This page separates reported trial results from statistical interpretation. The numerical trial information on this page is restricted to the ClinicalTrials.gov record data. The registry reports results for the posted outcome measures, but no formal statistical analyses are provided in the ClinicalTrials.gov record.
1. Trial at a Glance
OPERA II was a phase 2, completed, single-group study of long-term treatment with PF-00547659 in Crohn's Disease. The study enrolled 268 participants and had one intervention arm.
| Feature | OPERA II |
|---|---|
| Trial name | OPERA II |
| Brief title | A Study To Monitor Long-Term Treatment With PF-00547659 |
| Phase | Phase 2 |
| Condition | Crohn's Disease |
| Design model | Single-group |
| Allocation | NA |
| Primary purpose | Treatment |
| Enrollment | 268 |
| Arms | 1 |
| Intervention | PF-00547659 (drug) |
| Primary endpoint type | Binary |
| Trial status | Completed |
| Lead sponsor | Shire |
| Sponsor type | Industry |
| ClinicalTrials.gov | NCT01298492 |
2. Clinical Question
The registry describes OPERA II as a study to monitor long-term treatment with PF-00547659 in participants with Crohn's Disease. Because the design is single-group rather than randomized or controlled, the statistical question is fundamentally different from a two-arm comparative efficacy trial.
Population
Participants with Crohn's Disease enrolled in the phase 2 OPERA II study.
Intervention
PF-00547659.
Comparator
None. The registry identifies a single-group design and one study arm.
Primary question
To monitor the number of participants experiencing on-treatment adverse events, adverse events leading to withdrawal, and serious adverse events during the treatment period.
3. Trial Design
PF-00547659
- Intervention: PF-00547659 (drug)
- Study design: single-group
- Primary purpose: treatment
- Primary endpoint assessed through Week 72
No randomized comparator
- Allocation: NA
- Arms: 1
- No between-group treatment comparison is defined by the ClinicalTrials.gov record.
A single-group design can be appropriate when the principal objective is to characterize treatment exposure and safety rather than to estimate a comparative treatment effect. Its central statistical limitation is equally important: without a concurrent comparator, an observed event frequency cannot by itself establish how much of the observed event burden is attributable to the intervention relative to an alternative treatment or background rate.
4. Study Timeline
Study start
The OPERA II study began on July 1, 2011.
Primary completion
The registry lists July 27, 2016 as the primary completion date.
Primary endpoint time frame
The registered primary endpoint is assessed from the start of study treatment up to Week 72 during the treatment period.
5. Primary Endpoint
| Endpoint | Registry definition / time frame | Endpoint type |
|---|---|---|
| Number of Participants With On-Treatment Adverse Events (AEs), AEs Led to Withdrawal, and Serious Adverse Events (SAEs) | From start of study treatment up to Week 72 (Treatment Period) | Binary |
The registry definition states that adverse events included adverse drug reactions, illnesses with onset during the study, exacerbation of previous illnesses, clinically significant changes in physical examination findings, and abnormal objective test findings including ECG and laboratory findings.
The registry-reported definition also states that an SAE was defined as any AE at any dose that resulted in death, was life threatening with immediate risk of death, or required in-subject hospitalization or prolongation of hospitalization. The registry-reported trial-data excerpt ends during the SAE definition, so this page does not extend that definition beyond the text provided.
6. Results
ClinicalTrials.gov reports results for the study, but no formal statistical analyses were posted. Accordingly, the reported information can be described, but a treatment-effect estimate, confidence interval, or p-value cannot be attributed to the registry.
Primary Endpoint: Serious Adverse Events
Reported serious adverse events
PF-00547659 75 mg: affected / at risk
The ClinicalTrials.gov record reports 118 affected participants among 268 at risk for the PF-00547659 75 mg arm.
| Safety measure | Arm | Affected / at risk | Formal comparison |
|---|---|---|---|
| Serious adverse events | PF-00547659 75 mg | 118/268 | Not reported in registry-reported statistical analyses |
The reported 118/268 is a participant count: 118 of the 268 participants at risk were reported as affected by serious adverse events in the PF-00547659 75 mg arm. It is therefore a measure of the observed frequency of participants with the specified safety outcome during the reported study period.
It does not establish a comparative treatment effect, because OPERA II has one study arm and no concurrent comparator in the ClinicalTrials.gov record. It also does not mean that 118 participants experienced exactly one SAE each; the endpoint is framed around participants with events, while the ClinicalTrials.gov record does not provide the number of individual event episodes per participant.
No confidence interval or p-value is reported in the registry-reported statistical-analyses field. Consequently, there is no registry-reported formal inferential comparison to interpret here. A confidence interval for a single-group proportion could quantify sampling uncertainty, but it would not create a randomized treatment comparison that the study design does not contain.
Other Posted Outcome Measures
The registry data states that 3 outcome measures are posted and that results are posted for the primary endpoint. However, the registry-reported trial-data excerpt does not provide the definitions, numerical results, or formal statistical analyses for the other posted outcome measures. They therefore cannot be reconstructed here without adding information outside the ClinicalTrials.gov record.
7. Statistical Methodology
Descriptive analysis of a binary safety endpoint
The primary endpoint is identified in the registry as binary. For a single-group safety endpoint, the fundamental descriptive quantity is the proportion of participants experiencing the event during the prespecified observation period:
For OPERA II, the ClinicalTrials.gov record reports the serious-adverse-event result as 118/268. The registry does not supply a formal confidence interval or hypothesis test for this result.
This type of analysis answers a descriptive question: how many participants experienced the defined event during the specified period? It does not answer the comparative question: how does this frequency differ from what would have occurred under another treatment?
Why the binary endpoint is participant-based
A participant-level binary endpoint reduces a potentially complicated stream of individual safety events to an indicator for whether the participant met the endpoint definition during the observation period. This can make reporting straightforward, but it also discards information about the number, timing, duration, and recurrence of events unless those characteristics are reported separately.
Safety analysis population
The ClinicalTrials.gov record identifies an affected count and an at-risk denominator of 268 for PF-00547659 75 mg. Those two quantities define the reported denominator for this safety result. The ClinicalTrials.gov record does not provide an alternative analysis-population definition, such as an as-treated or per-protocol rule, so none is added here.
No comparative hazard ratio
Hazard ratios are useful when a randomized or otherwise comparative time-to-event analysis estimates the relative instantaneous event rate between groups. OPERA II is a single-group study in the ClinicalTrials.gov record. A hazard ratio comparing PF-00547659 with another treatment therefore has no defined treatment-group counterpart here.
No p-value without a specified null hypothesis
A p-value requires a null hypothesis. In a comparative randomized trial, that might be a null treatment effect. For a single-group safety proportion, a one-sample test could instead compare the observed proportion with a prespecified external benchmark.
8. Statistical Methods Explained
Why is this endpoint binary?
The registry identifies the primary endpoint type as Binary. In a binary safety endpoint, each participant is classified according to whether the specified event occurred during the defined treatment period. The resulting analysis is naturally expressed as a count and proportion rather than as a mean or median.
What does 118/268 tell us?
It tells us that the registry reports 118 affected participants among 268 at risk for serious adverse events in the PF-00547659 75 mg arm. The denominator is essential: the count alone would not communicate the frequency of the event within the study population.
Why can't this result establish a treatment effect?
Because there is no concurrent comparator in the registry-reported design. A single-group proportion can describe what occurred among treated participants, but attribution of an excess or reduction in events generally requires an appropriate reference, such as a randomized control group or a prespecified external benchmark with assumptions that can be justified.
Would a confidence interval be useful?
Yes. For a single-group proportion, a confidence interval can describe uncertainty around the underlying event probability under a specified statistical model. But the ClinicalTrials.gov record does not report such an interval, so this page does not calculate or substitute one.
Why does the observation period matter?
The endpoint has a specific time frame: from start of study treatment up to Week 72 (Treatment Period). Event frequency depends on how long participants are observed. Two studies with different follow-up windows cannot automatically be compared simply by comparing their percentages.
Why distinguish participants from events?
A participant can potentially experience more than one adverse event. A participant-level binary endpoint counts whether the participant experienced the defined category, whereas an event-level analysis could count individual episodes. The ClinicalTrials.gov record is an affected-participant count and does not provide the number of individual SAE episodes.
Why is a p-value not automatically appropriate?
A p-value is meaningful only relative to a specified null hypothesis and analysis model. The registry-reported OPERA II data contains no formal statistical analysis and no external null benchmark. Reporting a newly calculated p-value would therefore introduce an analytical assumption not contained in the trial data.
9. Safety Results
The ClinicalTrials.gov record provides one arm-specific serious-adverse-event result.
| Safety category | PF-00547659 75 mg |
|---|---|
| Serious adverse events | 118/268 affected / at risk |
The registry's primary endpoint also encompasses on-treatment adverse events and adverse events that led to withdrawal. Although these categories are part of the registered endpoint definition, the ClinicalTrials.gov record does not provide their numerical results. They are therefore not assigned values on this page.
What is reported
118 of 268 participants are reported as affected by serious adverse events in the PF-00547659 75 mg arm.
What is not reported here
The ClinicalTrials.gov record does not provide separate numerical counts for all AEs or AEs leading to withdrawal.
10. Single-Group Design and Statistical Interpretation
The most important statistical feature of OPERA II is not a complex model but the absence of randomization and a concurrent control group. The study's SINGLE_GROUP design means that the observed safety experience is interpreted primarily as a treatment-exposure description.
These are different estimands. A single-group safety frequency should not be described as a relative treatment effect without an appropriate comparator.
This distinction is especially important when reading safety data. The occurrence of an adverse event during treatment does not, by itself, prove that the intervention caused the event. Background disease activity, concomitant treatment, age, comorbidity, and other factors can contribute to observed safety outcomes. Establishing comparative causality requires a design and analysis capable of separating treatment exposure from those other influences.
11. Primary Endpoint: What Would Normally Be Analysed?
For a binary safety endpoint in a single-group study, a conventional analysis would begin with the number of participants with the event and the number evaluated. The event proportion would be the primary descriptive summary.
| Analytical component | Purpose | OPERA II the ClinicalTrials.gov record |
|---|---|---|
| Event count | Number of participants meeting the binary endpoint | 118 serious-AE-affected participants |
| Denominator | Number of participants at risk / evaluated | 268 |
| Event proportion | Describes observed frequency | Not separately reported as a percentage in the ClinicalTrials.gov record |
| Confidence interval | Quantifies sampling uncertainty | Not reported |
| Hypothesis test | Tests against a specified null value or comparator | No formal statistical analysis reported |
| Comparative effect estimate | Compares treatment with another group | Not defined by the single-group design |
If a prespecified external benchmark existed, a one-sample proportion framework could potentially test the observed rate against that benchmark. If the objective were simply estimation, an appropriate confidence interval for the proportion could be reported. Neither analysis is reported in the ClinicalTrials.gov record, so neither is presented as an OPERA II result.
12. Missing Data and Observation Time
The primary endpoint has a fixed registry time frame extending from the start of study treatment up to Week 72. In safety analyses, interpretation depends on who contributes to the denominator and for how long participants remain observable.
The ClinicalTrials.gov record does not specify a missing-data or imputation method, nor does it provide a detailed rule for handling participants who discontinue before Week 72. Accordingly, this page does not assign a specific imputation strategy to OPERA II.
13. Multiplicity and Endpoint Structure
The registry identifies 1 primary endpoint and reports 3 outcome measures. The ClinicalTrials.gov record does not provide a multiplicity adjustment, endpoint hierarchy beyond the designation of the single primary endpoint, or a formal testing strategy for the three posted outcome measures.
| Feature | What the ClinicalTrials.gov record supports |
|---|---|
| Primary endpoints | 1 |
| Primary endpoint type | Binary |
| Outcome measures posted | 3 |
| Formal statistical analyses | None reported |
| Multiplicity adjustment | Not reported in the ClinicalTrials.gov record |
The distinction between one primary endpoint and several posted outcomes is important. A collection of outcomes does not automatically imply that all were tested as confirmatory hypotheses, and the ClinicalTrials.gov record does not establish such a testing hierarchy.
14. Randomization, Stratification, Crossover, and Bayesian Methods
Randomization
The registry identifies the allocation as NA and the design model as single-group. No randomized allocation is reported in the ClinicalTrials.gov record.
Stratification
No stratification factors are provided in the ClinicalTrials.gov record.
Crossover
No crossover design or crossover analysis is provided in the ClinicalTrials.gov record.
Bayesian methods
No Bayesian method is reported in the statistical analyses posted on ClinicalTrials.gov.
These omissions are analytically meaningful. A comparative hazard ratio, stratified log-rank test, treatment-by-stratum interaction, crossover adjustment, or Bayesian posterior estimate would require design or analysis information that is not present in the ClinicalTrials.gov record.
15. Interpreting the Reported Safety Count
The reported 118/268 is an observed participant count for serious adverse events in the PF-00547659 75 mg arm. It summarizes how many participants were classified as affected among the reported at-risk population.
It does not quantify a randomized treatment effect, does not compare PF-00547659 with placebo or another treatment, and does not establish that every reported serious adverse event was caused by PF-00547659.
The denominator of 268 defines the population against which the affected count is interpreted. Without the denominator, the clinical scale of the count would be difficult to assess.
The ClinicalTrials.gov record does not report a confidence interval. A confidence interval would quantify statistical uncertainty around a single-group proportion, but it would not turn a single-group observation into a comparative treatment effect.
16. Limitations
- No concurrent comparator: the single-group design limits the ability to distinguish treatment-associated event frequency from background risk.
- No formal statistical analyses reported: the registry data provides no confidence intervals or p-values in the statistical-analyses field.
- Limited numerical results in the ClinicalTrials.gov record: the explicit numerical safety result provided is 118/268 for serious adverse events in the PF-00547659 75 mg arm.
- Incomplete outcome detail: three outcome measures are reported as posted, but their definitions and numerical results are not included in the registry-reported trial-data excerpt.
- Observation period: the primary endpoint is defined through Week 72, so safety frequencies depend on the specified treatment-period observation window.
- Participant-level endpoint: the binary endpoint does not, by itself, describe the number of individual adverse-event episodes or their severity distribution.
- Missing-data methods: the ClinicalTrials.gov record does not identify an imputation or incomplete-follow-up strategy.
- Causality: an adverse event occurring during treatment is not equivalent to proof that the intervention caused the event.
17. Why This Trial Matters Statistically
OPERA II is useful as a teaching case because it illustrates a statistical situation that is sometimes overshadowed by more elaborate randomized-trial methods: how to interpret safety evidence when there is no concurrent comparator.
| Concept | How it appears in OPERA II |
|---|---|
| Single-group design | One study arm and no randomized comparator are reported. |
| Binary endpoint | The primary endpoint is classified as binary. |
| Participant-level safety analysis | The endpoint counts participants with specified on-treatment safety events. |
| Denominator | The reported serious-AE result uses 268 participants at risk. |
| Descriptive inference | The reported count describes the observed safety experience within the study group. |
| Comparative inference | A treatment-effect comparison is not defined without a concurrent comparator. |
| Confidence intervals | Useful for quantifying uncertainty but not reported in the statistical analyses posted on ClinicalTrials.gov. |
| Hypothesis testing | Requires a prespecified null hypothesis or comparator; none is reported. |
| Observation time | The primary endpoint runs from treatment start through Week 72. |
| Safety interpretation | Observed events during treatment should be distinguished from proof of causation. |
The broader lesson is that statistical sophistication does not begin with choosing a complicated model. It begins with identifying the estimand that the design can legitimately support. In OPERA II, the ClinicalTrials.gov record supports a description of observed participant-level safety outcomes; it does not support a randomized relative treatment-effect estimate.
18. Related Tutorials
Learn more about the methods used in this trial:
19. Related Statistical Calculators
20. Sources
- ClinicalTrials.gov: OPERA II — NCT01298492.
- Linked publication: PubMed record for PMID 34427633.
Continue through the Clinical Biostats statistical library
Explore related biostatistics tutorials, statistical calculators, and clinical trial analyses.
21. Record Summary
OPERA II is a phase 2, single-group study of long-term treatment with PF-00547659 in Crohn's Disease, enrolling 268 participants. Its registered primary endpoint is the number of participants with on-treatment adverse events, adverse events leading to withdrawal, and serious adverse events from the start of study treatment through Week 72. The ClinicalTrials.gov record reports 118/268 participants affected by serious adverse events in the PF-00547659 75 mg arm, but the registry-reported statistical-analyses field contains no formal comparison, confidence interval, or p-value.
Statistically, the key lesson is the distinction between describing an observed safety frequency and estimating a comparative treatment effect. The former is supported by the single-group design and reported participant counts. The latter requires a suitable comparator or prespecified external benchmark and cannot be inferred from the reported 118/268 result alone.