A Study of Soticlestat as an Add-on Therapy in Children and Young Adults With Dravet Syndrome: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Soticlestat as an Add-on Therapy in Children and Young Adults With Dravet Syndrome (NCT04940624), a randomized phase 3 study. The primary outcome is Percent Change From Baseline in Convulsive Seizure Frequency Per 28 Days During the Full Treatment Period.

2. Clinical Question

The central statistical question concerns the between-group comparison, rather than whether either group changed from its own baseline. A change within one group alone cannot establish that the intervention caused a difference, because participants in both groups continued their standard antiseizure therapy and may also have experienced changes over time for other reasons.

The endpoint expresses each participant’s seizure frequency during treatment relative to their baseline frequency. A negative percentage change corresponds to a lower frequency during treatment than at baseline. The registry reports median changes for each group and a model-based comparison between groups. Those summaries address different aspects of the data: the group medians describe typical observed changes, while the comparison estimates the separation between groups under the reported analysis.

The question is therefore whether the posted evidence supports a difference in the distribution of seizure-frequency changes, and how uncertain that estimate remains. The registry record does not establish from these figures alone how the result should guide an individual treatment decision.

3. Trial Design

This was a phase 3, randomized, quadruple-masked, parallel-group study. The registry lists 144 participants enrolled and describes the study as completed. Its stated purpose was to evaluate efficacy, safety, and tolerability when soticlestat was added to standard antiseizure therapy in children and young adults with Dravet syndrome. The study began on 2021-10-28, and the primary completion date was 2024-04-11.

Participants were assigned to soticlestat or matching placebo while continuing their standard antiseizure therapy. The parallel-group design means the treatment groups were followed as separate groups rather than participants crossing over between study treatments. Masking is recorded as quadruple; the registry record does not identify in the information reported which specific roles were included in that masking designation.

The described treatment period was 16 weeks. The registry describes a titration period of up to 4 weeks and a 12-week maintenance period. Participants who met the extension study’s entry criteria could continue treatment in an extension study. For participants who wanted to stop, the summary describes gradual dose reduction over 1 week followed by 2 weeks of follow-up. These details describe the study’s stated treatment and follow-up arrangements, not additional outcome results.

The primary endpoint covered baseline and the full treatment period, identified as Weeks 1 to 16. The primary analysis population was modified Intent-to-Treat, rather than every randomized participant regardless of treatment exposure or seizure assessment. The registry gives the population definition but does not report a full participant-flow account in the information summarized for this analysis.

Random allocation and masking are design features intended to limit systematic differences in treatment assignment and assessment. They do not remove uncertainty from the estimates, nor do they by themselves establish that all potential sources of bias were prevented. The study is described as a parallel-group comparison of active add-on treatment against placebo add-on treatment.

4. Participant Flow and Accounting

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

Before assignment: A total of 144 participants with a diagnosis of Dravet Syndrome (DS) were randomized in a 1:1 ratio to receive either soticlestat or placebo.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Modified Intent-to-Treat (mITT) Analysis Set included all randomized participants who had received at least one dose of study drug and had been assessed for seizures for at least one day in the Full Treatment Period.

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.

Population: Safety Analysis Set included all participants who had taken at least one dose of study drug.

6. Primary Endpoint

The registry lists the primary outcome as Percent Change From Baseline in Convulsive Seizure Frequency Per 28 Days During the Full Treatment Period, measured over Baseline; Full Treatment Period: Weeks 1 to 16.

The primary endpoint was percent change from baseline in convulsive seizure frequency per 28 days during the full treatment period. The registry defines frequency by taking the number of convulsive seizures reported during a period, dividing by the number of days in that period when seizures were assessed, and scaling the result to a 28-day basis. This standardization allows periods with different assessment durations to be expressed on the same time scale.

Percent change compares the treatment-period frequency with the baseline frequency for each participant. Under the registry’s definition, the treatment-period value minus the baseline value is divided by the baseline value and multiplied by 100. Negative values therefore indicate a reduction relative to baseline, while positive values indicate an increase. A percentage-change endpoint puts participants with different baseline seizure frequencies on a relative scale, though it can also be affected by the baseline value used in the calculation.

Convulsive seizure frequency was selected because the study’s stated aim was to assess whether add-on soticlestat reduced convulsive seizures. The endpoint captures change over the full treatment period rather than a single visit. The registry reports group medians and an interquartile-range measure as the endpoint’s summary framework. A median describes the middle of the observed distribution and is less influenced by unusually large values than an arithmetic mean. The posted primary result does not include the group interquartile ranges, so their observed values are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is summarized with medians, while the registry-reported comparison uses an analysis identified as ANCOVA and reports a Hodges-Lehmann location shift estimate. These statistics should not be conflated. A median describes the center of one group’s observed values. A location-shift estimate summarizes the contrast between groups under its analysis framework. The registry does not give further details about the precise model specification, covariates, or assumptions, so those are not reported in the registry record.

What does the confidence interval communicate?

A confidence interval describes uncertainty around an estimated contrast under the analysis used. The 95% interval for the primary comparison runs from -31.30 to 0.24. Because it includes zero, the posted analysis does not rule out no location shift at that confidence level. The interval also includes negative values, which are consistent with a reduction favoring soticlestat under the endpoint’s sign convention. The interval is not a range containing a guaranteed proportion of individual participant outcomes, and it does not assign a probability to a particular true value.

The corresponding p-value is =0.061. A p-value measures how unusual data at least as discrepant as those observed would be under a specified null hypothesis and analysis assumptions. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. The registry reports the p-value as a value, rather than as evidence that a threshold has been met. Reading it alongside the estimate and interval gives a fuller picture than relying on a threshold alone.

Why report medians for the primary endpoint?

Seizure-frequency changes can vary across participants, and a small number of extreme observations can affect a mean more than a median. The median is the middle observed value when results are ordered. The interquartile range describes the middle half of a distribution, giving information about spread without requiring a symmetric distribution. Although the primary endpoint lists an interquartile range as its dispersion measure, group spread values are not reported in the registry record. The reported medians therefore describe typical observed changes, but they do not show the full distribution or how much individual results overlapped.

How should the secondary comparisons be understood?

For binary responder outcomes, the registry reports group percentages and its own odds-ratio analyses using the Cochran-Mantel-Haenszel method. An odds ratio compares the odds of the outcome between groups; it is not the same quantity as a difference in percentages. The secondary recalculations also express contrasts as risk differences, which compare the proportions directly. Because these are distinct measures and may use different analytic details, the recalculated risk difference and registry-reported odds ratio need not have identical numerical values or confidence intervals.

The registry-reported responder analyses have 95% confidence intervals for the odds ratios and p-values. The recalculated comparisons are unadjusted calculations from the posted group summaries. They are useful as a transparent descriptive check, but do not replace the registry-reported analysis. The quality-of-life secondary analysis is reported using MMRM and a least-square mean difference. MMRM is a model framework commonly used for repeated measurements; the registry record does not report the model’s detailed specification, covariance structure, or missing-data assumptions. The reported estimate and interval should therefore be interpreted as the registry’s analysis, without inferring unreported modeling choices.

What do the planning illustrations mean?

The sample-size illustrations are planning calculations based on observed effects, not the trial’s own design justification. Their stated assumptions are a two-sided significance level of 0.05, equal allocation between groups, and a normal approximation. The calculations consider the stated fractions of observed effects. Such illustrations answer a hypothetical planning question: how many participants per arm would be indicated under those assumptions for a specified target power and effect size. They are not observed power, do not show the probability that the completed trial detected an effect, and should not be read as the actual enrollment target or as evidence that the trial was planned using the observed result. The registry record does not report the trial’s sample-size justification or the assumptions used for the original design.

More generally, over-dispersion means that variation in observed data is greater than a model’s assumed variation. When present, an analysis that ignores it can understate uncertainty. That is a general statistical consideration; whether over-dispersion was assessed or affected this study is not reported in the registry record. Similarly, the registry record does not report interim analyses or enough details to reconstruct the full statistical analysis plan. Interpretation should remain tied to the estimates and methods that are actually described.

10. Analyses Reported by the Sponsor

The registry reports a primary analysis comparing the groups with ANCOVA and gives a Hodges-Lehmann location shift estimate of -15.64, a 95% confidence interval from -31.30 to 0.24, and p = 0.061. That is the reported primary inferential result and takes precedence over any simple unadjusted recalculation from group summaries. The record does not flag a discrepancy between the posted figures and a recalculation.

For the responder outcomes, the registry-reported analyses use the Cochran-Mantel-Haenszel method and report odds ratios. The accompanying simple recalculations use unadjusted comparisons of the posted group summaries, including risk differences. These recalculations are not the same analysis as the registry-reported model: they do not reproduce its method or necessarily its treatment of stratification or other model details. Their role is to provide a basic descriptive comparison, not to supersede the registry analysis.

For quality of life, the registry reports an MMRM comparison, while a simple difference calculated from the group means would not match that longitudinal model. The registry-reported least-square mean difference, confidence interval, and p-value should therefore be used when describing the formal posted analysis. In general, the model-based results and summary-statistic recalculations answer related but not identical questions. Where the registry gives its own analysis, that analysis is the appropriate primary account of the reported inferential result.

11. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates more than one outcome, time point, subgroup, or statistical comparison. Each additional test creates another opportunity for a result to appear notable by chance. This study lists multiple secondary outcomes, including responder measures across different periods, caregiver and investigator impressions, symptom domains, and a quality-of-life measure. The registry also lists 16 analyses. These features make it important to know whether the analyses were arranged in a prespecified hierarchy or adjusted to control the overall false-positive risk.

Multiplicity procedures, if used, can change how individual p-values should be interpreted. The registry record does not report a multiplicity adjustment or a testing hierarchy for the analyses discussed. It is therefore not possible to determine from the record whether the nominal p-values for the secondary outcomes control the chance of false-positive findings across the family of tests. This limitation does not erase the observed estimates, but it means secondary p-values should not be read as if each were the only hypothesis tested.

Missing-data handling is another important issue. The primary modified Intent-to-Treat population includes randomized participants who received at least one dose and had at least one day of seizure assessment during the full treatment period. This definition clarifies who contributed to the primary analysis, but does not explain how missing seizure assessments were handled, whether values were imputed, or how incomplete follow-up affected the endpoint. Those methods are not reported in the registry record.

Analysis denominators differ across posted outcomes. The primary endpoint uses 71 placebo and 73 soticlestat participants; the maintenance responder outcome uses 68 and 69; the full-period responder outcome uses 71 and 73; and the quality-of-life comparison uses 53 and 56. Different denominators may reflect outcome-specific availability or analysis populations, but the reason for each difference is not reported in the registry record. A smaller analysis population can also mean less precise estimates.

The registry record does not report interim analyses, a complete statistical analysis plan, detailed missing-data assumptions, or the original sample-size justification. Without those details, the available results support interpretation of the reported comparisons but do not allow full appraisal of all prespecified analytic decisions.

12. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: status epilepticus (5.6%), seizure cluster (2.7%), adenovirus infection (1.4%), covid-19 pneumonia (1.4%), infectious mononucleosis (1.4%).

The posted safety figures report serious adverse events for 10 of 71 participants in the placebo group and 7 of 73 participants in the soticlestat group. Other adverse events affected 37 of 71 participants in the placebo group and 40 of 73 participants in the soticlestat group. The registry reports 0 deaths in the placebo group and 1 in the soticlestat group. These are counts of participants, and they describe the posted groups rather than establishing that study treatment caused an event.

The registry lists status epilepticus at 5.6% and seizure cluster at 2.7% among the leading serious adverse-event terms. Adenovirus infection, COVID-19 pneumonia, infectious mononucleosis, and pneumonia are each listed at 1.4%. The information reported does not show whether participants could contribute to more than one term, nor does it provide event-level timing or attribution in this summary.

Serious adverse events and all other adverse events are different reporting categories and should not be treated as interchangeable measures of safety. The counts also do not establish comparative risk on their own: interpretation depends on event definitions, follow-up, severity, and attribution, among other details. The registry record does not provide enough information in this summary to determine causality or to support an individual safety judgment.

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

The primary point estimate is consistent with a larger reduction in convulsive seizure frequency in the soticlestat group than in the placebo group, but its 95% confidence interval includes zero and its p-value is =0.061. The posted primary analysis therefore leaves uncertainty about whether the groups differed. The estimate’s direction and its uncertainty both matter: neither should be omitted when describing the result.

The result does not mean that every participant’s seizures decreased, or that the group median describes each participant’s experience. Group summaries can conceal substantial variation. Nor does a negative median in both groups establish that the study treatment caused a reduction; changes occurred in both groups, which continued standard antiseizure therapy.

The secondary responder outcomes show a more favorable proportion in the soticlestat group, but those analyses are secondary and multiple outcomes were assessed. Because adjustment for multiplicity is not reported in the registry record, those findings should be treated as supportive rather than as definitive confirmation of the primary result. The quality-of-life comparison also has uncertainty and is a different outcome from seizure frequency.

These results do not establish that soticlestat works or does not work, and they do not determine whether it is appropriate for a particular person. The estimates describe a group comparison under the reported study design and analysis. Clinical importance cannot be inferred from statistical significance alone, and the registry information summarized does not give a threshold for a clinically meaningful change in the primary endpoint.

The study’s 16-week treatment period also limits what can be concluded about longer-term outcomes from this primary comparison. The registry notes that continuation could occur in an extension study for eligible participants, but extension results are not part of the primary result. Conclusions should stay within the measured population, endpoint, time period, and analyses that are reported.

14. Why This Trial Matters Statistically

This trial illustrates why a group’s change from baseline and the randomized between-group comparison answer different questions. Both primary group medians were negative, but the main inferential question is whether the groups differed. The reported contrast, its confidence interval, and its p-value are needed to assess that question; the two medians alone are not enough.

It also demonstrates the value of presenting uncertainty alongside a point estimate. The primary estimate favors a greater reduction in the soticlestat group under the endpoint’s sign convention, yet the interval includes zero. The evidence is therefore neither well described by the point estimate alone nor by a simple binary label. Confidence intervals help readers see the range of contrasts compatible with the analysis, while p-values provide a related test of a null hypothesis.

The different outcome types show why statistical results should not be collapsed into one overall verdict. A continuous seizure-frequency measure, threshold-based responder outcomes, and quality-of-life scores answer distinct questions. Their directions may align without having equal evidential weight. Multiple secondary tests also make interpretation depend on the analysis plan and any multiplicity control, details not reported in the registry record.

Finally, the registry offers a useful contrast between model-based analyses and simple recalculations from posted summaries. Recalculation can clarify how basic group proportions or means relate, but it does not recreate an analysis that accounts for a specified model. Treating those estimates as interchangeable would obscure the methods and uncertainty that define each result. These are general lessons for reading randomized clinical trial results, not conclusions about an individual participant.

15. 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.

Several trial-specific limitations constrain interpretation of the posted results. The primary outcome is a seizure-frequency measure based on reported seizures and days assessed. The registry defines the rate calculation, but the information reported does not describe how seizure reports were verified or how missing assessment days were handled. Because the endpoint depends on both event reporting and assessment time, those details would be relevant to judging measurement uncertainty.

The primary comparison is reported for a modified Intent-to-Treat population, which excludes randomized participants who did not meet the stated exposure and seizure-assessment requirements. The registry provides the definition but not a complete accounting of participants who were randomized, treated, assessed, or excluded from each analysis. The reasons for the outcome-specific denominators are also not reported in the registry record.

The primary period covers 16 weeks. That duration supports conclusions about the measured treatment period, but does not by itself establish persistence of effects or longer-term safety. The registry mentions possible extension-study participation, but corresponding extension results are not included in the results summarized.

The primary endpoint is reported using group medians, but the group interquartile-range values are not reported in the registry record. This limits assessment of the distribution and spread of individual changes. In addition, the record identifies ANCOVA and a Hodges-Lehmann location shift estimate but does not report model covariates or other detailed specifications.

For secondary outcomes, the registry reports several comparisons and does not report a multiplicity adjustment in the information summarized. Missing-data assumptions, interim analyses, and the original sample-size justification are also not reported in the registry record. These omissions do not invalidate the posted estimates, but they limit independent assessment of the full analysis strategy and how robust the conclusions are to alternative assumptions.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study of Soticlestat as an Add-on Therapy in Children and Young Adults With Dravet Syndrome — NCT04940624.
• Registry record last updated 2026-09-04; status last verified 2026-08.
• Results first posted 2025-01-01.
• 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.

19. Record Summary

The registry describes a completed, phase 3, randomized, quadruple-masked, parallel-group study of soticlestat as add-on therapy for children and young adults with Dravet syndrome. Enrollment was 144. Participants received standard antiseizure therapy with either soticlestat or matching placebo, with the stated full treatment period lasting 16 weeks.

The primary endpoint was percent change from baseline in convulsive seizure frequency per 28 days during the full treatment period. The registry-reported comparison gave a Hodges-Lehmann location shift estimate of -15.64, with a 95% confidence interval from -31.30 to 0.24 and p = 0.061. The interval includes zero, leaving uncertainty about the between-group difference.

Responder outcomes were directionally more favorable in the soticlestat group, while the quality-of-life comparison remained uncertain. The registry also reports serious adverse events, other adverse events, and deaths by group. Multiplicity procedures and several details needed to assess missing-data handling and the full analysis plan are not reported in the registry record.