Levetiracetam 1000 mg Under Non-Fasting Conditions: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Levetiracetam 1000 mg Under Non-Fasting Conditions (NCT00859430), a randomized phase 1 study. The primary outcome is Cmax - Maximum Observed Concentration.

2. Clinical Question

The randomized crossover arrangement means the study compared two treatment periods and two treatment sequences: test followed by reference, or reference followed by test. Both products were studied under fed conditions. The registry record identifies the study as unmasked and lists no masking. The primary analysis population is described as all subjects who completed the study, but the record does not report further details about the analysis population or the handling of incomplete observations.

The posted endpoint is a concentration measure, not a clinical outcome. The question addressed by these results is therefore a pharmacokinetic comparison of the recorded peak concentration after dosing. The registry record does not report a decision rule or equivalence bounds for interpreting the primary endpoint. The sponsor’s posted analysis reports a geometric mean ratio and confidence interval; an unadjusted recalculation from the treatment summaries describes a mean difference instead. Those are different summaries and should be read as such.

3. Trial Design

This completed phase 1 study was randomized, unmasked, and used a two-period, two-treatment, two-way crossover model. It enrolled 22 healthy participants. The study start and primary completion are both recorded as 2007-01. The sponsor was Teva Pharmaceuticals USA.

Each sequence included both products, with the order reversed between sequences. Participants assigned to the test-first sequence received Levetiracetam 1000 mg Tablet in the first period and Keppra® 1000 mg Tablet in the second. Participants in the other sequence received Keppra® 1000 mg Tablet first, followed by the test tablet. The study objective and official title describe dosing under fed conditions. The design therefore compares the two formulations within a two-period arrangement rather than assigning each treatment to a wholly separate parallel group.

The primary outcome was Cmax, or maximum observed concentration, measured in mcg/mL. Blood samples were collected over a 36 hour period. The primary analysis description says that data from all subjects who completed the study were included. The treatment-group summaries each report n=21, but the registry record does not explain the relationship between those summary counts and enrollment of 22, or give further participant-flow details.

The record does not report the washout duration, the method used to generate the random sequence, or further details of the analysis model. It also does not report the original design assumptions or a sample-size justification. Those elements cannot be inferred from the crossover label alone. The sequence descriptions establish the treatment order, but do not document whether period or sequence effects were assessed or how such effects were handled in the sponsor’s analysis.

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.

First Intervention

Reasons for not completing this period:

Second Intervention

Analysis population for the primary outcome, as defined in the registry: Data from all subjects who completed the study were 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 36 hour period.

The primary endpoint is Cmax, defined as maximum observed concentration. It records the highest concentration measured in the blood-sampling schedule after dosing. The registry gives mcg/mL as the unit and describes sampling over a 36 hour period. The stated primary-outcome description is “Bioequivalence based on Cmax.”

As a general pharmacokinetic concept, the peak concentration is used to characterize the concentration reached after a dose and can contribute information about the rate of absorption. The registry’s broader study objective also refers to comparing the rate and extent of absorption. Cmax is one concentration-based measure for that comparison; it is not itself a direct measure of a patient’s symptoms or clinical benefit.

The treatment summaries report means and standard deviations for the two formulations. The test formulation’s reported mean Cmax is 26.32 mcg/mL, with standard deviation 5.06; the Keppra® summary is 24.93 mcg/mL, with standard deviation 4.427. These summaries describe the observed endpoint values, while the sponsor’s reported comparison is expressed as a ratio of geometric means. The registry record does not explain the choice of transformation or give the detailed analysis model.

Interpreting a pharmacokinetic endpoint depends on the comparison method and on any prespecified criterion used to judge equivalence. The record does not report equivalence bounds or a decision rule for Cmax. The confidence interval can describe uncertainty around the sponsor’s reported ratio, but without stated bounds it does not establish whether a prespecified criterion was met.

7. Results Posted in the Registry

• Keppra®: mean 24.93 versus 26.32 on control; difference -1.39 (95% CI -4.27 to 1.48), standardized effect d = -0.29; z-test p = 0.342.

8. Statistical Interpretation

9. Statistical Methods Explained

The registry identifies Cmax as a continuous outcome and reports means with standard deviations for the treatment groups. A continuous measure can take values across a scale rather than being a simple event count. The mean describes the center of the observed values, while the standard deviation describes their spread around that mean. In these results, both are reported in mcg/mL.

What does the posted mean difference describe?

The unadjusted recalculation compares the treatment summaries as a mean difference. Its estimate is -1.393, with Keppra® as the comparison treatment relative to Levetiracetam, and the interval runs from -4.269 to 1.483. A difference is expressed in the original measurement scale, so it describes a contrast in Cmax rather than a ratio. The direction of the estimate depends on which treatment is subtracted from which; the stated orientation is essential to reading its sign.

This recalculation is labelled unadjusted and is based on posted summary statistics. It is not a reconstruction of the sponsor’s full analysis. The record identifies a crossover design, but does not report the sponsor’s statistical model, whether a paired analysis was used, or how period and sequence were handled. General statistical practice for a crossover may account for within-participant comparisons, but the registry record does not establish which such approach was used. The recalculated comparison should therefore be read as a simple descriptive statistical contrast, not as evidence that it reproduces the sponsor’s inferential procedure.

Why is the sponsor’s result on another scale?

The sponsor’s reported analysis is a ratio of geometric means, with a value of 105 and a 90% confidence interval from 98.6 to 112. A geometric mean is generally used for measurements analyzed on a logarithmic scale, where treatment contrasts are expressed as ratios after returning to the original scale. This is a general explanation; the registry record does not state the sponsor’s transformation or model. A ratio and an arithmetic mean difference are not interchangeable. One describes relative scale and the other absolute scale, and they do not share the same interpretation or uncertainty interval.

The record gives no p-value for the reported geometric-mean-ratio analysis and no method name. It also does not state the equivalence limits. A ratio interval can only be assessed against a decision criterion when that criterion is known. Since the registry does not report those limits, the result can be described without declaring whether it satisfies a formal equivalence rule.

What does a confidence interval contribute?

In general, a confidence interval gives a range of estimates compatible with the observed data under the analysis assumptions. It communicates both the direction and the precision of an estimate. For the recalculated mean difference, the interval extends from a negative to a positive value and includes zero. For the sponsor’s ratio, the interval is reported on the ratio scale. These intervals cannot be compared as if they measured the same contrast. Their different scales and reported confidence levels matter.

A confidence interval is not a statement that the true value has a particular probability of lying inside this one calculated interval. Rather, the confidence level describes the long-run behavior of the interval procedure under its assumptions. Those assumptions, including the exact crossover model for the sponsor’s analysis, are not fully described in the registry record. The intervals therefore aid interpretation but do not remove uncertainty about model choice or analysis details.

What does the p-value say?

The unadjusted p-value is 0.342. In general, a p-value measures how unusual data at least as discrepant as those observed would be under a specified null hypothesis and analysis model. It is not the probability that the null hypothesis is true, nor does it quantify the probability of bioequivalence. The registry does not specify the null hypothesis or test procedure behind this recalculated p-value beyond its unadjusted designation. It should be interpreted alongside the estimate and interval, not as a stand-alone conclusion.

For an equivalence question, a conventional test of a zero difference is not by itself the full decision process. Equivalence methods ordinarily require prespecified acceptable bounds and an analysis that evaluates the estimate and its interval against those bounds. This is general statistical context, not a description of a protocol that can be confirmed from this record. The registry’s lack of reported bounds means the formal equivalence criterion cannot be reconstructed.

What is and is not documented?

The registry reports randomized allocation, no masking, a crossover model, treatment sequences, the primary endpoint, and a completed-subject population description. It does not report a detailed statistical analysis plan, handling of missing observations, an interim-analysis plan, or multiplicity procedures. It also does not report the original sample-size assumptions. The simple recalculation and the planning illustrations are distinct from those unreported protocol details. They cannot establish what was planned before the study or how every analysis decision was made.

10. Confidence Intervals and Precision

The recalculated mean difference is -1.393, and its confidence interval extends from -4.269 to 1.483. The interval includes zero, so the posted summary statistics are compatible with a range of mean differences in either direction under this unadjusted calculation. The interval also conveys that the estimate is not exact: values across that range remain consistent with the stated uncertainty procedure. The unadjusted p-value is 0.342 and should be read alongside the estimate and interval, not as a substitute for them.

The sponsor’s analysis reports a ratio of geometric means of 105, with a 90% confidence interval from 98.6 to 112. This is a separate precision statement on a ratio scale. It cannot be merged with the mean-difference interval or read as though the same estimand were being estimated. The registry record does not report equivalence bounds, so the ratio interval cannot be judged against a formal acceptance range from the record alone.

As a general statistical principle, greater precision is reflected by a narrower interval when analyses and scales are comparable. Interval width also depends on variability, the amount of information, and the analysis model. Here, the posted summaries show standard deviations of 5.06 for the test tablet and 4.427 for Keppra®, while the exact sponsor model is not reported. The sample-size illustrations are planning calculations, not a direct measure of precision achieved in this trial. The available intervals are the clearest reported indication of uncertainty in the two different comparisons.

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 planning figures are participants per arm, not total events. They are illustrations computed from the observed effect, rather than the trial’s original design justification. The registry record does not report the original sample-size assumptions. The planning calculation assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to a stated fraction of the observed effect.

For planning at 100% of the observed effect, the illustration gives 183 participants per arm for 80% power and 245 participants per arm for 90% power. For planning at 75% of the observed effect, it gives 326 participants per arm for 80% power and 436 participants per arm for 90% power. For planning at 50% of the observed effect, it gives 732 participants per arm for 80% power and 980 participants per arm for 90% power.

In plain terms, the 80% power column describes the illustrated per-arm size associated with an 80% chance of detecting the specified effect under the planning assumptions, if that effect and those assumptions held. The 90% column uses a higher target chance of detection under the same stated framework. Smaller effects are harder to distinguish from random variation, so the illustration requires more participants per arm as the effect fraction becomes smaller. These values are not observed power and do not describe the probability that this completed trial found a real effect.

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 = -1.393; pooled standard deviation σ = 4.754.
2. n = 2 × (1.96 + 0.842)² × 4.754² / -1.393² = 182.8, rounded up to 183 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 <- -1.393                      # observed difference in means
sigma <- 4.75404717056951                     # 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(859430)
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 = 183 (80% power), 245 (90% power)
 75% of observed effect: n per arm = 326 (80% power), 436 (90% power)
 50% of observed effect: n per arm = 732 (80% power), 980 (90% power)
power.t.test: 183.8 
Simulated power at n = 183 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’s posted analysis reports a ratio of geometric means of 105, with a 90% confidence interval from 98.6 to 112. The separate recalculation uses the posted treatment means and standard deviations to give an unadjusted mean difference of -1.393, a confidence interval from -4.269 to 1.483, and an unadjusted p-value of 0.342. These results use different scales: a ratio of geometric means versus an absolute difference in means.

The difference in scale means the recalculation is not a replacement for the sponsor’s analysis and cannot be used to reproduce its result. The record does not report the sponsor’s analysis method, the model details, or the equivalence bounds. It also does not provide a p-value for the sponsor’s ratio analysis. The posted ratio and its interval should therefore be described as the sponsor’s reported comparison, while the arithmetic mean difference is an additional unadjusted summary-statistic calculation.

No discrepancy is flagged between the posted summaries and the reported analysis in the registry record. Even so, their distinct estimands and methods require separate interpretation. A determination against formal bioequivalence criteria cannot be made from the ratio alone because the registry record does not state the applicable bounds.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists three outcome measures and three analyses, but reports no additional secondary results. The posted primary endpoint is Cmax. The record does not explain the relationship among the listed outcome measures and analyses, identify all planned comparisons, or describe a prespecified hierarchy among outcomes.

Multiplicity arises when several hypotheses or outcomes are tested: the chance of at least one apparently unusual result can increase as the number of tests grows. This is a general statistical issue, not evidence that a particular correction was or was not applied in this study. The registry record does not report a multiplicity adjustment, a strategy for controlling the overall error rate, or a reason to treat the analyses as independent or dependent. It is therefore not possible to determine from the record whether multiplicity was planned for or addressed.

The primary analysis population is described as data from all subjects who completed the study. The treatment summaries each report n=21, while enrollment is 22. The registry record does not explain this difference, report a participant-flow account, or state how any incomplete observations were handled. It does not describe an imputation method, sensitivity analysis, or missing-data assumptions. No conclusion about the reason for the different counts should be inferred from the posted numbers.

The record also does not report whether the sponsor’s analysis used a per-protocol, intention-to-treat, or other named population, beyond the completed-subject description. Nor does it provide a detailed statistical analysis plan or identify any interim analyses. These omissions limit the ability to assess whether the posted analysis followed prespecified procedures and how robust the result may be to different analysis choices.

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

The posted results describe Cmax after a single dose of the test levetiracetam tablet and Keppra® under fed conditions in a randomized crossover study of healthy participants. The treatment means are 26.32 mcg/mL for the test formulation and 24.93 mcg/mL for the reference. The unadjusted recalculation gives a mean difference of -1.393, with a confidence interval from -4.269 to 1.483 and an unadjusted p-value of 0.342. The sponsor’s reported analysis gives a geometric mean ratio of 105, with a 90% confidence interval from 98.6 to 112.

These statistics describe the observed peak-concentration comparison and its uncertainty under the respective analyses. They do not establish that the formulations are exactly alike, nor do they alone establish a formal bioequivalence conclusion. The registry record does not report the equivalence bounds or decision rule needed to interpret the sponsor’s ratio against a prespecified criterion. The unadjusted mean difference is also not the same estimand as the sponsor’s ratio.

The results do not address whether either product improves health outcomes, nor do they provide a comparison of clinical effects. Cmax is a pharmacokinetic endpoint. The record does not establish what happens with repeated dosing, under conditions other than fed dosing, or in populations other than the healthy participants enrolled. It also does not provide detailed safety findings in the registry record reported for this analysis.

The confidence intervals are important because they show uncertainty rather than a single definitive value. The mean-difference interval includes zero, while the ratio interval is presented on another scale. Neither interval should be interpreted without its associated analysis and decision criterion. The available evidence supports a description of the posted measurements and comparisons, but not claims beyond the study’s endpoint and design.

16. Why This Trial Matters Statistically

This trial illustrates why the analysis scale matters in bioequivalence studies. The sponsor reports a ratio of geometric means, while the straightforward recalculation from treatment means is an absolute mean difference. Each can be useful for describing a comparison, but they answer different statistical questions. The ratio and difference cannot be substituted for one another, and the intervals attached to them must be interpreted on their own scales.

The crossover design is also central to statistical interpretation. The study used two treatment periods and reversed the order of the test and reference products across sequences. In general, crossover designs can make within-participant comparisons, but valid analysis depends on the model and assumptions used. The registry record does not report the detailed sponsor model or how period and sequence were considered. That makes the design clear while leaving important analysis details unavailable.

The endpoint reinforces the distinction between pharmacokinetic and clinical evidence. Cmax records the maximum observed concentration during sampling, and the study objective concerned the rate and extent of absorption. It does not directly measure patient experience or clinical outcome. The reported confidence intervals communicate precision, but formal interpretation of bioequivalence requires the relevant decision bounds, which are not reported in the registry record.

Finally, the planning illustrations distinguish design planning from post-study interpretation. They use fractions of the observed effect under stated assumptions and show why smaller effects call for larger planned groups. They are not evidence of the study’s original design process or its observed power. Together, these points show how study design, estimand, analysis method, and reporting detail shape what can be concluded from a numerical result.

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 leaves several trial-specific details unavailable. Although the study is described as a two-period, two-treatment, two-way crossover, it does not report the washout duration, the randomization procedure, or the sponsor’s detailed statistical model. It does not say how period or sequence effects were assessed, or how the within-participant structure was represented in the sponsor’s analysis. These omissions constrain independent evaluation of the reported ratio.

The registry reports enrollment of 22 and treatment summaries with n=21 for each product. The primary analysis description refers to subjects who completed the study, but the record does not give a participant-flow explanation or clarify the summary counts. It does not report how missing measurements were handled or whether sensitivity analyses were performed. These details cannot be inferred from the group summaries.

The study was unmasked, involved healthy participants, and evaluated a single dose under fed conditions. Its primary endpoint was Cmax, with blood samples collected over a 36 hour period. The record does not establish how results would generalize to other populations, dosing conditions, or repeated use. The registry entries available for this analysis do not include detailed safety outcomes or clinical endpoints.

Finally, the registry does not report the original sample-size rationale, an analysis plan, equivalence bounds, or multiplicity procedures. Those gaps are especially relevant when interpreting a reported ratio in an equivalence context. The unadjusted recalculation offers a separate summary, but it does not supply the missing protocol information or recreate the sponsor’s full analysis.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Levetiracetam 1000 mg Under Non-Fasting Conditions — NCT00859430.
• Registry record last updated 2024-08-20; status last verified 2024-08.
• Results first posted 2009-08-25.
• 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

NCT00859430 is a completed phase 1 randomized, unmasked crossover study sponsored by Teva Pharmaceuticals USA. The study enrolled 22 healthy participants and compared a single 1000 mg dose of a test levetiracetam tablet with Keppra® 1000 mg under fed conditions, using two treatment periods and reversed treatment sequences.

The primary endpoint was Cmax in mcg/mL, based on blood sampling over a 36 hour period. The treatment summaries report means of 26.32 for the test tablet and 24.93 for Keppra®. The sponsor’s reported geometric mean ratio is 105, with a 90% confidence interval from 98.6 to 112. An unadjusted recalculation reports a mean difference of -1.393, with a confidence interval from -4.269 to 1.483 and p=0.342.

The registry record does not report the sponsor’s detailed statistical method or equivalence bounds, so a formal equivalence decision cannot be reconstructed from the posted information alone.