This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
ACHIEVE II was a randomized, double-blind, phase 3 parallel-group trial evaluating the efficacy, safety, and tolerability of oral ubrogepant in the acute treatment of migraine with or without aura. The registry reports 1,686 enrolled participants across three arms and two primary binary efficacy endpoints assessed 2 hours after the initial dose.
| Feature | ACHIEVE II |
|---|---|
| Phase | Phase 3 |
| Condition | Migraine, with or without aura |
| Brief title | Efficacy, Safety, and Tolerability of Oral Ubrogepant in the Acute Treatment of Migraine |
| Design | Randomized, double-blind, parallel-group |
| Primary purpose | Treatment |
| Enrollment | 1,686 |
| Primary endpoints | Two binary endpoints assessed 2 hours after the initial dose |
| Primary analysis method | Logistic regression |
| Primary effect measure | Odds ratio |
| Hypothesis type | Superiority |
| Status | Completed |
2. Clinical Question
The central question was whether oral ubrogepant improved acute migraine outcomes compared with placebo when evaluated 2 hours after the initial dose. The registry specified two primary endpoints: pain freedom and absence of the participant-identified most bothersome migraine-associated symptom.
Population
Participants with migraine, with or without aura, enrolled in the phase 3 ACHIEVE II trial.
Intervention
Ubrogepant, evaluated at 25 mg and 50 mg doses.
Comparator
Placebo-matching ubrogepant.
Primary question
Does ubrogepant increase the odds of achieving the prespecified binary efficacy outcomes 2 hours after the initial dose?
3. Trial Design
Placebo
- Placebo-matching ubrogepant
- Randomized comparator arm
- Serious adverse events: 0/499
Ubrogepant 25 mg
- Ubrogepant 25 mg
- Compared with placebo for each posted primary analysis
- Serious adverse events: 1/478
Ubrogepant 50 mg
- Ubrogepant 50 mg
- Compared with placebo for each posted primary analysis
- Serious adverse events: 0/488
The registry describes the allocation as randomized, the design model as parallel, the masking as double, and the primary purpose as treatment. The registry does not report a separate randomization ratio in the ClinicalTrials.gov record.
4. Trial Timeline
Trial start
The registry lists August 26, 2016 as the study start date.
Primary completion
The registry lists January 25, 2018 as the primary completion date.
Registry status
The ClinicalTrials.gov record lists the study status as completed and reports statistical analyses and outcome measures.
5. Primary Endpoints
| Endpoint | Registry definition | Time frame | Type |
|---|---|---|---|
| Pain freedom | Percentage of participants with pain freedom at 2 hours after initial dose. Pain freedom was defined as a reduction in headache pain severity from moderate/severe at baseline to no pain at 2 hours after the initial dose of investigational product. | Baseline (Predose) to 2 hours after initial dose | Binary |
| Absence of most bothersome migraine-associated symptom | Percentage of participants with absence of the most bothersome migraine-associated symptom identified at baseline at 2 hours after initial dose. The symptom could be photophobia, phonophobia, or nausea and was identified by the participant as the most bothersome symptom at pre-dose baseline. | Baseline (Predose) to 2 hours after initial dose | Binary |
The pain-freedom endpoint was based on electronic diary headache-severity assessments. The most-bothersome-symptom endpoint used electronic diary recording of the presence or absence of migraine-associated symptoms.
6. Statistical Methodology
Modified intent-to-treat population
The primary pain-freedom analyses used a modified intent-to-treat population consisting of randomized participants who received at least 1 dose of study drug, recorded a baseline migraine headache severity measurement, and had at least 1 postdose migraine headache severity or migraine-associated symptom measurement at or before the 2-hour timepoint. The registry specifies last observation carried forward (LOCF) in this population definition.
The primary most-bothersome-symptom analyses similarly used a modified intent-to-treat population. The registry describes participants as randomized, exposed to investigational product, with baseline migraine headache severity recorded and at least 1 relevant postdose measurement at or before 2 hours, with LOCF.
Logistic regression
Each primary endpoint was binary: a participant either met the endpoint definition or did not. Logistic regression is therefore a natural modeling framework because it directly models the probability of a binary outcome through the log odds.
Exponentiating a treatment coefficient gives an odds ratio. In this trial, the registry reports odds ratios comparing each ubrogepant dose with placebo.
Covariate adjustment
For the pain-freedom analyses, the model included treatment group, historical triptan response, medication use for migraine prevention, and baseline headache severity as explanatory variables. For the most-bothersome-symptom analyses, the model additionally included the underlying symptom.
| Primary analysis | Explanatory variables |
|---|---|
| Pain freedom | Treatment group; historical triptan response; medication use for migraine prevention; baseline headache severity |
| Most bothersome symptom absence | Treatment group; historical triptan response; migraine prevention medication; baseline headache severity; underlying symptom |
Odds ratio as the effect measure
The registry reports odds ratios rather than risk ratios or absolute percentage-point differences. An odds ratio compares the odds of meeting the endpoint in one treatment group with the odds in the reference group.
An OR greater than 1 indicates higher estimated odds of the favorable binary endpoint in the ubrogepant group relative to placebo. The OR is not itself a percentage-point improvement in the endpoint rate.
Superiority framework
The posted primary analyses were labeled as superiority analyses. Thus, the statistical question was whether the odds of achieving the endpoint differed favorably from the placebo comparison rather than whether ubrogepant satisfied a prespecified non-inferiority margin.
7. Primary Results: Pain Freedom at 2 Hours
The first primary endpoint was the percentage of participants with pain freedom at 2 hours after the initial dose. The registry reports separate logistic-regression analyses comparing placebo with each ubrogepant dose.
| Comparison | Odds ratio | 95% CI | P-value | Hypothesis |
|---|---|---|---|---|
| Placebo vs Ubrogepant 25 mg | 1.56 | 1.09–2.22 | 0.0285 | Superiority |
| Placebo vs Ubrogepant 50 mg | 1.62 | 1.14–2.29 | 0.0129 | Superiority |
Ubrogepant 25 mg vs placebo
95% CI: 1.09–2.22 · P = 0.0285
The adjusted odds of pain freedom at 2 hours were estimated to be 1.56 times the placebo odds.
An odds ratio of 1.56 means that the modeled odds of pain freedom were 56% higher with ubrogepant 25 mg than with placebo. This is an odds comparison, not a statement that 56% more participants became pain-free.
The 95% CI of 1.09–2.22 describes the uncertainty around the estimated odds ratio under the analysis model. Because the interval is entirely above 1, the reported estimate is consistent with higher odds of pain freedom in the ubrogepant 25 mg group.
The P-value of 0.0285 addresses evidence against the null hypothesis in the specified statistical framework; it does not measure the size or clinical importance of the treatment effect. The magnitude of the estimated association is conveyed by the odds ratio and its confidence interval.
The result is based on the specified mITT population and adjusted logistic-regression model, including historical triptan response, migraine-prevention medication use, and baseline headache severity.
Ubrogepant 50 mg vs placebo
95% CI: 1.14–2.29 · P = 0.0129
The adjusted odds of pain freedom at 2 hours were estimated to be 1.62 times the placebo odds.
An odds ratio of 1.62 corresponds to estimated odds that were 62% higher for pain freedom with ubrogepant 50 mg than with placebo. Again, this should not be translated directly into a 62-percentage-point difference in pain-freedom rates.
The 95% CI of 1.14–2.29 gives the precision of the modeled odds-ratio estimate. It remains above 1 throughout the reported interval, while the width of the interval shows that the exact magnitude of the association is uncertain.
The P-value of 0.0129 provides evidence against the null hypothesis under the reported superiority analysis. It does not quantify how large the treatment effect is or how clinically important it is.
Because two dose-versus-placebo comparisons were made for this primary endpoint, interpretation also requires attention to the trial's multiplicity framework. The ClinicalTrials.gov record identifies the analyses as superiority tests but does not provide an additional multiplicity-adjustment description.
8. Primary Results: Absence of the Most Bothersome Migraine-Associated Symptom
The second primary endpoint assessed whether the migraine-associated symptom identified by the participant as most bothersome at baseline was absent at 2 hours after the initial dose.
| Comparison | Odds ratio | 95% CI | P-value | Hypothesis |
|---|---|---|---|---|
| Placebo vs Ubrogepant 25 mg | 1.37 | 1.02–1.83 | 0.0711 | Superiority |
| Placebo vs Ubrogepant 50 mg | 1.65 | 1.25–2.20 | 0.0129 | Superiority |
Ubrogepant 25 mg vs placebo
95% CI: 1.02–1.83 · P = 0.0711
The adjusted odds of absence of the most bothersome baseline symptom were estimated to be 1.37 times the placebo odds.
An odds ratio of 1.37 represents estimated odds 37% higher with ubrogepant 25 mg than with placebo. It does not mean that the endpoint rate was 37 percentage points higher.
The 95% CI of 1.02–1.83 is relatively close to the null value of 1 at its lower boundary. That indicates that the estimated treatment effect is less precisely separated from the null than the corresponding 50 mg result.
The P-value of 0.0711 is not a measure of effect size. It summarizes the statistical evidence against the null hypothesis under the specified model and analysis framework. The confidence interval and odds ratio should be considered alongside it.
The underlying symptom was included as an explanatory variable, which is particularly relevant because the most bothersome symptom could be photophobia, phonophobia, or nausea.
Ubrogepant 50 mg vs placebo
95% CI: 1.25–2.20 · P = 0.0129
The adjusted odds of absence of the most bothersome baseline symptom were estimated to be 1.65 times the placebo odds.
An odds ratio of 1.65 corresponds to estimated odds 65% higher with ubrogepant 50 mg than with placebo. It is not equivalent to a 65-percentage-point increase in the probability of symptom absence.
The 95% CI of 1.25–2.20 indicates that the estimate is separated from the null value of 1 throughout the reported interval, while still allowing a range of plausible effect magnitudes.
The P-value of 0.0129 addresses statistical evidence against the null under the reported superiority analysis. It does not tell us whether the effect is large or small in absolute terms.
Because the endpoint is defined around the symptom each participant identified as most bothersome at baseline, the result summarizes a participant-specific symptom target rather than one fixed symptom across all participants.
9. Primary Results at a Glance
The four primary analyses all produced point estimates above 1. The registry reports confidence intervals and P-values for each comparison. The estimates are adjusted odds ratios from logistic regression rather than crude event-rate ratios.
10. Secondary Endpoint Results
Pain Relief at 2 Hours
Pain relief at 2 hours was evaluated as a secondary binary endpoint. The registry reports logistic-regression odds ratios comparing each ubrogepant dose with placebo.
| Comparison | Odds ratio | 95% CI | P-value |
|---|---|---|---|
| Placebo vs Ubrogepant 25 mg | 1.65 | 1.25–2.17 | 0.0711 |
| Placebo vs Ubrogepant 50 mg | 1.77 | 1.35–2.32 | 0.0129 |
The analyses used the mITT population and logistic regression. The model included treatment group, historical triptan response, medication use for migraine prevention, and baseline headache severity.
Sustained Pain Relief From 2 to 24 Hours
| Comparison | Odds ratio | 95% CI | P-value |
|---|---|---|---|
| Placebo vs Ubrogepant 25 mg | 1.82 | 1.33–2.48 | 0.0711 |
| Placebo vs Ubrogepant 50 mg | 2.16 | 1.59–2.92 | 0.0129 |
This endpoint was assessed over the interval from 2 to 24 hours after the initial dose. The registry describes the analysis population as mITT participants with determinable cases.
Sustained Pain Freedom From 2 to 24 Hours
| Comparison | Odds ratio | 95% CI | P-value |
|---|---|---|---|
| Placebo vs Ubrogepant 25 mg | 1.62 | 1.04–2.53 | 0.0711 |
| Placebo vs Ubrogepant 50 mg | 1.85 | 1.20–2.83 | 0.0129 |
These estimates concern sustained pain freedom from 2 to 24 hours rather than the single 2-hour pain-freedom assessment. The distinction is statistically important because a sustained endpoint requires the favorable state to persist over a longer interval.
Absence of Photophobia at 2 Hours
| Comparison | Odds ratio | 95% CI | P-value |
|---|---|---|---|
| Placebo vs Ubrogepant 25 mg | 1.28 | 0.96–1.72 | 0.1833 |
| Placebo vs Ubrogepant 50 mg | 1.52 | 1.14–2.02 | 0.0167 |
Absence of Phonophobia at 2 Hours
| Comparison | Odds ratio | 95% CI | P-value |
|---|---|---|---|
| Placebo vs Ubrogepant 25 mg | 1.38 | 1.04–1.83 | 0.1066 |
| Placebo vs Ubrogepant 50 mg | 1.39 | 1.05–1.84 | 0.0440 |
Absence of Nausea at 2 Hours
| Comparison | Odds ratio | 95% CI | P-value |
|---|---|---|---|
| Placebo vs Ubrogepant 25 mg | 1.10 | 0.81–1.49 | 0.9522 |
| Placebo vs Ubrogepant 50 mg | 1.12 | 0.83–1.51 | 0.9522 |
The nausea analyses illustrate why individual secondary endpoints should not be collapsed into a single overall efficacy conclusion. The reported point estimates were close to 1, and the confidence intervals included 1 for both dose comparisons.
11. Secondary Results: Statistical Interpretation
The secondary analyses show a range of odds-ratio estimates rather than a single uniform effect. For example, the reported odds ratio for sustained pain relief was 2.16 for ubrogepant 50 mg versus placebo, while the corresponding estimate for absence of nausea was 1.12.
These measures address different clinical states. Sustained pain relief requires maintenance of relief from 2 to 24 hours, whereas absence of nausea asks whether a specific symptom is absent at the 2-hour assessment.
The confidence intervals are therefore especially useful for understanding precision. An interval that includes 1 is compatible with no difference under the model, while an interval entirely above 1 indicates that the reported estimate is separated from the null value.
12. Missing Data and Last Observation Carried Forward
Several analyses explicitly reference last observation carried forward (LOCF). For the primary pain-freedom endpoint, the mITT population required at least one relevant postdose measurement at or before the 2-hour timepoint, with LOCF included in the analysis-population description. The primary most-bothersome-symptom analyses use the same general approach.
LOCF replaces a missing later observation with a participant's most recent available observation. The attraction is simplicity, but the method relies on an assumption about what the carried-forward measurement represents after the missingness occurs.
The missing-data rule is part of the estimand definition because it can affect which participants are classified as meeting the binary endpoint. Consequently, the odds ratio reflects both the treatment comparison and the specified analysis population and missing-data procedure.
13. Why Logistic Regression Was Used
Both registered primary endpoints are binary percentages: pain freedom at 2 hours and absence of the most bothersome symptom at 2 hours. Logistic regression is designed for exactly this type of outcome.
Binary outcome
Each participant is classified according to whether the prespecified endpoint was achieved. Logistic regression models this probability while allowing covariate adjustment.
Covariate adjustment
The reported models included baseline and historical factors such as historical triptan response, migraine-prevention medication use, and baseline headache severity.
Odds ratio
The treatment coefficient can be exponentiated to produce an odds ratio comparing the treatment and placebo groups.
Interpretation
The odds ratio is a relative measure of the odds of the endpoint. It should not be confused with an absolute difference in endpoint percentages.
14. Statistical Methods Explained
Why was logistic regression appropriate for the primary endpoints?
Both primary endpoints are binary outcomes measured at a fixed postdose timepoint. Logistic regression allows the probability of meeting the endpoint to be related to treatment while simultaneously accounting for the explanatory variables listed in the registry.
What does an odds ratio of 1.62 mean?
An odds ratio of 1.62 means that the estimated odds of the endpoint in the ubrogepant 50 mg group were 1.62 times the odds in the placebo group. Equivalently, the modeled odds were 62% higher. This does not mean that the probability was 62% higher, because odds and probability are different quantities.
Why is an odds ratio different from a risk ratio?
A risk ratio compares probabilities directly, whereas an odds ratio compares odds. When an endpoint is common, the numerical difference between these measures can become substantial. Without the underlying event probabilities, an odds ratio should not be converted into a risk ratio by simple arithmetic.
Why does the confidence interval matter?
The confidence interval describes statistical uncertainty around the estimated treatment effect under the specified model. For example, the 95% CI of 1.14–2.29 for pain freedom with ubrogepant 50 mg gives a range of values compatible with the analysis framework and sampling uncertainty. It does not describe the range of individual patient responses.
Why does the P-value not measure effect size?
A P-value summarizes evidence against a null hypothesis under a specified model and sampling framework. It depends on both the observed effect and the amount of information in the analysis. The odds ratio describes the estimated relative effect, while the confidence interval describes its precision.
What does the mITT definition imply?
The mITT population is narrower than simply taking every randomized participant because participants needed to receive study drug and have the specified baseline and postdose measurements. This can improve alignment between the analysis population and the endpoint assessment, but it also means the primary analysis is not identical to an analysis of every randomized participant regardless of postdose data availability.
Why does the underlying symptom enter the most-bothersome-symptom model?
The most bothersome symptom could be photophobia, phonophobia, or nausea. Including the underlying symptom as an explanatory variable accounts for this baseline distinction in the logistic-regression analysis rather than treating all participants as though they started with the same symptom.
15. Odds Ratios, Probability, and Absolute Effects
A recurring statistical issue in binary-endpoint trials is the difference between an odds ratio and an absolute probability difference. Suppose two groups had probabilities p1 and p0. Their odds ratio would be:
Because the odds depend nonlinearly on probability, the same odds ratio can correspond to different absolute probability differences depending on the baseline probability in the placebo group.
This is why the registry's reported odds ratios should be read as relative model-based effects. An odds ratio alone does not tell the reader how many additional participants achieved the endpoint without knowing the underlying event probabilities.
For ACHIEVE II, the reported ORs allow the reader to compare the relative odds of each binary endpoint between ubrogepant and placebo. They do not, by themselves, provide the absolute treatment benefit in percentage points. Keeping those quantities conceptually separate prevents overinterpretation of the regression coefficient.
16. Comparing the 25 mg and 50 mg Estimates
The registry reports separate placebo comparisons for 25 mg and 50 mg ubrogepant. The point estimates are useful for describing the observed dose-specific comparisons, but they do not constitute a direct statistical test that 50 mg is superior to 25 mg.
| Primary endpoint | 25 mg OR | 50 mg OR |
|---|---|---|
| Pain freedom at 2 hours | 1.56 | 1.62 |
| Absence of most bothersome symptom at 2 hours | 1.37 | 1.65 |
17. Safety Results
The registry reports serious adverse events by randomized arm as affected participants divided by participants at risk.
| Arm | Serious adverse events | Affected / at risk |
|---|---|---|
| Placebo | Serious adverse events | 0/499 |
| Ubrogepant 25 mg | Serious adverse events | 1/478 |
| Ubrogepant 50 mg | Serious adverse events | 0/488 |
Placebo
0 affected participants among 499 at risk were reported for serious adverse events.
Ubrogepant 25 mg
1 affected participant among 478 at risk was reported for serious adverse events.
Ubrogepant 50 mg
0 affected participants among 488 at risk were reported for serious adverse events.
Statistical context
These counts are safety observations and are not equivalent to the efficacy odds ratios reported for the primary endpoints.
With only the reported affected/at-risk counts, the safety evidence is most appropriately described directly rather than converted into an unreported comparative safety model.
18. Randomization and Blinding
Randomization is central to the causal interpretation of the efficacy comparisons because it creates the treatment groups through an allocation process rather than through participant or investigator selection. The registry identifies the allocation as randomized and the masking as double.
Randomization helps balance measured and unmeasured prognostic factors in expectation. The resulting treatment comparison can therefore be interpreted as a randomized comparison rather than simply an observational association, subject to the trial's analysis population and conduct.
Double masking is also relevant to subjective or participant-reported outcomes. Pain severity and migraine-associated symptoms were recorded using electronic diaries, so knowledge of treatment assignment could potentially affect reporting. Masking is therefore a design feature directly relevant to interpretation of these endpoints.
19. Analysis Population and Estimand
The primary endpoints are framed around the effect of randomized treatment on binary outcomes assessed shortly after the initial dose, within the registry-defined mITT population and missing-data approach.
| Component | ACHIEVE II definition |
|---|---|
| Population | Modified intent-to-treat population defined by randomization, receipt of at least 1 dose, baseline assessment, and relevant postdose assessment |
| Treatment comparison | Placebo versus ubrogepant 25 mg or 50 mg |
| Outcome | Binary migraine outcome |
| Time point | 2 hours after initial dose for the primary endpoints |
| Effect measure | Odds ratio |
| Statistical model | Logistic regression |
This framework makes clear why the odds ratio should not be interpreted as a universal effect of ubrogepant on migraine over an unspecified period. It is an estimate tied to a particular population, treatment comparison, endpoint definition, time point, and model.
20. Primary Results: What the Confidence Intervals Say
| Analysis | Estimate | 95% CI | Precision observation |
|---|---|---|---|
| Pain freedom, 25 mg | 1.56 | 1.09–2.22 | Interval remains above 1 but spans a substantial range of possible odds ratios |
| Pain freedom, 50 mg | 1.62 | 1.14–2.29 | Interval remains above 1 |
| Most bothersome symptom absence, 25 mg | 1.37 | 1.02–1.83 | Lower confidence limit is close to 1 |
| Most bothersome symptom absence, 50 mg | 1.65 | 1.25–2.20 | Interval remains above 1 |
The confidence intervals should be read as measures of statistical precision rather than as a distribution of individual treatment responses. The width of each interval reflects uncertainty around the estimated odds ratio; the position of the interval relative to 1 describes its relationship to the null value for an odds ratio.
21. P-values in Context
The four primary analyses have reported P-values of 0.0285, 0.0129, 0.0711, and 0.0129. These values should be interpreted in the context of the endpoint definitions, treatment comparisons, logistic-regression models, and the overall multiplicity structure of the trial.
| Primary analysis | P-value | What it addresses |
|---|---|---|
| Pain freedom, 25 mg | 0.0285 | Evidence against the null under the reported logistic-regression superiority analysis |
| Pain freedom, 50 mg | 0.0129 | Evidence against the null under the reported logistic-regression superiority analysis |
| Most bothersome symptom absence, 25 mg | 0.0711 | Evidence against the null under the reported logistic-regression superiority analysis |
| Most bothersome symptom absence, 50 mg | 0.0129 | Evidence against the null under the reported logistic-regression superiority analysis |
The P-value does not say that a treatment effect is 0.0129 units large, nor does it state the probability that the treatment is effective. The odds ratio and its confidence interval carry the information about the estimated relative effect and its statistical precision.
22. Important Limitations and Interpretation Issues
- Odds ratios are not absolute effects: an OR cannot be read directly as a percentage-point difference in the probability of achieving the endpoint.
- mITT population: the primary efficacy analyses required treatment exposure and specified baseline and postdose measurements, so the analyzed population is narrower than a pure all-randomized ITT population.
- Missing-data handling: LOCF appears in the registry description of the primary analysis populations. Any missing-data method relies on assumptions about how missing observations relate to the outcome.
- Multiplicity: the trial includes multiple primary and secondary analyses. Individual nominal P-values need to be interpreted within the overall testing strategy.
- Dose comparison: separate placebo comparisons do not establish a statistically significant difference between 25 mg and 50 mg.
- Secondary endpoints: the secondary analyses address different clinical states and should not be combined into one summary effect.
- Safety counts: the reported serious-adverse-event counts are sparse, so descriptive reporting is more transparent than imposing an unreported comparative model.
- Time-specific interpretation: the primary endpoints concern the 2-hour assessment after the initial dose; the results should not automatically be generalized to an unspecified duration of treatment.
23. Why This Trial Matters Statistically
ACHIEVE II is a useful teaching case because it demonstrates how randomized clinical-trial statistics change when the primary outcome is binary rather than time-to-event. Instead of Kaplan-Meier estimation and Cox regression, the primary analysis uses logistic regression and odds ratios.
| Concept | How it appears in ACHIEVE II |
|---|---|
| Randomization | Randomized allocation across three parallel treatment arms |
| Blinding | Double-blind design |
| Binary endpoints | Pain freedom and symptom absence are defined as yes/no outcomes |
| Logistic regression | Primary and reported secondary binary outcomes are analyzed using logistic regression |
| Odds ratio | Reported effect measure for treatment comparisons |
| Confidence interval | 95% two-sided intervals quantify uncertainty around the OR estimates |
| P-value | Reported for each posted statistical comparison |
| mITT analysis | Primary efficacy analyses require treatment exposure and specified outcome measurements |
| Missing-data handling | LOCF appears in the registry description of several analysis populations |
| Multiplicity | Multiple primary and secondary treatment comparisons require interpretation within the overall testing strategy |
24. Statistical Interpretation vs Clinical Interpretation
Statistical interpretation
The posted primary analyses used logistic regression and reported odds ratios above 1 for both ubrogepant doses on both primary endpoints. The reported confidence intervals quantify uncertainty around those estimates, and the P-values quantify evidence against the corresponding null hypotheses under the specified analyses.
Clinical interpretation
The primary endpoints address whether participants were pain-free and whether their most bothersome baseline migraine-associated symptom was absent at 2 hours. Translating the odds ratios into clinical magnitude requires the underlying endpoint probabilities, which are not represented by the OR alone.
25. What the Primary Odds Ratios Do — and Do Not — Mean
The reported OR of 1.62 means that the modeled odds of pain freedom at 2 hours were 1.62 times the placebo odds for the ubrogepant 50 mg comparison. It does not mean that 62% of participants achieved pain freedom, nor that the absolute probability increased by 62 percentage points.
The reported OR of 1.65 means that the modeled odds of absence of the participant-identified most bothersome migraine-associated symptom were 1.65 times the placebo odds. It does not imply that 65% more participants experienced symptom absence.
The corresponding 95% confidence intervals of 1.14–2.29 and 1.25–2.20 show that the point estimates are not exact measurements. They quantify statistical uncertainty around the model-based estimates and should be considered alongside the point estimates rather than replaced by them.
26. Related Tutorials
Learn more about the methods used in this trial:
27. Related Calculators
28. Sources
- ClinicalTrials.gov: ACHIEVE II, NCT02867709.
- PubMed: PMID 39982105.
- PubMed: PMID 36125279.
- PubMed: PMID 35468729.
- PubMed: PMID 34874514.
- PubMed: PMID 33874756.
Continue through the Clinical Biostats knowledge graph
Connect this trial's binary endpoints, regression model, effect measures, and uncertainty concepts to deeper statistical tutorials and analytical tools.
29. Record Summary
ACHIEVE II provides a clear example of statistical analysis for randomized trials with binary efficacy endpoints. The two primary outcomes were assessed 2 hours after the initial dose and analyzed using logistic regression in a registry-defined modified intent-to-treat population. The reported effect measure was the odds ratio, with two-sided 95% confidence intervals and P-values for each dose-versus-placebo comparison.
The primary pain-freedom analyses produced odds ratios of 1.56 for ubrogepant 25 mg and 1.62 for ubrogepant 50 mg. For absence of the most bothersome migraine-associated symptom, the corresponding odds ratios were 1.37 and 1.65. The broader set of secondary analyses illustrates why effect estimates, confidence intervals, P-values, endpoint definitions, missing-data methods, and multiplicity all need to be considered together.