Study to Evaluate the Efficacy and Safety of Erenumab (AMG 334) in Migraine Prevention: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study to Evaluate the Efficacy and Safety of Erenumab (AMG 334) in Migraine Prevention (NCT01952574), a randomized phase 2 study. The primary outcome is Change From Baseline in Monthly Migraine Days at Week 12.

2. Clinical Question

3. Trial Design

The registry describes a phase 2, randomized, double-blind, placebo-controlled, parallel-group study of erenumab in migraine prevention. The masking field is recorded as triple. The condition was migraine, and the stated study objective was to evaluate change from baseline in monthly migraine days among participants with episodic migraine. The registry records enrollment of 483 and study status as completed.

During the double-blind treatment phase, participants received placebo or one of three monthly erenumab doses: 7 mg, 21 mg, or 70 mg. Injections were given on day 1 and at weeks 4 and 8. The primary outcome was assessed at week 12, using the last 4 weeks of that phase. The design also included an open-label treatment phase, with erenumab treatment described in the arm records, and a separate United States substudy comparing self-administration devices. Those later components are distinct from the placebo-controlled primary comparison.

The registry records a study start date of 2013-08-06 and a primary completion date of 2014-09-25. The original trial’s detailed sample-size justification, planned interim analyses, and full statistical analysis plan are not reported in the registry record. The posted numerical comparisons include the sponsor’s generalized linear mixed model analyses and simple unadjusted recalculations based on posted summary statistics; these are different forms of analysis and should not be treated as interchangeable.

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 3:2:2:2 to receive placebo, erenumab 7 mg, erenumab 21 mg, or erenumab 70 mg once a month (QM) in the double-blind phase. Randomization was stratified by region (North America vs. Europe). During the open-label treatment phase, participants in the United States (US) could participate in an optional Clinical Home Use (CHU) substudy.

Only the first 6 of 7 groups are shown.

Double-blind Treatment Phase

Reasons for not completing this period:

Open-label Treatment Phase

Reasons for not completing this period:

Clinical Home Use Substudy

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Participants who received at least 1 dose of investigational product (IP) and had ≥ 4 migraine days during the 4-week baseline phase (efficacy analysis set), and with at least one change from baseline value in monthly migraine days.

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 Monthly Migraine Days at Week 12, measured over 4-week baseline phase and the last 4 weeks of the 12-week double-blind treatment phase.

The primary endpoint was change from baseline in monthly migraine days at week 12. The registry defines a migraine day as a calendar day on which a participant experienced a qualified migraine headache, including onset, continuation, or recurrence. A qualified headache was defined as migraine with or without aura.

The change was calculated by subtracting the number of migraine days during the 4-week baseline phase from the number during the last 4 weeks of the 12-week double-blind treatment phase. A negative change therefore represents fewer recorded migraine days in the later measurement period than at baseline. The posted group estimates are least-squares means, with standard errors as the reported measure of dispersion.

Monthly migraine days provide a continuous measure of headache burden over time. General statistical context: using a continuous outcome retains information about the size of change across participants, rather than reducing each participant’s result to a yes-or-no threshold. The endpoint also gives a defined comparison period and a baseline reference. Its interpretation depends on the registry’s migraine-day definition, the analysis population, and the sponsor’s modeling approach. It describes the measured change over the specified periods; by itself, it does not establish how participants experienced that change or whether it was important to them.

7. Results Posted in the Registry

• Erenumab 7 mg QM: mean -2.18 versus -2.28 on control; difference 0.10 (95% CI -0.83 to 1.03), standardized effect d = 0.03; z-test p = 0.833.
• Erenumab 21 mg QM: mean -2.39 versus -2.28 on control; difference -0.11 (95% CI -1.07 to 0.85), standardized effect d = -0.03; z-test p = 0.823.
• Erenumab 70 mg QM: mean -3.40 versus -2.28 on control; difference -1.12 (95% CI -2.07 to -0.17), standardized effect d = -0.29; z-test p = 0.020.

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.

Percentage of Participants With at Least a 50% Reduction From Baseline in Monthly Migraine Days at Week 12

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Erenumab 7 mg QM: 28.8% (30/104) versus 29.9% (43/144) on control; risk difference -1.0 percentage points (95% CI -12.5 to 10.5), risk ratio 0.97 (0.65–1.43); unadjusted two-proportion z-test p = 0.863.
  • Erenumab 21 mg QM: 34.4% (32/93) versus 29.9% (43/144) on control; risk difference 4.5 percentage points (95% CI -7.7 to 16.8), risk ratio 1.15 (0.79–1.68); unadjusted two-proportion z-test p = 0.462.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Erenumab 7 mg QM and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Erenumab 21 mg QM and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in Monthly Migraine Attacks at Week 12

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Erenumab 7 mg QM: mean -1.07 versus -1.44 on control; difference 0.37 (95% CI -0.14 to 0.88), standardized effect d = 0.18; z-test p = 0.159.
  • Erenumab 21 mg QM: mean -1.42 versus -1.44 on control; difference 0.02 (95% CI -0.51 to 0.55), standardized effect d = 0.01; z-test p = 0.941.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Erenumab 7 mg QM and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Erenumab 21 mg QM and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists secondary outcomes covering a 50% reduction in monthly migraine days, change in monthly migraine attacks, treatment-emergent adverse events in double-blind and open-label periods, events in the self-administration substudy, and anti-erenumab antibodies. The available numerical efficacy summaries show a mixed pattern across doses and outcome definitions rather than a uniform shift across every measure. These outcomes offer context for the primary endpoint, but they do not replace it: responder status uses a threshold, while monthly attacks are a related but distinct measure from monthly migraine days.

The sponsor’s reported analyses for the responder outcome include comparisons for 70 mg and 21 mg. The sponsor reports an odds ratio of 2.00 for 70 mg versus placebo, with a 95% confidence interval from 1.17 to 3.42 and p = 0.011. For 21 mg versus placebo, the reported odds ratio is 1.25, with a 95% confidence interval from 0.71 to 2.18 and p = 0.44. The posted continuous summaries for monthly migraine attacks include sponsor comparisons for 70 mg and 21 mg; the reported intervals include zero. The registry record does not report a sponsor comparison for every secondary efficacy dose contrast.

General statistical context: a responder threshold can be useful for communicating the share of participants crossing a prespecified cutoff, but it discards differences among participants on the same side of that cutoff. Monthly migraine attacks and monthly migraine days may also capture overlapping but nonidentical aspects of the condition. Similarity or difference between these measures should not be treated as a validation test of the primary endpoint.

Secondary outcomes involve additional comparisons, and the registry record does not report a multiplicity adjustment strategy. Unadjusted secondary p-values therefore carry less weight than a single prespecified primary comparison: testing multiple outcomes and doses increases the chance of apparently notable findings arising by chance. These findings are best treated as supportive or exploratory evidence, interpreted alongside the primary analysis and with attention to which comparisons the sponsor actually reported.

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

• Percentage of Participants With at Least a 50% Reduction From Baseline in Monthly Migraine Days at Week 12
• Change From Baseline in Monthly Migraine Attacks at Week 12
• Number of Participants With Treatment-emergent Adverse Events in the Double-blind Treatment Phase
• Number of Participants With Treatment-emergent Adverse Events in the Open-label Treatment Phase
• CHU Substudy: Number of Participants With Treatment-emergent Adverse Events During the CHU Substudy
• Number of Participants Who Developed Anti-erenumab Antibodies During the Double-blind Treatment Phase
• Number of Participants Who Developed Anti-erenumab Antibodies During the Open-label Treatment Phase

9. Statistical Interpretation

10. Statistical Methods Explained

The posted results combine model-based sponsor analyses with simple unadjusted recalculations based on group summaries. Each answers a related but not identical statistical question. Keeping the distinction clear is especially important because the registry flags a discrepancy between the reported sponsor analyses and the recalculations.

What does a least-squares mean represent?

The primary outcome is reported as a least-squares mean, also called an LS mean. General statistical context: a least-squares mean is a model-based estimate of a group mean. Depending on the model, it can account for specified covariates or features of the analysis, rather than being merely the arithmetic average of observed values. The registry identifies the outcome measure and the sponsor’s model, but the full covariate specification and other model details are not reported in the registry record. The posted estimates should therefore be read as the sponsor’s model-based group summaries, not as unadjusted raw averages.

The primary group estimates are -2.28 for placebo, -2.18 for 7 mg, -2.39 for 21 mg, and -3.4 for 70 mg. These values describe estimated within-group change from baseline. The corresponding standard errors are 0.31, 0.36, 0.38, and 0.37, respectively. A standard error expresses uncertainty in an estimated mean; it is not a measure of the spread of individual participant outcomes.

What is the sponsor’s comparison?

The registry reports generalized linear mixed model analyses of the primary endpoint. These compare each erenumab group with placebo through a least-squares mean difference. General statistical context: a mixed model can account for structured relationships in data, such as repeated observations, when the model is specified to do so. The registry names the model but does not report enough detail to reconstruct its full specification. The reported contrasts and their confidence intervals and p-values are therefore the central inferential results for these comparisons.

A difference below zero in these reported contrasts indicates a more negative change estimate for the erenumab group than for placebo, given the reported subtraction direction. The interpretation is about the difference in estimated change, not about a direct difference in the number of migraine days at a single baseline or follow-up measurement.

What is recalculated?

The recalculations are simple, unadjusted comparisons using posted summary statistics. For continuous outcomes, the recalculated mean difference is formed from the posted group estimates, and its uncertainty uses an approximate pooled standard deviation derived from the reported standard errors and group sizes. The record marks this standard-deviation step as approximate. This approach does not reproduce the sponsor’s generalized linear mixed model, including any model structure or adjustment used in that analysis. Recalculated p-values are unadjusted and should not be substituted for sponsor results when the two differ.

For the primary outcome, the recalculated mean differences are 0.1 for 7 mg versus placebo, -0.11 for 21 mg versus placebo, and -1.12 for 70 mg versus placebo. The corresponding unadjusted p-values are 0.833, 0.823, and 0.02. These estimates are useful as transparent checks based on posted summaries, not as replacements for the sponsor’s own analysis. Because the discrepancy flag is true, the sponsor’s analysis takes precedence.

How should confidence intervals and p-values be read?

General statistical context: a confidence interval gives a range of values compatible with the data and the model under the interval procedure. Its width conveys precision, while its location conveys the direction and range of the estimated effect. A p-value measures how unusual data at least as incompatible with a specified null hypothesis would be under the assumptions of the test; it is not the probability that the null hypothesis is true, nor a measure of clinical importance.

The sponsor reports 95% intervals for the 70 mg and 21 mg comparisons, but a 92% interval for the 7 mg primary comparison. These reported levels should not be silently treated as identical. The recalculated intervals are a separate unadjusted analysis. The registry record does not report a full multiplicity plan, so the posted p-values should not be read as establishing that all comparisons were protected against the consequences of testing multiple doses and outcomes.

11. Confidence Intervals and Precision

The sponsor’s interval for erenumab 70 mg QM versus placebo runs from -2.06 to -0.17, around a reported difference of -1.12. It does not include zero under the sponsor’s model-based analysis. The interval gives a range of values compatible with that analysis; it does not guarantee that the underlying effect lies within the range, and its width reflects uncertainty as well as the modeling and data used.

For 21 mg, the sponsor’s 95% interval runs from -1.07 to 0.86 around -0.10. For 7 mg, the reported 92% interval runs from -0.83 to 1.05 around 0.11. Both intervals include zero and span values on either side of it. The 7 mg interval has a different stated confidence level from the other primary comparisons, so direct visual comparisons of interval widths should account for that difference.

The recalculated intervals are similar in broad direction but are not identical to the sponsor’s reported intervals. For example, the recalculated 70 mg interval is -2.066 to -0.174. Such differences can arise because the recalculation uses posted summary statistics and an approximate standard deviation rather than reproducing the sponsor’s model. The registry marks a discrepancy, so the sponsor’s analysis takes precedence.

General statistical context: an interval that includes zero does not prove that there is no difference; it indicates that the data and analysis are compatible with zero as well as with values on either side. Conversely, an interval excluding zero does not by itself establish clinical importance. The registry does not report a threshold for a clinically important change or a patient-centered interpretation of the interval. The estimates should therefore be read as measures of statistical precision for the stated comparisons, not as complete judgments about practical meaning.

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 sample-size figures are planning illustrations based on the observed effect, not a report of the trial’s original design calculation. For these comparisons, the figures are participants per arm. They are not total event counts. The stated planning 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 the stated fraction of the observed effect. The trial’s own original sample-size assumptions are not reported in the registry record.

The 80% power column gives the illustrative number of participants per arm for a design with an 80% chance of detecting the specified effect under those assumptions, if that effect and the assumptions were correct. The 90% power column gives the corresponding planning count for a 90% chance. These are design calculations, not guarantees about results in a future study. The plotted curve expresses the same planning relationship across effects, rather than documenting what the completed trial was powered to detect.

The rows describe planning for 100%, 75%, and 50% of the observed effect. Smaller effects require more participants because they are harder to distinguish from random variation under a fixed design framework. The table’s estimates vary greatly by comparison: the observed difference for 70 mg produces much smaller planning counts than the very small observed differences for 7 mg and 21 mg. That contrast is a mathematical consequence of using each comparison’s observed effect as the planning target, not evidence that a future effect will match the observed estimate.

These figures are not observed power. They should not be used to infer whether the completed study had adequate power from its observed p-values, or to replace the sponsor’s trial-specific design rationale, which is not reported in the registry record.

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

Worked calculation for this trial

1. Observed difference Δ = 0.100; pooled standard deviation σ = 3.782.
2. n = 2 × (1.96 + 0.842)² × 3.782² / 0.100² = 22,450.9, rounded up to 22,451 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 <- 0.0999999999999996                      # observed difference in means
sigma <- 3.78179666193436                     # 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(1952574)
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 = 22,451 (80% power), 30,056 (90% power)
 75% of observed effect: n per arm = 39,913 (80% power), 53,432 (90% power)
 50% of observed effect: n per arm = 89,804 (80% power), 120,222 (90% power)
power.t.test: 22451.9 
Simulated power at n = 22,451 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.

14. Analyses Reported by the Sponsor

The sponsor’s primary analyses are reported as generalized linear mixed model contrasts. They are the registry’s own inferential results and should be used in preference to the simple recalculations when the estimates differ, because the record flags a discrepancy. For 70 mg, the sponsor reported a difference of -1.12, a 95% confidence interval from -2.06 to -0.17, and p = 0.021. The unadjusted recalculation gives -1.12, an interval from -2.066 to -0.174, and p = 0.02. The broad result is similar, but the methods and reported precision are not identical.

For 21 mg, the sponsor reported -0.10, with a 95% interval from -1.07 to 0.86 and p = 0.83; the recalculated estimate is -0.11, with an unadjusted p-value of 0.823. For 7 mg, the sponsor reported 0.11, with a 92% interval from -0.83 to 1.05 and p = 0.82; the recalculated estimate is 0.1, with an unadjusted p-value of 0.833. The sponsor’s reported 7 mg interval also uses a different confidence level from the 95% intervals reported for the other primary comparisons.

The recalculations are checks derived from posted summary statistics, not replications of the sponsor’s model. They do not incorporate unreported model details. They can help readers understand the approximate scale and direction of the comparisons, but the sponsor’s own analysis takes precedence.

15. Multiplicity, Missing Data and Analysis Populations

The registry reports several dose comparisons for the primary endpoint and lists multiple secondary outcomes. It does not report a multiplicity procedure specifying how the family of tests was defined or whether p-values and confidence intervals were adjusted. Without that information, it is not possible to determine from the registry record how the sponsor controlled the overall chance of false-positive findings across the comparisons.

The sponsor labels its analyses as superiority comparisons and reports p-values and confidence intervals. Those labels do not, by themselves, establish how testing across doses and outcomes was handled. In particular, an unadjusted p-value from a secondary analysis should not be interpreted as though it were the sole test in the study. General statistical context: multiplicity matters because repeated testing creates more opportunities for a result to appear unusual by chance. This concern applies to the broader set of comparisons, not just to one endpoint in isolation.

The primary analysis population is described as participants who received at least one dose of investigational product, had at least 4 migraine days during the 4-week baseline phase, and had at least one change-from-baseline value in monthly migraine days. This is an efficacy analysis set, as named in the record. The registry does not report the complete participant flow for this analysis, nor does it provide enough information to assess how participants missing endpoint data were handled.

Specific missing-data methods, assumptions about missingness, imputation procedures, sensitivity analyses, and the handling of protocol deviations are not reported in the registry record. The record also does not report the full prespecified statistical analysis plan or interim-analysis arrangements. These absences limit assessment of robustness: different reasonable assumptions about incomplete outcome data can affect estimated treatment contrasts. They should be regarded as unreported methodological details, not as evidence that a particular procedure was or was not used.

16. Safety Results

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

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: vertigo (0.9%), migraine (0.9%), ovarian cyst ruptured (0.9%), appendicitis (0.8%), ligament rupture (0.8%).

The registry reports serious adverse event counts by treatment period and group. During the double-blind treatment phase, serious events were reported for 1 participant in placebo, 1 in erenumab 7 mg QM, 0 in erenumab 21 mg QM, and 1 in erenumab 70 mg QM. The corresponding numbers at risk were 153, 108, 105, and 106. During the open-label treatment phase, serious events were reported for 30 participants in the 70 mg group, 25 in the 140 mg group, and 49 in the 70/140 mg group; the numbers at risk were 383, 250, and 383, respectively.

In the self-administration substudy, no serious events were reported in either device group: the PFS group had 42 at risk, and the AI/Pen group had 41 at risk. The substudy total row reports 0 serious events among 83 at risk. The record lists serious-event terms and percentages, including vertigo, migraine, ovarian cyst ruptured, appendicitis, ligament rupture, and urinary tract infection. These are descriptive registry summaries; no causal attribution should be inferred from the counts alone. Death counts are not reported in the registry record.

General statistical context: serious adverse event counts are not a complete safety profile. Comparisons across double-blind and open-label periods are especially limited because those periods differ in treatment context and are not the same randomized comparison. The posted counts should be read alongside the stated groups and periods, not as direct evidence that one treatment caused or prevented an event.

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

The primary endpoint records change in monthly migraine days from a defined baseline period to the last 4 weeks of the double-blind phase. The sponsor’s model-based estimate for 70 mg versus placebo is -1.12, with a reported 95% confidence interval from -2.06 to -0.17 and p = 0.021. This describes the estimated contrast for the specified efficacy population and period. The sponsor’s analyses for 7 mg and 21 mg have intervals that include zero and reported p-values of 0.82 and 0.83, respectively.

These results do not establish that every participant experienced the group-average change, nor do they show how outcomes varied across individuals. The primary measure is a group-level estimate, and its precision is represented by the reported standard errors and confidence intervals. The trial registry does not report a threshold for a clinically important difference, so the statistical estimate alone cannot determine the practical importance of the contrast for participants.

The results also do not answer every question about treatment. The primary comparison concerns the double-blind phase, not the subsequent open-label treatment period or the device substudy. The record describes secondary efficacy and safety outcomes, but multiple comparisons, missing-data procedures, and the complete analysis plan are not reported in the registry record. These gaps constrain how broadly the findings can be interpreted.

General statistical context: a p-value is not a measure of effect size, certainty that a finding will replicate, or probability that an intervention is beneficial. A confidence interval communicates uncertainty under the analysis assumptions, while the endpoint definition sets the specific quantity being estimated. The appropriate reading is therefore bounded: these are posted estimates from a randomized, placebo-controlled study, analyzed over a specified period, with unresolved details about multiplicity and missing data. They do not by themselves establish a general conclusion beyond those data and conditions.

18. Why This Trial Matters Statistically

This trial illustrates why a randomized, placebo-controlled comparison is different from looking only at change within each treatment group. All four primary group estimates were negative, including placebo. A within-group decrease does not isolate the effect associated with assignment to a particular dose; the between-group contrast is the relevant comparison for the stated treatment question.

The example also shows the role of model-based analysis and transparent recalculation. The sponsor reports generalized linear mixed model contrasts, while the simple recalculations use posted summary values and an approximate standard deviation. Their close or differing estimates can help readers see which conclusions are robust to this simplified calculation and which details depend on the sponsor’s analysis. Because a discrepancy is flagged, the sponsor’s own results take precedence.

The trial provides a useful setting for understanding precision. The 70 mg estimate is accompanied by a confidence interval that excludes zero in the sponsor’s analysis, while the intervals for the other primary dose comparisons include zero. This contrast should be read together with interval width, the stated confidence level, and the fact that multiple doses and outcomes were examined. It is not a substitute for assessing clinical importance.

Finally, the record highlights the limits of registry reporting. The primary population and endpoint are defined, and several outcomes are listed, but some details needed for a full statistical audit are not reported in the registry record. Recognizing the difference between available results and unreported methods is an important part of sound interpretation, particularly when secondary analyses and safety summaries accompany a primary efficacy comparison.

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.

Interpretation is limited by the absence of the full statistical analysis plan in the registry record. The model is named as a generalized linear mixed model, but its detailed specification, including covariates and other implementation choices, is not reported. Multiplicity handling, planned interim analyses, and the trial’s original sample-size justification are also not reported in the registry record.

The primary outcome population is a defined efficacy analysis set rather than an unrestricted description of everyone enrolled. Participants had to meet the stated baseline migraine-day condition, receive at least one dose, and have at least one change-from-baseline value. The registry does not provide enough information to assess the full participant flow into the analysis or how missing endpoint values were handled. The missing-data assumptions and sensitivity analyses are not reported in the registry record.

The primary comparisons concern the 12-week double-blind phase and should not be extended to the open-label period. The open-label treatment summaries and the self-administration substudy address different contexts, without the same placebo-controlled comparison. The record also presents multiple doses and secondary outcomes; absent a reported multiplicity procedure, secondary p-values warrant caution.

The simple recalculations rely on posted summaries and an approximate pooled standard deviation derived from standard errors. They do not reproduce the sponsor’s model-based analysis. Since the registry flags a discrepancy, sponsor results take precedence. Finally, the registry record does not report a clinical-importance threshold for the primary contrast, limiting interpretation of the practical significance of the estimated difference.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Study to Evaluate the Efficacy and Safety of Erenumab (AMG 334) in Migraine Prevention — NCT01952574.
• Registry record last updated 2022-10-12; status last verified 2022-10.
• Results first posted 2018-07-10.
• 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.

23. Record Summary

The registry describes a completed phase 2 randomized, double-blind, placebo-controlled, parallel-group study of erenumab in migraine prevention, with enrollment of 483. The primary outcome was change from baseline in monthly migraine days at week 12, measured over the last 4 weeks of the 12-week double-blind phase relative to a 4-week baseline phase.

The sponsor reported generalized linear mixed model comparisons of erenumab 7 mg, 21 mg, and 70 mg QM with placebo. For 70 mg, the reported difference was -1.12, with a 95% confidence interval from -2.06 to -0.17 and p = 0.021. The record flags a discrepancy with simple recalculations, so the sponsor’s own analysis takes precedence. The registry also lists secondary efficacy and safety outcomes, an open-label phase, and a self-administration substudy.

The trial’s original sample-size justification, multiplicity procedure, detailed missing-data methods, and full analysis plan are not reported in the registry record.