This page separates reported trial results from statistical interpretation. Numerical trial results and trial-specific facts are taken from the ClinicalTrials.gov data posted on ClinicalTrials.gov for NCT00143507. The registry provides the formal statistical analyses for the posted outcome measures but does not provide every design or analysis detail that might appear in a full protocol or statistical analysis plan.
1. Trial at a Glance
BEAUTIFUL was a completed, randomized, parallel-group, quadruple-masked phase 3 trial evaluating ivabradine versus placebo in patients with coronary disease and left ventricular dysfunction. The registry reports 10,917 enrolled participants and a primary composite time-to-event endpoint analyzed using an adjusted Cox proportional-hazards model.
| Feature | BEAUTIFUL |
|---|---|
| Trial name | BEAUTIFUL |
| Brief title | The BEAUTIFUL Study: Effects of Ivabradine in Patients With Stable Coronary Artery Disease and Left Ventricular Systolic Dysfunction |
| Phase | Phase 3 |
| Status | Completed |
| Therapeutic area | Cardiology |
| Conditions | Coronary Disease; Ventricular Dysfunction, Left |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Quadruple |
| Primary purpose | Treatment |
| Enrollment | 10,917 |
| Arms | 2 |
| Interventions | Ivabradine; Placebo |
| Results posted | Yes |
| Outcome measures posted | 13 |
| Statistical analyses posted | 13 |
| Lead sponsor | Institut de Recherches Internationales Servier |
2. Clinical Question
The registered trial evaluated whether ivabradine, compared with placebo, affected the time to the first occurrence of a prespecified composite of cardiovascular death, hospitalization for acute myocardial infarction, or hospitalization for new onset or worsening heart failure in patients with coronary disease and left ventricular dysfunction.
Population
Patients represented by the registered conditions of coronary disease and left ventricular dysfunction. The ClinicalTrials.gov record identifies the brief title as referring specifically to stable coronary artery disease and left ventricular systolic dysfunction.
Intervention
Ivabradine.
Comparator
Placebo.
Primary question
Does ivabradine change the hazard of experiencing the first event in the registered primary composite endpoint compared with placebo?
3. Trial Design
Ivabradine
- Drug intervention.
- Compared with placebo in the randomized parallel-group design.
Placebo
- Drug-placebo comparator.
- Serves as the reference group for the reported hazard ratios.
4. Trial Timeline
Study start
The registered study began in December 2004.
Primary completion
The registered primary completion date was February 2008.
Results posted
ClinicalTrials.gov reports the trial as completed, with 13 outcome measures and 13 statistical analyses posted.
5. Endpoints
| Endpoint | Registry definition / time frame | Endpoint type |
|---|---|---|
| Primary Composite Endpoint | First event among cardiovascular death, hospitalisation for acute myocardial infarction (fatal or not), or hospitalisation for new onset or worsening heart failure (fatal or not). From the date of randomisation to the date of the first occurrence of the first event, up to 3 years. | Time-to-event |
| Cardiovascular Death | From the date of randomisation to death, up to 3 years. | Time-to-event |
| Hospitalisation for Acute Myocardial Infarction | From the date of randomisation to the date of first occurrence of the event, up to 3 years. | Time-to-event |
| Hospitalisation for New Onset or Worsening Heart Failure | From the date of randomisation to the date of first occurrence of the event, up to 3 years. | Time-to-event |
| All-cause of Mortality | From the date of randomisation to death, up to 3 years. | Time-to-event |
| Coronary Artery Disease Death | From the date of randomisation to death, up to 3 years. | Time-to-event |
| Hospitalisation for Coronary Revascularisation | From the date of randomisation to the date of first occurrence of the event, up to 3 years. | Time-to-event |
| Hospitalisation for Unstable Angina | From the date of randomisation to the date of first occurrence of the event, up to 3 years. | Time-to-event |
| Hospitalisation for Acute Coronary Syndrome (Unstable Angina or Acute Myocardial Infarction) | From the date of randomisation to the date of first occurrence of the first event, up to 3 years. | Time-to-event |
| Hospitalisation for Acute Coronary Syndrome, or Coronary Revascularisation | From the date of randomisation to the date of first occurrence of the first event, up to 3 years. | Time-to-event |
| Hospitalisation for Acute Coronary Syndrome, New Onset or Worsening Heart Failure or Coronary Revascularisation | From the date of randomisation to the date of first occurrence of the first event, up to 3 years. | Time-to-event |
| Cardiovascular Death, or Hospitalisation for New Onset or Worsening Heart Failure | From the date of randomisation to the date of first occurrence of the first event, up to 3 years. | Time-to-event |
| Cardiovascular Death, or Hospitalisation for Acute Myocardial Infarction | From the date of randomisation to the date of the first occurrence of the first event, up to 3 years. | Time-to-event |
6. Statistical Methodology
Cox proportional-hazards model
The registry identifies a Cox proportional-hazards model as the statistical method for all 13 posted analyses. For the primary endpoint, the reported estimate is based on an adjusted Cox proportional-hazards model with beta-blocker intake as a covariate.
A hazard ratio compares the modeled instantaneous event rates between treatment groups over follow-up. The reported HR is a relative time-to-event measure rather than an absolute probability of experiencing an event.
Covariate adjustment
The primary analysis adjusted the Cox model for beta-blocker intake. Adjustment incorporates that variable into the model rather than simply comparing unadjusted event times. The ClinicalTrials.gov record also identify beta-blocker intake as the factor used for stratification of the corresponding log-rank test.
Stratified log-rank testing
The primary p-value was obtained from a log-rank test stratified on beta-blocker intake factor. This means the time-to-event comparison was performed while accounting for the specified stratification factor.
Confidence intervals
The posted effect estimates are accompanied by 95% two-sided confidence intervals. For the primary endpoint, the interval is 0.91 to 1.10 around an estimated hazard ratio of 1.00.
Superiority hypothesis
The registry classifies the posted analyses as testing a superiority hypothesis. This differs conceptually from a non-inferiority analysis: there is no non-inferiority margin in the ClinicalTrials.gov record, and the interpretation of the p-value is therefore framed around evidence for a treatment difference rather than retention of a prespecified minimum effect.
7. Primary Result: Primary Composite Endpoint
The registered primary endpoint was the time from randomisation to the first occurrence of cardiovascular death, hospitalization for acute myocardial infarction, or hospitalization for new onset or worsening heart failure, with follow-up specified as up to 3 years.
Hazard ratio for the primary composite endpoint
95% CI: 0.91–1.10 · P = 0.945
Adjusted Cox proportional-hazards model with beta-blocker intake as a covariate; p-value from a log-rank test stratified on beta-blocker intake factor.
| Primary endpoint | Ivabradine vs Placebo | 95% CI | P-value |
|---|---|---|---|
| Primary Composite Endpoint | HR 1.00 | 0.91–1.10 | 0.945 |
The estimated hazard ratio of 1.00 is centered exactly at the null value of 1.00. In the context of this model, it means the estimated instantaneous rate of the first qualifying composite event was the same in the two randomized groups.
This does not mean that every participant had the same outcome, that the two groups had identical event histories at every time point, or that there was no uncertainty. The analysis is a population-level time-to-event comparison summarized through a model-based relative measure.
The 95% CI of 0.91–1.10 describes the statistical precision of the estimated hazard ratio under the analysis framework. It spans values below and above 1.00, so the data are compatible with a modestly lower or higher estimated hazard as well as the point estimate of 1.00.
The p-value of 0.945 is evidence against the null hypothesis under the specified test framework; it is not a measure of the size of the treatment effect. A p-value does not tell us that the effect is "94.5% null," nor does it quantify clinical importance.
The interpretation also depends on the Cox-model framework, including the proportional-hazards assumption and the handling of censoring. The registry reports the adjusted Cox model and stratified log-rank test but does not provide enough information in the ClinicalTrials.gov record to independently assess the proportional-hazards assumption.
8. Secondary Endpoint Results
The registry reports formal Cox-model analyses for 12 additional time-to-event outcomes. All use ivabradine versus placebo, an adjusted Cox proportional-hazards model with beta-blocker intake as a covariate, and a p-value from a log-rank test stratified on beta-blocker intake factor.
| Secondary endpoint | HR | 95% CI | P-value |
|---|---|---|---|
| Cardiovascular Death | 1.07 | 0.94–1.22 | 0.316 |
| Hospitalisation for Acute Myocardial Infarction | 0.87 | 0.72–1.06 | 0.159 |
| Hospitalisation for New Onset or Worsening Heart Failure | 0.99 | 0.86–1.13 | 0.850 |
| All-cause of Mortality | 1.04 | 0.92–1.16 | 0.547 |
| Coronary Artery Disease Death | 0.89 | 0.71–1.12 | 0.331 |
| Hospitalisation for Coronary Revascularisation | 0.83 | 0.67–1.02 | 0.078 |
| Hospitalisation for Unstable Angina | 1.08 | 0.83–1.40 | 0.583 |
| Hospitalisation for Acute Coronary Syndrome (Unstable Angina or Acute Myocardial Infarction) | 0.95 | 0.81–1.11 | 0.501 |
| Hospitalisation for Acute Coronary Syndrome, or Coronary Revascularisation | 0.90 | 0.78–1.04 | 0.141 |
| Hospitalisation for Acute Coronary Syndrome, New Onset or Worsening Heart Failure or Coronary Revascularisation | 0.96 | 0.86–1.06 | 0.411 |
| Cardiovascular Death, or Hospitalisation for New Onset or Worsening Heart Failure | 1.04 | 0.94–1.15 | 0.484 |
| Cardiovascular Death, or Hospitalisation for Acute Myocardial Infarction | 1.01 | 0.90–1.13 | 0.835 |
Reading the secondary hazard ratios
The secondary estimates range from an HR of 0.83 for hospitalization for coronary revascularisation to an HR of 1.08 for hospitalization for unstable angina. These estimates should be read together with their confidence intervals rather than in isolation.
For example, the HR of 0.83 for hospitalization for coronary revascularisation corresponds to an estimated 17% lower hazard because 1 − 0.83 = 0.17. However, its 95% CI of 0.67–1.02 extends across 1.00, so the interval includes both lower and slightly higher hazards relative to placebo.
Similarly, the HR of 1.07 for cardiovascular death represents an estimated 7% higher hazard in the ivabradine group under the model, but its 95% CI of 0.94–1.22 includes 1.00. The p-value of 0.316 is a test result, not an estimate of the magnitude or probability of a treatment effect.
Secondary endpoint interpretation table
| Pattern | What the reported estimate means | What it does not establish |
|---|---|---|
| HR below 1.00 | The modeled hazard is estimated to be lower with ivabradine than with placebo. | It does not establish a specific absolute risk reduction or guarantee a benefit for an individual participant. |
| HR above 1.00 | The modeled hazard is estimated to be higher with ivabradine than with placebo. | It does not establish that the treatment increases risk with certainty when the confidence interval includes 1.00. |
| CI crosses 1.00 | The reported uncertainty interval includes the null hazard ratio. | It does not prove that the true effect is exactly zero. |
| Small p-value | Provides stronger evidence against the specified null hypothesis under the test framework. | It does not measure effect size, clinical importance, or the probability that the treatment works. |
9. Primary Composite Endpoint: Why the Composite Matters
The primary endpoint combines three clinically distinct events: cardiovascular death, hospitalization for acute myocardial infarction, and hospitalization for new onset or worsening heart failure. The endpoint is triggered by the first occurrence of any of these events.
One statistical endpoint
The three components are treated as a single composite time-to-first-event outcome for the registered primary analysis.
Earliest event controls
Once a participant experiences a qualifying first event, that participant has reached the primary endpoint.
Different clinical events
Death, acute myocardial infarction hospitalization, and heart-failure hospitalization can have different clinical meanings and frequencies.
Component analyses matter
The registry separately reports cardiovascular death, acute myocardial infarction hospitalization, and new onset or worsening heart-failure hospitalization as secondary outcomes.
This structure is statistically important because a composite endpoint can be driven by its components in different ways. A hazard ratio for the composite should therefore not automatically be interpreted as if it described each component separately. The separate component analyses provide additional context, while still being secondary analyses in the registry structure.
10. Statistical Methods Explained
Why was a Cox proportional-hazards model used?
All 13 posted statistical analyses are identified as Cox regression analyses, and all are time-to-event outcomes. A Cox model is suited to outcomes where both whether an event occurs and when it occurs matter. It can incorporate censoring and produces a hazard ratio as a relative treatment-effect measure.
What does an HR of 1.00 mean?
An HR of 1.00 is the null value for a hazard ratio. For the primary composite endpoint, the estimated HR of 1.00 indicates no estimated difference in the modeled hazard between ivabradine and placebo at the point estimate. It is not a statement that the individual event histories were identical.
Why was beta-blocker intake included in the analysis?
The registry states that the primary hazard-ratio estimate came from an adjusted Cox proportional-hazards model with beta-blocker intake as a covariate. Adjustment allows the model to account statistically for that variable when estimating the treatment hazard ratio. The same factor was used to stratify the reported log-rank test.
What does the 95% confidence interval of 0.91–1.10 mean?
The interval expresses uncertainty around the estimated primary hazard ratio under the statistical model and sampling framework. Because it extends from below 1.00 to above 1.00, it contains the null value. The interval should not be interpreted as the range of individual patient treatment effects.
Why is the p-value not the same thing as the treatment effect?
The treatment effect is summarized by the hazard ratio, while the p-value evaluates evidence against a specified null hypothesis under the chosen test. The primary analysis has HR 1.00 and P = 0.945; those numbers answer different statistical questions and should not be substituted for one another.
Why does a time-to-event endpoint need censoring?
Not every participant necessarily experiences the endpoint during the observation period. A participant can therefore contribute information until the last time they are known to be event-free without being counted as having experienced the event. Survival-analysis methods such as Cox regression are designed to use this partial follow-up rather than simply discarding it.
What assumption is important for a Cox hazard ratio?
The standard Cox proportional-hazards interpretation assumes that the relative hazard is sufficiently represented by a common hazard ratio over the relevant follow-up. The ClinicalTrials.gov record identifies the Cox method but do not provide a formal assessment of that assumption, so this page does not claim that the assumption was empirically verified.
11. Confidence Intervals and Precision Across Endpoints
The confidence intervals provide an important complement to the point estimates. Looking across the 13 analyses, the intervals vary in width according to the statistical information available for each endpoint.
| Endpoint | HR | 95% CI width | Relationship to 1.00 |
|---|---|---|---|
| Primary Composite Endpoint | 1.00 | 0.19 | Includes 1.00 |
| Cardiovascular Death | 1.07 | 0.28 | Includes 1.00 |
| Hospitalisation for Acute Myocardial Infarction | 0.87 | 0.34 | Includes 1.00 |
| Hospitalisation for New Onset or Worsening Heart Failure | 0.99 | 0.27 | Includes 1.00 |
| All-cause of Mortality | 1.04 | 0.24 | Includes 1.00 |
| Coronary Artery Disease Death | 0.89 | 0.41 | Includes 1.00 |
| Hospitalisation for Coronary Revascularisation | 0.83 | 0.35 | Includes 1.00 |
| Hospitalisation for Unstable Angina | 1.08 | 0.57 | Includes 1.00 |
| Hospitalisation for Acute Coronary Syndrome (Unstable Angina or Acute Myocardial Infarction) | 0.95 | 0.30 | Includes 1.00 |
| Hospitalisation for Acute Coronary Syndrome, or Coronary Revascularisation | 0.90 | 0.26 | Includes 1.00 |
| Hospitalisation for Acute Coronary Syndrome, New Onset or Worsening Heart Failure or Coronary Revascularisation | 0.96 | 0.20 | Includes 1.00 |
| Cardiovascular Death, or Hospitalisation for New Onset or Worsening Heart Failure | 1.04 | 0.21 | Includes 1.00 |
| Cardiovascular Death, or Hospitalisation for Acute Myocardial Infarction | 1.01 | 0.23 | Includes 1.00 |
The confidence-interval widths in this table are simple descriptive differences between the registry-reported upper and lower bounds; they are shown only to illustrate the relative span of the reported intervals. They are not alternative statistical estimates.
12. P-values and Hypothesis Testing
The registry classifies the analyses as superiority analyses. The primary p-value was 0.945, obtained from a log-rank test stratified on beta-blocker intake factor.
The hypothesis test asks whether the observed treatment-group separation is inconsistent with the specified null model. The p-value should be interpreted alongside the hazard ratio and its confidence interval.
A high p-value does not establish that the treatment groups are biologically or clinically identical. It means that the observed data do not provide strong evidence against the specified null hypothesis under the test that was reported. Conversely, a low p-value would not by itself establish that an effect was clinically important.
13. Secondary Results: Component-by-Component Statistical Reading
Cardiovascular death
HR 1.07, 95% CI 0.94–1.22, P = 0.316. The point estimate is above 1.00, while the confidence interval includes 1.00.
Acute myocardial infarction hospitalization
HR 0.87, 95% CI 0.72–1.06, P = 0.159. The point estimate is below 1.00, with the interval extending across 1.00.
Heart-failure hospitalization
HR 0.99, 95% CI 0.86–1.13, P = 0.850. The point estimate is close to the null value of 1.00.
All-cause mortality
HR 1.04, 95% CI 0.92–1.16, P = 0.547. The interval includes both values below and above 1.00.
Coronary artery disease death
HR 0.89, 95% CI 0.71–1.12, P = 0.331.
Coronary revascularisation hospitalization
HR 0.83, 95% CI 0.67–1.02, P = 0.078. The point estimate is below 1.00, but the confidence interval includes 1.00.
Unstable angina hospitalization
HR 1.08, 95% CI 0.83–1.40, P = 0.583.
Acute coronary syndrome hospitalization
HR 0.95, 95% CI 0.81–1.11, P = 0.501.
Acute coronary syndrome or revascularisation
HR 0.90, 95% CI 0.78–1.04, P = 0.141.
Acute coronary syndrome, heart failure or revascularisation
HR 0.96, 95% CI 0.86–1.06, P = 0.411.
Cardiovascular death or heart-failure hospitalization
HR 1.04, 95% CI 0.94–1.15, P = 0.484.
Cardiovascular death or acute myocardial infarction
HR 1.01, 95% CI 0.90–1.13, P = 0.835.
14. Safety Results
The ClinicalTrials.gov record reports serious adverse events by treatment arm as the number affected divided by the number at risk.
Serious adverse events
Ivabradine vs Placebo
| Safety measure | Ivabradine | Placebo |
|---|---|---|
| Serious adverse events, affected / at risk | 1625/5477 | 1770/5430 |
The figures describe the number of participants affected and the number at risk for serious adverse events in each arm. They are not presented here as a formal hypothesis test because the ClinicalTrials.gov record does not provide a statistical comparison, confidence interval, or p-value for this safety measure.
15. Statistical Interpretation of the Primary Result
The primary HR of 1.00 is the null point estimate. Unlike an HR such as 0.83 or 1.07, it does not imply either a relative reduction or increase in the modeled hazard at the point estimate.
The 95% CI of 0.91–1.10 indicates that the estimate is not known with unlimited precision. The interval permits a modest reduction or increase in hazard relative to placebo within the statistical uncertainty represented by the interval.
The p-value of 0.945 is the reported result of the stratified log-rank test. It does not measure the magnitude of the HR and should not be interpreted as a probability that the null hypothesis is true.
The ClinicalTrials.gov record does not provide Kaplan-Meier event probabilities, median time-to-event estimates, event counts for the primary endpoint, or absolute risk differences. Those quantities are therefore not added to this analysis.
16. Analysis Population and What the Registry Reports
The ClinicalTrials.gov record identifies the enrollment as 10,917 and provides formal statistical analyses for all 13 posted outcome measures. It does not provide, in the ClinicalTrials.gov record, a detailed definition of the analysis population for each endpoint.
| Available information | Registry-supported description |
|---|---|
| Enrollment | 10,917 |
| Arms | 2 |
| Allocation | Randomized |
| Primary analysis method | Adjusted Cox proportional-hazards model with beta-blocker intake as a covariate |
| Primary test | Log-rank test stratified on beta-blocker intake factor |
| Effect measure | Hazard ratio |
| Hypothesis type | Superiority |
| Analysis population details | Not specified in the ClinicalTrials.gov record |
This distinction matters because a statistical result cannot be fully interpreted without knowing exactly which participants contributed to the analysis and how censoring and treatment deviations were handled. Since those details are not in the ClinicalTrials.gov record, they are not inferred from the enrollment figure.
17. Multiplicity and Multiple Secondary Endpoints
BEAUTIFUL has one registered primary endpoint and multiple secondary time-to-event analyses. The ClinicalTrials.gov record does not state a formal multiplicity-adjustment strategy for the secondary endpoints.
| Analysis family | Registry-supported interpretation |
|---|---|
| Primary Composite Endpoint | Registered primary endpoint; superiority analysis. |
| 12 secondary endpoints | Formal statistical analyses posted; all analyzed using Cox regression with the reported stratified log-rank p-value framework. |
| Multiplicity adjustment | Not specified in the ClinicalTrials.gov record. |
18. Stratification and Covariate Adjustment
The primary analysis contains two related but distinct statistical ideas involving beta-blocker intake:
Covariate adjustment
Beta-blocker intake was included as a covariate in the adjusted Cox proportional-hazards model used to estimate the hazard ratio.
Stratified testing
The p-value was obtained using a log-rank test stratified on the beta-blocker intake factor.
These are related approaches but should not be described as the same operation. A covariate-adjusted Cox model estimates treatment effects while including the factor in the regression model; a stratified log-rank test accounts for the factor through stratification of the comparison.
19. Kaplan-Meier Estimation and Time-to-Event Analysis
The registry identifies the endpoints as time-to-event outcomes and reports Cox proportional-hazards models. Kaplan-Meier estimation is a standard descriptive framework for these types of outcomes because it can represent the probability of remaining event-free over time while accommodating right-censored observations.
where di is the number of events at time ti and ni is the number at risk immediately before that time.
However, the ClinicalTrials.gov record does not report Kaplan-Meier event probabilities or median event times. Therefore, this page explains the method as the natural descriptive framework for a time-to-event endpoint but does not manufacture a survival curve or numerical Kaplan-Meier estimates.
20. Hazard Ratio: A Worked Interpretation
The primary HR is 1.00, while several secondary estimates are below or above 1.00. The meaning of an HR can be illustrated directly from the reported secondary results.
For hospitalization for coronary revascularisation, the reported HR of 0.83 corresponds to an estimated 17% lower hazard under the fitted model. This is a relative hazard interpretation, not a 17-percentage-point reduction in the probability of hospitalization.
The confidence interval of 0.67–1.02 is essential to that interpretation. It crosses 1.00, so the point estimate should not be read as though 0.83 were known without uncertainty.
Likewise, the primary HR of 1.00 should not be interpreted as proof that the treatment has no possible effect. The confidence interval provides the appropriate context for the uncertainty around the point estimate.
21. What the Primary Result Does — and Does Not — Mean
It means
The adjusted Cox model produced a primary hazard-ratio estimate of 1.00 for ivabradine versus placebo.
It does not mean
Every participant had the same outcome trajectory or identical event-free follow-up.
It means
The 95% confidence interval was 0.91–1.10 under the reported analysis framework.
It does not mean
The true treatment effect must lie inside the interval for every future population or individual patient.
The most direct statistical description of the primary result is that the reported adjusted hazard ratio was 1.00, with a two-sided 95% CI of 0.91–1.10 and a stratified log-rank p-value of 0.945. Clinical interpretation requires keeping those three quantities together and recognizing the limits of the information available in the registry record.
22. Important Limitations and Interpretation Issues
- Registry-level detail: the ClinicalTrials.gov record does not provide the complete protocol or statistical analysis plan, so some design details cannot be reconstructed.
- Analysis population: the ClinicalTrials.gov record does not specify the exact analysis population used for each posted endpoint.
- Absolute event measures: the ClinicalTrials.gov record does not report primary-endpoint event counts, Kaplan-Meier probabilities, median event times, or absolute risk differences.
- Proportional hazards: the Cox model relies on its usual proportional-hazards framework, but the ClinicalTrials.gov record does not report a formal assessment of that assumption.
- Multiplicity: 12 secondary analyses accompany the primary endpoint, but the ClinicalTrials.gov record does not specify a multiplicity-adjustment strategy.
- Component interpretation: the primary composite combines cardiovascular death, acute myocardial infarction hospitalization, and new onset or worsening heart-failure hospitalization. A composite HR should not automatically be treated as the HR for every component.
- Safety interpretation: serious adverse-event counts are reported by arm, but the ClinicalTrials.gov record does not provide a formal statistical comparison for this safety measure.
- Missing analysis details: the ClinicalTrials.gov record does not specify imputation procedures, crossover rules, interim-analysis methods, non-inferiority margins, factorial structure, or Bayesian methods.
23. Why This Trial Matters Statistically
BEAUTIFUL is a useful teaching case because the registry record connects a randomized parallel-group design with a composite time-to-event endpoint, covariate-adjusted Cox regression, stratified log-rank testing, hazard ratios, confidence intervals, and multiple secondary outcomes.
| Concept | How it appears in BEAUTIFUL |
|---|---|
| Randomization | The registered allocation is randomized. |
| Parallel-group design | The design model is parallel with two intervention arms. |
| Blinding | The registry classifies masking as quadruple. |
| Composite endpoint | The primary endpoint combines three first-event components. |
| Time-to-event analysis | The primary and posted secondary outcomes are time-to-event endpoints. |
| Hazard ratio | All posted statistical analyses use hazard ratios as the effect measure. |
| Cox regression | The registry reports Cox proportional-hazards models for all 13 analyses. |
| Covariate adjustment | Beta-blocker intake was included as a covariate in the adjusted Cox model. |
| Stratified analysis | The p-value was based on a log-rank test stratified on beta-blocker intake factor. |
| Confidence intervals | Each reported hazard ratio has a two-sided 95% confidence interval. |
| Superiority testing | The registry classifies the hypothesis type as superiority. |
| Multiple endpoints | One primary endpoint and 12 secondary statistical analyses are posted. |
24. A Statistical Reading of the Full Results Set
Viewed as a complete statistical record, the results show a primary HR of 1.00 and secondary HRs spanning values below and above 1.00. The primary confidence interval is relatively close to the null value, extending from 0.91 to 1.10. The secondary confidence intervals likewise include 1.00.
The important statistical lesson is that a collection of hazard ratios should be interpreted as a set of estimates with uncertainty rather than as a list of isolated p-values. The direction of a point estimate, the width and location of its confidence interval, the endpoint definition, and the analysis hierarchy all contribute to interpretation.
| Question | Reported BEAUTIFUL result | Statistical meaning |
|---|---|---|
| What was the primary point estimate? | HR 1.00 | The fitted model estimated equal hazard at the point estimate. |
| How precise was the primary estimate? | 95% CI 0.91–1.10 | The interval spans modestly lower and higher hazards. |
| What was the primary p-value? | 0.945 | Evidence against the specified null was weak under the reported stratified log-rank test. |
| Were secondary estimates identical? | No | They ranged from 0.83 to 1.08 among the analyses posted on ClinicalTrials.gov. |
| Did the secondary intervals all include 1.00? | Yes | Each registry-reported secondary 95% CI spans the null hazard ratio. |
| Was there formal multiplicity information? | Not reported | The secondary p-values should not be assigned an unreported multiplicity strategy. |
25. Related Tutorials
Learn more about the methods used in this trial:
26. Related Calculators
27. Sources
- ClinicalTrials.gov: BEAUTIFUL — NCT00143507. Trial design, endpoints, posted statistical analyses, effect estimates, confidence intervals, p-values, and safety data summarized on this page are based on the ClinicalTrials.gov record.
- Linked publication: PubMed — PMID 18757088.
Continue through the Clinical Biostats statistical pathway
Connect this trial's time-to-event endpoints to deeper tutorials and statistical calculation tools.
28. Record Summary
BEAUTIFUL provides a clear example of a randomized phase 3 time-to-event analysis in which a composite primary endpoint was evaluated using an adjusted Cox proportional-hazards model and a stratified log-rank test. The primary analysis reported an HR of 1.00, a two-sided 95% CI of 0.91–1.10, and a p-value of 0.945. Twelve additional secondary time-to-event analyses were also posted, each using the same general Cox-model framework with beta-blocker intake as a covariate and stratification factor for the log-rank test.
The most useful statistical reading is therefore not simply to look at the p-value. The primary HR establishes the direction and magnitude of the point estimate, the confidence interval describes its uncertainty, and the stratified log-rank p-value addresses evidence against the specified null hypothesis. The composite endpoint and its component outcomes should also be distinguished, while the multiple secondary analyses should be interpreted in light of the fact that the ClinicalTrials.gov record does not specify a multiplicity-adjustment strategy.