This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the ClinicalTrials.gov record.
1. Trial at a Glance
TRITON was a completed phase 3 randomized, parallel-group trial evaluating initial triple versus initial dual oral combination therapy in patients with newly diagnosed pulmonary arterial hypertension. The registry reports 247 participants and two treatment arms, with quadruple masking and a primary endpoint based on change in pulmonary vascular resistance from baseline to Week 26.
| Feature | TRITON |
|---|---|
| Trial name | TRITON |
| Phase | Phase 3 |
| Condition | Pulmonary Arterial Hypertension |
| Population description | Patients with newly diagnosed pulmonary arterial hypertension |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Quadruple |
| Primary purpose | Treatment |
| Enrollment | 247.0 |
| Interventions | Macitentan, tadalafil, selexipag |
| Results posted | Yes |
| Outcome measures posted | 10 |
| Statistical analyses posted | 9 |
| ClinicalTrials.gov | NCT02558231 |
2. Clinical Question
The central statistical question was whether initial triple oral combination therapy with macitentan, tadalafil, and selexipag differed from initial dual oral combination therapy with macitentan, tadalafil, and placebo in the registered primary measure of pulmonary vascular resistance at Week 26.
Population
Patients with newly diagnosed pulmonary arterial hypertension.
Intervention
Initial triple oral therapy with macitentan, tadalafil, and selexipag.
Comparator
Initial double oral therapy with macitentan, tadalafil, and placebo.
Primary question
What is the difference between the randomized groups in change from baseline to Week 26 in pulmonary vascular resistance?
3. Trial Design
Initial triple combination
- Macitentan
- Tadalafil
- Selexipag
Initial dual combination
- Macitentan
- Tadalafil
- Placebo
4. Endpoints
The registry lists one primary endpoint and reports statistical analyses for that endpoint plus eight secondary outcomes. The primary endpoint is a continuous ratio measure based on pulmonary vascular resistance.
| Endpoint | Time frame | Type | Analysis reported |
|---|---|---|---|
| Change From Baseline to Week 26 in Pulmonary Vascular Resistance (PVR) | Baseline, Week 26 | Continuous; ratio | ANCOVA |
| Change From Baseline to Week 26 in 6-minute Walk Distance (6MWD) | Baseline, Week 26 | Continuous; meter | ANCOVA |
| Change From Baseline to Week 26 in N-terminal Pro B-type Natriuretic Peptide (NT-proBNP) Levels | Baseline, Week 26 | Continuous; ratio | ANCOVA |
| Change From Baseline to Week 26 in Mean Pulmonary Arterial Pressure (mPAP) | Baseline, Week 26 | Continuous; mmHg | ANCOVA |
| Change From Baseline to Week 26 in Mean Right Atrial Pressure (mRAP) | Baseline, Week 26 | Continuous; mmHg | ANCOVA |
| Change From Baseline to Week 26 in Total Pulmonary Resistance | Baseline, Week 26 | Continuous; dynes*second per centimeter5 | ANCOVA |
| Change From Baseline to Week 26 in Cardiac Index | Baseline, Week 26 | Continuous; liters per minute per meter square | ANCOVA |
| Change From Baseline to Week 26 in Venous Oxygen Saturation (%) | Baseline, Week 26 | Binary in registry analysis record; percentage of oxygen saturation | ANCOVA |
| Number of Participants With Disease Progression Event | Week 26, Month 12, Month 18, Month 24, Month 30, and End of Analysis Period (up to 40 months) | Time-to-event | Log-rank test |
Primary endpoint definition
The registered primary endpoint was Change From Baseline to Week 26 in Pulmonary Vascular Resistance (PVR). The registry defines this as the ratio of the Week 26 PVR value to the baseline PVR value, using re-calculated PVR. PVR was determined by right heart catheterization.
The registry states that a geometric least square mean ratio below 1 corresponds to a reduction in PVR from baseline. Missing values were imputed using a last observation carried approach, as recorded in the registry endpoint definition.
5. Statistical Analysis Populations
The reported analyses identify the full analysis set (FAS) as the efficacy analysis population. The registry explicitly defines the FAS as including all randomized participants. For several outcomes, the registry notes that the reported N represents the number of participants analyzed for that outcome measure.
| Population | Registry description | Statistical role |
|---|---|---|
| Full analysis set (FAS) | Included all randomized participants. | Primary and reported secondary efficacy analyses |
| Intention-to-treat principle | The analysis records identify intention-to-treat analysis as an additional concept. | Preserves randomized treatment assignment as the basis of the comparison |
The ClinicalTrials.gov record does not provide separate arm-specific enrollment counts, baseline characteristics, or detailed numbers analyzed for every endpoint.
6. Primary Result: Pulmonary Vascular Resistance
The primary endpoint compared the Week 26-to-baseline PVR ratio between the triple-therapy and dual-therapy groups using ANCOVA in the full analysis set.
Ratio of geometric least square means
95% CI: 0.86–1.07 · P = 0.4239
Superiority hypothesis; two-sided 95% confidence interval.
| Primary endpoint | Triple vs dual therapy | 95% CI | P-value | Method |
|---|---|---|---|---|
| Change from baseline to Week 26 in PVR | Ratio of geometric LS means = 0.96 | 0.86–1.07 | 0.4239 | ANCOVA |
The reported ratio of 0.96 means that the estimated geometric least square mean Week 26-to-baseline PVR ratio in the triple-therapy group was 0.96 times that in the dual-therapy group under the reported ANCOVA framework. Because the ratio is below 1, its point estimate is directionally consistent with a lower PVR ratio for triple therapy.
That does not mean that PVR was reduced by exactly 4% in every participant, nor does it mean that every patient experienced a reduction. The estimate is a group-level model-based comparison of the ratio endpoint.
The 95% confidence interval of 0.86–1.07 describes uncertainty around the estimated treatment-group ratio under the statistical model and sampling framework. It spans 1, so the registry-reported confidence interval is compatible with both a lower and a higher ratio for triple therapy relative to dual therapy.
The P-value of 0.4239 is a measure of compatibility with the null hypothesis under the specified testing framework; it is not a measure of the magnitude or clinical importance of the observed estimate. A P-value should therefore be read together with the estimate and confidence interval rather than as a substitute for them.
The analysis used ANCOVA and the full analysis set. Interpretation also depends on the registry's handling of missing values, which used a last observation carried approach. Because the endpoint is based on Week 26 hemodynamic assessment, the result should not be interpreted as evidence about unmeasured later hemodynamic timepoints.
7. Secondary Results: Functional and Biomarker Outcomes
6-minute Walk Distance
LS mean difference
95% CI: -19.393 to 16.538 · P = 0.8758
The estimated LS mean difference was -1.43 meters, comparing triple oral therapy with dual oral therapy in change from baseline to Week 26. The negative point estimate indicates a lower estimated change in the triple-therapy group under the reported ANCOVA model.
The confidence interval, -19.393 to 16.538, is substantially wider than the point estimate and crosses zero. This indicates considerable uncertainty around the estimated between-group difference.
The P-value of 0.8758 does not quantify how large the difference is. It addresses the statistical evidence against the specified null hypothesis. It should not be interpreted as a probability that the treatment effect is zero.
NT-proBNP Levels
Ratio of geometric LS means
95% CI: 0.770–1.371 · P = 0.8529
The estimated ratio of geometric LS means was 1.03. For this ratio endpoint, a value of 1 represents equality of the estimated geometric LS means between the randomized groups. The point estimate is close to 1, while the 95% CI of 0.770–1.371 includes 1.
The interval shows that the reported estimate is not highly precise enough to isolate a narrow range of plausible between-group ratios. The P-value of 0.8529 is evidence about the statistical test, not a measure of the size or clinical importance of the observed 1.03 ratio.
8. Secondary Results: Hemodynamic Measures
| Endpoint | Estimate | 95% CI | P-value | Effect measure |
|---|---|---|---|---|
| Change from baseline to Week 26 in mean pulmonary arterial pressure (mPAP) | -0.72 mmHg | -2.834 to 1.386 | 0.4998 | LS mean difference |
| Change from baseline to Week 26 in mean right atrial pressure (mRAP) | -0.09 mmHg | -1.003 to 0.830 | 0.8528 | LS mean difference |
| Change from baseline to Week 26 in total pulmonary resistance | 2.40 dynes*second per centimeter5 | -69.368 to 74.178 | 0.9474 | LS mean difference |
| Change from baseline to Week 26 in cardiac index | 0.13 liters per minute per meter square | -0.066 to 0.328 | 0.1902 | LS mean difference |
| Change from baseline to Week 26 in venous oxygen saturation (%) | -1.20 percentage points | -2.737 to 0.327 | 0.1227 | LS mean difference |
All five outcomes in this group were analyzed using ANCOVA in the full analysis set. The registry describes the reported hypothesis type as superiority and the confidence intervals as two-sided 95% intervals.
Mean pulmonary arterial pressure
The estimated LS mean difference was -0.72 mmHg. The negative point estimate is directionally lower for the triple-therapy group, but the 95% CI of -2.834 to 1.386 crosses zero. The P-value of 0.4998 addresses the statistical test and should not be read as an effect-size measure.
Mean right atrial pressure
The estimated LS mean difference was -0.09 mmHg, with a 95% CI of -1.003 to 0.830. The estimate is close to zero and the interval includes zero. The P-value of 0.8528 likewise should not be interpreted as the probability that the treatment effect is absent.
Total pulmonary resistance
The reported LS mean difference was 2.40 dynes*second per centimeter5. Its 95% CI extends from -69.368 to 74.178, demonstrating substantial uncertainty around the point estimate. The P-value was 0.9474. This result should be understood as the reported model-based comparison, not as a statement that the two treatments are biologically identical.
Cardiac index
The estimated LS mean difference was 0.13 liters per minute per meter square, with a 95% CI of -0.066 to 0.328. Because the interval includes zero, the direction and magnitude of the between-group difference remain uncertain under this analysis. The P-value of 0.1902 is evidence for the hypothesis test rather than a direct measure of effect size.
Venous oxygen saturation
The reported LS mean difference was -1.20 percentage points, with a 95% CI of -2.737 to 0.327. The interval includes zero. The P-value of 0.1227 should be interpreted in conjunction with the estimate and confidence interval rather than as a measure of treatment benefit or harm.
9. Secondary Result: Disease Progression
Disease progression was evaluated as a time-to-event endpoint. The registry specifies assessments at Week 26, Month 12, Month 18, Month 24, Month 30, and the end of the analysis period, up to 40 months.
Hazard ratio for disease progression
95% CI: 0.32–1.09 · P = 0.0867
Two-sided 95% confidence interval; log-rank test; superiority hypothesis.
| Endpoint | Triple therapy | Dual therapy | Effect estimate | Method |
|---|---|---|---|---|
| Number of participants with disease progression event | Macitentan + tadalafil + selexipag | Macitentan + tadalafil + placebo | HR 0.59 (95% CI 0.32–1.09) | Log-rank test |
The reported hazard ratio of 0.59 means that the estimated instantaneous rate of disease-progression events in the triple-therapy group was approximately 59% of that in the dual-therapy group under the reported time-to-event analysis. Expressed as a simple relative interpretation of the point estimate, this corresponds to an estimated 41% lower hazard of disease progression.
That does not mean that 41% fewer participants necessarily experienced progression, nor does it mean that every individual had a 41% reduction in risk. A hazard ratio is a model-based relative measure of event rates over follow-up.
The 95% CI of 0.32–1.09 crosses 1. The interval therefore includes values corresponding to both a lower and a higher hazard in the triple-therapy group. The interval is also relatively wide, indicating uncertainty around the point estimate.
The P-value of 0.0867 is associated with the reported two-sided log-rank test. It does not measure the size of the hazard ratio and should not be converted into a probability that the treatment is effective or ineffective.
The registry specifically cautions that analyses of disease progression and mortality were exploratory because of the testing hierarchy and the post hoc nature of those analyses. It also states that the study was not powered to show significant differences for the disease-progression-related endpoint. These design considerations are important when interpreting the HR and P-value.
10. Statistical Methodology
ANCOVA for the Week 26 continuous outcomes
Seven of the reported analyses used analysis of covariance (ANCOVA): the primary PVR endpoint and the secondary analyses of 6MWD, NT-proBNP, mPAP, mRAP, total pulmonary resistance, cardiac index, and venous oxygen saturation.
The exact covariate specification beyond the information contained in the registry record is not provided in the ClinicalTrials.gov record. The important statistical feature is that the reported treatment comparison was generated with ANCOVA rather than by simply comparing raw Week 26 means.
For a baseline-to-Week-26 study, ANCOVA can account for baseline measurements while estimating the treatment-group difference in the outcome at the follow-up timepoint. This can improve statistical efficiency when baseline values are informative about follow-up values.
Geometric least square mean ratios
The primary PVR endpoint and NT-proBNP analysis were reported using ratios of geometric least square means. Ratio measures are particularly useful when the scientific question is naturally expressed on a multiplicative scale rather than an additive scale.
A ratio of 1 represents equality. A ratio below 1 indicates a lower estimated value in the numerator group, while a ratio above 1 indicates a higher estimated value.
For PVR, the registry additionally defines a Week 26-to-baseline ratio within each treatment group. That means two ratio concepts are involved in the primary endpoint: the within-participant-time ratio defining the outcome and the between-treatment ratio of geometric least square means used as the reported effect measure.
Log-rank test for disease progression
Disease progression was analyzed using a log-rank test, a standard method for comparing time-to-event distributions between randomized groups. Rather than reducing each participant's experience to a simple yes/no outcome at one fixed time, the analysis uses the timing of events and the information contributed by participants who have not yet experienced the event during observed follow-up.
The ClinicalTrials.gov record reports a hazard ratio alongside the log-rank analysis. The hazard ratio provides a relative measure of event rate between the treatment groups.
Intention-to-treat analysis
The statistical-analysis records identify intention-to-treat analysis as a concept and define the full analysis set as including all randomized participants. The basic principle is that randomized assignment remains the basis for the comparison rather than selectively analyzing only participants who completed treatment.
Missing-data handling
For the registered PVR endpoint, the registry states that missing values were imputed using a last observation carried approach. This is consequential because imputation replaces unavailable measurements with assumptions based on previously observed information. The resulting treatment comparison therefore depends not only on the observed Week 26 measurements but also on the prespecified approach to missing values.
11. Statistical Methods Explained
Why was ANCOVA used for the Week 26 outcomes?
ANCOVA provides a framework for comparing randomized groups while incorporating baseline information. For an endpoint measured at baseline and again at Week 26, this can be more informative than simply comparing the two groups' raw Week 26 averages. The TRITON registry reports ANCOVA for the primary PVR analysis and the listed secondary continuous outcomes.
What does a PVR ratio below 1 mean?
The registry explicitly defines a Week 26-to-baseline PVR ratio below 1 as corresponding to a reduction in PVR from baseline. This is a within-group ratio definition. The reported treatment effect is then a ratio of geometric least square means comparing those outcome ratios between the triple- and dual-therapy groups.
What does the primary ratio of 0.96 mean?
The reported 0.96 is the ratio of geometric least square means for the primary PVR endpoint. It indicates that the estimated geometric LS mean ratio was 0.96 in triple therapy relative to dual therapy. It does not mean that every participant had a 4% reduction, and it is not an absolute risk difference.
Why is the confidence interval important?
The primary 95% CI is 0.86–1.07. A confidence interval communicates uncertainty around the estimated effect. Because the interval includes 1, it does not isolate the direction of the treatment ratio at the conventional ratio null value. Confidence intervals also provide information that a P-value alone cannot: they show the range of effect estimates compatible with the statistical framework.
What does the disease-progression hazard ratio of 0.59 mean?
A hazard ratio of 0.59 means that the estimated instantaneous disease-progression rate in the triple-therapy group was 0.59 times the corresponding rate in the dual-therapy group under the reported analysis. It is not the same as saying that 59% of participants avoided progression or that the probability of progression was reduced by 41%.
Why does a P-value not measure effect size?
A P-value quantifies statistical evidence against a specified null hypothesis under the analysis assumptions. It depends on both the estimated effect and the amount of information in the data. The effect estimate and confidence interval are therefore needed to understand the magnitude and precision of a treatment comparison.
12. Confidence Intervals and the Scale of the Effect
TRITON provides an instructive mixture of additive and multiplicative effect measures. The primary PVR analysis and NT-proBNP analysis use ratios, while 6MWD and the hemodynamic measures use LS mean differences. Disease progression uses a hazard ratio.
| Scale | TRITON examples | Null value | Interpretation of direction |
|---|---|---|---|
| Ratio of geometric means | PVR, NT-proBNP | 1 | Below 1 indicates a lower estimated ratio in triple therapy; above 1 indicates a higher ratio. |
| Mean difference | 6MWD, mPAP, mRAP, total pulmonary resistance, cardiac index, venous oxygen saturation | 0 | Negative values indicate a lower estimated outcome in triple therapy; positive values indicate a higher estimated outcome. |
| Hazard ratio | Disease progression | 1 | Below 1 indicates a lower estimated instantaneous event rate in triple therapy. |
The null value changes with the effect-measure scale. This is why a confidence interval of 0.86–1.07 for a ratio is assessed relative to 1, while a confidence interval of -19.393 to 16.538 for a mean difference is assessed relative to 0.
13. Interpreting the Pattern of Reported Results
The ClinicalTrials.gov record shows several point estimates in different directions, but their interpretation must remain tied to the scale and uncertainty of each endpoint.
| Endpoint | Point estimate | Null value | Does the 95% CI include the null? |
|---|---|---|---|
| PVR ratio | 0.96 | 1 | Yes |
| 6MWD difference | -1.43 | 0 | Yes |
| NT-proBNP ratio | 1.03 | 1 | Yes |
| mPAP difference | -0.72 | 0 | Yes |
| mRAP difference | -0.09 | 0 | Yes |
| Total pulmonary resistance difference | 2.40 | 0 | Yes |
| Cardiac index difference | 0.13 | 0 | Yes |
| Venous oxygen saturation difference | -1.20 | 0 | Yes |
| Disease progression HR | 0.59 | 1 | Yes |
This table is a statistical summary rather than a ranking of endpoints. Each outcome measures a different aspect of pulmonary arterial hypertension, uses a different effect scale in some cases, and is subject to its own uncertainty. A common mistake is to treat all estimates as though they were measurements of the same underlying quantity.
14. Multiplicity and Testing Hierarchy
The ClinicalTrials.gov record identifies the hypothesis type for the analyses as superiority. They also state an important limitation: analyses of disease progression and mortality were exploratory because of the testing hierarchy and the post hoc nature of the analyses.
Primary endpoint
PVR change from baseline to Week 26 was the registered primary endpoint and had a formal ANCOVA analysis reported with an estimate, 95% CI, and P-value.
Secondary outcomes
The registry reports eight secondary statistical analyses, including functional, biomarker, hemodynamic, and disease-progression outcomes.
Exploratory progression analysis
The registry states that disease-progression and mortality analyses were exploratory because of the testing hierarchy and post hoc nature of those analyses.
Interpretation consequence
A nominal P-value for an exploratory analysis should not automatically be interpreted as though it were an independently powered confirmatory test with the same error-control status as the primary endpoint.
The ClinicalTrials.gov record does not specify an alpha-spending procedure, multiplicity-adjustment algorithm, or formal familywise error allocation. Those methods are therefore not inferred here.
15. Safety Results
The registry reports serious adverse events by randomized arm. The registry-reported safety field gives the number affected and the number at risk for each arm.
| Safety measure | Triple Oral Therapy | Double Oral Therapy |
|---|---|---|
| Serious adverse events | 58 / 119 | 48 / 127 |
The ClinicalTrials.gov record does not provide additional serious-adverse-event categories, grades, individual event terms, or a formal between-group safety hypothesis test. Accordingly, the numbers above are reported as affected participants over participants at risk rather than converted into an additional comparative statistic.
The reported serious-adverse-event counts provide an absolute description of the number of participants affected relative to the reported number at risk. They should not be interpreted as an efficacy result, and the ClinicalTrials.gov record does not provide enough information to establish a formal statistical comparison of serious adverse-event rates.
16. Disease Progression as a Time-to-Event Endpoint
Disease progression differs fundamentally from the Week 26 continuous outcomes. For PVR, 6MWD, NT-proBNP, and the hemodynamic measures, the analysis focuses on a prespecified follow-up measurement. Disease progression instead records whether and when an event occurs during follow-up.
Participants who have not experienced the event by their last available assessment can contribute information through that time. The log-rank test then compares the observed timing of events between randomized groups.
The TRITON registry analysis reports a hazard ratio of 0.59 with a 95% CI of 0.32–1.09. Because the confidence interval crosses 1, the point estimate should not be treated as a precise estimate of a single fixed treatment effect.
17. What the Primary Ratio Does — and Does Not — Mean
The primary estimate of 0.96 is a ratio of geometric least square means for the Week 26-to-baseline PVR endpoint. It is a group-level model-based effect measure. It does not describe the experience of an individual patient.
The 95% CI of 0.86–1.07 communicates uncertainty around the primary estimate. Because it includes 1, the interval does not establish a direction of the treatment ratio relative to the null value of equality.
The P-value of 0.4239 is a result of the hypothesis test. It does not tell the reader that the treatment effect is "42.39%" or that there is a 42.39% probability of a particular clinical outcome.
The primary endpoint uses a ratio, while several secondary endpoints use mean differences and disease progression uses a hazard ratio. These measures have different null values and should not be numerically compared as though they were on the same scale.
18. Statistical Interpretation vs Clinical Interpretation
Statistical interpretation
The primary PVR analysis reported a ratio of geometric LS means of 0.96 with a 95% CI of 0.86–1.07 and P = 0.4239. The secondary analyses reported mean differences or ratios for Week 26 outcomes, and the disease-progression analysis reported HR 0.59 with a 95% CI of 0.32–1.09.
Clinical interpretation
The registry results cover pulmonary vascular resistance, walking distance, NT-proBNP, hemodynamic measures, cardiac index, venous oxygen saturation, and disease progression. These endpoints describe different dimensions of the disease and should therefore be interpreted separately rather than collapsed into a single outcome.
19. Limitations and Interpretation Issues
- Exploratory disease-progression and mortality analyses: the registry states that these analyses were exploratory because of the testing hierarchy and the post hoc nature of the analyses.
- Power for disease progression: the registry states that the study was not powered to show significant differences for the disease-progression-related endpoint.
- Limited hemodynamic timing: hemodynamic assessments were available only for the Week 26 timepoint according to the registry limitation.
- Missing-data handling: the registered PVR endpoint states that missing values were imputed using a last observation carried approach. Such imputation can influence the resulting estimate.
- Multiple endpoints: one primary endpoint and eight reported secondary statistical analyses create a broader interpretive framework than the primary test alone.
- Post hoc analyses: an analysis described by the registry as post hoc should not automatically be treated as having the same confirmatory status as the registered primary endpoint.
- Hazard-ratio interpretation: the disease-progression HR is a relative time-to-event measure, not an absolute risk difference or a statement about the percentage of participants who benefit.
- Incomplete supporting detail in the ClinicalTrials.gov record: arm-specific randomized sample sizes, baseline characteristics, and detailed subgroup analyses are not provided in the ClinicalTrials.gov record and are not reconstructed here.
20. Why This Trial Matters Statistically
TRITON is a useful statistical teaching case because it combines randomized treatment allocation, a ratio-based primary endpoint, ANCOVA, geometric least square means, missing-data imputation, multiple secondary outcomes, and a time-to-event disease-progression analysis.
| Concept | How it appears in TRITON |
|---|---|
| Randomization | Randomized parallel-group phase 3 design |
| Masking | Quadruple masking |
| Intention-to-treat principle | The FAS included all randomized participants and the analysis records identify intention-to-treat analysis as a concept |
| ANCOVA | Used for the primary PVR analysis and the reported continuous secondary outcomes |
| Geometric mean ratio | Used as the reported effect measure for PVR and NT-proBNP |
| Confidence intervals | 95% two-sided intervals reported for the primary and secondary estimates |
| P-values | Reported for the nine statistical analyses |
| Log-rank test | Used for the disease-progression time-to-event analysis |
| Hazard ratio | Reported for disease progression as HR 0.59 |
| Missing-data handling | Last observation carried approach specified for the primary PVR endpoint |
| Multiplicity | Testing hierarchy affects interpretation of disease-progression and mortality analyses |
| Exploratory analysis | Disease-progression and mortality analyses described as exploratory in the registry limitations |
21. Trial Timeline
Trial start
The registry lists May 1, 2016 as the study start date.
Primary completion
The registry lists August 29, 2019 as the primary completion date.
Registry status
The trial is listed as completed, with results posted on ClinicalTrials.gov.
22. A Compact Statistical Summary
| Role | Endpoint | Estimate | 95% CI | P-value | Method |
|---|---|---|---|---|---|
| Primary | PVR, baseline to Week 26 | 0.96 | 0.86–1.07 | 0.4239 | ANCOVA |
| Secondary | 6MWD, baseline to Week 26 | -1.43 | -19.393 to 16.538 | 0.8758 | ANCOVA |
| Secondary | NT-proBNP, baseline to Week 26 | 1.03 | 0.770–1.371 | 0.8529 | ANCOVA |
| Secondary | mPAP, baseline to Week 26 | -0.72 | -2.834 to 1.386 | 0.4998 | ANCOVA |
| Secondary | mRAP, baseline to Week 26 | -0.09 | -1.003 to 0.830 | 0.8528 | ANCOVA |
| Secondary | Total pulmonary resistance, baseline to Week 26 | 2.40 | -69.368 to 74.178 | 0.9474 | ANCOVA |
| Secondary | Cardiac index, baseline to Week 26 | 0.13 | -0.066 to 0.328 | 0.1902 | ANCOVA |
| Secondary | Venous oxygen saturation, baseline to Week 26 | -1.20 | -2.737 to 0.327 | 0.1227 | ANCOVA |
| Secondary | Disease progression | HR 0.59 | 0.32–1.09 | 0.0867 | Log-rank test |
23. Related Tutorials
Learn more about the methods used in this trial:
24. Related Calculators
25. Sources
- ClinicalTrials.gov: TRITON — NCT02558231. Official trial registry record and source for the trial data summarized on this page.
- PubMed: PMID 34593120. Linked publication identified in the ClinicalTrials.gov record.
Continue through the Clinical Biostats statistical pathway
Explore the statistical methods that connect randomized trial design, continuous endpoints, confidence intervals, and time-to-event analysis.
26. Record Summary
TRITON provides a useful example of how several statistical frameworks can coexist within a single randomized clinical trial. The registered primary endpoint uses a Week 26-to-baseline PVR ratio analyzed with ANCOVA and reported as a ratio of geometric least square means. The registry also reports ANCOVA analyses for functional, biomarker, and hemodynamic outcomes, while disease progression is analyzed with a log-rank test and summarized with a hazard ratio.
The most important statistical distinction is between the point estimate and the uncertainty around it. The primary PVR estimate was 0.96, with a 95% CI of 0.86–1.07 and P = 0.4239. The disease-progression analysis reported HR 0.59, with a 95% CI of 0.32–1.09 and P = 0.0867. Neither estimate should be interpreted without its corresponding effect scale, confidence interval, analysis population, and testing context.