Safety, Tolerability and Immunogenicity Study of 2-dose Heterologous Regimens for Ebola Vaccines Ad26.ZEBOV/MVA-BN-Filo: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Safety, Tolerability and Immunogenicity Study of 2-dose Heterologous Regimens for Ebola Vaccines Ad26.ZEBOV/MVA-BN-Filo (NCT02598388), a randomized phase 2 study. The primary outcome is Part 1, Part 2 (Group 2): Number of Participants With Unsolicited Adverse Events.

2. Clinical Question

3. Trial Design

This was a randomized, observer-blind, placebo-controlled, parallel-group phase 2 study. The registry describes triple masking. Enrollment was 578, and the study status is completed. Its stated purpose was to assess the safety, tolerability, and immunogenicity of different two-dose heterologous schedules of Ad26.ZEBOV and MVA-BN-Filo in healthy and HIV-infected adults. Both vaccines were administered intramuscularly.

The study had two parts. Part 1 involved US participants. They received MVA-BN-Filo or placebo on Day 1, then Ad26.ZEBOV or placebo on Day 15. Part 2 involved African participants and had two groups. Group 1 received Ad26.ZEBOV or placebo on Day 1, followed by MVA-BN-Filo or placebo. Group 2 received MVA-BN-Filo or placebo on Day 1, followed by Ad26.ZEBOV or placebo on Day 15. Group 1's second-dose day is not reported in the registry record.

For the primary safety outcome, data from Part 1 and Part 2 Group 2 were pooled as a 14-day interval regimen and reported separately for healthy and HIV-infected cohorts. This pooled outcome therefore does not represent every regimen in the study: Part 2 Group 1 used the opposite vaccine order. The outcome population was the safety set, defined as randomized participants who received at least one study vaccine dose, whether or not protocol deviations occurred. The registry record does not report the randomization ratio, allocation sequence, or detailed masking procedures.

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 578 randomized participants, 574 received at least one dose of study vaccine.

Only the first 6 of 12 groups are shown.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Safety set was based on full analysis set (FAS) which included all participants who were randomized and received at least 1 dose of study vaccine, regardless of the occurrence of protocol deviations.

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.

Only the first 6 of 13 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Part 1, Part 2 (Group 2): Number of Participants With Unsolicited Adverse Events, measured over Up to 28 days post each dose (up to Day 43).

The primary endpoint is the number of participants with unsolicited adverse events in Part 1 and Part 2 Group 2, up to 28 days after each dose and through Day 43. An adverse event is defined in the registry as an untoward medical occurrence during the study that does not necessarily have a causal relationship with the study agent. Unsolicited events were reported voluntarily by participants or elicited through nondirected interviewing at study visits.

This endpoint counts participants who experienced an event rather than the total number of event episodes. A participant with more than one event is not represented as multiple participants in this measure. The registry presents the pooled 14-day-regimen outcome separately for healthy and HIV-infected participants, with vaccine and placebo groups distinguished. Counts are 33 of 79.0 vaccine recipients and 10 of 20.0 placebo recipients among healthy participants, and 32 of 60.0 vaccine recipients and 3 of 15.0 placebo recipients among HIV-infected participants.

Counting participants with at least one event is useful for describing how commonly participants were affected during the specified follow-up window. It does not by itself describe the number, duration, severity, or causal attribution of events. The registry record does not report a separate primary endpoint definition for event severity or relatedness within this outcome.

7. Results Posted in the Registry

• Healthy (Part 1, Part 2; Group 2): MVA-BN-Filo and Ad26.ZEBOV: 41.8% (33/79) versus 50.0% (10/20) on control; risk difference -8.2 percentage points (95% CI -32.7 to 16.2), risk ratio 0.84 (0.50–1.39); unadjusted two-proportion z-test p = 0.507.
• HIV-infected (Part 1, Part 2; Group 2): MVA-BN-Filo and Ad26.ZEBOV: 53.3% (32/60) versus 50.0% (10/20) on control; risk difference 3.3 percentage points (95% CI -22.0 to 28.6), risk ratio 1.07 (0.65–1.76); unadjusted two-proportion z-test p = 0.796.
• HIV-infected (Part 1, Part 2; Group 2): Placebo: 20.0% (3/15) versus 50.0% (10/20) on control; risk difference -30.0 percentage points (95% CI -59.8 to -0.2), risk ratio 0.40 (0.13–1.21); unadjusted two-proportion z-test p = 0.069.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is binary at the participant level: each participant either had at least one unsolicited adverse event during the stated observation period or did not. The registry reports the number affected and the number at risk in each group. This is a straightforward way to describe event occurrence, but it compresses potentially different experiences into one category. It does not distinguish one event from multiple events for the same participant, or summarize the timing and duration of events.

What does a risk difference describe?

A risk difference compares the proportion of participants with an event in one group with the proportion in a reference group. A value near zero represents similar observed proportions; a value below or above zero indicates the direction of the contrast according to the ordering used. It is an absolute contrast, not a relative one. The healthy vaccine-versus-placebo estimate is -0.082, while the HIV-infected vaccine-versus-placebo estimate is 0.033. Their confidence intervals include zero, so the posted data do not locate either contrast precisely enough to rule out no difference using those intervals.

Risk differences are interpretable in terms of absolute event frequency, but estimates can be unstable when group sizes are limited. That is especially relevant when one arm has fewer participants and relatively few affected participants. A change in the observed count can then materially alter the estimated contrast. Estimates should be read together with their uncertainty intervals and the actual group counts, not in isolation.

What do confidence intervals add?

A confidence interval expresses the uncertainty associated with an estimate under the method used to calculate it. In general, a wider interval indicates less precision, while an interval spanning zero for a risk difference is compatible with no difference as well as positive or negative contrasts. An interval is not a guarantee that the true value lies inside it, and it does not assign a probability to a particular treatment effect. Its interpretation depends on the analysis method and assumptions.

The registry record does not report the method used for the sponsor's analysis of this endpoint, and it lists no analyses for the outcome. The comparison statistics accompanying the posted counts are unadjusted recalculations. They should not be confused with a reported adjusted trial analysis. In particular, the registry record does not identify covariates, stratification factors, or model-based adjustments for these comparisons.

What is the role of a p-value?

A p-value describes how unusual data at least as incompatible with a specified null hypothesis would be under the assumptions of a statistical test. It is not the probability that the null hypothesis is true, nor does it measure the size or practical importance of an effect. The healthy cohort's unadjusted p-value is 0.507; the HIV-infected vaccine comparison's is 0.796. The separate placebo-to-placebo comparison has an unadjusted p-value of 0.069. These values must be read alongside the estimated contrasts and their precision, and no multiplicity adjustment is reported.

The HIV-infected placebo comparison has a listed interval entirely below zero while its unadjusted p-value is 0.069. The registry record does not report the methods underlying the interval and p-value, so it does not explain their different indications. They should not be treated as interchangeable summaries or as a single internally reconciled result.

Why can over-dispersion matter?

Over-dispersion is a general concern for models of event counts when variability exceeds what a model assumes. It matters most when the analysis models numbers of events, rather than the binary indicator of whether each participant had any event. The primary endpoint counts participants with an event, so an event-count model is not what the posted endpoint describes. If repeated events or event totals were analyzed, a model that accounts for extra variability could affect standard errors and intervals. The registry record does not report such a model for this endpoint.

Finally, combining Part 1 and Part 2 Group 2 makes the listed outcome relevant to the pooled 14-day schedule, not to every schedule in the trial. Cohort-specific reporting preserves the distinction between healthy and HIV-infected participants, but it does not eliminate other possible differences between study parts or sites. The registry record does not report a pooled adjusted analysis that accounts for those factors.

10. Confidence Intervals and Precision

Precision is limited by the sizes of the outcome groups and by the uncertainty intervals around the comparisons. The healthy cohort includes 79.0 vaccine recipients and 20.0 placebo recipients; the HIV-infected cohort includes 60.0 vaccine recipients and 15.0 placebo recipients. The smaller placebo groups mean that their observed counts provide a relatively limited basis for estimating event frequency. The registry counts should therefore be read as observations in these study groups, not as exact estimates of what would occur in a much broader population.

The confidence intervals for the healthy vaccine-versus-placebo contrast and the HIV-infected vaccine-versus-placebo contrast both include zero. In general, that means the posted intervals remain compatible with no difference and with contrasts in either direction. Their widths also indicate that the estimates are not highly precise. The record does not support treating either point estimate as a settled measure of the size or direction of a difference.

The separately listed HIV-infected placebo comparison has an interval entirely below zero, but it compares HIV-infected placebo participants with a healthy placebo reference group, not vaccine with placebo within the HIV-infected cohort. Its unadjusted p-value is 0.069. The registry record does not explain the choice of this comparison or report methods sufficient to reconcile the interval and p-value. It should not be used to infer a vaccine effect.

Confidence intervals describe statistical uncertainty under a calculation method; they do not account automatically for limitations such as selection, measurement, or differences among study groups. The registry record does not report adjusted estimates for these contrasts. The event counts, group sizes, and interval coverage should be considered together when judging how precisely the outcome is described.

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 figures are planning illustrations based on the observed effect, not the trial's original design assumptions. For these binary outcomes, the entries are participants per arm, not total events. The planning calculation assumes a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It uses the stated fraction of the observed effect as the effect to detect. The registry record does not report the trial's own original sample-size justification or design assumptions.

The 80% power column gives the illustrative number of participants per arm for a design with an 80% chance of detecting the specified effect under the calculation assumptions, if that effect were the true one and the assumptions held. The 90% power column gives the corresponding per-arm figure for a higher chance of detection. These are planning concepts, not probabilities that the trial result is correct. This is not observed power.

Rows that use 100% of the observed effect assume the full estimated contrast would be the target in a future planning exercise. The rows using 75% or 50% use smaller target effects. Smaller effects require more participants to distinguish them from chance variation at the same planned significance level and power. For example, the healthy cohort's illustration gives 575 participants per arm for 80% power at 100% of the observed effect, compared with 1,026 at 75% and 2,314 at 50%. The HIV-infected vaccine comparison's corresponding figures are 3,527, 6,275, and 14,123. These figures are particularly sensitive to the observed effect used, which itself is uncertain.

The illustrations are not a retrospective verdict on whether the completed study was adequately sized. They are not substitutes for the unreported original design rationale, and they do not account for features not included in the stated calculation assumptions. Read them as scenario-based planning aids only.

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 = 0.500; treatment proportion p₁ = 0.418; average p̄ = 0.459.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.7047 = 1.3812.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.7023 = 0.5911.
4. n = (1.3812 + 0.5911)² / (-0.0823)² = 574.6, rounded up to 575 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 <- 0.5                          # control proportion
p1 <- 0.417721518987342                          # 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(2598388)
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 = 575 (80% power), 769 (90% power)
 75% of observed effect: n per arm = 1,026 (80% power), 1,373 (90% power)
 50% of observed effect: n per arm = 2,314 (80% power), 3,097 (90% power)
power.prop.test: 574.6 
Simulated power at n = 575 per arm: 0.797 (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 outcome list includes 15 outcome measures, while the posted primary endpoint concerns unsolicited adverse events in pooled Part 1 and Part 2 Group 2 participants. One secondary outcome is listed as the number of participants with adverse events. The registry record does not report a multiplicity strategy, such as a hierarchy of tests, an adjustment to significance thresholds, or a designation of which analyses were confirmatory. The unadjusted p-values therefore should not be treated as multiplicity-controlled evidence across all outcomes or comparisons.

The record also lists no analyses for this outcome. It does not report the statistical analysis plan, a prespecified model, covariate adjustments, or the methods used to construct the posted comparison intervals and p-values. The registry record does not report how missing outcome observations were handled, whether any participants lacked follow-up, or whether sensitivity analyses assessed missingness. No assumptions about missing data should be inferred from the counts alone.

The safety set is described as participants who were randomized and received at least one dose, regardless of protocol deviations. This defines who could contribute to the safety outcome, but it does not give the number excluded from each cohort or explain whether all eligible participants had complete follow-up. The group sizes shown for this outcome are the denominators in the posted counts; they do not establish how missing observations were treated outside those denominators.

The pooled endpoint combines Part 1 and Part 2 Group 2, with results separated by healthy and HIV-infected cohorts. The registry record does not report an adjusted analysis accounting for differences among study parts, locations, or cohorts. It also does not report whether subgroup analyses beyond the health-cohort split were prespecified. These omissions limit assessment of how analysis choices, missingness, or multiple comparisons could affect interpretation.

14. Safety Results

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

Deaths

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: staphylococcal osteomyelitis (20.0%), cervical cord compression (20.0%), anaphylactic reaction (10.0%), panic attack (5.0%), appendicitis perforated (2.6%).

The serious adverse event summary lists numbers at risk and numbers affected separately for each study part, cohort, and intervention group. Across the listed groups, the number of participants with serious adverse events ranges from 0 to 2; the registry reports 0 deaths in each listed group. These figures are distinct from the primary endpoint, which counts participants with unsolicited adverse events and is restricted to pooled Part 1 and Part 2 Group 2.

The record also lists serious-event terms, including Staphylococcal Osteomyelitis at 20.0%, Cervical Cord Compression at 20.0%, Anaphylactic Reaction at 10.0%, Panic Attack at 5.0%, Appendicitis Perforated at 2.6%, and Postoperative Wound Infection at 2.6%. The registry record does not report the denominator or calculation details for these term percentages in the information available for this summary. The percentages should not be used to infer group-specific risks.

These counts describe events recorded in the study; they do not establish that a vaccine caused an event. The registry record does not report the serious-event observation window or event-level causal assessments in this summary.

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

The posted results describe unsolicited adverse-event counts in a defined safety population and follow-up window. They show that events were reported in both vaccine and placebo groups, among both healthy and HIV-infected participants. The unadjusted vaccine-versus-placebo estimates have intervals that include zero, so the comparisons are imprecise and do not establish a clear direction of difference from these data alone.

The findings do not show that the vaccine regimen caused the reported events or that all events were related to vaccination. The adverse-event definition explicitly allows events without a causal relationship to the study agent. Nor does the count endpoint capture all dimensions of safety: it counts affected participants, not total episodes, duration, intensity, or clinical consequences. The separate serious adverse event summary should be read as a distinct collection of information, not combined with the primary endpoint.

The primary endpoint applies to the pooled 14-day regimen in Part 1 and Part 2 Group 2. It does not directly summarize the alternative order used in Part 2 Group 1. The posted outcome also separates healthy and HIV-infected participants, so results for one cohort should not be substituted for the other. The third listed comparison is between HIV-infected and healthy placebo groups; it does not measure the vaccine-versus-placebo contrast within HIV-infected participants.

These results do not establish effectiveness against Ebola. Although immunogenicity was part of the study's purpose, the outcome is unsolicited adverse events, not immune response or disease prevention. The registry record does not report a sponsor analysis for this primary endpoint, a multiplicity plan, or a missing-data analysis. Interpretations should remain bounded by the reported counts, comparison estimates, and their uncertainty, rather than extending to unreported endpoints or causal conclusions.

16. Why This Trial Matters Statistically

This trial illustrates why the definition and scope of a clinical endpoint matter. A count of participants with at least one unsolicited event is easy to understand, but it does not represent the number or severity of event episodes. Keeping the participant-level endpoint distinct from serious adverse event summaries avoids conflating different safety measures.

It also demonstrates the importance of reading a pooled result alongside the design. Part 1 and Part 2 Group 2 shared the 14-day vaccine order described for the primary outcome, while Part 2 Group 1 used the opposite order. The pooled endpoint therefore answers a narrower question than the study's full set of schedules. Separate reporting for healthy and HIV-infected participants further signals that cohort context should be preserved when interpreting comparisons.

Precision and comparison choice are equally important. The vaccine-versus-placebo confidence intervals include zero, and the groups are not equally sized. A point estimate can look directional while remaining uncertain. The separate placebo-to-placebo contrast is a useful reminder that labels and reference groups must be checked before interpreting a numerical comparison as a treatment effect. Statistical summaries acquire meaning from the design and denominator behind them, not from a number alone.

Finally, planning illustrations based on observed effects are not the same as the trial's original design justification. Their sample-size requirements change as the target effect changes, emphasizing the connection between effect size, precision, and planning. Because the registry record does not report the original justification or analysis details, readers cannot reconstruct all design and analysis decisions from these results.

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 primary result covers only the pooled Part 1 and Part 2 Group 2 14-day regimen. It does not provide the same primary outcome comparison for Part 2 Group 1, which used the opposite vaccine order. Pooling the two included parts may also obscure differences between them; the registry record does not report a part-specific primary comparison or adjusted analysis.

The vaccine and placebo groups are different in size, particularly in the HIV-infected cohort. The posted estimates are unadjusted, and their confidence intervals are broad enough to include zero for both vaccine-versus-placebo comparisons. The record does not report a model accounting for cohort or study-part differences, nor does it report a multiplicity approach for the listed outcome measures and comparisons.

The third comparison is difficult to interpret as a treatment contrast because it compares HIV-infected placebo participants with a healthy placebo reference. The registry record does not explain why this comparison was listed. Its estimate should not be treated as an estimate of the vaccine regimen's effect in HIV-infected participants.

Further details needed to assess the outcome are not reported in the registry record: the original statistical analysis plan, the sponsor's analysis for this endpoint, the construction methods for the posted intervals and p-values, missing-data handling, and the original sample-size justification. The outcome is limited to unsolicited adverse events and does not, on its own, describe event severity, duration, or causal attribution. These limitations constrain interpretation to the posted counts and their stated comparison summaries.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Safety, Tolerability and Immunogenicity Study of 2-dose Heterologous Regimens for Ebola Vaccines Ad26.ZEBOV/MVA-BN-Filo — NCT02598388.
• Registry record last updated 2025-02-04; status last verified 2025-01.
• Results first posted 2022-02-18.
• 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

NCT02598388 is a completed randomized, observer-blind, placebo-controlled, parallel-group phase 2 study with enrollment of 578. It assessed safety, tolerability, and immunogenicity of two-dose heterologous Ad26.ZEBOV and MVA-BN-Filo schedules in healthy and HIV-infected adults.

The primary safety outcome counts participants with unsolicited adverse events through Day 43 in pooled Part 1 and Part 2 Group 2, reported separately by health cohort. The posted vaccine-versus-placebo comparisons are unadjusted recalculations; their confidence intervals include zero. The registry record lists no analyses for this endpoint and does not report the original sample-size rationale, multiplicity approach, or missing-data methods.