This page separates reported trial results from statistical interpretation. Numerical results on this page are restricted to the ClinicalTrials.gov data posted on ClinicalTrials.gov for CREST. Where the registry does not report a statistical method, p-value, or additional result, this page does not infer one from external trial knowledge.
1. Trial at a Glance
CREST was a randomized, parallel-design, double-masked clinical trial evaluating carotid endarterectomy (CEA) versus carotid artery stenting (CAS) for prevention in patients with carotid stenosis. The registered primary endpoint was a time-to-event composite observed during a 30-day peri-procedural period and through 4 years of follow-up.
| Feature | CREST |
|---|---|
| Trial name | Carotid Revascularization Endarterectomy Versus Stenting Trial |
| Therapeutic area | Neurology |
| Status | Completed |
| Start | 2000-12 |
| Primary completion | 2011-02 |
| Lead sponsor | Rutgers, The State University of New Jersey |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Double |
| Primary purpose | Prevention |
| Enrollment | 2,502 |
| Interventions | Carotid Endarterectomy (CEA); Carotid Artery Stenting (CAS) |
| Primary endpoint type | Time-to-event |
| Hypothesis type | Superiority |
| ClinicalTrials.gov | NCT00004732 |
2. Clinical Question
The registered primary aim was to assess whether the efficacy of CAS differs from that of CEA in preventing stroke, myocardial infarction, and death during a 30-day peri-procedural period, or ipsilateral stroke during follow-up in patients with symptomatic or asymptomatic extracranial carotid stenosis.
Population
Patients with symptomatic (≥50%) or asymptomatic (≥60%) extracranial carotid stenosis, within the condition framework registered for CREST.
Intervention
Carotid artery stenting (CAS), registered as a device intervention.
Comparator
Carotid endarterectomy (CEA), registered as a procedure intervention.
Primary question
Does the efficacy of CAS differ from that of CEA for the registered composite endpoint of peri-procedural stroke, myocardial infarction, or death and subsequent ipsilateral stroke?
3. Trial Design
Carotid-Artery Stenting
- Intervention: Carotid Artery Stenting (CAS)
- Intervention type: Device
Carotid Endarterectomy
- Intervention: Carotid Endarterectomy (CEA)
- Intervention type: Procedure
the ClinicalTrials.gov record does not provide arm-specific randomized sample sizes. The ClinicalTrials.gov record therefore support the overall enrollment of 2,502, but not an arm-level enrollment table.
4. Endpoints
The registered primary endpoint is a composite time-to-event outcome. Its design combines events occurring during the peri-procedural period with a specific postprocedural stroke outcome during longer follow-up.
| Endpoint | Registry definition / time frame | Type |
|---|---|---|
| Primary endpoint | Any Periprocedural Stroke, Myocardial Infarction, or Death During a 30-day Peri-procedural Period, and Postprocedural Ipsilateral Stroke Thereafter, up to 4-years. | Time-to-event |
| Primary endpoint time frame | 30 days and 4 years | Composite time-to-event |
The registry definition states that the primary aim was to assess whether the efficacy of CAS differs from CEA in preventing stroke, myocardial infarction, and death during the 30-day peri-procedural period, or ipsilateral stroke over follow-up, among patients with symptomatic (≥50%) or asymptomatic (≥60%) extracranial carotid stenosis.
5. Primary Endpoint Result
ClinicalTrials.gov reports one formal primary-endpoint statistical analysis. The comparison is CAS versus CEA, the effect measure is a hazard ratio, and the reported estimate is adjusted for age, sex, and symptomatic status.
Primary endpoint hazard ratio
95% CI: 0.81–1.51 · Two-sided 95% confidence interval
Adjusted for age, sex, and symptomatic status
| Endpoint | Comparison | Estimate | 95% CI | Adjustment | P-value |
|---|---|---|---|---|---|
| Any peri-procedural stroke, myocardial infarction, or death during 30 days, and postprocedural ipsilateral stroke thereafter, up to 4 years | CAS vs CEA | HR 1.11 | 0.81–1.51 | Age, sex, symptomatic status | Not reported in reported ClinicalTrials.gov statistical analysis |
An HR of 1.11 means that the fitted relative hazard for the primary time-to-event endpoint was estimated to be 11% higher for CAS than for CEA, based on the adjusted hazard-ratio model reported in the registry.
That statement does not mean that 11% more participants experienced the endpoint. A hazard ratio is a relative measure of the event rate over time; it is not an absolute risk difference, a risk ratio, or a difference in percentages of participants with events.
The 95% CI of 0.81–1.51 describes the uncertainty around the estimated hazard ratio under the statistical framework used for the analysis. Because the interval spans 1, the ClinicalTrials.gov record does not by itself establish that the two randomized strategies have different hazards for the primary endpoint at the conventional two-sided 5% level.
The ClinicalTrials.gov record does not report a p-value for this analysis, so a p-value should not be reconstructed from the confidence interval. In particular, the confidence interval is the reported measure of statistical uncertainty, while a p-value would address a separate hypothesis-testing question.
The analysis is adjusted for age, sex, and symptomatic status. Adjustment does not turn the hazard ratio into an absolute treatment effect; it specifies how those covariates enter the reported comparison.
6. What the Primary Hazard Ratio Means
Here, the first-listed group is CAS and the comparator is CEA. The reported estimate of 1.11 therefore lies above the null value of 1.
For a time-to-event endpoint, the hazard is an instantaneous event rate conditional on remaining event-free up to a particular point in time. A hazard ratio compares those modeled instantaneous rates between groups over the analysis.
The distinction matters because the CREST primary endpoint has a complicated temporal structure. During the first 30 days, the endpoint includes stroke, myocardial infarction, or death. After that peri-procedural period, the registered endpoint follows postprocedural ipsilateral stroke through 4 years. The resulting hazard ratio summarizes the specified time-to-event comparison rather than representing a single fixed percentage difference in event probabilities.
What HR 1.11 does mean
Within the reported adjusted time-to-event analysis, the estimated hazard for CAS relative to CEA was 1.11.
What HR 1.11 does not mean
It does not mean that 11% of participants had an event or that the absolute event probability differed by 11 percentage points.
What the CI tells us
The 0.81–1.51 interval quantifies uncertainty around the estimated hazard ratio; it does not give the range of outcomes for individual participants.
What the p-value tells us
No p-value is in the ClinicalTrials.gov record in the data provided, so the page does not assign a formal p-value interpretation to the result.
7. Secondary Endpoint Result
The ClinicalTrials.gov record contains one formal secondary statistical analysis. It evaluates differential efficacy in male and female participants using the primary endpoint, with the reported statistical result specifically identified as the HR for women comparing CAS with CEA.
Women: primary endpoint hazard ratio
95% CI: 0.82–2.23 · Two-sided 95% confidence interval
Adjusted for age and symptomatic status
| Secondary analysis | Comparison | Estimate | 95% CI | Adjustment | P-value |
|---|---|---|---|---|---|
| Differential efficacy of CAS and CEA in male and female participants in the primary endpoint | Women: CAS vs CEA | HR 1.35 | 0.82–2.23 | Age, symptomatic status | Not reported in reported ClinicalTrials.gov statistical analysis |
The reported HR of 1.35 for women means that, in the specified adjusted time-to-event analysis, the estimated hazard for CAS relative to CEA was 1.35. On the hazard-ratio scale, this is an estimate above 1.
The 95% CI of 0.82–2.23 is substantially wider than the primary endpoint's interval of 0.81–1.51. That wider interval indicates greater statistical uncertainty around the female-participant estimate than around the overall primary comparison.
Importantly, a subgroup estimate should not be interpreted as evidence that the treatment effect differs between sexes merely because its point estimate differs from the overall estimate. A formal claim of effect modification requires an appropriate interaction or heterogeneity analysis. The registry-reported statistical analysis identifies this as an analysis of differential efficacy, but it does not provide an interaction p-value.
The registry data also do not provide the corresponding male-specific HR in the registry-reported statistical analysis. Therefore, the page does not infer one from the female estimate or from the overall result.
8. Primary vs Secondary Analysis
| Analysis | Role | HR | 95% CI | Covariate adjustment |
|---|---|---|---|---|
| Primary endpoint | Primary, superiority | 1.11 | 0.81–1.51 | Age, sex, symptomatic status |
| Women: primary endpoint | Secondary, differential efficacy | 1.35 | 0.82–2.23 | Age, symptomatic status |
These are not interchangeable estimates. The primary analysis is the overall CAS-versus-CEA comparison adjusted for age, sex, and symptomatic status. The secondary analysis is specifically the reported estimate for women, adjusted for age and symptomatic status.
The point estimates differ, but their confidence intervals overlap substantially. More importantly, a difference between point estimates is descriptive rather than a formal test of interaction. The ClinicalTrials.gov record contains no interaction statistic or interaction p-value.
9. Statistical Methodology
The registry identifies the primary endpoint as a time-to-event outcome and reports a hazard ratio as its effect measure. It also identifies covariate adjustment for the primary and secondary analyses. The specific statistical method is otherwise recorded as "Not reported" in the ClinicalTrials.gov record.
| Methodological element | What the ClinicalTrials.gov record establishes |
|---|---|
| Endpoint framework | Time-to-event |
| Effect measure | Hazard ratio |
| Primary hypothesis | Superiority |
| Primary adjustment | Age, sex, symptomatic status |
| Secondary adjustment for women | Age, symptomatic status |
| Formal method named in statistical analysis | Not reported |
| Primary p-value | Not reported in registry-reported statistical analysis |
| Secondary p-value | Not reported in registry-reported statistical analysis |
Kaplan-Meier estimation
For a time-to-event endpoint, Kaplan-Meier estimation is a standard way to describe the event-free probability over follow-up while accounting for right censoring. The registry-reported CREST statistical-analysis record does not explicitly name Kaplan-Meier estimation, so it is presented here as the typical statistical framework for this endpoint type, not as a claim about an unreported registry method.
Here, di is the number of events at event time ti, and ni is the number at risk immediately before that time.
Cox proportional-hazards modeling
A hazard ratio is commonly estimated with a Cox proportional-hazards model. In a covariate-adjusted model, the treatment comparison is estimated while accounting for specified covariates. For CREST, the registry-reported analysis notes explicitly identify adjustment for age, sex, and symptomatic status in the primary analysis.
This equation is an educational representation of how a covariate-adjusted proportional-hazards model can be expressed. The ClinicalTrials.gov record does not specify the exact fitted model formulation.
Covariate adjustment
The primary analysis reports an HR adjusted for age, sex and symptomatic status. The secondary women-specific analysis reports an HR adjusted for age and symptomatic status. Adjustment can improve precision or account for prognostic covariates, but it does not change the randomized assignment itself.
Right censoring
Time-to-event analyses can include participants who have not experienced the endpoint by the end of their observed follow-up. Such observations are censored rather than treated as if the endpoint never occurred. The ClinicalTrials.gov record does not report the number or timing of censored observations, so no censoring distribution is inferred.
10. Statistical Methods Explained
Why is this endpoint analyzed as time-to-event rather than simply as a percentage?
The registered endpoint has a time structure: it distinguishes events during a 30-day peri-procedural period from postprocedural ipsilateral stroke thereafter, with follow-up extending to 4 years. A time-to-event analysis can use both the timing of observed events and the follow-up contributed by participants, rather than reducing the entire study to a single proportion.
What does an HR of 1.11 mean?
It means the estimated hazard for CAS relative to CEA was 1.11 in the reported adjusted analysis. On this scale, 1 is the null value. An HR of 1.11 is therefore an estimate above the null, but it should not be translated into an 11-percentage-point difference in event probability.
Why was covariate adjustment used?
The primary analysis explicitly reports adjustment for age, sex, and symptomatic status. In a time-to-event regression model, adjustment incorporates these covariates into the estimated treatment comparison. The purpose is to estimate the treatment association conditional on the specified covariates, rather than reporting an unadjusted comparison.
Why is the confidence interval important?
The point estimate alone gives only one estimated hazard ratio. The 95% CI of 0.81–1.51 communicates the uncertainty surrounding that estimate. The interval includes values below 1 and above 1, so the ClinicalTrials.gov record does not establish a statistically distinguishable hazard ratio from the null value of 1 on the basis of the interval alone.
Does a wider subgroup confidence interval mean the subgroup result is "worse"?
No. A wider interval means less precision, not a direction of treatment benefit or harm. The women-specific estimate has a 95% CI of 0.82–2.23, which spans a wider range of plausible hazard ratios than the overall estimate's 0.81–1.51 interval. Precision and the direction of the point estimate are separate concepts.
Why can't the women-specific HR establish a sex difference by itself?
A subgroup point estimate answers a subgroup-specific question. To determine whether treatment effects differ between subgroups, the relevant statistical question is an interaction or treatment-by-subgroup effect. The registry-reported CREST analysis does not provide an interaction estimate or p-value, so the difference between 1.35 in women and 1.11 overall should not be treated as a formal test of sex-based effect modification.
Why is the p-value not reported here?
The ClinicalTrials.gov record provides the hazard-ratio estimates and two-sided 95% confidence intervals but does not provide p-values. A p-value should therefore not be invented or reverse-engineered and presented as if it were reported by the registry.
11. Composite Endpoint Interpretation
Composite endpoints require careful interpretation because they combine multiple clinical event types. CREST's registered primary endpoint contains stroke, myocardial infarction, and death during the 30-day peri-procedural period, followed by postprocedural ipsilateral stroke thereafter through 4 years.
Early period
The first component covers a 30-day peri-procedural period and includes three event types: stroke, myocardial infarction, or death.
Later period
After the peri-procedural period, the registered endpoint follows postprocedural ipsilateral stroke through 4 years.
Why this matters
A single HR summarizes the composite time-to-event endpoint; it does not tell us from the ClinicalTrials.gov record which component drove the estimate.
What is not reported
The ClinicalTrials.gov record does not provide component-specific event estimates or p-values for the individual components.
This distinction prevents a common analytical mistake: interpreting the primary HR as though it were an effect estimate for stroke alone, myocardial infarction alone, or death alone. The reported estimate belongs to the registered composite endpoint.
12. Safety Results
The ClinicalTrials.gov record reports serious adverse events by randomized arm as affected participants divided by participants at risk.
| Arm | Serious adverse events affected | At risk |
|---|---|---|
| Carotid-Artery Stenting | 322 | 1,262 |
| Carotid Endarterectomy | 292 | 1,240 |
These safety data should be kept conceptually separate from the primary efficacy hazard ratio. The primary endpoint is a time-to-event efficacy comparison, whereas the registry-reported safety field is an affected/at-risk summary. They are different statistical quantities and should not be combined into a single measure.
13. Randomization and Covariate Adjustment
Randomization is the core design feature that creates the basis for comparing CAS and CEA. Because treatment assignment is randomized, the treatment groups are defined by their assigned intervention rather than by an investigator's selection of which procedure a participant should receive.
The primary statistical analysis nevertheless adjusts the hazard ratio for age, sex, and symptomatic status. This is compatible with the distinction between randomization and regression adjustment: randomization establishes the treatment comparison, while covariate adjustment specifies a statistical model for estimating that comparison.
| Feature | Primary analysis | Women-specific secondary analysis |
|---|---|---|
| Comparison | CAS vs CEA | CAS vs CEA among women |
| Effect measure | Hazard ratio | Hazard ratio |
| Age adjustment | Yes | Yes |
| Sex adjustment | Yes | Not applicable to the women-only estimate |
| Symptomatic-status adjustment | Yes | Yes |
| Estimate | 1.11 | 1.35 |
| 95% CI | 0.81–1.51 | 0.82–2.23 |
The difference in adjustment sets is important. The overall analysis adjusts for sex because the overall population contains male and female participants. The reported women-specific analysis does not adjust for sex because it is already restricted to women.
14. Superiority Hypothesis
the ClinicalTrials.gov record identifies the hypothesis type as superiority. In a superiority framework, the statistical question is whether the randomized intervention groups differ rather than whether one treatment remains within a prespecified non-inferiority margin.
This is an educational representation of the usual two-sided hazard-ratio testing framework. The ClinicalTrials.gov record identifies the hypothesis type as superiority and the confidence interval as two-sided, but do not provide the formal testing statement or p-value.
The reported primary 95% CI is 0.81–1.51. Because this interval includes the null hazard ratio of 1, the confidence interval alone does not provide evidence of a statistically distinguishable treatment difference at the conventional two-sided 5% level.
This is different from saying that CAS and CEA are proven identical. A confidence interval containing 1 is compatible with a range of effects, including values below and above the null. The interval therefore describes uncertainty rather than establishing exact equivalence.
15. Censoring and Time-to-Event Follow-up
The primary endpoint has a maximum registered follow-up of 4 years, with a distinct 30-day peri-procedural period. Time-to-event methods are useful in this setting because participants may contribute different amounts of observable follow-up.
A participant who remains free of the endpoint through the last available observation does not necessarily provide information equivalent to a participant observed for the full 4-year period. Survival-analysis methods account for the amount of follow-up available before censoring.
Event time
The analysis uses the timing of the registered endpoint rather than only whether an event ever occurred.
Censoring
Participants without an observed event by the end of their available follow-up can be censored in a time-to-event framework.
Four-year horizon
The registered primary endpoint follows participants up to 4 years.
Data limitation
The ClinicalTrials.gov record does not report event counts by time, censoring times, or a Kaplan-Meier table.
16. Subgroup Analysis: Women
The registry-reported secondary analysis specifically addresses differential efficacy in male and female participants, but the posted statistical estimate reported here is the women-specific hazard ratio.
| Subgroup analysis | HR, CAS vs CEA | 95% CI | Adjustment |
|---|---|---|---|
| Women | 1.35 | 0.82–2.23 | Age and symptomatic status |
The width of 0.82–2.23 illustrates why subgroup estimates need careful interpretation. The estimate is compatible with a lower hazard, a hazard close to the null, or a higher hazard for CAS relative to CEA within the range represented by the confidence interval.
17. Multiplicity and Other Design Features
The registry-reported CREST data support discussion of the primary endpoint, one secondary analysis, covariate adjustment, randomization, masking, and time-to-event methodology. They do not provide information about a non-inferiority margin, crossover, factorial design, interim analysis, alpha spending, missing-data imputation, or Bayesian methods.
| Design / analysis topic | What can be established from the ClinicalTrials.gov record |
|---|---|
| Superiority | Yes — identified in the trial data. |
| Non-inferiority margin | Not reported. |
| Crossover | Not reported. |
| Factorial design | No — the registered design model is parallel. |
| Interim analysis | Not reported. |
| Multiplicity procedure | Not reported. |
| Missing-data / imputation method | Not reported. |
| Bayesian methods | Not reported. |
| Covariate adjustment | Yes — age, sex and symptomatic status for the primary analysis. |
This distinction is important for statistical interpretation. Absence of a method in the ClinicalTrials.gov record is not evidence that the method was or was not used elsewhere in the trial's complete statistical documentation. It means only that the ClinicalTrials.gov record does not establish it.
18. Understanding the Confidence Intervals
The primary HR of 1.11 has a 95% CI of 0.81–1.51. The women-specific HR of 1.35 has a 95% CI of 0.82–2.23. Both intervals are two-sided.
| Analysis | Point estimate | Lower 95% CI | Upper 95% CI | Null value included? |
|---|---|---|---|---|
| Primary endpoint | 1.11 | 0.81 | 1.51 | Yes |
| Women | 1.35 | 0.82 | 2.23 | Yes |
A confidence interval should be read as an uncertainty statement about the estimated parameter under the model and sampling framework. It should not be interpreted as saying that there is a 95% probability that the true hazard ratio lies inside this particular interval.
The confidence intervals also show why point estimates should not be overinterpreted. The women-specific estimate of 1.35 is farther from 1 than the overall estimate of 1.11, but its upper limit extends to 2.23. The interval therefore provides substantially less precision about the size and direction of the subgroup effect.
19. Clinical Biostats Interpretation of the Complete Statistical Record
The reported primary comparison is CAS versus CEA for a composite time-to-event endpoint. The adjusted HR is 1.11, with a two-sided 95% CI of 0.81–1.51. The registry does not provide a p-value in the ClinicalTrials.gov record.
The HR point estimate lies above 1, but the confidence interval spans both sides of the null. The appropriate statistical description is therefore the reported estimate together with its uncertainty, rather than treating the point estimate alone as a definitive difference between the procedures.
For women, the reported HR is 1.35 with a 95% CI of 0.82–2.23. This estimate is less precise than the overall estimate and does not, by itself, establish that sex modifies the treatment effect.
The ClinicalTrials.gov record reports serious adverse events in 322/1262 participants in the CAS arm and 292/1240 in the CEA arm. These are affected/at-risk summaries, not the same statistical quantity as the primary hazard ratio.
20. Important Limitations and Interpretation Issues
- Method reporting: The ClinicalTrials.gov record does not name the formal statistical method used to obtain the hazard ratio. Kaplan-Meier and Cox modeling are discussed as standard educational methods for this endpoint type, not asserted as registry-reported methods.
- No reported p-values: The registry-reported primary and secondary analyses contain estimates and confidence intervals but no p-values. No p-value is inferred.
- Composite endpoint: The primary HR applies to the registered composite endpoint rather than to stroke, myocardial infarction, death, or postprocedural ipsilateral stroke individually.
- Time structure: The endpoint changes its event definition after the 30-day peri-procedural period, which should be considered when interpreting the single hazard ratio.
- Subgroup precision: The women-specific confidence interval of 0.82–2.23 is wide, indicating substantial uncertainty around the subgroup estimate.
- Interaction testing: The ClinicalTrials.gov record does not report an interaction statistic or p-value, so the difference between the overall and women-specific HRs should not be treated as proof of treatment-effect heterogeneity.
- Proportional-hazards assumption: If a Cox proportional-hazards model was used, the interpretation of a single HR depends on the model's assumptions, including proportionality of hazards. The ClinicalTrials.gov record does not report a proportional-hazards diagnostic.
- Censoring: The ClinicalTrials.gov record does not provide censoring counts or timing, so the censoring pattern cannot be independently evaluated from this record.
- Arm-level enrollment: Overall enrollment is 2,502, but the ClinicalTrials.gov record does not provide randomized enrollment by arm.
- Safety inference: Serious-adverse-event counts are reported by arm, but no inferential comparison is posted on ClinicalTrials.gov for this safety measure.
- Unreported design details: Non-inferiority margins, interim analyses, multiplicity procedures, missing-data methods, crossover, and Bayesian methods are not established by the ClinicalTrials.gov record.
21. Why This Trial Matters Statistically
CREST is a useful teaching case because its statistical record connects randomized treatment assignment with a clinically structured time-to-event endpoint and covariate-adjusted hazard-ratio estimation.
| Concept | How it appears in CREST |
|---|---|
| Randomization | The trial uses randomized allocation. |
| Parallel design | The registered design model is parallel. |
| Double masking | The masking field is registered as double. |
| Time-to-event endpoint | The primary endpoint is a composite time-to-event outcome through 4 years. |
| Hazard ratio | The primary effect measure is HR 1.11. |
| Confidence interval | The primary 95% CI is 0.81–1.51. |
| Covariate adjustment | The primary HR is adjusted for age, sex and symptomatic status. |
| Superiority testing | The registered hypothesis type is superiority. |
| Composite endpoint | Peri-procedural stroke, myocardial infarction, or death is combined with postprocedural ipsilateral stroke thereafter. |
| Subgroup analysis | A women-specific HR of 1.35 is reported as a secondary analysis. |
| Precision | The women-specific 95% CI is 0.82–2.23. |
| Safety reporting | Serious adverse events are reported as affected/at-risk by arm. |
The most important statistical lesson is that a clinical trial result is not just a single number. The interpretation depends on the estimand, endpoint definition, time horizon, treatment comparison, adjustment variables, uncertainty interval, analysis population, and the distinction between primary and secondary analyses.
22. Statistical Concepts in This Trial
Learn more about the methods used in this trial:
23. Related Statistical Calculators
24. Sources
- ClinicalTrials.gov: NCT00004732 — CREST.
- PubMed: PMID 20505173.
- PubMed: PMID 20876500.
- PubMed: PMID 21307169.
- PubMed: PMID 21606394.
- PubMed: PMID 22821614.
Continue through the Clinical Biostats statistical pathway
Connect this trial's randomized design, time-to-event endpoint, hazard ratio, confidence interval, and covariate adjustment to deeper statistical tutorials and calculation tools.
25. Record Summary
CREST is a randomized, parallel, double-masked trial with 2,502 participants and two intervention arms: carotid artery stenting and carotid endarterectomy. Its registered primary endpoint is a time-to-event composite covering any peri-procedural stroke, myocardial infarction, or death during 30 days and postprocedural ipsilateral stroke thereafter, up to 4 years.
The reported primary analysis gives an adjusted hazard ratio of 1.11 for CAS versus CEA, with a two-sided 95% CI of 0.81–1.51. The adjustment variables are age, sex, and symptomatic status. A secondary women-specific analysis reports an HR of 1.35, with a 95% CI of 0.82–2.23, adjusted for age and symptomatic status. Neither registry-reported analysis includes a reported p-value.
The statistical interpretation therefore centers on the hazard-ratio scale, the uncertainty represented by the confidence intervals, the composite and time-dependent structure of the primary endpoint, and the distinction between an overall randomized comparison and a subgroup estimate. The safety record separately reports serious adverse events of 322/1262 for CAS and 292/1240 for CEA.