This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Efficacy and Safety of Eslicarbazepine Acetate as Preventive Therapy for Subjects With Migraine (NCT01820559), a randomized phase 2 study. The primary outcome is Absolute Change From Baseline in the Frequency of Migraine Attacks.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT01820559 |
| Phase | Phase 2 |
| Condition | Migraine |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Double |
| Enrollment | 452 |
| Arms in the primary outcome | 3 |
| Lead sponsor | Bial - Portela C S.A. |
| Results posted | Yes |
2. Clinical Question
Population
Subjects with migraine, with or without aura, were eligible for this multinational study. The registry record lists Migraine as the condition. It does not report further eligibility details or a more specific description of the population used for the primary outcome.
Intervention
Participants were assigned to eslicarbazepine acetate, abbreviated ESL in the registry summary, at 800 mg/day or 1200 mg/day, each taken once daily. The treatment was described as prophylactic therapy, and the primary outcome concerned migraine-attack frequency during the Maintenance Period.
Comparator
The comparator was placebo tablets. The registry summary states that participants were randomized in a 1:1:1 ratio among placebo, ESL 800 mg/day, and ESL 1200 mg/day. The posted primary outcome includes a placebo mean and a separate mean for each ESL dose group.
Primary question
The primary question was whether the absolute change from baseline in migraine-attack frequency, standardized to 4 weeks, differed between each ESL dose group and placebo during the Maintenance Period. The outcome was recorded in participant diaries. The posted arm means are −0.8 attacks per participant for placebo and −1.0 for each ESL dose. Recalculated unadjusted mean differences compare each active group with placebo; those estimates are −0.2 attacks per participant for both comparisons. These descriptive and recalculated results address the recorded outcome, but the registry record does not report a sponsor inferential analysis for it.
3. Trial Design
This was a phase 2, randomized, double-blind, placebo-controlled, parallel-group, multicentre clinical trial. The registry title and brief summary describe a multinational study, and the brief summary states that the study evaluated efficacy, safety, and tolerability of multiple ESL doses as prophylactic treatment for subjects with migraine with or without aura. The sponsor is listed as Bial - Portela C S.A. The registry record lists enrollment of 452, a start in 2009-04, primary completion in 2010-06, and status as completed.
Randomization assigned participants in a 1:1:1 ratio to placebo, ESL 800 mg/day once daily, or ESL 1200 mg/day once daily. The parallel-group structure means each participant was assigned to one group rather than crossing between the treatment groups. Double masking means the trial was described as double-blind; further details of who was masked and how masking was maintained are not reported in the registry record.
The posted primary outcome gives group counts of 133 for placebo, 134 for ESL 800 mg, and 136 for ESL 1200 mg. These are the counts attached to the outcome results; they are not identical in presentation to the overall enrollment figure, and the registry record does not explain the relationship between them. The adverse-event denominators are separately listed as 136, 135, and 139, respectively. The registry record does not report additional details about treatment duration, study visits, allocation concealment, site count, or the original sample-size justification. The record lists one outcome measure and no posted analysis entries.
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.
Overall Study
| Stage | Placebo | ESL 800 mg | ESL 1200 mg |
|---|---|---|---|
| STARTED | 136 | 135 | 139 |
| Randomized and Treated | 136 | 135 | 139 |
| Completed Tutration Period | 130 | 134 | 131 |
| Entered Maintenance Period | 129 | 133 | 129 |
| Completed Maintenance Period | 122 | 123 | 110 |
| COMPLETED | 122 | 122 | 110 |
| NOT COMPLETED | 14 | 13 | 29 |
Reasons for not completing this period:
| Reason | Placebo | ESL 800 mg | ESL 1200 mg |
|---|---|---|---|
| Adverse Event | 4 | 6 | 13 |
| Withdrawal by Subject | 1 | 2 | 5 |
| Protocol Violation | 2 | 1 | 3 |
| Subject's non-compliance | 1 | 0 | 2 |
| Lost to Follow-up | 3 | 2 | 2 |
| At sponsor request | 0 | 0 | 1 |
| Lack of Efficacy | 3 | 2 | 3 |
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.
| Characteristic | Unit | Placebo | ESL 800 mg | ESL 1200 mg | Total |
|---|---|---|---|---|---|
| Participants analysed | 136 | 135 | 139 | 410 | |
| Age, Customized: <=65 years | participants | 135 | 135 | 136 | 406 |
| Age, Customized: >65 years | participants | 1 | 0 | 3 | 4 |
| Sex: Female, Male: Female | Participants | 121 | 114 | 111 | 346 |
| Sex: Female, Male: Male | Participants | 15 | 21 | 28 | 64 |
6. Primary Endpoint
The registry lists the primary outcome as Absolute Change From Baseline in the Frequency of Migraine Attacks, measured over 4 weeks.
| Element | Registry entry |
|---|---|
| Result type | Least Squares Mean |
| Measure of spread | Standard Error |
| Unit | number of migraine attacks/participant |
| Endpoint classification | Continuous Endpoint |
The primary endpoint was the absolute change from baseline in the frequency of migraine attacks, standardized to 4 weeks and measured during the Maintenance Period. The unit is number of migraine attacks per participant. The registry describes the group results as least-squares means and gives a standard error as the measure of dispersion. The posted values are −0.8 for placebo and −1.0 for each ESL dose group.
The outcome relied on participant diaries. The registry description specifies how separate migraine events were counted: when less than 24 h separated the end of one migraine event from the start of the next, the events were considered one migraine attack. For two attacks to be counted separately in the statistical analysis, the diary had to record at least 24 h free from headache, pain, and migraine symptoms between them. Those rules matter because the endpoint is a count of attacks, not a count of every recorded event or symptom.
General: a standardized observation period makes attack frequency more comparable when diary observation periods may differ. Change from baseline expresses the direction and size of within-participant change over the measured period, while a comparison with placebo helps distinguish change during treatment from change in the control group. The registry record does not report baseline attack-frequency values, so the absolute changes cannot be used to recover the starting frequencies.
7. Results Posted in the Registry
- ESL 800 mg: mean -1.00 versus -0.80 on control; difference -0.20 (95% CI -0.60 to 0.20), standardized effect d = -0.12; z-test p = 0.322.
- ESL 1200 mg: mean -1.00 versus -0.80 on control; difference -0.20 (95% CI -0.59 to 0.19), standardized effect d = -0.12; z-test p = 0.320.
Mean difference: ESL 800 mg
95% CI: -0.60 to 0.20
Recalculated from the posted summary statistics
8. Statistical Interpretation
The posted least-squares mean change was −0.8 migraine attacks per participant in the placebo group and −1.0 in each ESL group over the 4-week timeframe. The negative signs indicate decreases relative to baseline as the endpoint is defined. The difference between the displayed group means is reflected in the recalculated unadjusted mean differences: −0.2 attacks per participant for ESL 800 mg versus placebo and −0.2 for ESL 1200 mg versus placebo. These are small differences on the endpoint's stated scale, and the two posted active-group means are the same at the displayed precision.
For ESL 800 mg, the recalculated unadjusted 95% confidence interval for the mean difference is −0.596 to 0.196, with p = 0.322. For ESL 1200 mg, the recalculated unadjusted p-value is 0.32; its confidence interval also crosses zero. The latter interval is not repeated numerically because its endpoint values are reported to more than 3 decimal places. In both comparisons, the interval includes no mean difference, so these recalculated summaries do not pin down the direction of the difference with precision sufficient to exclude zero.
The p-values are unadjusted recalculations, not results of an inferential analysis posted by the sponsor in the registry. They should be read with that distinction in mind. The registry record lists no analysis entries and no secondary outcomes. It does not report whether the primary comparison was part of a broader analysis plan, whether any multiplicity adjustment was intended, or what analysis population was used. The posted least-squares means also do not come with confidence intervals in the registry outcome listing.
These results describe the posted outcome and its recalculated unadjusted comparisons. They do not establish that either dose changes migraine outcomes in a way that would matter clinically, and they do not establish equivalence between treatment groups. A confidence interval that includes zero can be compatible with a range of effects in either direction, including no difference. The available summary statistics do not resolve that uncertainty. No conclusion about safety or overall benefit follows from the primary endpoint alone; the adverse-event figures are described separately and also have limitations.
The trial is recorded as completed, but completion does not itself indicate what the primary comparison concluded. The registry record does not report sponsor-authored statistical test results, an adjusted model estimate, or a detailed interpretation of the endpoint. The most direct reading is therefore descriptive: the average recorded change was negative in all three groups, and the recalculated unadjusted differences between each ESL group and placebo were −0.2 attacks per participant, with intervals that include zero.
9. Statistical Methods Explained
The primary outcome is a continuous measure summarized as a least-squares mean. General: a mean describes the average value in a group, while a least-squares mean is typically a model-based group mean adjusted for terms in a statistical model. The registry record does not state the model terms, covariates, or estimation details behind the posted least-squares means. It is therefore appropriate to describe the posted values as least-squares means without assuming what adjustments were made.
The arm-level dispersion is reported as a standard error. General: a standard error describes uncertainty in an estimated mean, not the spread of individual participant outcomes. A smaller standard error usually indicates a more precisely estimated mean, while a larger one indicates greater sampling uncertainty. The displayed standard errors are 0.143 for placebo, 0.143 for ESL 800 mg, and 0.142 for ESL 1200 mg. Those values are not repeated in the interpretation because the requested number-format limit excludes values with more than 3 decimal places.
The pairwise mean differences and confidence intervals are recalculated, unadjusted comparisons based on the posted summary statistics. The recalculation uses the placebo mean of −0.8 and the relevant active-group mean of −1.0, with the posted group counts. General: a difference of means expresses how far apart two group averages are, on the original outcome scale. Here, a negative difference means the active-group mean is lower than the placebo mean because the subtraction is active group minus placebo. It does not, by itself, indicate whether the difference is statistically or clinically important.
The calculations use an approximate pooled standard deviation reconstructed from the reported standard errors and group counts. The posted pooled standard deviations are 1.651 for the ESL 800 mg comparison and 1.65 for the ESL 1200 mg comparison, and the record marks the approximation as used. This is not the same as an adjusted model comparison. General: converting standard errors back to an estimate of participant-level spread relies on assumptions about how the standard errors were obtained; when the underlying model details are not available, the resulting spread and comparisons are approximate. This limitation matters when interpreting the recalculated intervals and p-values.
A confidence interval describes a range of values compatible with the estimate under the calculation's assumptions. In repeated use of the same procedure, intervals constructed in this way would cover the true parameter at the stated confidence level at the specified rate, provided the assumptions hold. It is not a probability statement that the particular true value lies in the interval. The recalculated 95% interval for ESL 800 mg versus placebo runs from −0.596 to 0.196. The ESL 1200 mg interval also spans zero. Both estimates are −0.2, but the uncertainty around each estimate means the data do not sharply distinguish a negative difference from zero or a positive difference.
The p-values are likewise recalculated from the summary statistics and are unadjusted: 0.322 for ESL 800 mg and 0.32 for ESL 1200 mg. General: a p-value measures how unusual data at least as incompatible with a specified null hypothesis would be if that null hypothesis and the analysis assumptions were true. It is not the probability that the null hypothesis is true, and it does not measure the size or importance of an effect. Because these are recalculated unadjusted values and not posted sponsor analysis results, they should not be represented as the trial's original inferential findings.
Why does over-dispersion matter for count outcomes?
General: count outcomes often vary more across participants than a simple Poisson model allows. This is called over-dispersion. If a model assumes less variability than the data actually contain, standard errors can be too small and confidence intervals too narrow. A count outcome may be analyzed with methods designed for counts, but the registry record does not report the model used for the primary endpoint or whether over-dispersion was assessed. The arm summaries alone cannot establish that the relevant count-data assumptions were met.
What can be inferred from the reported analysis information?
The registry lists one outcome measure and zero analysis entries. It does not report the original statistical analysis plan, model specification, handling of repeated observations, treatment of missing diary entries, or definition of the analysis population. The posted primary means are available, but the absence of those details limits assessment of how the sponsor's planned analysis was carried out. The recalculated comparisons provide a transparent, simple unadjusted reading of the posted summaries; they cannot replace an analysis that accounts for any prespecified design or data features not reported in the registry record.
10. Confidence Intervals and Precision
The arm means are displayed as −0.8 for placebo and −1.0 for each ESL dose. The recalculated mean differences are −0.2 attacks per participant for both comparisons. These point estimates are concise summaries, but an estimate alone does not show how much uncertainty surrounds it. That is why the confidence intervals are important: they indicate a range of values compatible with the estimate under the recalculation assumptions.
For ESL 800 mg versus placebo, the recalculated unadjusted confidence interval is −0.596 to 0.196. Its span includes zero and values on either side of zero. The ESL 1200 mg confidence interval also crosses zero; its exact endpoints are reported with more than 3 decimal places and are not reproduced. The posted p-values, 0.322 and 0.32, are consistent with the same broad reading: the recalculated comparisons do not provide a precise distinction from zero under an unadjusted analysis.
General: a wide interval signals that the data are compatible with a relatively broad range of possible effects. A narrow interval indicates greater precision, though it does not guarantee that the estimate is free of bias or that the underlying model is appropriate. Precision depends in part on the amount of information and the variability in the outcome. The approximate pooled standard deviations are 1.651 and 1.65 for the two comparisons, while the outcome counts are 133, 134, and 136. The posted information does not include the individual participant data needed to examine distributional shape or outliers.
The registry record does not provide confidence intervals for the arm-level least-squares means, nor does it report an adjusted treatment contrast with its own interval. The recalculated intervals are therefore useful for understanding uncertainty in simple unadjusted comparisons, but they should not be mistaken for sponsor-reported model-based intervals. Their interpretation depends on the summary-statistic approximation and its assumptions.
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.
| Effect to detect | Participants per arm, 80% power | Participants per arm, 90% power |
|---|---|---|
| 100% of the observed effect | 1,070 | 1,432 |
| 75% of the observed effect | 1,901 | 2,545 |
| 50% of the observed effect | 4,277 | 5,725 |
The planning figures are participants per arm for 80% power and participants per arm for 90% power. They are not total event counts. The planning illustration uses a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. For each scenario, the effect to detect is the stated fraction of the observed effect. The trial's own original sample-size assumptions and justification are not reported in the registry record.
For the ESL 800 mg comparison, planning for 100% of the observed effect gives 1,070 participants per arm for 80% power and 1,432 participants per arm for 90% power. Planning for 75% of the observed effect gives 1,901 participants per arm for 80% power and 2,545 participants per arm for 90% power. Planning for 50% of the observed effect gives 4,277 participants per arm for 80% power and 5,725 participants per arm for 90% power. The ESL 1200 mg comparison has corresponding planning figures of 1,069 and 1,430 for 100% of the observed effect, 1,899 and 2,542 for 75%, and 4,273 and 5,720 for 50%, respectively.
General: power is the chance that a specified analysis will reject its null hypothesis if a particular effect is the true effect and the planning assumptions hold. The 80% and 90% columns describe different target probabilities under those assumptions; the higher target generally requires more participants. Smaller effects need more participants because they are harder to distinguish from sampling variation. The curve should be read as a visual representation of how the required sample size changes with the effect being targeted, not as observed trial results. These are planning illustrations based on the observed effect, not observed power.
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 z1−α/2 = 1.960) and the desired power (z1−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
- Observed difference Δ = -0.200; pooled standard deviation σ = 1.651.
- n = 2 × (1.96 + 0.842)2 × 1.6512 / -0.2002 = 1,069.1, rounded up to 1,070 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 <- -0.2 # observed difference in means
sigma <- 1.65052774915085 # 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(1820559)
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 = 1,070 (80% power), 1,432 (90% power)
75% of observed effect: n per arm = 1,901 (80% power), 2,545 (90% power)
50% of observed effect: n per arm = 4,277 (80% power), 5,725 (90% power)
power.t.test: 1070.1
Simulated power at n = 1,070 per arm: 0.797 (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. Multiplicity, Missing Data and Analysis Populations
The registry record lists one outcome measure and no secondary outcomes, with zero analysis entries. This does not establish that the trial measured only one outcome or that only one statistical test was planned; it describes what is posted in the registry record. The record does not report whether the primary outcome was analyzed at more than one timepoint, whether multiple treatment contrasts were part of the original plan, or whether an adjustment for multiple comparisons was planned or used.
The recalculated p-values are unadjusted. Because two active-dose comparisons with placebo are shown, readers should not treat those values as multiplicity-adjusted evidence. General: when several hypotheses are tested, the chance of at least one small p-value under a collection of null hypotheses can increase. Statistical plans may address that issue through procedures such as adjustment or a prespecified testing sequence. Whether such methods were used in this trial is not reported in the registry record. No adjustment should be inferred from the posted summaries.
Missing-data handling is also not reported in the registry record. The primary outcome depended on participant diaries, but the record does not explain how incomplete diaries, missing visits, or participants without a usable outcome were handled. It does not state whether imputation was performed or what assumptions were made about missing observations. The outcome counts are 133, 134, and 136 across the three groups, while overall enrollment is listed as 452; the record does not explain the difference in these reported counts. These figures alone do not establish why participants are not represented in the outcome results.
The analysis population is not specified. The registry record does not say whether the results are based on all randomized participants, participants who received treatment, participants with available diary data, or another defined group. Nor does it report whether the analysis followed assigned treatment regardless of adherence. These omissions limit assessment of potential selection effects and of how closely the posted outcome represents the randomized comparison.
14. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Placebo | 1 | 136 | 0.7% |
| ESL 800 mg | 1 | 135 | 0.7% |
| ESL 1200 mg | 4 | 139 | 2.9% |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 2% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Placebo | 46 of 136 |
| ESL 800 mg | 67 of 135 |
| ESL 1200 mg | 74 of 139 |
Most frequent serious adverse event terms
| Event term | Organ system | Placebo | ESL 800 mg | ESL 1200 mg |
|---|---|---|---|---|
| VIRAL INFECTION | Infections and infestations | 1 of 136 | 0 of 135 | 1 of 139 |
| STATUS MIGRAINOSUS | Nervous system disorders | 0 of 136 | 1 of 135 | 1 of 139 |
| VENTRICULAR ARRHYTHMIA | Cardiac disorders | 0 of 136 | 0 of 135 | 1 of 139 |
| ADJUSTMENT DISORDER | Psychiatric disorders | 0 of 136 | 0 of 135 | 1 of 139 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Placebo | ESL 800 mg | ESL 1200 mg |
|---|---|---|---|---|
| Somnolence | Nervous system disorders | 1 of 136 | 9 of 135 | 13 of 139 |
| Dizziness | Nervous system disorders | 2 of 136 | 6 of 135 | 14 of 139 |
| Nausea | Gastrointestinal disorders | 4 of 136 | 6 of 135 | 10 of 139 |
| Vertigo | Ear and labyrinth disorders | 4 of 136 | 4 of 135 | 11 of 139 |
| Nasopharyngitis | Infections and infestations | 6 of 136 | 5 of 135 | 4 of 139 |
| Migraine | Nervous system disorders | 3 of 136 | 2 of 135 | 6 of 139 |
| Vomiting | Gastrointestinal disorders | 3 of 136 | 2 of 135 | 3 of 139 |
| Fatigue | General disorders | 1 of 136 | 4 of 135 | 3 of 139 |
| Influenza | General disorders | 2 of 136 | 3 of 135 | 3 of 139 |
| Abdominal pain upper | Gastrointestinal disorders | 3 of 136 | 3 of 135 | 2 of 139 |
Serious events with the highest arm-level rate: status migrainosus (0.7%), viral infection (0.7%), ventricular arrhythmia (0.7%), adjustment disorder (0.7%).
The registry lists serious adverse events affecting 1 participant in the placebo group, 1 in the ESL 800 mg group, and 4 in the ESL 1200 mg group. The denominators at risk are 136, 135, and 139, respectively. These are participant counts, not counts of all serious-event episodes; the record does not provide further detail about event timing, duration, severity beyond the serious classification, or relationship to study treatment. The registry record does not report death counts.
Other adverse events are listed as affecting 46 participants in the placebo group, 67 in the ESL 800 mg group, and 74 in the ESL 1200 mg group, using the same group-specific denominators. The record does not provide a detailed definition of this category or state whether a participant could be counted more than once for different events. Those figures should therefore be read as participants affected, not as event totals.
The registry lists the serious-event terms STATUS MIGRAINOSUS, VIRAL INFECTION, VENTRICULAR ARRHYTHMIA, and ADJUSTMENT DISORDER, each with 0.7%. The record does not assign these terms to particular groups in the posted term list. These summary counts do not establish causation or provide a complete comparative safety analysis. The registry record does not report a formal statistical comparison of adverse-event rates.
15. What the Result Does, and Does Not, Mean
The posted outcome documents a decrease from baseline in average migraine-attack frequency in all three groups over the 4-week timeframe. The placebo least-squares mean change was −0.8 attacks per participant, and the mean change was −1.0 in each ESL dose group. The recalculated unadjusted difference for each active dose compared with placebo was −0.2 attacks per participant. The confidence intervals cross zero, so the recalculated summaries remain compatible with no mean difference as well as differences in either direction.
This is a description of the posted trial outcome, not a conclusion that an intervention does or does not work. A point estimate does not establish whether a difference is important to participants, and the registry record does not provide a threshold for a clinically meaningful change. It also does not report the baseline attack frequency, a sponsor-reported inferential comparison, or the model details needed to interpret adjusted treatment effects. The posted data therefore do not support a definitive claim about the clinical importance of the observed differences.
The findings do not establish that the treatment groups are equivalent. Equivalence requires a defined margin and an analysis designed to assess whether the uncertainty lies within that margin. No equivalence margin or such analysis is reported in the registry record. Likewise, failure of a recalculated unadjusted comparison to exclude zero is not proof that the groups have identical outcomes. It indicates that these summaries do not precisely distinguish the estimated differences from zero under the assumptions used.
The safety counts are separate descriptive information. They identify the number of participants affected by reported serious and other adverse events, but they do not by themselves determine whether an event was caused by treatment or how the overall balance of benefits and risks should be judged. The limited registry reporting on analysis populations, missing diary data, and safety-event detail further constrains interpretation. Conclusions should remain specific to the posted measures and summaries.
16. Why This Trial Matters Statistically
Statistically, this trial is useful because it presents a randomized, double-blind, placebo-controlled comparison with two active dose groups and a diary-based continuous endpoint derived from migraine-attack counts. The endpoint's diary rules make explicit how closely spaced events were grouped into attacks and what symptom-free period was needed for attacks to be treated as separate in the analysis. Those operational definitions are important for reproducibility in studies based on participant-recorded events.
The outcome also illustrates the distinction between a point estimate and its precision. The displayed mean changes are close across groups, and the recalculated differences are −0.2 attacks per participant for both active-dose comparisons. Their confidence intervals cross zero. Readers therefore need the interval and the underlying calculation assumptions, rather than relying only on the direction of a mean or on a p-value. The arm-level least-squares means are posted, but the registry does not include the model specification behind them or sponsor-reported analysis entries.
The study also highlights how registry completeness affects statistical interpretation. The record identifies the design and key endpoint, yet it does not report the original sample-size justification, a detailed analysis plan, the treatment of missing diary data, or the analysis population. Planning illustrations based on the observed effect show that required sample sizes depend strongly on the effect targeted, but they cannot reveal the original design rationale or substitute for it. The safety summaries likewise provide participant counts without a full comparative analysis.
These features make the trial a useful case for separating posted descriptive results from recalculated comparisons and from unreported design or analysis choices. That separation helps readers understand what can be learned from a registry record and which questions remain unanswered.
17. Limitations
- The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
- Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.
The primary outcome is based on diary entries, but the registry record does not report diary completion rates, missing-data rules, or how uncertain event classifications were handled. Although the record gives explicit rules for merging closely spaced migraine events, it does not describe adherence checks or how the diary data were processed. Missing or incomplete recording could affect the observed attack-frequency measure, but the available information does not quantify that issue.
The registry lists enrollment of 452, whereas the outcome results give arm counts of 133, 134, and 136. The record does not explain the relationship between the overall enrollment and these outcome counts. Adverse-event denominators are 136, 135, and 139, and the record does not explain why these differ from the outcome counts. These differences should not be interpreted as evidence of a particular exclusion or loss-to-follow-up process without further information.
The registry does not provide the trial's original sample-size assumptions, the statistical analysis plan, the model used for the least-squares means, or the population included in the primary analysis. It also does not report a sponsor inferential analysis, secondary outcome results, or detailed multiplicity and missing-data procedures. The unadjusted recalculations rely on an approximate pooled standard deviation inferred from summary statistics and should not be confused with a model-based sponsor analysis.
The listed adverse-event counts lack a full account of event definitions, severity detail, treatment relatedness, and comparative statistical analysis. Death counts are not reported in the registry record. The summary therefore supports description of the listed counts, but not a comprehensive assessment of safety or causal attribution.
18. Related Tutorials
19. Related Statistical Calculators
20. Sources
- ClinicalTrials.gov: Efficacy and Safety of Eslicarbazepine Acetate as Preventive Therapy for Subjects With Migraine — NCT01820559.
- Registry record last updated 2013-05-24; status last verified 2013-04.
- Results first posted 2013-05-24.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
21. Record Summary
The registry record describes NCT01820559 as a completed phase 2 randomized, double-blind, placebo-controlled, parallel-group trial sponsored by Bial - Portela C S.A. It lists enrollment of 452 and a study start in 2009-04, with primary completion in 2010-06. Participants with migraine, with or without aura, were assigned to placebo, ESL 800 mg/day once daily, or ESL 1200 mg/day once daily in a 1:1:1 ratio.
The primary outcome was absolute change from baseline in migraine-attack frequency standardized to 4 weeks during the Maintenance Period. Posted least-squares mean changes were −0.8 for placebo and −1.0 for each ESL dose. Recalculated unadjusted mean differences were −0.2 attacks per participant for both active-dose comparisons, and the confidence intervals cross zero. The registry lists serious and other adverse-event counts but does not report a detailed comparative safety analysis. It also does not report the original sample-size justification, the detailed statistical analysis plan, missing-data procedures, or the analysis population.