This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. The ClinicalTrials.gov record contains the registered primary endpoint and trial-level design information, but does not provide a formal statistical-analysis result table for that endpoint.
1. Trial at a Glance
ACT I was a randomized, parallel-group clinical trial comparing carotid artery stenting with filter against carotid artery endarterectomy in asymptomatic subjects who were at standard risk for carotid endarterectomy and had significant extracranial carotid stenotic disease. The trial enrolled 1663 subjects and had 2 intervention arms.
| Feature | ACT I |
|---|---|
| Trial name | ACT I |
| NCT ID | NCT00106938 |
| Therapeutic area | Cardiovascular |
| Conditions | Carotid Artery Disease; Carotid Stenosis; Stroke; Atherosclerosis |
| Phase | NA |
| Status | TERMINATED |
| Allocation | RANDOMIZED |
| Design model | PARALLEL |
| Masking | NONE |
| Primary purpose | OTHER |
| Enrollment | 1663 |
| Lead sponsor | Abbott Medical Devices |
| Sponsor type | INDUSTRY |
2. Clinical Question
The trial addresses a comparative procedural question in subjects with significant extracranial carotid stenotic disease who were described in the registry as asymptomatic and at standard risk for carotid endarterectomy. The randomized comparison was between a catheter-based carotid artery stenting strategy and surgical carotid artery endarterectomy.
Population
Asymptomatic subjects who were at standard risk for carotid endarterectomy with significant extracranial carotid stenotic disease.
Intervention
Carotid artery stenting with filter, classified in the registry as an interventional device intervention.
Comparator
Carotid artery endarterectomy, classified in the registry as a surgical procedure.
Primary question
How does the randomized comparison of the two procedures perform on the registered composite of death, stroke, and myocardial infarction through 30 days post-procedure, plus ipsilateral stroke from 31 to 365 days?
3. Trial Design
Carotid artery stenting with filter
- Interventional device intervention
- Carotid artery stenting with filter
- Randomized comparison against carotid artery endarterectomy
Carotid artery endarterectomy
- Surgical procedure
- Carotid artery endarterectomy
- Randomized comparison against carotid artery stenting with filter
The registry identifies the allocation as randomized and the design model as parallel. Masking is listed as none, so this was an open-label procedural comparison rather than a blinded treatment assignment.
4. Trial Timeline
Trial start
The ACT I registry record gives 2005-04 as the study start.
Primary completion
The registry gives 2013-03 as the primary completion date.
Terminated
The ClinicalTrials.gov record lists the study status as TERMINATED.
5. Registered Primary Endpoint
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Primary endpoint | Composite of Death, Stroke (Ipsilateral or Contralateral; Major or Minor) and Myocardial Infarction (DSMI) Through 30 Days Post-procedure, Plus Ipsilateral Stroke 31 to 365 Days. | 0 to 365 days |
The registry identifies one primary endpoint and describes its type as “Other / unclear.” The endpoint is nevertheless interpretable as a composite clinical outcome with an early post-procedure component and an additional later ipsilateral-stroke component.
The important statistical feature is that the primary endpoint is not simply “stroke by 365 days.” It combines several clinical events during the first 30 days with a specific stroke outcome during days 31 through 365.
6. Understanding the Composite Endpoint
A composite endpoint combines multiple clinically defined events into a single primary outcome. In ACT I, the registered composite includes death, stroke, and myocardial infarction through 30 days after the procedure, followed by ipsilateral stroke during days 31 to 365.
Early component
Death, ipsilateral or contralateral stroke, major or minor stroke, and myocardial infarction are incorporated through 30 days post-procedure.
Later component
From 31 to 365 days, the registered endpoint adds ipsilateral stroke.
Single primary outcome
The composite creates one prespecified endpoint rather than requiring separate primary comparisons for every component.
Component interpretation
A composite result can be driven disproportionately by one component, so a complete statistical interpretation normally examines the components as well as the overall composite.
The ClinicalTrials.gov record does not provide component-specific event counts, treatment-group event rates, effect estimates, confidence intervals, or p-values for the primary endpoint.
7. Planned Analysis
ClinicalTrials.gov lists the primary endpoint as having results posted, but the ClinicalTrials.gov record contains no formal statistical analyses. Consequently, there is no registry-reported treatment-effect estimate, confidence interval, or p-value to report for the primary endpoint on this page.
How this endpoint would typically be analysed
For a randomized two-arm comparison with a binary composite outcome assessed over a prespecified period, a conventional analysis would compare the proportion of subjects experiencing the composite endpoint between randomized groups. Depending on the statistical analysis plan, the comparison could be expressed as a risk difference, risk ratio, odds ratio, or a time-to-event measure if the analysis explicitly incorporated event timing and censoring.
Because the registered endpoint has a defined 30-day component followed by a 31-to-365-day component, a time-to-event framework could also be appropriate when exact event times and censoring information are available. However, the registry-reported ACT I data do not identify the formal analysis method used for the primary endpoint, so no particular model should be attributed to the trial.
8. What a Formal Treatment Effect Would Mean
If a formal comparison of the primary composite had been reported, its interpretation would depend on the effect measure. An absolute risk difference would describe the difference in the proportion experiencing the endpoint. A risk ratio would compare those proportions multiplicatively. A hazard ratio, if a time-to-event analysis were prespecified, would instead compare estimated instantaneous event rates over follow-up.
The absence of a reported effect estimate in the ClinicalTrials.gov record is itself important. It means that this page cannot legitimately state that one procedure reduced, increased, or had the same rate of the primary endpoint relative to the other. The appropriate conclusion from the ClinicalTrials.gov record is limited to the trial's randomized design and registered endpoint definition.
A p-value addresses compatibility of the observed data with a specified null hypothesis under the statistical model. It does not quantify the size of a treatment effect or its clinical importance. A complete result would ideally present an effect estimate together with a confidence interval and the corresponding hypothesis-test result.
A confidence interval would describe statistical uncertainty around the estimated treatment effect under the chosen analysis framework. A narrow interval would indicate greater precision than a wide interval, but neither width nor statistical significance alone establishes clinical importance.
9. Safety Results
The ClinicalTrials.gov record includes serious adverse events by randomized arm. These data are presented as affected subjects over subjects at risk.
| Safety measure | Carotid artery stenting with filter | Carotid artery endarterectomy |
|---|---|---|
| Serious adverse events | 376 / 1089 | 118 / 364 |
These are the safety counts reported in the ClinicalTrials.gov record. They should not be converted here into percentages because the task requires reported numbers to remain exactly as reported in the registry and prohibits recomputation except for simple derived statements. More importantly, the affected/at-risk counts do not by themselves provide a formal statistical comparison.
10. Analysis Populations and Randomization
The registry identifies the allocation as RANDOMIZED and the design model as PARALLEL. These are central statistical features because randomization establishes the basis for comparing outcomes according to assigned intervention, while the parallel design means participants remain in their assigned comparison groups rather than sequentially receiving both interventions as part of a crossover design.
| Design feature | Registry information |
|---|---|
| Allocation | RANDOMIZED |
| Design model | PARALLEL |
| Masking | NONE |
| Arms | 2 |
| Enrollment | 1663 |
The ClinicalTrials.gov record does not provide an analysis-population definition such as intention-to-treat, per-protocol, modified intention-to-treat, or as-treated populations. It would therefore be inappropriate to assign one of these labels to the primary efficacy analysis without additional source information.
11. Statistical Methodology
Randomized comparison
Randomization is the principal design mechanism that supports a causal comparison between the two intervention strategies. In a properly conducted randomized trial, assignment is determined by the randomization procedure rather than by observed prognostic characteristics. This helps balance measured and unmeasured factors in expectation.
The actual statistical expression depends on the chosen endpoint and effect measure. For ACT I, the ClinicalTrials.gov record establishes the two randomized strategies but do not supply the formal effect estimate for the primary composite.
Binary-event analysis
If the primary composite were analysed simply as whether an event occurred during the prespecified observation period, the analysis would reduce the outcome to an event indicator for each participant. Common measures would include a risk difference, risk ratio, or odds ratio.
For example, if \(p_1\) and \(p_0\) represented the event probabilities in the two randomized groups, the risk difference would be \(p_1-p_0\), while the risk ratio would be \(p_1/p_0\). These measures answer different questions and should not be treated as interchangeable.
Time-to-event analysis
If exact event times and censoring information were incorporated, the primary composite could instead be evaluated using survival-analysis methods. Kaplan-Meier estimation could describe the event-free probability over time, while a log-rank test could compare event-time distributions and a Cox model could estimate a hazard ratio.
Here \(S(t)\) represents the probability of remaining free of the event beyond time \(t\). The appropriate estimand for ACT I would depend on the prespecified statistical analysis plan and the way the composite endpoint was operationalized.
Composite endpoint analysis
The statistical analysis of a composite should preserve the registered component definitions. Combining death, stroke, and myocardial infarction into one endpoint can improve statistical efficiency when the components are clinically relevant and expected to occur sufficiently often, but it also means that the overall result can reflect different components to different degrees.
Confidence intervals
A confidence interval should accompany the primary effect estimate when available. It provides information about precision and helps distinguish a precisely estimated small effect from a highly uncertain estimate. The registry-reported ACT I data contain no confidence interval for the primary endpoint.
12. Statistical Methods Explained
Why was randomization important in ACT I?
Randomization makes the treatment groups comparable in expectation and provides the foundation for interpreting between-group differences as consequences of the assigned intervention rather than simply differences in who received each procedure. It does not guarantee identical groups in every characteristic, but it removes systematic treatment assignment based on known or unknown prognostic factors when the randomization is properly implemented.
Why is the primary endpoint a composite?
The registered endpoint combines several clinically relevant events into one outcome. Statistically, this can increase the number of subjects experiencing the primary endpoint compared with relying on a single component, potentially improving the efficiency of the comparison. The tradeoff is interpretational: the composite must be understood component by component because a treatment effect on the composite does not necessarily imply the same magnitude of effect on every component.
Why does the 30-day boundary matter?
The registry explicitly defines the first component as occurring through 30 days post-procedure and then specifies ipsilateral stroke from 31 to 365 days. This creates a structured endpoint rather than a generic one-year event rate. An analysis that ignored these definitions could measure a different outcome from the registered primary endpoint.
Why distinguish ipsilateral and contralateral stroke?
The endpoint definition treats stroke later in follow-up differently from stroke during the first 30 days. During the first 30 days, the composite includes ipsilateral or contralateral stroke, major or minor. From 31 to 365 days, the registered component is ipsilateral stroke. The distinction therefore matters directly to endpoint construction.
What would a hazard ratio mean if the trial used one?
A hazard ratio would compare estimated instantaneous event rates between the randomized groups over time. A value below 1 would indicate a lower estimated hazard in the numerator group, while a value above 1 would indicate a higher estimated hazard. It would not mean that the same percentage of subjects experienced a corresponding reduction in absolute probability.
Why is censoring important in survival analysis?
In a time-to-event analysis, some subjects may reach the end of observation without experiencing the event or may have incomplete follow-up. Censoring allows their available follow-up to contribute information without pretending that their event status after the censoring time is known. Valid interpretation depends on the censoring assumptions and prespecified analysis rules.
13. Missing Data and Censoring
The ClinicalTrials.gov record does not specify a missing-data strategy, imputation method, censoring rules, or estimand for incomplete follow-up. These details can materially affect a clinical-trial analysis and therefore should not be inferred.
Missing outcome data
The ClinicalTrials.gov record does not state how missing primary-endpoint information was handled.
Imputation
No imputation method is identified in the ClinicalTrials.gov record.
Censoring
No specific censoring rules are posted on ClinicalTrials.gov for the primary endpoint.
Estimand
No formal estimand framework is provided in the ClinicalTrials.gov record.
This is particularly important when distinguishing a fixed-period binary analysis from a time-to-event analysis. The former focuses on whether an event occurred during the defined period, whereas the latter uses the timing of events and censoring information.
14. Multiplicity and Endpoint Structure
The registry lists 1 primary endpoint. That simplifies the primary-endpoint multiplicity problem relative to a trial with several co-primary or multiple primary endpoints. However, the primary endpoint itself contains several component events.
| Feature | ACT I | Statistical implication |
|---|---|---|
| Registered primary endpoints | 1 | One registered primary endpoint is identified. |
| Primary endpoint structure | Composite | Component interpretation remains important. |
| Outcome measures posted | 33 | Many posted outcomes do not automatically become primary confirmatory endpoints. |
| Formal statistical analyses reported | None | No reported treatment-effect estimate can be attributed to the primary endpoint from the ClinicalTrials.gov record. |
The presence of 33 posted outcome measures does not mean that 33 hypothesis tests were primary confirmatory tests. Statistical multiplicity depends on the prespecified inferential hierarchy and analysis plan, neither of which is reported here.
15. Non-Inferiority, Crossover, and Bayesian Methods
The registry-reported ACT I data do not identify a non-inferiority margin, crossover design, interim-analysis rule, or Bayesian analysis. These topics therefore should not be presented as trial methods.
| Design topic | What the ClinicalTrials.gov record supports |
|---|---|
| Non-inferiority margin | Not reported in the ClinicalTrials.gov record. |
| Crossover | Not reported in the ClinicalTrials.gov record. |
| Interim analysis | Not reported in the ClinicalTrials.gov record. |
| Bayesian methods | Not reported in the ClinicalTrials.gov record. |
| Missing-data / imputation strategy | Not reported in the ClinicalTrials.gov record. |
| Stratification factors | Not reported in the ClinicalTrials.gov record. |
16. Serious Adverse Events: Statistical Interpretation
The registry supplies serious adverse-event counts as affected subjects divided by subjects at risk: 376/1089 for the carotid artery stenting group and 118/364 for the carotid artery endarterectomy group.
They identify the number of subjects affected by serious adverse events and the corresponding number at risk in each reported arm. They are useful for describing the posted safety data at a basic level.
They do not provide a confidence interval, p-value, adjusted effect estimate, time-to-event comparison, or formal causal comparison. They also do not establish whether the serious adverse events were distributed similarly across individual clinical categories.
A rigorous safety analysis would normally consider the definition of the safety population, exposure duration, event timing, recurrent events, seriousness criteria, and whether the comparison is based on subjects with at least one event or on event counts. Those details are not reported here.
17. Results Interpretation: What Can and Cannot Be Concluded
Supported by the ClinicalTrials.gov record
ACT I was randomized, had 2 parallel arms, enrolled 1663 subjects, and registered a composite primary endpoint assessed over 0 to 365 days.
Supported safety information
Serious adverse-event counts are reported as 376/1089 for carotid artery stenting with filter and 118/364 for carotid artery endarterectomy.
Not reported
No formal primary-endpoint effect estimate, confidence interval, or p-value appears in the posted results.
Therefore
This distinction is fundamental to independent statistical review. A trial can have posted results without every formal statistical comparison being available in a structured analysis record. The correct response is to preserve the boundary between what is documented and what would ordinarily be analysed.
18. Limitations
- No formal statistical analyses reported: no formal statistical analyses were posted, so no primary treatment-effect estimate, confidence interval, or p-value is available for reporting.
- Composite endpoint complexity: the primary outcome combines several early events and a later ipsilateral-stroke component. Interpretation would ideally examine the individual components as well as the composite.
- Limited analysis-method detail: the ClinicalTrials.gov record does not identify the formal statistical model, analysis population, stratification factors, or estimand.
- Missing-data strategy not reported:
- No subgroup estimates reported: subgroup treatment effects cannot be evaluated from the provided statistical-analysis data.
- No formal multiplicity framework reported: although one primary endpoint is registered, the handling of the 33 posted outcome measures is not specified.
- Safety interpretation is limited: serious adverse-event counts are reported, but no formal statistical comparison or detailed event-category analysis is included.
- Open-label design: masking is listed as none. The implications for outcome assessment depend on procedures not described in the ClinicalTrials.gov record.
19. Why This Trial Matters Statistically
ACT I is a useful teaching example because its design makes several core clinical-trial concepts visible without requiring a large collection of secondary statistical models. The trial compares two procedural strategies under randomization, uses a clinically structured composite endpoint, follows subjects across two distinct time windows, and reports safety by randomized arm.
| Concept | How it appears in ACT I |
|---|---|
| Randomization | Allocation is explicitly listed as RANDOMIZED. |
| Parallel design | The design model is PARALLEL with 2 arms. |
| Composite endpoint | The primary endpoint combines death, stroke, and myocardial infarction through 30 days with ipsilateral stroke from 31 to 365 days. |
| Fixed follow-up window | The registered primary endpoint has a 0 to 365 day time frame. |
| Component interpretation | Early and later endpoint components have different definitions. |
| Open-label design | Masking is listed as NONE. |
| Safety analysis | Serious adverse events are reported by arm. |
| Statistical reporting discipline |
20. A Statistical Reading of the Primary Endpoint
The most important statistical lesson from ACT I is that endpoint definition precedes endpoint analysis. Before calculating a risk difference, risk ratio, odds ratio, or hazard ratio, the analyst must know exactly what counts as an event and during which period it counts.
For the first 30 days, the registered composite includes death, ipsilateral or contralateral stroke, major or minor stroke, and myocardial infarction. From day 31 through day 365, the registered endpoint includes ipsilateral stroke. These rules create a specific composite estimand rather than a generic measure of one-year procedural safety.
Skipping the first steps can produce a numerically precise answer to the wrong clinical question. ACT I illustrates why statistical analysis begins with the registered endpoint definition.
21. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The trial provides a randomized framework for comparing two procedural strategies, but the ClinicalTrials.gov record does not contain the formal estimate for the registered primary endpoint.
Clinical interpretation
The registry identifies the clinical outcomes incorporated into the primary endpoint and provides serious adverse-event counts by arm. Those facts describe what was measured, not which procedure produced a better outcome.
This distinction prevents an important analytical error: treating trial design as if it were trial evidence. Randomization determines how a comparison can be interpreted; the observed effect estimate and its uncertainty determine what the reported data actually show.
22. What a Complete Results Table Would Normally Contain
For an endpoint such as the ACT I primary composite, a complete statistical results presentation would normally make several pieces of information visible together.
| Component | Purpose | Available in the ClinicalTrials.gov record? |
|---|---|---|
| Event count by randomized arm | Shows the observed number of primary-endpoint events. | No |
| Subjects analysed / at risk | Defines the denominator for the efficacy comparison. | No primary-endpoint denominator reported |
| Effect estimate | Quantifies the between-group difference or ratio. | No |
| 95% confidence interval | Quantifies statistical uncertainty around the estimate. | No |
| p-value | Provides the hypothesis-test result under the specified analysis. | No |
| Analysis method | Identifies the statistical model or test used. | No formal method reported |
| Component results | Shows whether the composite is driven by particular event types. | No |
23. Sources
- ClinicalTrials.gov: ACT I — NCT00106938. Official registry record for the trial and the source of the trial data used on this page.
- Linked publication: PubMed PMID 28982691.
- Linked publication: PubMed PMID 26886419.
24. Related Tutorials
Learn more about the statistical concepts that form the ACT I learning pathway:
25. Related Statistical Calculators
26. Record Summary
ACT I is a randomized, parallel-group cardiovascular trial comparing carotid artery stenting with filter against carotid artery endarterectomy in asymptomatic subjects at standard risk for carotid endarterectomy with significant extracranial carotid stenotic disease. The trial enrolled 1663 subjects, used 2 arms, and registered one primary composite endpoint covering 0 to 365 days.
The primary endpoint is statistically distinctive because its definition changes across the follow-up period: death, stroke, and myocardial infarction are included through 30 days post-procedure, while ipsilateral stroke is included from 31 to 365 days. That structure illustrates why a statistical analysis must follow the registered endpoint definition rather than substituting a simpler one-year event measure.
The ClinicalTrials.gov record also report serious adverse events by arm as 376/1089 for carotid artery stenting with filter and 118/364 for carotid artery endarterectomy. However, the ClinicalTrials.gov record contains no formal analyses, so no treatment-effect estimate, confidence interval, or p-value for the primary endpoint is reported here.
Continue through the Clinical Biostats statistical pathway
Use the related tutorials and calculators to explore randomization, composite endpoints, effect measures, confidence intervals, and time-to-event analysis.