A Study to Evaluate the Efficacy, Safety and Immunogenicity of a Vaccine Designed to Protect Against Infection With Shigella Sonnei in Healthy Adults: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Evaluate the Efficacy, Safety and Immunogenicity of a Vaccine Designed to Protect Against Infection With Shigella Sonnei in Healthy Adults (NCT03527173), a randomized phase 2 study. The primary outcome is Percentage of Subjects With at Least One Episode of Shigellosis According to the Protocol Primary Case Definition.

2. Clinical Question

3. Trial Design

This was a randomized, quadruple-masked, parallel-group Phase 2 study sponsored by GlaxoSmithKline. Its official title describes evaluation of the efficacy, safety and immunogenicity of the GVGH Shigella sonnei vaccine, 1790GAHB, in a human challenge study of healthy, non-immune adults. The registry title also identifies the vaccine as designed to protect against infection with Shigella sonnei. The registry record lists bacillary dysentery as the condition.

The study enrolled 71 participants and is marked completed. The recorded start date is 2018-08-29, and the primary completion date is 2019-05-08. The arm descriptions specify adults aged 18 to 50 years at the first vaccination. The vaccine and placebo groups each received two intramuscular injections, on Day 1 and Day 29, followed by challenge on Day 57 with the S. sonnei 53G strain. The primary endpoint observation period extended through the end of the inpatient stay on Day 64.

The primary efficacy analysis population was the per-protocol set. It included participants who received the assigned vaccine or placebo according to protocol, were not excluded for reasons determined before unblinding or analysis, and received the challenge agent. The results list 32 participants in the vaccine group and 28 in the placebo group for the primary endpoint. Those analysis denominators should not be confused with the total enrollment, because the efficacy result is reported for the specified per-protocol population.

The registry record describes several safety, immunogenicity and disease outcomes, but it does not report the full protocol, all operational details, the original design assumptions or the complete statistical analysis plan. Details such as the method used to generate the random sequence, allocation concealment, a planned sample-size justification, and the procedures for handling missing endpoint data are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: All 71 subjects enrolled in the study, received a study vaccination and were included in the All Exposed Set.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The analysis is performed on the Per-protocol set (PPS) for efficacy, which includes all subjects who received the study vaccine or placebo according to protocol, were not excluded due to other reasons defined prior to unblinding or analysis and who received the challenge agent.

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 Percentage of Subjects With at Least One Episode of Shigellosis According to the Protocol Primary Case Definition, measured over Starting with the challenge visit (Day 57) and lasting up to the end of the inpatient stay (Day 64).

The primary endpoint was the percentage of participants with at least one episode of shigellosis meeting the protocol primary case definition after challenge. It was assessed from Day 57 through Day 64 and expressed as an attack rate, with a 90% confidence interval calculated by the Clopper-Pearson method. The efficacy analysis used the per-protocol set, which required receipt of the assigned injections according to protocol, inclusion under the prespecified eligibility for analysis, and receipt of the challenge agent.

The case definition required shedding of S. sonnei 53G accompanied by moderate or severe diarrhea, or shedding accompanied by an oral temperature of at least 38.5°C. Moderate diarrhea was defined as 4 to 5 loose or watery stools, or 400 to 800 grams of such stools, within 24 hours. Severe diarrhea was defined as 6 or more loose or watery stools, more than 800 grams of such stools within 24 hours, or a need for medical intervention. The definition of intervention in severe diarrhea included intravenous fluids or anticipation of antibiotic treatment before the fifth day after challenge.

This composite definition combines evidence of infection with clinically defined illness. It is used to identify participants who met a common protocol threshold rather than counting every positive challenge or symptom as an endpoint event. The registry also lists related secondary definitions, including a CHIM Working Group case definition and outcomes for more severe disease, symptoms and stool measures. The primary endpoint is therefore a specific protocol-defined measure, not a general measure of every possible manifestation of infection.

7. Results Posted in the Registry

• S. Sonnei Group: 46.9% (15/32) versus 42.9% (12/28) on control; risk difference 4.0 percentage points (95% CI -21.2 to 29.2), risk ratio 1.09 (0.62–1.92); unadjusted two-proportion z-test p = 0.755.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Percentage of Subjects With at Least One Episode of Shigellosis According to Controlled Human Infection Model (CHIM) Working Group Case Definition for Shigellosis

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • S. Sonnei Group: 46.9% (15/32) versus 32.1% (9/28) on control; risk difference 14.7 percentage points (95% CI -9.7 to 39.2), risk ratio 1.46 (0.76–2.80); unadjusted two-proportion z-test p = 0.245.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between S. Sonnei Group and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Subjects With at Least One Episode of Shigellosis

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • S. Sonnei Group: 46.9% (15/32) versus 35.7% (10/28) on control; risk difference 11.2 percentage points (95% CI -13.6 to 35.9), risk ratio 1.31 (0.71–2.44); unadjusted two-proportion z-test p = 0.382.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between S. Sonnei Group and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Subjects With at Least One Episode of More Severe Shigellosis

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • S. Sonnei Group: 15.6% (5/32) versus 14.3% (4/28) on control; risk difference 1.3 percentage points (95% CI -16.7 to 19.4), risk ratio 1.09 (0.33–3.68); unadjusted two-proportion z-test p = 0.885.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between S. Sonnei Group and the control. The interval gives the range of effects that remain compatible with the data.

The listed secondary outcomes cover alternative definitions of shigellosis, more severe illness, specific symptoms, stool frequency and weight, onset timing, and adverse events. The registry record reports numerical results for several related binary shigellosis endpoints, with further outcome names listed but numerical results not reported in the registry record. These outcomes add different views of illness after challenge, but they do not replace the primary protocol-defined endpoint.

The secondary pattern does not materially change the broad uncertainty in the primary comparison. The secondary estimates use the same challenge setting and generally relate to whether participants experienced illness or a particular severity threshold. Their interpretation depends on their own definitions and analysis populations. The record does not report enough information to assume that a secondary case definition is interchangeable with the primary definition, or that all secondary analyses used the same method.

General explanation: When several outcomes or comparisons are examined, the chance of obtaining at least one seemingly notable result by chance can increase. Multiplicity procedures can address this by adjusting the testing strategy or by defining a hierarchy of hypotheses. The registry record does not report whether an adjustment or hierarchy was planned for the secondary outcomes. Unadjusted secondary results therefore carry less weight as confirmatory evidence: they are most useful as descriptive or exploratory evidence, and should be considered alongside the primary outcome rather than as independent proof.

The posted secondary outcomes also include measures that may be correlated because they describe related aspects of the same illness episode. Correlation means the results are not necessarily independent signals. Without a prespecified analysis plan and multiplicity details, a pattern across related outcomes should be interpreted cautiously rather than counted as multiple separate confirmations of a single conclusion.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Percentage of Subjects With at Least One Episode of Shigellosis According to Controlled Human Infection Model (CHIM) Working Group Case Definition for Shigellosis
• Percentage of Subjects With at Least One Episode of Shigellosis
• Percentage of Subjects With at Least One Episode of More Severe Shigellosis
• Percentage of Subjects With Specific Disease Symptoms
• Mean Number of Grade 3-5 Stools Per Subject
• Weight of Grade 3-5 Stools
• Time (Days) From Challenge Administration to Onset of Shigellosis, According to the Primary Case Definition
• Number of Subjects With Any Solicited Local Adverse Events (AEs)

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is binary at the participant level: each participant either met the case definition at least once during the observation window or did not. A percentage is a direct way to describe the frequency of such an event in each group. The arm-specific percentages are accompanied by 90% confidence intervals, calculated using the Clopper-Pearson method according to the endpoint description.

General explanation: A confidence interval expresses uncertainty in an estimate by identifying a range of values compatible with the observed data under the method’s assumptions. For a binary outcome, the interval reflects both the event proportion and the amount of information in the group. Small groups generally yield less precise estimates than larger groups, especially when the event count is limited. A confidence interval is not a guarantee that the true value lies in the interval, and it does not by itself answer whether a treatment effect is clinically important.

The Clopper-Pearson method is an exact method for estimating a confidence interval for a binomial proportion. It is often used when the outcome is a count of participants with and without an event. The registry specifies this method for the attack-rate intervals. It does not report a rationale for choosing it, and the method description should not be taken as evidence about other parts of the analysis plan.

The recalculated comparison of the group percentages is an unadjusted summary comparison. A p-value is a measure of how compatible the observed data are with a specified null model, given the method used. It is not the probability that the null hypothesis is true, and it does not measure the size or practical importance of a difference. The posted recalculation’s p-value is 0.755. The registry record also reports a sponsor analysis using a one-sided Barnard Unconditional Exact Test; that is a distinct analysis and is discussed separately.

For binary outcomes, the risk difference describes the difference between two group event probabilities, while the risk ratio compares those probabilities as a ratio. The odds ratio compares the odds of an event between groups. These are different effect measures and can convey different numerical scales for the same data. The re-calculated comparisons use unadjusted summaries based on the posted group counts. The sponsor-reported vaccine-efficacy rate instead has its own reported estimate and confidence interval. No new effect estimate should be inferred by combining these presentations.

General explanation: A test of non-inferiority asks whether the data are sufficiently inconsistent with the possibility that one group is worse than the other by more than a prespecified margin. Such a conclusion depends on the chosen margin, direction of the test, analysis population and statistical procedure. The sponsor analysis is labelled non-inferiority in the registry record, but the margin and complete decision criteria are not reported in the registry record. The label alone is not enough to establish whether a formal non-inferiority criterion was met.

Randomization is intended to make treatment assignment independent of participants’ baseline characteristics, on average. Masking can reduce the influence of knowledge of assignment on participant behavior, clinical assessment or other study processes. This study is recorded as randomized and quadruple-masked. The registry record does not report the allocation procedure, the roles included in quadruple masking, or whether masking was assessed after the challenge.

The primary efficacy population is per protocol rather than all enrolled participants. Its definition requires protocol-consistent receipt of the assigned study product, no exclusion for reasons defined before unblinding or analysis, and receipt of the challenge agent. This defines who contributes to the primary analysis, but the registry record does not report a participant flow or explain how missing endpoint assessments were treated. General statistical practice distinguishes a per-protocol analysis from an intention-to-treat analysis, which retains participants according to randomized assignment; no intention-to-treat primary result is reported in the registry record.

The posted planning illustrations use a two-sided significance level of 0.05, equal allocation between two arms and a normal approximation. They examine the observed effect and specified fractions of it. These are planning calculations based on observed results, not evidence about the original trial’s design or its planned sample-size justification. The trial’s own original sample-size assumptions are not reported in the registry record.

General explanation: Over-dispersion arises when outcome variability is greater than a statistical model assumes. In count outcomes, it can make standard errors too small if a model assumes less variability than the data contain. The primary outcome is a participant-level binary event, and no over-dispersion model is reported for it. The registry lists other outcomes involving stool counts and symptoms, but the statistical methods for those outcomes are not reported in the registry record.

11. Confidence Intervals and Precision

The primary arm-specific estimates have 90% confidence intervals of 31.5% to 62.7% for the vaccine group and 26.9% to 60.0% for the placebo group. The intervals are wide, reflecting substantial uncertainty around the percentages observed in the analyzed groups. The reported range for each arm includes values noticeably lower and higher than its point estimate.

The intervals overlap. Overlap is not a formal test of the group contrast, and the degree of overlap should not replace a direct comparison. The posted unadjusted comparison has a p-value of 0.755, while the sponsor’s separate vaccine-efficacy analysis reports a 90% confidence interval from -96.7 to 33.7. Each interval belongs to its particular estimate and method; they should not be combined or treated as if they describe the same quantity.

General explanation: Precision concerns the range of plausible values around an estimate, whereas statistical significance concerns compatibility with a particular null hypothesis under a specified test. A result can have a point estimate that appears meaningfully different while remaining imprecise, particularly when the number of participants or events is limited. Conversely, a narrow interval can still include values that matter to a decision. The width and location of the interval both matter, as does the clinical or policy threshold relevant to interpretation.

The primary efficacy denominators are 32 in the vaccine group and 28 in the placebo group. Those group sizes, together with the observed event frequencies, help explain why the estimates do not pin down a narrow range. The registry record does not report a prespecified precision target or an original sample-size justification, so it is not possible to determine from the posted result whether the study was designed to attain a particular confidence-interval width.

The precision should also be considered in light of the analysis population and study setting. The intervals describe the per-protocol challenge-study groups, not a broader target population. Confidence intervals quantify sampling uncertainty under their statistical method; they do not account for every possible source of bias, missing data, limits in generalizability or uncertainty about whether the case definition captures all relevant illness.

12. 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 illustrations give participants per arm for the binary outcomes. They are not total-event counts. The calculation 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 a stated fraction of the observed effect. The original trial’s own design assumptions and sample-size justification are not reported in the registry record.

The 80% power column gives the planned number of participants per arm for a hypothetical study to have an 80% chance of detecting the specified effect if that effect were the true one, under the calculation assumptions. The 90% power column gives the corresponding per-arm planning figure for a 90% chance. These are planning probabilities for a future study under a model, not statements about the certainty of the observed trial result.

The rows use the observed effect and smaller fractions of that effect. For each outcome, the planning figures increase as the effect to detect is reduced: a smaller difference is harder to distinguish from random variation, so more participants are needed under the stated assumptions. The illustrations are particularly sensitive to the observed estimate. When the observed effect is small or uncertain, planning directly from it can produce very large or unstable figures; the calculations do not establish which effect size should be considered important.

The figures are illustrations derived from the observed results, not the study’s original design plan. They do not adjust for attrition, departures from equal allocation, a different analysis method, multiplicity or other design considerations unless such adjustments are explicitly part of the stated assumptions. This is not observed power: the calculations do not measure the probability that the completed study detected an effect after seeing its results.

13. 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 = 0.429; treatment proportion p₁ = 0.469; average p̄ = 0.449.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.7034 = 1.3786.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.7028 = 0.5915.
4. n = (1.3786 + 0.5915)² / (0.0402)² = 2,404.2, rounded up to 2,405 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 <- 0.428571428571429                          # control proportion
p1 <- 0.46875                          # 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(3527173)
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 = 2,405 (80% power), 3,219 (90% power)
 75% of observed effect: n per arm = 4,266 (80% power), 5,711 (90% power)
 50% of observed effect: n per arm = 9,577 (80% power), 12,820 (90% power)
power.prop.test: 2404.2 
Simulated power at n = 2,405 per arm: 0.800 (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.

14. Analyses Reported by the Sponsor

The sponsor reported a vaccine efficacy rate of -9.4, with a 90% confidence interval from -96.7 to 33.7. The analysis is identified as non-inferiority and uses a one-sided Barnard Unconditional Exact Test; its reported p-value is 0.427. The registry record does not report the non-inferiority margin or all details required to independently assess the sponsor’s decision rule.

The separate recalculation compares the observed proportions in the two per-protocol groups using an unadjusted comparison and reports a p-value of 0.755. It is not the sponsor’s exact test, and it is not presented as the same vaccine-efficacy estimand. Differences in effect measure, test, confidence-interval construction or hypothesis can lead to different numerical summaries without making them direct substitutes.

The record does not flag a discrepancy between the posted sponsor result and the recalculation. The sponsor’s own analysis therefore remains the relevant reported analysis for its stated non-inferiority question, while the recalculation provides a simple descriptive comparison of the posted group results. Neither summary should be interpreted without the missing non-inferiority margin and complete analysis plan.

15. Multiplicity, Missing Data and Analysis Populations

The registry record lists 19 outcome measures and 18 posted secondary outcomes. It reports one sponsor analysis for the primary endpoint and provides several secondary outcome definitions and results. However, the statistical plan for handling multiplicity is not reported in the registry record. It does not state whether secondary hypotheses were arranged in a testing hierarchy, whether p-values were adjusted, or which results were intended to be confirmatory.

General explanation: Testing multiple outcomes increases the opportunity for chance findings. A prespecified family of hypotheses and an adjustment method can help control the overall false-positive risk. Without that information, unadjusted secondary p-values should be treated as descriptive and interpreted with particular caution. The primary endpoint is the clearest stated efficacy focus, but the registry record does not supply a complete multiplicity strategy for the full set of outcomes.

The record specifies the per-protocol set for primary efficacy, but does not report a parallel primary intention-to-treat analysis or an account of participants excluded from the efficacy population. The number and reasons for any exclusions, withdrawals or missing endpoint assessments are not reported in the registry record. Nor are the methods for handling missing binary outcomes, protocol deviations or incomplete challenge observations. These omissions limit assessment of how sensitive the result may be to analysis-population choices.

The analysis plan, including any planned interim analyses, is not reported in the registry record. The registry record also does not report whether endpoint definitions or analysis procedures changed after study initiation, or whether any changes occurred before or after unblinding. Without the protocol and statistical analysis plan, the degree to which reported analyses followed a prespecified plan cannot be assessed from the registry record alone.

The reported non-inferiority analysis has a one-sided test, but the margin and full decision rule are not reported in the registry record. These details are essential to interpret a non-inferiority claim. The absence of those details is distinct from the multiplicity issue: even a well-defined single test needs a specified margin, while a set of multiple tests also needs an account of how their combined error is handled.

16. 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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: carotid artery aneurysm (2.9%), deep vein thrombosis (2.9%), haematoma (2.9%), pelvic venous thrombosis (2.9%).

The safety summaries list 36 participants at risk for serious adverse events in the vaccine group and 35 in the placebo group. Serious adverse events were reported for 0 participants in the vaccine group and 1 participant in the placebo group. Deaths were reported for 0 participants in each group. The registry also lists four serious adverse event terms, each with a reported frequency of 2.9%: carotid artery aneurysm, deep vein thrombosis, haematoma and pelvic venous thrombosis.

These terms and percentages should be read as the registry’s serious-event summary, not as evidence that the vaccine caused an event. The record does not report a causal attribution for each term, the timing or duration of the events, or detailed clinical narratives. The at-risk denominators in this safety summary differ from the efficacy denominators because safety and efficacy analyses may use different populations; the registry record does not provide a participant-flow explanation for that distinction.

The summary is a description of events recorded in this study, not a complete characterization of safety in other populations or settings. A small controlled challenge study cannot by itself settle the frequency of uncommon events. The registry record does not report additional detail needed to assess event relatedness or the handling of adverse events beyond these counts and terms.

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

The results describe how often participants in the per-protocol challenge groups met a specific shigellosis case definition during the defined inpatient observation window. The observed percentages were 46.9% in the vaccine group and 42.9% in the placebo group. The confidence intervals are broad, and the recalculated unadjusted comparison does not provide a clear distinction between the groups. These findings leave considerable uncertainty about the magnitude and direction of any difference under this endpoint definition.

The results do not establish that the groups are equivalent. A non-significant comparison is not proof of no effect, and an equivalence or non-inferiority conclusion requires prespecified margins and a suitable analysis. Although the sponsor’s reported analysis is labelled non-inferiority, the registry record does not report its margin or complete decision criteria. The sponsor’s estimate and the simple unadjusted comparison answer different statistical questions and should be interpreted in their own terms.

The controlled human challenge setting supports a focused comparison under the study’s conditions, but it is not the same as observing naturally occurring exposure in everyday settings. Participants were healthy, non-immune adults within the recorded age range, and challenge used the S. sonnei 53G strain. The results do not establish effects in populations not represented by this study, nor do they describe outcomes beyond the protocol’s observation window.

The endpoint is a defined combination of shedding and illness criteria. It does not count every possible symptom or every instance of exposure as a primary event. Secondary endpoints provide other descriptions, but their statistical interpretation is limited by the missing multiplicity plan and incomplete reporting of their results.

In short, the posted data support a description of the recorded challenge outcomes and their uncertainty. They do not, on their own, establish a definitive conclusion about efficacy, equivalence, non-inferiority, safety in broader use or performance in different populations. Such conclusions would require the full protocol and analysis plan, complete participant accounting and evidence from settings appropriate to the intended population.

18. Why This Trial Matters Statistically

This study illustrates why a clinical trial result cannot be reduced to a single percentage or p-value. The primary endpoint is a protocol-defined binary outcome, its estimate has a confidence interval, and the same results can be expressed through different effect measures. Each summary answers a slightly different question. Reading the endpoint definition, analysis population and uncertainty together is essential to understanding what the comparison represents.

It also demonstrates the distinction between a sponsor’s inferential analysis and a simple recalculation from posted group summaries. The sponsor’s vaccine-efficacy analysis is labelled non-inferiority and uses a one-sided exact test. The recalculation is an unadjusted comparison of observed proportions. Those analyses should not be substituted for each other, and the missing non-inferiority margin is consequential for interpretation.

The study’s challenge design provides a controlled way to examine outcomes after a defined exposure, while also limiting generalization to other exposure settings and populations. The per-protocol efficacy set further makes the population represented by the primary result narrower than the full enrollment. These design and analysis features are part of the estimand, not minor technical details.

Finally, the registry record provides useful outcome and safety summaries but does not report several elements needed for a complete statistical appraisal: the original sample-size justification, full analysis plan, missing-data methods, multiplicity strategy and non-inferiority margin. The case is a practical example of how a careful reading can distinguish what the posted data show from what remains uncertain.

19. 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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The primary efficacy result comes from a per-protocol analysis, and the registry record does not provide a participant-flow account or explain exclusions from that set. The number and reasons for missing endpoint assessments, withdrawals, protocol deviations and exclusions are not reported in the registry record. Without those details, the potential influence of participant selection on the reported percentages cannot be fully assessed.

The study involved healthy, non-immune adults aged 18 to 50 years and a controlled challenge with S. sonnei 53G. This restricts direct generalization to people outside the recorded population or to settings with different exposure conditions. The endpoint observation period was limited to the challenge visit through the end of the inpatient stay, as defined in the registry. Longer-term outcomes are not reported in the registry record.

The primary endpoint depends on a composite case definition involving bacterial shedding and specified illness criteria. The registry gives the definition, but does not provide additional information on endpoint adjudication, assessment procedures or blinding of those assessing the endpoint. Those operational details are not reported in the registry record.

The sponsor’s analysis is identified as non-inferiority, but its margin and complete decision rule are not reported in the registry record. The registry also does not report the original sample-size justification, planned precision, interim analyses, multiplicity strategy or the complete statistical analysis plan. These gaps limit assessment of prespecification and of the confirmatory status of secondary outcomes.

The safety summary reports counts and serious-event terms but does not give detailed narratives, event timing or causal attribution. The trial’s size and setting also constrain what can be learned about uncommon safety events. These limitations do not invalidate the posted summaries, but they define the scope of conclusions that can be drawn from them.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study to Evaluate the Efficacy, Safety and Immunogenicity of a Vaccine Designed to Protect Against Infection With Shigella Sonnei in Healthy Adults — NCT03527173.
• Registry record last updated 2020-07-28; status last verified 2020-07.
• Results first posted 2020-05-15.
• 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.

23. Record Summary

NCT03527173 is a completed, randomized, quadruple-masked, parallel-group Phase 2 human challenge study sponsored by GlaxoSmithKline. It enrolled 71 healthy adults; the arm descriptions specify an age range of 18 to 50 years at first vaccination. Participants received two doses of S. sonnei vaccine or placebo, followed by challenge with S. sonnei 53G.

The primary endpoint was the percentage meeting the protocol primary case definition for shigellosis from Day 57 through Day 64. In the per-protocol efficacy analysis, the reported percentages were 46.9% for the vaccine group and 42.9% for placebo. The estimates have broad 90% confidence intervals. The sponsor separately reported a non-inferiority vaccine-efficacy analysis, but the margin is not reported in the registry record.

The registry record does not report the full protocol or statistical analysis plan, the original sample-size justification, missing-data methods or multiplicity strategy. These omissions and the controlled challenge population should be considered when interpreting the efficacy and safety summaries.