← Clinical Trials
Pulmonary Arterial Hypertension Phase 3 Time-to-Event Analysis NCT00660179

SERAPHIN: Complete Statistical Analysis of Macitentan in Pulmonary Arterial Hypertension

An independent statistical review of the randomized phase 3 SERAPHIN trial evaluating macitentan in patients with symptomatic pulmonary arterial hypertension, with emphasis on its morbidity-or-mortality time-to-event endpoint, hazard ratios, confidence intervals, and log-rank analyses.

SERAPHIN  ·  Randomized  ·  Quadruple-masked  ·  Completed
Scope of this record

This page separates reported trial results from statistical interpretation. The numerical results on this page are restricted to the ClinicalTrials.gov trial data posted on ClinicalTrials.gov for SERAPHIN. Where the ClinicalTrials.gov record does not provide a result or design detail, it is not inferred.

Registry note: This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. View the SERAPHIN registry record.

Trial at a Glance

SERAPHIN was a randomized, parallel-group, quadruple-masked phase 3 trial evaluating macitentan in patients with symptomatic pulmonary arterial hypertension. The registry reports a total enrollment of 742 participants across three arms, with a primary time-to-event endpoint assessing the time to first confirmed morbidity or mortality event up to the end of treatment.

742
Enrollment
All randomized patients in the reported analyses
3
Arms
Two macitentan doses and placebo
0.547
Primary HR
10 mg vs placebo
<0.0001
P-value
10 mg vs placebo
FeatureSERAPHIN
PhasePhase 3
ConditionPulmonary Arterial Hypertension
PopulationPatients with symptomatic pulmonary arterial hypertension
DesignRandomized, parallel-group, quadruple-masked, treatment-purpose study
AllocationRandomized
ArmsMacitentan 3 mg, macitentan 10 mg, and placebo
Enrollment742
Primary endpointTime to first confirmed morbidity or mortality event up to the end of treatment
Primary endpoint typeTime-to-event
Primary analysis methodLog-rank test
Primary effect measureHazard ratio
Hypothesis typeSuperiority
Study start2008-05
Primary completion2012-03
Lead sponsorActelion
Sponsor typeIndustry

Clinical Question

The central statistical question was whether macitentan reduced the time to a first confirmed morbidity or mortality event compared with placebo in patients with symptomatic pulmonary arterial hypertension.

Population

Patients with symptomatic pulmonary arterial hypertension enrolled in a phase 3 randomized study.

Intervention

Macitentan at the reported study dose levels of 3 mg and 10 mg.

Comparator

Placebo.

Primary question

Does macitentan change the time to a first confirmed morbidity or mortality event relative to placebo?

Trial Design

01
Randomize 742 participants
02
Three arms 3 mg, 10 mg, placebo
03
Follow-up Time-to-event assessment
04
Event assessment Morbidity or mortality
05
Analysis Log-rank and hazard ratio
MACITENTAN · 3 MG

ACT-064992 3 mg

  • Macitentan (ACT-064992)
  • Randomized treatment arm
  • Primary comparison reported against placebo
  • Secondary time-to-event comparisons also reported
MACITENTAN · 10 MG

ACT-064992 10 mg

  • Macitentan (ACT-064992)
  • Randomized treatment arm
  • Primary comparison reported against placebo
  • Secondary time-to-event comparisons also reported
CONTROL

Placebo

  • Placebo drug
  • Randomized control arm
  • Reference group for the reported hazard ratios
DESIGN CHARACTERISTICS

Parallel, quadruple-masked study

  • Randomized allocation
  • Parallel design model
  • Quadruple masking
  • Primary purpose: treatment

The ClinicalTrials.gov record does not report a crossover procedure, factorial structure, non-inferiority margin, Bayesian analysis, or a formal missing-data/imputation method. Those features are therefore not incorporated into the statistical interpretation of this page.

Endpoints

Endpoint roleRegistered endpointTime frameType
Primary Time to First Confirmed Morbidity or Mortality Event up to the End of Treatment (Kaplan-Meier Estimate of Patients Without a Morbidity or Mortality Event) Up to end of treatment (data presented up to month 36) Time-to-event
Secondary Time to Death Due to PAH or Hospitalisation for PAH up to the End of Treatment (Kaplan-Meier Estimate of Patients Without an Event) Up to end of treatment (data presented up to month 36) Time-to-event
Secondary Time to Death Due to Any Cause up to the End of Treatment (Kaplan-Meier Estimate of Patients Without an Event) Up to end of treatment (data presented up to month 36) Time-to-event
Secondary Time to Death Due to Any Cause up to the End of Study (Kaplan-Meier Estimate of Patients Without an Event) Up to end of study (data presented up to month 36) Time-to-event

What counted as a morbidity or mortality event?

The registry definition states that morbidity or mortality events included death, atrial septostomy, lung transplantation, initiation of intravenous or subcutaneous prostanoids, or other worsening of pulmonary arterial hypertension.

The registry further defines other worsening of pulmonary arterial hypertension using a combined set of clinical criteria, beginning with at least a 15% decrease in the 6 minute walk distance from baseline. The registry-reported trial extract ends during that definition, so this page does not reconstruct additional criteria that are not present in the ClinicalTrials.gov record.

Endpoint perspective: the primary outcome is a composite time-to-first-event endpoint. That means the analysis asks when the first qualifying event occurs, rather than separately treating every component as an independent primary outcome.

Statistical Methodology

Kaplan-Meier estimation

The registry describes the primary endpoint as a Kaplan-Meier estimate of patients without a morbidity or mortality event. Kaplan-Meier estimation is appropriate for time-to-event outcomes because participants may have different lengths of observed follow-up and may be censored before experiencing the event.

Kaplan-Meier concept
S(t) = probability of remaining event-free through time t

The Kaplan-Meier curve estimates the probability of remaining free of the event over time while accounting for right-censored observations.

Log-rank test

The registry reports the log-rank test for the primary and secondary time-to-event comparisons. The log-rank test evaluates whether the observed event-time distributions differ between randomized groups over follow-up.

Hazard ratio

The reported effect measure is the hazard ratio. A hazard ratio compares the estimated instantaneous event rates between treatment groups within the time-to-event analysis.

Interpretation of the hazard ratio
HR < 1  →  lower estimated instantaneous event rate in the macitentan group than in the placebo group

The hazard ratio is a relative time-to-event measure. It is not the same as an absolute reduction in event probability, a difference in Kaplan-Meier percentages, or a statement that a particular proportion of participants benefited.

Analysis population

The statistical analyses posted on ClinicalTrials.gov specify all randomized patients as the analysis population for the reported primary and secondary efficacy comparisons. This is important because treatment assignment, rather than subsequent treatment behavior, anchors the randomized comparison.

Superiority framework

The registered hypothesis type for the reported analyses is superiority. The reported comparisons therefore assess whether the time-to-event outcome differs between macitentan and placebo, rather than whether macitentan remains within a prespecified non-inferiority margin.

Results: Primary Endpoint

The registry reports two formal primary-endpoint analyses, both comparing a macitentan dose with placebo. Both use the log-rank test, report hazard ratios, use all randomized patients, and specify superiority as the hypothesis type.

Macitentan 3 mg vs placebo

Time to first confirmed morbidity or mortality event

HR 0.704

97.5% CI: 0.516–0.960   ·   P = 0.0108

Log-rank analysis; all randomized patients; data presented up to month 36.

Clinical Biostats interpretation

The hazard ratio of 0.704 means that the estimated instantaneous rate of experiencing the first confirmed morbidity or mortality event was lower in the ACT-064992 3 mg group than in the placebo group within the reported time-to-event analysis. Expressed as a simple relative interpretation, the estimate corresponds to an approximately 29.6% lower estimated hazard relative to placebo.

The estimate does not mean that 29.6% of participants avoided an event, that each participant had exactly the same reduction in risk, or that the absolute event probability was reduced by that amount.

The 97.5% confidence interval of 0.516–0.960 describes statistical uncertainty around the estimated hazard ratio. Because the interval remains below the null value for a hazard ratio, the reported interval is consistent with a lower estimated event hazard for ACT-064992 3 mg than placebo under the stated analysis.

The P-value of 0.0108 addresses the evidence against the null hypothesis in the reported statistical test. It does not measure the magnitude or clinical importance of the treatment effect. Interpretation also depends on the time-to-event analysis, censoring, and the assumptions underlying a hazard-ratio summary.

Macitentan 10 mg vs placebo

Time to first confirmed morbidity or mortality event

HR 0.547

97.5% CI: 0.392–0.762   ·   P < 0.0001

Log-rank analysis; all randomized patients; data presented up to month 36.

Clinical Biostats interpretation

The hazard ratio of 0.547 indicates a lower estimated instantaneous rate of the first confirmed morbidity or mortality event in the ACT-064992 10 mg group relative to placebo under the reported analysis. As a direct mathematical interpretation of the reported estimate, this corresponds to an approximately 45.3% lower estimated hazard.

The hazard ratio does not mean that 45.3% of participants were protected from an event, nor does it describe an absolute difference in the probability of an event by a particular time.

The 97.5% confidence interval of 0.392–0.762 indicates the statistical precision of the reported hazard-ratio estimate. The interval lies below the null value, so the reported confidence interval is consistent with a lower estimated event hazard for ACT-064992 10 mg than placebo.

The P-value of <0.0001 describes the strength of evidence against the null hypothesis under the reported test. It should not be interpreted as a probability that the treatment works, a measure of effect size, or a measure of clinical importance.

Primary comparisonAnalysis populationMethodHazard ratio97.5% CIP-value
Placebo vs ACT-064992 3 mg All randomized patients Log-rank 0.704 0.516–0.960 0.0108
Placebo vs ACT-064992 10 mg All randomized patients Log-rank 0.547 0.392–0.762 <0.0001
Educational note: the ClinicalTrials.gov record provides hazard ratios, confidence intervals, and P-values but do not provide the underlying individual participant event and censoring data. A Kaplan-Meier curve is therefore not reconstructed here.

Results: Secondary Endpoints

Time to death due to PAH or hospitalisation for PAH

ComparisonHazard ratio97.5% CIP-valueMethod
Placebo vs ACT-064992 3 mg 0.669 0.462–0.970 0.0146 Log-rank
Placebo vs ACT-064992 10 mg 0.500 0.335–0.747 <0.0001 Log-rank

These analyses use all randomized patients and assess time to death due to pulmonary arterial hypertension or hospitalisation for pulmonary arterial hypertension up to the end of treatment, with data presented up to month 36.

The ACT-064992 3 mg hazard ratio of 0.669 corresponds to an approximately 33.1% lower estimated hazard relative to placebo. The ACT-064992 10 mg hazard ratio of 0.500 corresponds to an approximately 50.0% lower estimated hazard relative to placebo. These are relative hazard interpretations, not absolute reductions in the probability of death or hospitalisation.

Time to death due to any cause up to the end of treatment

ComparisonHazard ratio97.5% CIP-valueMethod
Placebo vs ACT-064992 3 mg 0.971 0.477–1.976 0.9249 Log-rank
Placebo vs ACT-064992 10 mg 0.638 0.287–1.418 0.2037 Log-rank

The reported hazard ratio for ACT-064992 3 mg versus placebo is 0.971, with a 97.5% confidence interval of 0.477–1.976 and P = 0.9249. The corresponding ACT-064992 10 mg comparison has a hazard ratio of 0.638, a 97.5% confidence interval of 0.287–1.418, and P = 0.2037.

For both comparisons, the confidence intervals include the null value. The statistical interpretation is therefore different from the primary composite endpoint: the ClinicalTrials.gov record does not provide evidence from these reported tests that the time to death from any cause up to the end of treatment differed between either macitentan dose and placebo.

Time to death due to any cause up to the end of study

ComparisonHazard ratio97.5% CIP-valueMethod
Placebo vs ACT-064992 3 mg 1.046 0.653–1.673 0.8312 Log-rank
Placebo vs ACT-064992 10 mg 0.771 0.464–1.282 0.2509 Log-rank

The reported ACT-064992 3 mg hazard ratio is 1.046, with a 97.5% confidence interval of 0.653–1.673 and P = 0.8312. The ACT-064992 10 mg hazard ratio is 0.771, with a 97.5% confidence interval of 0.464–1.282 and P = 0.2509.

Both confidence intervals include the null value. These estimates therefore do not establish a statistically detectable difference in time to death due to any cause up to the end of study under the reported analyses.

Important endpoint distinction: the primary endpoint was a composite morbidity-or-mortality outcome, whereas these secondary analyses focus on death due to any cause. Different endpoints can produce different hazard-ratio estimates because they count different events and therefore answer different clinical questions.

Safety Results

The ClinicalTrials.gov record reports serious adverse events by randomized arm as the number of affected participants divided by the number at risk.

ArmSerious adverse events affected / at risk
Placebo137 / 249
ACT-064992 3 mg130 / 250
ACT-064992 10 mg109 / 242

These counts should be kept separate from the efficacy analyses. The primary efficacy analysis was based on all randomized patients, whereas serious adverse events are presented here as affected participants divided by participants at risk in each arm.

The ClinicalTrials.gov record does not provide a formal statistical comparison, confidence interval, or P-value for serious adverse events. Accordingly, the affected/at-risk counts are descriptive and are not converted into an inferential treatment comparison on this page.

Statistical Methods Explained

Why was a time-to-event endpoint appropriate for SERAPHIN?

The primary outcome is defined by the time until the first confirmed morbidity or mortality event. Time-to-event analysis preserves information about when events occur and can accommodate participants who have not yet experienced an event by the end of their observed follow-up.

What does the hazard ratio represent in this trial?

A hazard ratio summarizes the relative instantaneous event rate between a macitentan group and placebo during the analyzed follow-up. A value below the null value indicates a lower estimated event hazard in the macitentan group under the model used to summarize the time-to-event data.

Why is the hazard ratio not the same as a risk reduction?

The hazard ratio is a relative rate measure over time. It is not an absolute probability difference. For example, the primary hazard ratio of 0.547 should not be read as saying that the absolute probability of a morbidity or mortality event was reduced by 45.3%. The correct interpretation is that the estimated instantaneous event rate was lower by that relative amount under the reported hazard-ratio framework.

What does the confidence interval add?

The confidence interval communicates uncertainty around the estimated hazard ratio. A narrow interval indicates greater statistical precision than a wide interval, while the location of the interval relative to the null value helps describe the evidence from the corresponding statistical test.

Why use a log-rank test?

The log-rank test is designed for comparing time-to-event distributions between groups. It uses information from event times throughout follow-up rather than reducing the outcome to whether each participant eventually experienced an event.

What does the P-value tell us?

The P-value measures the degree of compatibility between the observed data and the null hypothesis under the specified statistical test. It does not measure effect size, clinical importance, or the probability that the treatment hypothesis is true.

Why does the analysis population matter?

The registry specifies all randomized patients for the reported primary and secondary analyses. Maintaining the randomized treatment assignment in the efficacy analysis helps preserve the balance created by randomization and avoids redefining the comparison according to treatment received after randomization.

Understanding the Primary Endpoint More Deeply

The primary endpoint combines several clinically important events into a single time-to-first-event outcome. This has an important statistical consequence: the hazard ratio describes the first qualifying event, regardless of which component of the composite occurs first.

Composite endpoint

The event definition includes death and several forms of clinical worsening, including atrial septostomy, lung transplantation, initiation of specified prostanoids, and other worsening of pulmonary arterial hypertension.

Time-to-first-event

Only the first qualifying morbidity or mortality event contributes as the primary event for an individual participant.

Kaplan-Meier perspective

The analysis estimates the probability of remaining free from the composite event as follow-up progresses.

Hazard-ratio perspective

The hazard ratio summarizes the relative instantaneous event rate between randomized treatment groups.

Interpreting the Pattern of Reported Results

The analyses posted on ClinicalTrials.gov show a consistent pattern for the primary composite endpoint: both macitentan comparisons have hazard ratios below the null value, with the reported confidence intervals remaining below the null value and the reported P-values indicating evidence against the corresponding null hypotheses.

The secondary endpoint results illustrate why a trial should not be summarized by one statistic alone. For time to death due to pulmonary arterial hypertension or hospitalisation for pulmonary arterial hypertension, the reported hazard ratios are also below the null value. For time to death due to any cause, however, the reported confidence intervals include the null value for both dose comparisons in the end-of-treatment analysis and in the end-of-study analysis.

Statistical interpretation

The primary endpoint and the secondary endpoints are related but not interchangeable. A treatment effect on a composite morbidity-or-mortality endpoint does not automatically imply that every individual component, including death from any cause, will show the same estimated effect.

Why absolute measures still matter

The ClinicalTrials.gov record reports hazard ratios and confidence intervals rather than the underlying Kaplan-Meier event-free percentages at specific time points. Consequently, this page does not manufacture absolute risk differences from the reported hazard ratios.

Why statistical significance is not effect size

The P-values for the primary comparisons provide evidence from the specified statistical tests, but they do not quantify the size of the treatment effect. The hazard ratio and its confidence interval are the appropriate reported quantities for describing the relative magnitude and statistical precision of the primary treatment comparisons.

Assumptions and Cautions in Time-to-Event Analysis

Hazard ratios are useful summaries of time-to-event treatment effects, but their interpretation depends on the structure of the survival data and the model used to obtain the estimate. A hazard ratio is not a direct statement about cumulative risk at a particular time point.

Design Features That Affect Statistical Interpretation

Randomization
The trial used randomized allocation, providing the basis for comparing outcomes between the macitentan and placebo groups.
Masking
The registry describes the study as quadruple-masked, reducing opportunities for knowledge of treatment assignment to influence trial conduct or outcome assessment.
Parallel design
The study used a parallel design, so participants were assigned to a randomized treatment arm rather than sequentially receiving study treatments as part of a crossover design.
Superiority
The registry identifies the hypothesis type as superiority for the reported primary and secondary statistical analyses.

The ClinicalTrials.gov record does not report stratification factors, interim-analysis boundaries, a non-inferiority margin, crossover rules, a factorial structure, Bayesian methods, or a prespecified missing-data method. Those topics are therefore not presented as trial features.

Limitations

Why This Trial Matters Statistically

SERAPHIN is a useful statistical teaching case because the trial combines randomized treatment allocation with a clinically meaningful composite time-to-event endpoint and several layers of survival analysis. The registry results allow the treatment comparison to be examined through the hazard ratio, confidence interval, log-rank test, and analysis population rather than through a single P-value.

Statistical conceptHow it appears in SERAPHIN
RandomizationRandomized parallel-group phase 3 design
MaskingQuadruple-masked study
Kaplan-Meier estimationPrimary and secondary endpoints reported as Kaplan-Meier time-to-event measures
Hazard ratioPrimary and secondary treatment effects reported as hazard ratios
Confidence interval97.5% confidence intervals reported for the formal statistical analyses
Log-rank testReported method for the primary and secondary time-to-event comparisons
Composite endpointPrimary endpoint combines morbidity and mortality events
Superiority testingReported hypothesis type for the statistical analyses
Analysis populationAll randomized patients for the reported efficacy analyses
Safety analysisSerious adverse events reported descriptively by randomized arm

Clinical Biostats Interpretation of the Evidence

What the primary analysis shows

The reported primary hazard ratios are below the null value for both macitentan dose comparisons, and both corresponding 97.5% confidence intervals remain below the null value.

What the secondary analyses show

The reported hazard ratios for death due to pulmonary arterial hypertension or hospitalisation for pulmonary arterial hypertension are below the null value for both dose comparisons. The reported all-cause mortality analyses have wider uncertainty and confidence intervals that include the null value.

What cannot be concluded from the ClinicalTrials.gov record

The registry extract does not support conclusions about subgroup heterogeneity, median survival, absolute event probabilities, or a formal multiplicity strategy.

Why the confidence intervals matter

The intervals show that statistical precision differs across endpoints. The all-cause mortality estimates have substantially wider intervals than the primary composite estimates.

Related Tutorials

Learn more about the methods used in this trial:

Related Calculators

Sources

Continue through the Clinical Biostats statistical pathway

Connect the trial's time-to-event endpoints to deeper statistical tutorials and practical analysis tools.

Record Summary

SERAPHIN provides a clear example of how a randomized clinical trial can evaluate a composite morbidity-or-mortality endpoint using time-to-event methods. The registry analyses use all randomized patients, Kaplan-Meier estimation, log-rank testing, hazard ratios, and 97.5% confidence intervals. The primary comparisons report hazard ratios of 0.704 for ACT-064992 3 mg versus placebo and 0.547 for ACT-064992 10 mg versus placebo, while the reported all-cause mortality analyses show wider uncertainty. The statistical story is therefore best understood through the combination of the endpoint definition, analysis population, hazard-ratio estimates, confidence intervals, P-values, and the distinction between the composite primary endpoint and the secondary mortality endpoints.

Clinical Biostats methodology: A trial-results page should not merely repeat the registry output. The goal is to explain how the endpoint was constructed, how the reported statistical method answers the clinical question, what the effect estimates mean, and where the registry-reported evidence does not support additional conclusions.