Trial of PDE4 Inhibition With Roflumilast for the Management of Atopic Dermatitis (INTEGUMENT-II): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Trial of PDE4 Inhibition With Roflumilast for the Management of Atopic Dermatitis (INTEGUMENT-II) (NCT04773600), a randomized phase 3 study. The primary outcome is Achievement of Validated Investigator Global Assessment for Atopic Dermatitis (vIGA-AD) Success at Week 4.

2. Clinical Question

This is a question about a defined clinical assessment at a specified time, not about every possible symptom or longer-term outcome. The registry record also lists secondary outcomes involving earlier vIGA-AD assessments, itch, eczema severity, and clear or almost clear status. Those measures add context, but they do not change the primary question: how often participants met the prespecified vIGA-AD success definition at Week 4 in each randomized arm.

3. Trial Design

INTEGUMENT-II is a Phase 3 randomized, parallel-group, double-blind, vehicle-controlled study. The registry record reports enrollment of 683 participants and a completed trial. Participants were assigned to roflumilast cream 0.15% or vehicle cream, each applied once daily for 4 weeks. The record describes masking as triple; it does not identify the specific roles covered by that masking description.

The primary outcome is a binary response: each participant either achieved vIGA-AD success at Week 4 or did not. Its reported unit is the percentage of participants. The primary outcome description states that all randomized participants are included and that missing observations are handled with multiple imputation. The details of the imputation procedure are not reported in the registry record.

The arm-level primary results are based on 451 participants in the roflumilast group and 232 in the vehicle group. The adverse-event record uses at-risk counts of 452 and 232, respectively. These are distinct registry entries for different result categories; the reason for the difference in the active-arm counts is not reported in the registry record.

The registry record gives the study start as 2021-02-24 and primary completion as 2022-09-29. It does not report the randomization sequence method, allocation concealment procedure, stratification factors, detailed visit schedule, or the trial’s original sample-size justification. These details should not be inferred from the randomized, masked, parallel-group description alone.

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: All randomized participants 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 Achievement of Validated Investigator Global Assessment for Atopic Dermatitis (vIGA-AD) Success at Week 4, measured over Week 4.

The primary endpoint is the percentage of participants achieving validated Investigator Global Assessment for Atopic Dermatitis success at Week 4. Success requires a vIGA-AD score of clear or almost clear and an improvement of 2 grades from baseline. The scale is an ordinal assessment of overall atopic dermatitis severity, with integer scores from 0 to 4; higher scores indicate greater severity.

This endpoint combines a status at the assessment visit with improvement from baseline. Requiring both means that a low final score alone is not sufficient: the participant must also meet the specified change criterion. A binary success outcome makes the result directly interpretable as the proportion meeting the definition in each randomized arm, although it compresses the range of possible scores and changes into success versus non-success.

In general, a standardized investigator assessment can support consistent comparisons across study arms when the assessment is applied in the same way. The registry record identifies this measure as the primary outcome and reports the Week 4 percentages with multiple imputation for missing observations. It does not report further details about assessor training, assessment reliability, or the imputation model.

7. Results Posted in the Registry

• Roflumilast Cream 0.15%: 28.8% (130/451) versus 12.1% (28/232) on control; risk difference 16.8 percentage points (95% CI 10.8 to 22.7), risk ratio 2.39 (1.64–3.48); unadjusted two-proportion 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.

Achievement of vIGA-AD Success at Week 4 in Participants With "Moderate" Baseline Scores

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Roflumilast Cream 0.15%: 33.0% (113/342) versus 12.8% (23/179) on control; risk difference 20.2 percentage points (95% CI 13.2 to 27.2), risk ratio 2.57 (1.71–3.88); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Roflumilast Cream 0.15% than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

vIGA-AD Success at Week 2

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Roflumilast Cream 0.15%: 17.7% (80/451) versus 5.2% (12/232) on control; risk difference 12.6 percentage points (95% CI 8.0 to 17.1), risk ratio 3.43 (1.91–6.16); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Roflumilast Cream 0.15% than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

vIGA-AD Success at Week 1

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Roflumilast Cream 0.15%: 6.0% (27/451) versus 3.0% (7/232) on control; risk difference 3.0 percentage points (95% CI -0.1 to 6.1), risk ratio 1.98 (0.88–4.49); unadjusted two-proportion z-test p = 0.091.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Roflumilast Cream 0.15% and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes extend the assessment beyond the primary Week 4 result. They include earlier vIGA-AD success assessments, a subgroup defined by moderate baseline scores, itch reductions among participants with a specified baseline itch score, an EASI response, and clear or almost clear status. Taken together, these measures address related but not identical aspects of severity and symptoms, and they use different response definitions and assessment times.

The posted secondary results available in the record include an overall pattern broadly consistent with the primary Week 4 finding at some assessments, while the earliest assessment is less conclusive. A result in a subgroup with moderate baseline scores offers related context, but it is not the same population as the all-randomized primary analysis. The secondary endpoints should therefore be treated as supportive context rather than as substitutes for the primary comparison.

In general, testing many outcomes creates more opportunities for chance findings. Unless multiplicity is addressed, a nominal p-value for an individual secondary endpoint does not have the same confirmatory interpretation as a prespecified primary analysis with a controlled testing strategy. The registry record does not report a multiplicity adjustment or a hierarchy for secondary hypotheses. Unadjusted secondary results consequently carry less weight, particularly when endpoints, time points, and subgroups are considered together.

The registry lists 10 posted secondary outcomes. The exact relationship among all analyses, any prespecified ordering, and the treatment of missing values for each secondary outcome are not reported in the registry record. The overall pattern can help assess coherence across related measures, but it does not eliminate uncertainty about multiplicity, selection, or differences in endpoint definitions.

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

• Achievement of vIGA-AD Success at Week 4 in Participants With "Moderate" Baseline Scores
• vIGA-AD Success at Week 2
• vIGA-AD Success at Week 1
• Achievement of a 4-point Reduction at Week 4 in the Average, Weekly Worst Itch Numeric Rating Scale (WI-NRS) in Participants With Baseline WI-NRS Score ≥4
• Achievement of a 4-point Reduction at Week 2 in the WI-NRS in Participants With Baseline WI-NRS Score ≥4
• Achievement of a 4-point Reduction at Week 1 in the WI-NRS in Participants With Baseline WI-NRS Score ≥4
• Achievement of ≥75% Decrease From Baseline at Week 4 in the Eczema Area and Severity Index (EASI-75)
• Achievement of vIGA-AD Score of 'Clear' or 'Almost Clear' at Week 4

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is binary, because vIGA-AD success is recorded as achieved or not achieved. The registry reports the percentage achieving success in each arm and a between-arm odds ratio from a Cochran-Mantel-Haenszel analysis. These measures answer related but distinct questions: percentages describe the proportion meeting the endpoint in each group, while an odds ratio compares the odds of success between the groups.

What does the odds ratio describe?

In general, odds are the probability of an event divided by the probability of no event. An odds ratio compares those odds between two groups. An odds ratio above 1 indicates higher odds in the numerator group relative to the comparison group, under the defined ordering. The reported primary odds ratio is 3.19 for roflumilast cream relative to vehicle cream. It should not be described as a percentage-point difference or as the ratio of the two response percentages.

The reported method is Cochran-Mantel-Haenszel. In general, this method combines information across strata to estimate an association while accounting for the stratification used in an analysis. The registry record does not report the strata or other implementation details, so it is not possible to identify what factors, if any, were incorporated into the reported comparison. The arm-level recalculation is unadjusted and uses only posted summary statistics; it is not a reproduction of every feature of the reported analysis.

How should the p-value be read?

In general, a p-value assesses how compatible the observed data are with a specified null hypothesis under the statistical model and its assumptions. The primary reported p-value is p < 0.001. The separate unadjusted recalculation reports p < 0.001. Those values belong to different analyses and should be kept with their respective methods. A p-value is not the probability that the treatment has no effect, nor does it measure the size or practical importance of a difference.

The registry labels the reported comparison as a superiority analysis. That label describes the statistical objective of testing for a difference in a direction consistent with superiority; it does not substitute for examining the effect estimate, uncertainty, endpoint definition, and design limitations. The registry record does not report the trial’s full statistical analysis plan, including the specified significance threshold for the primary analysis or details of any planned interim analyses.

What does multiple imputation contribute?

The primary outcome description says that missing observations were handled using multiple imputation. In general, multiple imputation replaces missing values with a set of plausible values drawn using a model, analyzes each completed dataset, and combines the results to reflect imputation uncertainty. It can make use of observed information, but its validity depends on the imputation model and assumptions about why data are missing. The registry record does not report the model, variables used, number of imputations, missingness pattern, or sensitivity analyses. The displayed percentages therefore should be understood in light of the stated method, without assuming unreported implementation details.

The endpoint population is all randomized participants, according to the registry description. This indicates the intended analysis population for the primary result, but details such as exclusions, protocol deviations, and how each such case was handled are not reported in the registry record. The arm counts and the all-randomized description should be read as registry-reported information rather than as a complete account of participant flow.

Why are confidence intervals useful?

In general, a confidence interval gives a range of effect estimates compatible with the data under the method used, while reflecting sampling uncertainty. It conveys more information about estimate precision than a p-value alone. The primary reported interval is for the odds ratio, so it is on the odds-ratio scale; it should not be interpreted as an interval for the difference in percentages. The interval does not account automatically for every design limitation, model assumption, or unreported analysis decision.

The unadjusted recalculation uses the posted arm-level success proportions to compare the arms on a risk-difference scale. This is a straightforward summary, but it cannot recover information not contained in the aggregate counts and percentages. In general, recalculations from summary statistics can differ from a trial analysis because methods, populations, missing-data handling, stratification, or covariate adjustment may differ. The registry record does not report enough implementation detail to attribute any numerical difference to a particular factor.

11. Confidence Intervals and Precision

The registry-reported primary estimate is an odds ratio of 3.19, with a 95% confidence interval from 1.962 to 5.183. The interval is entirely above 1, the null value for an odds ratio. In general, a confidence interval describes the range of values compatible with the data and the method under repeated sampling assumptions; it is not a guarantee that the true value lies in the particular interval, nor does it incorporate every possible source of bias.

The interval gives information about precision on the odds-ratio scale. It should be interpreted alongside the arm percentages, 28.9% and 12.0%, because odds ratios are not direct percentage-point differences. The interval spans multiple possible magnitudes of association, so the point estimate alone should not be treated as exact. The registry does not give a confidence interval around the arm percentages in the posted primary results.

The unadjusted recalculation uses a risk-difference measure and has its own uncertainty interval in the statistical output. That interval is distinct from the registry-reported odds-ratio interval, both in scale and method. In general, uncertainty intervals from different estimands should not be compared as though they were measuring the same quantity. The record does not report a direct comparison between the recalculated risk difference and the reported odds ratio.

Precision is also affected by the number of observed outcomes, missing observations, outcome definition, and analysis assumptions. The registry states that missing primary observations were handled with multiple imputation, but details needed to assess the resulting uncertainty are not reported in the registry record. Thus, the confidence interval is informative about sampling uncertainty under the reported analysis, but it cannot resolve questions about missing-data assumptions or generalizability.

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 planning figures are illustrations derived from the observed effect, not the trial’s original design calculation. The registry record does not report the original sample-size justification or its design assumptions. For non-time-to-event outcomes such as this binary endpoint, the figures are participants per arm, rather than total events. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The assumed effect to detect is the stated fraction of the observed effect.

The 80% power column gives the illustrative participants per arm for a design with an 80% chance of detecting the assumed effect under those planning assumptions, if that effect were the true effect. The 90% power column uses a higher target probability of detecting it. For the primary outcome, planning for 100% of the observed effect gives 90 participants per arm for 80% power and 120 participants per arm for 90% power. Planning for 75% of the observed effect gives 148 participants per arm for 80% power and 198 participants per arm for 90% power. Planning for 50% of the observed effect gives 304 participants per arm for 80% power and 406 participants per arm for 90% power.

The pattern reflects a general design principle: smaller effects are harder to distinguish from random variation, so detecting them requires more participants under otherwise identical assumptions. These figures are not observed power. They are planning illustrations, and they do not establish that the trial was designed using these assumptions or that the observed result has a particular probability of replication.

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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.121; treatment proportion p₁ = 0.288; average p̄ = 0.204.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.5704 = 1.1179.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5579 = 0.4696.
4. n = (1.1179 + 0.4696)² / (0.1676)² = 89.8, rounded up to 90 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
p0 <- 0.120689655172414                          # control proportion
p1 <- 0.288248337028825                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^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 (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(4773600)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 90 (80% power), 120 (90% power)
 75% of observed effect: n per arm = 148 (80% power), 198 (90% power)
 50% of observed effect: n per arm = 304 (80% power), 406 (90% power)
power.prop.test: 89.8 
Simulated power at n = 90 per arm: 0.806 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-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 registry-reported primary analysis is a Cochran-Mantel-Haenszel odds ratio of 3.19, with a 95% confidence interval from 1.962 to 5.183 and p < 0.001. The arm-level result is also expressed as percentages: 28.9% for roflumilast cream 0.15% and 12.0% for vehicle cream. These are complementary descriptions, not identical statistical measures.

The recalculated comparison uses the posted arm-level summary statistics and is unadjusted. It expresses the comparison as a risk difference and gives p < 0.001. The reported analysis uses an odds ratio and the Cochran-Mantel-Haenszel method. As a result, the two p-values and effect measures should not be expected to match exactly or be substituted for one another. The registry record does not report enough detail about the reported analysis’s strata or implementation to explain any difference in numerical terms.

The recalculation is useful as a transparent check based on arm-level results, but it is not a replacement for the trial’s reported method. No discrepancy is flagged between the reported and recalculated results in the record. The estimates should be read according to the scale and method attached to each one, with the registry-reported analysis representing the trial’s stated analysis.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates more than one outcome, time point, or subgroup. In general, the more statistical tests conducted, the greater the chance that at least one result will appear notable by chance if each is interpreted separately. A prespecified testing hierarchy, adjustment procedure, or clearly designated confirmatory endpoint can help control this issue. The registry identifies a primary endpoint and lists secondary outcomes, but a multiplicity adjustment or testing hierarchy is not reported in the registry record.

The primary result is described as using multiple imputation for missing observations. This is a stated feature of the primary endpoint analysis, but the registry does not report the imputation model, variables included, number of completed datasets, missingness by arm, or sensitivity analyses. In general, multiple imputation relies on assumptions about the relationship between missing values and observed data. Without those details, the robustness of the result to alternative missing-data assumptions cannot be assessed from the registry record alone.

The primary population is described as all randomized participants. That description is consistent with an intention-to-treat approach in broad terms, but the registry record does not report detailed rules for inclusion, exclusions, protocol deviations, or handling of participants without an evaluable assessment. No specific analysis population beyond the stated all-randomized group should be inferred.

The record also does not report the complete statistical analysis plan, the original sample-size justification, planned interim analyses, or whether any secondary-outcome testing adjustments were conducted. These omissions do not negate the reported result, but they limit how fully the analysis choices can be evaluated and how confidently the secondary findings can be interpreted as confirmatory.

16. Safety Results

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

Deaths

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

Serious events with the highest arm-level rate: nerve compression (0.2%), staphylococcal scalded skin syndrome (0.2%), general physical health deterioration (0.2%), dermatitis atopic (0.2%).

The serious adverse-event record reports 4 serious adverse events among 452 participants at risk in the roflumilast cream 0.15% group and 0 among 232 at risk in the vehicle group. The registry reports 0 other adverse events in each group and 0 deaths in each group. The at-risk counts for safety are the counts recorded in the safety results and should be distinguished from the primary outcome arm counts.

The listed serious-event terms are nerve compression, staphylococcal scalded skin syndrome, general physical health deterioration, and atopic dermatitis; each is reported as 0.2%. These entries describe terms recorded in the registry and do not, by themselves, establish whether an event was caused by treatment. The registry record does not provide further details about event timing, severity beyond the serious classification, outcome, or relatedness assessment.

In general, safety counts are descriptive and should be read in the context of the study duration, the numbers at risk, and the types of events collected. This trial’s treatment period was 4 weeks, so these entries do not characterize longer-term safety. A count of zero in a group means no event was reported for that category in the posted record; it does not establish that an event is impossible.

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

The posted primary results show different Week 4 vIGA-AD success percentages in the randomized arms: 28.9% with roflumilast cream 0.15% and 12.0% with vehicle cream. The registry-reported comparison is an odds ratio of 3.19, with a 95% confidence interval from 1.962 to 5.183 and p < 0.001. These results describe the trial’s defined endpoint and the statistical analysis reported for it.

They do not establish what will happen for a particular person. A group-level percentage cannot predict an individual response, and an odds ratio is not an individual probability. The endpoint also requires a particular combination of final assessment and change from baseline; it does not encompass every possible change in eczema severity, itch, or daily functioning.

The result is limited to the 4-week treatment period and the enrolled study population. The record does not supply evidence about longer-term outcomes. Nor does the primary outcome by itself answer every question raised by the listed secondary measures. Those address other symptoms, definitions, time points, or subgroups and are subject to the limitations of multiple testing.

Statistical evidence should be considered alongside precision and analysis assumptions. The confidence interval indicates uncertainty around the reported odds ratio under the stated method, while the details of missing-data imputation and other analysis decisions are not fully reported. The result therefore supports a between-arm difference on the primary endpoint as analyzed, but it should not be expanded into claims about unmeasured outcomes, duration beyond the study, or individual benefit.

These statistical findings are not medical advice. Decisions about care require an individual clinical context and information beyond the registry outcome summaries.

18. Why This Trial Matters Statistically

This trial illustrates why an outcome definition matters. vIGA-AD success is not simply a score at Week 4: it requires clear or almost clear status as well as a 2-grade improvement from baseline. That combined rule determines who counts as a responder and shapes the percentages, odds ratio, and planning illustrations. Statistical interpretation depends on understanding the endpoint before interpreting its effect estimate.

It also demonstrates how the same comparison can be reported on different scales. The arm percentages are accessible summaries of response frequency, while the registry-reported odds ratio comes from a Cochran-Mantel-Haenszel analysis. An unadjusted recalculation provides another summary, but it does not reproduce the reported method. Keeping the estimand and analysis method attached to each number prevents misleading comparisons.

Finally, the record shows the distinction between trial results and design assumptions. Planning figures based on the observed effect can illustrate how sample size responds to assumed effect size and desired power, but they are not the trial’s original justification and are not observed power. The original design assumptions are not reported in the registry record.

For researchers, the case emphasizes the value of reporting analysis details alongside results: missing-data methods, multiplicity procedures, analysis populations, and the construction of reported comparisons all affect interpretation. The primary result is clear as a registry-reported statistical finding, while several details needed for a full audit of the analysis are not reported.

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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The registry record does not report the full statistical analysis plan, the original sample-size justification, or whether interim analyses were planned or conducted. It also does not provide the randomization sequence method, allocation concealment details, or the specific roles covered by triple masking. These omissions limit assessment of design and analysis procedures beyond the broad randomized, masked, parallel-group description.

For the primary outcome, multiple imputation is named, but its model, variables, missingness by arm, and sensitivity analyses are not reported in the registry record. The all-randomized population is stated, but detailed participant-flow and analysis-inclusion rules are not included in the registry record for interpretation. The safety at-risk count in the active arm differs from the primary outcome arm count, and the reason is not reported in the registry record.

The reported primary measure is a dichotomized assessment at Week 4. As a general limitation of binary endpoints, participants with different score patterns can be grouped together if they meet the same success rule, while small changes that do not meet the rule count as non-success. The registry record does not provide additional information about assessor reliability or how borderline assessments were handled.

The secondary outcomes cover multiple measures and time points, but a multiplicity adjustment or hierarchy is not reported. This limits the confirmatory weight of individual secondary findings. The treatment period was 4 weeks, so the posted results do not establish longer-term outcomes. Generalizability beyond the enrolled population also cannot be assessed fully from the registry summary alone.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Trial of PDE4 Inhibition With Roflumilast for the Management of Atopic Dermatitis (INTEGUMENT-II) — NCT04773600.
• Registry record last updated 2024-10-04; status last verified 2024-09.
• Results first posted 2024-10-04.
• 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

INTEGUMENT-II (NCT04773600) is a completed Phase 3 randomized, triple-masked, parallel-group, vehicle-controlled study of roflumilast cream 0.15% for atopic dermatitis. Participants applied roflumilast cream or vehicle once daily for 4 weeks. Enrollment was 683.

The primary endpoint was vIGA-AD success at Week 4, defined as clear or almost clear status plus a 2-grade improvement from baseline. The registry reports success percentages of 28.9% with roflumilast cream and 12.0% with vehicle, with a Cochran-Mantel-Haenszel odds ratio of 3.19 (95% CI 1.962 to 5.183; p < 0.001). Missing primary observations were handled using multiple imputation, although implementation details are not reported in the registry record.

The record lists secondary outcomes and serious adverse-event information. The original sample-size justification, full analysis plan, multiplicity procedure, and detailed missing-data methods are not reported in the registry record.