This page separates reported trial information from statistical interpretation. The ClinicalTrials.gov record reports results for the registered endpoints, but the ClinicalTrials.gov record contains no formal statistical-analysis entries with estimates, confidence intervals, or p-values.
1. Trial at a Glance
EXCURSION is a completed phase 3, single-group trial assessing dupilumab in children with asthma. The registry describes a treatment-purpose study with no masking and one arm, with 378 participants enrolled.
| Feature | EXCURSION |
|---|---|
| Trial name | EXCURSION |
| Brief title | Assessment of the Safety and Efficacy of Dupilumab in Children With Asthma (Liberty Asthma Excursion) |
| Phase | Phase 3 |
| Status | COMPLETED |
| Condition | Asthma |
| Allocation | NA |
| Design model | SINGLE_GROUP |
| Masking | NONE |
| Primary purpose | TREATMENT |
| Enrollment | 378.0 |
| Arms | 1 |
| Lead sponsor | Sanofi |
| Sponsor type | INDUSTRY |
| ClinicalTrials.gov | NCT03560466 |
2. Clinical Question
The registry describes EXCURSION as an assessment of the safety and efficacy of dupilumab in children with asthma. Because the registered design is single-group with one arm and no masking, the central statistical structure is different from a randomized controlled trial: the study does not provide a randomized concurrent comparator against which a between-arm treatment effect can be estimated.
Population
Children with asthma enrolled in the EXCURSION phase 3 study.
Intervention
Dupilumab (SAR231893/REGN668), together with asthma controller therapies including prednisone/prednisolone and asthma reliever therapies as listed in the registry.
Comparator
The registry specifies a SINGLE_GROUP design and one arm. No randomized comparator arm is identified in the ClinicalTrials.gov record.
Primary questions
How many participants experience treatment-emergent adverse events, and how does pre-bronchodilator percent predicted FEV1 change from baseline to Week 12 in the Japan sub-study?
3. Trial Design
Dupilumab-based treatment
- Dupilumab (SAR231893/REGN668)
- Asthma controller therapies, including prednisone/prednisolone
- Asthma reliever therapies
No randomized comparator
- Allocation listed as NA
- Design model listed as SINGLE_GROUP
- Masking listed as NONE
- One arm reported in the ClinicalTrials.gov record
Why the single-group design changes the statistical question
In a two-arm randomized trial, the primary estimand is often a contrast between treatment groups. In EXCURSION, the registry-reported design does not contain such a randomized contrast. Consequently, the statistical interpretation must focus on quantities that can be estimated within the study population, such as the number or proportion of participants with an event and change in a physiological measurement over time.
This distinction is important because a change observed after treatment cannot automatically be interpreted as the causal difference between treatment and no treatment. Without a concurrent randomized comparator, changes may also reflect natural disease variation, background therapy, measurement variation, regression to the mean, or other factors.
4. Endpoints
The registry lists two primary endpoints. Both have results posted in the ClinicalTrials.gov record, but the ClinicalTrials.gov record contains no formal analysis entries with effect estimates, confidence intervals, or p-values.
| Primary endpoint | Time frame | Registry endpoint type |
|---|---|---|
| Main Study: Number of Participants With Any Treatment-Emergent Adverse Events (TEAEs) | From first dose of study treatment (Day 1) up to 112 days post last dose of study treatment, approximately 64 weeks | Binary |
| Japan Sub-study: Change From Baseline to Week 12 in Pre-Bronchodilator Percent Predicted Forced Expiratory Volume in 1 Second (FEV1) | Baseline (Day 1) to Week 12 | Binary, as classified in the ClinicalTrials.gov record |
Primary Endpoint 1: Treatment-Emergent Adverse Events
The registry defines an adverse event as any untoward medical occurrence in a participant or clinical investigation participant administered a pharmaceutical product that does not necessarily have a causal relationship with the treatment. Treatment-emergent adverse events are adverse events that developed or worsened or became serious during the TEAE period.
The registered outcome is therefore an event-occurrence endpoint: whether a participant had any TEAE during the specified period from first dose through 112 days after the last dose, approximately 64 weeks.
Primary Endpoint 2: Change in Pre-Bronchodilator Percent Predicted FEV1
The Japan sub-study endpoint is defined as change from baseline to Week 12 in pre-bronchodilator percent predicted FEV1. The registry definition describes FEV1 as the volume of air, in liters, exhaled from the lungs in the first second of a forced expiration as measured by spirometer. Spirometry was performed after a washout period for bronchodilators according to their action duration.
The registry defines baseline as the last available measurement prior to the first study-treatment dose if the participant was treated, or the last available value up to the point described in the registry definition.
5. Planned Analysis
The ClinicalTrials.gov record reports that results are posted for both primary endpoints, but its no formal statistical analyses were posted. There are therefore no formal statistical-analysis estimates, confidence intervals, p-values, or named inferential models in the ClinicalTrials.gov record to reproduce.
The appropriate interpretation is consequently a planned-analysis framework: describe what the registry measures, identify the statistical approach ordinarily suited to each endpoint, and avoid presenting an unreported comparison or numerical treatment effect.
TEAEs: binary safety endpoint
For a binary endpoint defined as whether a participant experienced at least one TEAE during a prespecified period, a typical analysis would summarize the number and proportion of participants with an event. A confidence interval for the event proportion can also quantify sampling uncertainty.
For EXCURSION, the registry endpoint is "Number of Participants With Any Treatment-Emergent Adverse Events (TEAEs)." The ClinicalTrials.gov record does not provide the formal statistical-analysis record needed to establish which estimator or confidence-interval method was prespecified.
Because this is a single-group study, a proportion of participants with TEAEs would describe the observed safety experience in the enrolled population. It would not, by itself, establish how that rate differs from an untreated population or from another treatment.
FEV1: change-from-baseline endpoint
For a continuous change-from-baseline endpoint, a typical analysis might summarize baseline FEV1, Week 12 FEV1, and the within-participant change, followed by an appropriate model for the change or follow-up value. Depending on the prespecified design and available covariates, approaches can include analysis of covariance (ANCOVA) or longitudinal models.
Those methods are described here as general statistical approaches, not as methods documented for EXCURSION. The ClinicalTrials.gov record does not identify an ANCOVA model, mixed model, covariance adjustment, hypothesis-testing framework, or confidence-interval method.
What cannot be inferred from the empty analysis record
- The ClinicalTrials.gov record does not identify a formal null hypothesis for either primary endpoint.
- They do not specify a treatment-effect estimator with a confidence interval.
- They do not provide p-values.
- They do not identify a multiplicity-adjustment strategy.
- They do not identify an imputation method for missing FEV1 measurements.
- They do not identify a prespecified covariance structure or longitudinal model.
- They do not provide a randomized comparator from which a between-group treatment effect could be calculated.
6. Statistical Methods Explained
Why is a single-group study different from a randomized trial?
A randomized trial estimates a treatment contrast by assigning participants to different interventions. In a single-group study, all enrolled participants belong to the same study arm. The analysis can describe outcomes within that group, but a randomized concurrent counterfactual is not available from the design itself.
That makes causal interpretation more limited. For example, an improvement in FEV1 from baseline can establish that the measured value changed over time in the study population, but the change alone does not establish how much of that change was caused by dupilumab.
How should a TEAE endpoint be analysed?
The registered TEAE endpoint is binary at the participant level: a participant either has at least one treatment-emergent adverse event during the defined period or does not. A typical descriptive analysis therefore reports the event count and corresponding proportion. Confidence intervals can be used to express uncertainty around the observed proportion.
Importantly, "any TEAE" does not measure severity, duration, frequency of individual events, or causal attribution. Those are separate safety questions.
Why does the FEV1 endpoint say "change from baseline"?
Change from baseline converts each participant's follow-up measurement into a within-participant difference. Conceptually, if \(Y_0\) is baseline percent predicted FEV1 and \(Y_{12}\) is the Week 12 value, the change is:
The actual inferential method depends on the statistical analysis plan. The registry-reported EXCURSION data do not specify which model or estimator was used.
Why is pre-bronchodilator FEV1 important to the endpoint definition?
The endpoint specifically uses pre-bronchodilator FEV1. That makes the measurement condition part of the endpoint definition rather than an incidental detail. The registry also specifies a bronchodilator washout period according to the action duration of the bronchodilator.
Standardizing the measurement condition can reduce variation caused by recent bronchodilator exposure. It does not, however, eliminate all measurement variability or establish causality in a single-group study.
What does a confidence interval add to a binary endpoint?
A confidence interval places the observed event proportion in the context of sampling uncertainty. It is not a range containing the true event status of individual participants, nor does it measure clinical importance.
For EXCURSION, the ClinicalTrials.gov record contains no confidence interval for the TEAE endpoint, so no interval should be fabricated from the enrollment count.
Why should results not be reconstructed from the enrollment number?
Knowing that 378 participants were enrolled does not tell us how many participants experienced a TEAE, how FEV1 changed, or how missing observations were handled. A sample size is a denominator-related design quantity, not an outcome estimate. Producing an efficacy or safety percentage from enrollment alone would create a result that is not contained in the trial data.
7. Safety Results
The ClinicalTrials.gov record provides serious adverse events by arm/category. Because the study is single-group but the registry safety data are reported under the labels below, those labels are preserved exactly rather than being collapsed into an inferred treatment comparison.
| Safety category | Affected / at risk |
|---|---|
| Main Study: Placebo-Dupilumab | 1/125 |
| Main Study: Dupilumab-Dupilumab | 6/240 |
| Japan Sub-study: Dupilumab | 4/13 |
These are counts of participants with serious adverse events and the corresponding numbers at risk as reported in the ClinicalTrials.gov record. They should not be converted into a comparative treatment effect because the ClinicalTrials.gov record does not define these categories as randomized comparator arms and do not provide a statistical analysis comparing them.
The serious-adverse-event entries provide useful descriptive safety information, but they answer a narrower question than a randomized comparative safety analysis. An affected/at-risk count describes how many participants in the specified registry category experienced a serious adverse event. It does not establish a relative risk versus an external or randomized control group.
The denominators also differ across the listed categories. Directly comparing the raw counts would therefore be misleading. Even comparing proportions would require careful attention to how the registry-defined categories were constructed and whether follow-up and exposure were comparable.
8. Understanding the Two Primary Endpoints Together
EXCURSION combines a safety endpoint with a pulmonary-function endpoint. These endpoints measure fundamentally different aspects of the study experience.
| Endpoint | What it measures | Statistical structure |
|---|---|---|
| Any TEAE | Whether a participant experienced at least one treatment-emergent adverse event during the specified period | Binary event occurrence |
| Change in pre-bronchodilator percent predicted FEV1 | Change from baseline to Week 12 in a pulmonary-function measurement | Change-from-baseline measurement; registry classification registry-reported as Binary |
A safety endpoint and a physiological efficacy endpoint should not be reduced to a single summary statistic. A treatment can have one pattern for adverse events and another for pulmonary function. Proper interpretation therefore requires keeping the endpoints distinct and preserving their different measurement scales.
9. Statistical Interpretation of a Single-Group FEV1 Change
A positive change from baseline in pre-bronchodilator percent predicted FEV1 would describe an increase in that measured pulmonary-function outcome over the specified period. A negative change would describe a decrease.
In a single-group study, a baseline-to-Week-12 change is not equivalent to a randomized treatment effect. Without a concurrent randomized comparator, the observed change cannot isolate the effect of dupilumab from other temporal or clinical influences.
A point estimate of mean or median change, if reported, would summarize the observed sample. Its confidence interval would describe statistical uncertainty around the corresponding population quantity under the chosen analysis. The ClinicalTrials.gov record contains no such estimate or interval, so none is presented here.
10. Analysis Population and Missing Data Considerations
The ClinicalTrials.gov record identifies enrollment of 378 participants and report results as posted, but they do not provide formal definitions for an intention-to-treat population, per-protocol population, safety population, or FEV1 evaluable population.
This matters particularly for a change-from-baseline endpoint. Participants may have baseline measurements but not a Week 12 measurement, or may have measurements that do not meet the endpoint's assessment conditions. The statistical handling of those observations can affect the resulting estimate.
11. Multiplicity, Interim Analysis, and Other Design Features
The ClinicalTrials.gov record identifies two registered primary endpoints but do not describe a multiplicity-adjustment strategy, interim-analysis boundary, alpha-spending procedure, Bayesian method, non-inferiority margin, or crossover procedure.
| Design topic | What the ClinicalTrials.gov record supports |
|---|---|
| Multiplicity | Two registered primary endpoints are identified; no multiplicity procedure is reported. |
| Interim analysis | No interim-analysis method is reported. |
| Bayesian methods | No Bayesian analysis is identified. |
| Non-inferiority margin | No non-inferiority design or margin is identified. |
| Crossover | No crossover procedure is reported. |
| Stratification | No statistical stratification factors are reported. |
| Missing-data method | No formal imputation or missing-data strategy is reported. |
Why the absence of a comparator is especially important
Multiplicity is only one part of trial interpretation. The more fundamental design issue here is that the ClinicalTrials.gov record identifies a single-group study. Even if a highly sophisticated statistical model were applied to the data, a model cannot create the randomized control group that the design does not contain.
12. Timeline
EXCURSION begins
The trial start date recorded in the ClinicalTrials.gov record is 2018-06-21.
Single-group treatment study
The study is classified as phase 3, SINGLE_GROUP, with one arm, no masking, and a primary purpose of treatment.
Primary completion
The recorded primary completion date is 2025-04-01. The ClinicalTrials.gov record identifies the trial status as COMPLETED.
13. What the Registry Results Can and Cannot Tell Us
What is documented
The registry identifies 378 enrolled participants, one study arm, two registered primary endpoints, 16 posted outcome measures, and serious-adverse-event counts for the listed registry categories.
What is not documented in the ClinicalTrials.gov record
No formal statistical-analysis records, effect estimates, confidence intervals, p-values, or inferential model specifications are reported.
Why that distinction matters
Posted outcome measures are not the same thing as a complete statistical-analysis record.
Appropriate interpretation
The page can explain the endpoint structure and the statistical methods ordinarily appropriate for those endpoints, while keeping unreported numerical efficacy results out of the analysis.
14. Important Limitations and Interpretation Issues
- Single-group design: the registry design does not provide a randomized concurrent comparator, limiting causal interpretation of within-study changes.
- No masking: the registry identifies masking as NONE. For outcomes involving clinical assessment or participant-reported experience, lack of masking can be relevant to potential measurement or reporting bias.
- Endpoint-scale discrepancy: the registry classification labels the FEV1 change endpoint as Binary even though the endpoint is expressed as a change in a pulmonary-function measurement. This classification should be preserved rather than silently corrected.
- Missing statistical-analysis record: no formal statistical analyses were posted. Formal model specifications, confidence intervals, and p-values therefore cannot be reproduced from the ClinicalTrials.gov record.
- Safety denominators: the reported serious-adverse-event categories have different denominators. Raw event counts should not be treated as directly comparable measures of safety burden.
- Missing-data uncertainty: no formal approach to missing FEV1 measurements is reported, so the effect of missingness on any change-from-baseline estimate cannot be evaluated from the provided information.
- Multiplicity: two primary endpoints are registered, but no multiplicity procedure is reported. The relationship between the two endpoints for formal type I error control cannot therefore be reconstructed.
- Generalizability: the enrolled population is described only at the level reported by the registry data here. No additional baseline characteristics should be inferred.
15. Why This Trial Matters Statistically
EXCURSION is a useful teaching case because it illustrates a statistical issue that is sometimes obscured when attention is focused only on endpoint names: study design determines what an endpoint can support as an inference.
| Concept | How it appears in EXCURSION |
|---|---|
| Single-group design | One study arm is reported, with allocation listed as NA. |
| Unmasked study | Masking is listed as NONE. |
| Binary safety endpoint | Any treatment-emergent adverse event is registered as a binary endpoint. |
| Change from baseline | Japan sub-study FEV1 is measured from baseline to Week 12. |
| Physiological endpoint | Pre-bronchodilator percent predicted FEV1 provides a pulmonary-function measure. |
| Safety denominators | Serious adverse events are reported as affected/at-risk counts for specific registry categories. |
| Statistical-analysis transparency | No formal statistical analyses were posted despite posted outcome measures. |
| Registry interpretation | Posted results and formal inferential analyses are distinct pieces of evidence. |
The trial therefore provides a practical example of how to read a registry without overinterpreting it. A careful statistical analysis asks not only "What endpoint was measured?" but also "What comparison does the design permit?" and "What statistical evidence has actually been reported?"
16. Related Tutorials
Learn more about the methods used in this trial:
17. Related Statistical Calculators
18. Sources
- ClinicalTrials.gov: EXCURSION, NCT03560466.
- Linked publication: PubMed record for PMID 42379525.
- Linked publication: PubMed record for PMID 39016623.
- Linked publication: PubMed record for PMID 37956679.
Continue through the Clinical Biostats statistical library
Explore statistical tutorials and calculators covering clinical-trial design, safety endpoints, change-from-baseline analysis, and related methods.
19. Record Summary
EXCURSION is a completed phase 3, single-group study of dupilumab in children with asthma, with 378 participants enrolled and one reported study arm. Its two registered primary endpoints address different statistical questions: any treatment-emergent adverse event from first dose through 112 days after the last dose, approximately 64 weeks, and change from baseline to Week 12 in pre-bronchodilator percent predicted FEV1 in the Japan sub-study.
The ClinicalTrials.gov record reports that results are posted for both endpoints and provide serious-adverse-event counts for three registry categories. However, the ClinicalTrials.gov record contains no formal analysis entries. The appropriate statistical interpretation is therefore to distinguish descriptive registry information from inferential results that have not been specified in the ClinicalTrials.gov record.
The most important methodological lesson is the relationship between design and inference. A single-group study can quantify outcomes within the enrolled population, but it does not create a randomized concurrent comparator. Consequently, a baseline-to-Week-12 FEV1 change, even if numerically available, would describe within-study change rather than a randomized treatment effect. Similarly, TEAE counts describe the observed safety experience and do not by themselves establish comparative risk.