This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. The ClinicalTrials.gov record identifies the primary endpoint and report that results are posted, but does not provide a formal statistical analysis result for that endpoint. Accordingly, this page distinguishes the registered endpoint from the statistical analysis that would ordinarily be used to evaluate it.
1. Trial at a Glance
LIBERATE was a randomized, parallel-group, unmasked clinical trial evaluating Pulmonx endobronchial valves used in the treatment of emphysema. The registry lists 190 enrolled participants, two arms, and a primary binary endpoint assessing the percentage of participants meeting a clinically significant FEV1 improvement threshold at 1 year.
| Feature | LIBERATE |
|---|---|
| Trial name | LIBERATE |
| Brief title | Pulmonx Endobronchial Valves Used in Treatment of Emphysema (LIBERATE Study) |
| Condition | Emphysema |
| Phase | NA |
| Status | COMPLETED |
| Allocation | RANDOMIZED |
| Design model | PARALLEL |
| Masking | NONE |
| Primary purpose | TREATMENT |
| Enrollment | 190 |
| Arms | 2 |
| Interventions | EBV (device); Optimal Medical Management (other) |
| Lead sponsor | Pulmonx Corporation |
| Sponsor type | INDUSTRY |
| Results posted | Yes |
| Outcome measures posted | 4 |
2. Clinical Question
The registered primary question concerns whether the percentage of participants achieving a clinically significant improvement in FEV1 at 1 year differs between the Zephyr Valve EBV treatment arm and the control arm receiving optimal medical management.
Population
Study participants with emphysema enrolled in the LIBERATE study.
Intervention
EBV (device), described in the registered primary endpoint as the Zephyr Valve EBV treatment arm.
Comparator
Optimal Medical Management.
Primary question
What percentage of participants in each arm meets the clinically significant FEV1 improvement threshold at 1 year, and how do those percentages compare?
3. Trial Design
Zephyr Valve EBV treatment
- EBV (device)
- The registered primary endpoint refers to this as the Zephyr Valve EBV treatment arm.
- FEV1 was obtained immediately following bronchodilator therapy for the primary endpoint.
Optimal Medical Management
- Optimal Medical Management (other)
- The primary endpoint compares the percentage meeting the FEV1 threshold with the EBV treatment arm.
The registry describes the allocation as randomized and the design model as parallel. Masking is listed as none. The primary purpose is listed as treatment. These features make the treatment groups the fundamental unit of the primary comparison.
4. Trial Timeline
Study begins
The registered trial start is 2013-07.
Primary completion
The registered primary completion date is 2017-09.
Trial completed
The ClinicalTrials.gov record lists the study status as COMPLETED and indicates that results have been posted.
5. Primary Endpoint
| Endpoint | Registered definition | Time frame | Endpoint type |
|---|---|---|---|
| Forced Expiratory Volume in 1-second (FEV1) | The percentage of study participants in the Zephyr Valve EBV (Endobronchial Valves) treatment arm meeting the clinically significant threshold of >15% improved forced expiratory volume in one second (FEV1), obtained immediately following bronchodilator therapy, as compared to the percentage in the control arm. | 1 year | Binary |
What makes the endpoint binary?
The underlying measurement, FEV1, is continuous. The registered primary endpoint converts that continuous measurement into a yes/no outcome using a threshold: a participant either meets the specified criterion of >15% improvement or does not meet it.
The statistical comparison is therefore based on the proportion of participants classified as meeting the clinically significant threshold, rather than on a comparison of mean FEV1 values alone.
This distinction is important. A binary endpoint discards some of the numerical detail contained in the original continuous FEV1 measurement, but it can provide a clinically interpretable classification when a threshold has been prespecified.
6. Statistical Methodology
The ClinicalTrials.gov record identifies the primary endpoint as binary but does not provide a populated the posted statistical analyses record. Therefore, the formal statistical test, effect estimate, confidence interval, and p-value used for the registered primary endpoint cannot be established from the ClinicalTrials.gov record.
Primary comparison of proportions
For a binary endpoint, the most direct analysis compares the proportion of participants meeting the >15% FEV1 improvement threshold in the EBV group with the corresponding proportion in the optimal medical management group.
If the observed proportions were denoted by pEBV and pOMM, useful effect measures could include an absolute risk difference, risk ratio, or odds ratio:
These quantities answer different questions. The ClinicalTrials.gov record does not report values for them.
Potential hypothesis test
A two-group comparison of proportions can be evaluated using a chi-square test or, when appropriate for sparse cell counts, Fisher's exact test. A model-based approach such as logistic regression can also be used, particularly when adjustment for prespecified covariates or stratification factors is required.
Because the ClinicalTrials.gov record does not identify which procedure was prespecified or used, none of these methods should be presented as the actual LIBERATE primary analysis.
Confidence intervals
A confidence interval around a risk difference, risk ratio, or odds ratio quantifies statistical uncertainty in the estimated treatment contrast. The choice of interval method should be aligned with the effect measure and the analysis model.
Analysis population
The ClinicalTrials.gov record does not specify whether the primary endpoint was analyzed under an intention-to-treat population, a modified intention-to-treat population, a per-protocol population, or another prespecified population. That distinction matters for a binary endpoint because exclusions or missing endpoint assessments can change the denominator and therefore the estimated proportions.
7. Planned Analysis
The primary endpoint is the percentage of study participants meeting the clinically significant threshold of >15% improved FEV1 at 1 year, comparing the Zephyr Valve EBV treatment arm with the control arm. The registry identifies this as a binary endpoint and states that results are posted, but the ClinicalTrials.gov record does not contain a formal comparison.
What would be measured?
For each study arm, the percentage of participants whose FEV1 met the registered >15% improvement criterion at 1 year.
What would be compared?
The proportion meeting the criterion in the EBV group would be compared with the corresponding proportion in the optimal medical management group.
Typical effect measures
A risk difference, risk ratio, or odds ratio could express the between-group contrast, accompanied by a confidence interval.
Typical hypothesis test
A two-group test of proportions or an appropriate regression model could evaluate the null hypothesis of no between-group difference.
8. Statistical Methods Explained
Why is the primary endpoint binary if FEV1 is continuous?
FEV1 itself is a quantitative physiological measurement. The registered endpoint transforms it into a binary classification based on a clinically significant threshold of >15% improvement. This makes the primary outcome a proportion rather than a mean or median change.
What does a proportion measure?
A proportion answers the question: what fraction of participants met the prespecified criterion? For example, if 30 out of 100 participants met a criterion, the observed proportion would be 30%. The registry-reported LIBERATE data do not provide the observed numerator and denominator for the primary endpoint, so no such percentage can be calculated here.
Why might a risk difference be useful?
The risk difference expresses the absolute separation between the two treatment-group proportions. It is directly interpretable in percentage-point terms. If one group had a 40% response proportion and another had a 25% response proportion, the risk difference would be 15 percentage points. Those numbers are only an illustration and are not LIBERATE results.
What does a risk ratio measure?
The risk ratio compares the probability of meeting the binary endpoint in one group with the probability in the other. A risk ratio of 1 would indicate equal observed proportions. Values above or below 1 describe relative differences in the probability of meeting the endpoint.
What does an odds ratio measure?
An odds ratio compares the odds of meeting the endpoint between treatment groups. It is not the same as a risk ratio, particularly when the endpoint is common. Logistic regression commonly produces odds ratios when the binary outcome is modeled directly.
Why does the analysis population matter?
Binary endpoint analyses depend on both the numerator and denominator. If participants with missing assessments are excluded, or if eligibility for analysis differs from the randomized population, the estimated proportion can change. The ClinicalTrials.gov record does not specify the formal primary analysis population or missing-data rule.
Why should the statistical test not be inferred from the endpoint alone?
Several statistically valid methods can analyze a binary endpoint. The appropriate choice depends on the prespecified estimand, analysis population, covariate structure, sample size, missing-data strategy, and trial statistical analysis plan. Knowing that an endpoint is binary is not enough to establish which method was actually used.
9. Safety
The ClinicalTrials.gov record reports serious adverse events by arm using affected participants over participants at risk. These data provide a safety count for each treatment group, but they do not provide a formal statistical comparison or a broader adverse-event table.
| Study arm | Serious adverse events | At risk |
|---|---|---|
| Zephyr Valve Treatment and Optimal Medic | 77 | 128 |
| Optimal Medical Management | 24 | 62 |
The affected/at-risk counts correspond to the safety information reported in the ClinicalTrials.gov record. They should not be converted into a formal comparative effect estimate here because the ClinicalTrials.gov record does not report the analysis method, confidence interval, p-value, follow-up definition, or statistical estimand for serious adverse events.
10. Statistical Interpretation of the Primary Endpoint
The primary endpoint does not ask how much FEV1 changed on average. It asks whether each participant crossed a specified threshold of >15% improvement. The resulting analysis therefore focuses on the proportion of participants classified as meeting that threshold.
A risk difference would describe the absolute difference in the percentages meeting the threshold. A risk ratio would describe the relative probability of meeting it. An odds ratio would describe the relative odds. These measures should not be treated as interchangeable.
A confidence interval would describe uncertainty around the estimated between-group contrast under the chosen statistical model and sampling framework. It would not describe the range of individual FEV1 responses.
If a formal hypothesis test were reported, its p-value would quantify the compatibility of the observed data with the specified null hypothesis under the chosen test. It would not measure the magnitude or clinical importance of the treatment effect.
The ClinicalTrials.gov record contains no formal statistical analyses. Reporting a risk difference, risk ratio, odds ratio, confidence interval, or p-value would therefore require information not present in the provided trial dataset.
11. Design Features and What the Registry Supports
| Design topic | What the ClinicalTrials.gov record establishes | Statistical implication |
|---|---|---|
| Randomization | Allocation is listed as RANDOMIZED. | The primary treatment comparison is based on randomized treatment assignment. |
| Parallel design | Design model is PARALLEL. | Participants are evaluated within one of two parallel study arms rather than a crossover design. |
| Masking | Masking is listed as NONE. | The ClinicalTrials.gov record does not describe a blinded treatment assignment. |
| Primary purpose | TREATMENT. | The registered study purpose is treatment evaluation. |
| Multiplicity | One registered primary endpoint is reported. | No multiplicity procedure is reported in the ClinicalTrials.gov record. |
| Interim analysis | Not specified in the ClinicalTrials.gov record. | No interim-monitoring method should be inferred. |
| Missing-data method | Not specified in the ClinicalTrials.gov record. | No imputation procedure should be attributed to the trial. |
| Stratification | Not specified in the ClinicalTrials.gov record. | No stratified analysis should be attributed to the trial. |
| Bayesian methods | Not specified in the ClinicalTrials.gov record. | No Bayesian analysis should be attributed to the trial. |
| Non-inferiority margin | Not specified; the registry-reported primary endpoint is presented as a binary comparison. | No non-inferiority interpretation should be applied. |
| Crossover | Not specified in the ClinicalTrials.gov record. | No crossover adjustment should be assumed. |
This distinction between what is documented and what is merely statistically possible is central to a rigorous trial analysis. A statistical method should be attributed to LIBERATE only when the registry-reported evidence establishes that the method was actually part of the trial analysis.
12. Understanding the Endpoint Through a 2×2 Table
A binary primary endpoint can be organized into a simple two-by-two table. The four cells represent whether a participant was assigned to the EBV or optimal medical management group and whether the participant did or did not meet the >15% FEV1 improvement criterion.
| Meets >15% FEV1 threshold | Does not meet threshold | |
|---|---|---|
| EBV | Numerator for EBV response proportion | Remaining EBV participants in the analysis population |
| Optimal Medical Management | Numerator for control response proportion | Remaining control participants in the analysis population |
The actual cell counts for the primary endpoint are not included in the ClinicalTrials.gov record. Consequently, this table is a statistical framework rather than a reconstruction of LIBERATE's observed endpoint results.
13. Why This Trial Matters Statistically
LIBERATE provides a useful teaching example because its primary endpoint illustrates an important issue in clinical-trial analysis: a continuous physiological measurement can be transformed into a clinically defined binary outcome, after which the statistical problem becomes one of comparing proportions between randomized groups.
| Concept | How it appears in LIBERATE |
|---|---|
| Randomization | The allocation is listed as RANDOMIZED. |
| Parallel-group design | The design model is PARALLEL with 2 arms. |
| Binary endpoint | The primary endpoint classifies participants according to whether FEV1 improvement exceeds 15%. |
| Continuous-to-binary transformation | FEV1 is a continuous physiological measure, while the registered primary outcome is a threshold-based classification. |
| Proportion comparison | The endpoint is expressed as the percentage of participants meeting the threshold in each arm. |
| Effect measures | Risk difference, risk ratio, and odds ratio are possible ways to express a binary treatment contrast. |
| Confidence intervals | Would quantify uncertainty around a reported binary treatment-effect estimate. |
| Analysis populations | The denominator used for the primary endpoint is important, but the ClinicalTrials.gov record does not specify the formal analysis population. |
| Missing data | The ClinicalTrials.gov record does not specify how missing 1-year FEV1 assessments were handled. |
| Safety analysis | Serious adverse events are reported as affected participants over participants at risk by arm. |
14. Limitations
- No formal statistical-analysis entries: the ClinicalTrials.gov record reports that results are posted but contain no formal statistical analyses. Therefore, formal effect estimates, confidence intervals, and p-values for the primary endpoint cannot be reproduced from the ClinicalTrials.gov recordset.
- No primary endpoint cell counts: the actual numbers meeting and not meeting the >15% FEV1 threshold are not reported.
- Analysis population not specified: the ClinicalTrials.gov record does not establish the exact population used for the primary endpoint analysis or the rule for handling participants without an evaluable 1-year endpoint.
- No missing-data method: no imputation strategy, censoring rule, or other missing-data procedure is posted on ClinicalTrials.gov for the primary binary outcome.
- No stratification information: the ClinicalTrials.gov record does not identify any stratification factors or a stratified primary analysis.
- No multiplicity procedure: the ClinicalTrials.gov record does not identify a multiplicity-control strategy for the four posted outcome measures.
- No interim-analysis information: there is no registry-reported basis for attributing an interim monitoring procedure or alpha-spending method to the trial.
- Binary threshold: reducing a continuous FEV1 measurement to a threshold outcome can improve interpretability but does not preserve all information contained in the original measurement.
- Safety information is limited: the ClinicalTrials.gov record contains serious adverse-event counts by arm but not a complete safety analysis.
15. What Can and Cannot Be Concluded From the ClinicalTrials.gov record
Supported
The trial was randomized, parallel, unmasked, completed, enrolled 190 participants, used EBV and optimal medical management as its two interventions, and registered a binary primary FEV1 endpoint assessed at 1 year.
Also supported
The registry data report serious adverse events by arm as 77/128 for Zephyr Valve Treatment and Optimal Medic and 24/62 for Optimal Medical Management.
Not reported
The formal primary treatment-effect estimate, confidence interval, p-value, exact analysis population, missing-data method, and prespecified statistical test.
Therefore
The appropriate educational approach is to explain how the registered binary endpoint would ordinarily be analyzed without attributing an unsupported statistical result to LIBERATE.
16. Related Tutorials
Learn more about the methods used in this trial:
17. Related Statistical Calculators
18. Sources
- ClinicalTrials.gov: NCT01796392 — LIBERATE.
- PubMed: PMID 39515624.
- PubMed: PMID 32223724.
- PubMed: PMID 29787288.
Continue with the statistical methods behind clinical trials
Explore the underlying concepts used to design, analyze, and interpret randomized clinical trials, including binary endpoints, effect measures, confidence intervals, and randomization.
19. Record Summary
LIBERATE is a completed randomized, parallel-group trial of EBV versus optimal medical management in emphysema. The registry lists 190 enrolled participants and identifies a single primary binary endpoint: the percentage of participants meeting a clinically significant threshold of >15% improved FEV1, obtained immediately following bronchodilator therapy, at 1 year.
From a statistical perspective, the key feature is the conversion of a continuous physiological measurement into a threshold-based binary outcome. The resulting primary comparison is naturally expressed through proportions and can be summarized using measures such as a risk difference, risk ratio, or odds ratio, with corresponding uncertainty intervals and an appropriate hypothesis test.
However, the ClinicalTrials.gov record contains no formal statistical analyses. Although results are listed as posted, the available data do not provide the primary endpoint's observed proportions, treatment-effect estimate, confidence interval, p-value, formal analysis population, or missing-data method.
The ClinicalTrials.gov record does establish serious adverse-event counts of 77/128 for Zephyr Valve Treatment and Optimal Medic and 24/62 for Optimal Medical Management. These are descriptive safety data and should not be converted into an unsupported formal treatment comparison.