Topiramate Treatment of Hazardous and Harmful Alcohol Use in Veterans With TBI: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Topiramate Treatment of Hazardous and Harmful Alcohol Use in Veterans With TBI (NCT01750268), a randomized phase 4 study. The primary outcome is Change in the Number of Drinking Days Per Week as Assessed by the Timeline Followback (TLFB).

2. Clinical Question

This is a question about a measured average change between randomized groups, not whether every participant changed in the same way. TLFB uses a calendar and participants’ retrospective estimates of daily drinking over a specified period. The primary endpoint therefore represents reported drinking frequency across the assessment period, expressed as drinking days per week.

The registry describes the project as a pilot controlled clinical trial intended to address co-occurring TBI and hazardous or harmful alcohol use in veterans. The principal posted comparison concerns drinking frequency. The registry also lists a secondary measure of TBI symptom severity. Results from a pilot study can describe observed group outcomes and their uncertainty, but they do not by themselves establish a precise treatment effect or settle how the result would generalize beyond the enrolled participants.

3. Trial Design

The registry identifies this as a Phase 4, randomized, double-masked, parallel-group trial, with an enrollment of 32. The status is completed. The listed study period began in 2012-11, and primary completion is recorded as 2015-10. The sponsor is the University of California, San Francisco. The registry describes the project as a pilot controlled clinical trial of topiramate for veterans with co-occurring TBI and hazardous or harmful alcohol use.

The two arms with posted primary outcome results are topiramate and placebo. The topiramate arm included 15 participants in that outcome result, and the placebo arm included 17. The arm descriptions specify daily capsules, up to 300 mg, for both drugs. Medical Management Counseling is listed among the interventions, but the registry record does not report enough detail to establish how counseling was assigned, standardized, or incorporated into each arm.

Randomization means that treatment assignment is intended to reduce systematic differences between groups at the time of assignment. Double masking indicates that the record identifies the study as masked at two levels, but it does not state which people were masked or how masking was maintained. Allocation concealment procedures, eligibility details, recruitment methods, and participant flow beyond the reported enrollment and outcome denominators are not reported in the registry record.

The trial is small in the context of estimating a between-group mean difference: the primary comparison uses 15 and 17 participants. A small pilot can be useful for describing feasibility and obtaining preliminary outcome estimates, but the registry record does not report the original sample-size justification, intended recruitment target beyond the enrollment figure, or a prespecified precision goal. Those details should not be inferred from planning illustrations based on the observed effect.

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:

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 in the Number of Drinking Days Per Week as Assessed by the Timeline Followback (TLFB), measured over Baseline to Week 12.

The primary endpoint is change in the number of drinking days per week, assessed by Timeline Followback (TLFB) from baseline to week 12. It is expressed as a mean and standard deviation in each arm. The registry description says participants use a calendar to provide retrospective estimates of daily drinking over a specified period.

Drinking days per week captures frequency rather than total alcohol consumed, drinking intensity, or the consequences of drinking. It provides a count-based summary of the number of days on which drinking was reported, expressed on a weekly scale. Comparing change from baseline to week 12 is intended to describe how that measure shifted over the assessment period in each randomized group.

As a general statistical point, a continuous endpoint summarized by a mean can be useful for comparing average outcomes between groups. The mean does not show the full distribution: participants may have had very different individual experiences, and a group average does not describe each person’s outcome. The standard deviation gives information about variation among participants, but it is not a confidence interval for the group mean.

The registry record does not report the detailed TLFB recall window used for each assessment, rules for handling incomplete calendars, or a fuller scoring specification for the change measure. The posted results should therefore be interpreted as the registry’s reported endpoint summaries, without assuming unreported measurement or data-processing details.

7. Results Posted in the Registry

• Topiramate: mean 2.20 versus 1.60 on control; difference 0.60 (95% CI -0.75 to 1.95), standardized effect d = 0.31; z-test p = 0.384.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Change in TBI Symptom Severity as Assessed by the Neurobehavioral Symptom Inventory (NSI)

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Topiramate: mean 16.30 versus 19.30 on control; difference -3.00 (95% CI -12.77 to 6.77), standardized effect d = -0.21; z-test p = 0.547.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Topiramate and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists TBI symptom severity assessed with the Neurobehavioral Symptom Inventory as a secondary outcome, and numerical results are posted for that measure. Across the posted primary and secondary comparisons, the estimates have confidence intervals that include zero. Both are uncertain estimates rather than precise evidence of a directional difference. Their point estimates are on different outcome scales, so their numerical sizes should not be compared directly.

The secondary outcome relates to the primary result by addressing a distinct aspect of the participants’ condition. It does not act as a confirmation of the drinking-days result, and it cannot resolve the uncertainty in the primary comparison. The registry lists three outcome measures overall, while numerical results for one secondary outcome are posted; results for the other outcome measures are not reported in the registry record.

Unadjusted secondary results generally carry less evidential weight when several outcomes or analyses are considered. As a general statistical issue, examining multiple outcomes increases the chance that at least one estimate may appear more notable simply through sampling variation. This does not mean a secondary result is uninformative, but its interpretation depends on whether it was prespecified, how many other comparisons were examined, and whether multiplicity was addressed. Those details are not reported in the registry record.

The registry reports no formal analyses and does not describe a hierarchy of endpoints or a multiplicity procedure. The secondary result should therefore be treated as an additional descriptive finding, with its own uncertainty, rather than as independent proof that strengthens or overturns the primary comparison. It also should not be read as an exhaustive account of every outcome listed for the trial.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Change in TBI Symptom Severity as Assessed by the Neurobehavioral Symptom Inventory (NSI)

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous and reported as a mean with a standard deviation in each group. The comparison is expressed as a mean difference: the topiramate mean minus the placebo mean. On the endpoint’s original scale, the estimate is 0.6 drinking days per week. This direct scale is usually the most interpretable for understanding the size and direction of a difference in the measured outcome.

What does the mean difference represent?

In general, a difference between two sample means estimates the difference between the corresponding group averages in the populations from which the samples arose. A value of zero would represent no difference in the two averages. A positive estimate means the first group’s reported mean is higher; a negative estimate means it is lower. Here, the registry comparison defines the order as topiramate minus placebo. Because the endpoint is described as change, interpreting its direction also depends on the coding convention for change; details beyond the reported summaries are not reported in the registry record.

What does the confidence interval add?

In general, a confidence interval gives a range of values compatible with the data under the method used to calculate it. It reflects sampling uncertainty and is not a range containing a fixed proportion of individual participant outcomes. The primary 95% confidence interval is -0.751 to 1.951 drinking days per week. Since it includes zero, the data are compatible with no group difference as well as with differences in either direction. The interval should be considered alongside the point estimate rather than treating the estimate alone as conclusive.

What does the p-value say?

In general, a p-value measures how unusual the data, or more extreme data, would be under a specified null model. It is not the probability that the null hypothesis is true, and it does not quantify the probability that a result occurred by chance. The posted primary p-value is unadjusted and equals 0.384. The registry record does not report a multiplicity adjustment or a detailed test specification. A p-value should not replace consideration of the effect estimate, its scale, and its confidence interval.

Why do standard deviations matter?

In general, a standard deviation describes the spread of individual observations around their group mean. Greater variability can make a mean difference harder to estimate precisely when sample sizes are unchanged. It is distinct from the standard error, which describes the sampling variability of an estimate. The primary group standard deviations are 1.8 and 2.1. The comparison also reports a pooled standard deviation of 1.966, used for a standardized effect description. Standardization expresses a difference relative to variability, but it does not replace the original-scale estimate when the endpoint’s units are meaningful.

Why can over-dispersion matter?

As a general concept, over-dispersion arises when observed variation exceeds what a statistical model assumes. It matters because a model that understates variability can produce standard errors that are too small and intervals that are too narrow. The primary results are presented as means and standard deviations, and the record does not report a model-based count analysis or an assessment of over-dispersion. Since drinking days are count-like, alternative modeling choices could be relevant in some analyses, but no unreported model or diagnostic should be attributed to this trial.

What can be inferred from the posted summaries?

The arm means, standard deviations, sample sizes, mean difference, confidence interval, and unadjusted p-value describe the reported comparison. The record does not report adjusted estimates, covariate selection, missing-data methods, or an analysis population such as intention-to-treat. It also reports no formal analyses in the registry. Accordingly, the interpretation should stay with the posted unadjusted group summaries and their uncertainty rather than assuming additional modeling or design features.

As a general principle, randomization supports a causal comparison when assignment is implemented as intended and outcomes are analyzed appropriately, but the registry summaries alone do not document all those conditions. The record identifies randomization and double masking, yet does not report allocation procedures, adherence, crossover, attrition details, or masking checks. These omissions limit assessment of how closely the reported estimates reflect the intended randomized comparison.

11. Confidence Intervals and Precision

The primary mean difference is 0.6 drinking days per week, with a 95% confidence interval from -0.751 to 1.951. The interval includes zero and extends from a negative difference to a positive difference. This means the posted data do not precisely establish the direction or magnitude of the between-group difference. The estimate is best read together with the interval, rather than as a standalone value.

A confidence interval describes uncertainty in an estimated group difference, not the range of individual participants’ drinking-day changes. The standard deviations of 1.8 and 2.1 describe observed variability in the two arms. The distinction matters: high participant-to-participant variation and small group sizes can leave an average difference uncertain, even when a point estimate is available.

The primary comparison contains 15 participants in the topiramate arm and 17 in the placebo arm. With these reported group sizes, the confidence interval is wide relative to the point estimate. The registry record does not report a prespecified precision target or the original sample-size justification, so the interval should not be judged against an unreported design goal.

The secondary NSI mean difference is -3.0 scores, and its 95% confidence interval extends from -12.77 to 6.771. It also includes zero and is wide enough to encompass differences in either direction. The registry does not report a multiplicity-adjusted interval. These intervals are descriptive of the posted comparisons and should not be interpreted as proof of equivalence or as guarantees about future estimates.

12. 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 events. This is a continuous-outcome comparison, so the entries describe the illustrative number of participants needed in each of two equally allocated arms. The associated curve represents how the planning requirement changes as the assumed effect changes; it should be read as a design illustration, not as a count of participants actually enrolled.

The calculations use a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set at the stated fraction of the observed effect: 100%, 75%, or 50%. The 80% power column means the illustrative per-arm sample size for an 80% probability of detecting the assumed effect under those planning assumptions; the 90% column uses a higher target probability of detection. These are conventional design probabilities under a specified model, not guarantees for a future study.

For the primary endpoint, the listed figures are 169 and 226 participants per arm for 100% of the observed effect, 300 and 401 for 75% of the observed effect, and 674 and 903 for 50% of the observed effect, for the 80% and 90% targets respectively. Smaller assumed effects require more participants because they are harder to distinguish from sampling variation.

These figures are planning illustrations based on the observed effect, not the trial’s original design assumptions; those original assumptions are not reported in the registry record. This is not observed power. The calculations do not show the probability that the completed trial’s result is true, and they should not be used to reinterpret the observed p-value.

13. 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 = 0.600; pooled standard deviation σ = 1.966.
2. n = 2 × (1.96 + 0.842)² × 1.966² / 0.600² = 168.5, rounded up to 169 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- 0.6                      # observed difference in means
sigma <- 1.9657059800489                     # 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(1750268)
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 = 169 (80% power), 226 (90% power)
 75% of observed effect: n per arm = 300 (80% power), 401 (90% power)
 50% of observed effect: n per arm = 674 (80% power), 903 (90% power)
power.t.test: 169.5 
Simulated power at n = 169 per arm: 0.798 (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.

14. Multiplicity, Missing Data and Analysis Populations

The registry identifies a primary outcome and lists secondary outcomes, but it does not report a statistical analysis plan describing how multiple outcomes or analyses were handled. No formal analyses are reported in the registry. The record does not state whether a multiplicity adjustment, a prespecified testing hierarchy, or another strategy was used. The posted p-values are unadjusted.

As a general statistical concern, when several hypotheses are tested, the chance of at least one apparently unusual result can increase. Adjustments or a prespecified testing order can help control interpretation across multiple comparisons. Without the planned procedure and the full set of analyses, it is not possible to determine how the registry’s individual p-values fit into the trial’s intended inferential framework. This limitation is especially relevant to secondary outcomes, which are often interpreted as supportive or exploratory.

Missing-data handling is not reported in the registry record. The primary outcome denominators are 15 in the topiramate arm and 17 in the placebo arm, but the record does not describe whether these are all randomized participants, participants with an observed week 12 assessment, or another analysis population. It does not report the amount or pattern of missing outcome data, reasons for missingness, imputation procedures, or sensitivity analyses.

The analysis population is also not reported. The record does not say whether the posted comparison follows an intention-to-treat approach, a per-protocol approach, or another defined population. Details on adherence, treatment discontinuation, and crossover are not reported. Without those details, readers cannot fully assess how exclusions or missing observations might affect the estimate.

The registry also does not report the trial’s original sample-size justification, interim analyses, or the prespecified statistical test and model. These omissions do not invalidate the posted descriptive statistics, but they restrict assessment of whether the reported results match the original analysis plan and how robust the findings are to alternative reasonable choices.

15. 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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: inpatient medical hospitalization for an episode of chest pain in the context of cocaine use (6.7%), voluntary psych inpatient admission due to homicidal ideation in the context of heavy cocaine use (6.7%), inpatient medical hospitalization due to alcohol withdrawal (6.7%), inpatient medical hospitalization for planned distal pancreatectomy and splenectomy (5.9%).

The registry reports serious adverse events among 3 of 15 participants at risk in the topiramate group and 1 of 17 at risk in the placebo group. Other adverse events are recorded for 15 of 15 participants at risk in the topiramate group and 17 of 17 in the placebo group. The recorded death count is 0 in each group. These are registry counts and should be read in relation to the reported numbers at risk.

The serious-event terms listed include inpatient medical hospitalization for chest pain in the context of cocaine use, voluntary psychiatric inpatient admission due to homicidal ideation in the context of heavy cocaine use, inpatient hospitalization due to alcohol withdrawal, and inpatient hospitalization for planned distal pancreatectomy and splenectomy. The registry lists percentages of 6.7%, 6.7%, 6.7%, and 5.9% for these terms, respectively. The record does not provide a treatment-relatedness assessment for these events in the information.

Adverse-event counts alone do not establish that an event was caused by an assigned treatment. The small group sizes also mean that these counts are descriptive, not precise comparative estimates of safety. The registry record does not report further detail on event timing, duration, severity beyond serious-event classification, or whether individuals experienced more than one event.

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

The posted primary result is an estimated difference in mean change in drinking days per week between two randomized groups, with considerable uncertainty. The mean difference is 0.6, and its 95% confidence interval ranges from -0.751 to 1.951. The interval includes zero, so the result does not establish a clear between-group difference. It also does not establish equivalence or show that any possible difference is necessarily unimportant.

The unadjusted p-value of 0.384 is not a probability that the treatment effect is absent. It is a summary of how the observed data relate to a null model under the analysis used. The estimate, confidence interval, outcome definition, and size of the groups all matter for interpretation. The secondary NSI estimate is also uncertain and should not be treated as a substitute for the primary endpoint.

These results do not show what happened to every participant, because group means summarize outcomes across participants. Nor do they identify the mechanism behind any observed changes. The registry record does not report adherence, missing-data methods, baseline balance, an adjusted analysis, or a detailed analysis population. Those omissions limit conclusions about the robustness and scope of the comparison.

The study is recorded as a pilot trial with 32 enrolled participants. Its estimates may be useful as preliminary information, but the planning illustrations based on the observed effect are not a substitute for the trial’s original design assumptions, which are not reported. The observed data do not by themselves determine what a larger study would find.

Nothing in these statistical results should be interpreted as individual medical advice or as a definitive judgment about treatment. They describe the registry’s posted group-level outcomes and the uncertainty around those outcomes for the study population. Generalizing to other people or settings requires evidence beyond the summaries reported in this record.

17. Why This Trial Matters Statistically

This trial raises several important statistical issues in a compact example: a small randomized comparison, a continuous primary outcome, a mean difference on a clinically interpretable scale, and broad uncertainty around the estimate. It demonstrates why the point estimate alone is not enough. Reading the confidence interval alongside the observed difference makes clear that the data are compatible with a range of possibilities.

The distinction between observed variability and estimate precision also matters. Standard deviations describe how outcomes varied among participants, while the confidence interval describes uncertainty in the between-group difference. Both help explain why small group sizes can yield estimates that remain uncertain even when complete summary statistics are posted.

The registry’s planning illustrations also highlight a common design principle: detecting smaller effects generally requires larger samples. However, planning based on an observed effect is different from the original prospective design rationale. The registry does not report its original sample-size assumptions, so the illustrative planning entries should not be mistaken for the study’s actual planning process or for evidence of observed power.

Finally, the record shows why analysis transparency matters. Information on multiplicity, missing data, analysis populations, and the statistical plan is needed to assess how the posted comparisons were selected and how robust they may be. The results are informative as preliminary summaries, but those missing details constrain stronger inference.

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.

Several limitations are specific to interpreting this trial’s registry results. First, the primary comparison has 15 participants in the topiramate arm and 17 in the placebo arm. This leaves the estimated mean difference relatively uncertain, as reflected in the confidence interval. The registry describes the project as a pilot but does not report a prospective sample-size justification or target precision.

Second, the posted comparison is unadjusted. The registry record does not report baseline characteristics, adjustment for baseline outcome values, or other covariate analyses. It is therefore not possible to assess from the posted results whether the groups differed on important baseline measures or whether an adjusted estimate would differ from the reported comparison.

Third, the record does not describe missing observations, participant flow by assessment, adherence, treatment discontinuation, crossover, or the analysis population. Although the primary outcome has reported group denominators, their relationship to randomization and follow-up is not detailed. The effects of missingness or exclusions cannot be evaluated from these summaries.

Fourth, the outcome is based on retrospective participant estimates using a calendar. The record does not report the recall period, validation procedures, or how incomplete reports were handled. The intervention listing includes Medical Management Counseling, but the record does not explain its implementation or distribution between arms.

Finally, the registry does not report a detailed statistical analysis plan, multiplicity approach, or results for all listed outcome measures. The available results therefore offer a limited view of the trial’s full analytical context. These limitations do not alter the arithmetic of the posted summaries, but they restrict how confidently those summaries can be generalized or interpreted as definitive evidence.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Topiramate Treatment of Hazardous and Harmful Alcohol Use in Veterans With TBI — NCT01750268.
• Registry record last updated 2020-10-14; status last verified 2020-09.
• Results first posted 2020-10-14.
• 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 describes a completed Phase 4 randomized, double-masked, parallel pilot trial of topiramate versus placebo in veterans with TBI and hazardous or harmful alcohol use. Enrollment was 32. The primary endpoint was change in drinking days per week from baseline to week 12, assessed by TLFB.

The posted primary mean difference was 0.6 drinking days per week, with a 95% confidence interval from -0.751 to 1.951 and an unadjusted p-value of 0.384. A secondary NSI outcome was also posted, with an uncertain mean difference. The registry does not report the original sample-size justification, analysis population, missing-data methods, multiplicity plan, or detailed statistical analysis plan.