This page separates reported trial results from statistical interpretation. Numerical trial results are taken only from the ClinicalTrials.gov record for NCT01082367. Where the registry does not provide a result or methodological detail, this page does not infer one from outside sources.
1. Trial at a Glance
EARLY was a randomized, quadruple-masked, phase 3 crossover study in cystic fibrosis patients between 3 months and less than 7 years. The trial evaluated TOBI and placebo, with the primary endpoint defined as the percentage of participants who were P Aeruginosa-free after completion of the first treatment cycle.
| Feature | EARLY |
|---|---|
| Trial name | EARLY |
| NCT ID | NCT01082367 |
| Phase | Phase 3 |
| Therapeutic area | Pulmonology |
| Condition | Treatment of Early Pulmonary Infections With P. Aeruginosa in Cystic Fibrosis Patients |
| Population | Cystic fibrosis patients between 3 months and less than 7 years |
| Allocation | Randomized |
| Design model | Crossover |
| Masking | Quadruple |
| Primary purpose | Treatment |
| Enrollment | 50 |
| Interventions | TOBI; Placebo |
| Primary endpoint type | Binary |
| Primary analysis | Logistic regression |
| Effect measure | Odds ratio |
| Hypothesis type | Superiority |
| Lead sponsor | Novartis Pharmaceuticals |
| Sponsor type | Industry |
| Status | Completed |
2. Clinical Question
The primary statistical question was whether the randomized treatment sequences involving TOBI differed from the placebo/TOBI sequence with respect to the probability of being P Aeruginosa-free after completion of the first treatment cycle.
Population
Cystic fibrosis patients between 3 months and less than 7 years enrolled in the EARLY phase 3 trial.
Intervention
TOBI, a study intervention administered within the randomized crossover sequences.
Comparator
Placebo, administered within the randomized crossover sequences.
Primary question
Does the treatment sequence involving TOBI versus the placebo/TOBI sequence differ in the odds of being P Aeruginosa-free after the first treatment cycle?
3. Trial Design
EARLY used randomized allocation with a crossover design and quadruple masking. The registry identifies two interventions, TOBI and placebo, and the posted primary analysis compares the sequences TOBI/Placebo versus Placebo/TOBI.
TOBI / Placebo
- TOBI during the first treatment cycle
- Placebo in the subsequent crossover sequence
Placebo / TOBI
- Placebo during the first treatment cycle
- TOBI in the subsequent crossover sequence
4. Primary Endpoint
| Endpoint | Registry definition | Time frame | Type |
|---|---|---|---|
| Percentage of Participants P Aeruginosa-free After Completion of the First Treatment Cycle | Sputum/throat swab cultures were assessed. | Day 29 | Binary |
The primary endpoint is binary at the participant level: each participant is classified according to whether they were P Aeruginosa-free after completion of the first treatment cycle. The registry describes the underlying assessment as sputum/throat swab cultures.
The analysis therefore models the odds of the binary endpoint rather than directly modeling a continuous percentage. The reported treatment effect is an odds ratio comparing the two randomized sequences.
5. Analysis Population and Statistical Method
The posted primary analysis used an intent-to-treat (ITT) population. The registry defines the ITT population as all randomized participants who received at least one dose of study treatment.
| Component | Reported specification |
|---|---|
| Analysis population | Intent-to-treat: all randomized participants who received at least one dose of study treatment |
| Groups compared | TOBI/Placebo vs Placebo/TOBI |
| Outcome | Percentage of participants P Aeruginosa-free after completion of the first treatment cycle |
| Time frame | Day 29 |
| Outcome type | Binary |
| Regression method | Logistic regression |
| Effect measure | Odds ratio |
| Hypothesis | Superiority |
| Confidence interval | 95%, two-sided |
The ClinicalTrials.gov record does not specify additional covariates, stratification factors, interaction terms, or a detailed model-building strategy. Accordingly, this page does not attribute any such features to the EARLY analysis.
6. Primary Result
The registry contains one formal statistical analysis for the primary endpoint. It reports a logistic regression comparison in the ITT population between the TOBI/Placebo and Placebo/TOBI sequences.
Odds ratio for being P Aeruginosa-free
95% CI: 4.67–99.52 · P < 0.001
Two-sided 95% confidence interval · Superiority hypothesis
| Primary endpoint | Analysis population | Comparison | Effect | 95% CI | P-value |
|---|---|---|---|---|---|
| P Aeruginosa-free after completion of first treatment cycle, Day 29 | ITT | TOBI/Placebo vs Placebo/TOBI | OR 21.55 | 4.67–99.52 | <0.001 |
The reported odds ratio of 21.55 means that the estimated odds of being P Aeruginosa-free at Day 29 were 21.55 times as high for the TOBI/Placebo sequence as for the Placebo/TOBI sequence in the reported logistic regression analysis.
This is an odds ratio, not a probability ratio. It does not mean that 21.55 times as many participants were P Aeruginosa-free, and it cannot be translated into a percentage-point difference without the underlying event proportions.
The 95% confidence interval of 4.67–99.52 is wide. It indicates that, under the statistical model and sampling framework, there is substantial uncertainty about the precise magnitude of the odds ratio even though the entire reported interval is above 1.
The P-value < 0.001 addresses evidence against the null hypothesis represented by an odds ratio of 1 under the specified statistical test. It does not measure the size of the treatment effect, the probability that the treatment is effective, or the clinical importance of the result.
The analysis was performed in an ITT population, as defined by the registry. Because this is a crossover trial, interpretation also depends on the fact that the reported primary endpoint concerns the first treatment cycle, rather than an aggregate outcome across both treatment periods.
7. Understanding the Odds Ratio
For a binary endpoint, logistic regression works with the odds of the event. If p is the probability of being P Aeruginosa-free, the odds are:
The odds ratio compares those odds between two groups. An odds ratio of 1 represents equal odds. An odds ratio greater than 1 indicates higher odds in the numerator group, while an odds ratio below 1 indicates lower odds.
For EARLY, the reported odds ratio of 21.55 is therefore a relative comparison on the odds scale. The large numerical value should not be confused with a 2,055% increase in probability or any particular absolute increase in the percentage of participants who were culture-negative.
What 21.55 means
The estimated odds of the binary endpoint were 21.55 times higher for TOBI/Placebo than for Placebo/TOBI in the reported model.
What 21.55 does not mean
It does not mean that the probability of being P Aeruginosa-free was 21.55 times higher or that 21.55 additional participants benefited.
8. Confidence Interval and Precision
The reported 95% confidence interval extends from 4.67 to 99.52. Both limits are substantially above 1, while the interval itself spans a large range of possible odds-ratio magnitudes.
The lower endpoint indicates an estimated odds ratio of at least 4.67 within the reported confidence interval, while the upper endpoint reaches 99.52. The width of the interval is important: statistical evidence against an odds ratio of 1 does not imply that the exact magnitude of the effect has been estimated with high precision.
This distinction is especially important when the sample size is limited. the ClinicalTrials.gov record reports enrollment of 50 participants. The registry result does not provide the underlying Day 29 event counts for the two randomized sequences, so this page does not reconstruct the odds ratio or calculate an alternative confidence interval.
9. P-values and Superiority Testing
The registry classifies the primary hypothesis as superiority and reports a P-value of <0.001 for the logistic regression analysis.
What the P-value addresses
It quantifies how incompatible the observed data and test statistic are with the null hypothesis used for the statistical comparison, under the assumptions of the analysis.
What it does not address
It does not quantify the size of the odds ratio, the precision of the estimate, or the probability that the null hypothesis is true.
For interpretation, the P-value and confidence interval should be considered together with the effect estimate. Here, the reported OR of 21.55 and its 95% CI of 4.67–99.52 provide the effect-size and uncertainty information that the P-value alone cannot provide.
10. Safety Results
The ClinicalTrials.gov record includes serious adverse-event counts by treatment period and analysis context. These counts should be read as affected participants over the corresponding participants at risk; they are not equivalent to an incidence percentage unless the denominator is explicitly used for that purpose.
| Safety context | Affected / at risk |
|---|---|
| DB TOBI | 0 / 32 |
| DB Placebo | 1 / 38 |
| OL TOBI (Core) | 0 / 22 |
| Off-treatment (Core) | 2 / 50 |
| OL TOBI (Follow-up) | 0 / 10 |
| Off-treatment (Follow-up) | 4 / 29 |
The available safety data support reporting the serious-adverse-event counts above. The ClinicalTrials.gov record does not provide a broader table of all adverse events, severity categories, discontinuations, or a formal statistical comparison of serious adverse events between randomized sequences.
11. Crossover Design: Why It Matters
A crossover trial differs from a conventional parallel-group trial because treatment assignment specifies a sequence through which participants may receive different interventions at different stages. EARLY's posted primary comparison is explicitly expressed as TOBI/Placebo versus Placebo/TOBI.
Participants receive TOBI in the first cycle and placebo in the later sequence.
Participants receive placebo in the first cycle and TOBI in the later sequence.
For the primary endpoint, the time frame is Day 29 after completion of the first treatment cycle. That feature is statistically important: the primary analysis does not require this page to assume that outcomes after the crossover can simply be treated as additional independent observations.
12. Quadruple Masking
the ClinicalTrials.gov record identifies the masking level as quadruple. Masking is a design feature intended to reduce the potential for knowledge of treatment assignment to influence trial conduct or assessment.
For a binary microbiologic endpoint based on sputum/throat swab cultures, maintaining masking can be relevant to the integrity of treatment administration, participant management, and assessment procedures. The registry data identify the trial as quadruple-masked but do not specify the individual roles included in the masking designation.
What is documented
The trial's masking designation is quadruple.
What is not assumed
The specific four masked roles are not reported in the ClinicalTrials.gov record and are not reported here.
13. Intention-to-Treat Analysis
The registry states that the primary analysis used an intent-to-treat population consisting of all randomized participants who received at least one dose of study treatment.
The purpose of analyzing participants according to their randomized sequence is to preserve the comparison generated by randomization as much as possible. The specific EARLY registry definition is narrower than a purely theoretical definition that would necessarily include every randomized participant regardless of treatment exposure, so the registry wording is retained here.
The ITT designation is particularly relevant in a crossover study because participants are assigned to sequences rather than simply to one treatment for the entire trial. The posted analysis nevertheless identifies the comparison and endpoint in a way that can be interpreted at the first-cycle Day 29 assessment.
14. Statistical Methods Explained
Why was logistic regression used?
The primary endpoint is binary: participants are classified according to whether they were P Aeruginosa-free after the first treatment cycle. Logistic regression is designed for binary outcomes and expresses the treatment comparison through an odds ratio.
What does an odds ratio of 21.55 mean?
An odds ratio of 21.55 means that the estimated odds of being P Aeruginosa-free were 21.55 times higher for the TOBI/Placebo sequence than for the Placebo/TOBI sequence in the reported analysis. It is not a 21.55-fold increase in probability.
Why can an odds ratio be much larger than the difference in percentages?
Odds and probabilities are different scales. When an event is uncommon in one group, converting a difference in probabilities to odds can produce a substantially different relative measure. Without the underlying event counts, the absolute probability difference cannot be recovered from the reported odds ratio alone.
Why is the confidence interval so important here?
The estimate is 21.55, but the 95% CI ranges from 4.67 to 99.52. That wide interval indicates that the precise magnitude of the odds ratio is uncertain even though the interval remains above 1.
Why does the P-value not measure effect size?
The P-value of <0.001 describes evidence against the relevant null hypothesis under the specified analysis. It does not tell us whether the odds ratio is 2, 20, or 200, nor does it describe the absolute percentage of participants who were P Aeruginosa-free.
Why does the crossover design matter?
Participants were assigned to treatment sequences, TOBI/Placebo or Placebo/TOBI. Because the primary endpoint is measured after the first treatment cycle at Day 29, the posted analysis can be interpreted around that first-cycle comparison without assuming that the later treatment period is statistically interchangeable with the first.
Why should the first-cycle endpoint be distinguished from later treatment?
The registry defines the primary endpoint specifically as being P Aeruginosa-free after completion of the first treatment cycle. Extending the interpretation to later periods would introduce information not contained in the registered primary endpoint or its posted statistical analysis.
15. What the Primary Odds Ratio Does — and Does Not — Mean
The reported odds ratio of 21.55 indicates substantially higher estimated odds of the binary primary endpoint for TOBI/Placebo relative to Placebo/TOBI in the posted logistic regression analysis.
The corresponding 95% CI of 4.67–99.52 indicates uncertainty around the effect estimate. The interval does not describe the range of individual participant responses.
The odds ratio does not provide the percentage of participants who were P Aeruginosa-free in either sequence, the absolute risk difference, the number needed to treat, or the probability that an individual participant would be culture-negative.
The point estimate is large, but the upper confidence limit is also large. A statistically strong result and a precisely estimated treatment effect are different concepts. The EARLY result illustrates why both the point estimate and the confidence interval belong in the primary result display.
16. Secondary Endpoints and Additional Results
the ClinicalTrials.gov record states that 3 outcome measures were posted and that 1 statistical analysis was posted. The formal statistical analysis reported in the ClinicalTrials.gov record is for the primary endpoint only.
| Result category | Information available in the ClinicalTrials.gov record |
|---|---|
| Primary endpoint | Formal logistic regression analysis with odds ratio, 95% CI, and P-value |
| Secondary endpoints | Outcome measures are reported as posted, but no additional formal statistical analyses are reported |
| Statistical analyses posted | 1 |
| Primary-Endpoint Analyses | 1 |
| Primary analyses with estimate + CI | 1 |
For a binary secondary endpoint, a logistic regression analysis and an odds ratio could be appropriate depending on the prespecified statistical analysis plan, but the ClinicalTrials.gov record does not establish that such an analysis was used for the other outcomes.
17. Missing Data, Imputation, and Analysis Assumptions
The ClinicalTrials.gov record does not describe a missing-data strategy, imputation method, missingness assumptions, or sensitivity analysis for the primary endpoint.
Similarly, the registry does not provide the full logistic-regression specification. This page therefore explains the implications of logistic regression generally but does not attribute particular covariates, coding choices, interaction terms, or sensitivity analyses to EARLY.
18. Design Features Not Reported in the ClinicalTrials.gov record
| Design / analysis topic | What the ClinicalTrials.gov record supports |
|---|---|
| Non-inferiority margin | Not reported; the hypothesis type is superiority. |
| Factorial design | Not reported; the design model is crossover. |
| Interim analysis | Not reported in the ClinicalTrials.gov record. |
| Alpha-spending method | Not reported in the ClinicalTrials.gov record. |
| Multiplicity adjustment | Not reported in the ClinicalTrials.gov record. |
| Bayesian methods | Not reported; the posted method is logistic regression. |
| Missing-data imputation | Not reported in the ClinicalTrials.gov record. |
| Stratification factors | Not reported in the registry-reported statistical analysis. |
| Carryover analysis | Not reported in the ClinicalTrials.gov record. |
| Period-effect analysis | Not reported in the ClinicalTrials.gov record. |
Keeping these distinctions explicit is important. A statistical analysis page should explain what a design permits researchers to investigate without silently converting standard methodological possibilities into claims about what this particular trial actually did.
19. Trial Timeline
Trial start
The registered trial start was April 2010.
Primary completion
The registered primary completion date was June 2015.
Trial status
The registry identifies EARLY as completed.
The ClinicalTrials.gov record does not provide additional interim data-cut dates or a sequence of formal efficacy analyses. Accordingly, the statistical story presented here centers on the one posted primary analysis.
20. Safety and Efficacy Should Be Kept Separate
The primary efficacy analysis and serious-adverse-event counts answer different questions. The efficacy analysis asks whether the treatment sequences differ on the Day 29 binary microbiologic endpoint. Safety counts describe participants experiencing serious adverse events within specified treatment or follow-up contexts.
Efficacy
The formal primary analysis reports OR 21.55, 95% CI 4.67–99.52, and P < 0.001 for the Day 29 binary endpoint.
Safety
The registry provides serious-adverse-event counts such as 0/32 for DB TOBI and 1/38 for DB Placebo, with additional open-label and follow-up contexts.
Neither component should be reduced to a single overall numerical judgment. The treatment-effect estimate describes the primary endpoint, while the safety counts describe a separate dimension of the trial experience.
21. Important Limitations and Interpretation Issues
- Small enrolled population: the ClinicalTrials.gov record reports enrollment of 50 participants. A limited sample can contribute to uncertainty in effect estimation.
- Wide confidence interval: the primary OR is 21.55, but the 95% CI extends from 4.67 to 99.52. The statistical evidence and the precision of the magnitude estimate are therefore distinct issues.
- Crossover design: treatment sequences introduce design considerations that are different from a parallel-group study, including the possibility of period and carryover effects.
- First-cycle endpoint: the registered primary endpoint is specifically assessed after completion of the first treatment cycle at Day 29. It should not be generalized automatically to later treatment periods.
- Binary endpoint: the odds ratio is not a risk ratio or probability ratio. Without the underlying event counts, an absolute effect cannot be reconstructed from the reported OR alone.
- Model specification: the ClinicalTrials.gov record identifies logistic regression but does not provide the full model specification, covariates, or sensitivity analyses.
- Missing-data strategy: no imputation or missing-data handling rule is provided in the ClinicalTrials.gov record.
- Secondary analyses: only one formal statistical analysis is reported, so additional outcome measures should not be assigned unreported effect estimates.
- Safety context: serious-adverse-event counts are posted on ClinicalTrials.gov for distinct treatment and follow-up contexts and should not be pooled without an appropriate analysis framework.
22. Why This Trial Matters Statistically
EARLY is a useful statistical teaching case because it combines a randomized crossover design with a binary primary endpoint and a logistic-regression treatment comparison. The resulting odds ratio illustrates several concepts that are important when interpreting clinical trial results beyond simply asking whether a P-value is below a conventional threshold.
| Concept | How it appears in EARLY |
|---|---|
| Randomization | Participants were randomized in a phase 3 treatment trial. |
| Crossover design | The trial used a crossover model with TOBI/Placebo and Placebo/TOBI sequences. |
| Quadruple masking | the ClinicalTrials.gov record identifies quadruple masking. |
| Binary endpoint | The primary endpoint classifies participants as P Aeruginosa-free or not after the first treatment cycle. |
| Logistic regression | The posted formal analysis used logistic regression. |
| Odds ratio | The treatment effect was reported as OR 21.55. |
| Confidence interval | The 95% two-sided CI was 4.67–99.52. |
| P-value | The reported P-value was <0.001. |
| ITT analysis | The posted analysis used the registry-defined ITT population. |
| Superiority testing | The registered hypothesis type was superiority. |
| Safety denominators | Serious-adverse-event counts were reported separately by treatment and follow-up context. |
The most instructive feature is the difference between statistical evidence and effect-size precision. The P-value is reported as <0.001, yet the confidence interval is broad. A rigorous interpretation therefore reports the odds ratio, its confidence interval, and the P-value together rather than allowing the P-value to stand in for the entire analysis.
23. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The posted logistic regression analysis produced an odds ratio of 21.55 with a two-sided 95% CI of 4.67–99.52 and P < 0.001 for the Day 29 primary binary endpoint in the reported ITT population.
Clinical interpretation
The endpoint concerns being P Aeruginosa-free after completion of the first treatment cycle, assessed using sputum/throat swab cultures. The reported statistical result describes the odds comparison for that specific endpoint and time frame.
The distinction matters because a statistical association on the odds scale does not by itself describe absolute benefit, duration of benefit, later treatment-period outcomes, or the complete safety profile.
24. Statistical Concepts in This Trial
Learn more about the methods used in this trial:
25. Related Statistical Calculators
The most relevant calculator pathway for this analysis focuses on binary outcomes, logistic regression, effect-size interpretation, and uncertainty around estimates.
26. Sources
- ClinicalTrials.gov: NCT01082367 — EARLY.
- PubMed: Publication associated with PMID 29685813: PubMed record for PMID 29685813.
Continue through Clinical Biostats
Explore related biostatistics tutorials, statistical calculators, and additional clinical trial analyses.
27. Record Summary
EARLY is a phase 3 randomized crossover trial in cystic fibrosis patients between 3 months and less than 7 years, using TOBI and placebo with quadruple masking. Its registered primary endpoint was the percentage of participants who were P Aeruginosa-free after completion of the first treatment cycle, assessed at Day 29 using sputum/throat swab cultures.
The posted formal analysis used logistic regression in the registry-defined ITT population and compared the TOBI/Placebo and Placebo/TOBI sequences. The reported odds ratio was 21.55, with a two-sided 95% confidence interval of 4.67–99.52 and a P-value <0.001 under a superiority hypothesis.
The principal statistical lesson is that the three reported quantities answer different questions: the odds ratio describes the estimated relative effect on the odds scale, the confidence interval describes uncertainty around that estimate, and the P-value describes evidence against the relevant null hypothesis under the analysis. None of these quantities alone supplies the absolute probability of being P Aeruginosa-free.