Randomized, Double-blind, Placebo-controlled, Multicenter, Cross-over Study to Assess the Effects of a 3 Week Therapy Each With QVA149 Versus Placebo on Pulmonary Function and Average Physical Activity Levels in Patients With COPD.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Randomized, Double-blind, Placebo-controlled, Multicenter, Cross-over Study to Assess the Effects of a 3 Week Therapy Each With QVA149 Versus Placebo on Pulmonary Function and Average Physical Activity Levels in Patients With COPD. (NCT01996319), a randomized phase 3 study. The primary outcome is Change From Baseline in Peak Inspiratory Capacity (IC) Comparison Between QVA149 and Placebo.

2. Clinical Question

The central comparison is within a randomized cross-over study rather than a conventional parallel-group trial in which each participant receives only one assigned treatment. Each sequence included both treatments in turn. The registry describes the primary measurement at day 22 or day 57, with the corresponding baseline measured at day 1 or day 36. This makes the study question a comparison of response under the two treatment periods, while the interpretation of the statistical estimate depends on the analysis method used for the repeated, within-participant measurements.

The registry reports 194 participants enrolled. The primary endpoint summaries list 189 participants for QVA149 and 184 for placebo. These endpoint counts are not the same as the enrollment count, and the registry’s definition of the Full Analysis Set additionally requires receipt of at least one dose and availability of the required measurements. The reasons individual participants were not represented in each endpoint summary are not reported in the registry record.

3. Trial Design

This was a randomized, double-blind, placebo-controlled, multicenter, phase 3 cross-over trial sponsored by Novartis Pharmaceuticals. The registry record gives the trial identifier NCT01996319 and the study name MOVE. It lists COPD as the condition and describes the target population as patients with moderate to severe COPD. The trial status is completed, with start listed as 2014-04 and primary completion listed as 2015-02.

The two treatment sequences were QVA149 followed by placebo, and placebo followed by QVA149. The first treatment period lasted 21 days; the second lasted up to 21 days in the first sequence and 21 days in the second sequence, as described in the record. Both treatments were taken once daily. Randomization and double masking are recorded, but the method used to generate the allocation sequence, the concealment procedures, and who specifically remained masked are not reported in the registry record.

For the primary outcome, the record describes spirometric IC measurements and the mean of 3 acceptable measurements. Measurements were collected at the study periods’ baseline and follow-up visits: baseline and day 22, then baseline day 36 and day 57. The change for each period was defined by subtracting its corresponding baseline from the follow-up value. The primary outcome is described as a least-squares mean with a 95% confidence interval.

The primary analysis population was the Full Analysis Set, defined as randomized patients who applied at least one dose during at least one period. The description says that patients were included in the analyses when the required baseline and day 22 data, and baseline day 36 and day 57 data, were available. The registry reports 194 enrolled participants, while the primary endpoint summaries include 189 for QVA149 and 184 for placebo. The trial’s original sample-size rationale, treatment-period washout details, and protocol-specific handling of period or sequence effects are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full Analysis Set (FAS): all randomized patients who applied at least one dose of study medication during at least one study period. all patients were included in these analyses when baseline and day 22 data plus baseline day 36 and day 57 data were available.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Peak Inspiratory Capacity (IC) Comparison Between QVA149 and Placebo, measured over Baseline, day 22, baseline day 36, day 57.

The primary endpoint was change from baseline in peak inspiratory capacity (IC), comparing QVA149 with placebo. IC is a lung-volume measurement obtained by spirometry. The registry states that measurements followed internationally accepted standards and that the mean of 3 acceptable measurements was reported in liters. For each treatment period, the follow-up result was compared with its corresponding period baseline: day 22 with day 1, or day 57 with day 36.

Change from baseline expresses how the measurement at the end of a treatment period differs from the start of that period. This can account for the participant’s starting measurement when describing the treatment-period change. The primary endpoint is focused on peak IC, while other listed lung-function outcomes include trough IC and FEV1 measures. The record also lists average steps per day and duration of at least moderate activity per day as secondary outcomes.

In general, a prespecified primary endpoint gives a trial a central outcome around which the main comparison is organized. That does not, by itself, establish how the analysis handled every feature of a cross-over design. The registry’s reported primary comparison is an ANCOVA result, and the endpoint description identifies the analysis population and measurement schedule. Details such as the covariates included in the model and the handling of period, sequence, or carryover effects are not reported in the registry record.

The endpoint summaries show a least-squares mean change of 0.379 liters for QVA149 and 0.177 liters for placebo. The sponsor-reported comparison estimates a difference of 0.202 liters, with a 95% confidence interval from 0.158 to 0.246. These values describe the posted primary analysis; they do not alone establish the clinical importance of the change, for which the registry record does not provide a threshold.

7. Results Posted in the Registry

• QVA149: mean 0.38 versus 0.18 on control; difference 0.20 (95% CI 0.15 to 0.25), standardized effect d = 0.81; z-test p = <0.001.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Change in the Comparison of QVA149 vs. Placebo on the Average Number of Steps Per Day

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • QVA149: mean 30.70 versus -320.70 on control; difference 351.40 (95% CI 5.11 to 697.69), standardized effect d = 0.21; z-test p = 0.047.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for QVA149 than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Change in the Duration of at Least Moderate Activity Per Day Comparison of QVA149 Versus Placebo

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • QVA149: mean -5.50 versus -13.00 on control; difference 7.50 (95% CI -3.39 to 18.39), standardized effect d = 0.14; z-test p = 0.177.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between QVA149 and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in Peak IC Comparison Between QVA149 and Placebo on Day 1.

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • QVA149: mean 0.49 versus 0.21 on control; difference 0.28 (95% CI 0.23 to 0.33), standardized effect d = 1.13; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for QVA149 than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes cover physical activity and additional measures of lung function. Across the numerical secondary results, the unadjusted estimates are not uniform in their precision: the average-steps comparison has a positive interval that does not include zero, while the duration-of-moderate-activity interval includes zero. The additional peak IC comparison has a positive estimate with an interval above zero. These results complement the primary endpoint by addressing related but distinct measures; they are not interchangeable with the sponsor-reported primary ANCOVA.

The registry lists other secondary outcomes, including trough IC and FEV1 comparisons, but numerical results for those outcomes are not reported in the registry record described in the posted data. The primary result is therefore the clearest sponsor-reported analysis, while the numerical secondary comparisons are simple unadjusted recalculations. The secondary results should not be treated as a single coherent confirmatory test of treatment effect.

In general, examining multiple outcomes creates more opportunities for nominally notable results to occur, even when no adjustment is made for the number of comparisons. A nominal p-value is calculated for an individual comparison and does not account for the wider set of outcomes. The registry record does not report a multiplicity procedure, a hierarchy for secondary outcomes, or sponsor analyses for the posted secondary comparisons. That makes the secondary findings more exploratory in interpretation and gives them less weight than the designated primary analysis.

The measures also differ in scale and variability. Lung-function changes are measured in liters, steps per day in steps/day, and activity duration in minutes. A statistical difference on one measure does not establish a difference on another, nor does it by itself show that a difference is meaningful to patients. The record does not provide clinical-importance thresholds for these secondary measures.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Change in the Comparison of QVA149 vs. Placebo on the Average Number of Steps Per Day
• Change in the Duration of at Least Moderate Activity Per Day Comparison of QVA149 Versus Placebo
• Change From Baseline in Peak IC Comparison Between QVA149 and Placebo on Day 1.
• Change From Baseline in the Trough IC Comparison Between QVA149 and Placebo
• Peak Forced Expiratory Volume 1 (FEV1) Comparison Between QVA149 and Placebo at Day 1
• Trough FEV1 Comparison Between QVA149 and Placebo After 22 Days

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous: change in peak IC is measured in liters, and the reported comparison is a difference in means. The sponsor’s posted analysis is an ANCOVA, with a reported estimate of 0.202 liters and a 95% confidence interval from 0.158 to 0.246. The record identifies the method as ANCOVA but does not report the complete model specification, the covariates, or the handling of cross-over period and sequence effects.

What does the mean difference describe?

In general, a mean difference is the average outcome in one treatment condition minus the average outcome in the comparator condition, according to the stated analysis. A positive difference in the primary comparison indicates that the estimated change was larger under QVA149 than under placebo on the outcome’s scale. Its size is expressed in liters, so the estimate remains tied to the IC measurement rather than being a standardized quantity. The interpretation of the estimate depends on the trial’s design and analysis, not only on the arithmetic contrast.

The posted primary arm summaries are least-squares means. In general, a least-squares mean is a model-based estimate, often adjusted for terms in the model, and is not necessarily identical to a raw arithmetic mean. The registry does not list the adjustment terms for this ANCOVA. The reported arm values are 0.379 liters for QVA149 and 0.177 liters for placebo, while the sponsor’s comparison is 0.202 liters. Since the comparison is reported as ANCOVA, its inference should be understood as model-based rather than as a simple subtraction of two independently observed group averages.

Why is a simple recalculation different from the trial analysis?

The unadjusted recalculation uses the posted endpoint summaries to estimate a mean difference and its uncertainty. For the primary outcome, it reports a difference of 0.202 liters, a 95% confidence interval from 0.151 to 0.253, and p < 0.001. The recalculation uses an approximate pooled standard deviation and does not reproduce the sponsor’s ANCOVA model. Although its point estimate matches the sponsor’s reported estimate, its confidence interval is not identical. Different model structures and standard-error calculations can yield different uncertainty estimates even when the point estimate is the same.

In general, a pooled standard deviation combines variability information across groups under simplifying assumptions. The registry marks the primary recalculation’s pooled standard deviation as approximate. In a cross-over trial, measurements from the same participant under the two treatments are related. A reconstruction from separate arm-level summaries cannot recover the within-participant correlation or all the information in paired data. It should therefore be treated as a descriptive, unadjusted recalculation, not as a replacement for the reported analysis.

How should the confidence interval and p-value be read?

In general, a confidence interval gives a range of values compatible with the data and the method used, at the stated confidence level. A narrower interval indicates greater precision under that analysis, while a wider interval indicates more uncertainty. It is not a guarantee that the true effect lies within the particular interval. The primary sponsor interval runs from 0.158 to 0.246 liters. Its endpoints are both positive, indicating that zero is outside the interval under the reported model.

In general, a p-value measures how unusual data at least as discrepant as those observed would be if the null hypothesis and analysis assumptions were true. It does not measure the probability that the null hypothesis is true, the probability that the result occurred by chance, or the size or importance of an effect. The sponsor reports p < 0.001 for the primary ANCOVA comparison. The recalculated p-value is also reported as less than a threshold, but its calculation is unadjusted and based on summary statistics.

What does over-dispersion mean for activity outcomes?

In general, over-dispersion means that observed variability is larger than a model would expect under a simpler distributional assumption. It can make estimates less precise and can affect standard errors and interval estimates if not accounted for. The steps-per-day summaries have standard deviations of 1662.0 and 1648.0, compared with mean changes of 30.7 and -320.7. Those figures indicate substantial spread around the reported means, but the registry does not identify a specific over-dispersion model for this outcome. The recalculated interval is wide relative to its point estimate, so the estimated difference is imprecise.

For duration of at least moderate activity, the recalculated mean difference is 7.5 minutes, and the 95% confidence interval runs from -3.392 to 18.39. In general, an interval spanning zero indicates that the direction of the difference is uncertain under the calculation used. The registry does not report a sponsor analysis for this secondary outcome. The same caution applies to the other unadjusted results: they are not documented as model-adjusted trial analyses.

The record reports 8 outcome measures and 2 analyses overall, and lists six secondary outcomes. It contains a sponsor-reported ANCOVA for the primary endpoint. The secondary outcomes are not accompanied by sponsor-reported analyses in the posted data described for those results. The registry does not report the trial’s statistical analysis plan, the original sample-size justification, interim analyses, or multiplicity procedures. These omissions limit what can be inferred about whether the reported tests were prespecified, how they were prioritized, or how the broader set of outcomes was intended to be interpreted.

11. Confidence Intervals and Precision

The primary estimate is 0.202 liters, and the sponsor-reported 95% confidence interval extends from 0.158 to 0.246 liters. This interval communicates the precision of the ANCOVA estimate under the model used. Its endpoints are relatively close to the estimate, and both are positive. The record’s reported p-value is p < 0.001. These statistics indicate a comparatively precise positive estimate within the stated analysis, but do not establish a threshold for clinical importance.

The summary-based recalculation gives the same point estimate, 0.202 liters, with a 95% confidence interval from 0.151 to 0.253. Its limits are wider than those in the reported ANCOVA result. This difference is consistent with the fact that the calculations use different approaches: the registry’s recalculation is unadjusted and uses an approximate pooled standard deviation, while the sponsor result is from ANCOVA. The recalculation cannot capture all the information or assumptions in the trial’s model.

In general, a confidence interval describes uncertainty around an estimate under a particular statistical method and its assumptions. It is not an interval containing a fixed proportion of individual participant responses, and it does not show the full range of possible outcomes. It also does not establish that a difference is important to patients. The registry record does not report a clinically important threshold for peak IC, steps per day, or activity duration.

Precision varies across the secondary estimates. The interval for steps per day extends from 5.107 to 697.7 steps/day, while the interval for activity duration extends from -3.392 to 18.39 minutes. The latter includes zero; the former does not, but spans a broad range of positive values. These intervals show why a point estimate alone can be misleading: plausible values under the calculation cover a range of magnitudes, and for the activity-duration comparison include both a small negative and a positive difference. Secondary intervals are unadjusted, and the record does not report a multiplicity correction.

In general, uncertainty in a cross-over study depends on the paired nature of the observations as well as on the variation in each treatment condition. The summary-based recalculations do not contain all participant-level pairing information. The sponsor ANCOVA interval is therefore the most direct registry-reported precision estimate for the primary endpoint.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The sample-size figures are planning illustrations based on the observed effect, not the original enrollment plan. For these continuous outcomes, the listed figures are participants per arm, not total events. The illustrations assume a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The effect to detect is the stated fraction of the observed effect. The trial’s own original sample-size assumptions and justification are not reported in the registry record.

The 80% power column gives the number of participants per arm in the illustration for a design with an 80% chance of detecting the specified effect under the stated assumptions, if that effect and assumptions were correct. The 90% power column gives the corresponding number per arm for a 90% chance. These are hypothetical planning quantities; they are not a statement about the probability that the completed trial found a true effect. This is not observed power.

Each set of rows changes the effect size used for the illustration. Planning for 100% of the observed effect uses the full observed estimate; planning for 75% or 50% uses progressively smaller effects. Smaller effects require more participants to achieve the same target power under the same assumptions. This is why the participant counts increase as the assumed fraction of the observed effect decreases.

The displayed planning exercise uses a simplified equal-allocation, two-arm framework. The clinical trial itself was a cross-over study in which participants received both QVA149 and placebo in different sequences. The illustration should not be mistaken for a reconstruction of the trial’s original design or its actual analysis. Its purpose is to show how assumed effect size and target power affect a hypothetical participant count.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 0.202; pooled standard deviation σ = 0.250.
2. n = 2 × (1.96 + 0.842)² × 0.250² / 0.202² = 24.0, rounded up to 25 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- 0.2021                      # observed difference in means
sigma <- 0.250026938371953                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(1996319)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 25 (80% power), 33 (90% power)
 75% of observed effect: n per arm = 43 (80% power), 58 (90% power)
 50% of observed effect: n per arm = 97 (80% power), 129 (90% power)
power.t.test: 25 
Simulated power at n = 25 per arm: 0.805 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

The sponsor’s reported primary analysis is an ANCOVA. It estimates a QVA149-versus-placebo difference of 0.202 liters in change from baseline in peak IC, with a 95% confidence interval from 0.158 to 0.246 and p < 0.001. This is the trial’s posted model-based result and should be distinguished from the simple recalculation based on endpoint summaries.

The unadjusted recalculation estimates the same difference, 0.202 liters, but gives a 95% confidence interval from 0.151 to 0.253 and p < 0.001. The point estimates agree, while the uncertainty estimates are not identical. The recalculation uses an approximate pooled standard deviation and does not reproduce the ANCOVA. It cannot account for unreported model details or the full cross-over data structure.

The registry does not report the ANCOVA’s covariates or full specification, so the difference between the two intervals cannot be attributed to a specific model term. The recalculation is useful as a check on the scale and direction of the primary estimate, but it is not an alternative sponsor analysis. For the primary result, the sponsor’s posted ANCOVA is the relevant reported analysis.

15. Multiplicity, Missing Data and Analysis Populations

The registry identifies change in peak IC as the primary outcome and lists six secondary outcomes. The record reports an ANCOVA analysis for the primary comparison. It does not report how the trial controlled the overall chance of false-positive findings across the primary and secondary outcomes, whether a hierarchy of testing was used, or whether any secondary results were adjusted for multiple comparisons. The numerical secondary comparisons described in the posted results are unadjusted recalculations, not sponsor-reported analyses.

In general, multiplicity arises when several hypotheses or outcomes are tested. If each is assessed separately without adjustment, the collection of tests can produce a nominally significant result even when the underlying effects are absent. Adjustment procedures or prespecified testing hierarchies can address this, but the appropriate approach depends on the analysis plan. Since the plan and procedure are not reported in the registry record, the nominal secondary p-values should be interpreted as individual, unadjusted comparisons rather than as family-wise confirmatory evidence.

The primary endpoint’s p-value and confidence interval come from the reported ANCOVA, but the complete model specification is not reported. The registry does not state whether the model included period or sequence terms, how it handled potential carryover, or how missing repeated measurements were treated. The trial’s interim-analysis plan and any stopping rules are also not reported in the registry record. These are unknowns rather than evidence that such methods were or were not used.

The Full Analysis Set is defined as randomized patients who received at least one dose during at least one study period, with the relevant baseline and follow-up measurements available for inclusion. The primary endpoint summaries include 189 participants under QVA149 and 184 under placebo. The registry does not explain the differences between those endpoint counts and total enrollment of 194, or provide a detailed account of missingness by reason, treatment sequence, or period. It therefore is not possible from this record to assess whether missing measurements were related to participant characteristics or outcomes.

In general, missing-data methods can affect both estimates and uncertainty, especially when repeated measurements are required. The record’s population definition gives some information about who contributed to the primary analysis, but the specific imputation or model-based handling of incomplete records is not reported in the registry record. Consequently, the results should be read as applying to the analyzed measurements and population as described, without assuming a particular missing-data strategy.

16. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 1% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: incisional hernia (0.5%), oesophageal carcinoma (0.5%), myocardial infarction (0.5%), cellulitis (0.5%), bone cancer (0.5%).

The registry reports serious adverse events by treatment sequence. In the QVA149-then-placebo sequence, 4 of 193 participants were recorded with a serious adverse event; in the placebo-then-QVA149 sequence, 2 of 188 were recorded with a serious adverse event. The record also lists other adverse events for 27 of 193 participants in the first sequence and 19 of 188 in the second. These are sequence-level counts, not separate event counts for each treatment period, so they do not isolate events occurring under QVA149 from those occurring under placebo.

The registry lists five serious-event terms, each shown as 0.5%: incisional hernia, oesophageal carcinoma, myocardial infarction, cellulitis, and bone cancer. The record does not provide enough detail in these figures to determine treatment-period timing or attribute an event to a particular treatment. Death counts are not reported in the registry record.

In general, adverse-event summaries describe what was recorded during a study; they do not, by themselves, establish causation. These figures should be read in light of the sequence-level reporting, the study’s duration, and the absence of treatment-specific event attribution in the record. They do not support a broader comparison of safety between the two treatments.

17. What the Result Does, and Does Not, Mean

The posted primary analysis estimates a positive difference in change in peak IC between QVA149 and placebo. The sponsor-reported ANCOVA estimate is 0.202 liters, with a 95% confidence interval from 0.158 to 0.246 and p < 0.001. The result describes the measured primary outcome in the analyzed population and under the reported model. It does not, on its own, establish how important that difference is in clinical terms.

The record does not report a threshold for a clinically important change in IC. Statistical evidence about whether an estimate differs from zero and clinical interpretation of its magnitude are separate questions. Nor does a confidence interval describe the range of individual participant responses. The estimate and interval summarize uncertainty about an average comparison under the analysis method.

The trial was cross-over, with participants receiving QVA149 and placebo in different sequences. That design differs from a parallel-group trial, and interpretation depends in part on how measurements across periods are modeled. The registry reports the sponsor’s ANCOVA result but not the full model specification or details of the handling of period, sequence, or potential carryover effects. The summary-based recalculation does not reproduce that model.

The secondary comparisons relate to physical activity and other lung-function measures, but they are not a substitute for the primary result. Their unadjusted status, the absence of a reported multiplicity procedure, and varying precision mean that they should be interpreted cautiously. The record does not provide clinical-importance thresholds for these outcomes, and numerical results for some listed secondary endpoints are not reported in the registry record.

The registry’s safety counts are reported by treatment sequence rather than by individual treatment period, and the record does not report death counts. Those figures therefore do not establish treatment-specific risks. The study results describe the participants, measurements, and analyses recorded; they do not answer every question about clinical benefit, long-term outcomes, or comparative safety.

18. Why This Trial Matters Statistically

This trial is statistically notable because it combines a cross-over design with a continuous primary endpoint and a sponsor-reported adjusted analysis. The same participants were assigned to sequences that exposed them to both QVA149 and placebo. That design makes within-participant comparisons relevant, while also making the analysis dependent on how repeated measurements and treatment periods are handled.

The primary comparison illustrates why it is useful to distinguish a model-based trial analysis from a recalculation using aggregate summaries. Both produce a point estimate of 0.202 liters, but their confidence intervals differ. The sponsor’s ANCOVA is the posted primary analysis; the unadjusted reconstruction uses an approximate pooled standard deviation and cannot recover all the participant-level pairing information. Understanding the methods behind a number is as important as reading its size.

The secondary outcomes also show why statistical evidence should be considered alongside precision and multiplicity. Different measures have different scales and variability, and their confidence intervals do not convey the same degree of certainty. Since the registry does not report a multiplicity procedure, secondary p-values should not be treated as though each arose from an isolated, prespecified confirmatory test.

Finally, the registry record makes some features clear, including the primary endpoint, analysis population, and sponsor-reported ANCOVA estimate, while leaving other features unreported, including the original sample-size rationale and full model details. This distinction helps readers separate what the trial reports from what cannot be established from the record. It is a practical example of reading estimates, uncertainty, design, and reporting completeness together.

19. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record does not report the original sample-size justification, so the enrollment cannot be evaluated against the trial’s intended effect size or target power from this information. The planning illustrations based on observed effects are not a substitute for the protocol’s original assumptions. They use a simplified equal-allocation, two-arm framework, whereas the study itself used a cross-over design.

The primary analysis is identified as ANCOVA, but its full specification is not reported. The record does not identify the covariates or explain whether and how period, sequence, or potential carryover effects were handled. This limits independent assessment of the model assumptions and of the primary confidence interval. The summary-based recalculation is unadjusted, uses an approximate pooled standard deviation, and does not reproduce the participant-level cross-over analysis.

The primary endpoint summaries include different counts for QVA149 and placebo, and the registry does not explain the reasons for the difference from enrollment or between endpoint summaries. The Full Analysis Set definition requires relevant measurements to be available, but detailed missing-data reasons and methods are not reported in the registry record. The potential impact of incomplete measurements therefore cannot be assessed from these registry entries alone.

For the secondary comparisons, the posted numerical estimates are unadjusted recalculations, and the registry does not report a multiplicity method or a hierarchy for interpreting the outcomes. Numerical results for some listed secondary endpoints are not reported in the registry record. The record also does not provide clinical-importance thresholds for the primary or secondary measures, limiting interpretation of the magnitude of estimated differences.

Safety information is summarized by treatment sequence rather than treatment period, so the listed counts do not establish which treatment was being taken when each event occurred. Death counts are not reported in the registry record. The reported figures should not be used to make treatment-specific causal comparisons.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Randomized, Double-blind, Placebo-controlled, Multicenter, Cross-over Study to Assess the Effects of a 3 Week Therapy Each With QVA149 Versus Placebo on Pulmonary Function and Average Physical Activity Levels in Patients With COPD. — NCT01996319.
• Registry record last updated 2016-04-05; status last verified 2016-03.
• Results first posted 2016-04-05.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

23. Record Summary

NCT01996319 is a completed randomized, double-blind, placebo-controlled, multicenter, phase 3 cross-over trial sponsored by Novartis Pharmaceuticals. It enrolled 194 participants with moderate to severe COPD and compared once-daily QVA149 with placebo in two treatment sequences.

The primary outcome was change from baseline in peak IC, measured in liters. The sponsor-reported ANCOVA estimate was 0.202 liters, with a 95% confidence interval from 0.158 to 0.246 and p < 0.001. The registry also contains unadjusted summary-based comparisons for selected secondary outcomes, with no reported multiplicity procedure.

The record defines the Full Analysis Set and describes the primary measurement schedule, but does not report the full ANCOVA specification, the original sample-size rationale, or detailed missing-data methods. Safety counts are reported by treatment sequence rather than by treatment period.