← Clinical Trials
Pulmonary Arterial Hypertension Phase 3 Superiority NCT00325442

FREEDOM-C: Complete Statistical Analysis of Oral Treprostinil in Pulmonary Arterial Hypertension

An independent statistical review of the randomized phase 3 FREEDOM-C trial evaluating oral treprostinil (UT-15C) sustained-release tablets versus placebo in participants with pulmonary arterial hypertension receiving background endothelin receptor antagonist and/or phosphodiesterase-5 inhibitor therapy.

Trial start: October 2006  ·  Primary completion: September 2008  ·  Enrollment: 354
Scope of this record

This page separates reported trial results from statistical interpretation. The numerical results and trial characteristics presented here are restricted to the ClinicalTrials.gov record for FREEDOM-C.

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

1. Trial at a Glance

FREEDOM-C was a randomized, parallel-group, quadruple-masked phase 3 treatment trial with 354 participants and two study arms. The primary endpoint was change in six minute walk distance (6MWD) from Baseline to Week 16, analyzed using ANCOVA with a reported Hodges-Lehmann effect estimate.

354
Enrollment
ClinicalTrials.gov record
2
Study Arms
Placebo vs active
11.0
Primary Estimate
6MWD, meters
0.072
Primary P-value
Two-sided 95% CI
FeatureFREEDOM-C
TrialFREEDOM-C
ClinicalTrials.gov identifierNCT00325442
Therapeutic areaPulmonology
ConditionPulmonary Hypertension
PhasePhase 3
StatusCOMPLETED
AllocationRANDOMIZED
Design modelPARALLEL
MaskingQUADRUPLE
Primary purposeTREATMENT
Enrollment354
InterventionsOral treprostinil (UT-15C) sustained release tablets; placebo
Lead sponsorUnited Therapeutics
Sponsor typeINDUSTRY

2. Clinical Question

The primary statistical question was whether oral treprostinil produced a different change in six minute walk distance from Baseline to Week 16 compared with placebo. The registered endpoint describes placebo-corrected change in 6MWD as an objective measure of patient functional status and notes its relationship to the historical clinical standard for assessing functional status in pulmonary arterial hypertension.

Population

Participants in a phase 3 treatment trial for pulmonary hypertension, with background pulmonary arterial hypertension therapy represented by endothelin receptor antagonist and/or phosphodiesterase-5 inhibitor subgroups in the reported analyses.

Intervention

Oral treprostinil (UT-15C) sustained-release tablets.

Comparator

Placebo.

Primary question

Does oral treprostinil improve the change in 6MWD from Baseline to Week 16 relative to placebo?

3. Trial Design

01
Randomize354 enrolled
02
Parallel armsActive vs placebo
03
Masked treatmentQuadruple masking
04
Assess changeBaseline to Week 16
05
Compare groupsANCOVA
ACTIVE

Oral treprostinil

  • Oral treprostinil (UT-15C)
  • Sustained-release tablets
  • Compared with placebo
CONTROL

Placebo

  • Placebo tablets
  • Parallel comparison
  • Used as the reference group for the reported analyses

The registry identifies the allocation as randomized and the design model as parallel. Masking is listed as quadruple. The trial was intended to evaluate treatment rather than a diagnostic or prevention question, and the registered hypothesis type for the posted analyses is superiority.

4. Endpoints

RoleOutcome measureTime frameTypeReported method
Primary Six Minute Walk Distance (6MWD) Baseline and 16 Weeks Continuous ANCOVA
Secondary Borg Dyspnea Score Baseline and 16 Weeks Continuous Wilcoxon rank-sum test
Secondary Dyspnea-Fatigue Index Baseline and 16 Weeks Continuous Wilcoxon sum-rank test
Secondary Six Minute Walk Distance (6MWD) Baseline and 12 weeks Continuous ANCOVA
Secondary Six Minute Walk Distance (6MWD) Baseline and 8 weeks Continuous ANCOVA
Secondary Six Minute Walk Distance (6MWD) Baseline and 4 weeks Continuous ANCOVA
Secondary Clinical Worsening Assessment Baseline and 16 Weeks Binary Fisher Exact

The registered primary endpoint is defined as placebo corrected change in six minute walk distance (6MWD) from Baseline to Week 16. The registry describes 6MWD as correlating with the historical clinical standard for assessing patient functional status in pulmonary arterial hypertension and as an objective measure of patient functional status by the American Thoracic Society.

Endpoint timing matters. The trial reported 6MWD at several prespecified time frames in the posted analyses: 4, 8, 12, and 16 weeks. These are separate comparisons and should not be interpreted as repeated versions of a single 16-week result without considering their distinct analysis time points.

5. Statistical Methodology

Endpoint / analysisStatistical methodEffect measureHypothesis
Primary 6MWD, Week 16 ANCOVA Hodges-Lehmann Superiority
Borg Dyspnea Score, Week 16 Wilcoxon rank-sum test Hodges-Lehmann Superiority
Dyspnea-Fatigue Index, Week 16 Wilcoxon sum-rank test Hodges-Lehmann Superiority
6MWD, Week 12 ANCOVA Hodges-Lehmann Superiority
6MWD, Week 8 ANCOVA Hodges-Lehmann Superiority
6MWD, Week 4 ANCOVA Hodges-Lehmann Superiority
Clinical Worsening Assessment Fisher Exact Not reported Superiority

The statistical-method profile is notably different from a time-to-event trial. There is no hazard ratio or Kaplan-Meier analysis in the statistical analyses posted on ClinicalTrials.gov. The principal continuous endpoint was analyzed with ANCOVA, while several patient-reported or scale-based measures were analyzed with a Wilcoxon procedure. The binary Clinical Worsening Assessment used Fisher's exact test.

Primary endpoint analysis

For the primary 6MWD analysis, the registry reports ANCOVA and an effect measure of Hodges-Lehmann. The reported comparison is between the placebo arm and the active arm, with change in 6MWD from Baseline to Week 16 as the analysis target.

Primary statistical result
Hodges-Lehmann estimate = 11.0 meters

95% two-sided confidence interval: 0 to 22 meters; P = 0.072.

ANCOVA is a linear-model approach for comparing continuous outcomes while accounting for relevant baseline information. For a change-from-baseline endpoint, this is important because participants can begin at different baseline levels of functional capacity. Adjusting for baseline information can reduce residual variability and produce a more informative between-group comparison than simply comparing raw post-treatment means.

6. Primary Endpoint Result: Six Minute Walk Distance

The primary endpoint was Six Minute Walk Distance (6MWD), measured at Baseline and 16 Weeks. The outcome was continuous and measured in meters.

Hodges-Lehmann estimate for change in 6MWD

11.0 meters

95% CI: 0 to 22 meters   ·   P = 0.072

ANCOVA; placebo arm vs active arm; superiority hypothesis.

Primary endpointReported result
OutcomeSix Minute Walk Distance (6MWD)
Time frameBaseline and 16 Weeks
Outcome unitmeters
ComparisonPlacebo Arm vs Active
MethodANCOVA
Effect measureHodges-Lehmann
Estimate11.0
95% CI0 to 22
P-value0.072
HypothesisSuperiority
Clinical Biostats interpretation

The reported Hodges-Lehmann estimate of 11.0 meters represents the estimated difference associated with the active-versus-placebo comparison for change in 6MWD from Baseline to Week 16 under the reported analysis.

The estimate does not mean that every participant receiving oral treprostinil increased their walking distance by exactly 11.0 meters, nor does it describe an individual patient's expected change. It is a group-level effect estimate from the randomized comparison.

The two-sided 95% confidence interval of 0 to 22 meters indicates the statistical uncertainty around the reported estimate. Because the lower endpoint is exactly 0, the interval reaches the null difference.

The P-value of 0.072 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 effect. A P-value should therefore not be substituted for the estimated 11.0-meter effect or its confidence interval.

Finally, the registry reports ANCOVA as the method and Hodges-Lehmann as the effect measure. The ClinicalTrials.gov record does not provide the model's complete covariate specification or additional analysis-population details, so the statistical interpretation should remain limited to what was reported.

7. Secondary Endpoint Results

The posted statistical analyses include six secondary comparisons. These extend the assessment beyond the primary 16-week 6MWD analysis to dyspnea, fatigue, earlier 6MWD measurements, and clinical worsening.

Borg Dyspnea Score

Change from Baseline to Week 16

0.0

95% CI: -1.0 to 0.0   ·   P = 0.062

Wilcoxon rank-sum test; Hodges-Lehmann estimate.

One subject in the placebo arm did not have a Baseline Borg score value. The analysis compared the placebo and active groups using the Wilcoxon rank-sum test.

Clinical Biostats interpretation

The Hodges-Lehmann estimate was 0.0, with a two-sided 95% confidence interval from -1.0 to 0.0. The P-value was 0.062. The estimate describes the reported group comparison; it does not establish that individual patients had no change in dyspnea. The confidence interval is also more informative than the P-value alone because it shows the range of effect estimates compatible with the analysis.

Dyspnea-Fatigue Index

Change from Baseline to Week 16

0.0

95% CI: 0.0 to 1.0   ·   P = 0.011

Wilcoxon sum-rank test; Hodges-Lehmann estimate.

Five subjects in the placebo arm and three subjects in the active arm did not have a Baseline dyspnea-fatigue index score. The reported comparison used the Wilcoxon sum-rank test.

Clinical Biostats interpretation

The reported Hodges-Lehmann estimate was 0.0, with a 95% confidence interval of 0.0 to 1.0 and P = 0.011. The statistical result should be interpreted in the context of the scale's direction and the exact construction of the change variable; the ClinicalTrials.gov record does not provide enough information to assign a clinical direction to the numerical value independently. The P-value indicates evidence against the null hypothesis used for the reported superiority comparison, but it does not quantify the magnitude of benefit.

Six Minute Walk Distance at Week 12

Change from Baseline to Week 12

13.0 meters

95% CI: 3.0 to 23.0 meters   ·   P = 0.015

ANCOVA; Hodges-Lehmann estimate.

Clinical Biostats interpretation

The reported estimate of 13.0 meters is the between-group effect estimate for change in 6MWD from Baseline to Week 12. The 95% confidence interval, 3.0 to 23.0 meters, does not include 0. The P-value of 0.015 indicates evidence against the null hypothesis under the reported superiority analysis. It should not, however, be interpreted as saying that 98.5% of the effect is real or that the probability of the treatment hypothesis being true is 98.5%.

Six Minute Walk Distance at Week 8

Change from Baseline to Week 8

9.0 meters

95% CI: 0.0 to 18.0 meters   ·   P = 0.051

ANCOVA; Hodges-Lehmann estimate.

Clinical Biostats interpretation

The estimated difference was 9.0 meters, with a two-sided 95% confidence interval from 0.0 to 18.0 meters and P = 0.051. The estimate suggests a positive between-group difference in the reported analysis, while the confidence interval reaches the null value. This illustrates why an effect estimate and confidence interval should be reported alongside a P-value rather than converting the result into a binary label based solely on whether P is above or below a threshold.

Six Minute Walk Distance at Week 4

Change from Baseline to Week 4

4 meters

95% CI: -2.4 to 12.0 meters   ·   P = 0.238

ANCOVA; Hodges-Lehmann estimate.

Clinical Biostats interpretation

The Week 4 estimate was 4 meters. Its 95% confidence interval ranged from -2.4 to 12.0 meters, crossing 0, and the P-value was 0.238. The result is therefore compatible with a range of underlying differences that includes both a small negative difference and a positive difference. It would be inappropriate to interpret the P-value as evidence that the treatment has exactly zero effect.

Clinical Worsening Assessment

Baseline to Week 16

P = 0.491

95% confidence level reported   ·   Fisher Exact

Binary outcome; placebo arm vs active; superiority hypothesis.

The registry reports Clinical Worsening Assessment as a binary endpoint involving participants. The posted analysis used Fisher Exact testing, with P = 0.491. No effect estimate is reported in the ClinicalTrials.gov record for this analysis.

Clinical Biostats interpretation

Fisher's exact test is appropriate for a comparison of categorical outcomes when exact inference is useful, particularly when sample counts may be limited. Here, the reported P-value of 0.491 summarizes the evidence from the reported exact test; it does not quantify the difference in clinical-worsening probabilities because no effect estimate is provided in the ClinicalTrials.gov record.

8. Secondary Results Summary

EndpointTime frameMethodEstimate95% CIP-value
Borg Dyspnea ScoreBaseline and 16 WeeksWilcoxon rank-sum0.0-1.0 to 0.00.062
Dyspnea-Fatigue IndexBaseline and 16 WeeksWilcoxon sum-rank0.00.0 to 1.00.011
6MWDBaseline and 12 weeksANCOVA13.03.0 to 23.00.015
6MWDBaseline and 8 weeksANCOVA9.00.0 to 18.00.051
6MWDBaseline and 4 weeksANCOVA4-2.4 to 12.00.238
Clinical Worsening AssessmentBaseline and 16 WeeksFisher ExactNot reported95% confidence level reported0.491
Multiple posted analyses require context. The trial data contain 17 statistical analyses, including the primary analysis, six secondary analyses, and ten post-hoc subgroup analyses. The ClinicalTrials.gov record does not report a multiplicity-adjustment procedure or alpha-allocation scheme for these analyses. Consequently, individual nominal P-values from secondary and post-hoc analyses should not automatically be treated as independent confirmatory tests.

9. Post-Hoc 6MWD Analyses by Baseline 6MWD Quartile

The registry contains four post-hoc analyses dividing the study population into quartiles according to Baseline 6MWD. Each comparison used ANCOVA and reported a Hodges-Lehmann estimate for change in 6MWD from Baseline to Week 16.

Baseline 6MWD subgroupEstimate95% CIP-value
Quartile 1: 126-302 Meters240 to 450.121
Quartile 2: 303-362 Meters15-7 to 410.069
Quartile 3: 363-397 Meters4-15 to 240.889
Quartile 4: 398-450 Meters0-23 to 240.966

The estimates vary across the four baseline-distance groups, from 24 meters in Quartile 1 to 0 meters in Quartile 4. The confidence intervals are relatively wide, particularly compared with the primary analysis, which is expected when a randomized population is divided into smaller subgroups.

How to read subgroup analyses

A subgroup-specific estimate does not by itself demonstrate that treatment effects differ between subgroups. To establish effect modification, one generally needs a formal treatment-by-subgroup interaction analysis. The ClinicalTrials.gov record does not report such an interaction analysis, so these four estimates should be viewed as post-hoc descriptive comparisons rather than proof of differential treatment effect across baseline 6MWD quartiles.

10. Post-Hoc Analyses by Lowest Dose Strength Available at Randomization

Three post-hoc analyses examined 6MWD according to the lowest study-drug dose strength available at randomization. Each used ANCOVA and a Hodges-Lehmann effect estimate for change from Baseline to Week 16.

Dose-strength subgroupEstimate95% CIP-value
Smallest Dose Available 1 mg5-16 to 280.853
Dose Strength 0.5 mg7-7 to 210.327
Study Drug Dose 0.25 mg29.51 to 730.085

The registry defines these subgroups by the minimum tablet strength available for initiation of study-drug dosing and dose titration. The estimates are 5, 7, and 29.5 meters, respectively. Their confidence intervals differ substantially in width, with the 0.25-mg subgroup having a particularly wide interval of 1 to 73 meters.

Statistical interpretation

These analyses are post-hoc subgroup analyses rather than the registered primary endpoint. Smaller subgroups generally provide less precise estimates, and the ClinicalTrials.gov record does not provide subgroup sample sizes or an interaction test. Therefore, the numerical differences among these estimates should not be interpreted as establishing a dose-strength interaction.

11. Post-Hoc Analyses by Background PAH Therapy

Three additional post-hoc analyses evaluated change in 6MWD according to background pulmonary arterial hypertension therapy. The registry defines the background-therapy subgroups according to treatment received for 90 days or greater at randomization.

Background therapy subgroupDefinition in registryEstimate95% CIP-value
ERA Receiving treatment with an ERA for 90 days or greater at randomization 5 -12 to 24 0.615
PDE5-I Receiving treatment with a PDE5-I for 90 days or greater at randomization 17 -6 to 40 0.230
ERA and PDE5-I Receiving treatment with an ERA and PDE5-I for 90 days or greater at randomization 10 -6 to 28 0.209

All three comparisons used ANCOVA and a Hodges-Lehmann effect estimate. The confidence intervals for all three background-therapy analyses include 0.

Why background therapy needs careful interpretation

These analyses ask a narrower question than the primary randomized comparison. They describe the reported treatment comparison within selected background-therapy groups. Because the analyses are post-hoc and the ClinicalTrials.gov record does not include interaction tests or subgroup sample sizes, they cannot independently establish that the treatment effect differs according to background therapy.

12. All Post-Hoc Analyses at a Glance

Post-hoc analysisEstimate95% CIP-value
Baseline 6MWD Quartile 1: 126-302 Meters240 to 450.121
Baseline 6MWD Quartile 2: 303-362 Meters15-7 to 410.069
Baseline 6MWD Quartile 3: 363-397 Meters4-15 to 240.889
Baseline 6MWD Quartile 4: 398-450 Meters0-23 to 240.966
Lowest dose strength: 1 mg5-16 to 280.853
Lowest dose strength: 0.5 mg7-7 to 210.327
Lowest dose strength: 0.25 mg29.51 to 730.085
Background PAH therapy: ERA5-12 to 240.615
Background PAH therapy: PDE5-I17-6 to 400.230
Background PAH therapy: ERA and PDE5-I10-6 to 280.209

13. Statistical Methods Explained

Why was ANCOVA used for 6MWD?

6MWD is a continuous outcome, and the registered primary endpoint concerns change from Baseline. ANCOVA provides a linear-model framework for comparing treatment groups while incorporating baseline information. This can improve precision when baseline functional capacity varies among participants. The trial data identify ANCOVA as the reported method but do not provide the complete model specification.

What is a Hodges-Lehmann estimate?

The Hodges-Lehmann approach provides an estimate of a treatment-group difference that is closely associated with rank-based inference. In this trial it appears as the reported effect measure for the ANCOVA analyses as well as for the Wilcoxon analyses. The important distinction is that the effect estimate communicates the magnitude of the group comparison, while the P-value communicates evidence against a null hypothesis under the specified test.

Why use the Wilcoxon rank-sum test for the Borg and Dyspnea-Fatigue endpoints?

The Wilcoxon rank-sum test is a nonparametric method for comparing two independent groups. It is useful when an outcome is ordinal, non-normally distributed, or otherwise better handled through ranks rather than a conventional normal-theory comparison. The registry reports this method for the Borg Dyspnea Score and Dyspnea-Fatigue Index.

Why is Fisher's exact test used for Clinical Worsening Assessment?

Clinical Worsening Assessment is reported as a binary endpoint. Fisher's exact test evaluates the association between treatment group and a binary outcome using an exact calculation rather than relying on a large-sample approximation. The registry specifically identifies Fisher Exact as the method for this endpoint.

What does a 95% confidence interval tell us here?

A two-sided 95% confidence interval describes the uncertainty around the reported effect estimate under the relevant statistical framework. For the primary 6MWD result, the interval is 0 to 22 meters around an estimate of 11.0 meters. It does not describe the range of individual patient responses and does not mean that there is a 95% probability that the true effect lies inside this particular interval.

Why should the P-value not be treated as the effect size?

A P-value depends on the observed data and the statistical model or test, including the amount of information available. It does not directly measure the size of a treatment effect. For FREEDOM-C, the primary analysis illustrates this distinction clearly: the estimated effect is 11.0 meters, its 95% confidence interval is 0 to 22 meters, and the P-value is 0.072. These three quantities answer different statistical questions.

Why are the post-hoc subgroup results not equivalent to primary results?

The registry labels these analyses as post-hoc. Dividing participants into baseline 6MWD quartiles, dose-strength groups, or background-therapy groups creates smaller analysis populations and increases the number of comparisons. Without a reported interaction test and a reported multiplicity-control strategy, a difference in one subgroup's P-value from another's does not establish treatment-effect heterogeneity.

14. Missing Baseline Measurements

The statistical analyses posted on ClinicalTrials.gov explicitly identify missing Baseline values for two secondary endpoints.

EndpointMissing Baseline information
Borg Dyspnea Score One subject in the placebo arm did not have a Baseline Borg score value.
Dyspnea-Fatigue Index Five subjects in the placebo arm and three subjects in the active arm did not have a Baseline dypsnea-fatigue index score.
Missing-data limitation: The ClinicalTrials.gov record identifies these missing Baseline observations but do not report an imputation method or a sensitivity analysis. It is therefore not appropriate to infer how the missing values were handled beyond the analysis-population information reported in the ClinicalTrials.gov record.

15. Safety

The ClinicalTrials.gov record reports serious adverse events by treatment arm. The affected counts are presented against the corresponding numbers at risk.

Safety measurePlacebo armActive arm
Serious adverse events33/17632/174
Serious adverse events: affected / at risk
Placebo
33/176
Active
32/174

These counts should be kept separate from the efficacy analysis. A serious-adverse-event count is a safety outcome, whereas the primary 6MWD analysis evaluates a continuous functional endpoint. The ClinicalTrials.gov record does not provide a formal statistical comparison or confidence interval for serious adverse events, so no such comparison is added here.

16. Confidence Intervals Across the Trial

The FREEDOM-C results provide a useful illustration of how confidence intervals help distinguish an estimated effect from its uncertainty.

AnalysisEstimate95% CIWhat the interval indicates
Primary 6MWD, Week 1611.00 to 22Includes the null difference at 0
Borg Dyspnea, Week 160.0-1.0 to 0.0Includes the null difference at 0
Dyspnea-Fatigue Index, Week 160.00.0 to 1.0Begins at the null value
6MWD, Week 1213.03.0 to 23.0Entire interval is above 0
6MWD, Week 89.00.0 to 18.0Includes the null difference at 0
6MWD, Week 44-2.4 to 12.0Includes both negative and positive differences

Confidence intervals are particularly useful when several time points are reported. The Week 4 estimate is 4 meters with an interval extending from -2.4 to 12.0, while the Week 12 estimate is 13.0 meters with an interval from 3.0 to 23.0. These are not interchangeable summaries: they correspond to different time frames and different statistical estimates.

17. Interpreting the Primary Result Without Overreading It

What the estimate says

The reported primary Hodges-Lehmann estimate is 11.0 meters for the active-versus-placebo comparison in change in 6MWD from Baseline to Week 16.

What it does not say

It does not mean that every participant improved by 11.0 meters, nor does it describe an individual patient's expected response.

What the CI says

The two-sided 95% confidence interval is 0 to 22 meters, showing uncertainty around the estimated group difference.

What the P-value says

P = 0.072 summarizes evidence against the reported null hypothesis; it is not a probability that the treatment works or a measure of effect size.

The primary result should therefore be read as a complete statistical object: estimate + confidence interval + P-value + endpoint definition + analysis method. Removing any one of these elements makes the result easier to misinterpret.

18. Comparing the Time Course of 6MWD Results

The posted analyses allow the 6MWD treatment comparison to be examined at four time points. These are distinct estimates rather than a reconstructed longitudinal curve.

Hodges-Lehmann estimate for change in 6MWD
Week 4
4
Week 8
9.0
Week 12
13.0
Week 16
11.0

The point estimates are 4 meters at Week 4, 9.0 meters at Week 8, 13.0 meters at Week 12, and 11.0 meters at Week 16. These values should not be interpreted as a fitted trajectory because the ClinicalTrials.gov record provides separate summary estimates rather than participant-level longitudinal measurements.

Statistical caution: comparing P-values across time points is not a substitute for formally testing a treatment-by-time interaction or modeling repeated measurements. The ClinicalTrials.gov record does not report such an analysis.

19. Why This Trial Matters Statistically

FREEDOM-C is a useful teaching example because it combines randomized treatment allocation, masking, a continuous functional endpoint, covariate-adjusted analysis, nonparametric testing, exact categorical testing, multiple assessment times, and post-hoc subgroup analyses.

ConceptHow it appears in FREEDOM-C
RandomizationThe trial is registered as randomized with two parallel arms.
BlindingMasking is registered as quadruple.
Continuous outcome6MWD is measured in meters and analyzed as a continuous endpoint.
ANCOVAUsed for the primary 6MWD analysis and several secondary/post-hoc 6MWD analyses.
Hodges-Lehmann estimationReported as the effect measure for the 6MWD and nonparametric analyses.
Wilcoxon testUsed for Borg Dyspnea Score and Dyspnea-Fatigue Index.
Fisher exact testUsed for the binary Clinical Worsening Assessment.
Confidence intervalsTwo-sided 95% intervals are reported for the primary and several secondary/post-hoc continuous analyses.
Multiple time points6MWD was analyzed at Weeks 4, 8, 12, and 16.
Post-hoc subgroup analysis6MWD was examined by baseline-distance quartile, dose strength, and background PAH therapy.
Missing baseline valuesThe registry identifies missing Baseline measurements for Borg Dyspnea Score and Dyspnea-Fatigue Index.
Safety analysisSerious adverse events are reported by randomized treatment arm.

20. Limitations

21. Statistical Interpretation vs Clinical Interpretation

Statistical interpretation

The primary analysis estimated an 11.0-meter active-versus-placebo difference in change in 6MWD at Week 16, with a two-sided 95% CI of 0 to 22 meters and P = 0.072.

Clinical interpretation

Clinical significance cannot be assigned from the P-value alone. The ClinicalTrials.gov record does not specify a minimally important difference for 6MWD, so no threshold-based clinical conclusion is added.

This distinction is important. Statistical evidence concerns the uncertainty and compatibility of the observed data with a specified hypothesis. Clinical importance requires a separate judgment about the magnitude of the outcome, the patient population, the endpoint's meaning, and an appropriate clinical threshold. The registry-reported FREEDOM-C data do not provide such a threshold.

22. What the P-values Across Endpoints Illustrate

AnalysisEstimateP-valueStatistical lesson
Primary 6MWD, Week 1611.00.072Effect estimates and CIs should accompany the P-value.
Borg Dyspnea Score0.00.062A P-value near a conventional threshold does not quantify effect magnitude.
Dyspnea-Fatigue Index0.00.011Statistical evidence does not by itself describe clinical magnitude.
6MWD, Week 1213.00.015Estimate and CI provide information beyond the P-value.
6MWD, Week 89.00.051A small change in P-value does not imply a categorical change in effect.
6MWD, Week 440.238A nonsignificant P-value does not prove absence of an effect.
Clinical Worsening AssessmentNot reported0.491Without an effect estimate, the P-value does not communicate the size of the group difference.

23. A Practical Reading of the FREEDOM-C Statistical Record

A disciplined reading starts with the registered endpoint rather than the most favorable secondary or subgroup result. The primary endpoint was change in 6MWD from Baseline to Week 16. Its reported estimate was 11.0 meters, the two-sided 95% confidence interval was 0 to 22 meters, and the P-value was 0.072.

The next layer is the secondary evidence. At Week 12, the 6MWD estimate was 13.0 meters with a 95% confidence interval of 3.0 to 23.0 meters and P = 0.015. At Week 8, the estimate was 9.0 meters with a confidence interval of 0.0 to 18.0 meters and P = 0.051. At Week 4, the estimate was 4 meters with a confidence interval of -2.4 to 12.0 meters and P = 0.238.

Finally, the post-hoc analyses should be kept in their proper statistical category. They explore whether the 16-week 6MWD comparison differs according to baseline 6MWD, minimum dose strength, or background therapy. They are informative as exploratory analyses, but the ClinicalTrials.gov record does not provide the interaction tests or multiplicity framework needed to turn these subgroup comparisons into definitive claims about heterogeneity.

The key statistical lesson: a clinical-trial result is not a single P-value. The endpoint definition, randomized comparison, analysis method, effect estimate, confidence interval, analysis population, multiplicity context, and timing all contribute to the interpretation.

24. Related Tutorials

Learn more about the methods used in this trial:

25. Related Calculators

26. Sources

Continue through the Clinical Biostats statistical pathway

Use the methods in FREEDOM-C as a starting point for deeper study of ANCOVA, nonparametric tests, confidence intervals, randomization, blinding, and categorical-data analysis.

27. Record Summary

FREEDOM-C provides a compact but statistically rich example of randomized clinical-trial analysis. The study used randomized parallel allocation and quadruple masking to compare oral treprostinil with placebo in a phase 3 treatment setting. Its registered primary endpoint was change in 6MWD from Baseline to Week 16, analyzed using ANCOVA with a reported Hodges-Lehmann estimate of 11.0 meters, a two-sided 95% confidence interval of 0 to 22 meters, and P = 0.072.

The broader statistical record includes nonparametric analyses of Borg Dyspnea Score and Dyspnea-Fatigue Index, earlier 6MWD assessments at Weeks 4, 8, and 12, and Fisher exact testing of Clinical Worsening Assessment. Ten additional post-hoc 6MWD analyses examine baseline-distance quartiles, lowest dose strength, and background PAH therapy.

The most important interpretive discipline is to distinguish the registered primary analysis from the secondary and post-hoc analyses. Estimates, confidence intervals, and P-values should be read together, while subgroup findings require particular caution because the ClinicalTrials.gov record does not report interaction testing or a multiplicity-control strategy. The same principle applies to safety: serious adverse events are reported as 33/176 in the placebo arm and 32/174 in the active arm, but the ClinicalTrials.gov record does not provide a formal statistical comparison.

Clinical Biostats methodology: The purpose of a trial-results page is not simply to reproduce individual registry fields. It is to reconstruct the statistical structure of the study—what was randomized, what was measured, how it was analyzed, what the estimates and uncertainty mean, and which interpretations are supported by the reported evidence.