This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. The numerical results on this page are restricted to the ClinicalTrials.gov record.
1. Trial at a Glance
PINNACLE was a completed phase 3, randomized, parallel-group, triple-masked treatment trial in subjects with moderate to very severe COPD. The ClinicalTrials.gov record reports 1615 enrolled subjects, 4 intervention arms, 1 registered primary endpoint, 6 posted outcome measures, and 6 posted statistical analyses.
| Feature | PINNACLE |
|---|---|
| Trial name | PINNACLE |
| Brief title | Multi-Center Study to Assess the Efficacy and Safety of PT003, PT005, and PT001 in Subjects With Moderate to Very Severe COPD (PINNACLE 2) |
| Phase | Phase 3 |
| Status | Completed |
| Therapeutic area | Pulmonology |
| Condition | Chronic Obstructive Pulmonary Disease (COPD) |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Triple |
| Primary purpose | Treatment |
| Enrollment | 1615 |
| Interventions | GFF MDI (PT003); GP MDI (PT001); FF MDI (PT005); Placebo |
| Lead sponsor | Pearl Therapeutics, Inc. |
| Sponsor type | Industry |
| ClinicalTrials.gov | NCT01854658 |
2. Clinical Question
The registry describes a phase 3 treatment study of three active interventions—GFF MDI (PT003), GP MDI (PT001), and FF MDI (PT005)—and placebo in subjects with moderate to very severe COPD. The statistical question represented by the posted analyses is whether each active intervention produced a different FEV1 value from placebo at specified early Day 1 time points.
Population
Subjects with moderate to very severe chronic obstructive pulmonary disease.
Interventions
GFF MDI (PT003), GP MDI (PT001), and FF MDI (PT005).
Comparator
Placebo.
Primary registered question
Change from baseline in morning pre-dose trough FEV1 at Week 24.
3. Trial Design
Study arms
GFF MDI (PT003)
- Drug intervention
- Compared with placebo in the posted onset-of-action analyses
GP MDI (PT001)
- Drug intervention
- Compared with placebo in the posted onset-of-action analyses
FF MDI (PT005)
- Drug intervention
- Compared with placebo in the posted onset-of-action analyses
Placebo
- Drug intervention listed as placebo
- Comparator for the posted Day 1 FEV1 analyses
4. Trial Timeline
Trial start
The registry lists the study start as 2013-07.
Primary completion
The registry lists primary completion as 2015-03.
Registry status
The trial is listed as completed, with results posted.
5. Registered Primary Endpoint
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Change From Baseline in Morning Pre-dose Trough FEV1 | Change from baseline in morning pre-dose trough forced expiratory volume in 1 second (FEV1) at Week 24. | At Week 24 |
This is the single registered primary endpoint in the ClinicalTrials.gov record. Its endpoint type is classified in the ClinicalTrials.gov record as “Other / unclear.” The ClinicalTrials.gov record does not contain a formal analysis of this Week 24 primary endpoint.
Conceptually, the endpoint compares each subject's Week 24 morning pre-dose trough FEV1 with that subject's baseline value. The registry wording should remain the reference definition for the actual trial endpoint.
6. Secondary Endpoint Analysed in the Posted Results
All 6 registry-reported statistical analyses concern the same secondary outcome measure: Onset of Action as Assessed by FEV1. The time frame is Day 1, the outcome unit is liters, and the analysis population is the Intent-to-Treat population with evaluable data and no imputation.
| Outcome | Time frame | Unit | Analysis population | Method |
|---|---|---|---|---|
| Onset of Action as Assessed by FEV1 | Day 1 | Liters | Intent-to-Treat population with evaluable data (no imputation) | ANCOVA |
The posted analyses evaluate the outcome at 5 min post dose and 15 min post dose. Each active treatment is compared separately with placebo.
7. Statistical Methodology
Analysis of covariance (ANCOVA)
The registry reports ANCOVA for all 6 statistical analyses. ANCOVA is a linear-model framework that compares a continuous outcome between treatment groups while accounting for one or more covariates. In a clinical trial, this can be useful when the outcome is continuous and baseline information is relevant to the comparison.
The treatment coefficient represents the adjusted difference associated with treatment after accounting for the covariate(s) included in the model. The exact covariates used in the PINNACLE ANCOVA models are not specified in the registry-reported statistical-analysis text, so they are not added here.
Continuous outcome
FEV1 is reported in liters in the analyses posted on ClinicalTrials.gov, making the analysed outcome continuous. For a continuous endpoint, an ANCOVA treatment comparison can provide an adjusted comparison on the original measurement scale.
Intention-to-treat analysis
The analyses posted on ClinicalTrials.gov use the Intent-to-Treat population with evaluable data for this outcome measure. The registry explicitly states “no imputation.” Thus, the analysis population is not described as every randomized subject regardless of evaluability; it is the ITT population restricted to subjects with evaluable data for the particular outcome.
Superiority hypothesis
Each posted analysis is classified as a superiority test. In a superiority framework, the statistical question is whether the active treatment and comparator differ in the specified direction of the tested endpoint rather than whether they remain within a prespecified non-inferiority margin.
Multiple treatment comparisons
The ClinicalTrials.gov record contains three active-versus-placebo comparisons, each evaluated at two Day 1 time points. That produces 6 posted analyses. Because six formal comparisons are represented, multiplicity is an important interpretive issue. The ClinicalTrials.gov record does not specify an adjustment procedure, alpha allocation, or hierarchical testing strategy, so none is assumed here.
8. Results
The ClinicalTrials.gov record contains 6 statistical analyses, all for the secondary endpoint Onset of Action as Assessed by FEV1. The reported method is ANCOVA, the hypothesis type is superiority, and every registry-reported P-value is <0.0001.
| Comparison | Time point | Method | Analysis population | P-value |
|---|---|---|---|---|
| FF MDI (PT005) vs Placebo MDI | Day 1, 5 min post dose | ANCOVA | ITT with evaluable data; no imputation | <0.0001 |
| FF MDI (PT005) vs Placebo MDI | Day 1, 15 min post dose | ANCOVA | ITT with evaluable data; no imputation | <0.0001 |
| GP MDI (PT001) vs Placebo MDI | Day 1, 5 min post dose | ANCOVA | ITT with evaluable data; no imputation | <0.0001 |
| GP MDI (PT001) vs Placebo MDI | Day 1, 15 min post dose | ANCOVA | ITT with evaluable data; no imputation | <0.0001 |
| GFF MDI (PT003) vs Placebo MDI | Day 1, 5 min post dose | ANCOVA | ITT with evaluable data; no imputation | <0.0001 |
| GFF MDI (PT003) vs Placebo MDI | Day 1, 15 min post dose | ANCOVA | ITT with evaluable data; no imputation | <0.0001 |
Posted statistical signal
All 6 reported ANCOVA analyses have P-values below 0.0001.
The ClinicalTrials.gov record does not provide the corresponding treatment-effect estimates, confidence intervals, standard errors, adjusted means, or adjusted mean differences.
The P-values indicate that each of the six specified ANCOVA comparisons produced a result with a P-value below 0.0001 under its corresponding statistical test. This is evidence against the null hypothesis used for each individual comparison within the stated testing framework.
However, the P-value does not tell us how large the FEV1 difference was. A very small P-value can occur with either a clinically large or clinically small effect when the data are sufficiently informative. Because the ClinicalTrials.gov record does not provide an effect estimate or confidence interval, the magnitude and precision of the FEV1 difference cannot be quantified from these records alone.
The P-value also does not represent the probability that the null hypothesis is true. It is calculated under a null model and describes how incompatible the observed data are with that model. Finally, six analyses are reported here, so the interpretation of individual P-values should be distinguished from a claim about the overall family of tests. The ClinicalTrials.gov record does not state whether a multiplicity adjustment was used.
Primary endpoint versus posted analyses
9. Interpreting the Six ANCOVA Comparisons
The six analyses form a simple 3 × 2 comparison structure: three active interventions are each compared with placebo at two Day 1 time points.
| Active treatment | 5 min post dose | 15 min post dose |
|---|---|---|
| FF MDI (PT005) | P < 0.0001 | P < 0.0001 |
| GP MDI (PT001) | P < 0.0001 | P < 0.0001 |
| GFF MDI (PT003) | P < 0.0001 | P < 0.0001 |
This structure is useful statistically because it separates which treatment is being compared from when the FEV1 outcome was measured. The fact that the same P-value threshold appears in all six analyses means the registry reports consistent statistical evidence across these six individual tests. It does not, by itself, establish that the size of the treatment effect was identical across treatments or time points.
10. Safety
The ClinicalTrials.gov record provides serious adverse events by treatment arm. These counts should be interpreted as the number affected divided by the number at risk within each listed arm.
| Arm | Serious adverse events | Affected / at risk |
|---|---|---|
| FF MDI (PT005) | Serious adverse events | 37/438 |
| GP MDI (PT001) | Serious adverse events | 37/439 |
| GFF MDI (PT003) | Serious adverse events | 36/510 |
| Placebo MDI | Serious adverse events | 15/223 |
The registry data do not provide a statistical comparison, confidence interval, or P-value for these serious adverse-event counts in the ClinicalTrials.gov record. Accordingly, this page reports the affected/at-risk counts without converting them into additional derived rates.
Safety counts and efficacy P-values answer different questions. The FEV1 analyses assess a continuous pulmonary-function outcome, whereas serious adverse events are event counts within treatment arms. A treatment cannot be characterized statistically from the efficacy P-values alone; safety requires its own descriptive and, where prespecified, comparative analysis.
11. Statistical Methods Explained
Why was ANCOVA used?
ANCOVA is appropriate for comparing a continuous outcome across treatment groups while incorporating relevant covariate information. In this trial, the reported outcome is FEV1 measured in liters. The registry explicitly identifies ANCOVA as the method for all six posted statistical analyses. The ClinicalTrials.gov record does not specify the individual covariates in the fitted models, so the precise model specification should not be inferred.
What does “no imputation” mean here?
The analysis population is described as the Intent-to-Treat population with evaluable data for the outcome measure, with no imputation. In practical terms, missing outcome values were not replaced by statistically generated values for these analyses. Subjects without evaluable data for the particular outcome therefore were not reported with an imputed FEV1 value in the reported analysis.
Why is the intention-to-treat population important?
Randomization establishes the treatment groups before outcome data are observed. An ITT framework preserves that randomized assignment as the basis for the comparison. Here, the registry specifies an ITT population with evaluable data for the particular outcome, which is more specific than simply saying that every randomized subject contributed an observed value.
What does P < 0.0001 tell us?
It tells us that the reported statistical test produced a P-value below 0.0001. It does not give the size of the treatment difference, its clinical importance, or the width of a confidence interval. Those quantities are needed to describe magnitude and precision.
Why are there six P-values?
There are three active-versus-placebo comparisons and two specified Day 1 assessment times: 5 min post dose and 15 min post dose. Each treatment-time combination has a separate ANCOVA result, giving 6 posted analyses.
Does P < 0.0001 prove that the treatments are clinically important?
No. Statistical significance and clinical importance are different concepts. A P-value addresses compatibility with a statistical null hypothesis; clinical importance depends on the magnitude of the observed difference, its precision, the clinical context, and the endpoint itself. The ClinicalTrials.gov record does not provide the treatment-effect estimates or confidence intervals needed for that magnitude assessment.
Why should the Week 24 primary endpoint be kept separate?
The registered primary endpoint is change from baseline in morning pre-dose trough FEV1 at Week 24, whereas the statistical analyses posted on ClinicalTrials.gov concern Day 1 onset of action. These are different endpoint definitions and time frames. A P-value from one should not be presented as the result for the other.
12. Randomization and Blinding
Randomization
The study used randomized allocation. Randomization is the design mechanism that creates the treatment groups before outcome information is known, supporting a causal comparison under the trial protocol.
Triple masking
The registry classifies the study as triple-masked. Masking is intended to reduce opportunities for knowledge of treatment assignment to influence trial conduct or assessment.
These design features matter even though the posted statistical method is ANCOVA. Statistical modeling cannot completely compensate for systematic bias introduced by a poorly controlled treatment assignment or outcome-assessment process. Randomization and masking therefore form part of the statistical credibility of the trial rather than being separate from the analysis.
13. Missing Data and Analysis Population
The ClinicalTrials.gov record provides unusually specific wording about the analysis population: “Intent-to-Treat population with evaluable data (no imputation) for this outcome measure.”
| Feature | registry registry information |
|---|---|
| Underlying population | Intent-to-Treat population |
| Outcome requirement | Evaluable data for the outcome measure |
| Imputation | No imputation |
| Outcome | Onset of Action as Assessed by FEV1 |
| Time frame | Day 1 |
The absence of imputation does not automatically mean that missing data are irrelevant. Excluding subjects without evaluable outcome data can affect the composition of the analysed population. The statistical consequences depend on why data are missing and whether missingness is related to treatment or other factors.
14. Multiplicity
Multiplicity is particularly relevant to the six posted analyses because the registry reports three active-versus-placebo comparisons at two different time points.
| Dimension | Comparisons represented |
|---|---|
| Active treatments | FF MDI (PT005); GP MDI (PT001); GFF MDI (PT003) |
| Comparator | Placebo MDI |
| Time points | 5 min post dose; 15 min post dose |
| Total posted analyses | 6 |
| Hypothesis type | Superiority |
| Reported method | ANCOVA |
| Reported P-values | <0.0001 for all 6 analyses |
When multiple hypotheses are tested, the probability of obtaining at least one apparently significant result can differ from the nominal error rate of an individual test. Common trial strategies include hierarchical testing, Bonferroni-type adjustments, or other multiplicity-control procedures. The registry-reported PINNACLE data do not specify which, if any, multiplicity adjustment was applied to these six analyses.
15. What the P-value Does — and Does Not — Mean
A P-value of <0.0001 means that the corresponding statistical test produced a P-value below 0.0001. Under the null hypothesis and assumptions of that test, the observed result is sufficiently inconsistent with the null model to cross that threshold.
It does not mean that the treatment has a particular percentage improvement, that there is a <0.01% probability the null hypothesis is true, or that the observed effect is necessarily clinically large.
A confidence interval around an ANCOVA treatment contrast would show the range of effect sizes compatible with the data under the specified statistical framework.
The outcome unit is liters. An effect estimate on the FEV1 scale would therefore describe an adjusted difference in pulmonary-function measurement. The clinical meaning of such a difference depends on its magnitude and context, neither of which can be recovered from the registry-reported P-values alone.
16. Results Reporting: What Is Known and What Is Not
| Question | Supported by the ClinicalTrials.gov record? | What can be stated |
|---|---|---|
| Were statistical analyses posted? | Yes | 6 analyses were posted. |
| What method was used? | Yes | ANCOVA for all 6 analyses. |
| What was tested? | Yes | Onset of Action as Assessed by FEV1 on Day 1. |
| Which comparisons were made? | Yes | Each of FF MDI, GP MDI, and GFF MDI versus placebo. |
| What were the time points? | Yes | 5 min and 15 min post dose. |
| What were the P-values? | Yes | <0.0001 for all 6 analyses. |
| What was the treatment-effect estimate? | No | Not included in the ClinicalTrials.gov record. |
| What was the confidence interval? | No | Not included in the ClinicalTrials.gov record. |
| What was the Week 24 primary-endpoint estimate? | No | Not included in the ClinicalTrials.gov record. |
| Was multiplicity adjustment specified? | No | Not included in the ClinicalTrials.gov record. |
This distinction is central to responsible trial reporting. A statistical result is not fully characterized by its P-value alone. For a continuous endpoint, an effect estimate and confidence interval are usually the most informative companions to the hypothesis test. The registry extract does not provide those quantities for these analyses.
17. Serious Adverse Events by Arm
The visual above is intentionally a count display rather than a rate comparison. The denominators differ across the four listed arms, and the ClinicalTrials.gov record does not include a formal statistical test of serious adverse events. The counts should therefore be read descriptively.
18. Limitations
- Primary endpoint results are not reported in the posted statistical-analysis records: the registered primary endpoint is Week 24 change from baseline in morning pre-dose trough FEV1, while all six registry-reported analyses concern Day 1 onset of action.
- No treatment-effect estimates: the analyses posted on ClinicalTrials.gov provide P-values but no adjusted mean differences or other effect estimates.
- No confidence intervals: the ClinicalTrials.gov record does not report confidence intervals for the six ANCOVA comparisons.
- Incomplete model specification: ANCOVA is identified as the method, but the ClinicalTrials.gov record does not identify the covariates or exact model specification.
- No imputation: the analyses use evaluable data and explicitly state no imputation. The consequences of missing outcome data therefore cannot be assessed from the ClinicalTrials.gov record alone.
- Multiplicity: six treatment-time comparisons are reported, but the ClinicalTrials.gov record does not specify a multiplicity adjustment or testing hierarchy.
- Safety inference: serious adverse-event counts are reported, but no comparative safety analysis is provided.
- Population detail: the ClinicalTrials.gov record identifies the disease severity range and enrollment but does not provide a baseline-characteristics table or subgroup results.
- Generalizability: the registry identifies a population with moderate to very severe COPD; applicability to other populations cannot be inferred from the ClinicalTrials.gov record alone.
19. Why This Trial Matters Statistically
PINNACLE is a useful teaching example because the ClinicalTrials.gov record place several core clinical-trial concepts next to one another: randomized allocation, triple masking, a continuous physiological endpoint, ANCOVA, intention-to-treat analysis, evaluable-data restrictions, multiple treatment comparisons, and interpretation of very small P-values.
| Concept | How it appears in PINNACLE |
|---|---|
| Randomization | The study uses randomized allocation. |
| Parallel-group design | The registry identifies a parallel design with 4 arms. |
| Blinding | The registry classifies the study as triple-masked. |
| Continuous endpoint | Onset of action is assessed using FEV1 measured in liters. |
| ANCOVA | All 6 registry-reported statistical analyses use ANCOVA. |
| ITT analysis | The analysis population is an Intent-to-Treat population with evaluable data. |
| No imputation | The registry-reported analysis notes explicitly state no imputation. |
| Superiority testing | All 6 posted analyses use superiority hypotheses. |
| Multiple comparisons | Three active treatments are each compared with placebo at two time points. |
| P-value interpretation | All 6 analyses report P < 0.0001, without registry-reported effect estimates or confidence intervals. |
| Endpoint hierarchy | The registered primary endpoint is Week 24 trough FEV1 change, whereas the analyses posted on ClinicalTrials.gov are secondary Day 1 onset-of-action analyses. |
| Safety analysis | Serious adverse events are reported as affected/at-risk counts by arm. |
20. Statistical Concepts in This Trial
Learn more about the methods used in this trial:
21. Related Statistical Calculators
22. Sources
- ClinicalTrials.gov: PINNACLE (NCT01854658).
- PubMed: Publication associated with PMID 42551404.
- PubMed: Publication associated with PMID 35815359.
- PubMed: Publication associated with PMID 32450869.
- PubMed: Publication associated with PMID 32021148.
- PubMed: Publication associated with PMID 28720336.
Continue through the Clinical Biostats statistical pathway
Use the trial's methods as a starting point for deeper study of ANCOVA, clinical-trial design, intention-to-treat analysis, P-values, and randomization.
23. Record Summary
PINNACLE is a completed randomized phase 3 trial with 1615 enrolled subjects, 4 parallel treatment arms, triple masking, and a registered primary endpoint of change from baseline in morning pre-dose trough FEV1 at Week 24. The ClinicalTrials.gov record concern a secondary Day 1 endpoint, “Onset of Action as Assessed by FEV1,” analysed using ANCOVA in the Intent-to-Treat population with evaluable data and no imputation.
The six posted analyses comprise FF MDI (PT005), GP MDI (PT001), and GFF MDI (PT003), each compared with placebo at 5 min and 15 min post dose. Every registry-reported analysis reports a superiority hypothesis and a P-value of <0.0001. The ClinicalTrials.gov record does not provide treatment-effect estimates or confidence intervals, so the magnitude and precision of those FEV1 differences cannot be quantified from the ClinicalTrials.gov record.
The most important statistical lesson is therefore the distinction between statistical evidence and effect-size information. The six P-values provide evidence from the reported ANCOVA tests, but they do not substitute for the corresponding adjusted treatment differences and confidence intervals. The distinction between the secondary Day 1 analyses and the registered Week 24 primary endpoint is equally important: results for one endpoint should not be transferred to another simply because both involve FEV1.