IV Keppra in the Emergency Department for Prevention of Early Recurrent Seizures: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for IV Keppra in the Emergency Department for Prevention of Early Recurrent Seizures (NCT00510783), a randomized phase 4 study. The primary outcome is Number of Participants Who Experienced a Recurrent Seizure After Treatment..

2. Clinical Question

This comparison is not a simple contrast between levetiracetam and one specific comparator drug: the control arm combined participants treated with either phenytoin or fosphenytoin. The registry record does not report how the choice between those two control treatments was made or how many participants received each one. The primary outcome therefore addresses the posted comparison of levetiracetam against that combined control arm, rather than separate comparisons against each control medicine.

3. Trial Design

The registry describes NCT00510783 as a completed, phase 4, randomized, parallel-group study sponsored by Emory University. Its enrollment is listed as 158. The recorded start is 2007-07, and primary completion is recorded as 2008-12. The trial is described as having no masking, so it was open-label according to the registry entry. The record does not give further details about who knew each participant’s assigned treatment or whether outcome assessment was masked.

One arm received intravenous levetiracetam. The other arm received either intravenous phenytoin or intravenous fosphenytoin. The registry describes the levetiracetam infusion as 1 gram in 100 mL diluent over 15 minutes. It describes phenytoin as 1 gram infused at 25 mg/min or slower depending on vital signs, and fosphenytoin as 1 gram infused at 15 mg/min or slower depending on vital signs. These descriptions identify the assigned treatment options; details about allocation procedures, treatment adherence, crossovers, or treatment after the initial infusion are not reported in the registry record.

The posted primary outcome is the number of participants with a recurrent seizure within 24 hours of treatment in the emergency department. The registry lists 82.0 participants in the phenytoin/fosphenytoin arm and 76.0 in the levetiracetam arm. It reports 2.0 and 3.0 participants with the endpoint, respectively. The registry record does not describe the trial’s original sample-size justification, planned statistical analysis, or how the combined control treatment was allocated. It also does not report analysis populations or whether all randomized participants were included in the endpoint 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.

Before assignment: study completed

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 Number of Participants Who Experienced a Recurrent Seizure After Treatment., measured over 24 hours.

The primary endpoint counts participants who experienced a recurrent seizure after treatment in the emergency department. The registry defines a recurrent seizure as a seizure within 24 hours of treatment and gives the timeframe as 24 hours. This is a binary outcome: each participant either met the endpoint definition during the stated follow-up period or did not. The posted results give the number of participants with an event in each arm, along with the arm sizes.

A short, defined follow-up window makes the endpoint directly focused on early recurrence after emergency-department treatment. It also means that the result describes only the recorded events within that window; it does not describe recurrence beyond 24 hours or other aspects of seizure control. The registry record does not report further operational details about how seizures were identified, confirmed, or adjudicated, nor does it report whether follow-up was complete for every participant.

The registry summary says the investigators were also looking for possible side effects. However, the posted outcome information lists one outcome measure, and a secondary outcome measure is not reported in the registry record. The adverse-event figures are listed separately from the primary endpoint and should not be interpreted as additional efficacy outcomes.

7. Results Posted in the Registry

• Levetiracetam: 3.9% (3/76) versus 2.4% (2/82) on control; risk difference 1.5 percentage points (95% CI -4.0 to 7.0), risk ratio 1.62 (0.28–9.42); unadjusted two-proportion z-test p = 0.588.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a count of participants with a binary event. Each participant is classified according to whether a recurrent seizure occurred within 24 hours of emergency-department treatment. The registry reports event counts and arm sizes, then gives comparative measures for levetiracetam versus the phenytoin/fosphenytoin control arm. The posted comparison is unadjusted; further details of the analysis method are not reported in the registry record.

What does the risk difference describe?

The risk difference expresses the absolute contrast between the event risks in two arms. A value of zero corresponds to equal estimated risks. A positive value for the stated levetiracetam comparison indicates a higher estimated event risk in that arm; a negative value indicates a lower estimated risk. The reported estimate is 0.015, with a confidence interval from -0.04 to 0.070. Since the interval includes zero, the data are compatible with no difference as well as with effects in either direction. General explanation: an absolute measure is often easier to relate to event frequency than a relative measure, but it can still be uncertain when events are sparse.

What do the relative measures add?

The risk ratio compares the event risk in one arm with the event risk in the other. A value of one represents equal risks. The posted risk ratio is 1.618, with a confidence interval from 0.278 to 9.424. The odds ratio compares event odds rather than risks; its posted estimate is 1.644, with a confidence interval from 0.267 to 10.12. For an uncommon outcome, odds and risks can be close, but they are not the same measure. General explanation: relative measures can communicate proportional contrasts, while their interpretation may be less intuitive when readers want the absolute change in event frequency.

Both relative-effect intervals include one, the value corresponding to no relative difference. Their wide ranges show that the data do not pin down the relative contrast precisely. The small number of events contributes to this uncertainty: with few endpoint events, changing the event count in either arm can appreciably affect an estimate. This is a general feature of sparse binary outcomes, not evidence by itself that the analysis was incorrectly performed.

How should the p-value be read?

The unadjusted p-value is 0.588. In general, a p-value measures how unusual data at least as incompatible with a specified null hypothesis would be if that hypothesis and the model assumptions were true. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. Here, the p-value does not indicate a statistically significant contrast under the conventional framework. It also cannot demonstrate that the arms are equivalent or that an important difference is absent.

The record labels the p-value unadjusted. It does not report whether the analysis included covariate adjustment, how missing endpoint observations were handled, or whether any prespecified analysis differed from the posted estimate. The record lists zero posted analyses, so the reported comparison should not be mistaken for a fully described statistical analysis plan. Further details about the method used to calculate the confidence intervals and p-value are not reported in the registry record.

Why can over-dispersion matter?

General explanation: over-dispersion is a concern for some count-data models when observed variability exceeds what the model expects. It is not a separate issue established by these results. The primary endpoint is reported as participants with an event, not as a rate of repeated events per participant. The registry does not describe a count-data model or report an over-dispersion assessment. Therefore, an over-dispersion adjustment should not be inferred from this record.

What do the planning illustrations represent?

The planning figures are illustrations based on the observed effect, not the trial’s original design assumptions. They use a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect. The record presents participants per arm for 80% and 90% power at the listed effect fractions. These are planning calculations; they are not observed trial outcomes or evidence about the probability that the trial’s result is correct. The original trial’s sample-size justification and design assumptions are not reported in the registry record.

General explanation: confidence intervals and p-values address uncertainty in different but related ways. The interval displays a range of effect estimates compatible with the data under the calculation method, while the p-value summarizes evidence against a specified null value. Neither replaces examination of the event counts, the endpoint definition, the comparator structure, and the limitations of the record. In this comparison, the few reported primary events and broad intervals are central to interpreting the estimates.

10. Confidence Intervals and Precision

The risk-difference estimate is 0.015, and its confidence interval runs from -0.04 to 0.070. The interval crosses zero, so it includes the value corresponding to no absolute difference. It also includes estimates on either side of zero, meaning the posted data do not precisely determine the direction or size of the difference. The interval is broad relative to the point estimate, which makes the estimate uncertain.

The relative estimates show similar uncertainty. The risk ratio is 1.618, with an interval from 0.278 to 9.424. The odds ratio is 1.644, with an interval from 0.267 to 10.12. Each interval includes one, the value corresponding to no relative difference, and each spans a wide range of possible relative effects. The relatively small number of recorded endpoint events is important when considering this imprecision.

General explanation: a confidence interval is a range produced by a statistical procedure that, under its assumptions, reflects uncertainty in an estimate. It is not a guarantee that the true effect lies within a particular interval, nor does it say that values inside the interval are equally likely. The registry record does not report the interval-calculation method, so the limits should be read as posted rather than as a basis for reconstructing the analysis.

The unadjusted p-value is 0.588, but the interval is more informative for seeing the range of effects compatible with the posted estimate. Neither the p-value nor the interval establishes equivalence. A conclusion that treatments are equivalent would require an appropriate design and a prespecified margin, neither of which is 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 planning figures are expressed as participants per arm, not as a total across both arms and not as total events. The calculations use the observed effect as a starting point and show planning for 100% of that effect, 75% of the observed effect, and 50% of the observed effect. For each effect assumption, the record gives the number of participants per arm for 80% power and for 90% power.

In plain language, the 80% power column gives the planning number per arm under which the stated statistical assumptions would yield an 80% chance of detecting the specified effect if that effect were the true one. The 90% power column gives the corresponding planning number per arm for a 90% chance under those assumptions. These descriptions concern hypothetical future planning, not the probability that the posted finding is true. This is not observed power.

The listed 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 table shows 2,132 participants per arm for 80% power and 2,854 for 90% power when planning for 100% of the observed effect; 3,574 and 4,784 for 75%; and 7,553 and 10,110 for 50%.

Smaller effects require more participants to detect under the same planning assumptions, which is why the illustrative sample sizes increase as the assumed fraction of the observed effect decreases. The figures are not the original trial’s sample-size calculation. The trial’s own design assumptions and sample-size justification are not reported in the registry record. The curve should be read as a visual presentation of this planning relationship, not as an additional observed result.

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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.024; treatment proportion p₁ = 0.039; average p̄ = 0.032.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.2486 = 0.4873.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.2484 = 0.2091.
4. n = (0.4873 + 0.2091)² / (0.0151)² = 2,131.7, rounded up to 2,132 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
p0 <- 0.024390243902439                          # control proportion
p1 <- 0.0394736842105263                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^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 (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(510783)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 2,132 (80% power), 2,854 (90% power)
 75% of observed effect: n per arm = 3,574 (80% power), 4,784 (90% power)
 50% of observed effect: n per arm = 7,553 (80% power), 10,110 (90% power)
power.prop.test: 2131.7 
Simulated power at n = 2,132 per arm: 0.803 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-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. Multiplicity, Missing Data and Analysis Populations

The registry record lists one posted outcome measure, the number of participants with a recurrent seizure after treatment, and lists zero posted analyses. It also records zero posted secondary outcomes. The brief summary says the investigators were looking for possible side effects, but the registry entry does not provide a set of posted secondary efficacy outcomes or their analysis results. The number and scope of all outcomes assessed during the trial are not reported in the registry record.

Multiplicity arises when several outcomes, treatment contrasts, time points, or analytical choices are examined. General explanation: examining many comparisons can increase the chance of finding an apparently notable result by chance, and statistical plans may address this through prespecification or adjustment. The registry record does not report a multiplicity plan, adjustment method, hierarchy of outcomes, or whether any additional comparisons were undertaken. No correction or absence of correction should be inferred from the single posted comparison.

The record provides arm-level endpoint counts and an unadjusted comparison, but does not report the analysis population, whether the endpoint analysis followed intention-to-treat principles, or how participants with missing outcome data were handled. It also does not state whether participants were excluded after randomization or whether endpoint ascertainment was complete. Consequently, the posted counts cannot be assumed to represent a particular analysis population beyond the arm sizes and outcome counts recorded.

Further information about protocol deviations, missing data, imputation, sensitivity analyses, or prespecified versus exploratory analyses is not reported in the registry record. These omissions limit assessment of potential bias and of how closely the posted analysis followed the trial’s planned analysis. They do not demonstrate that data were missing or that analyses were changed; they mean those details are unavailable in the record.

14. Safety Results

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

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent non-serious adverse event terms

The registry lists serious adverse events for 0 of 82 participants in the phenytoin/fosphenytoin arm and 0 of 76 participants in the levetiracetam arm. The adverse-event record also lists other, non-serious adverse events for 40 participants in the control arm and 19 participants in the levetiracetam arm. These are counts of affected participants as recorded, not descriptions of event severity beyond the serious versus other classification.

The record does not list serious adverse-event terms, and it does not provide a death count. The death values are not reported in the registry record. No further event descriptions, timing, relationship assessments, or severity details are included in the posted safety information. The listed counts should therefore be read as a limited registry summary, not as a complete account of all safety information collected in the study.

General explanation: a zero count in a trial arm means no events of that category were recorded in that arm in the posted data; it does not show that the event is impossible or that risk is absent. Sparse safety data can be imprecise, and counts alone do not establish whether an event was caused by treatment. The registry’s figures should be interpreted alongside their limited detail and should not be used to make individual treatment decisions.

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

The result is a comparison of recorded recurrent seizures within 24 hours after emergency-department treatment. The registry reports 2.0 events among 82.0 participants in the phenytoin/fosphenytoin arm and 3.0 among 76.0 in the levetiracetam arm. The risk difference is 0.015, with a confidence interval from -0.04 to 0.070, and the unadjusted p-value is 0.588. The confidence interval includes no difference and effects in both directions, so the estimate is not precise enough to settle the direction or magnitude of the contrast.

This result does not show that the treatments are equivalent. Equivalence requires a defined margin and an analysis designed to assess whether differences fall within that margin; such details are not reported in the registry record. Nor does the result establish that the observed numerical contrast is a true treatment effect. With few primary events and wide intervals, chance variation remains a plausible explanation for the observed difference.

The comparison is also specific to the enrolled study population, the registry’s endpoint definition, and the stated 24-hour timeframe. It does not address later recurrence, other seizure outcomes, or outcomes not posted in the registry. The control arm pooled participants who received phenytoin or fosphenytoin, and the record does not report separate results for those medicines. Therefore, the result cannot be read as a distinct comparison with each control drug.

The registry’s open-label design and missing details about endpoint assessment, allocation procedures, analysis populations, and missing data further constrain interpretation. These are uncertainties about what the record documents; they do not establish that bias occurred. The posted figures are best understood as an imprecise registry comparison, rather than a definitive statement about treatment effects or safety.

16. Why This Trial Matters Statistically

This trial is statistically notable because its primary endpoint is a binary event recorded in a short follow-up window, and the posted event counts are small. When only a few participants experience an outcome, estimates can be sensitive to the observed counts and confidence intervals can remain wide. That makes precision, not just the direction of a point estimate or a p-value, central to interpretation.

The comparison also illustrates why absolute and relative measures should be considered together. The risk difference gives an absolute contrast, while the risk ratio and odds ratio express relative contrasts. In this record, all of the intervals indicate considerable uncertainty. Looking at the measures together discourages an overly confident conclusion based on a single point estimate.

The control arm’s structure matters as well. It combines two active medicines, phenytoin and fosphenytoin, so the result answers a comparison against that combined arm rather than separate drug-specific questions. Without the distribution of control treatments or separate outcomes, the reader cannot determine whether the pooled result reflects similar patterns across the two medicines.

Finally, the record provides only limited information about the planned analysis and handling of missing data, while its sample-size illustrations are based on the observed effect rather than the trial’s original design. Those illustrations show how assumptions about the effect size influence planning, but they do not replace a trial-specific sample-size justification. Taken together, these features make transparent reporting of uncertainty and limitations essential to statistical interpretation.

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 result is based on few primary endpoint events, and the confidence intervals are wide. The estimates therefore leave substantial uncertainty about the direction and size of the treatment contrast. The unadjusted p-value does not resolve that uncertainty, and the record does not report an equivalence framework.

The comparator is a combined phenytoin/fosphenytoin arm. The registry does not report how participants were distributed between those options or provide separate estimates for each drug. That limits drug-specific interpretation. The study is also open-label, with masking listed as none. The registry record does not describe whether outcome assessment was independently masked or how recurrent seizures were confirmed.

Important methodological details are not reported in the registry record, including the allocation procedure, the original sample-size justification, analysis populations, treatment of missing endpoint data, and the statistical analysis plan. The record lists no posted analyses and no posted secondary outcome measures. The brief summary mentions possible side effects, but the available safety figures do not include detailed event terms or relationship assessments; deaths are not reported in the registry record.

The primary outcome applies only to recurrence within 24 hours after treatment in the emergency department. The record does not provide results for outcomes beyond that period. The registry also does not provide details needed to assess whether follow-up was complete or whether all randomized participants contributed to the posted outcome counts. These limits define what can be concluded from the registry record; they do not establish that a particular bias or data problem occurred.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: IV Keppra in the Emergency Department for Prevention of Early Recurrent Seizures — NCT00510783.
• Registry record last updated 2013-12-12; status last verified 2013-11.
• Results first posted 2013-08-05.
• 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

NCT00510783 is recorded as a completed, randomized, parallel-group, phase 4 study with no masking. The sponsor is Emory University. The recorded study start is 2007-07, primary completion is 2008-12, and enrollment is 158.

The trial compared intravenous levetiracetam with a control arm receiving either intravenous phenytoin or intravenous fosphenytoin. Its primary endpoint was participants with a recurrent seizure within 24 hours of emergency-department treatment. The registry reports 2.0 events among 82.0 participants in the control arm and 3.0 among 76.0 in the levetiracetam arm. The reported risk difference is 0.015, with a confidence interval from -0.04 to 0.070 and an unadjusted p-value of 0.588.

The record reports 0 serious adverse events in each arm, while other adverse events are listed for 40 control-arm participants and 19 levetiracetam-arm participants. Further details about the analysis plan, missing data, analysis populations, and separate outcomes for phenytoin and fosphenytoin are not reported in the registry record.