This page separates reported trial information from statistical interpretation. The numerical and design information presented here is limited to the ClinicalTrials.gov record data. The registry reports that results have been posted, but the ClinicalTrials.gov record does not provide formal statistical-analysis entries with treatment estimates, confidence intervals, or p-values for the registered primary endpoint.
1. Trial at a Glance
SHINE was a randomized, parallel-group, quadruple-masked phase 3 trial evaluating once-daily QVA149 in patients with moderate to severe COPD. The registry identifies trough forced expiratory volume in one second (FEV1) after 26 weeks of treatment as the primary endpoint.
| Feature | SHINE |
|---|---|
| Trial name | SHINE |
| NCT ID | NCT01202188 |
| Phase | Phase 3 |
| Condition | Chronic Obstructive Pulmonary Disease (COPD) |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Quadruple |
| Primary purpose | Treatment |
| Enrollment | 2144 |
| Status | Completed |
| Lead sponsor | Novartis Pharmaceuticals |
| Sponsor type | Industry |
2. Clinical Question
The SHINE trial addresses the efficacy, safety, and tolerability of once-daily QVA149 in patients with moderate to severe COPD. The ClinicalTrials.gov record identifies QVA149 as the combination of indacaterol and glycopyrronium and identifies the primary efficacy endpoint as trough FEV1 after 26 weeks of treatment.
Population
Patients with moderate to severe chronic obstructive pulmonary disease (COPD).
Intervention
Indacaterol and glycopyrronium (QVA149), administered once daily.
Comparators
Glycopyrronium (NVA237), indacaterol (QAB149), tiotropium, and placebo.
Primary question
How does once-daily QVA149 perform on the registered trough FEV1 endpoint after 26 weeks of treatment in the randomized SHINE population?
3. Trial Design
Study arms
QVA149
- Indacaterol and glycopyrronium
- Listed as a drug intervention
Glycopyrronium
- Glycopyrronium (NVA237)
- Listed as a drug intervention
Indacaterol
- Indacaterol (QAB149)
- Listed as a drug intervention
Tiotropium
- Tiotropium
- Listed as a drug intervention
Placebo
- Placebo
- Listed as a drug intervention
The ClinicalTrials.gov record does not provide the randomized sample size assigned to each of the five arms. The safety data do provide affected/at-risk counts for serious adverse events, which are reported separately below.
4. Trial Timing and Status
Study start
The registry lists September 2010 as the study start.
Primary completion
The registry lists February 2012 as the primary completion date.
Current trial status in the ClinicalTrials.gov record
The trial is classified as completed, and results are posted in ClinicalTrials.gov.
5. Endpoints
The ClinicalTrials.gov record identifies one registered primary endpoint. Its definition is unusually specific about the timing of the spirometry measurements and the mixed-model framework.
| Endpoint | Registry definition / time frame | Statistical information reported |
|---|---|---|
| Trough Forced Expiratory Volume In One Second (FEV1) After 26 Weeks of Treatment | 23 hours 15 minutes and 23 hour 45 minute post-dose Week 26 | Spirometry according to internationally accepted standards; trough FEV1 defined as the mean of the 23 hour 15 minute and 23 hour 45 minute post-dose values. A mixed model was used with treatment as a fixed effect and baseline FEV1, FEV1 prior to inhalation, and FEV1 60 minutes post inhalation of two short acting bronchodilators as covariate information. |
What "trough FEV1" means here
FEV1 is the forced expiratory volume in one second, a spirometric measure of pulmonary function. The registry defines the primary trough measurement specifically as the mean of two post-dose measurements: one obtained 23 hours 15 minutes after dosing and one obtained 23 hours 45 minutes after dosing at Week 26.
This definition is important statistically because the primary endpoint is not simply one spirometry reading. It is a prespecified summary of two closely timed measurements. That reduces dependence on either individual measurement and creates a single continuous outcome for the primary model.
6. Statistical Methodology
Mixed-model analysis
The registry states that a mixed model was used for the primary trough FEV1 endpoint, with treatment as a fixed effect and baseline and specified bronchodilator-response measurements included as covariate information.
The displayed expression is a conceptual representation of the registry-described covariate-adjusted model. The ClinicalTrials.gov record does not provide enough information to reconstruct the complete model specification, including all random-effect, covariance, interaction, or estimation details.
Treatment as a fixed effect
In a mixed-model framework, describing treatment as a fixed effect means that the model estimates the treatment-group differences of interest directly. Treatment is therefore part of the systematic component of the model rather than being treated as a random sample from a broader distribution of treatment effects.
Covariate adjustment
The registry specifically identifies baseline FEV1, FEV1 prior to inhalation, and FEV1 60 minutes after inhalation of two short acting bronchodilators as covariate information. Adjustment for baseline measurements can improve statistical precision when the covariates are predictive of the outcome.
Importantly, adjustment does not change the randomized nature of the treatment comparison. Randomization establishes the primary basis for comparing groups; covariate adjustment can then account for prognostic information and potentially reduce residual variation.
Repeated or closely related measurements
The endpoint definition uses two measurements at different post-dose times and averages them to form trough FEV1. The registry's stated mixed-model approach should be distinguished from simply averaging the observations and then applying an unadjusted comparison: the ClinicalTrials.gov record specifically identifies a treatment fixed effect and covariate adjustment within a mixed model.
7. Results
Primary endpoint: trough FEV1 after 26 weeks
The registry identifies the primary endpoint as trough FEV1 after 26 weeks of treatment, measured at 23 hours 15 minutes and 23 hours 45 minutes post-dose at Week 26. Trough FEV1 was defined as the mean of those two measurements.
| Primary endpoint component | Reported information |
|---|---|
| Endpoint | Trough Forced Expiratory Volume In One Second (FEV1) After 26 Weeks of Treatment |
| Time frame | 23 hours 15 minutes and 23 hour 45 minute post-dose Week 26 |
| Measurement | Mean of the two post-dose FEV1 values |
| Measurement standard | Spirometry according to internationally accepted standards |
| Model | Mixed model |
| Treatment specification | Treatment as a fixed effect |
| Covariate information | Baseline FEV1, FEV1 prior to inhalation, and FEV1 60 minutes post inhalation of two short acting bronchodilators |
| Formal treatment estimate reported | Not provided in the posted results |
| Confidence interval reported | Not provided in the posted results |
| P-value reported | Not provided in the posted results |
The registry tells us exactly what the primary outcome was intended to measure and how it was modeled, but the ClinicalTrials.gov record does not provide the numerical treatment contrast. Therefore, it is not possible from the ClinicalTrials.gov record to state how much QVA149 changed trough FEV1 relative to any comparator, how precise that estimated difference was, or whether a formal hypothesis test met a particular significance threshold.
For a continuous endpoint such as trough FEV1, the natural treatment effect would ordinarily be expressed as an adjusted difference in mean FEV1, or another model-derived contrast, accompanied by a confidence interval and hypothesis-test result.
A confidence interval, when available, would describe uncertainty around the estimated treatment contrast; it would not describe the range of FEV1 values experienced by individual patients. A p-value, when available, would quantify evidence against a specified null hypothesis under the model and assumptions; it would not measure the magnitude or clinical importance of the treatment effect.
Secondary outcomes
The ClinicalTrials.gov record states that 23 outcome measures have been posted, but it does not provide their individual definitions, numerical estimates, confidence intervals, or p-values in the ClinicalTrials.gov record. They are therefore not reproduced or summarized here.
For continuous pulmonary-function outcomes, a model-based comparison such as the mixed-model framework described for the primary endpoint would be a natural analytical approach when supported by the prespecified statistical analysis plan. The appropriate method for a particular secondary endpoint depends on its scale, distribution, timing, and repeated-measure structure.
8. Safety Results
The ClinicalTrials.gov record provides serious adverse-event counts by treatment arm. These are reported as affected patients over patients at risk. Because the denominators differ among the arms, the counts should be interpreted together with their corresponding at-risk populations.
| Treatment arm | Serious adverse events | Affected / at risk |
|---|---|---|
| Indacaterol and Glycopyrronium (QVA149) | Serious adverse events | 22/474 |
| Indacaterol (QAB149) | Serious adverse events | 26/476 |
| Glycopyrronium (NVA237) | Serious adverse events | 29/473 |
| Tiotropium | Serious adverse events | 19/480 |
| Placebo | Serious adverse events | 13/232 |
9. Statistical Methods Explained
Why was a mixed model used for the primary endpoint?
The registry explicitly identifies a mixed model for the primary trough FEV1 endpoint. Mixed models are useful in clinical-trial settings because they can incorporate treatment effects together with covariates and, depending on the complete model specification, account for correlation among observations. The ClinicalTrials.gov record does not provide enough information to reproduce every component of the fitted model.
Why was treatment specified as a fixed effect?
The treatment groups are the prespecified interventions being compared in this randomized trial. Treating treatment as a fixed effect allows the model to estimate the differences associated with those specific treatment assignments. The goal is therefore a direct comparison among the trial's treatment groups rather than estimation of a distribution of hypothetical treatments.
Why define trough FEV1 as the mean of two measurements?
The registry defines trough FEV1 as the mean of the measurements obtained 23 hours 15 minutes and 23 hours 45 minutes after dosing at Week 26. Using the prespecified mean creates one endpoint from two closely timed observations and avoids making either single measurement solely responsible for the primary value.
Why adjust for baseline FEV1?
Baseline FEV1 is a clinically relevant continuous measurement and is explicitly included in the registry-described model. Covariate adjustment can account for baseline prognostic variation and can improve precision when the covariate is related to the outcome.
What is the role of the bronchodilator measurements?
The primary endpoint definition specifically lists FEV1 prior to inhalation and FEV1 60 minutes after inhalation of two short acting bronchodilators as covariate information. These measurements characterize pulmonary-function information available around the reversibility assessment and are incorporated into the described mixed-model analysis.
Why can the safety counts not be interpreted as a treatment comparison by themselves?
Affected/at-risk counts describe how many participants experienced serious adverse events within the registry-reported safety populations. A formal comparison would require a defined estimand and statistical method, along with consideration of denominators, follow-up, multiplicity, and the prespecified safety-analysis framework. The ClinicalTrials.gov record does not provide those formal comparisons.
Why should the missing treatment estimate not be reconstructed?
The primary endpoint is a model-derived continuous outcome. Without the underlying participant-level data, complete model specification, covariance assumptions, and the appropriate analysis population, recomputing an adjusted treatment contrast from enrollment or safety counts would not reproduce the registered analysis.
10. Blinding and Randomization as Statistical Design Features
Randomization
Randomized allocation is the central design mechanism supporting causal comparison among the five treatment arms. It reduces systematic differences in prognostic factors that could otherwise confound the treatment comparison.
Quadruple masking
The registry classifies SHINE as quadruple masked. Masking can reduce the influence of knowledge of treatment assignment on participant behavior, clinical management, outcome assessment, and other trial processes.
These design features are especially relevant for a subjective or behavior-sensitive endpoint. FEV1 is measured objectively by spirometry, but the broader trial still includes safety and tolerability assessments for which knowledge of treatment assignment could influence reporting or assessment.
11. What Can and Cannot Be Concluded From the registry Results
| Question | What the ClinicalTrials.gov record supports |
|---|---|
| Was the trial randomized? | Yes. The allocation is listed as randomized. |
| Was it a parallel-group trial? | Yes. The design model is listed as parallel. |
| Was the trial quadruple masked? | Yes. The masking classification is quadruple. |
| What was the primary endpoint? | Trough FEV1 after 26 weeks of treatment. |
| How was trough FEV1 defined? | The mean of the 23 hour 15 minute and 23 hour 45 minute post-dose Week 26 values. |
| Was a mixed model specified? | Yes. Treatment was a fixed effect, with the listed FEV1 covariates. |
| Is a numerical primary treatment effect reported? | No. |
| Is a confidence interval reported? | No. |
| Is a p-value reported? | No. |
| Are serious adverse-event counts reported? | Yes, by arm as affected/at risk. |
| Can a formal efficacy conclusion be reconstructed from the ClinicalTrials.gov record? | No. The necessary numerical statistical-analysis results are not included in the posted results. |
12. Limitations
- No formal statistical-analysis entries reported: although the registry indicates that results have been posted, no formal statistical analyses were posted. Numerical treatment estimates, confidence intervals, and p-values therefore cannot be reported for the primary endpoint from the ClinicalTrials.gov record.
- Incomplete model specification: the registry excerpt identifies a mixed model and several covariates but does not provide the complete model structure, covariance specification, estimation method, degrees-of-freedom method, or all analysis details needed for exact reproduction.
- Secondary endpoint detail is unavailable: the ClinicalTrials.gov record states that 23 outcome measures are posted but does not provide their individual results or definitions.
- Safety comparisons are limited: serious adverse events are reported as affected/at-risk counts, but no formal comparative analysis is provided.
- No arm-level randomized enrollment is reported: the overall enrollment is 2144.0 and there are five arms, but the ClinicalTrials.gov record does not state the randomized sample size for each arm.
- No additional design assumptions are reported: the ClinicalTrials.gov record does not document a non-inferiority margin, crossover procedure, factorial structure, interim-analysis strategy, multiplicity procedure, missing-data imputation method, stratification factors, or Bayesian methodology. These topics are therefore not assigned characteristics that the ClinicalTrials.gov record does not establish.
13. Why This Trial Matters Statistically
SHINE is a useful teaching example because its primary endpoint illustrates how a clinical trial can combine a precisely defined physiologic measurement with covariate-adjusted model-based inference. The statistical story is not simply "compare two FEV1 means." It involves randomized allocation, masking, a five-arm parallel design, a prespecified definition of trough FEV1, and a mixed-model framework.
| Concept | How it appears in SHINE |
|---|---|
| Randomization | The allocation is randomized across five treatment arms. |
| Parallel design | The registry identifies the design model as parallel. |
| Blinding | The masking classification is quadruple. |
| Continuous endpoint | The primary outcome is trough FEV1 after 26 weeks. |
| Repeated measurement summary | Trough FEV1 is defined as the mean of two post-dose measurements. |
| Mixed model | The registry identifies a mixed model with treatment as a fixed effect. |
| Covariate adjustment | Baseline FEV1 and specified pre/post-bronchodilator FEV1 measurements are included as covariate information. |
| Multi-arm comparison | QVA149 is evaluated alongside glycopyrronium, indacaterol, tiotropium, and placebo. |
| Safety analysis | Serious adverse-event counts are reported by treatment arm. |
14. How to Think About the Primary Estimand
A useful way to understand the SHINE primary endpoint is to distinguish the outcome from the estimand and the estimator.
Outcome
Trough FEV1 after 26 weeks, defined as the mean of the two specified post-dose measurements.
Comparison
The randomized treatment groups provide the basis for estimating differences associated with the interventions.
Estimator
The registry identifies a mixed model with treatment as a fixed effect and specified FEV1 covariates.
Uncertainty
The ClinicalTrials.gov record does not report the resulting confidence interval or other numerical measure of uncertainty for the primary treatment contrast.
This distinction prevents a common statistical mistake: treating the endpoint definition itself as if it were the treatment effect. "Trough FEV1 after 26 weeks" tells us what was measured. The mixed model tells us how the treatment comparison was estimated. A treatment estimate and its confidence interval would then tell us what numerical difference was estimated and how precise that estimate was. The ClinicalTrials.gov record stop before that final step.
15. Interpreting a Future Model-Based Treatment Estimate
If the complete posted analysis were available, the most informative primary result would be the model-derived treatment contrast for trough FEV1, accompanied by its confidence interval and hypothesis-test result. Interpretation should begin with the direction and magnitude of the adjusted difference rather than the p-value alone.
An adjusted treatment difference above zero for QVA149 versus a specified comparator would indicate higher model-estimated trough FEV1 for QVA149 under the prespecified analysis framework. A difference below zero would indicate lower model-estimated trough FEV1. The magnitude would describe the estimated difference on the FEV1 scale.
The confidence interval would indicate statistical precision around that estimated difference. A narrow interval would provide more precise information than a wide interval, all else equal. The interval would not describe the range of individual patient responses.
A p-value would address the evidence against the relevant null hypothesis under the specified statistical model. It would not by itself indicate whether the observed difference was clinically important, nor would it quantify the size of the treatment effect.
16. Safety Interpretation
The serious adverse-event information illustrates another important distinction between descriptive safety reporting and comparative safety inference.
The ClinicalTrials.gov record identifies 22/474 for QVA149, 26/476 for indacaterol, 29/473 for glycopyrronium, 19/480 for tiotropium, and 13/232 for placebo. These figures establish the reported affected/at-risk counts. They do not, on their own, establish a statistically significant difference between any pair of treatment groups.
17. Sources
- ClinicalTrials.gov: NCT01202188 — SHINE.
- Linked publication: PubMed record for PMID 25609942.
Continue through the Clinical Biostats statistical learning pathway
Use the trial's design features to deepen your understanding of randomized and masked clinical-trial methodology.
18. Related Tutorials
Learn more about the methods used in this trial:
19. Related Calculators
20. Record Summary
SHINE is a completed randomized, parallel-group, quadruple-masked phase 3 trial in 2144 participants with moderate to severe COPD. It evaluates five interventions: indacaterol and glycopyrronium (QVA149), glycopyrronium (NVA237), indacaterol (QAB149), tiotropium, and placebo. Its registered primary endpoint is trough FEV1 after 26 weeks of treatment, measured at 23 hours 15 minutes and 23 hours 45 minutes post-dose at Week 26 and defined as the mean of those two values.
Statistically, the most important feature of the ClinicalTrials.gov record is the use of a mixed model with treatment as a fixed effect and specified FEV1 measurements as covariate information. The registry also reports that results have been posted and lists 23 outcome measures. However, the ClinicalTrials.gov record contains no formal analysis entries. Consequently, the appropriate conclusion from the registry record is methodological rather than numerical: the trial's primary endpoint and analysis framework can be described precisely, but a treatment-effect estimate, confidence interval, or p-value cannot be responsibly from the provided record.
The ClinicalTrials.gov record provides serious adverse-event counts by arm, reported as affected/at risk. These figures are useful descriptive information, but they do not constitute a formal comparative safety analysis. This distinction between what the registry defines, what the statistical model is designed to estimate, and what numerical results are actually available is central to rigorous interpretation of clinical-trial evidence.