Topiramate 25 mg Capsule Mixed With Applesauce Under Fasting Conditions: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Topiramate 25 mg Capsule Mixed With Applesauce Under Fasting Conditions (NCT00905346), a randomized phase 1 study. The primary outcome is Cmax - Maximum Observed Concentration.

2. Clinical Question

3. Trial Design

This was a randomized, two-way crossover study in healthy subjects, registered as phase 1 and completed. The registry record lists enrollment of 24. The study dates are 2002-06 for the start and 2002-07 for primary completion. The sponsor was Teva Pharmaceuticals USA.

In a crossover design, participants receive each study treatment in a planned sequence, so comparisons can be made within participants. The registry record identifies two sequences: Test First and Reference First. It describes the test as topiramate capsules and the reference as Topamax® capsules. Both products were administered as capsule contents mixed with applesauce under fasting conditions, as described in the study objective. The registry record does not provide further details of the sequence schedule, timing between treatments, or other procedural features.

Masking is listed as none. The primary outcome description states that blood samples were collected over a 168 hour period and that data from all subjects who completed the study were included in the statistical analysis. The posted results report 21 subjects for each treatment arm. They do not explain the difference between the overall enrollment and the number included in the primary analysis.

The design was intended to compare the rate and extent of absorption. The primary outcome listed is Cmax, while the registry record lists three outcome measures and three analyses. Additional details about those other outcomes and analyses are not reported in the registry record. The sponsor’s original sample-size assumptions and other design-stage statistical details are also 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.

Period 1

Reasons for not completing this period:

Period 2

Analysis population for the primary outcome, as defined in the registry: Data from all subjects who completed the study was included in the statistical analysis.

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 Cmax - Maximum Observed Concentration, measured over Blood samples collected over 168 hour period.

Cmax is the maximum observed concentration of a drug in blood after administration. It describes the peak concentration captured among the collected samples, rather than the total amount absorbed over time. In this study, the registry record identifies Cmax as the primary endpoint and gives its unit as ng/mL. Blood samples were collected over a 168 hour period.

Cmax is used in pharmacokinetic comparisons because the peak concentration is one aspect of the concentration-time profile. The study objective also refers to comparing the rate and extent of absorption, but the posted primary endpoint specifically concerns the maximum observed concentration. Cmax should therefore be interpreted as one measure of absorption, not as a complete description of every feature of the concentration-time profile.

The posted arm results give mean Cmax values and standard deviations for the test and reference products. The sponsor’s analysis reports the ratio of least-squares means for test versus reference, expressed as a percentage, with a 90% confidence interval. The registry record does not state the equivalence limits used to interpret that interval. Consequently, the numerical estimate and its interval can be described, but the registry record alone does not establish whether a prespecified bioequivalence criterion was met.

7. Results Posted in the Registry

• Topiramate (Test): mean 630.45 versus 631.49 on control; difference -1.04 (95% CI -99.24 to 97.16), standardized effect d = -0.01; z-test p = 0.983.

8. Statistical Interpretation

9. Statistical Methods Explained

The registry record describes a randomized, two-way crossover study. In a crossover design, each participant receives the study treatments according to a sequence, allowing treatment comparisons to use information from the same participant. This can be useful when participant-to-participant variability is substantial, because within-participant comparisons can account for stable differences between people. A sound analysis must reflect the treatment sequence and the repeated observations within participants. The registry record does not report the sponsor’s statistical model or details of how those features were handled.

What does the reported ratio represent?

The sponsor’s result is a ratio of least-squares means for test versus reference, multiplied by 100. A ratio expresses the test result relative to the reference on a proportional scale. The reported estimate is 98.39, and the reported 90% confidence interval is 94.36 to 102.59. Least-squares means are model-based estimates; their interpretation depends on the model and analysis specification. The registry record does not identify the method used to calculate them, so the model cannot be described in more detail.

The ratio is not the same quantity as a difference between arithmetic means. A ratio compares relative levels, while a difference expresses an absolute gap in the endpoint’s units. Which scale is appropriate depends on the prespecified analysis and scientific question. For a bioequivalence assessment, ratio-based estimates are commonly interpreted against prespecified limits, but the applicable limits for this study are not reported in the registry record. The interval alone therefore cannot establish whether the study met a formal equivalence criterion.

What does the recalculation test?

The unadjusted recalculation uses posted arm summary statistics to estimate a difference in mean Cmax. It reports a mean difference of -1.04, a 95% confidence interval from -99.24 to 97.16, and an unadjusted p-value of 0.983. This calculation is deliberately distinct from the sponsor’s least-squares-mean ratio. It does not reconstruct the crossover model from participant-level observations, and the registry record does not include the participant-level data needed to do so.

A p-value summarizes how compatible the observed result is with a specified null hypothesis under the calculation’s assumptions. A large p-value is not proof that treatments are equivalent, and it does not quantify the probability that the null hypothesis is true. Equivalence requires a prespecified range of differences or ratios considered acceptable, together with an analysis showing sufficient precision relative to that range. The registry record does not report such a range.

Why can crossover summaries and an unadjusted comparison differ in meaning?

The posted means and standard deviations are marginal summaries by treatment. They do not show the correlation between each participant’s measurements under the two treatments, nor do they show sequence, period, or participant-level data. In a crossover trial, those within-participant relationships may matter to the treatment estimate and its uncertainty. The unadjusted recalculation is therefore a simple descriptive re-analysis of posted summaries, not a replacement for the sponsor’s analysis.

Standard deviation describes the spread of individual observations around a group mean. It is not the same as the uncertainty around the estimated treatment effect. The standard errors listed in the recalculation are derived for the unadjusted summary-statistic comparison and likewise should not be interpreted as the sponsor model’s uncertainty estimates. The registry record does not report the sponsor’s standard errors or the details needed to verify them.

What can be said about assumptions?

For the planning illustration, the stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is specified as a fraction of the observed effect. These are assumptions for the planning illustration, not documented assumptions for the original trial. The original sample-size justification and its assumptions are not reported in the registry record.

More generally, statistical conclusions depend on the analysis population, missing-data handling, the model, and the endpoint definition. The primary outcome description says that all subjects who completed the study were included in the statistical analysis. It does not state how incomplete observations were handled, whether any participants were excluded from particular analyses, or how missing data were treated. Those details are not reported in the registry record. The record also does not report interim analyses, multiplicity procedures, or a detailed statistical analysis plan. These omissions limit independent assessment of the analysis, even though the posted estimates and intervals can still be described as recorded.

10. Confidence Intervals and Precision

A confidence interval describes a range of values compatible with the data and the assumptions of the analysis. It conveys uncertainty around an estimate; it is not a guarantee that the true value lies inside the interval, nor does it show the probability that the estimate is correct. A narrower interval generally indicates greater precision for the parameter on the stated scale.

For the unadjusted difference in mean Cmax, the recalculated estimate is -1.04, with a 95% confidence interval from -99.24 to 97.16. This interval includes differences on both sides of zero and extends across a broad range of possible differences. It therefore does not pin down the size or direction of a small difference with high precision. The interval pertains to the unadjusted summary-statistic calculation, not to the sponsor’s crossover model.

The sponsor-reported ratio of least-squares means is 98.39, with a 90% confidence interval from 94.36 to 102.59. This interval includes values below and above 100, the reference point for equal estimated levels on a ratio scale. Whether the interval is sufficiently narrow for a formal bioequivalence conclusion depends on prespecified acceptance limits. Those limits are not reported in the registry record, so precision relative to the formal criterion cannot be evaluated from the record alone.

The sponsor interval and the recalculated interval use different scales, confidence levels, and analyses; they should not be compared as if they were interchangeable. The registry record gives no detailed sponsor method, and the posted arm summaries do not include the within-participant information required to reconstruct the crossover analysis. The intervals describe uncertainty under their respective analyses, while the missing design and analysis details constrain a fuller assessment.

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 curve are planning illustrations based on the observed effect, not the trial’s original design. Their figures are participants per arm, rather than total events. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; the effect to detect is the stated fraction of the observed effect. The original trial’s sample-size assumptions are not reported in the registry record.

The 80% power column gives the illustrated number of participants per arm for a design with an 80% chance of detecting the specified effect under the planning assumptions, if that effect and the assumptions were true. The 90% power column gives the corresponding planning figure for a 90% chance. These are prospective planning quantities, not measures of the probability that the completed study found a real effect.

The rows distinguish planning for 100%, 75%, and 50% of the observed effect. The listed figures increase as the effect used for planning becomes smaller. This reflects a general statistical principle: when a target effect is smaller relative to variability, more participants are needed to distinguish it from random variation at a chosen power level. The figures are not a recommendation for this crossover design; the illustration specifies equal allocation between two arms and does not document the original trial’s design assumptions or analysis model. This is not observed power.

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 = -1.040; pooled standard deviation σ = 162.360.
2. n = 2 × (1.96 + 0.842)² × 162.360² / -1.040² = 382,584.4, rounded up to 382,585 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 <- -1.03999999999996                      # observed difference in means
sigma <- 162.359667559403                     # 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(905346)
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 = 382,585 (80% power), 512,173 (90% power)
 75% of observed effect: n per arm = 680,151 (80% power), 910,529 (90% power)
 50% of observed effect: n per arm = 1,530,338 (80% power), 2,048,689 (90% power)
power.t.test: 382585.4 
Simulated power at n = 382,585 per arm: 0.799 (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 result is a ratio of least-squares means for test versus reference of 98.39, with a 90% confidence interval from 94.36 to 102.59. It is labeled in the registry record as a legacy non-inferiority or equivalence analysis. The method and the equivalence limits are not reported in the registry record, and the sponsor result has no posted p-value.

The recalculation is a separate, unadjusted difference in means from posted summary statistics. It gives a mean difference of -1.04, a 95% confidence interval from -99.24 to 97.16, and an unadjusted p-value of 0.983. Because the scales, confidence levels, and analytical approaches differ, this recalculation is not a direct replication or correction of the sponsor result. In particular, the summary statistics do not capture the participant-level pairing that a crossover analysis may use.

The discrepancy indicator is false, but that does not make the two analyses identical. The sponsor’s own analysis takes precedence for describing the sponsor-reported conclusion. Its reported estimate and interval can be stated as recorded; the missing method and acceptance limits prevent an independent determination from the registry record about the formal equivalence criterion.

14. Multiplicity, Missing Data and Analysis Populations

The registry record lists three outcome measures and three analyses, while the primary endpoint is Cmax. It does not describe a multiplicity adjustment, a hierarchy among endpoints, or how the overall chance of false-positive findings was controlled. Those procedures are not reported in the registry record. The presence of multiple listed outcomes does not, by itself, establish that an adjustment was required or omitted; the planned analysis strategy would be needed to assess that question.

The primary outcome description says that data from all subjects who completed the study were included in the statistical analysis. The posted arm results each report 21 subjects, while overall enrollment is 24. The registry record does not explain why the analyzed number differs from enrollment, give reasons for non-completion, or describe the handling of missing measurements. It also does not state whether the analysis population was defined in other ways for the sponsor’s ratio analysis.

The record does not report interim analyses, a detailed statistical analysis plan, or the sponsor’s model and its assumptions. It does not give participant-level data or information about the correlation between repeated measurements in the crossover. These absences limit assessment of the sponsor’s uncertainty calculation and the handling of sequence or period effects. They do not erase the posted results, but they mean the registry record is insufficient to reproduce the sponsor’s analysis independently.

The unadjusted recalculation is based on arm-level summary statistics and should not be treated as though it resolves these omissions. The available results support description of the estimates and their intervals; assessment of missing-data choices, multiplicity procedures, and analysis-population decisions requires information that is not reported in the registry record.

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

The results describe a pharmacokinetic comparison of topiramate test capsules and Topamax® reference capsules in healthy subjects, with both products administered as contents mixed with applesauce under fasting conditions. The primary endpoint, Cmax, concerns maximum observed blood concentration. The posted means are 630.5 ng/mL for the test and 631.5 ng/mL for the reference. The sponsor’s ratio estimate is 98.39, with a 90% confidence interval from 94.36 to 102.59.

These figures address the measured endpoint in this study setting. They do not establish effects on symptoms, clinical outcomes, safety, or other populations; such conclusions are not evaluated by the reported primary endpoint. They also do not establish that either product is better or worse. No medical recommendation follows from these results.

The sponsor’s interval may be considered in relation to a formal bioequivalence criterion only if the prespecified acceptance limits and analysis method are known. Those details are not reported in the registry record. The unadjusted recalculation is not a substitute for the sponsor’s analysis, and its p-value of 0.983 is not evidence of equivalence. A lack of a statistically detectable difference and a demonstration that products fall within an acceptable equivalence range are different conclusions.

Interpretation is further limited by the absence of details about the original sample-size justification, missing-data handling, multiplicity, and sponsor model. The posted data therefore permit a careful account of the measured concentrations and reported estimates, but not a complete reconstruction of the planned statistical argument or an independent verification of a formal bioequivalence conclusion.

16. Why This Trial Matters Statistically

This study illustrates how the design and estimand shape interpretation. In a crossover study, the same participants receive both treatments in different sequences. That design can make within-participant comparisons informative, but it also means that a simple comparison of treatment-arm means does not necessarily reproduce the planned analysis. The posted summary statistics support a basic recalculation, while the sponsor’s least-squares-mean ratio represents a separate analysis.

It also demonstrates why equivalence questions require more than a non-significant difference test. The unadjusted p-value is 0.983, but that value does not show that the treatments are equivalent. The sponsor’s ratio and confidence interval are more directly connected to a ratio-based equivalence assessment, yet the acceptance limits and detailed method are not reported in the registry record. Without them, the boundary between an interval compatible with equivalence and one that is not cannot be assessed.

Finally, the record distinguishes observed results from planning calculations. The sample-size illustration uses fractions of the observed effect under stated assumptions, but it does not reveal the original design rationale and is not observed power. Together, these features make the study a useful example of why outcome scale, trial structure, precision, analysis specification, and transparent reporting must be considered together when reading clinical-trial statistics.

17. 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.

The registry record reports enrollment of 24 and 21 subjects in each posted arm result, but does not explain the difference or provide reasons for non-completion. The primary analysis description includes subjects who completed the study; it does not explain how missing samples or incomplete outcome measurements were handled. The registry record also does not provide participant-level data, which prevents independent reconstruction of the crossover comparison.

Although the record identifies the study as randomized and crossover, it does not report details of the sponsor’s statistical model, sequence and period handling, or other analysis assumptions. The sponsor’s result is a ratio of least-squares means with a 90% confidence interval, but the method and formal bioequivalence limits are not reported. This prevents a registry-only assessment of whether the reported interval met the study’s intended criterion.

The record lists three outcome measures and three analyses, but does not describe a multiplicity strategy or identify the details of the additional analyses. It also does not report interim analyses or the original sample-size justification. The planning illustration is not a substitute for those design details, since it is based on the observed effect and stated planning assumptions rather than the trial’s original rationale.

Finally, this was a study in healthy subjects under fasting conditions, using capsule contents mixed with applesauce. The posted primary measure is Cmax, not a clinical outcome. Conclusions should remain within that population, administration setting, and measured endpoint. The registry record does not provide further information to extend the interpretation beyond them.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Topiramate 25 mg Capsule Mixed With Applesauce Under Fasting Conditions — NCT00905346.
• Registry record last updated 2024-08-20; status last verified 2024-08.
• Results first posted 2009-09-23.
• 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.

21. Record Summary

NCT00905346 is a completed phase 1 randomized, two-way crossover study sponsored by Teva Pharmaceuticals USA. It enrolled 24 healthy subjects and compared topiramate 25 mg capsules with Topamax® 25 mg capsules, administered as capsule contents mixed with applesauce under fasting conditions. The primary endpoint was Cmax, with blood samples collected over a 168 hour period.

The posted primary results report 21 subjects for each product. The sponsor-reported test/reference ratio of least-squares means is 98.39, with a 90% confidence interval from 94.36 to 102.59. An unadjusted recalculation reports a mean difference of -1.04, a 95% confidence interval from -99.24 to 97.16, and an unadjusted p-value of 0.983. The sponsor’s method and equivalence limits are not reported in the registry record, so the formal bioequivalence criterion cannot be independently assessed from this record alone.