Treadmill Cardiovascular Safety Study of Erenumab (AMG 334): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Treadmill Cardiovascular Safety Study of Erenumab (AMG 334) (NCT02575833), a randomized phase 2 study. The primary outcome is Change From Baseline in Total Exercise Time.

2. Clinical Question

The primary comparison concerns a continuous outcome: the time participants exercised during the treadmill test. The registry reports group least-squares means for change from baseline and a treatment difference. The posted treatment difference is negative, indicating that the estimated change was lower in the erenumab group than in the placebo group for this comparison. The uncertainty around that estimate is substantial, and the registry also reports a sponsor analysis framed as a non-inferiority analysis.

Time to onset of exercise-induced angina and time to onset of at least 1 mm ST-segment depression are listed as secondary outcomes. They are distinct from the primary exercise-time question. The record does not report results for those secondary outcomes in the registry record for this analysis, so their findings cannot be described.

3. Trial Design

This was a phase 2a, multicenter, randomized, double-blind, placebo-controlled study of erenumab, also identified as AMG 334, in adults with stable angina. The registry lists a parallel-group model and quadruple masking. These design features describe how the trial was organized; the registry record does not give further details about the masking roles or allocation procedures.

Participants received a single dose by intravenous infusion on day 1: erenumab 140 mg or placebo. The primary outcome was assessed at baseline and on day 1 after dosing. The registry describes enrollment as 89. For the primary outcome, the posted group counts are 44 for placebo and 42 for erenumab. The outcome population is defined as participants who received study drug and completed both the baseline and post-randomization treadmill tests. The registry record does not report why the outcome counts differ from each other or from the enrollment figure.

The study is marked completed. Its start date is 2015-11-23 and its primary completion date is 2017-01-23. The sponsor is Amgen. The stated objective was to evaluate the effect of erenumab compared with placebo on exercise time during an exercise treadmill test. The record lists three outcome measures and nine analyses, but the registry entries available for this explanation include the primary outcome results and two secondary outcome names, not results for those secondary outcomes.

Further details about the original sample-size justification, planned analysis populations, interim analyses, and allocation concealment are not reported in the registry record. Those omissions limit how fully the design and analysis decisions can be evaluated.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Participants were randomized in a 1:1 ratio to receive either a single dose of erenumab 140 mg or placebo intravenously (IV) prior to starting an exercise treadmill test. Randomization was stratified by the total exercise time average (\< 7 minutes or ≥ 7 minutes) of the 2 qualifying exercise treadmill tests performed during screening.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Participants who received study drug and completed the baseline and post-randomization exercise treadmill tests

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Total Exercise Time, measured over Baseline and day 1, after dosing.

The primary endpoint is change from baseline in total exercise time, measured in seconds. Exercise time was assessed with an exercise treadmill test using the standard Bruce protocol. In that protocol, a participant walks on an uphill treadmill while speed and incline increase at regular stages; electrodes on the chest monitor cardiac function. The registry describes seven possible stages and a maximum possible exercise time of 21 minutes.

As a continuous measure, exercise time records a quantity rather than sorting participants into outcome categories. Change from baseline expresses the post-dose measure relative to each participant’s baseline measurement. Comparing changes between randomized groups is intended to describe whether the groups’ measured exercise times changed differently over the assessment period. The posted outcome is a group-level least-squares mean, not a report of each participant’s individual change.

In general, treadmill exercise time can be used to assess exercise performance in a controlled test, while the protocol also monitors cardiac response. The registry selected this measure to evaluate the study question about exercise time after dosing. The endpoint alone does not describe every aspect of cardiovascular safety or the secondary outcomes, which include time to onset of exercise-induced angina and time to onset of at least 1 mm ST-segment depression. Results for those secondary outcomes are not reported in the registry record details.

7. Results Posted in the Registry

• Erenumab: mean -2.90 versus 8.10 on control; difference -11.00 (95% CI -51.47 to 29.47), standardized effect d = -0.11; z-test p = 0.594.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous, and the registry reports least-squares means with standard errors. A least-squares mean is a model-based group mean; depending on the model, it can account for specified factors rather than simply reporting the raw arithmetic mean. The registry record does not identify the model or covariates used to obtain these means. The standard error describes the estimated uncertainty in a group mean, not the spread of individual participant measurements.

What does the treatment difference describe?

The treatment difference compares the estimated change in the erenumab group with the estimated change in the placebo group. The posted values are -2.9 seconds for erenumab and 8.1 seconds for placebo, and the reported difference is -11.0 seconds. A negative value means the estimated change was lower in the erenumab group under this subtraction order. The result is expressed in seconds, keeping the estimate in the same units as the endpoint.

In general, a difference in means is easier to interpret in the original measurement units than a standardized effect. A standardized effect expresses the contrast relative to a measure of variability and can support comparisons across scales, but it does not replace the seconds-based estimate. The posted standardized effect is -0.115. Its interpretation depends on the variability estimate, which is approximate in this re-calculation.

What is the role of the unadjusted re-calculation?

The re-calculation is a simple, unadjusted analysis based on the posted summary statistics. It gives a mean difference of -11.0 seconds, an interval from -51.47 to 29.47 seconds, and an unadjusted p-value of 0.594. The registry also gives standard errors of 14.4 and 14.8 seconds for the group estimates. The re-calculation uses approximate dispersion derived from these posted summaries, as indicated by the record’s approximation flag. It does not reproduce an undocumented sponsor model or account for covariates that are not listed.

In general, a confidence interval describes a range of values compatible with the data under the analysis method and its assumptions. It is not a range containing a fixed proportion of individual participants, nor is it a probability statement about the particular treatment difference after the data have been observed. The re-calculated interval crosses zero, so it includes no difference as well as negative and positive differences.

How should the p-value be read?

In general, a p-value is calculated relative to a null hypothesis and a statistical model. The unadjusted p-value of 0.594 does not quantify the size or clinical importance of the estimated difference. It also does not establish equivalence between groups. Interpretation requires attention to the point estimate, interval, endpoint, analysis population, and design. The sponsor analysis has no p-value reported in the registry record details used for this explanation.

Why distinguish the sponsor analysis?

The registry labels the sponsor’s reported analysis as non-inferiority and gives a treatment difference of -11.0 with a 90% confidence interval from -44.9 to 22.9. A non-inferiority analysis generally evaluates whether an outcome is not worse than a prespecified margin in a particular direction. The margin, direction, model, and other details needed to assess that question are not reported in the registry record. Therefore, the interval alone cannot establish the sponsor’s non-inferiority conclusion. The posted label identifies the analysis type, but it does not supply the missing decision criterion.

The sponsor result and the re-calculation should be kept distinct because they are not presented as the same analysis. Their interval limits differ, and the sponsor interval is explicitly identified as 90%. The registry record does not state the confidence level for the re-calculated interval in the registry record. A difference in interval width can reflect analysis choices and confidence level; without the full methods, it should not be attributed to a particular cause.

What does variability contribute?

Variability affects precision. In general, greater variation among observations makes a group mean harder to estimate precisely when other factors are held fixed. The standard errors summarize uncertainty around the group estimates, while the approximate pooled standard deviation of 95.71 seconds describes the scale of variation used in the re-calculation. These are different quantities: a standard deviation concerns spread in measurements, whereas a standard error concerns uncertainty in an estimated mean.

All statistical summaries depend on assumptions. The registry record does not report the full analysis model, distributional checks, covariate adjustment, handling of incomplete measurements, or details of any sensitivity analyses. The re-calculated quantities are therefore best read as a transparent, unadjusted comparison of the posted summaries, not as a replacement for the sponsor’s analysis or a complete reconstruction of the trial’s statistical plan.

10. Confidence Intervals and Precision

The unadjusted re-calculated estimate is a mean difference of -11.0 seconds, with an interval from -51.47 to 29.47 seconds. This interval is wide relative to the point estimate and extends from negative to positive values. It therefore allows for a range of possible group contrasts, including no difference, a lower change in the erenumab group, or a higher change in that group. The data do not pin down the direction or size of the difference with high precision.

The sponsor’s reported treatment difference is also -11.0, with a 90% confidence interval from -44.9 to 22.9. This is a distinct reported analysis. The confidence level differs from the standard confidence-interval interpretation often used for re-calculation; the registry entries for the unadjusted interval do not state its confidence level. The two intervals should not be treated as directly identical summaries or combined.

In general, an interval’s width reflects both the amount of information and the chosen confidence level, along with the analysis method. A narrower interval indicates greater precision under its assumptions; it does not by itself show that an effect is clinically important. Conversely, a wide interval indicates uncertainty, not proof that the true difference is zero. The registry does not report a threshold for a clinically important change in exercise time, so the clinical significance of values within either interval cannot be determined from these results alone.

The arm-level standard errors are 14.4 seconds for placebo and 14.8 seconds for erenumab. They describe uncertainty around the reported group estimates, not ranges of individual responses. Because the primary outcome population comprises participants who completed both treadmill assessments, missing outcome information could affect precision or interpretation; the amount and handling of missing data are not reported in the registry record.

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 illustrations translate the observed effect into planning examples. Because the primary endpoint is continuous, the figures are participants per arm for 80% power and participants per arm for 90% power, not total events. The illustrations assume a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The effect used for each illustration is the stated fraction of the observed effect: 100%, 75%, or 50%. These are planning assumptions for the illustrations, not the original trial’s design assumptions. The original sample-size justification is not reported in the registry record.

The 80% power column describes the number of participants per arm that the planning calculation estimates would be needed to have an 80% chance of detecting the specified effect under the stated assumptions. The 90% power column describes the corresponding planning target for a 90% chance. In general, higher planned power requires more participants when other assumptions remain fixed.

The rows show that smaller effects need more participants. Planning for 75% of the observed effect requires more participants per arm than planning for 100% of the observed effect, and planning for 50% requires more still. This is a feature of the calculation, not evidence that a smaller effect was observed in a separate group. Because the observed effect is used as the starting point, the estimates are sensitive to that observed value and its uncertainty. The figures are planning illustrations, not observed power, and they should not be read as the trial’s actual enrollment target or as evidence about the trial’s success.

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, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -11.000; pooled standard deviation σ = 95.712.
2. n = 2 × (1.96 + 0.842)² × 95.712² / -11.000² = 1,188.5, rounded up to 1,189 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -11                      # observed difference in means
sigma <- 95.7123667185027                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(2575833)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 1,189 (80% power), 1,592 (90% power)
 75% of observed effect: n per arm = 2,113 (80% power), 2,829 (90% power)
 50% of observed effect: n per arm = 4,754 (80% power), 6,365 (90% power)
power.t.test: 1189.4 
Simulated power at n = 1,189 per arm: 0.801 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Analyses Reported by the Sponsor

The sponsor’s own analysis is reported as a treatment difference of -11.0, with a 90% confidence interval from -44.9 to 22.9, and is labelled non-inferiority. The registry does not provide a p-value for this result or the non-inferiority margin, model details, or other information needed to independently evaluate its conclusion. A confidence interval without the decision margin does not establish whether a non-inferiority criterion was met.

The separate unadjusted re-calculation gives the same point estimate, -11.0, but an interval from -51.47 to 29.47 and an unadjusted p-value of 0.594. It is based on posted summary statistics and approximate dispersion, rather than a reconstruction of the sponsor’s full analysis. The sponsor result and the re-calculation therefore answer related but not necessarily identical statistical questions.

The registry does not flag a discrepancy between these results. That does not make the methods interchangeable: the sponsor’s result remains the trial’s reported analysis, while the re-calculation is a supplementary, unadjusted comparison. Without further analysis details, the difference between their interval limits should not be assigned to a specific modeling choice.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists three outcome measures and nine analyses. The primary endpoint is change from baseline in total exercise time. The two named secondary outcomes are time to onset of exercise-induced angina and time to onset of at least 1 mm ST-segment depression. The results for those secondary outcomes are not reported in the registry record details available for this analysis.

When a study examines multiple outcomes or analyses, the chance of obtaining at least one apparently notable result can increase if each result is assessed separately without accounting for multiplicity. In general, multiplicity methods can adjust interpretation or decision thresholds across a family of tests. The registry record does not report whether multiplicity was addressed, which analyses formed any testing family, or whether a hierarchy was prespecified. The unadjusted p-value should therefore be understood as an unadjusted result, not as evidence that multiplicity was handled.

The primary outcome population includes participants who received study drug and completed both baseline and post-randomization treadmill tests. This is the population definition reported for that endpoint. The registry record does not describe the full analysis population, whether an intention-to-treat analysis was conducted, or whether other analysis populations were used. It also does not report the number or pattern of missing treadmill measurements, reasons for missingness, imputation methods, or sensitivity analyses. No specific explanation for the difference between the posted outcome group counts can be established from the available record.

These omissions matter because analysis populations and missing-data methods can affect the estimated treatment contrast and its uncertainty. In general, results based on participants with complete measurements may differ from results based on other populations, especially if missingness is related to outcomes or treatment. The record does not provide enough information to assess whether that concern applies in this trial. Likewise, the registry does not report interim analyses or the original statistical analysis plan, so those features cannot be inferred from the posted primary result.

15. 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 3% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: atrial fibrillation (2.3%).

The registry reports serious adverse events for 44 participants at risk in each group. There was 1 participant with a serious adverse event in the placebo group and 0 in the erenumab group. The posted figures for other adverse events are 6 in the placebo group and 4 in the erenumab group. No deaths are reported in either group. These are descriptive counts; they do not by themselves establish a causal relationship with study treatment.

Atrial fibrillation is listed among the top serious event terms at 2.3%. The record details do not identify which group this term belongs to, so it should not be assigned to either group. The available information also does not give event-level detail, timing, severity beyond the serious-event classification, or a full account of how events were assessed.

In general, adverse-event counts are useful for describing what was recorded during a study, but small counts provide limited precision for comparing groups. Differences in counts alone do not establish that one treatment caused more or fewer events. The registry record does not provide enough detail to estimate comparative risk or to draw broader safety conclusions. These figures should be read alongside the study’s size, duration, population, and the limits of the reported safety information.

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

The primary result is an estimate of the difference between groups in change from baseline in treadmill exercise time on day 1 after dosing. The posted group estimates are 8.1 seconds for placebo and -2.9 seconds for erenumab, with a re-calculated difference of -11.0 seconds. The interval around the re-calculated estimate extends from -51.47 to 29.47 seconds, and the unadjusted p-value is 0.594. Together, these values indicate substantial uncertainty about the size and direction of the group contrast.

The result does not show that the groups are equivalent. A non-significant p-value is not proof of no difference, and an interval crossing zero does not establish that all clinically relevant differences have been ruled out. The registry does not report a threshold for a clinically important difference, so the clinical importance of the estimate and interval cannot be determined from the endpoint results alone.

The registry’s sponsor analysis is labelled non-inferiority and reports a 90% interval from -44.9 to 22.9 around the treatment difference of -11.0. The non-inferiority margin and analysis details are not reported in the registry record. Without them, the sponsor’s label and interval cannot independently establish what decision criterion was applied or whether it was met.

The endpoint concerns a standardized treadmill test, not all aspects of stable angina, cardiovascular function, or daily activity. The primary assessment is at baseline and on day 1 after dosing; these registry entries do not describe longer-term effects. Secondary outcome names are recorded, but their results are not reported in the registry record. Safety counts are also descriptive and do not establish causation.

Accordingly, the posted statistics characterize one primary outcome in this trial and the precision of its estimated group difference. They do not establish that erenumab works or does not work, and they are not medical advice. Broader interpretation would require the full analysis methods, secondary results, missing-data information, and relevant clinical context, some of which are not reported in the registry record.

17. Why This Trial Matters Statistically

This trial illustrates why the point estimate, interval, and analysis definition need to be read together. The treatment difference is -11.0 seconds, but the unadjusted interval extends well into both negative and positive values. The point estimate supplies a direction and scale; the interval shows that the data do not estimate the contrast precisely. Treating either the point estimate or the p-value as a complete summary would omit important information.

It also illustrates the distinction between a sponsor analysis and a simple re-calculation. Both report a treatment difference of -11.0, but their intervals differ, and the sponsor result is explicitly a 90% interval from an analysis labelled non-inferiority. The unadjusted re-calculation is based on posted summaries and does not reproduce the sponsor’s analysis. Clear reporting of the margin, model, analysis population, and missing-data approach would help readers assess the sponsor result more fully.

The study’s endpoint is measured under a standardized treadmill protocol and is expressed in seconds, which makes the estimated difference interpretable in the original measurement scale. Yet the registry does not give a threshold for clinical importance, and the primary assessment is limited to baseline and day 1 after dosing. This keeps the statistical question focused: the estimates concern a short-term test measure, not every possible clinical outcome.

Finally, the sample-size illustrations show how planning depends on the effect selected for detection and the assumed power. Because they are based on the observed effect, they are not a record of the trial’s original design rationale. Distinguishing illustrative planning calculations from the actual study design prevents a retrospective calculation from being mistaken for evidence about the original planning process.

18. 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 contains limited detail about the statistical analysis. It does not report the model used for the least-squares means, covariate adjustment, the original sample-size justification, interim analyses, or the full analysis plan. The sponsor result is labelled non-inferiority, but the margin and decision details are not reported. These gaps limit independent assessment of the sponsor’s analysis.

The primary outcome population is defined by receipt of study drug and completion of both treadmill tests. The record does not report the amount or reasons for missing data, how missing measurements were handled, or whether sensitivity analyses were conducted. The posted primary outcome counts are 44 for placebo and 42 for erenumab, while the safety counts at risk are 44 in each group; the record does not explain this difference. No reason should be inferred from the counts alone.

The outcome is measured at baseline and on day 1 after dosing. The registry entries do not provide later primary endpoint assessments. The sample-size illustrations are based on the observed effect and stated planning assumptions rather than the original design justification. They therefore do not resolve questions about the study’s intended precision.

The results available for this explanation include the primary exercise-time endpoint and names of two secondary outcomes, but not secondary outcome results. Safety information is limited to the posted serious-event and other-event counts, deaths, and a top serious event term. These details do not permit a complete assessment of all outcomes, event timing, or causal attribution.

These limitations do not invalidate the posted estimates, but they constrain what can be concluded from them. The interval indicates uncertainty in the primary contrast, while absent analysis details limit evaluation of the assumptions underlying both the sponsor result and the re-calculation.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Treadmill Cardiovascular Safety Study of Erenumab (AMG 334) — NCT02575833.
• Registry record last updated 2018-08-24; status last verified 2018-07.
• Results first posted 2018-07-05.
• 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.

22. Record Summary

NCT02575833 is a completed phase 2a randomized, double-blind, placebo-controlled, parallel-group study of erenumab in adults with stable angina. Participants received a single intravenous infusion of erenumab 140 mg or placebo on day 1. The primary endpoint was change from baseline in total exercise time, measured in seconds with a treadmill test.

The posted group estimates were 8.1 seconds for placebo and -2.9 seconds for erenumab. The unadjusted re-calculated treatment difference was -11.0 seconds, with an interval from -51.47 to 29.47 and an unadjusted p-value of 0.594. The sponsor’s reported non-inferiority analysis gave a treatment difference of -11.0 and a 90% confidence interval from -44.9 to 22.9. The registry does not report the non-inferiority margin or full analysis details.

The primary result is imprecise, and missing-data, multiplicity, and analysis-plan details are not reported in the registry record. The secondary outcomes are named, but their results are not included in the registry record for this explanation.