Single and Multiple Ascending Dose Study of AMG 133 in Participants With Obesity: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Single and Multiple Ascending Dose Study of AMG 133 in Participants With Obesity (NCT04478708), a randomized phase 1 study. The primary outcome is Number of Participants With Treatment-emergent Adverse Events (TEAEs).

2. Clinical Question

3. Trial Design

This was a Phase 1, randomized, double-blind, placebo-controlled study with quadruple masking recorded in the registry. Its model is listed as sequential. The study evaluated single ascending doses and multiple ascending doses of maridebart cafraglutide, also identified as AMG 133, in participants with obesity. The registry lists Amgen as sponsor and the study status as completed.

Part A included up to 7 single ascending dose cohorts, and Part B included up to 4 multiple ascending dose cohorts. Both parts had AMG 133 and placebo groups. Part C included up to 2 open-label multiple ascending dose cohorts treated with doses previously studied in Parts A and B; no placebo arm is listed for Part C. The posted results also separate some placebo groups by route or cohort grouping, so the observed groups should not be collapsed into a single comparison without a defined analysis.

The registry records a study start date of 2020-08-07 and a primary completion date of 2022-11-18. The primary endpoint timeframe varies by part and cohort: Part A is listed as 150 days; Part B as 207 days; Part C cohort 12 as 193 days; and Part C cohort 13 as 207 days. These are distinct follow-up windows, so crude event counts across cohorts do not necessarily represent equivalent observation periods.

The record describes a safety and tolerability aim alongside pharmacokinetic and pharmacodynamic assessment. The seven posted secondary outcomes include measures of maximum concentration, time to maximum concentration, area under the concentration-time curve, and anti-AMG 133 antibody formation. The registry record does not report the trial’s original sample-size justification or full statistical 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: This trial comprised three parts: Parts A and B were randomized, double-blind, placebo-controlled, single ascending dose (SAD) and multiple ascending dose (MAD) phases, respectively; and part C was an open-label, modified, MAD phase.

Only the first 6 of 17 groups are shown.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full analysis set: all randomized participants who received at least one dose of AMG 133 or placebo.

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.

Only the first 6 of 18 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants With Treatment-emergent Adverse Events (TEAEs), measured over Part A: 150 days; Part B: 207 days; Part C Cohort 12: 193 days; Part C Cohort 13: 207 days.

The primary endpoint is the number of participants with treatment-emergent adverse events, abbreviated TEAEs in the registry. It counts participants who experienced at least one qualifying event during the specified observation period, rather than counting every episode or measuring event severity as a single numerical score. The registry defines a qualifying event as an adverse event that started on or after the first dose of AMG 133. Clinically significant changes in laboratory safety tests, vital signs, and 12-lead electrocardiograms are included as adverse events under that definition.

The endpoint is reported separately for study groups and cohort-specific follow-up periods. Its stated population is the full analysis set: all randomized participants who received at least one dose of AMG 133 or placebo. This definition makes the endpoint relevant to describing observed safety events among treated randomized participants, while excluding randomized participants who received no dose from the specified population.

For a safety-focused early-phase study, a count of participants with events is a direct descriptive measure of how many people experienced at least one event. It does not, by itself, describe how many events each participant had, how long events lasted, or whether an event was caused by treatment. Nor does the endpoint alone establish comparative safety: interpretation depends on the groups being compared, their follow-up periods, and the precision of the estimates. Those details matter especially in this study because it used multiple cohorts, included both single-dose and multiple-dose parts, and had open-label cohorts in Part C.

7. Results Posted in the Registry

• Part B MAD Cohort 10 AMG 133: 80.0% (8/10) versus 25.0% (3/12) on control; risk difference 55.0 percentage points (95% CI 20.1 to 89.9), risk ratio 3.20 (1.14–8.94); unadjusted two-proportion z-test p = 0.010.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a participant count measure?

The primary outcome is a binary participant-level measure: a participant either had at least one TEAE during the defined period or did not. The posted number is therefore a count of participants with one or more events. It is not the number of event episodes, and it does not distinguish a participant with one event from a participant with several. General statistical explanation: binary outcomes are often compared by describing the event proportion in each group and then expressing the contrast as a risk difference, risk ratio, or odds ratio.

The registry’s outcome definition sets the event window relative to the first dose and includes clinically significant changes in laboratory safety tests, vital signs, and electrocardiograms. Different follow-up windows are reported for the study parts and Part C cohorts. General statistical explanation: when observation time differs, event proportions may reflect both event occurrence and duration of observation. Comparisons are easiest to interpret when groups have comparable follow-up and the same endpoint definition.

How should the reported comparison be read?

The posted comparison reports a risk difference of 0.55. The associated event counts are 8 among 10 in the Part B cohort 10 AMG 133 group and 3 among 12 in the comparison group. A risk difference expresses the absolute difference between the event proportions represented by those two groups. It is distinct from a risk ratio, which compares the proportions multiplicatively, and an odds ratio, which compares event odds.

The registry reports a risk ratio of 3.2 and an odds ratio of 12.0 alongside the risk difference. Because event outcomes are common in some of these small cohorts, the odds ratio should not be read as if it were the same quantity as a risk ratio. General statistical explanation: odds ratios and risk ratios can differ substantially when events are common, even when they point in the same direction. The comparison summary’s group counts should be used as posted; the record does not label the comparison group as the cohort 10 placebo arm. A separate cohort 10 placebo group is listed with 0 TEAEs among 3 participants.

What does the p-value convey?

The reported p-value is unadjusted and equals 0.01. General statistical explanation: a p-value measures how incompatible the observed data are with a specified null model, under the assumptions of the analysis. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of a difference. The registry record does not report the full model or testing procedure behind this comparison, so its assumptions and scope cannot be assessed from the posted summary alone.

Because the record lists seven secondary outcomes and a large set of cohort-specific outcome counts, the status of the comparison among all analyses matters. The registry record does not report a multiplicity adjustment or describe whether this comparison was prespecified. General statistical explanation: when many hypotheses are tested, the chance of at least one small p-value by chance can increase unless the testing strategy accounts for multiplicity. The unadjusted value should therefore be interpreted as one result, not as a multiplicity-controlled finding.

Why do confidence intervals matter?

A confidence interval describes a range of effect values compatible with the data under a stated statistical procedure. The registry reports intervals for the risk ratio and odds ratio, and an interval for the risk difference. The ratio intervals extend across a wide range of values, indicating considerable uncertainty about the magnitude of the relative contrast. General statistical explanation: interval width depends on information in the data, including the number of participants and events, as well as the method used to calculate the interval.

For these cohort sizes, a small change in the number of participants with events could materially alter an estimate. A confidence interval is not a guarantee that the true effect lies in the interval, and it does not address bias from cohort selection, differences in follow-up, or choice of comparator. The registry record does not state the interval method.

What can be said about analyses and populations?

The registry identifies the full analysis set for the primary endpoint as randomized participants who received at least one dose of AMG 133 or placebo. The posted counts are presented by study cohort and do not show an analysis of all randomized participants as one pooled group. The record lists no analyses in the summary details, and does not report missing-data handling, a detailed statistical model, interim analyses, or the original sample-size justification.

General statistical explanation: an analysis population defines which participants contribute to an estimate. Missing outcome information can affect interpretation if the reasons for missingness are related to treatment or event status. Without reported missing-data methods or participant-level disposition for this endpoint, the extent of that issue cannot be evaluated. Similarly, randomization and masking can support fair comparisons, but their protections do not remove uncertainty arising from small cohort sizes, multiple comparisons, or differences between study parts.

The study’s sequential structure is important context. General statistical explanation: dose-escalation studies commonly report results by cohort because dose and study progression are connected to cohort membership. Cohort-level event counts are useful descriptions, but they are not interchangeable with a single pooled estimate unless a suitable pooled analysis is reported. No such pooled primary-endpoint comparison is reported in the registry record.

10. Confidence Intervals and Precision

Precision is limited by the small numbers in the individual dose cohorts. The posted primary-endpoint groups range from 2 participants in the Part A intravenous placebo group to 10 in the Part B cohort 10 AMG 133 group. Such group sizes mean that each participant’s outcome can have a noticeable influence on the observed count and on an estimated contrast.

For the reported cohort 10 comparison, the risk difference is 0.55, with a reported confidence interval whose lower and upper bounds are 0.202 and 0.898. The interval is broad, so the estimate does not pinpoint the size of the absolute difference precisely. The risk ratio is 3.2, with an interval from 1.145 to 8.944; the odds ratio is 12.0, with an interval from 1.581 to 91.08. These ranges likewise allow substantial uncertainty about the magnitude of the relative association.

General statistical explanation: confidence intervals are useful because they show how much uncertainty surrounds an estimate, not just whether an estimate differs from a null value. Wide intervals are common when the amount of information is limited. They may include a range of effects that would have different practical implications. An interval also reflects only the statistical method and data used; it does not account automatically for bias, unreported analyses, or limitations in endpoint ascertainment.

The registry record does not report the confidence-interval method or a complete analysis plan. The reported values should therefore be interpreted as registry estimates rather than independently verified calculations. The unadjusted p-value of 0.01 does not remove the uncertainty reflected in the intervals, nor does it address the multiplicity of outcomes or comparisons.

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 planning display expresses sample sizes as participants per arm, not as total events. Its assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect. These are planning assumptions for an illustration based on the observed comparison, not the assumptions documented for the original trial design.

For planning at 100% of the observed effect, the display gives 12 participants per arm for 80% power and 16 participants per arm for 90% power. For planning at 75% of the observed effect, it gives 22 participants per arm for 80% power and 29 participants per arm for 90% power. For planning at 50% of the observed effect, it gives 49 participants per arm for 80% power and 64 participants per arm for 90% power.

In plain language, the 80% power column describes the planned number of participants per arm under which the specified test would have an 80% chance of detecting the chosen effect if that effect were the true one and the assumptions held. The 90% column uses a higher target chance of detection, which requires more participants per arm for each listed effect. The table illustrates the general pattern that smaller effects need more participants to detect reliably.

These values are not observed power and should not be read as evidence that the completed study had a particular probability of detecting an effect. The trial’s own original sample-size justification is not reported in the registry record. The planning values are illustrative, use the observed effect as their starting point, and do not account for every feature of this sequential, multi-cohort study.

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 = 0.250; treatment proportion p₁ = 0.800; average p̄ = 0.525.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.7062 = 1.3842.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5895 = 0.4961.
4. n = (1.3842 + 0.4961)² / (0.5500)² = 11.7, rounded up to 12 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 <- 0.25                          # control proportion
p1 <- 0.8                          # 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(4478708)
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 = 12 (80% power), 16 (90% power)
 75% of observed effect: n per arm = 22 (80% power), 29 (90% power)
 50% of observed effect: n per arm = 49 (80% power), 64 (90% power)
power.prop.test: 11.7 
Simulated power at n = 12 per arm: 0.850 (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 lists seven secondary outcomes, in addition to the primary TEAE endpoint. Those secondary outcomes include concentration and exposure measures, time to maximum concentration, and the percentage of participants with anti-AMG 133 antibody formation. The posted results also contain cohort-specific primary-endpoint counts across multiple study parts. A single unadjusted comparison is reported for the Part B cohort 10 comparison.

The registry record does not report whether the comparison was prespecified, how it was selected from other possible comparisons, or whether any multiplicity adjustment was planned or applied. It also does not report a hierarchy for the primary and secondary outcomes. General statistical explanation: when multiple outcomes, doses, or subgroup comparisons are examined, the probability of observing at least one apparently unusual result can rise. A prespecified testing strategy or multiplicity adjustment can help control that issue; without a reported strategy, an unadjusted p-value should be treated cautiously.

The endpoint population is defined as randomized participants who received at least one dose of AMG 133 or placebo. The record does not report a detailed missing-data method for the primary endpoint, nor does it describe how participants with unavailable outcome information were handled. It does not provide enough information to evaluate whether any missing outcomes could differ systematically across groups.

Although the study is described as randomized and double-blind, Part C is explicitly open-label. The record does not report a detailed analysis plan explaining how Part C was handled in the primary analysis, and Part C has no placebo group listed. The posted cohort counts should therefore be read as separate descriptive results unless an analysis combining them is reported. A pooled estimate, interim-analysis plan, and full statistical analysis plan are not reported in the registry record.

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 5% reporting threshold.

Most frequent non-serious adverse event terms

The serious adverse event display reports zero participants with serious adverse events in every listed group. The death count is also zero in every listed group. The registry’s serious-event listing contains no top serious event terms. These entries describe the counts recorded for the listed cohorts and their follow-up periods; they are not evidence that serious events are impossible or that risk is absent.

General statistical explanation: zero observed events in a small group still leaves uncertainty about the frequency of uncommon events. The study has small cohort groups, and the follow-up windows vary by part and cohort. The serious adverse event counts are therefore best interpreted as descriptive safety observations from this study, not as precise estimates of rare-event risk.

The primary TEAE outcome is broader than serious adverse events. It includes participants with any qualifying event under the registry definition, including clinically significant changes in laboratory safety tests, vital signs, or electrocardiograms. Thus, a group can have TEAEs while reporting no serious adverse events. The displayed counts should not be combined across cohorts or compared without accounting for the different cohort structures and observation periods.

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

The results describe safety events observed among participants assigned to the study cohorts and included in the stated analysis population. They show that TEAEs were recorded in both AMG 133 and placebo groups, with counts differing across cohorts. The one reported inferential comparison has a risk difference of 0.55 and an unadjusted p-value of 0.01, based on the posted counts of 8 among 10 and 3 among 12. Its confidence intervals, particularly for the relative measures, leave substantial uncertainty about the size of the contrast.

These findings do not establish that AMG 133 caused the observed events or that the same pattern would occur in a different group of participants. A reported difference in event counts is an association within the comparison as posted; causal interpretation depends on the exact comparison, design implementation, outcome ascertainment, and analysis. The registry summary does not identify the 3-of-12 comparison group as the cohort 10 placebo arm, while the separately listed cohort 10 placebo group has 0 events among 3. That distinction should be preserved when describing the result.

The findings also do not establish a single safety profile across every dose and study part. The study has multiple ascending-dose cohorts, single-dose cohorts, and open-label Part C cohorts, with different follow-up durations and differing control arrangements. The record does not report a pooled analysis that would support treating all these groups as one estimate.

Zero serious adverse events and zero deaths in the listed groups are reassuring only as observations within those limited groups and follow-up windows; they do not rule out uncommon harms. Likewise, the posted TEAE counts do not describe event severity, recurrence, or duration. No medical conclusion about an individual’s treatment can be drawn from these registry-level summaries.

The study is an early-phase evaluation with a stated safety and tolerability focus and additional pharmacokinetic and pharmacodynamic outcomes. Its data are useful for describing the observed cohort results and for generating questions for further evaluation. They are not, on their own, a definitive answer about comparative safety across broader populations or clinical use.

16. Why This Trial Matters Statistically

This trial illustrates why clinical-trial results should be interpreted at the level at which they were collected and analyzed. The registry presents many small cohorts arranged by dose and study part, rather than one simple two-group comparison. Distinguishing single-dose from multiple-dose cohorts, placebo groupings, and open-label cohorts is necessary to avoid treating unlike groups as interchangeable.

It also demonstrates the difference between a participant count, an effect estimate, and a measure of uncertainty. The primary endpoint counts people with at least one TEAE. The posted risk difference, risk ratio, and odds ratio express the cohort 10 comparison in different ways, while their confidence intervals show that the magnitude remains uncertain. A p-value alone cannot convey that uncertainty or resolve questions about the comparison group.

The record’s unadjusted comparison is especially instructive because seven secondary outcomes are listed and results are reported across multiple cohorts. Without a reported multiplicity strategy or complete analysis plan, a nominal p-value should not be detached from the broader analysis context. Statistical evidence is shaped not only by the numerical result but also by which outcomes and comparisons were planned and how they were selected.

Finally, the study highlights the limits of interpreting zero serious adverse event counts in small groups. The counts are important observations, but they cannot precisely characterize rare events. A careful reading keeps descriptive safety information, inferential comparison, and planning illustrations separate. This distinction helps prevent overstatement while preserving the statistical information that the registry does report.

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 primary-endpoint results are presented by cohort, and individual groups are small. A count in one cohort may therefore be sensitive to the outcome of a single participant. The registry also reports different observation periods across parts and Part C cohorts, which makes unqualified comparisons across all groups difficult.

The reported statistical comparison does not identify the 3-of-12 comparison group by a cohort label in the comparison summary. The separate cohort 10 placebo group is reported as 0 among 3, so those summaries should not be treated as the same comparator. The registry record does not report a detailed explanation of the selected comparison or its place in the analysis plan.

The study includes open-label Part C cohorts, while the randomized, masked portions include placebo groups. The open-label groups do not have a listed placebo comparator. The registry entries summarized do not report a pooled analysis, missing-data method, interim-analysis plan, multiplicity procedure, or full statistical analysis plan. The trial’s original sample-size justification is also not reported in the registry record.

The posted endpoint is a count of participants with at least one TEAE; it does not communicate the number of events, duration, recurrence, or severity beyond the separately reported serious-event counts. The listed secondary outcomes are not accompanied by posted numerical results. Finally, registry-level summaries cannot establish event causation, and the small group sizes limit the ability to characterize uncommon safety events.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Single and Multiple Ascending Dose Study of AMG 133 in Participants With Obesity — NCT04478708.
• Registry record last updated 2025-12-10; status last verified 2025-11.
• Results first posted 2025-12-10.
• 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

NCT04478708 is a completed Phase 1 randomized study of maridebart cafraglutide, also called AMG 133, in participants with obesity. It evaluated single ascending doses in Part A, multiple ascending doses in Part B, and open-label multiple ascending doses in Part C. The registry lists enrollment of 110, quadruple masking, and placebo-controlled groups in Parts A and B.

The primary endpoint was the number of participants with TEAEs. Counts vary across the listed cohorts, and the registry reports one comparison for Part B cohort 10 with a risk difference of 0.55 and an unadjusted p-value of 0.01. Serious adverse event counts and death counts are zero in every listed group. The cohort-level results, broad uncertainty intervals, varied follow-up, and unreported analysis details limit conclusions beyond the posted descriptive findings.