Roflumilast Plus Alogliptin Proof-of-Mechanism Study in Type2 Diabetes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Roflumilast Plus Alogliptin Proof-of-Mechanism Study in Type2 Diabetes (NCT01664624), a randomized phase 1 study. The primary outcome is Change From Baseline in Postprandial Area Under the Curve From Time 0 to 8 Hours (AUC[0-8]) for Active Glucagon-like Peptide-1.

2. Clinical Question

3. Trial Design

This was a randomized, parallel-group Phase 1b study with four treatment groups. The registry reports 40 enrolled participants and a completed status. The listed study dates are 2012-07 for study start and 2012-11 for primary completion. Participants were described as having type 2 diabetes inadequately controlled on a stable dose of metformin.

The four groups received roflumilast plus alogliptin, alogliptin alone, roflumilast alone, or exenatide. The combination group received roflumilast 500 μg and alogliptin 25 mg orally once a day for 11 days. The single-medicine groups received the specified active medicine and a placebo for the other medicine, also once a day for 11 days. The exenatide group received 5 μg by subcutaneous injection twice a day for 11 days. These treatment descriptions establish the short duration of exposure in this study; they do not describe longer-term treatment.

The official title describes the study as randomized, double-blind, and controlled with an open-label exenatide comparator. The registry’s masking classification is quadruple. The exact roles covered by that classification, and how the open-label comparator was handled in masking procedures, are not reported in the registry record. Allocation concealment and other operational details are also not reported in the registry record.

The primary endpoint analysis used an analysis of covariance (ANCOVA) model, with treatment as a fixed effect and baseline postprandial active GLP-1 AUC as a continuous covariate. The endpoint record specifies a full analysis set defined as all randomized participants included in the safety analysis, while also requiring baseline and post-baseline data for inclusion in the endpoint analysis. The primary results have arm counts of 10, 10, 10, and 9. The reasons for those counts, and the original sample-size justification or design assumptions, are not reported in the registry record.

The posted results include three sponsor-reported primary-endpoint contrasts involving the combination group. The registry also lists secondary outcomes concerning post-meal glucose, C-peptide, insulin, appetite sensation, and 24-hour average plasma glucose. The record does not report a detailed protocol, an interim-analysis plan, or a prespecified multiplicity strategy.

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.

Before assignment: Participants with a diagnosis of type 2 diabetes mellitus and inadequate glycemic control were enrolled equally in 1 of 4 treatment groups, roflumilast + alogliptin, alogliptin alone, roflumilast alone, or exenatide.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full analysis set defined as all randomized participants included in the safety analysis. Only participants with data at both Baseline and post-baseline visits are included.

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 Postprandial Area Under the Curve From Time 0 to 8 Hours (AUC[0-8]) for Active Glucagon-like Peptide-1, measured over Baseline and Day 11; samples were taken at -15 min and -5 min (pre-meal), and 15 min, 30 min, and 1, 2, 3, 4, 6, and 8 hours (post-meal)..

The primary endpoint measures change from baseline to Day 11 in active glucagon-like peptide-1 (GLP-1) after eating. Blood concentrations were measured before the meal and at post-meal time points through 8 hours; the area under the concentration-time curve (AUC) was calculated using the linear trapezoidal rule. The unit is pmol/L*hr. AUC combines the measured concentrations across the observation period into one summary of the post-meal response, rather than treating each blood draw as a separate primary result.

GLP-1 is the biological marker named in the study’s primary objective. The endpoint was used to compare the change in this postprandial marker between treatment groups. The analysis describes least-squares means, which are model-based group estimates adjusted for the baseline AUC covariate. The sponsor’s ANCOVA therefore addresses differences in change while accounting for the measured baseline value in the model.

The endpoint is not itself a direct summary of every aspect of glucose control or of longer-term clinical outcomes. The registry separately lists postprandial plasma glucose and 24-hour average plasma glucose among secondary outcomes. A change in the GLP-1 AUC alone does not establish what happened on those other measures, nor does it answer questions beyond the short study period. The meaning of the endpoint is bounded by the reported meal-related measurement window, the Day 11 assessment, the analysis population, and the uncertainty in the between-group estimates.

7. Results Posted in the Registry

• Alogliptin Alone: mean 31.60 versus 26.50 on control; difference 5.10 (95% CI -6.06 to 16.26), standardized effect d = 0.40; z-test p = 0.370.
• Roflumilast Alone: mean 3.80 versus 26.50 on control; difference -22.70 (95% CI -33.76 to -11.64), standardized effect d = -1.80; 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 From Baseline in AUC(0-8) of Postprandial Plasma Glucose

• 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 errors.
• Result:
  • Alogliptin Alone: mean -9.80 versus -13.40 on control; difference 3.60 (95% CI -3.90 to 11.10), standardized effect d = 0.42; z-test p = 0.347.
  • Roflumilast Alone: mean -9.00 versus -13.40 on control; difference 4.40 (95% CI -3.10 to 11.90), standardized effect d = 0.51; z-test p = 0.250.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Alogliptin Alone and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Roflumilast Alone and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in Postprandial AUC(0-8) of C-peptide

• 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 errors.
• Result:
  • Alogliptin Alone: mean 1.40 versus 1.00 on control; difference 0.40 (95% CI -2.80 to 3.60), standardized effect d = 0.11; z-test p = 0.807.
  • Roflumilast Alone: mean -0.70 versus 1.00 on control; difference -1.70 (95% CI -4.92 to 1.52), standardized effect d = -0.46; z-test p = 0.300.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Alogliptin Alone and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Roflumilast Alone and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in Postprandial AUC(0-8) of Insulin

• 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 errors.
• Result:
  • Alogliptin Alone: mean 107.80 versus -49.50 on control; difference 157.30 (95% CI -344.18 to 658.78), standardized effect d = 0.27; z-test p = 0.539.
  • Roflumilast Alone: mean 26.90 versus -49.50 on control; difference 76.40 (95% CI -424.56 to 577.36), standardized effect d = 0.13; z-test p = 0.765.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Alogliptin Alone and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Roflumilast Alone and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists secondary outcomes covering postprandial plasma glucose, C-peptide, insulin, appetite sensation, and 24-hour average plasma glucose. Posted numerical results are available for postprandial plasma glucose, C-peptide, and insulin. Numerical results for appetite sensation and 24-hour average plasma glucose are not reported in the registry record, despite their listing among the secondary outcomes.

Across the posted secondary outcomes, the combination group’s estimated changes do not show one uniform pattern relative to the single-medicine groups. The simple recalculated contrasts for the posted glucose, C-peptide, and insulin outcomes have intervals that include zero. These secondary calculations are unadjusted comparisons from summary statistics; they are not sponsor-reported ANCOVA results. The registry’s primary endpoint is active GLP-1 AUC, so the secondary measures provide related but distinct information about other biological and appetite measures. They should not be substituted for the primary result.

Secondary outcomes can help describe the range of measurements collected, but their interpretation is more tentative when many outcomes or contrasts are considered. If a family of tests is examined without adjustment, the chance of at least one apparently unusual result can increase. The registry record does not report a multiplicity adjustment or a complete prespecified hierarchy for secondary testing. Therefore, unadjusted secondary comparisons carry less weight as confirmatory evidence than a prespecified primary analysis.

The registry describes six outcome measures and three analyses, but the detailed reporting is incomplete for some listed secondary endpoints. The missing numerical results, the absence of a reported multiplicity strategy, and the small groups constrain any broad synthesis across outcomes. No claim about a consistent effect across all secondary measures follows from the posted results. In particular, the available secondary results do not resolve the imprecision in the primary comparison of the combination with alogliptin alone.

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

• Change From Baseline in AUC(0-8) of Postprandial Plasma Glucose
• Change From Baseline in Postprandial AUC(0-8) of C-peptide
• Change From Baseline in Postprandial AUC(0-8) of Insulin
• Change From Baseline to Day 11 in AUC(0-8) of Appetite Sensation
• Change From Baseline to Day 11 in 24-hour Average Plasma Glucose

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is a continuous measure: change in postprandial active GLP-1 AUC from baseline to Day 11. A continuous outcome can take values across a measurement scale, rather than being limited to categories such as event or no event. The registry reports least-squares means and standard errors for the groups. Least-squares means are adjusted estimates from a statistical model; they are not necessarily identical to unadjusted averages calculated directly from the observed values.

How was the endpoint constructed?

GLP-1 concentrations were measured before and after a meal, with observations extending through 8 hours after eating. The registry states that AUC was calculated with the linear trapezoidal rule. In general, this method approximates the area under a curve by joining adjacent measured points with straight lines and adding the resulting areas. It provides one summary of the measured concentration profile over the specified time window. The primary outcome is the change in this summary from baseline to Day 11.

The endpoint record describes the analysis as ANCOVA, with treatment as a fixed effect and baseline postprandial AUC as a continuous covariate. In general, ANCOVA estimates treatment-group differences while adjusting for a baseline measurement. Including baseline can account for differences in starting values and can improve precision when baseline is related to the follow-up outcome. The registry names the model components but does not report further details such as diagnostic checks, the precise handling of model assumptions, or any additional covariates.

What does the reported contrast represent?

The sponsor reports least-squares mean differences for treatment comparisons. A difference is expressed in the endpoint’s units, pmol/L*hr, and represents the estimated contrast between the specified groups under the ANCOVA model. Its sign depends on the order in which the groups are contrasted. The registry’s sponsor-reported contrast of the combination with alogliptin alone is -5.1. The separately recalculated unadjusted contrast has the opposite sign because it is based on a simple comparison of posted group summary values. As the registry record flags a discrepancy, the sponsor’s own analysis takes precedence.

A p-value describes how compatible the data are with a specified null model, under the assumptions of the analysis. It is not the probability that the null hypothesis is true, nor does it measure the importance of an effect. A small p-value can indicate that the observed data would be unusual under that null model, while a larger p-value does not demonstrate that groups are the same. Interpretation also depends on the estimate, its confidence interval, design features, and whether the analysis was prespecified.

How should the confidence interval be understood?

A confidence interval expresses uncertainty around an estimated contrast using a method with a stated long-run coverage property. In general, a wider interval means the estimate is less precise, while a narrower interval indicates greater precision under the model. For the sponsor-reported combination versus alogliptin contrast, the 95% confidence interval ranges from -16.9 to 6.7. Because it includes zero, the posted result is compatible with a range of contrasts in both directions. That does not establish that the treatments are equivalent; equivalence requires a defined margin and a design and analysis suited to that question.

The sponsor’s reported ANCOVA estimates and the recalculations from posted summaries answer related but different questions. The recalculations are simple unadjusted comparisons and cannot reconstruct the adjusted model from group means and standard errors alone. The registry record does not report the individual-level observations needed to refit or independently verify the ANCOVA. In general, adjustment, model specification, and the covariance between baseline and outcome can affect an estimate and its uncertainty.

Why does over-dispersion matter in general?

For continuous outcomes, variability describes how spread out observations are around a group estimate. In general, greater variability makes a treatment contrast harder to estimate precisely, all else being equal. Over-dispersion is most often discussed for count outcomes when observed variation exceeds that assumed by a simple count model; the primary endpoint is continuous, so that count-model issue is not the central method. The relevant general point is that the chosen model should represent the observed variation adequately. The registry reports standard errors but does not report detailed model diagnostics or a separate assessment of distributional assumptions.

The sample-size illustrations use observed contrasts and a normal approximation under stated planning assumptions. Such illustrations are conditional calculations, not a substitute for a prospective design justification. The original sample-size assumptions, target effect, and rationale for the enrolled sample are not reported in the registry record. An illustration based on an observed effect can be unstable when the observed groups are small, because the effect estimate itself is uncertain.

The registry identifies a full analysis set tied to the safety analysis and requires baseline and post-baseline data for the endpoint. In general, excluding participants without both measurements can affect the analysis if missingness relates to treatment or outcome. The reasons for missing endpoint data, any imputation approach, and sensitivity analyses are not reported in the registry record. Thus, the statistical interpretation relies on the posted analysis definition and should not assume that missing observations were absent at random or handled in a particular way.

11. Confidence Intervals and Precision

Precision concerns how tightly the data locate a treatment contrast. The sponsor’s primary comparison of roflumilast plus alogliptin with alogliptin alone is -5.1 pmol/L*hr, with a 95% confidence interval from -16.9 to 6.7. The interval spans zero and includes values in both directions. Its width indicates that the estimate does not sharply distinguish a negative contrast from a positive one. The p-value is 0.388, which is consistent with the interval’s inclusion of zero under the reported analysis.

The contrasts with roflumilast alone and exenatide have 95% confidence intervals from 11.1 to 34.3 and from 17.1 to 40.5, respectively. Each interval is entirely positive. These intervals still cover a range of possible effect sizes, rather than identifying a single exact value. The point estimate is not a guarantee of the true contrast, and statistical evidence about these comparisons is distinct from evidence about other outcomes or longer-term effects.

Precision is shaped by the amount of information in the data, including the number of participants and the variability of measurements. The primary analysis group counts are 10, 10, 10, and 9. These small groups provide limited information for estimating between-group differences and variability. The standard errors reported for group estimates describe uncertainty in those estimates, while the confidence intervals for contrasts express uncertainty in the between-group comparisons.

A confidence interval is not a range containing a fixed proportion of individual participants’ outcomes. In general, it is an interval estimate constructed so that, across repeated use of the same method under its assumptions, a stated proportion of such intervals would cover the target parameter. Its practical interpretation depends on the model and design. The registry record does not report diagnostics or sensitivity analyses that would show how robust the interval estimates are to alternative assumptions.

For the comparison with alogliptin alone, the posted interval does not support a precise conclusion about the direction of the adjusted difference. It should not be interpreted as evidence of no difference or equivalence. A meaningful equivalence conclusion would require a prespecified margin and a suitable design; these are not reported in the registry record. The other intervals are more directionally informative for their stated contrasts, but the absence of reported multiplicity handling and the small study size remain important when considering the set of results.

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 illustrations describe participants per arm for 80% power and participants per arm for 90% power. They are not total event counts: this is a continuous-outcome analysis, and the figures refer to participants in each of two equally allocated comparison arms. They are planning illustrations calculated from the observed effect, not the original trial’s sample-size design. The original design assumptions and sample-size justification are not reported in the registry record.

The listed planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. For each comparison, the effect used is a stated fraction of the observed effect. The planning rows describe using 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect. The 80% power column gives the per-arm sample size associated with an 80% chance of detecting the assumed effect under those planning assumptions, if that effect and assumptions held. The 90% power column gives the corresponding per-arm sample size for a 90% chance.

Smaller assumed effects require more participants because they are harder to distinguish from random variation. This pattern is reflected in the larger planning figures for smaller fractions of the observed effect. The illustrations are specific to each estimated comparison and should not be treated as a general sample-size recommendation. Estimates derived from small groups may themselves be uncertain, so planning around an observed effect can yield unstable projections.

These figures are not observed power. They do not tell how much power the completed study actually had in light of its realized data, and they do not replace interpretation of the estimates and confidence intervals. The registry does not report the original target effect, its design-stage assumptions, or a rationale for selecting the enrolled sample size.

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 = 5.100; pooled standard deviation σ = 12.728.
2. n = 2 × (1.96 + 0.842)² × 12.728² / 5.100² = 97.8, rounded up to 98 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 <- 5.1                      # observed difference in means
sigma <- 12.7282559685135                     # 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(1664624)
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 = 98 (80% power), 131 (90% power)
 75% of observed effect: n per arm = 174 (80% power), 233 (90% power)
 50% of observed effect: n per arm = 392 (80% power), 524 (90% power)
power.t.test: 98.7 
Simulated power at n = 98 per arm: 0.796 (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 own primary-endpoint analyses are ANCOVA comparisons reported as least-squares mean differences, with 95% confidence intervals and p-values. For the combination compared with alogliptin alone, the sponsor reports -5.1 pmol/L*hr, a 95% confidence interval from -16.9 to 6.7, and p = 0.388. The sponsor reports 22.7 pmol/L*hr for the combination compared with roflumilast alone, with a 95% confidence interval from 11.1 to 34.3 and p < 0.001; the combination versus exenatide estimate is 28.8 pmol/L*hr, with a 95% confidence interval from 17.1 to 40.5 and p < 0.001.

Recalculated comparisons use posted arm-level summary statistics and are simple, unadjusted calculations. They do not reproduce the sponsor’s baseline-adjusted ANCOVA. For the combination versus alogliptin alone, the recalculated mean difference has the opposite sign from the sponsor’s least-squares mean difference. The registry record flags a discrepancy, so the sponsor’s own analysis takes precedence. The difference illustrates why model-adjusted results should not be treated as interchangeable with simple calculations from group summaries.

The registry record does not provide participant-level data or enough model detail to independently refit the ANCOVA. The recalculations can be useful as a transparent check of the posted summaries, but they are not a replacement for the sponsor’s specified analysis. The interpretation should therefore center on the sponsor-reported adjusted estimates, their intervals, and the limits created by the small analysis groups and unreported multiplicity strategy.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when several outcomes, treatment comparisons, time points, or analyses are examined. In general, conducting many tests can increase the chance that at least one result appears statistically notable by chance, even when the underlying effects are absent. Approaches such as a prespecified testing hierarchy or statistical adjustment can address this issue, but the registry record does not report a multiplicity strategy for the primary contrasts or secondary outcomes.

The primary endpoint has three sponsor-reported comparisons involving the combination group. The record also lists several secondary outcomes, some with posted numerical results and some without. Without a reported plan for controlling multiplicity or defining which results are confirmatory, the individual p-values should be read as results for their stated comparisons rather than as a fully controlled family of claims. The primary outcome remains the central endpoint, but the several comparisons still matter when judging the strength of the overall evidence.

The endpoint population is described as all randomized participants included in the safety analysis, with only participants having both baseline and post-baseline data included. This clarifies the stated rule for inclusion in the primary endpoint analysis, but does not explain why some participants may lack data. Reasons for missing endpoint measurements, methods for handling missing data, and sensitivity analyses are not reported in the registry record. The registry lists 40 enrolled participants, while primary endpoint group counts are 10, 10, 10, and 9. The reason for that difference is not reported in the registry record.

In general, complete-case analysis can be vulnerable to bias if the presence of observed measurements is related to treatment or outcomes. That possibility cannot be assessed from group counts alone. No particular missingness mechanism, imputation method, or participant disposition should be assumed without documentation. The record also does not report a detailed analysis population beyond the stated definition, such as additional protocol-deviation rules or sensitivity populations.

The model description names treatment as a fixed effect and baseline AUC as a continuous covariate. Further details on model diagnostics, covariate handling beyond baseline AUC, and robustness checks are not reported in the registry record. The absence of those details does not show that such procedures were or were not undertaken; it means their role cannot be established from the registry record.

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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: influenza (10.0%).

The registry reports serious adverse events by treatment group. One serious adverse event was reported among 10 participants in the roflumilast plus alogliptin group. The reported count was 0 among 11 participants receiving alogliptin alone, 0 among 10 receiving roflumilast alone, and 0 among 9 receiving exenatide. The registry lists influenza at 10.0% among the top serious-event terms. The record does not provide enough detail to assign that term to a particular treatment group, so it should not be interpreted as an arm-specific rate.

These counts are descriptive. The groups are small, and a difference in event counts does not by itself establish a treatment-related cause or a comparative safety effect. The adverse-event denominators also differ from some primary endpoint group counts. The deaths fields are not populated with a numerical value; deaths are not reported in the registry record, so the blank fields should not be read as a confirmed count of zero.

The registry also reports other adverse events affecting 6 of 10 participants in the combination group, 5 of 11 in the alogliptin-alone group, 8 of 10 in the roflumilast-alone group, and 7 of 9 in the exenatide group. These are separate from serious adverse events. The short treatment period of 11 days limits what these observations can say about longer-term safety.

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

The posted results describe short-term changes in a post-meal active GLP-1 measure among participants assigned to four treatment groups. The sponsor-reported adjusted comparison of the combination with alogliptin alone is uncertain in direction, with a confidence interval that includes both negative and positive values. The reported comparisons with roflumilast alone and exenatide have positive estimates and confidence intervals above zero. Those statements concern the registered primary endpoint and its reported analysis, not an overall judgment about treatment.

A statistical contrast is not, by itself, a conclusion about patient benefit or harm. The primary endpoint is a biological measure of GLP-1 response over a defined post-meal period. The study also listed glucose and other secondary outcomes, but not all have numerical results in the registry. No broader clinical outcome or long-term follow-up result is reported in the registry record. Consequently, the posted primary result cannot establish effects beyond the measured endpoint and study duration.

The study enrolled 40 participants and treated participants for 11 days. The primary endpoint analyses include small groups, and the registry does not explain why the outcome-analysis counts differ from total enrollment. These limitations affect how much confidence can be placed in the estimates and how widely they can be generalized. The original sample-size justification and trial-design assumptions are not reported in the registry record.

The imprecise comparison with alogliptin alone does not prove that the combination and alogliptin alone are the same. A finding of equivalence would require a defined margin and an appropriate design; neither is reported in the registry record. Likewise, statistical evidence in the other two sponsor-reported comparisons does not establish a clinical consequence or resolve the absence of a reported multiplicity strategy.

The most defensible reading is therefore specific: the registry reports differing adjusted GLP-1 AUC contrasts across the specified treatment comparisons, with varying precision. The results are informative as short-term, early-phase observations, but they do not answer every question about glucose outcomes, longer treatment, or clinical outcomes.

18. Why This Trial Matters Statistically

This study matters statistically because it places several treatment comparisons around a clearly defined continuous endpoint and reports both adjusted estimates and uncertainty intervals. The primary measure combines post-meal active GLP-1 observations across a stated time window, and the ANCOVA model adjusts for baseline AUC. These details make it possible to distinguish the sponsor’s adjusted comparisons from simple differences calculated from arm summaries.

The contrast with alogliptin alone is especially instructive. Its sponsor-reported confidence interval includes zero, while a simple unadjusted recalculation has an opposite sign from the sponsor’s adjusted estimate. The registry flags this discrepancy, making clear why an analyst should not silently substitute a re-calculation for the reported model result. Statistical estimates depend on the analysis definition, not only on the displayed group means.

The study also illustrates the limits of small-group comparisons. The primary endpoint counts are 10, 10, 10, and 9, and the intervals vary in how much they constrain the treatment contrasts. A point estimate alone can obscure this uncertainty. Reading the estimate together with its confidence interval and p-value gives a more complete account, while still requiring attention to design and analysis limitations.

Finally, the posted record highlights how outcome families and incomplete reporting shape interpretation. Multiple sponsor-reported contrasts and several secondary outcomes are present, but a multiplicity strategy is not reported. Numerical results are not reported for every listed secondary measure. These features make careful separation of primary, secondary, adjusted, and recalculated results important. The registry record is most useful when read as a report of specific estimates and their limits, rather than as a single definitive answer to a broad clinical question.

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 study was short, with each treatment description specifying 11 days of dosing. Its primary endpoint was a postprandial biomarker measure assessed at baseline and Day 11, so the registry results do not establish longer-term patterns. The study was a Phase 1b investigation, and the enrolled sample was 40 participants. Primary endpoint group counts were 10, 10, 10, and 9; the reason for the difference between enrollment and analysis counts is not reported in the registry record.

The sponsor-reported primary comparison of the combination with alogliptin alone is imprecise. Its 95% confidence interval spans zero and includes contrasts in both directions. The registry record flags a discrepancy between the sponsor’s ANCOVA estimate and the simple recalculation from posted summary statistics. The sponsor’s own analysis takes precedence, but the record does not provide individual-level data for independent model verification.

The registry lists several secondary outcomes, but numerical results are not reported for appetite sensation or 24-hour average plasma glucose. It does not report a multiplicity adjustment or a complete prespecified testing hierarchy. Consequently, the strength of evidence across the multiple comparisons cannot be assessed as though they formed a clearly controlled family of tests.

The original sample-size rationale and design assumptions are not reported in the registry record. Missing-data reasons, any imputation method, and sensitivity analyses are also not reported in the registry record. Details of allocation concealment, masking procedures, model diagnostics, and analysis robustness are not reported in the registry record. These omissions limit assessment of potential bias and model dependence; they do not establish that a particular procedure was or was not performed.

The exenatide comparator is described as open-label, while the registry classifies masking as quadruple. How these descriptions were reconciled is not reported in the registry record. The safety results are descriptive and come from small groups with denominators that differ across listed analyses. These registry entries constrain both comparative interpretation and generalization beyond the specific short-term study.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Roflumilast Plus Alogliptin Proof-of-Mechanism Study in Type2 Diabetes — NCT01664624.
• Registry record last updated 2017-02-01; status last verified 2016-09.
• Results first posted 2014-02-20.
• 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

The registry describes a completed, randomized, parallel-group Phase 1b study of roflumilast plus alogliptin in people with type 2 diabetes inadequately controlled on a stable dose of metformin. It lists 40 enrolled participants and four treatment groups: the combination, alogliptin alone, roflumilast alone, and exenatide. Treatment was specified for 11 days.

The primary endpoint was change from baseline to Day 11 in postprandial active GLP-1 AUC from 0 to 8 hours. The sponsor used ANCOVA with treatment as a fixed effect and baseline AUC as a continuous covariate. The combination versus alogliptin comparison was -5.1 pmol/L*hr, with a 95% confidence interval from -16.9 to 6.7 and p = 0.388. The registry flags a discrepancy with a simple unadjusted recalculation, so the sponsor’s analysis takes precedence.

The record lists secondary outcomes concerning glucose, C-peptide, insulin, appetite sensation, and average plasma glucose. It does not report numerical results for every listed secondary outcome, a multiplicity strategy, the original sample-size rationale, or details of missing-data handling. The findings should be interpreted as short-term estimates for the registered measures and comparisons, with uncertainty and reporting limitations kept in view.