An Efficacy and Tolerability Study of Topiramate in Participants With Migraine: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for An Efficacy and Tolerability Study of Topiramate in Participants With Migraine (NCT01060111), a randomized phase 4 study. The primary outcome is Percentage Decrease in Migraine Episodes.

2. Clinical Question

The registry describes this as a Phase 4, randomized, single-masked, parallel-group study, with enrollment of 250. The central statistical question is not just which group had the largest posted mean. It is how large the observed differences were, how uncertain those estimates are, and how the sponsor's reported analyses relate to the simple pairwise re-calculations. The available results do not establish that one titration approach is preferable. They show group summaries and analyses whose uncertainty and scope need to be read alongside the trial's design and incomplete reporting.

3. Trial Design

This completed Phase 4 study compared three parallel groups: Topiramate Standard, Topiramate Slow, and Topiramate Slow and Propranolol Booster. The registry lists randomized allocation and single masking. It records enrollment of 250 and a study status of completed. The record gives a start date of 2006-07 and a primary completion date of 2008-07. The sponsor is Janssen Korea, Ltd., Korea.

The treatment schedules differed in the pace of topiramate dose increases. The standard arm increased the dose at 1-week intervals, while the slow-titration arms used 2-week intervals. Both descriptions specify a maintenance dose of 50 mg to 100 mg twice daily up to Week 10, with dosing at physician discretion. The booster-arm description ends after “Propranolol 80”; its full schedule and dose units are not reported in the registry record, so the incomplete text cannot be used to characterize the propranolol regimen more fully.

The primary outcome is described as percentage decrease in migraine episodes, measured from baseline using a headache diary. Its timeframe is the maintenance period, Weeks 7 to 10. The endpoint description says migraine was diagnosed in accordance with International Headache Society guidelines. The primary population is described as participants who took topiramate at least once and had migraine improvement data at Week 6. The record does not explain how the Week 6 data requirement relates to the outcome timeframe of Weeks 7 to 10.

Three secondary outcomes are listed: change from baseline in migraine frequency, MIDAS score, and VAS score at Week 6. Although the record lists four outcome measures and five analyses, details for all measures and analyses are not available. The exact masking procedures, randomization method, allocation concealment, and original design assumptions are not reported in the registry record.

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: Out of 250 consented participants, 243 participants were assigned to study treatment.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The intent-to-treat (ITT) population included all the participants who took topiramate at least once and had migraine improvement data at the Week 6. Here 'N' signifies participants who were evaluated for this outcome measure.

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: Baseline characteristics were measured for Intent-to-treat (ITT) population which was defined as participants who took topiramate at least once and had migraine improvement data at the Week 6.

6. Primary Endpoint

The registry lists the primary outcome as Percentage Decrease in Migraine Episodes, measured over Maintenance period (Weeks 7 to 10).

The primary endpoint is the percentage decrease in migraine episodes from baseline during the maintenance period, identified in the registry as Weeks 7 to 10. The description says episodes were measured using a headache diary, and that migraine diagnosis followed International Headache Society guidelines. The outcome is summarized as a mean with standard deviation, making it a continuous measure rather than a count of participants who crossed a threshold.

A percentage change can summarize the direction and magnitude of change in episode frequency relative to baseline. In a randomized comparison, the group means can be compared to describe whether the average decrease differed across treatment approaches. This type of endpoint retains more information about the amount of change than a simple yes-or-no classification, although the interpretation still depends on how baseline values and diary data were handled.

The posted endpoint has an important timing detail: the timeframe is Weeks 7 to 10, whereas the population description requires migraine improvement data at Week 6. The registry record does not explain that distinction or give further details about diary completion, endpoint derivation, or missing observations. The outcome description also says that the reported N is the number evaluated for the endpoint, so the displayed group sizes should be understood as evaluated counts rather than automatically as every randomized participant.

7. Results Posted in the Registry

• Topiramate Slow: mean 57.00 versus 48.00 on control; difference 9.00 (95% CI -13.67 to 31.67), standardized effect d = 0.15; z-test p = 0.436.
• Topiramate Slow and Propranolol Booster: mean 46.00 versus 48.00 on control; difference -2.00 (95% CI -30.00 to 26.00), standardized effect d = -0.03; z-test p = 0.889.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous and reported as a mean with standard deviation. A mean summarizes the group average, while the standard deviation describes how spread out individual measurements are around that average. The posted standard deviations are large relative to the mean differences shown in the pairwise comparisons. In general, greater within-group variability makes group differences harder to estimate precisely because observations are less tightly clustered around the group mean.

What does a mean difference represent?

The pairwise re-calculations subtract the standard-arm mean from the comparison-arm mean. A positive estimate means the comparison group's posted mean is higher; a negative estimate means it is lower. The re-calculated difference for slow titration is 9.0, and the difference for the booster arm is -2.0. These are unadjusted comparisons derived from the posted group summaries, not adjusted estimates from the sponsor's full analysis. The record does not give covariate-adjusted estimates for these comparisons.

How should the confidence intervals be read?

A confidence interval gives a range of estimates compatible with the data and the method used to calculate it. It communicates uncertainty around an estimated difference rather than a range containing a specified proportion of individual participants. For the slow-versus-standard comparison, the 95% confidence interval is -13.67 to 31.67. For the booster-versus-standard comparison, it is -30.0 to 26.0. Both intervals include zero, the value corresponding to no mean difference. Their width indicates that the estimates do not narrowly distinguish among possible differences. The interval calculation is a general statistical summary and does not resolve limitations in the underlying data or analysis choices.

What do the p-values say?

The unadjusted p-values for the two pairwise re-calculations are 0.436 and 0.889. A p-value describes how unusual data at least as discrepant as those observed would be under a specified null model and its assumptions. It is not the probability that the null hypothesis is true, and it does not measure the size or practical importance of a difference. The sponsor's reported ANOVA p-value is 0.621 and concerns a three-group analysis. It should not be treated as interchangeable with either pairwise p-value: they address different comparisons and are reported from different analyses.

What is the sponsor's ratio analysis?

The sponsor reported a one-sided t-test for standard topiramate versus the booster arm, expressed as a ratio of percentage decrease. Its estimate is 0.950, with a 95% confidence interval from 0.519 to 1.737 and p = 0.5. A ratio compares quantities on a relative scale, unlike the mean differences, which are expressed on an additive scale. The registry labels this analysis non-inferiority or equivalence, but does not report the decision margin, the precise hypothesis, or enough detail to reproduce the sponsor's test. A confidence interval that includes the ratio null value does not, by itself, establish equivalence or non-inferiority.

Why does dispersion matter?

Standard deviation describes observed variability within each group. In general, over-dispersion means that data vary more than a chosen statistical model expects. When a model does not account for that extra variability, standard errors can be too small and confidence intervals too narrow. The primary outcome is a continuous percentage measure, and the record reports standard deviations rather than a count-model dispersion parameter. It does not report a special over-dispersion assessment. The broader lesson is to check whether the model matches the outcome and its observed variation; the record does not provide enough detail to evaluate all model assumptions.

How do the analysis population and timing matter?

The registry defines the primary population as participants who took topiramate at least once and had migraine improvement data at Week 6. It also identifies the endpoint timeframe as Weeks 7 to 10. The record does not explain why Week 6 data determine inclusion for an outcome described over a later period, or whether the population definition was applied as written in the sponsor analysis. The endpoint description says the reported N represents those evaluated. That makes the group counts relevant to interpreting the estimates, but it does not reveal how many randomized participants lacked evaluable outcome data or how missingness was handled.

Overall, the displayed pairwise calculations are useful descriptive re-analyses of the posted means and standard deviations. They are not a reconstruction of the complete sponsor analysis. The registry record does not report enough detail about model specification, transformations, covariates, analysis assumptions, or missing-data methods to reproduce the full analysis independently.

10. Confidence Intervals and Precision

The confidence intervals show that the estimated pairwise differences are not precise enough to isolate a narrow range of effects. For slow titration compared with standard titration, the estimate is 9.0 and the 95% confidence interval extends from -13.67 to 31.67. For the booster approach compared with standard titration, the estimate is -2.0 and the interval extends from -30.0 to 26.0. Each interval includes zero, so neither estimate rules out no mean difference under the calculation used.

Precision is about the width of the range of estimates, not whether an estimate is favorable or unfavorable. Here, the intervals allow differences in either direction. The group means alone can give an overly definite impression if the uncertainty is ignored. The reported standard deviations—69.0, 48.0, and 69.0—also show appreciable variation in the endpoint within groups, which contributes to uncertainty in comparisons.

The sponsor's ratio estimate for standard topiramate versus the booster arm is 0.950, with a 95% confidence interval from 0.519 to 1.737. This interval spans a broad range on the relative scale and includes the ratio value corresponding to no relative difference. The registry labels this as a non-inferiority or equivalence analysis, but the margin and full decision rules are not reported. Therefore the interval should be read as an estimate with substantial uncertainty, not as a stand-alone declaration of equivalence.

Confidence intervals quantify sampling uncertainty under their calculation methods. They do not account automatically for every potential limitation, such as missing outcome data, analysis choices, or selective reporting. Details about those issues are not reported in the registry record.

11. Planning a Similar Trial: Sample Size

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

The sample-size table and its associated curve are planning illustrations based on the observed effects, not the trial's original design assumptions. The figures are participants per arm for these continuous-outcome comparisons; they are not total event counts. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each row uses the stated fraction of the observed effect as the effect to detect.

The 80% power column describes the planned number of participants per arm that, under those assumptions and for that target effect, would give an 80% chance of detecting the effect using the specified test framework. The 90% power column uses a higher target chance of detection and consequently calls for a larger planned sample. The rows show planning for 100%, 75%, and 50% of the observed effect. As the target effect gets smaller, more participants are needed to distinguish it from random variation with the same planned power.

The illustrations are especially sensitive to the observed effect used as their starting point. For the slow-versus-standard comparison, the table gives 672 participants per arm for 80% power and 899 for 90% power when planning for 100% of the observed effect. For 75% of the observed effect, it gives 1,194 and 1,598; for 50%, it gives 2,685 and 3,594. For the booster-versus-standard comparison, the corresponding entries are 18,685 and 25,013 for 100% of the observed effect, 33,217 and 44,468 for 75%, and 74,738 and 100,052 for 50%.

These calculations are not observed power. They do not show the probability that the completed study succeeded, nor do they replace a confidence interval or analysis of the trial results. The original trial's sample-size justification and design assumptions are not reported in the registry record. The curve should be read as another view of the same planning relationship, not as additional trial evidence.

12. How the Sample Size Is Derived

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

Formula

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

Worked calculation for this trial

1. Observed difference Δ = 9.000; pooled standard deviation σ = 58.847.
2. n = 2 × (1.96 + 0.842)² × 58.847² / 9.000² = 671.1, rounded up to 672 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 <- 9                      # observed difference in means
sigma <- 58.8469392929738                     # 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(1060111)
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 = 672 (80% power), 899 (90% power)
 75% of observed effect: n per arm = 1,194 (80% power), 1,598 (90% power)
 50% of observed effect: n per arm = 2,685 (80% power), 3,594 (90% power)
power.t.test: 672.1 
Simulated power at n = 672 per arm: 0.801 (20000 replicates)

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

13. Analyses Reported by the Sponsor

The sponsor reported two analyses relevant to the primary endpoint. The first was an ANOVA across the standard, slow, and booster groups, with p = 0.621. The second was a one-sided t-test comparing standard topiramate with the booster arm. That analysis reported a ratio of percentage decrease of 0.950, a 95% confidence interval from 0.519 to 1.737, and p = 0.5, and was classified in the registry as non-inferiority or equivalence.

The pairwise re-calculations use posted means, standard deviations, and evaluated group counts to produce unadjusted mean differences and confidence intervals. They use an additive scale, whereas the sponsor's second analysis uses a ratio. The pairwise re-calculations are therefore not reproductions of the sponsor's non-inferiority analysis. The sponsor's own analysis should be interpreted in its stated framework, but the registry does not report the non-inferiority margin, full hypothesis specification, or sufficient method detail to determine the precise decision rule.

The re-calculated estimates offer a transparent comparison of the posted summary statistics; they do not override the sponsor's analysis. Neither set of results should be interpreted without its uncertainty and the limits of the registry reporting. The registry does not report a sponsor analysis for the slow-versus-standard pairwise contrast that can be directly matched to the simple re-calculation.

14. Multiplicity, Missing Data and Analysis Populations

The record lists four outcome measures and five analyses. The primary endpoint is percentage decrease in migraine episodes, while the listed secondary outcomes are change from baseline in migraine frequency, MIDAS score, and VAS score at Week 6. The posted results include a three-group ANOVA and a sponsor-reported one-sided t-test, alongside the pairwise calculations based on posted summary statistics.

When a study evaluates several outcomes or comparisons, the chance of at least one apparently unusual result can increase if each is interpreted without regard to the others. General statistical approaches to this issue include pre-specifying a primary analysis, adjusting significance thresholds, or treating secondary results as exploratory. The registry record does not report a multiplicity adjustment or a detailed plan for controlling multiple comparisons. It also does not report whether the five analyses were pre-specified, how the secondary outcomes were analyzed, or how the reported tests relate to all planned comparisons. Accordingly, the individual p-values should not be treated as if they necessarily represent a single, fully documented testing strategy.

The primary population is described as intent-to-treat participants who took topiramate at least once and had migraine improvement data at Week 6. That description is more restricted than an unrestricted all-randomized population, but the record does not explain the population construction or provide an account of participants without the required data. Although the primary endpoint is described as covering Weeks 7 to 10, the eligibility detail refers to Week 6; the handling of that timing difference is not reported in the registry record.

The record does not report the amount or pattern of missing primary-outcome data, imputation methods, sensitivity analyses, or reasons for missingness. It does not provide a full statistical analysis plan, interim-analysis information, or the original sample-size justification. These absences limit assessment of whether the analysis choices were appropriate and how robust the results are to missing data or multiple testing. They do not establish that any particular method was or was not used; they mean the information is not reported in the registry record.

15. Safety Results

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

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 2% reporting threshold.

Most frequent non-serious adverse event terms

The registry reports serious adverse events for the three groups. For Topiramate Standard, 0 serious adverse events were reported among 81 participants at risk. For Topiramate Slow, 0 were reported among 83 participants at risk. For Topiramate Slow and Propranolol Booster, 0 were reported among 79 participants at risk. These are the posted serious-event counts and denominators; they are not the same counts as the primary-outcome evaluated group sizes.

A zero count in a trial group means no serious adverse events were recorded in that group during the period represented by the registry entry. It does not establish that the underlying risk is zero. Event counts can be informative about what was observed in the study, but their interpretation depends on the amount of observation and the completeness of reporting. The registry does not give serious-event terms, and the deaths field is not reported in the registry record.

The record also lists other, non-serious adverse events affecting 62, 62, and 54 participants in the standard, slow, and booster groups, respectively. Those figures are separate from the serious-event counts. This page reports the registry's safety summaries without drawing a treatment recommendation or comparing safety in a way that the available reporting cannot support.

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

The posted primary-outcome means are 48.0 for standard topiramate, 57.0 for slow titration, and 46.0 for slow titration with a propranolol booster. The group standard deviations are 69.0, 48.0, and 69.0. The means are descriptive summaries for participants evaluated for the outcome; they are not, on their own, proof that the approaches differ. The pairwise re-calculated confidence intervals are broad and include zero, indicating that these data do not precisely determine the direction or size of the mean differences.

The sponsor's three-group ANOVA reported p = 0.621. The sponsor's one-sided comparison of standard topiramate with the booster arm reported a ratio of 0.950 and a 95% confidence interval from 0.519 to 1.737. The registry calls the latter analysis non-inferiority or equivalence, but does not report the margin or complete decision criteria. It would therefore be inappropriate to treat the estimate or p-value alone as proof of non-inferiority or equivalence.

These results do not answer every question about the treatment schedules. They address the posted migraine-frequency endpoint and the analyses described in the record. The registry lists secondary measures of migraine frequency, disability, and VAS score at Week 6, but the results and analysis details for those measures are not included in the registry record. The available record also does not explain the relationship between the endpoint's Weeks 7 to 10 timeframe and the Week 6 data requirement for the primary population.

In general, a non-significant test is not evidence that groups are identical, and a numerical difference is not itself evidence of a reliable effect. Here, the point estimates and intervals communicate both the observed pattern and considerable uncertainty. Missing details about multiplicity, missing data, analysis methods, and the non-inferiority margin further limit the strength of conclusions that can be drawn from the registry record. These findings are statistical descriptions, not medical advice.

17. Why This Trial Matters Statistically

This trial illustrates why group means and p-values should not be read in isolation. The three posted means are relatively close compared with the reported within-group standard deviations, and the simple pairwise confidence intervals span zero as well as differences in either direction. Looking at the interval alongside the estimate helps distinguish the observed result from the range of effects that remains compatible with the data.

It also highlights the difference between a global comparison and a pairwise one. The sponsor's ANOVA considers all three groups together, while the re-calculations compare one treatment group with the standard arm at a time. The sponsor's separate ratio analysis uses a relative scale and is labeled as non-inferiority or equivalence. Those analyses answer different statistical questions and should not be conflated.

The sample-size illustrations show how planning requirements depend on the assumed effect to detect. In general, a smaller target effect requires a larger sample for the same planned power. Because these illustrations use observed effects, they are not the original sample-size rationale and cannot establish the trial's planning quality. The original assumptions are not reported in the registry record.

Finally, this record makes visible the importance of complete analysis reporting. The endpoint timeframe and population timing are stated differently, and details about missing data, multiplicity, and the sponsor's non-inferiority margin are absent. These are not merely technical footnotes: they shape how confidently a reader can connect a numerical result to the question the trial intended to answer.

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 specific to this registry entry affect interpretation. First, the primary endpoint is assigned to the maintenance period, Weeks 7 to 10, but the primary population description requires migraine improvement data at Week 6. The record does not explain this timing relationship. The primary outcome's definition and the population used to evaluate it therefore cannot be fully reconciled from the registry text alone.

Second, the propranolol booster arm description is truncated after “Propranolol 80”. The complete dose and schedule are not reported in the registry record. The registry also says the topiramate maintenance dose was at physician discretion, which means the described range does not specify the dose each participant received.

Third, the posted primary outcome is based on participants evaluated, with counts of 51, 57, and 43 across the groups. The registry enrollment is 250, but the record does not provide a full flow of participants from randomization through outcome evaluation. Reasons for unavailable outcomes, missing-data procedures, and sensitivity analyses are not reported in the registry record.

Fourth, the sponsor's non-inferiority or equivalence analysis has a reported ratio estimate and confidence interval, but its margin and full decision rules are not reported. The registry does not give enough detail to reconstruct that analysis. Similarly, a detailed statistical analysis plan, multiplicity handling, exact randomization procedures, masking procedures, and the original sample-size rationale are not reported in the registry record.

Finally, the secondary outcomes are listed, but their results are not included. The record also reports no serious adverse events in the listed risk groups, while deaths are not reported in the registry record. These limitations restrict interpretation of the available summary; they do not establish what unreported analyses or procedures may have been performed.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: An Efficacy and Tolerability Study of Topiramate in Participants With Migraine — NCT01060111.
• Registry record last updated 2013-08-22; status last verified 2013-08.
• Results first posted 2013-08-06.
• 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

This completed Phase 4 study randomized participants with migraine to standard topiramate titration, slower topiramate titration, or slower titration with a propranolol booster. It was listed as single-masked and parallel-group, with enrollment of 250. The primary outcome was percentage decrease in migraine episodes, described for the maintenance period, Weeks 7 to 10.

The posted means were 48.0, 57.0, and 46.0 across the three groups. The unadjusted pairwise re-calculations have broad confidence intervals that include zero. The sponsor reported a three-group ANOVA with p = 0.621 and a separate one-sided ratio analysis for standard topiramate versus the booster arm. The latter's non-inferiority margin is not reported in the registry record.

Interpretation is constrained by incomplete reporting of the endpoint timing and analysis population, missing-data and multiplicity methods, and the full booster regimen. The sample-size figures are planning illustrations based on observed effects, not observed power or the trial's original design assumptions.