Eslicarbazepine Acetate (BIA 2 093) as Therapy for Refractory Partial Seizures in Children: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Eslicarbazepine Acetate (BIA 2 093) as Therapy for Refractory Partial Seizures in Children (NCT00988156), a randomized phase 3 study. The primary outcome is Change From Baseline in Seizure Frequency.

2. Clinical Question

3. Trial Design

The registry identifies this as a phase 3, randomized, parallel-group study with triple masking. Its listed status is completed. The overall enrollment is recorded as 304. The posted primary-outcome summaries list 129 participants in the placebo group and 134 in the eslicarbazepine acetate group; the safety summaries use the same group counts. These entries describe different record fields, and the registry does not explain the relationship between the overall enrollment and the outcome-specific group counts.

The intervention was eslicarbazepine acetate (BIA 2-093) added to concomitant therapy, compared with placebo added to concomitant therapy. The arm descriptions do not specify the background drugs, treatment doses, dose adjustment procedures, or duration of treatment beyond the primary outcome’s stated baseline-to-Visit 7 timeframe and its 12-week maintenance period. Those details are not reported in the registry record.

The study start is listed as 2007-12-07, and primary completion is listed as 2012-08-20. Bial - Portela C S.A. is the named sponsor. The record lists two outcome measures and two analyses, but gives no posted secondary outcomes. The primary endpoint concerns relative reduction in standardised 4-week seizure frequency from baseline to the maintenance period. Details about allocation procedures, the roles included in triple masking, the original sample-size justification, and any interim analyses are not reported in the registry record. The parallel randomized structure provides the basis for comparing the two treatment groups, but the listed design features alone do not establish how every aspect of conduct or analysis was implemented.

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

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 Seizure Frequency, measured over Baseline up to Visit 7.

The primary endpoint is titled “Change From Baseline in Seizure Frequency.” Its description specifies the relative reduction in standardised 4-week seizure frequency between the baseline period and the 12-week maintenance period; the timeframe is baseline up to Visit 7. The outcome is presented as a continuous measure, with a mean and standard deviation for each group. In general, a continuous endpoint retains information about the amount of change across participants rather than reducing results to a simple event/no-event classification.

Seizure frequency is relevant to the study question because the record concerns treatment for refractory partial seizures and describes the aim as examining whether adjunctive eslicarbazepine acetate changes seizure frequency. Standardising frequency to a 4-week period allows the endpoint to be expressed on a consistent time basis, as described in the registry. However, the record does not give the precise calculation used to form each participant’s relative reduction, how baseline and maintenance observations were combined, or how incomplete observation periods were handled. Those details are not reported in the registry record.

The posted group means are 62.0 for placebo and 36.6 for eslicarbazepine acetate. The posted mean-difference estimate is -25.4 for eslicarbazepine acetate relative to placebo. The measurement unit is seizures/month. The registry labels the measure as a mean, but the endpoint description refers to relative reduction; interpreting the clinical direction therefore requires care with the exact endpoint coding and calculation, which are not fully described in the record.

7. Results Posted in the Registry

• Esl (BIA 2-093): mean 36.60 versus 62.00 on control; difference -25.40 (95% CI -59.79 to 8.99), standardized effect d = -0.18; z-test p = 0.148.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a continuous measure summarized with a mean and standard deviation. A mean describes the average value in a group, while a standard deviation describes how dispersed individual observations are around that average. These summaries can be used to estimate a difference between group means. They do not show the full distribution of observations, and they may not capture unusual features such as strong skew or extreme values.

What does the mean difference represent?

The unadjusted mean difference compares the posted group means by expressing the eslicarbazepine acetate mean relative to the placebo mean. The registry’s re-calculated estimate is -25.4, with a 95% confidence interval from -59.79 to 8.993. The direction and meaning of this estimate depend on how the endpoint is coded. Although the title refers to change from baseline and the description refers to relative reduction, the record does not fully specify the participant-level calculation or sign convention. The estimate should therefore be reported as a difference in the posted endpoint, without assuming more detailed clinical interpretation than the record supports.

What does a confidence interval convey?

In general, a confidence interval describes the uncertainty around an estimate under the statistical model used. A narrower interval indicates greater precision, while a wider interval indicates less precision. For this re-calculated difference, the interval includes zero, so it includes no difference as well as values on either side of zero. The interval does not assign a probability to any particular treatment effect, and it does not confirm that the true effect lies at the point estimate.

How should the p-values be read?

The unadjusted p-value is 0.148. In general, a p-value measures how incompatible the observed data are with a specified null model, assuming that model and the analysis assumptions are appropriate. It is not the probability that the null hypothesis is true, nor a measure of clinical importance. The registry’s sponsor-reported ANCOVA p-value is 0.249. Because this analysis is adjusted and the re-calculation is not, the two p-values are based on different methods. The ANCOVA estimate, covariates, and other implementation details are not reported in the registry record, so the adjusted result cannot be reconstructed from the posted information.

Why can variability matter?

The posted standard deviations are 186.2 for placebo and 72.47 for eslicarbazepine acetate. In general, greater variation within groups makes the estimated mean difference less precise, particularly when group sizes are limited. Variation may arise from differences between participants, measurement variability, or features of the outcome distribution; the registry does not identify its sources. When variability is substantial or differs between groups, statistical model assumptions and the choice of analysis may matter. The record does not report diagnostic checks or alternative analyses.

What is ANCOVA?

In general, analysis of covariance (ANCOVA) compares groups while adjusting for one or more covariates. If an appropriate baseline measure is included, adjustment can account for baseline differences and may improve precision. The registry identifies ANCOVA as the method for the sponsor-reported analysis, but does not identify the covariates, specify whether baseline seizure frequency was included, or provide the adjusted estimate and its confidence interval. It is therefore possible to describe the method named in the record, but not to assess the detailed adjustment or reproduce its result.

What does a planning calculation mean?

In general, sample-size planning uses assumptions about the effect to detect, variability, significance level, allocation, and statistical model to estimate the participants needed for a chosen power. The posted planning illustration uses the observed effect and specified fractions of it, with equal allocation, a two-sided significance level of 0.05, and a normal approximation. It is a planning exercise based on the observed summary result, not evidence about the design assumptions originally used for the trial. The original sample-size rationale is not reported in the registry record.

These methods address different questions. Descriptive summaries report what was observed in the posted groups; a mean-difference estimate quantifies their contrast; a confidence interval describes uncertainty under a model; and a p-value measures compatibility with a null model. None of these alone establishes clinical importance, causation, or the completeness of the evidence. The registry’s limited account of analysis populations, missing data, and endpoint construction constrains how confidently the calculations can be assessed.

10. Confidence Intervals and Precision

The unadjusted mean difference is -25.4, with a 95% confidence interval from -59.79 to 8.993. This interval is wide and crosses zero. In general, a confidence interval gives a range of values compatible with the observed estimate under the chosen statistical model and its assumptions. Its inclusion of zero means the posted unadjusted summary statistics do not rule out no difference, while its endpoints also allow differences in either direction. The interval should not be read as a range guaranteed to contain the true value.

The width reflects limited precision, but the registry record does not permit attributing that uncertainty to a single cause. The endpoint has considerable dispersion in the placebo group, with a standard deviation of 186.2, and the eslicarbazepine acetate standard deviation is 72.47. The outcome-specific group counts are 129 and 134. In general, both within-group variability and the amount of information available affect precision. Other considerations, including the endpoint’s distribution and missing observations, cannot be assessed from the posted summary.

The confidence interval belongs to the unadjusted re-calculation, not the sponsor-reported ANCOVA. The registry gives the ANCOVA p-value as 0.249 but does not provide its estimate or confidence interval. Consequently, the adjusted result’s precision cannot be judged from the record. The re-calculated interval is useful for understanding uncertainty in the simple mean comparison, but it does not replace the sponsor’s analysis or establish the likely clinical size of an effect.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm, not total event counts. In general, the 80% power column gives the planned number of participants per arm for a calculation designed to have an 80% chance of detecting the assumed effect if that effect is present and the planning assumptions hold. The 90% power column uses a higher target chance of detection and therefore calls for more participants per arm under the same assumptions.

The illustration uses the observed effect as a reference, and also considers planning for 75% and 50% of that observed effect. The listed 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. For 100% of the observed effect, the figures are 479 participants per arm for 80% power and 641 participants per arm for 90% power. For 75% of the observed effect, they are 852 and 1,140 participants per arm. For 50% of the observed effect, they are 1,916 and 2,564 participants per arm.

Smaller effects are harder to distinguish from random variation, so detecting them at the same target power generally requires more participants. These numbers are planning illustrations based on the observed effect, not observed power. The original sample-size assumptions and justification for the trial are not reported in the registry record, and the illustration should not be interpreted as a reconstruction of the trial’s original design.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -25.400; pooled standard deviation σ = 140.277.
2. n = 2 × (1.96 + 0.842)² × 140.277² / -25.400² = 478.8, rounded up to 479 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 <- -25.4                      # observed difference in means
sigma <- 140.276717229025                     # 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(988156)
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 = 479 (80% power), 641 (90% power)
 75% of observed effect: n per arm = 852 (80% power), 1,140 (90% power)
 50% of observed effect: n per arm = 1,916 (80% power), 2,564 (90% power)
power.t.test: 479.7 
Simulated power at n = 479 per arm: 0.804 (20000 replicates)

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

13. Analyses Reported by the Sponsor

The registry reports the sponsor’s own primary-outcome analysis as ANCOVA, with p = 0.249. No adjusted effect estimate, confidence interval, covariates, or further analysis details are posted. The summary statistics also allow a simple unadjusted re-calculation: the mean difference is -25.4, its 95% confidence interval is -59.79 to 8.993, and its p-value is 0.148.

These analyses are not identical. The re-calculation compares the posted group means without adjustment, whereas ANCOVA adjusts for covariates. Differences between their p-values can arise because the methods use the data differently; the record does not supply enough detail to determine the specific contribution of each modeling choice. The re-calculation is therefore a supplementary check on the posted summaries, not a replacement for the sponsor’s own analysis. The discrepancy indicator is false, so the record does not flag a conflict requiring the sponsor analysis to take precedence over a discrepant recalculation. Even so, the sponsor’s ANCOVA is the analysis reported by the registry, and its missing estimate and methodological details limit independent interpretation.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists two outcome measures and two analyses, but no secondary outcome results are posted. It does not identify a multiplicity adjustment, explain how the primary endpoint was prioritized relative to other outcomes, or state whether any outcome or analysis was selected after examining the data. These matters are not reported in the registry record. Without a documented testing plan, it is not possible to determine how the sponsor handled the possibility of multiple statistical comparisons.

The primary outcome’s population field is blank. The record does not identify whether the analysis used an intention-to-treat population, a modified intention-to-treat population, a per-protocol population, or another analysis set. Although outcome and safety groups have posted counts, those counts alone do not establish the population definitions or how participants were included in each analysis.

Missing-data methods are also not reported in the registry record. The record does not say how missing seizure-frequency observations were treated, whether observations were imputed, whether participants with incomplete follow-up were excluded, or whether sensitivity analyses were conducted. These choices can affect a continuous endpoint, especially if missingness is related to seizure frequency or treatment experience. That is a general statistical concern, not evidence that a particular problem occurred in this trial.

The sponsor analysis is identified as ANCOVA, but the registry does not give its covariates, model specification, assumption checks, or adjusted estimate. The timing and conduct of interim analyses are not reported. The original statistical analysis plan and the trial’s own sample-size justification are also not reported in the registry record. As a result, the posted material supports a description of the named analysis and its p-value, but not a full audit of prespecification, missing-data handling, multiplicity control, or analysis-population choices.

15. 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 3% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pneumonia (2.3%), status epilepticus (2.2%), convulsion (1.6%), device malfunction (1.5%), bronchopneumonia (1.5%).

The safety summaries list serious adverse events for 9 of 129 participants in the placebo group and 15 of 134 participants in the eslicarbazepine acetate group. Other adverse events are listed for 94 of 129 and 112 of 134, respectively. These are the counts reported for the safety set part I. They should be read as descriptive group summaries; the registry does not provide a statistical comparison for these counts or establish that treatment caused any particular event.

The listed serious adverse event terms and percentages include pneumonia, 2.3%; status epilepticus, 2.2%; convulsion, 1.6%; device malfunction, 1.5%; bronchopneumonia, 1.5%; and irritability, 0.8%. The record does not explain whether terms can overlap within participants or provide fuller event-level details, so these listings should not be added or treated as mutually exclusive categories. Death counts are not reported in the registry record. Information about event attribution, severity criteria beyond the serious-event classification, and the completeness of follow-up is not reported in the registry record.

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

The posted outcome summaries describe a lower mean endpoint value in the eslicarbazepine acetate group than in the placebo group: 36.6 compared with 62.0. The unadjusted mean difference is -25.4, with a 95% confidence interval from -59.79 to 8.993 and an unadjusted p-value of 0.148. The interval includes zero and spans values on both sides, so the unadjusted result is not precise enough to establish a clear difference. The sponsor-reported ANCOVA p-value is 0.249, but its adjusted estimate and confidence interval are not posted.

These statistics do not establish whether the treatment works or does not work. A non-significant p-value does not demonstrate equivalence or prove that there is no effect. Conversely, a numerical difference between group means does not by itself establish a treatment effect. The confidence interval gives useful information about uncertainty, but the endpoint’s coding and sign convention are not fully explained, which limits detailed interpretation of the numerical direction.

The trial is described as randomized and triple-masked, features that in general can help reduce bias when properly implemented. The registry does not describe the allocation procedure, identify the masking roles, or provide enough methods detail to assess implementation. It also does not report secondary outcome results, the primary analysis population, missing-data methods, or the ANCOVA covariates. These omissions restrict conclusions about robustness and the full balance of efficacy and safety evidence.

The serious and other adverse event summaries are descriptive counts. They do not establish comparative risk or causality, and deaths are not reported in the registry record. The results should therefore be understood as a limited account of the posted endpoint and safety data, rather than a complete evaluation of treatment benefit or harm.

17. Why This Trial Matters Statistically

This study raises important statistical issues common to clinical trials of variable continuous outcomes. The primary endpoint concerns seizure frequency summarized over a standardized period and described as relative reduction from baseline to maintenance. Standardization can make measurements more comparable, but interpretation depends on the precise participant-level calculation, the handling of baseline values, and how the endpoint is coded. Those details are not reported in the registry record, making clear endpoint definition essential when reading the estimate.

The posted standard deviations differ substantially between groups, and the placebo group’s value is particularly large. In general, over-dispersion or a strongly skewed distribution can make a simple mean-based model less representative of the data and can affect standard errors and confidence intervals. The registry does not report distributional checks or alternative modeling. This makes the sponsor’s named ANCOVA and the unadjusted re-calculation useful to distinguish, while the absent adjusted estimate prevents a fuller comparison.

The record also illustrates why a p-value alone is incomplete. The unadjusted estimate is accompanied by a confidence interval that includes zero and allows values in both directions. That interval communicates uncertainty and possible effect sizes more informatively than a binary significant/not-significant label. The sponsor’s p-value is reported without an effect estimate or confidence interval, limiting assessment of its magnitude and precision.

Finally, registry completeness matters. No secondary results, multiplicity plan, missing-data methods, or analysis-population definitions are posted. Those elements are important for determining how securely a primary result can be interpreted and how it fits with other measured outcomes. The case demonstrates the value of reporting the estimand, analysis assumptions, population, and uncertainty together, rather than relying on a single test result.

18. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The primary endpoint is described as relative reduction in standardised 4-week seizure frequency, while its title says change from baseline and its measure is reported as a mean in seizures/month. The registry does not fully explain the calculation, transformation, or sign convention. This makes it difficult to translate the posted difference into a more specific clinical interpretation.

The primary outcome population field is blank, and analysis-population definitions are not reported. The overall enrollment is 304, while the posted outcome and safety summaries list 129 in placebo and 134 in eslicarbazepine acetate. The record does not explain the relationship between those figures. Missing-data handling, follow-up completeness, and sensitivity analyses are not reported, limiting assessment of potential bias in the continuous endpoint.

The sponsor-reported ANCOVA result is incomplete for independent interpretation: only p = 0.249 is posted, without the adjusted estimate, confidence interval, covariates, or model details. The unadjusted re-calculation is based on summary statistics and cannot reproduce that adjusted analysis. The registry also does not report multiplicity procedures, secondary outcome results, interim analyses, or the trial’s original sample-size justification.

Safety information is limited to group counts and a list of serious event terms and percentages. Deaths are not reported in the registry record, and detailed attribution or event-level information is absent. These limitations concern what can be established from the registry account; they do not show that unreported procedures or analyses were not conducted.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Eslicarbazepine Acetate (BIA 2 093) as Therapy for Refractory Partial Seizures in Children — NCT00988156.
• Registry record last updated 2018-11-13; status last verified 2018-10.
• Results first posted 2014-07-17.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

22. Record Summary

The registry identifies NCT00988156 as a completed, randomized, triple-masked, parallel-group phase 3 study sponsored by Bial - Portela C S.A. It examined eslicarbazepine acetate added to concomitant therapy compared with placebo added to concomitant therapy for refractory partial seizures in children and adolescents. Overall enrollment is listed as 304; the posted primary-outcome groups are 129 for placebo and 134 for eslicarbazepine acetate.

The primary outcome concerns change from baseline in seizure frequency, described as relative reduction in standardised 4-week seizure frequency from baseline to the 12-week maintenance period. The unadjusted mean difference is -25.4, with a 95% confidence interval from -59.79 to 8.993 and p = 0.148. The sponsor-reported ANCOVA p-value is 0.249. The record also posts serious and other adverse event counts, but deaths are not reported in the registry record. Important analysis details, including the primary analysis population, missing-data methods, and ANCOVA covariates, are not reported.