A Study of Oral Dosing of Gabapentin Enacarbil in Japanese Restless Legs Syndrome Patients: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Oral Dosing of Gabapentin Enacarbil in Japanese Restless Legs Syndrome Patients (NCT03053427), a randomized phase 4 study. The primary outcome is Change From Baseline in International Restless Legs Syndrome Rating Scale (IRLS) Score at Week 12.

2. Clinical Question

The principal statistical question is about the estimated average change in symptom score under each assigned treatment, and the uncertainty around the between-group contrast. A larger reduction corresponds to a more favorable change in this scale because higher IRLS scores represent greater disease activity. The result therefore needs to be read as a comparison of changes, rather than as a comparison of end-of-study scores alone.

The registry record describes the study as a post-marketing clinical study in Japanese participants and lists restless legs syndrome as the condition. It does not report a broader population or provide details that would establish how well these findings apply outside the study population. The clinical question is consequently specific to the population, treatments, outcome definition, and follow-up described in the record.

3. Trial Design

The registry record identifies this as a phase 4, randomized, quadruple-masked, parallel-group study comparing gabapentin enacarbil with placebo. The study title specifies Japanese participants with restless legs syndrome; the brief summary describes moderate-to-severe idiopathic restless legs syndrome. The recorded enrollment is 375. The registry lists the study as completed, with start and primary-completion dates of 2017-03-30 and 2018-06-25.

Both groups received their assigned study treatment orally once daily after the evening meal. The available description for gabapentin enacarbil includes a dosing adjustment related to estimated creatinine clearance at the start of the run-in period, but the full dosing description is not reported in the registry record. The masking label is quadruple; the identities of the masked groups or roles are not reported in the registry record.

The primary outcome was change from baseline in IRLS score at week 12. The primary analysis used a Mixed Model of Repeated Measurements (MMRM), with treatment group, baseline IRLS score, age category, estimated creatinine-clearance category, time point, and the treatment-group-by-time-point interaction among its explanatory variables. Compound symmetry was specified as the covariance structure. The primary outcome population was the Full Analysis Set, defined in the record by receipt of study drug and evaluation for at least one efficacy endpoint during the treatment period.

The registry record lists secondary outcomes covering repeated IRLS assessments, investigator- and patient-rated global impressions, sleep, pain, health status, and adverse events. The record does not report the randomization sequence method, allocation concealment, a sample-size justification, or further details of the planned analysis strategy. Those omissions limit assessment of design choices beyond the features explicitly described.

4. Participant Flow and Accounting

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

Before assignment: Participants received single-blind placebo for 1 week (run-in period). Subsequently, eligible participants were randomized to receive gabapentin enacarbil or placebo orally once daily after dinner for 12 weeks (treatment period).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full Analysis Set (FAS) consisted of all participants who received the study drug for the treatment period and were evaluated for at least one efficacy (either primary or secondary) endpoint during the 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 (SAF) consisted of all participants given at least one dose of the study drug for the treatment period.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in International Restless Legs Syndrome Rating Scale (IRLS) Score at Week 12, measured over Baseline and week 12.

The primary endpoint is change from baseline in the International Restless Legs Syndrome Rating Scale score at week 12. The IRLS is a 10-item measure of restless legs syndrome severity. Each item ranges from 0 (no symptoms) to 4 (very severe symptoms), and the total score ranges from 0 to 40. Higher total scores indicate greater disease activity.

Change from baseline is used to express how the measured score shifts during the study relative to each participant’s starting score. This can make the treatment comparison more informative than an end-of-study score alone when baseline symptom levels vary. In this trial, the primary analysis did not simply subtract group averages: it used an MMRM model that included baseline IRLS score and other listed explanatory variables, and it estimated least-squares means and their difference.

Because a higher IRLS score indicates greater disease activity, a more negative change represents a greater reduction in measured symptoms. The reported group estimates were -10.5 for placebo and -11.7 for gabapentin enacarbil. These are model-based least-squares mean changes, with 95% confidence intervals of -11.4 to -9.5 and -12.6 to -10.7, respectively. The treatment comparison is the key inferential quantity: it describes the estimated separation between groups while conveying uncertainty around that estimate.

The scale captures symptom severity as defined by the IRLS; it does not by itself represent every dimension of restless legs syndrome or establish a broader effect on health. The record lists other outcomes, including sleep, pain, global impressions, and health status, which address different aspects of participants’ experience.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is continuous: it is a numeric change in IRLS score, analyzed as a least-squares mean. The posted group estimates are -10.5 for placebo and -11.7 for gabapentin enacarbil. The comparison is reported as an LSM difference of -1.2, with a 95% confidence interval from -2.6 to 0.2 and p = 0.088. A confidence interval describes the uncertainty around an estimate under the assumptions of the analysis. It helps distinguish a relatively precise estimate from one that remains compatible with a wider range of effects.

Why use a repeated-measures model?

Participants’ scores can be recorded at multiple time points. An MMRM uses the repeated observations together in a model, rather than treating every follow-up measurement as unrelated. The registry specifies compound symmetry as the covariance structure. In general, a covariance structure describes how measurements from the same participant are related across assessment times. Compound symmetry represents a particular pattern of within-participant correlation; whether that pattern fits these data cannot be judged from the information reported in the registry record.

The model’s listed explanatory variables were treatment group, baseline IRLS score, age category, estimated creatinine-clearance category, time point, and the interaction of treatment group and time point. The interaction permits the modeled treatment contrast to vary over time. The primary endpoint focuses on week 12, while the record also lists change from baseline at each time point as a secondary outcome. Adjusting for baseline score and the other listed variables is part of the stated model, but the record does not provide enough detail to reconstruct the full model or its diagnostics.

What do least-squares means represent?

Least-squares means are model-based estimates of group averages, accounting for the variables included in the model. The difference between them is the model-based treatment contrast. It is therefore not necessarily identical to a direct subtraction of unadjusted observed means. The registry reports the primary group estimates and sponsor contrast as least-squares quantities. The interval for the contrast is particularly important: its lower bound is -2.6 and its upper bound is 0.2, so it includes 0.

In general, a confidence interval gives a range of values that are reasonably compatible with the data and model at the stated confidence level. It does not assign a probability that the true effect lies within this specific calculated interval. Nor does overlap or non-overlap among separate group intervals provide the same test as an interval calculated directly for the treatment contrast. For this reason, interpretation should center on the reported contrast interval and the reported analysis rather than on an informal comparison of the group intervals.

How should the p-value be read?

In general, a p-value measures how incompatible the observed data are with a specified null model, assuming that model and the analysis assumptions hold. It is not the probability that the null hypothesis is true, and it does not quantify the size or clinical importance of an effect. The reported primary p-value is 0.088. The registry record labels the analysis as a superiority comparison but does not report the full decision rule or a multiplicity strategy. Thus, the p-value should be considered alongside the point estimate, confidence interval, endpoint definition, and analysis plan details that are available.

What do the planning illustrations mean?

The planning illustrations elsewhere in the posted results are calculations based on observed secondary-outcome effects, not the trial’s own design justification. Their stated assumptions are a two-sided significance level of 0.05, equal allocation, a normal approximation, and detection of a stated fraction of the observed effect. Such estimates are prospective planning illustrations: they describe sample sizes under their assumptions, not the power observed in this completed trial. The registry record does not report that these calculations informed the study design. Planning estimates based on an observed effect can also be sensitive to uncertainty in that effect, particularly when the estimated contrast is imprecise.

Different measures require different statistical summaries. For binary response outcomes, a risk difference describes a difference in response proportions. For continuous outcomes such as health status, a mean difference is expressed in the outcome’s scale units. Recalculated comparisons from posted summary statistics are simple, unadjusted calculations and may not reproduce a model-based sponsor analysis. They can be useful for orientation, but they do not replace the trial’s specified analysis or establish that model assumptions are satisfied.

Over-dispersion is a general concern when count or binary data vary more than the assumed model allows; it can make standard errors too small if ignored. It is not a stated feature of the primary continuous IRLS analysis, and the registry record does not report an over-dispersion assessment. No conclusion about such a diagnostic should be inferred from the reported primary estimates.

10. Analyses Reported by the Sponsor

The sponsor-reported primary analysis used a Mixed Model of Repeated Measurements and reported an LSM difference of -1.2 for gabapentin enacarbil versus placebo. Its 95% confidence interval was -2.6 to 0.2, and its p-value was 0.088. This is the main analysis to use when describing the trial’s primary comparison because it is the analysis reported for the study, with the model specified in the registry record.

The posted group estimates are also model-based least-squares mean changes: -10.5 for placebo and -11.7 for gabapentin enacarbil. They are consistent in direction with the negative sponsor-reported contrast, but the contrast itself should not be recreated by subtracting rounded group estimates. The model can incorporate covariates and repeated measurements, and rounding can also affect apparent arithmetic. No recalculated primary contrast is listed in the posted results.

For secondary outcomes, some posted unadjusted re-calculations are based on summary statistics, whereas the registry reports its own analyses for the response outcomes using Fisher Exact tests. Differences between methods or analysis populations can lead to differences in estimates, confidence intervals, and p-values. These recalculations are descriptive checks, not replacements for the sponsor’s analysis. The registry record does not report a discrepancy for the primary comparison.

11. Multiplicity, Missing Data and Analysis Populations

The registry record identifies one primary endpoint at week 12 and lists multiple secondary outcomes, with 14 analyses and nine outcome measures in the record. The posted results include eight secondary outcomes. When many outcomes, time points, or comparisons are analyzed, the chance of at least one apparently notable result can increase even when there is no underlying treatment difference. Statistical procedures such as an adjustment for multiplicity or a prespecified testing sequence can help control this issue. Whether such a procedure was planned or applied is not reported in the registry record.

The record describes the primary MMRM and its explanatory variables, but it does not report the complete statistical analysis plan, any interim analyses, or the rule used to decide whether the primary analysis met its decision criterion. The reported p-value and confidence interval can therefore be interpreted as the estimates and uncertainty from the analysis described, but they cannot establish how the trial’s full testing strategy was handled.

The Full Analysis Set included participants who received study drug during the treatment period and were evaluated for at least one efficacy endpoint during that period. This definition is narrower than a simple statement that every randomized participant was analyzed, and the registry record does not provide a full accounting of exclusions or analysis-population flow. The record gives different analysis denominators for some efficacy and safety summaries; those should be read as the denominators attached to each posted analysis, not reconciled by adding or assuming participants across groups.

Missing-data handling is especially relevant to repeated-measures analyses because not every participant necessarily contributes observations at every assessment. The MMRM approach is stated, but the assumptions about missingness, imputation or other handling, reasons for missing observations, and sensitivity analyses are not reported in the registry record. Without those details, the robustness of the primary estimate to missing data cannot be assessed from the posted record alone.

The record also does not report a sample-size justification or the assumptions used to plan enrollment. The separate planning illustrations are based on observed effects and stated calculation assumptions; they should not be mistaken for the original design rationale or for observed power. These gaps limit a complete appraisal of the inferential plan, but they do not alter the reported endpoint estimates and intervals.

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: colitis ischaemic (0.5%), anaphylactic reaction (0.5%), femur fracture (0.5%).

The safety summaries report serious adverse events for 0 of 186 participants in the placebo group and 3 of 189 participants in the gabapentin enacarbil group. Deaths were reported for 0 participants in each group. The listed serious adverse event terms were colitis ischaemic, anaphylactic reaction, and femur fracture, each reported as 0.5%.

Other adverse events were reported for 32 of 186 participants in the placebo group and 59 of 189 participants in the gabapentin enacarbil group. These are descriptive counts within the reported safety denominators. They do not, by themselves, establish that a study treatment caused an event, nor do they quantify event severity, duration, or clinical course. The registry record does not report additional detail about those events in the registry record.

Safety comparisons require attention to the denominator, the event definition, follow-up, and how events were collected and judged. The listed serious-event terms and counts are useful signals to describe, but rare-event summaries can remain uncertain and should not be interpreted as a complete profile of risk. No inference about causality or comparative safety beyond the posted counts is warranted from these figures alone.

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

The primary result estimates a greater average reduction in IRLS score in the gabapentin enacarbil group than in the placebo group, but the between-group estimate is imprecise. The reported LSM difference is -1.2, with a 95% confidence interval from -2.6 to 0.2. Because that interval includes 0, the data are compatible with no between-group difference as well as with effects in either direction within the interval. The reported p-value is 0.088. Under a conventional p < 0.05 reference, that result does not meet the threshold, but the registry record does not provide the full planned decision rules.

This does not mean that the treatments have been shown to be equivalent. Equivalence requires a design and analysis with a defined margin and an appropriate interval-based assessment; such a margin is not reported in the registry record. Nor does the result prove that gabapentin enacarbil has no effect. The estimate points toward a greater reduction, while the uncertainty leaves the size and direction of the true contrast unresolved.

The group-level changes should also be distinguished from the comparative effect. The estimates of -10.5 and -11.7 are changes from baseline within the respective groups. The causal question is not answered by observing improvement in either arm alone; the randomized comparison is the relevant basis for estimating a treatment contrast. The placebo group also showed a substantial mean change, so within-group change cannot be attributed solely to the active treatment.

The result applies to the endpoint, time point, model, and analysis population stated in the registry record. It does not establish effects on all symptoms, long-term outcomes, or populations not described. The secondary outcomes and safety summaries provide additional context, but they do not remove the uncertainty in the primary comparison. The clearest reading is therefore a directionally favorable point estimate with an interval that remains compatible with no difference, rather than a definitive conclusion about efficacy or absence of efficacy.

14. Why This Trial Matters Statistically

This trial illustrates why a treatment comparison should be interpreted through its estimated contrast and uncertainty, rather than through within-group change alone. Both groups had lower mean IRLS scores at week 12, but the primary question is whether their changes differed. The sponsor’s model-based contrast and its confidence interval address that question more directly than either arm’s change considered in isolation.

The example also shows the value of matching the statistical summary to the design. Repeated measurements were analyzed with an MMRM, and the registry specifies covariates and a covariance structure. That makes the result a model-based estimate rather than a simple comparison of endpoint averages. Readers need to distinguish this sponsor analysis from an unadjusted re-calculation using summary values, which may not reproduce the fitted model.

Uncertainty matters even when a point estimate has a clear direction. Here, the confidence interval includes 0, and the p-value is 0.088. Those results caution against treating the apparent direction as a settled conclusion, while also avoiding the opposite mistake of declaring equivalence or no effect. Effect estimation and interval width communicate more than a binary label alone.

Finally, the study includes several secondary measures and safety summaries. These can illuminate different aspects of the outcome, but multiplicity and missing-data details affect how much weight to place on individual findings. The registry record’s lack of reported information on those issues is itself important: statistical interpretation depends not only on the reported estimates, but also on knowing how the analyses were planned and how incomplete observations were handled.

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 features limit what can be concluded from the registry record. The primary contrast is uncertain: its 95% confidence interval extends from -2.6 to 0.2, and the p-value is 0.088. The record does not report a prespecified significance threshold or the full analysis plan, so the result cannot be assessed against a complete protocol-defined decision framework.

The population is described as Japanese participants with moderate-to-severe idiopathic restless legs syndrome, and the primary analysis uses the Full Analysis Set definition in the record. The applicability of the results to other populations is not established by these details. The full gabapentin enacarbil dosing schedule is also not reported in the available arm description, which limits a complete description of the intervention as administered.

The record does not report sample-size justification, multiplicity procedures, interim analyses, or detailed missing-data methods and sensitivity analyses. Those omissions restrict appraisal of precision planning, the interpretation of secondary tests, and the robustness of the repeated-measures result. The posted planning illustrations are based on observed effects and are not evidence of the original design assumptions or observed power.

The safety information is summarized as event counts and a short list of serious-event terms. It does not provide the event-level clinical context needed to assess causality, duration, severity, or timing. Denominators differ across some efficacy and safety summaries, and the record does not provide enough participant-flow detail to explain all such differences. These limitations do not invalidate the posted estimates, but they narrow the conclusions that can be drawn from them.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study of Oral Dosing of Gabapentin Enacarbil in Japanese Restless Legs Syndrome Patients — NCT03053427.
• Registry record last updated 2024-12-12; status last verified 2024-11.
• Results first posted 2019-07-10.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

The registry record describes a completed phase 4 randomized, quadruple-masked, placebo-controlled, parallel-group study of oral gabapentin enacarbil in Japanese participants with restless legs syndrome. The recorded enrollment is 375. The primary endpoint was change from baseline in IRLS score at week 12, analyzed with an MMRM in the Full Analysis Set.

The reported LSM difference for gabapentin enacarbil versus placebo was -1.2, with a 95% confidence interval from -2.6 to 0.2 and p = 0.088. The point estimate favors a greater reduction in the active-treatment group, while the interval includes 0 and leaves the magnitude uncertain. Secondary outcomes cover global impressions, sleep, pain, health status, and adverse events. The registry record does not report a multiplicity strategy, detailed missing-data methods, or the sample-size justification.