This page separates reported trial results from statistical interpretation. Numerical results are restricted to information contained in the ClinicalTrials.gov record. The registry reports one formal statistical analysis for the primary time-to-event endpoint.
1. Trial at a Glance
SAMMPRIS was a randomized, parallel, open-label phase 3 trial in ischemic stroke evaluating intracranial angioplasty and stenting versus intensive medical management. The registry reports enrollment of 451 participants and a primary time-to-event analysis based on the time to a composite stroke-or-death/revascularization-related outcome or territory-specific ischemic stroke.
| Feature | SAMMPRIS |
|---|---|
| Trial name | SAMMPRIS |
| NCT identifier | NCT00576693 |
| Therapeutic area | Neurology |
| Condition | Ischemic Stroke |
| Phase | Phase 3 |
| Status | Completed |
| Enrollment | 451 |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None |
| Primary purpose | Treatment |
| Lead sponsor | Medical University of South Carolina |
| Interventions | Intracranial angioplasty and stenting (device); intensive medical management (other) |
2. Clinical Question
The statistical question is whether the two randomized treatment strategies differ in time to the registered primary outcome. The registry identifies the hypothesis type as superiority and reports a log-rank comparison of the two treatment groups.
Population
Participants enrolled in the SAMMPRIS phase 3 trial with the registered condition of ischemic stroke.
Intervention
Intracranial angioplasty and stenting.
Comparator
Intensive medical management alone.
Primary question
Does time to the registered primary outcome differ between intensive medical management plus stenting and intensive medical management alone?
3. Trial Design
Intracranial angioplasty and stenting
- Intracranial angioplasty and stenting
- Intervention classified in the registry as a device
- Compared with intensive medical management alone
Medical management
- Intensive medical management
- Intervention classified in the registry as other
- Compared with intensive medical management plus stenting
Randomization creates the principal design basis for comparing the two treatment strategies.
The registry identifies the study as a parallel-design trial.
The registry identifies the study as unmasked.
The registry classifies the primary purpose as treatment.
4. Endpoints
The registry lists one primary endpoint. It is a composite time-to-event endpoint, and the registered time frame states that the mean length of follow-up was 2.4 years.
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Primary endpoint | Any Stroke or Death Within 30 Days of Enrollment or Any Revascularization Procedure OR an Ischemic Stroke in the Territory of the Symptomatic Intracranial Artery Beyond 30 Days After Enrollment. | Mean length of follow-up was 2.4 years |
Understanding the composite endpoint
The endpoint combines several clinically distinct events into one time-to-first-event outcome: stroke or death within 30 days of enrollment, stroke or death within 30 days of a qualifying revascularization procedure performed during follow-up, or an ischemic stroke in the territory of the symptomatic intracranial artery beyond 30 days after enrollment.
The registry definition specifies that the stroke component includes ischemic stroke, parenchymal brain hemorrhage, subarachnoid hemorrhage, or intraventricular hemorrhage. The primary endpoint therefore does not represent one single type of event; it is a prespecified composite that treats qualifying events as occurrences of the same primary endpoint.
5. Analysis Population
The registry states explicitly that all patients enrolled in the study were included in the primary outcome analysis. This is important because the formal comparison was therefore not restricted to a subset defined by treatment received or by completion of follow-up.
| Population | Role in the posted analysis |
|---|---|
| All patients enrolled | Included in the primary outcome analysis |
| Primary comparison | Intensive Medical Management Plus Stenting vs Intensive Medical Management Alone |
| Analysis endpoint | Time to the registered primary outcome |
6. Primary Results
ClinicalTrials.gov reports one formal statistical analysis for the primary endpoint. The analysis compared the two treatment groups with respect to time to the primary outcome using a log-rank test.
Primary time-to-event comparison
Log-rank test · Superiority hypothesis
Analysis population: all patients enrolled in the study
| Primary endpoint | Analysis method | Hypothesis | P-value |
|---|---|---|---|
| Any Stroke or Death Within 30 Days of Enrollment or Any Revascularization Procedure OR an Ischemic Stroke in the Territory of the Symptomatic Intracranial Artery Beyond 30 Days After Enrollment. | Log-rank test | Superiority | 0.0252 |
The reported P = 0.0252 is the result of the prespecified-type log-rank comparison recorded in the registry. Under the superiority hypothesis, the p-value quantifies how unusual the observed separation in event-time experience would be under the null hypothesis represented by the statistical test.
The p-value does not measure the size of the treatment effect. A p-value of 0.0252 cannot be translated into a percentage reduction in hazard, a percentage reduction in stroke, or a particular clinical benefit.
The registry does not provide a confidence interval for an effect estimate, so the precision of the treatment effect cannot be quantified from the posted statistical-analysis record. In particular, the absence of a reported hazard ratio means that the direction and magnitude of the treatment difference should not be inferred from the p-value alone.
For a time-to-event endpoint, interpretation also depends on how patients are censored and on the pattern of event occurrence over follow-up. The registry analysis identifies the log-rank test but does not provide enough additional information to evaluate proportional hazards or reconstruct the underlying survival curves.
7. Statistical Methodology
Log-rank test
The registry identifies the log-rank test as the statistical method used for the primary analysis. The log-rank test is designed to compare the distribution of time to an event between groups while incorporating information from participants who are followed for different lengths of time.
The log-rank procedure compares the observed pattern of events with the pattern expected under the null hypothesis across the observed follow-up times.
Why a time-to-event method fits the endpoint
The SAMMPRIS primary endpoint is defined not simply as whether a participant experienced an event, but as a composite event occurring within specified time windows and during subsequent follow-up. A time-to-event framework can account for both whether an event occurs and when it occurs.
This distinction matters when follow-up is not identical for every participant. A participant who remains event-free through their available observation period still contributes information up to the point at which follow-up ends. That observation is different from treating the participant as if they had been followed indefinitely without an event.
Kaplan-Meier estimation
Kaplan-Meier estimation is the standard descriptive companion to a log-rank comparison of time-to-event outcomes. It estimates the probability of remaining event-free over time while accounting for right-censored observations.
Here, di represents events at an event time and ni represents participants at risk immediately before that time.
The ClinicalTrials.gov record does not identify Kaplan-Meier estimation as a posted statistical method, nor do they provide the event and censoring data needed to reconstruct a Kaplan-Meier curve. It is therefore appropriate to explain the method as the usual descriptive framework for this endpoint without presenting a reconstructed curve as if it were reported by the registry.
What the log-rank test does not provide by itself
A log-rank p-value establishes a statistical comparison but does not itself provide a readily interpretable effect size. For clinical interpretation, an effect estimate such as a hazard ratio, together with a confidence interval and absolute event-time summaries, is generally useful. Those additional quantities are not present in the ClinicalTrials.gov record.
8. Primary Endpoint: How to Read the Result
The p-value
0.0252 is the reported probability measure from the log-rank comparison. It addresses compatibility with the null hypothesis under the statistical framework; it is not a measure of treatment magnitude.
The missing effect estimate
No hazard ratio or other effect estimate is reported in the registry-reported analysis. The p-value alone cannot establish how large the difference between groups was.
The missing confidence interval
No confidence interval for an effect estimate is reported. Consequently, precision around the magnitude of the treatment difference cannot be quantified from this record.
The analysis population
All patients enrolled in the study were included in the primary outcome analysis, according to the registry's analysis notes.
9. Safety Results
The ClinicalTrials.gov record reports serious adverse events by treatment arm as affected participants divided by participants at risk. These figures should be kept separate from the primary efficacy endpoint because they describe a safety outcome rather than time to the registered primary event.
| Safety measure | Intensive Medical Management Plus Stenting | Intensive Medical Management Alone |
|---|---|---|
| Serious adverse events | 146/224 | 121/227 |
The registry reports serious adverse events as affected / at risk for each arm. The denominators are therefore important: they indicate the number of participants represented in the reported safety measure, while the numerators identify participants affected by a serious adverse event.
These data should not be converted into a causal safety conclusion by simply comparing the two fractions without knowing the complete safety-analysis definitions, observation periods, event ascertainment rules, and whether participants could contribute more than one serious adverse event. The ClinicalTrials.gov record does not provide those additional details.
10. Trial Timeline
Trial start
The SAMMPRIS trial began in October 2008.
Enrollment
The completed ClinicalTrials.gov record reports enrollment of 451 participants across 2 treatment arms.
Mean follow-up
The registered primary endpoint has a time frame of mean length of follow-up of 2.4 years.
Primary completion
The trial's primary completion date was April 2013.
11. Statistical Methods Explained
Why was a log-rank test used?
The primary endpoint is explicitly classified in the ClinicalTrials.gov record as time-to-event. The log-rank test is designed for comparing event-time distributions between randomized groups, making it directly aligned with the structure of the registered endpoint.
What does a log-rank p-value of 0.0252 mean?
It is the probability measure reported by the log-rank comparison under the statistical testing framework. A smaller p-value indicates greater incompatibility with the null hypothesis of no difference in event-time distributions. The p-value does not tell us how large the treatment effect is.
Why can't the p-value be converted into a hazard ratio?
A p-value and a hazard ratio answer different statistical questions. The p-value summarizes evidence against a null hypothesis, whereas the hazard ratio is an effect-size measure comparing estimated instantaneous event rates. Many different effect sizes can be associated with statistically significant or nonsignificant p-values depending on sample size, event frequency, censoring, and other features of the data.
What does randomization contribute?
Randomization assigns participants to treatment groups using a mechanism intended to balance prognostic factors in expectation. It provides the fundamental design basis for a causal comparison between the randomized strategies. The ClinicalTrials.gov record identifies SAMMPRIS as randomized but do not provide the randomization ratio or stratification factors.
Why is the endpoint called time-to-event?
Because the endpoint concerns events occurring within specified time windows after enrollment or a qualifying revascularization procedure and an ischemic stroke occurring beyond 30 days after enrollment. The timing of an event therefore forms part of the endpoint rather than merely its presence or absence.
Why is a composite endpoint statistically important?
A composite combines several event types into a single primary outcome. This can allow multiple clinically relevant outcomes to contribute to one time-to-event analysis. Interpretation requires attention to which component events actually contribute to the composite, because the statistical result describes the combined endpoint rather than necessarily having the same meaning for every component.
Why does follow-up matter in this analysis?
The registry reports a mean follow-up of 2.4 years. In a time-to-event analysis, follow-up determines how much opportunity participants have to experience the endpoint and how much information is available about event timing. It also determines the amount of censoring and therefore affects the precision and interpretation of the survival analysis.
12. Statistical Interpretation of the Primary Analysis
The SAMMPRIS registry record provides a formal log-rank P = 0.0252 for the primary endpoint. This supplies evidence from the specified statistical test, but it does not supply an effect-size estimate.
A complete clinical interpretation normally benefits from knowing both the statistical evidence and the magnitude of the observed difference. Here, the latter cannot be reconstructed from the registry analysis without introducing external numerical results.
Confidence intervals communicate uncertainty around an estimated treatment effect. Because the registry-reported statistical analysis does not report a hazard ratio or confidence interval, the precision of the effect magnitude cannot be assessed from the posted analysis alone.
The p-value by itself does not identify which treatment group had the longer event-free time. The comparison is between intensive medical management plus stenting and intensive medical management alone, but the registry-reported analysis does not provide an effect estimate or survival summary from which the direction can be established numerically.
Time-to-event methods generally allow participants who have not experienced the endpoint by the end of their available observation to contribute information up to censoring. The ClinicalTrials.gov record identifies the endpoint as time-to-event but does not provide the detailed censoring rules or censoring counts needed for a deeper assessment.
13. What This Analysis Can and Cannot Establish
| Question | What the ClinicalTrials.gov record establishes |
|---|---|
| Was the trial randomized? | Yes. Allocation is recorded as randomized. |
| What was compared? | Intensive Medical Management Plus Stenting versus Intensive Medical Management Alone. |
| What was the primary endpoint? | The registered composite stroke/death/revascularization-related or territory-specific ischemic stroke endpoint. |
| What type of endpoint was it? | Time-to-event. |
| What statistical test was used? | Log-rank test. |
| What was the reported p-value? | 0.0252. |
| What was the analysis population? | All patients enrolled in the study. |
| What was the hazard ratio? | Not reported in the ClinicalTrials.gov record. |
| What was the confidence interval? | Not reported in the ClinicalTrials.gov record. |
| What was the median event-free time? | Not reported in the ClinicalTrials.gov record. |
14. Limitations and Interpretation Issues
- Effect estimate not reported: the registry analysis gives a log-rank p-value but no hazard ratio or other quantitative treatment-effect estimate.
- Confidence interval not reported: there is no confidence interval for the treatment effect in the ClinicalTrials.gov record, limiting direct assessment of precision.
- Composite endpoint: the primary outcome combines several event definitions, so the overall statistical result should not automatically be interpreted as evidence about each individual component.
- Mean follow-up is not a median event time: the registered time frame reports a mean length of follow-up of 2.4 years. This should not be interpreted as the median time to the primary endpoint.
- No masking: the registry identifies the study as unmasked. The ClinicalTrials.gov record does not provide additional information about how lack of masking affected individual endpoint ascertainment.
- No reconstruction of survival curves: the ClinicalTrials.gov record does not contain the participant-level event and censoring information needed to recreate Kaplan-Meier estimates.
- Limited safety detail: serious adverse events are reported as affected/at risk by arm, but the ClinicalTrials.gov record does not provide additional safety definitions or statistical comparisons.
- Multiplicity information is absent: the ClinicalTrials.gov record does not report an alpha-spending strategy, multiplicity adjustment, or a hierarchy of multiple statistical tests. No such methods should be inferred.
- Stratification information is absent: the ClinicalTrials.gov record does not identify stratification factors for the primary analysis.
- Missing-data and imputation methods are absent: the ClinicalTrials.gov record does not specify an imputation strategy for missing observations, and none should be assumed.
15. Why This Trial Matters Statistically
SAMMPRIS is a useful teaching case because the registry data bring together several core principles of randomized clinical-trial statistics without requiring a complicated modeling framework to understand the primary comparison.
| Concept | How it appears in SAMMPRIS |
|---|---|
| Randomization | The trial uses randomized allocation to compare two treatment strategies. |
| Parallel design | The registry identifies the design model as parallel. |
| Time-to-event endpoint | The primary outcome is analyzed according to time until a qualifying event. |
| Composite endpoint | The primary endpoint combines several stroke, death, and revascularization-related event definitions. |
| Log-rank test | The registry reports the log-rank test for the primary comparison. |
| P-value | The primary analysis reports P = 0.0252. |
| Analysis population | All patients enrolled in the study were included in the primary outcome analysis. |
| Follow-up | The registered primary-endpoint time frame reports a mean length of follow-up of 2.4 years. |
| Safety analysis | Serious adverse events are reported as affected/at risk for each treatment arm. |
16. A Deeper Look at Time-to-Event Thinking
Why event timing contains information
Suppose two randomized groups had exactly the same number of primary events but the events occurred at different times. A simple binary analysis at a single fixed time could obscure that difference. A time-to-event analysis instead uses the observed timing of events throughout follow-up.
This is particularly relevant to SAMMPRIS because the primary endpoint is explicitly defined through several time windows: events within 30 days of enrollment, events within 30 days of qualifying revascularization procedures during follow-up, and territory-specific ischemic stroke beyond 30 days after enrollment.
Why censoring cannot simply be treated as no event
A participant who leaves observation without experiencing the endpoint is not equivalent to a participant who is known to remain event-free for the entire study period. Time-to-event methods preserve the information available up to the censoring point rather than assigning an unobserved future outcome.
Why a survival curve would add information
A Kaplan-Meier curve would provide a visual description of how the estimated event-free probability changes over time in each randomized group. It would complement the log-rank p-value by showing the temporal pattern of separation. However, the ClinicalTrials.gov record does not contain the underlying event/censoring information needed to reproduce such a curve accurately.
17. The Primary Endpoint as a Statistical Object
Event definition
The endpoint is not simply "stroke." It contains specified stroke, death, and revascularization-related timing rules plus territory-specific ischemic stroke beyond 30 days.
Time origin
The registry definition explicitly uses enrollment as a time origin for several components and incorporates qualifying revascularization procedures into another component.
Statistical comparison
The two randomized treatment groups were compared according to time to the primary outcome using the log-rank test.
Hypothesis
The registry classifies the hypothesis as superiority, meaning the statistical objective was to evaluate whether the treatment groups differed rather than to establish equivalence or non-inferiority.
18. Statistical Concepts in This Trial
Learn more about the methods used in this trial:
19. Related Statistical Calculators
20. Sources
- ClinicalTrials.gov: NCT00576693 — SAMMPRIS.
- PubMed: PMID 22977278.
- PubMed: PMID 22991350.
- PubMed: PMID 22984008.
- PubMed: PMID 21899409.
- PubMed: PMID 24168957.
Continue through the Clinical Biostats statistical pathway
Explore the statistical concepts behind randomized trials, time-to-event endpoints, log-rank testing, Kaplan-Meier estimation, and clinical-trial analysis.
21. Record Summary
SAMMPRIS provides a focused example of randomized time-to-event analysis. The phase 3 trial enrolled 451 participants and compared intensive medical management plus stenting with intensive medical management alone in a randomized parallel design. The registered primary endpoint is a composite time-to-event outcome incorporating stroke, death, revascularization-related events, and territory-specific ischemic stroke. The registry reports a log-rank test with P = 0.0252 under a superiority hypothesis, with all enrolled patients included in the primary outcome analysis and a mean follow-up of 2.4 years.
The most important statistical limitation is equally clear: the registry analysis does not report a hazard ratio, confidence interval, median event time, or treatment-specific primary endpoint event counts. The p-value therefore provides evidence from the formal comparison but cannot, by itself, establish the magnitude or numerical direction of the treatment effect. Keeping that distinction explicit is essential when interpreting clinical-trial statistics.