Evaluating the Effects of Tradipitant vs. Placebo in Atopic Dermatitis (EPIONE2): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Evaluating the Effects of Tradipitant vs. Placebo in Atopic Dermatitis (EPIONE2) (NCT04140695), a randomized phase 3 study. The primary outcome is Worst Itch Numeric Rating Scale (WI-NRS) Responder Rate at Week 2.

2. Clinical Question

3. Trial Design

EPIONE2 (NCT04140695) is identified in the registry as a Phase 3 randomized, parallel-group study of tradipitant versus placebo in participants with atopic dermatitis. The official title describes an efficacy study of the neurokinin-1 receptor antagonist VLY-686. The registry record names Vanda Pharmaceuticals as the sponsor. The brief summary describes a multicenter study conducted in the United States.

The design is listed as randomized, quadruple-masked, and parallel. The primary outcome description says participants were randomized 1:1 to tradipitant or placebo. Tradipitant and placebo are both described as oral capsules. The registry record does not specify the allocation-generation procedure, concealment method, masking roles, or how the study maintained masking; those details are not reported in the registry record.

The registry record reports enrollment of 87 participants. The primary outcome results include 42.0 participants in the tradipitant group and 43.0 in the placebo group. The primary endpoint was assessed at 2 weeks. The record lists a start date of 2019-10-21, a primary completion date of 2020-10-14, and a status of terminated. The reason for termination, the timing of the termination decision, and any consequences for recruitment or follow-up are not reported in the registry record.

One primary outcome measure is listed, and the record contains no posted secondary outcome measures. There are zero analysis entries in the registry. This means the registry record does not present a detailed statistical analysis plan or an analysis entry explaining how the comparison was conducted. The reported arm summaries and comparison estimates can be described, but the design and analysis procedures beyond those explicitly listed should not be inferred.

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

Analysis population for the primary outcome, as defined in the registry: All participants randomized (1:1) to placebo or tradipitant.

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 Worst Itch Numeric Rating Scale (WI-NRS) Responder Rate at Week 2, measured over 2 weeks.

The primary endpoint is the Worst Itch Numeric Rating Scale responder rate at Week 2. It is a binary outcome: each participant either meets the responder definition or does not. A responder was defined as achieving at least 4 points reduction from baseline in the weekly average of diary WI-NRS. The diary score runs from 0, described as no itch, to 10, described as the worst itch imaginable.

Using a responder definition turns a continuous or repeated symptom score into a count of participants who meet a prespecified threshold. That can make a result easier to interpret as the frequency of participants reaching a defined degree of improvement, while discarding information about score changes that fall short of the threshold and about the range of changes among responders. This is a general feature of dichotomizing a measurement, not a claim about the study’s rationale.

The endpoint is reported as counts of participants: 2.0 in the tradipitant arm and 4.0 in the placebo arm. The reported risk difference is -0.045, defined in the comparison as tradipitant relative to placebo. Its interval extends from -0.153 to 0.063. The registry record does not report the underlying individual diary scores, the handling of missing diary entries, or further detail about how weekly averages were calculated. Those details matter for understanding exactly how participants were classified, but are not reported in the registry record.

7. Results Posted in the Registry

• Tradipitant: 4.8% (2/42) versus 9.3% (4/43) on control; risk difference -4.5 percentage points (95% CI -15.4 to 6.3), risk ratio 0.51 (0.10–2.65); unadjusted two-proportion z-test p = 0.414.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is binary because it classifies each participant as either a WI-NRS responder or a non-responder at Week 2. For a binary outcome, the arm-level response proportion is the number of responders divided by the number assessed in that arm. The registry reports the event counts and arm sizes, as well as the comparison quantities. The comparison uses placebo as the reference and reports the tradipitant-minus-placebo risk difference. A negative risk difference means that the observed responder proportion was lower in the tradipitant arm; a positive value would mean it was higher. The estimate is -0.045.

A risk difference expresses an absolute contrast between two proportions. Unlike a relative measure, it is stated on the probability scale and can be read in relation to the observed frequency of the outcome. Its interpretation depends on which outcome is counted and on the stated subtraction order. Here, the outcome is meeting the responder definition, and the registry’s comparison is tradipitant relative to placebo. The accompanying interval runs from -0.153 to 0.063. Since this interval crosses zero, it includes no absolute difference as well as differences in either direction.

The registry also reports a risk ratio and an odds ratio. A risk ratio compares the probability of response in one arm with the probability in the other. A value of 1 represents equal response probabilities; values below or above 1 represent a lower or higher probability in the intervention arm under the stated comparison. The reported risk ratio is 0.512, with an interval from 0.099 to 2.648. An odds ratio compares odds rather than probabilities. Its reported estimate is 0.487, with an interval from 0.084 to 2.816. These relative measures can differ numerically from the risk difference because they use different scales. With few observed responders, relative estimates can be especially sensitive to the counts and their uncertainty.

A confidence interval is a range generated by a statistical procedure to express uncertainty around an estimate. In general, an interval’s width reflects both the amount of information in the data and the method used to construct it. Narrow intervals indicate greater precision; wide intervals leave more uncertainty about plausible effect sizes. An interval should not be read as a list of values that are equally likely, and it is not a guarantee that the true effect lies within its endpoints. The registry record does not report the confidence level or the interval-construction method, so those features should not be inferred from the endpoint values alone.

The unadjusted p-value for the comparison is 0.414. In general, a p-value evaluates how unusual the observed data, or data more extreme, would be under a specified null model. It is not the probability that the null hypothesis is true, and it does not measure the magnitude or clinical importance of an effect. The record does not state the test statistic, test procedure, or the assumptions used for this p-value. The word unadjusted indicates that the posted value is not described as adjusted for multiple comparisons, but the registry record does not explain whether any other adjustment was contemplated.

The comparison statistics accompany the arm summaries, but there are zero analysis entries in the registry record. The exact analysis population beyond the primary outcome description, the treatment of incomplete outcome data, any covariate adjustment, and the method for estimating the risk difference are not reported in the registry record. The primary outcome description identifies all randomized participants as the population, but the posted outcome arm sizes are 42.0 and 43.0. The record does not explain whether these counts reflect participants with observed outcome data, a particular analysis set, or another reporting convention.

For binary outcomes, few events can lead to unstable estimates, particularly for relative measures. This is a general statistical issue: when the number of events is small, a change in only a few classifications can materially alter the estimated proportions and relative contrasts. The observed counts are 2.0 and 4.0, so the numerical estimates should be read with their broad intervals rather than treated as precise measurements. The data do not support a detailed exploration of subgroup patterns or adjusted relationships because such analyses are not posted.

Several methodological details remain unknown. The registry record does not report the trial’s original sample-size justification, the planned statistical analysis method, whether interim analyses were planned or conducted, the prespecified significance level for the primary analysis, or the intended approach to multiplicity. It also does not report a missing-data method or additional analysis populations. These omissions prevent a full reconstruction of the planned analysis and make it important to distinguish the posted descriptive comparison from an independently documented protocol analysis. General definitions explain how these statistics are usually interpreted; they do not establish which particular procedures were used in this trial.

10. Confidence Intervals and Precision

The primary risk-difference estimate is -0.045, and its reported interval extends from -0.153 to 0.063. The interval includes zero and reaches in both negative and positive directions. Thus, the point estimate alone does not convey the full uncertainty: the data are compatible with a range of absolute differences, including a lower responder proportion with tradipitant, little or no difference, or a higher responder proportion. The registry record does not state the confidence level or the procedure used to construct this interval.

The relative measures are also imprecise. The risk ratio is 0.512, with an interval from 0.099 to 2.648. The odds ratio is 0.487, with an interval from 0.084 to 2.816. For both ratios, the interval includes 1, the value corresponding to no relative difference. Their ranges extend far on either side of that value, so these estimates do not tightly identify the relative contrast.

The outcome counts are small: 2.0 responders among 42.0 participants in the tradipitant arm and 4.0 among 43.0 in the placebo arm. As a general principle, limited event information tends to produce less precise estimates, particularly for binary outcomes and relative measures. The reported unadjusted p-value is 0.414, but it should be read together with the effect estimates and intervals rather than as a stand-alone verdict.

Precision is not the same as whether a result is statistically significant, and neither alone determines the practical importance of an effect. The interval makes the uncertainty visible; it does not identify which value within the range is correct. More detailed precision assessment is constrained because the registry record does not report the interval method, confidence level, or analysis details.

11. 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 trial’s own design justification. For this binary outcome, the figures refer to participants per arm, not total events. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each row considers detecting a stated fraction of the observed effect: 100% of the observed effect, 75% of the observed effect, or 50% of the observed effect.

The 80% power column gives the estimated number of participants per arm for a planning scenario designed to have 80% probability of detecting the specified effect under the listed assumptions, if that effect were the true effect and the assumptions held. The 90% power column expresses the corresponding planning target for 90% probability. These are planning concepts, not guarantees that a future study will detect an effect. The listed figures are 497 and 665 participants per arm for planning for 100% of the observed effect; 950 and 1,271 for planning for 75% of the observed effect; and 2,284 and 3,057 for planning for 50% of the observed effect, respectively.

Smaller effects require more participants to distinguish them from random variation at a given significance level and power target. The figures increase as the assumed effect becomes a smaller fraction of the observed effect. Because the observed effect in this trial comes from a small number of responders and is uncertain, using it as a planning input is only illustrative; it should not be treated as a stable estimate of a future effect. These calculations are not observed power. The trial’s original sample-size assumptions and justification are not reported in the registry record.

12. 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 = 0.093; treatment proportion p₁ = 0.048; average p̄ = 0.070.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.3616 = 0.7087.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.3602 = 0.3031.
4. n = (0.7087 + 0.3031)² / (-0.0454)² = 496.6, rounded up to 497 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 <- 0.0930232558139535                          # control proportion
p1 <- 0.0476190476190476                          # 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(4140695)
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 = 497 (80% power), 665 (90% power)
 75% of observed effect: n per arm = 950 (80% power), 1,271 (90% power)
 50% of observed effect: n per arm = 2,284 (80% power), 3,057 (90% power)
power.prop.test: 496.6 
Simulated power at n = 497 per arm: 0.807 (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.

13. Multiplicity, Missing Data and Analysis Populations

The registry record lists one primary outcome and no posted secondary outcome measures. It reports an unadjusted p-value of 0.414 for the primary comparison. No analysis entries are posted. The record does not describe a multiplicity procedure, such as an adjustment or a hierarchy for testing multiple outcomes, treatment groups, time points, or interim looks. It is therefore not possible to determine from the record whether multiplicity was relevant to the original analysis plan or how it would have been handled.

Multiplicity is a general concern when several statistical tests are performed: conducting more tests can increase the chance of at least one apparently unusual result by chance alone. This concept does not establish that multiple tests were performed in EPIONE2. The available registry entries do not identify planned interim analyses, exploratory analyses, or other testing families, so none should be assumed. The posted p-value is described as unadjusted, and no adjusted value is reported.

The endpoint description identifies all randomized participants as the primary population, but the record does not specify the analysis-set rules or how participants without an observed Week 2 outcome were handled. The outcome counts and denominators are reported, but there is no account of missing diary records, withdrawals, exclusions, or imputation. A complete-case analysis, an imputed analysis, or another approach cannot be inferred.

These omissions matter because the analysis population and missing-data approach can affect both the estimated response proportions and their uncertainty. They also limit evaluation of whether the reported outcome counts align with the stated population of all randomized participants. The registry record does not provide enough information to assess that alignment or to reconstruct a missing-data sensitivity analysis.

14. 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 2% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: procedural complication (2.3%), obstructive pancreatitis (2.3%).

The registry reports serious adverse events for 1 of 43 participants at risk in the tradipitant group and 1 of 44 participants at risk in the placebo group. Deaths are reported as 0 in each group. These are event counts and group denominators as recorded; they are not a comparison adjusted for differences between participants or for exposure time.

Other adverse events are reported for 13 of 43 participants at risk in the tradipitant group and 10 of 44 participants at risk in the placebo group. The record does not provide event-level detail for these counts in the information used for this interpretation, nor does it specify whether events were judged related to study treatment. The top serious-event terms listed are Procedural Complication and Obstructive Pancreatitis, each shown as 2.3%. The record does not state in this listing how the terms map to the group counts.

Safety counts should be read descriptively. They do not by themselves establish that an intervention caused an event or that a difference between groups is meaningful. The registry record does not report additional details needed to interpret timing, severity beyond the serious-event designation, exposure duration, or the method for collecting and classifying adverse events.

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

The posted primary outcome describes how many participants met the specified itch-responder definition at Week 2. There were 2.0 responders among 42.0 participants in the tradipitant arm and 4.0 among 43.0 in the placebo arm. The risk-difference estimate is -0.045, with an interval from -0.153 to 0.063, and the unadjusted p-value is 0.414. Together, these figures describe a lower observed responder proportion in the tradipitant group, alongside substantial uncertainty about the size and direction of the underlying difference.

The result does not show that the groups are equivalent. A failure to demonstrate a clear difference is not the same as evidence that any difference is absent. The broad intervals include values in both directions, so the comparison does not precisely distinguish a lower response proportion, no difference, or a higher response proportion. Conversely, the point estimate alone is not evidence that the observed difference represents a reliable treatment effect.

The endpoint reflects a threshold-based classification at one time point. It does not summarize every participant’s score trajectory, the magnitude of change among participants who did not meet the threshold, or outcomes beyond Week 2. The registry record does not report secondary outcomes, individual diary data, or longer-term results. It also does not report the analysis details needed to assess missing diary data or other analysis decisions.

The record identifies the trial as terminated, but the reason for termination is not reported in the registry record. The termination status should not be interpreted as evidence about the outcome, nor should the outcome be used to infer why the trial ended. The posted statistics are limited to the recorded endpoint and safety counts. They do not establish a causal conclusion or provide a basis for medical advice. They should be understood as a small, uncertain body of trial evidence rather than a definitive answer about the intervention.

16. Why This Trial Matters Statistically

This trial illustrates why reporting an effect estimate together with its uncertainty is important. The primary outcome is a binary responder measure, and the posted counts are 2.0 in the tradipitant group and 4.0 in the placebo group. The risk-difference estimate is negative, but its interval crosses zero and spans values on both sides. Reporting only the point estimate would conceal how much uncertainty remains; reporting only the p-value would also leave the possible range of effects unclear.

The endpoint’s threshold is straightforward to describe, but converting a symptom scale into a responder category has statistical consequences. Participants with score changes near the threshold can be assigned to different categories despite similar underlying changes, while variation among participants within each category is hidden. This is a general property of threshold-based outcomes. The registry record does not provide individual data that would allow the continuous score pattern to be assessed alongside the responder result.

The trial also highlights the importance of recording analysis details. The registry identifies the population and outcome definition, but does not provide a detailed analysis plan, missing-data method, interval method, or multiplicity approach. Those details affect how readers can evaluate the reported comparison and whether it follows a prespecified plan. The record’s status as terminated adds context, but the reason for termination is not reported and should not be guessed.

Finally, the sample-size illustrations show how planning depends on assumptions about the effect to detect. Smaller assumed effects require larger planned groups under the stated calculation assumptions. Because the observed estimate is imprecise, planning directly from it would carry substantial uncertainty. Taken together, these features make EPIONE2 a useful example of the distinction between a numerical result, its statistical precision, and the additional design and analysis information needed to interpret it fully.

17. 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 main limitation for interpretation is the small number of observed responders: 2.0 in the tradipitant group and 4.0 in the placebo group. The resulting estimates have broad intervals, so the magnitude and direction of the underlying response difference remain uncertain. This limits the strength of conclusions that can be drawn from the primary outcome.

The registry lists 87 participants enrolled, while the primary outcome arm sizes are 42.0 and 43.0. The record does not explain the relationship between enrollment and those outcome denominators or whether any randomized participants lacked outcome data. The outcome description identifies all randomized participants as the population, but the record does not report exclusions, withdrawals, missing diary information, or how such cases were analyzed.

The registry status is terminated. The reason for termination, whether recruitment stopped early, and whether termination affected follow-up or outcome collection are not reported in the registry record. Without that context, the status cannot be used to explain the observed sample or result.

The available analysis reporting is limited: there are zero analysis entries and no posted secondary outcomes. The statistical test, interval construction method, confidence level, covariate adjustment, original sample-size justification, interim-analysis plan, and multiplicity strategy are not reported in the registry record. Consequently, the reported comparison cannot be fully checked against a documented analysis plan using the registry entries alone.

The primary outcome is a threshold-based symptom response at 2 weeks. The record does not provide individual scores, other time points, longer-term outcomes, or detail about diary completion. Safety reporting is also limited to the posted counts and listed serious-event terms, without treatment-relatedness or exposure-time detail. These constraints restrict the scope of interpretation to the reported endpoint and safety summaries.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Evaluating the Effects of Tradipitant vs. Placebo in Atopic Dermatitis (EPIONE2) — NCT04140695.
• Registry record last updated 2024-05-14; status last verified 2024-04.
• Results first posted 2024-05-14.
• 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.

21. Record Summary

EPIONE2 (NCT04140695) is a Phase 3 randomized, quadruple-masked, parallel-group study of tradipitant versus placebo in participants with atopic dermatitis. The registry lists Vanda Pharmaceuticals as sponsor, enrollment of 87, and a terminated status. The study start date is 2019-10-21, and the primary completion date is 2020-10-14.

The primary outcome is the Week 2 WI-NRS responder rate, defined as achieving at least 4 points reduction from baseline in weekly average diary score. The record reports 2.0 responders among 42.0 participants in the tradipitant arm and 4.0 among 43.0 in the placebo arm. The risk difference is -0.045, with an interval from -0.153 to 0.063; the unadjusted p-value is 0.414.

No secondary outcomes or analysis entries are posted. Serious adverse events are reported for 1 of 43 participants at risk with tradipitant and 1 of 44 with placebo; deaths are 0 in each group. The reason for termination and further statistical-analysis details are not reported in the registry record.