 was a randomized, parallel, open-label phase 4 trial in coronary artery disease comparing two durations of dual antiplatelet therapy: 1-month DAPT and 12-month DAPT. The registry reports enrollment of 3045 participants and one registered primary endpoint assessed at 12 months.
| Feature | STOPDAPT-2 |
|---|---|
| Study acronym | STOPDAPT-2 |
| Full title | ShorT and OPtimal Duration of Dual AntiPlatelet Therapy-2 Study |
| Phase | Phase 4 |
| Condition | Coronary Artery Disease |
| Enrollment | 3045 |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None |
| Primary purpose | Treatment |
| Interventions | 1-month DAPT; 12-month DAPT |
| Lead sponsor | Kyoto University, Graduate School of Medicine |
| Status | Completed |
| Start date | 2015-12 |
| Primary completion date | 2018-12-08 |
| ClinicalTrials.gov | NCT02619760 |
2. Clinical Question
The central clinical question is whether a strategy of 1-month dual antiplatelet therapy differs from a strategy of 12-month dual antiplatelet therapy with respect to the registered 12-month composite endpoint in patients with coronary artery disease.
Population
Participants with coronary artery disease enrolled in the STOPDAPT-2 trial.
Intervention
1-month DAPT.
Comparator
12-month DAPT.
Primary question
How do the two randomized DAPT-duration strategies compare on the registered 12-month composite cardiovascular, ischemic, and bleeding endpoint?
3. Trial Design
STOPDAPT-2 used a randomized, parallel-group design with two arms. The registry identifies the masking status as none, meaning the trial was not masked. Its primary purpose was treatment.
1-month DAPT
- Dual antiplatelet therapy for 1 month.
- Randomized treatment strategy.
12-month DAPT
- Dual antiplatelet therapy for 12 months.
- Randomized comparator strategy.
4. Enrollment and Trial Timeline
Study start
The registry lists December 2015 as the study start date.
Randomized treatment comparison
The study used randomized allocation and a parallel two-arm design to compare 1-month and 12-month DAPT strategies.
Primary endpoint assessment
The registered primary endpoint was a composite event assessed over a 12-month time frame.
Primary completion
The registry lists December 8, 2018 as the primary completion date.
5. Primary Endpoint
| Endpoint | Time frame | Registry definition |
|---|---|---|
| Composite event of cardiovascular death/myocardial infarction/definite stent thrombosis/stroke/bleeding | 12-month | Composite event of cardiovascular death/myocardial infarction/definite stent thrombosis/stroke/bleeding defined as major or minor under the definition of Thrombolysis in Myocardial Infarction (TIMI) Study group |
The endpoint combines several clinically distinct outcomes into a single composite. Four components are ischemic or mortality outcomes—cardiovascular death, myocardial infarction, definite stent thrombosis, and stroke—while bleeding is incorporated as another component, with bleeding classified as major or minor under the TIMI Study group definition.
6. Why the Composite Endpoint Requires Careful Statistical Interpretation
Composite endpoints are common in cardiovascular trials because several clinically relevant events can be evaluated within one prespecified outcome. Statistically, however, the composite is meaningful only in relation to the components that make it up and the frequency with which those components occur.
Different clinical events
Cardiovascular death, myocardial infarction, stent thrombosis, stroke, and bleeding are not interchangeable outcomes. A composite treats their occurrence as a single endpoint for the primary comparison.
Event frequency matters
If one component occurs much more frequently than the others, that component can contribute disproportionately to the total number of composite events.
Direction matters
The clinical meaning of a composite depends on how its individual components behave. A combined endpoint should not be interpreted as though all components represent the same type of clinical benefit or harm.
Component results add context
Component-specific estimates are useful for understanding whether the composite reflects mortality, ischemic events, bleeding, or a combination of these outcomes.
7. Statistical Methodology
The registry identifies the primary outcome as a composite event measured at 12 months, but the registry record does not post a formal statistical analysis for this endpoint. Consequently, the appropriate statistical framework must be described at the level of the endpoint rather than represented as a reported trial result.
Binary composite analysis at 12 months
Because the registered time frame is 12 months, one natural analysis is to classify each randomized participant according to whether the composite event occurred by 12 months. The resulting event proportions can then be compared between the two randomized groups.
A between-group comparison can be expressed using an absolute risk difference, a risk ratio, or another prespecified measure. The choice should follow the trial's statistical analysis plan rather than being selected after examining the results.
Time-to-event analysis
If the statistical analysis plan treats the composite as a time-to-first-event endpoint, survival methods such as Kaplan-Meier estimation and a time-to-event comparison may be appropriate. Such an approach uses the timing of events and accommodates participants who are censored before experiencing the endpoint.
The registry endpoint description itself does not specify a Kaplan-Meier analysis, Cox model, log-rank test, hazard ratio, or censoring rule. Those methods therefore should not be presented as reported STOPDAPT-2 results.
Intention-to-treat principle
For a randomized comparison, the primary efficacy analysis is commonly anchored to randomized treatment assignment. This preserves the balance created by randomization and answers the question associated with assignment to the treatment strategy.
Composite-event analysis
For a composite endpoint, the statistical analysis should define the first qualifying event according to the prespecified endpoint rules when a time-to-first-event framework is used. A participant experiencing more than one component requires rules that specify how multiple events are handled. Those rules are part of the statistical analysis plan and are not described in the registry endpoint entry.
8. Statistical Methods Explained
What is a composite endpoint?
A composite endpoint combines multiple clinical events into one outcome. In STOPDAPT-2, the registered composite includes cardiovascular death, myocardial infarction, definite stent thrombosis, stroke, and bleeding. A participant who experiences any qualifying component can therefore contribute an event to the composite endpoint.
Why combine cardiovascular and bleeding outcomes?
DAPT duration can affect different types of clinical events in different directions. A composite that includes both ischemic outcomes and bleeding is intended to summarize these clinically important event types within one prespecified endpoint. Its interpretation therefore requires attention to the individual components as well as the combined result.
What is an absolute risk difference?
An absolute risk difference compares the proportion of participants experiencing the endpoint in the two groups. If one group has risk \(p_1\) and the other has risk \(p_0\), the risk difference is \(p_1-p_0\). It is expressed in percentage points and directly describes the difference in event probability over the specified follow-up period.
Unlike a relative measure, an absolute risk difference describes how much the event probability differs between treatment strategies over the specified period.
What is a risk ratio?
A risk ratio compares the probability of the composite event between treatment groups. A ratio of 1 indicates equal observed risks, a value below 1 indicates lower observed risk in the numerator group, and a value above 1 indicates higher observed risk in the numerator group.
Why might Kaplan-Meier analysis be used?
If the endpoint is analyzed according to time until the first qualifying event, Kaplan-Meier estimation can account for differing follow-up times and censoring. It estimates the probability of remaining free of the event over time rather than reducing every participant's experience to a simple yes-or-no status at one time point.
Why is a hazard ratio different from a risk ratio?
A risk ratio compares cumulative event probabilities over a specified period. A hazard ratio compares modeled instantaneous event rates over time. They answer different statistical questions and should not be substituted for one another without justification.
Why are component-specific analyses important?
A composite endpoint can conceal differences among its components. For example, a combined result could be influenced primarily by bleeding rather than cardiovascular death or myocardial infarction. Reporting the individual components helps determine what the composite actually represents clinically.
9. Randomization and Causal Interpretation
Randomization is the central design feature that supports a causal comparison between the two DAPT-duration strategies. With randomized allocation, treatment assignment is determined independently of participants' baseline characteristics in expectation, allowing differences in outcomes between randomized groups to be interpreted as differences associated with the assigned strategies, subject to the usual assumptions of randomized-trial analysis.
| Design feature | STOPDAPT-2 | Statistical relevance |
|---|---|---|
| Allocation | Randomized | Supports comparison of outcomes by assigned treatment strategy. |
| Design model | Parallel | Each participant contributes to one randomized treatment arm. |
| Masking | None | Knowledge of assignment can matter for some clinical assessments and treatment decisions. |
| Primary purpose | Treatment | The trial evaluates alternative treatment strategies. |
| Arms | 2 | The primary comparison is between two DAPT-duration strategies. |
10. Analysis Population and Estimand Considerations
The registry reports randomized allocation and a two-arm parallel design but does not provide an analysis-population definition in the ClinicalTrials.gov record. For randomized trials, several populations can be considered, including an intention-to-treat population and treatment-exposed populations. Their interpretation differs.
Intention-to-treat
Participants are analyzed according to randomized assignment. This preserves the treatment comparison created by randomization and is commonly central to efficacy analyses.
Per-protocol
A per-protocol analysis focuses on participants who sufficiently followed the assigned strategy. It can address a different question but is more vulnerable to post-randomization selection.
As-treated
Participants are classified according to treatment actually received. This can be useful for some safety questions but may weaken the protection against confounding provided by randomization.
Estimand
The statistical estimand defines the treatment effect being targeted, including the population, endpoint, time frame, and handling of post-randomization events.
The appropriate primary analysis population and treatment-effect estimand should be taken from the trial's prespecified statistical documents. The registry entry identifies the randomized design but does not provide those additional specifications.
11. Missing Data, Censoring, and Follow-Up
A 12-month clinical endpoint requires clear rules for participants whose status at 12 months is unavailable or whose follow-up ends earlier. If the endpoint is treated as a fixed-time binary outcome, missing outcome status requires a prespecified approach. If it is analyzed as a time-to-event outcome, censoring rules become central.
| Issue | Why it matters |
|---|---|
| Incomplete 12-month follow-up | Participants without ascertainable endpoint status may not contribute the same information as participants followed through the full assessment period. |
| Loss to follow-up | Can introduce bias if the probability of incomplete follow-up is related to treatment or outcome risk. |
| Censoring | In a time-to-event analysis, censoring determines when a participant stops contributing observed event-free follow-up. |
| Multiple events | Composite endpoints require predefined rules for how multiple component events are counted. |
| Competing event types | Death and nonfatal clinical events can interact statistically when time-to-event methods are used. |
The registry record does not specify a missing-data method, imputation method, censoring rule, or competing-risk framework. These should therefore be treated as statistical-analysis-plan details rather than assumed features of the registry analysis.
12. Multiplicity and the Composite Endpoint
Multiplicity concerns arise whenever a trial makes several formal statistical comparisons. A single registered composite endpoint is different from separately testing every component as though each were an independent primary endpoint.
Primary composite
The registered primary endpoint is one composite consisting of five clinical event categories.
Component analyses
Analyses of cardiovascular death, myocardial infarction, stent thrombosis, stroke, or bleeding individually answer different questions from the primary composite.
Multiple testing
If several component-specific hypotheses are tested formally, the statistical analysis plan should specify how type I error is controlled.
Interpretation
A nominally significant component result should not automatically be interpreted as another confirmatory primary finding.
13. Non-Inferiority, Superiority, and Effect Direction
The registry information identifies the two treatment strategies and the composite endpoint but does not specify a non-inferiority margin, superiority hypothesis, or formal hypothesis-testing framework. Those details are fundamental to interpreting a comparative DAPT-duration trial.
For a superiority analysis, the null hypothesis generally concerns no difference between treatment strategies. For a non-inferiority analysis, the question is instead whether the shorter strategy is not unacceptably worse than the comparator according to a prespecified margin. The margin must be established before examining the observed treatment effect.
Because the registry entry does not report such a margin or testing framework, no non-inferiority conclusion can be assigned from the registry endpoint description alone.
14. Planned Analysis
The registry states that the primary endpoint is the composite event of cardiovascular death/myocardial infarction/definite stent thrombosis/stroke/bleeding, assessed over a 12-month time frame. The composite includes bleeding defined as major or minor according to the TIMI Study group definition.
For an endpoint of this type, the analysis would typically compare the incidence of the composite between the randomized treatment strategies. If analyzed as a fixed 12-month binary endpoint, the comparison could use an absolute risk difference and/or risk ratio with a confidence interval and a prespecified hypothesis test. If the statistical analysis plan defines the endpoint as time to first composite event, Kaplan-Meier estimation and a time-to-event treatment comparison could instead be used.
A complete interpretation would also identify how missing follow-up, multiple events, censoring, and the individual components of the composite were handled.
No formal statistical analyses are posted to ClinicalTrials.gov for the registered primary endpoint. Results have not been posted on ClinicalTrials.gov.
15. Safety and the Composite Endpoint
Bleeding is explicitly included as a component of the registered primary composite endpoint. This creates an important statistical distinction between the overall composite and a conventional safety analysis.
| Outcome category | Role in STOPDAPT-2 |
|---|---|
| Cardiovascular death | Component of the primary composite |
| Myocardial infarction | Component of the primary composite |
| Definite stent thrombosis | Component of the primary composite |
| Stroke | Component of the primary composite |
| Bleeding | Component of the primary composite; major or minor under the TIMI Study group definition |
The registry record does not provide serious adverse-event counts by treatment arm. It therefore does not support a separate numerical safety comparison here. The presence of bleeding within the primary composite nevertheless makes the distinction between ischemic and bleeding outcomes central to interpretation of the trial's endpoint.
16. Why This Trial Matters Statistically
STOPDAPT-2 is a useful statistical case study because it combines randomized treatment assignment with a clinically heterogeneous composite endpoint. The design makes it possible to examine how treatment-duration questions are translated into a prespecified endpoint and how the choice of statistical estimand affects interpretation.
| Concept | How it appears in STOPDAPT-2 |
|---|---|
| Randomization | Participants were randomly allocated to two treatment strategies. |
| Parallel-group design | The study has two parallel treatment arms. |
| Open-label treatment | The registry lists masking as none. |
| Composite endpoint | Five clinically distinct event categories are combined into one primary endpoint. |
| Fixed follow-up | The registered primary endpoint has a 12-month time frame. |
| Risk comparison | A fixed-time composite can be summarized with absolute and relative measures of event probability. |
| Time-to-event methods | If specified in the statistical analysis plan, event timing and censoring can be incorporated into survival analysis. |
| Component analysis | The combined endpoint requires attention to cardiovascular death, myocardial infarction, stent thrombosis, stroke, and bleeding individually. |
| Multiplicity | Formal testing of additional component hypotheses requires a prespecified multiplicity framework. |
| Estimand | The intended treatment effect depends on the population, endpoint, time frame, and handling of post-randomization events. |
17. Statistical Interpretation of a Composite Treatment Effect
If the 1-month DAPT strategy produced a lower estimated 12-month composite-event risk than the 12-month strategy, that result would describe the difference in the combined endpoint between randomized strategies. It would not by itself establish that every individual component was reduced.
A composite treatment effect would not mean that cardiovascular death, myocardial infarction, definite stent thrombosis, stroke, and bleeding each had identical treatment effects. Those outcomes have different clinical meanings and can occur at different frequencies.
Relative measures can describe the proportional difference between treatment groups, but an absolute risk difference directly describes the difference in event probability over the 12-month period. Both perspectives can be useful when interpreting a composite endpoint.
A confidence interval describes uncertainty around the estimated treatment effect under the specified statistical framework. A narrow interval indicates greater precision than a wide interval, while the location of the interval relative to the null value determines the range of effects compatible with the analysis.
A p-value addresses the compatibility of the observed data with a specified null hypothesis. It does not measure the magnitude or clinical importance of the treatment effect. Effect estimates and their confidence intervals are needed to describe magnitude and precision.
18. Important Limitations and Interpretation Issues
- Composite heterogeneity: the primary endpoint combines cardiovascular death, myocardial infarction, definite stent thrombosis, stroke, and bleeding, which are clinically different outcomes.
- Component weighting: the composite is driven by the events that occur, so the most frequent components can contribute more observations to the combined endpoint.
- Open-label design: masking is listed as none, which is relevant for outcomes or clinical decisions that may be influenced by knowledge of treatment assignment.
- Analysis specification: the registry endpoint description does not state whether the primary analysis uses a fixed-time risk comparison or a time-to-event framework.
- Missing-data rules: the registry record does not specify how missing 12-month endpoint assessments are handled.
- Multiplicity: formal testing of individual components or additional hypotheses requires an appropriate prespecified error-control strategy.
- Non-inferiority interpretation: the ClinicalTrials.gov record does not specify a non-inferiority margin, so such a margin cannot be used to interpret the endpoint from the registry description alone.
- Component-specific interpretation: a composite result should not be assumed to represent equal effects across all of its constituent outcomes.
19. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The trial is structured to compare two randomized DAPT-duration strategies on a prespecified 12-month composite endpoint. The key statistical task is to estimate the difference between those strategies while respecting the endpoint definition and analysis population.
Clinical interpretation
The endpoint combines mortality, ischemic events, stent thrombosis, stroke, and bleeding. Clinical interpretation therefore requires understanding which components contribute to the overall endpoint rather than treating the composite as a single homogeneous clinical event.
20. What Would Make the Statistical Result Most Informative?
A complete interpretation of the STOPDAPT-2 primary endpoint would ideally present the overall composite event frequency in each randomized group together with an absolute risk difference, a relative effect measure, a confidence interval, and the prespecified hypothesis test.
| Information | Why it matters |
|---|---|
| Composite event count | Shows how many participants experienced the primary endpoint. |
| Composite event proportion | Shows the absolute frequency of the endpoint by treatment strategy. |
| Absolute risk difference | Quantifies the difference in event probability between groups. |
| Relative effect estimate | Describes the proportional difference between treatment strategies. |
| Confidence interval | Quantifies uncertainty around the estimated treatment effect. |
| P-value | Addresses the prespecified hypothesis test but does not quantify effect magnitude. |
| Component counts | Shows which clinical events contribute to the composite. |
| Analysis population | Defines which randomized participants contribute to the primary analysis. |
| Missing-data and censoring rules | Clarifies how incomplete follow-up is incorporated into the analysis. |
21. Record Summary
STOPDAPT-2 is a randomized phase 4 parallel-group trial in coronary artery disease comparing 1-month versus 12-month DAPT. Its registered primary endpoint is a 12-month composite of cardiovascular death, myocardial infarction, definite stent thrombosis, stroke, and bleeding, with bleeding defined as major or minor under the TIMI Study group definition.
From a statistical perspective, the central challenge is interpreting a heterogeneous composite endpoint within a randomized comparison. The most informative analysis would combine the overall treatment effect with its confidence interval, the absolute event risks, and the individual component events. The distinction between a fixed-time risk analysis and a time-to-event analysis is also important because the registry endpoint description does not specify which framework was used.
The trial's randomized allocation supports a causal comparison of the assigned DAPT strategies, while the open-label design and composite nature of the endpoint are important considerations when interpreting the evidence. Results have not been posted on ClinicalTrials.gov, so the registry record supports analysis of the trial's design and planned statistical framework rather than a numerical treatment-effect conclusion.
22. Sources
- ClinicalTrials.gov: STOPDAPT-2, NCT02619760.
Continue through the Clinical Biostats clinical trial library
Clinical trial analysis connects study design, endpoint definitions, statistical estimands, and interpretation into a structured statistical framework.