A Study to Evaluate Rocatinlimab (AMG 451) in Adolescent Participants With Moderate-to-severe Atopic Dermatitis (AD): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Evaluate Rocatinlimab (AMG 451) in Adolescent Participants With Moderate-to-severe Atopic Dermatitis (AD) (NCT05704738), a randomized phase 3 study. The primary outcome is Number of Participants Who Achieved vIGA-AD 1 Response With Presence of Only Barely Perceptible Erythema or vIGA-AD 0 Response (Revised Investigator's Global Assessment [rIGA] 0/1) at Week 24.

2. Clinical Question

3. Trial Design

This was a randomized, double-blind, phase 3 study of rocatinlimab in adolescents with moderate-to-severe atopic dermatitis. The record describes a sequential model and reports enrollment of 532. The study evaluated rocatinlimab as monotherapy and in combination with topical corticosteroids or topical calcineurin inhibitors. The registry record does not report how participants were distributed between those treatment cohorts.

The initial period ran from Week 0 to Week 24. Participants received placebo, rocatinlimab 150mg, or rocatinlimab 300mg, each on an every-4-weeks schedule with a loading dose at Week 2. In the combination-therapy cohort, topical treatment could be used alongside the assigned study treatment. The primary endpoint was assessed at Week 24, at the end of this initial period.

The maintenance period ran from Week 24 to Week 52 and involved treatment reassignment based on response status and initial assignment. The record describes responders in some groups as re-randomized or reassigned to continued treatment schedules, including every 4 weeks or every 8 weeks. Nonresponders could be reassigned to open-label rocatinlimab. These later treatment paths mean the Week 24 primary comparison should be distinguished from maintenance-period comparisons: the latter follow a different treatment structure and selected response groups.

The study is listed as completed. Its start date is 2023-04-20 and its primary completion date is 2025-02-27. The registry record does not report the trial’s original sample-size justification, detailed randomization procedure, stratification factors, or interim-analysis plan.

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: Based on achieving validated investigator's global assessment for AD score of 0 (clear) or 1 (almost clear) with a ≥ 2-point reduction from baseline (vIGA-AD 0/1) or achievement of ≥ 75% reduction from baseline in eczema area and severity index score (EASI 75) at Week 24, and whether therapies consistent with AD rescue were used, participants were assigned to either maintenance period or open-label period.

Only the first 6 of 16 groups are shown.

Initiation Period (Baseline - Week 24)

Reasons for not completing this period:

Maintenance Period (Week 24 - Week 52)

Reasons for not completing this period:

OLRT Period (Week 25 - Week 52)

Reasons for not completing this period:

Open-label Period (Week 24 - Week 52)

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: FAS: All randomized participants. Participants were analyzed according to randomized treatment.

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.

Population: FAS: All randomized participants. Participants were analyzed according to the randomized treatment.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants Who Achieved vIGA-AD 1 Response With Presence of Only Barely Perceptible Erythema or vIGA-AD 0 Response (Revised Investigator's Global Assessment [rIGA] 0/1) at Week 24, measured over Baseline and Week 24.

The primary endpoint counts participants who achieved a revised investigator’s global assessment (rIGA) response at Week 24. The underlying vIGA-AD assessment rates atopic dermatitis severity on a 5-point scale, from 0, clear, to 4, severe; higher scores indicate greater severity. A score of 0 or 1 is not sufficient by itself in every case: a participant with a score of 1 also had to meet the specified barely perceptible erythema and other skin-feature criteria. A score of 0 counted as a response. The endpoint description also specifies a reduction of at least 2 points from baseline for the clear or almost-clear assessment.

This kind of global assessment is used to turn a clinician-rated severity measure into a binary outcome: each participant either meets the response definition or does not. The resulting count can be compared between randomized groups. It is an interpretable endpoint for describing how many participants crossed the prespecified clinical-assessment threshold, but it does not capture every aspect of atopic dermatitis or the full range of change within each response category.

Participants who initiated rescue therapy for atopic dermatitis were classified as nonresponders at subsequent time points. The registry record does not report additional detail about how other missing Week 24 assessments were handled. The endpoint is therefore best read according to its stated response and rescue rules, with the limits of unreported missing-data procedures kept in view.

7. Results Posted in the Registry

• Initiation Period: Roca 150mg Q4W ± TCS/TCI: 35.4% (56/158) versus 10.8% (17/157) on control; risk difference 24.6 percentage points (95% CI 15.7 to 33.5), risk ratio 3.27 (1.99–5.37); unadjusted two-proportion z-test p = <0.001.
• Initiation Period: Roca 300mg Q4W ± TCS/TCI: 36.4% (79/217) versus 10.8% (17/157) on control; risk difference 25.6 percentage points (95% CI 17.5 to 33.6), risk ratio 3.36 (2.08–5.44); 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.

Number of Participants Who Achieved EASI 75 at Week 16

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Initiation Period: Roca 150mg Q4W ± TCS/TCI: 52.5% (83/158) versus 27.4% (43/157) on control; risk difference 25.1 percentage points (95% CI 14.7 to 35.6), risk ratio 1.92 (1.43–2.58); unadjusted two-proportion z-test p = <0.001.
  • Initiation Period: Roca 300mg Q4W ± TCS/TCI: 52.1% (113/217) versus 27.4% (43/157) on control; risk difference 24.7 percentage points (95% CI 15.0 to 34.3), risk ratio 1.90 (1.43–2.53); 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 Initiation Period: Roca 150mg Q4W ± TCS/TCI than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Initiation Period: Roca 300mg Q4W ± TCS/TCI than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Number of Participants Who Achieved vIGA-AD 0/1 at Week 16

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Initiation Period: Roca 150mg Q4W ± TCS/TCI: 21.5% (34/158) versus 12.7% (20/157) on control; risk difference 8.8 percentage points (95% CI 0.5 to 17.0), risk ratio 1.69 (1.02–2.80); unadjusted two-proportion z-test p = 0.039.
  • Initiation Period: Roca 300mg Q4W ± TCS/TCI: 27.6% (60/217) versus 12.7% (20/157) on control; risk difference 14.9 percentage points (95% CI 7.0 to 22.8), risk ratio 2.17 (1.37–3.45); 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 Initiation Period: Roca 150mg Q4W ± TCS/TCI than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Initiation Period: Roca 300mg Q4W ± TCS/TCI than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Number of Participants Who Achieved a ≥ 4-point Reduction From Baseline in Weekly Average of Daily Worst Pruritus Numeric Rating Scale (NRS) Score at Week 16 in Participants With Baseline Weekly Average of Daily Worst Pruritus NRS Score ≥ 4

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Initiation Period: Roca 150mg Q4W ± TCS/TCI: 27.0% (38/141) versus 13.8% (20/145) on control; risk difference 13.2 percentage points (95% CI 3.9 to 22.4), risk ratio 1.95 (1.20–3.19); unadjusted two-proportion z-test p = 0.006.
  • Initiation Period: Roca 300mg Q4W ± TCS/TCI: 30.8% (60/195) versus 13.8% (20/145) on control; risk difference 17.0 percentage points (95% CI 8.4 to 25.5), risk ratio 2.23 (1.41–3.53); 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 Initiation Period: Roca 150mg Q4W ± TCS/TCI than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Initiation Period: Roca 300mg Q4W ± TCS/TCI than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The available secondary results cover several related aspects of atopic dermatitis response: EASI 75, vIGA-AD 0/1, and a clinically defined reduction in worst-pruritus score among participants whose baseline score met the stated threshold. Across these listed Week 16 outcomes, the reported response counts are higher in both rocatinlimab groups than in placebo. The sponsor’s comparisons report p-values below 0.05 for each of these available comparisons, with the exact reported values varying by outcome and dose.

This general pattern is directionally consistent with the primary Week 24 finding: the primary outcome and the listed secondary outcomes all compare response frequencies between initial-period treatment groups. They are not interchangeable measures, however. The global assessment, EASI response, and symptom-score reduction capture different features of disease status and have different response definitions and time points. Agreement in direction across measures can add context, but it does not make them equivalent or establish a result beyond the outcomes actually assessed.

The record lists additional secondary outcomes, while detailed results in the analysis materials are available for only a subset. An overall claim about every secondary endpoint would therefore go beyond the results described. Secondary outcomes also carry less evidential weight than the primary endpoint when their multiplicity handling is not reported. Testing many outcomes can increase the chance of nominally small p-values, and the registry record does not say whether adjustment or a prespecified testing sequence was used. The secondary findings are best treated as supportive, outcome-specific evidence whose interpretation depends on that unresolved multiplicity context.

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

• Number of Participants Who Achieved EASI 75 at Week 16
• Number of Participants Who Achieved vIGA-AD 0/1 at Week 16
• Number of Participants Who Achieved a ≥ 4-point Reduction From Baseline in Weekly Average of Daily Worst Pruritus Numeric Rating Scale (NRS) Score at Week 16 in Participants With Baseline Weekly Average of Daily Worst Pruritus NRS Score ≥ 4
• Number of Participants Who Achieved vIGA-AD 0/1 at Week 24
• Number of Participants Who Achieved a ≥ 4-point Reduction From Baseline in Weekly Average of Daily Worst Pruritus NRS Score at Week 24 in Participants With Baseline Weekly Average of Daily Worst Pruritus NRS Score ≥ 4
• Number of Participants Who Achieved ≥ 90% Reduction From Baseline in EASI Score (EASI 90) at Week 24
• Number of Participants Who Achieved ≥ 4-point Reduction From Baseline in Weekly Average of AD Skin Pain NRS Score at Week 24 in Participants With Baseline Weekly Average of AD Skin Pain NRS Score ≥ 4
• Percentage of Participants Who Initiated Rescue Therapy for AD at or Before Week 16

9. Statistical Interpretation

10. Statistical Methods Explained

For a binary endpoint, each participant is classified as meeting or not meeting the response definition. The basic group summaries are counts and denominators. A risk difference compares the response probabilities between two groups by subtracting one probability from the other. It is expressed as a difference in probability, not as a ratio. A value above zero means the response probability is higher in the first named treatment group than in the comparator under that contrast; a value below zero would indicate the reverse.

What does the confidence interval describe?

A confidence interval describes uncertainty around an estimated effect under the assumptions of the analysis. For the sponsor’s primary analyses, the registry reports 95% confidence intervals for the common risk differences. The lower and upper limits indicate a range of effect values compatible with the analysis and observed data at that confidence level. The interval is not a range containing a fixed proportion of individual participant outcomes, and it does not remove uncertainty about the trial’s design, conduct, or applicability.

The unadjusted re-calculations use the posted response counts and group sizes to estimate simple risk differences. These calculations do not reproduce a stratified or adjusted sponsor analysis. They are useful as a transparent check of the direction and approximate magnitude of the group contrast, but they are not a substitute for the sponsor’s prespecified analysis. The registry record does not report the factors, if any, used to define strata in the sponsor’s Cochran-Mantel-Haenszel analysis.

Why use a Cochran-Mantel-Haenszel analysis?

In general, the Cochran-Mantel-Haenszel method combines information across strata to estimate a common association or effect while accounting for the stratification structure. This can be useful when a trial’s analysis is designed to account for groups or factors within the randomized comparison. The particular stratification factors and details of the sponsor’s model are not reported in the registry record, so the precise adjustment cannot be reconstructed from the record alone. The sponsor-reported estimates should therefore be read as the official reported analysis, with the simple unadjusted recalculations treated as a separate descriptive check.

A p-value measures how unusual data at least as incompatible with a specified null hypothesis would be if that null hypothesis and the analysis assumptions held. Here, the sponsor’s primary comparison p-values are reported as < 0.001. This is not the probability that the null hypothesis is true, nor does it measure the size or importance of the effect. The effect estimate and confidence interval are needed to describe magnitude and precision alongside the p-value.

The analysis population is described as the full analysis set, consisting of all randomized participants analyzed according to randomized treatment. This approach preserves the randomized group assignment for the stated primary analysis. The endpoint description further specifies that rescue-therapy initiation led to classification as a nonresponse at subsequent time points. Other missing-data methods, such as the handling of unavailable assessments not related to rescue therapy, are not reported in the registry record.

Because the primary outcome is binary, methods for continuous outcomes, such as comparing group means, do not describe this endpoint directly. Counts and response probabilities are the relevant quantities. The registry record reports the endpoint as number of participants, while the comparative analyses report risk differences. Those two forms of reporting answer related but distinct questions: the count describes how many participants met the definition, while the risk difference describes the contrast in response probability between groups.

The registry lists numerous outcome measures and analyses, but the multiplicity adjustment procedure is not reported in the registry record. When many endpoints or comparisons are assessed, the chance of observing at least one small p-value by chance can increase. Whether and how that concern was addressed in the trial’s statistical plan is unknown from the record. Primary and secondary results should therefore not be assigned equal evidential weight solely because they have p-values.

11. Confidence Intervals and Precision

The primary sponsor-reported 95% confidence intervals quantify uncertainty in the Week 24 common risk differences. For the 150mg comparison, the interval runs from 15.6 to 33.2 percentage points; for the 300mg comparison, it runs from 17.8 to 33.7 percentage points. Both intervals are entirely above zero, consistent with the reported p-values below 0.001. The interval limits also show that the effect is not known as a single exact value: a range of positive differences remains compatible with each analysis.

The re-calculated, unadjusted intervals are reported on the probability-difference scale: 0.157 to 0.335 for the 150mg group and 0.175 to 0.336 for the 300mg group. These intervals are not identical to the sponsor’s intervals because the analyses differ. It is important not to interpret small differences between the estimates as proof that one dose is superior to the other. The registry reports comparisons of each dose with placebo, not a direct randomized comparison between the two active doses.

Confidence intervals describe statistical precision conditional on the analysis and data. They do not account for every possible source of uncertainty, such as unreported analysis choices, missing-data handling, or generalizability beyond the trial population. Nor do they show the distribution of individual outcomes. The number of participants in each group and the number meeting the response definition help contextualize the estimates, but the main uncertainty summary is the interval around the comparative effect.

For the secondary outcomes, confidence intervals are likewise relevant to judging precision, but their interpretation also depends on the endpoint definitions and multiplicity context. The registry record does not report a multiplicity adjustment. The reported intervals should therefore be considered as estimates for their individual comparisons rather than as a simultaneous guarantee across all assessed outcomes.

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 participants per arm for the binary outcomes in the sample-size illustration. They are not total event counts. The illustration uses a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The target effect is set to a stated fraction of the observed effect. These are planning assumptions for an illustration based on observed results; the trial’s original design assumptions and sample-size justification are not reported in the registry record.

The 80% power column gives the illustrative number of participants per arm needed for an 80% chance of detecting the chosen effect under those assumptions, if the assumed effect and model are correct. The 90% power column gives the corresponding number per arm for a higher chance of detection. Neither column describes what happened in the completed trial. These figures are not observed power.

Rows planning for 100% of the observed effect use the observed estimate as the target effect. Rows planning for 75% or 50% use smaller effects than that estimate. Smaller effects are harder to distinguish from random variation, so they generally require more participants per arm under the same planning assumptions. The figures are conditional illustrations, not recommendations or a reconstruction of the original protocol’s calculations.

The curve should be read as a visual representation of how the illustrative sample requirement changes as the target effect changes. It does not add information about the trial’s actual design process. In particular, the registry record does not report whether the original study used other assumptions, allowances, or design considerations. The values are most useful for understanding the relationship between target effect, desired power, and sample size under the stated simplified planning setup.

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 = 0.108; treatment proportion p₁ = 0.354; average p̄ = 0.231.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.5964 = 1.1689.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5704 = 0.4801.
4. n = (1.1689 + 0.4801)² / (0.2462)² = 44.9, rounded up to 45 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 <- 0.10828025477707                          # control proportion
p1 <- 0.354430379746835                          # 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(5704738)
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 = 45 (80% power), 60 (90% power)
 75% of observed effect: n per arm = 73 (80% power), 97 (90% power)
 50% of observed effect: n per arm = 145 (80% power), 194 (90% power)
power.prop.test: 44.9 
Simulated power at n = 45 per arm: 0.826 (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 reports sponsor analyses using the Cochran-Mantel-Haenszel method. For the primary endpoint, those analyses give a common risk difference of 24.7 percentage points for rocatinlimab 150mg every 4 weeks versus placebo and 26.1 percentage points for rocatinlimab 300mg every 4 weeks versus placebo. The corresponding reported 95% confidence intervals are 15.6 to 33.2 and 17.8 to 33.7, and both p-values are < 0.001.

The posted re-calculations are simple, unadjusted risk differences from the group counts. They give estimates of 0.246 and 0.256, with intervals of 0.157 to 0.335 and 0.175 to 0.336. The difference in presentation reflects that the sponsor’s reported common risk difference comes from a Cochran-Mantel-Haenszel analysis, while the recalculation does not apply that method. The registry record does not report enough detail about the sponsor’s stratification structure to reconstruct its analysis fully.

The registry record does not flag a discrepancy between the sponsor-reported results and the recalculated summaries. The sponsor’s analysis remains the primary reported analysis; the unadjusted calculations are a supplementary check, not a replacement or an alternative official result. Both analyses point in the same direction, but their estimates and intervals should be kept distinct rather than blended.

15. Multiplicity, Missing Data and Analysis Populations

The registry record reports a primary endpoint, numerous secondary outcome measures, and multiple analyses. It does not report a multiplicity-control procedure, such as an adjustment method or a hierarchy for testing outcomes. Without that information, it is not possible to determine from the registry record how the trial’s formal statistical plan addressed the increased chance of false-positive findings when many comparisons are evaluated. The primary outcome should be distinguished from secondary findings, and nominal p-values for secondary outcomes should not automatically be read as confirmatory evidence.

The primary analysis population is described as the full analysis set: all randomized participants analyzed according to randomized treatment. The endpoint rules specify that participants initiating rescue therapy for atopic dermatitis were classified as nonresponders at subsequent time points. The registry record does not report how other missing Week 24 observations were handled, whether imputation was used, or whether sensitivity analyses assessed alternative assumptions about missing outcomes.

The record also does not report an interim-analysis plan or whether any interim efficacy analyses were conducted. This is not evidence that interim analyses did or did not occur; the information is not reported in the registry record. Likewise, details of the original statistical analysis plan, the trial’s sample-size justification, and the factors used in the Cochran-Mantel-Haenszel analysis are not reported. These omissions limit independent reconstruction of the sponsor’s analysis and assessment of all prespecified decision rules.

The explicit rescue rule is important because it defines how rescue treatment affects later endpoint classification. It does not resolve the separate question of how an unavailable assessment for another reason was treated. Interpretation should follow the stated analysis population and response rules while recognizing that missing-data and multiplicity details remain unreported.

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

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: appendicitis perforated (3.0%), gastritis (2.9%), wrist fracture (2.3%), upper respiratory tract infection (2.1%), inflammatory myofibroblastic tumour (1.9%).

The registry lists serious adverse event counts by treatment period and group. In the initial period, serious adverse events were reported for 6 participants in the placebo group, 2 in the rocatinlimab 150mg group, and 4 in the rocatinlimab 300mg group. The corresponding at-risk group sizes were 157, 156, and 216. The record lists zero deaths in each of these groups.

For maintenance and later open-label periods, the record gives additional group-specific counts and denominators. Those groups reflect treatment reassignment, response status, or treatment history, so they are not equivalent to the original randomized comparisons. Later-period figures should be read within their own period and group context rather than combined informally with the initial-period results.

Counts of participants with at least one serious adverse event are not the same as counts of all events, and they do not establish whether an event was caused by study treatment. The registry record lists selected serious-event terms, but the displayed group summaries alone do not provide a complete safety evaluation, comparative risk analysis, or causal assessment. Zero reported deaths in the listed groups is a descriptive observation within those recorded groups, not proof of absence of risk.

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

The primary results show that, in this trial’s Week 24 analysis, the sponsor-reported response probability was higher in each rocatinlimab initial-period group than in the placebo group under the specified rIGA 0/1 definition. The reported common risk differences are positive, their 95% confidence intervals are above zero, and the reported p-values are < 0.001. The raw response counts and denominators provide descriptive context for those comparative estimates.

That conclusion is limited to the trial’s participants, endpoint definition, time point, and randomized initial-period comparisons. It does not establish what an individual participant will experience, how the result would compare with treatments not studied, or whether differences persist beyond the reported endpoint. It also does not demonstrate that one rocatinlimab dose is better than the other: the reported primary analyses compare each dose with placebo, not the doses directly with each other.

A statistically distinguishable group difference is not the same as a complete judgment of clinical importance. The endpoint records a threshold response, and participants who do not meet that threshold may still have had changes that the binary endpoint does not capture. Conversely, meeting the endpoint does not describe every symptom, quality-of-life outcome, or safety consideration. Those broader outcomes cannot be inferred from the primary response count alone.

The results do not establish a treatment recommendation or prove safety. Adverse-event counts are descriptive, and the record does not provide a causal interpretation of individual events. Generalization beyond adolescents with the trial’s moderate-to-severe atopic dermatitis population, or beyond the treatment regimens and follow-up reported, is not established by these data. Missing-data procedures, multiplicity control, and several other analysis-plan details are not reported in the registry record, which further constrains interpretation.

18. Why This Trial Matters Statistically

This trial illustrates how a binary clinical endpoint can make a randomized comparison interpretable while still requiring careful statistical reading. The response definition combines a clinician-rated severity scale with additional criteria for a score of 1, and the rescue-therapy rule affects subsequent classification. Understanding those operational details is essential: the endpoint is not simply any improvement or any low severity score.

The registry reports both raw counts and comparative risk differences. Counts alone are not enough when group sizes differ. The comparative estimates give a more direct summary of the difference in response probability, while confidence intervals show the uncertainty around those estimates. Reporting the sponsor’s stratified method alongside simple unadjusted recalculations also illustrates why analyses based on the same posted counts need not produce identical estimates.

The maintenance design adds another statistical lesson. Week 24 responders and nonresponders follow different paths after the initial period, and some responders are re-randomized or reassigned. The initial randomized comparison therefore answers a different question from later-period summaries. Later treatment groups should not be treated as though they retain the same simple randomized comparison structure as the initial groups.

Finally, the record lists many outcomes but does not report multiplicity handling, missing-data details, or the original sample-size justification. Those omissions matter when judging how confirmatory secondary findings are and how fully the analysis can be reproduced. The trial is therefore useful for discussing not just effect estimates, but also endpoint definitions, analysis populations, uncertainty, and the limits of registry reporting.

19. Limitations

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

The registry record does not report the original sample-size justification, so the planning illustration based on observed effects cannot be taken as a reconstruction of the trial’s design. Nor does the record report a multiplicity procedure, interim-analysis plan, or detailed statistical analysis plan. These omissions limit assessment of the complete confirmatory testing strategy.

The record identifies the full analysis set as all randomized participants analyzed according to randomized treatment, and it specifies that rescue-therapy initiation results in nonresponse classification at subsequent time points. Other missing-data handling is not reported. Without those details, the potential influence of unavailable assessments cannot be evaluated from the registry record.

The primary sponsor analysis uses the Cochran-Mantel-Haenszel method, but the analysis strata or other relevant implementation details are not reported. The unadjusted recalculations are consequently useful only as simple comparisons from posted summaries; they cannot reproduce every feature of the sponsor’s model.

The study includes combination-therapy cohorts and a maintenance period with re-randomization or reassignment based on initial-period response. The available initial-period comparisons should not be generalized to every later treatment path. The registry summary does not report all subgroup-specific details needed to assess whether effects differ across monotherapy and combination-therapy cohorts.

Finally, the primary endpoint is a threshold-based clinician assessment at a specified time point. It does not encompass every disease feature, symptom, or long-term outcome. The adverse-event listings are also descriptive and do not by themselves establish causal relationships or provide a full comparative safety assessment. These are limits on what can be concluded from the reported record, rather than evidence that an unreported analysis had a particular result.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study to Evaluate Rocatinlimab (AMG 451) in Adolescent Participants With Moderate-to-severe Atopic Dermatitis (AD) — NCT05704738.
• Registry record last updated 2026-08-28; status last verified 2026-07.
• Results first posted 2026-08-28.
• 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

NCT05704738 is a completed, randomized, double-blind phase 3 trial of rocatinlimab in adolescent participants with moderate-to-severe atopic dermatitis. The initial-period groups received placebo, rocatinlimab 150mg every 4 weeks, or rocatinlimab 300mg every 4 weeks, with topical therapy in the combination-therapy cohort. The primary endpoint was the Week 24 rIGA 0/1 response.

The sponsor’s reported comparisons favored each rocatinlimab group over placebo on the primary endpoint, with positive common risk differences and p-values < 0.001. Selected Week 16 secondary outcomes show a broadly similar direction, but multiplicity handling is not reported. The record also describes a maintenance period with treatment reassignment, lists serious adverse-event counts, and leaves several analysis details—including the original sample-size rationale and non-rescue missing-data handling—not reported in the registry record.