The Effects of Topiramate on Alcohol Use in Alcohol Dependent Subjects: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for The Effects of Topiramate on Alcohol Use in Alcohol Dependent Subjects (NCT00329407), a non-randomized phase 4 study. The primary outcome is The Primary Outcome Measure Will be Subjects Ethanol Consumption Over the Course of the Drug Treatment Period as Assessed by the Timeline Followback Method.

2. Clinical Question

3. Trial Design

This was a Phase 4, non-randomized, single-group study sponsored by Boston University. The registry records 10 enrolled subjects and a completed status. The study began in 2003-09, with primary completion in 2008-06. These dates describe the record; they do not supply the duration of treatment for any individual subject.

The treatment was topiramate, identified in the record as Topamax. The study was open label, with no masking, and had no placebo control. All subjects received the active medication. The dosing description gives a stepwise escalation, including 50 mg q PM on Days 1-3 and 150 mg BID on Days 20-23. The entry ends with “Days 24-27” without a dose, so the remaining schedule is not reported in the registry record.

The primary outcome was framed as a comparison of alcohol consumption within the treated group. The analysis description says that least-squares values for baseline and week 10 were compared using a t test with Dunnett-Hsu adjustment. The registry identifies an intention-to-treat approach, but does not describe how that population was defined or how missing observations were handled.

The registry lists two outcome measures and two analyses. It names the phonetic portion of the Controlled Word Association Test as a secondary outcome. It does not give the secondary outcome results or enough detail to determine how the analyses were grouped for multiplicity adjustment. The design can describe change during treatment, but, without a concurrent control group, it cannot isolate a treatment effect from other changes over time.

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

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: An ITT approach was used in the analysis. Least squares value for the baseline and 10 week of treatment were compared using a t test with Dunnett-Hsu adjustment. Differences between these means are presented as the result.

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 The Primary Outcome Measure Will be Subjects Ethanol Consumption Over the Course of the Drug Treatment Period as Assessed by the Timeline Followback Method, measured over 70 days.

The primary endpoint was mean daily consumption of standard alcoholic drinks, with a standard drink defined as 14 g of alcohol. Alcohol consumption was assessed using the Timeline Followback method. The registry describes the principal contrast as baseline-week consumption versus consumption during week 10, the final week when subjects were on one maintenance dose of topiramate. The stated timeframe for the outcome was 70 days.

A mean gives a measure of the group’s average consumption, making it possible to express the before-and-after contrast in the original outcome units: standard drinks per day. This can be easier to interpret than a test statistic alone. The registry says that differences between least-squares means were to be presented as the result, but it does not report the estimated difference, its uncertainty, or the underlying means.

Repeated assessment of alcohol consumption is relevant to the study’s stated aim of assessing whether topiramate reduces ethanol consumption. However, the endpoint alone does not establish why consumption might change. In a single-group study, the baseline comparison does not account for changes that could occur over time for reasons other than treatment. The available record therefore describes the intended measure and contrast, not a reported finding about consumption.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes a continuous primary outcome: mean daily consumption of standard drinks. For a continuous measure, a mean summarizes the central level of observations in the group. In this study the intended units were standard drinks per day, with a standard drink defined as 14 g of alcohol. The mean is interpretable in those units, but it does not show how much individual subjects varied around the group average. The record does not report a standard deviation or other observed measure of spread for this endpoint.

What does a baseline-to-week-10 comparison estimate?

In general, a before-and-after comparison contrasts measurements taken at different times in the same study group. Here, the registry specifies the baseline week and week 10. The described result is the difference between the corresponding least-squares means. A least-squares mean is a model-based mean; depending on the analysis, it can account for terms included in the model. The registry does not describe the model terms or other adjustments, so the precise basis for these means is not reported in the registry record.

The t test is a method for assessing a contrast relative to its estimated uncertainty under the test’s assumptions. A test result, such as a p value, does not give the size of the change by itself. An estimate in the original units and an interval expressing uncertainty would help readers judge both magnitude and precision. The registry does not report an estimate, confidence interval, or p value for the primary endpoint. A confidence interval is, in general, a range of values compatible with the data and the method’s assumptions; it is not reported in the registry record for this outcome.

Why is Dunnett-Hsu adjustment mentioned?

In general, multiplicity adjustment addresses the increased chance of a false-positive finding when several statistical comparisons are considered together. Dunnett-Hsu is an adjustment approach used in settings involving multiple comparisons. The registry states that the t test used a Dunnett-Hsu adjustment, but does not identify the full family of comparisons, the hypotheses included, or the adjusted test result. Although the record lists two outcome measures and two analyses, it does not establish exactly which analyses were included together for adjustment. The adjustment’s practical effect on the primary result cannot be assessed without the analysis details and results.

Adjustment for multiplicity is distinct from adjustment for confounding. A multiplicity procedure addresses the interpretation of multiple statistical tests; it does not create a control group or remove the limitations of a single-group study. General statistical reasoning therefore distinguishes the question of false-positive risk across tests from the question of whether changes over time can be attributed to treatment.

What does intention-to-treat mean in this record?

Intention-to-treat generally refers to analyzing participants according to the group to which they were assigned, regardless of adherence or treatment changes. In a single-group study, the usual between-group allocation distinction is not present. The registry says an ITT approach was used but does not explain how that approach was operationalized, which observations were included, or how participants with missing endpoint measurements were treated. The label alone is not enough to reconstruct the analysis.

Missing-data methods matter because the observed endpoint may not be available for every participant. In general, an analysis that excludes missing measurements and one that imputes or models them can depend on different assumptions. The registry does not report the amount or pattern of missing primary-outcome data, or a method for handling it. No assumption about missingness or participant completion can therefore be made.

How should sample-size planning illustrations be read?

General sample-size calculations estimate the amount of information a future study might need under specified assumptions. The registry’s planning assumptions describe a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and effects expressed as fractions of the observed effect. Such calculations are planning illustrations, not evidence about the precision or power actually achieved in this trial. They should not be described as observed power.

Those assumptions also describe a two-arm planning setup, whereas the registered trial is single group and non-randomized. A two-arm illustration is not a description of the actual treatment allocation in this study. The record does not report the planning outputs or the trial’s own sample-size justification, so no conclusion about the adequacy of the enrolled sample for the planned analysis can be drawn.

Finally, statistical significance and clinical importance are different concepts in general. A test addresses compatibility with a statistical null under specified assumptions; an effect estimate in meaningful units helps assess size. Neither can be evaluated for the primary outcome from the registry information because the outcome estimate and its uncertainty are not reported.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Topiramate Treatment")
n   <- c(10)
m   <- c(-4.9)
sp  <- c(0.86)
sd  <- sp * sqrt(n)      # standard errors converted to SDs
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Topiramate Treatment  10  -4.900  2.720  -6.845  -2.955

11. Multiplicity, Missing Data and Analysis Populations

The registry states that an intention-to-treat approach was used and that the baseline and week-10 least-squares values were compared with a t test using Dunnett-Hsu adjustment. It also lists two outcome measures and two analyses. These details indicate that the analysis involved an adjustment for multiplicity, but do not establish which comparisons made up the adjustment family or how the two listed analyses were related. The adjusted result is not reported in the registry record.

The analysis population is not fully described. The ITT label appears in the record, but its implementation is not specified. In a single-group study, the record would need to clarify which enrolled subjects contributed to each time point and how the stated analysis population was defined. That information is not reported in the registry record.

Missing data are another important uncertainty. The record does not state how many primary endpoint measurements were available at baseline or week 10, whether any observations were missing, or what method was used to address missingness. In general, missing-data approaches rely on assumptions about the relationship between observed and unobserved values; different assumptions can affect estimates and uncertainty. No such assumptions or sensitivity analyses are reported.

The record does not describe an interim analysis, stopping rule, or sample-size justification. It also does not report a prespecified statistical analysis plan beyond the stated comparison and adjustment. Those details should not be inferred from the listed methods. As a result, the degree to which the analysis followed a more detailed plan, and the robustness of the finding to alternative missing-data or multiplicity choices, cannot be assessed from the registry record.

12. 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 5% reporting threshold.

Most frequent non-serious adverse event terms

The registry lists 10 subjects at risk for serious adverse events and 0 affected. It lists 10 at risk for other adverse events and 2 affected. No serious adverse-event terms are listed. The deaths field has no reported value, so deaths are not reported in the registry record.

These entries are counts within the topiramate treatment group. They are not comparisons against placebo or another treatment, because the study was single group and had no control arm. The counts do not establish whether an event was caused by treatment. The record does not provide event descriptions for the other adverse events, their severity, timing, duration, or whether any subjects stopped treatment because of them.

In general, an adverse-event count describes events recorded under the study’s reporting definitions; it does not by itself provide a complete account of safety. The registry’s safety fields should therefore be read as limited descriptive reporting. The absence of listed serious-event terms does not supply details about how events were assessed or collected, and the record does not report a comparative safety estimate.

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

The study’s stated purpose was to assess topiramate in relation to ethanol consumption among alcohol-dependent subjects, with safety and possible abuse also identified as goals. The primary outcome was mean daily standard-drink consumption at baseline compared with week 10. The registry also names phonetic verbal fluency on the Controlled Word Association Test as a secondary outcome.

The posted record does not report the primary outcome result. It gives no baseline or week-10 mean, estimated difference, confidence interval, or p value. Consequently, it does not show the direction, size, or precision of any change in alcohol consumption. The presence of a planned outcome and statistical test cannot be taken as evidence that consumption increased, decreased, or remained similar.

The study design also limits what a change could mean. All subjects received topiramate in an open-label, single-group study, without a placebo or concurrent control. In general, a change from baseline can reflect treatment and other influences that coincide with time. A within-group comparison alone cannot distinguish among those explanations. The record therefore does not support a causal conclusion about the effect of topiramate on alcohol consumption.

The adverse-event counts are descriptive and do not resolve this limitation or establish comparative safety. The registry reports 0 subjects affected by serious adverse events and 2 affected by other adverse events, but does not provide detailed event information or a control group.

The appropriate reading is narrow: the registry documents a completed study, its planned outcome, its stated analysis approach, and limited adverse-event counts. It does not report enough outcome information to judge the primary result, and its design does not provide a randomized treatment comparison. Broader claims about benefit, lack of benefit, or comparative safety are not supported by the registry record.

14. Why This Trial Matters Statistically

This trial is statistically notable for the distinction between measuring change and attributing change. The study was designed around a before-and-after contrast in mean daily alcohol consumption. Such a contrast can describe the observed course of an outcome in a treated group, but a single-group design does not provide a concurrent reference for what might have happened without treatment. That distinction remains important even when the outcome and analysis are clearly specified.

The record also illustrates why an analysis plan and a result must be read separately. It names Timeline Followback, identifies baseline and week 10, and specifies a t test with Dunnett-Hsu adjustment. Yet it does not report the primary estimate or its uncertainty. Without those quantities, statistical precision and the magnitude of any observed difference cannot be evaluated.

Multiplicity is another relevant statistical issue. The record names more than one outcome measure and analysis and reports an adjustment, but does not explain the comparison family. This shows why naming an adjustment method is not enough to determine which conclusions it protects or how to interpret an adjusted result.

Finally, the registry’s planning assumptions describe an equal-allocation, two-arm calculation, while the registered study itself was single group. Planning illustrations can be useful for considering future studies, but they are not observed power and do not change the design that generated the data. The distinction helps prevent a planning exercise from being mistaken for evidence about the precision or strength of this completed study.

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 affect interpretation. The study was non-randomized, open label, and single group. All subjects received topiramate, with no placebo or active control. Thus, the planned baseline-to-week-10 contrast is not a randomized comparison and cannot separate treatment-associated change from other influences that may occur during the study period.

The registry lists enrollment of 10, but does not report how many subjects contributed data at each assessment. It does not provide the primary outcome estimate, the baseline and week-10 means, variability, a confidence interval, or a p value. The primary result’s direction, magnitude, and precision therefore cannot be assessed. The record also does not report the secondary verbal-fluency result.

Participant characteristics and detailed eligibility criteria are not reported in the registry record. The dosing schedule is incomplete: it lists doses through Days 20-23 and then an unfinished entry for Days 24-27. The record does not provide enough information to reconstruct the remaining schedule or assess adherence and treatment exposure.

The stated ITT approach is not described in operational detail, and missing-data handling is not reported. The exact family of tests covered by the Dunnett-Hsu adjustment is also unclear. These gaps limit assessment of the analysis and its robustness. Finally, adverse-event information is sparse: the registry gives counts but does not describe the other events, their severity, or their timing. These limitations restrict what can be concluded from the posted record; they do not establish what the unreported results would have shown.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: The Effects of Topiramate on Alcohol Use in Alcohol Dependent Subjects — NCT00329407.
• Registry record last updated 2010-09-08; status last verified 2010-08.
• Results first posted 2010-08-04.
• 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 4, non-randomized, open-label, single-group study sponsored by Boston University. It records enrollment of 10 and identifies topiramate as the intervention. All subjects received active medication, with no placebo control.

The primary endpoint was mean daily standard-drink consumption, assessed by Timeline Followback and compared between the baseline week and week 10. The planned analysis used least-squares values, a t test, and Dunnett-Hsu adjustment, with an ITT approach stated. The primary estimate and its uncertainty are not reported in the registry record. Phonetic verbal fluency on the Controlled Word Association Test is listed as a secondary outcome, but its result is not reported.

The safety entry lists 0 affected by serious adverse events and 2 affected by other adverse events. The record does not give further event details or a comparative group. The single-group design and missing outcome results limit interpretation of efficacy and safety.