A Study to Evaluate Safety, Tolerability, & Immunogenicity of Multiple Formulations of BNT162b2 Against COVID-19 in Healthy Adults: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Evaluate Safety, Tolerability, & Immunogenicity of Multiple Formulations of BNT162b2 Against COVID-19 in Healthy Adults (NCT04816669), a randomized phase 3 study. The primary outcome is Geometric Mean Titers (GMTs) of Full-Length S-Binding IgG Concentrations of Lyophilized Formulation SDVs and Frozen-Liquid Formulation in MDVs 1 Month After Dose 2.

2. Clinical Question

The registry describes a broader study than this primary comparison. It also describes safety and immunogenicity assessment of frozen-liquid formulations with lipid nanoparticle size at the upper end of specification and of a ready-to-use formulation. Separately, it describes an additional frozen-liquid dose for participants who had already received a two-dose lyophilized series. These components do not change the primary endpoint’s specified comparison: the lyophilized and frozen-liquid formulations after dose 2.

The registry identifies COVID-19 and SARS-CoV-2 infection as conditions. The primary outcome is an antibody measurement, not a clinical disease endpoint. It therefore addresses a difference in measured immune response between formulations in the defined population; it does not directly estimate comparative rates of COVID-19 or SARS-CoV-2 infection.

3. Trial Design

This was a Phase 3 randomized, observer-blind, parallel study. The registry lists triple masking. The study title and summary describe an evaluation of safety, tolerability, and immunogenicity of multiple BNT162b2 formulations in healthy adults 18 through 55 years of age. The posted enrollment is 629.

The primary comparison was between lyophilized BNT162b2 in SDVs and frozen-liquid BNT162b2 in MDVs. The latter is identified as the control for the lyophilized formulation. The study summary describes a two-dose schedule separated by 21 days and a dose of 30µg. The reported primary endpoint is measured one month after dose 2. The study was to be conducted in the United States, and the expected duration for each participant was approximately 2 months with 3 visits in total.

Other study components covered frozen-liquid BNT162b2 with lipid nanoparticle size at the upper end of specification, ready-to-use BNT162b2, and a frozen-liquid dose offered as dose 3 after a primary series of lyophilized BNT162b2. These are distinct from the main two-formulation comparison. The registry lists an overall enrollment of 629, while the primary immunogenicity results have group counts of 238 for lyophilized SDV and 226 for frozen-liquid MDV. The registry’s primary outcome population has additional eligibility, dosing, testing, and sample-timing conditions, so the outcome group counts should be read in that context rather than assumed to represent all enrolled participants.

The record gives a start date of 2021-04-01, a primary completion date of 2021-12-01, and a completed status. It does not report the randomization sequence or allocation ratio. The protocol’s detailed operational procedures and the rationale for the enrollment target 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: This study was conducted in two parts: Part 1 and Part 2. Participants who received 2 doses of lyophilized formulation of BNT162b2 in Part 1, were offered to receive Dose 3 of frozen-liquid formulation.

Part 1

Reasons for not completing this period:

Part 2

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Evaluable immunogenicity population: Participants who were eligible and randomized, received 2 doses of vaccine, with Dose 2 received within 19 to 42 days after Dose 1, had at least 1 valid immunogenicity result within 28 to 42 days after Dose 2, were negative for both SARS-CoV-2 tests at both Day 1 and 1 month post-Dose 2 visits, and had no other important protocol deviations as determined by clinician.

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.

Population: Safety population set included all randomized participants who received at least 1 dose of the study intervention.

6. Primary Endpoint

The registry lists the primary outcome as Geometric Mean Titers (GMTs) of Full-Length S-Binding IgG Concentrations of Lyophilized Formulation SDVs and Frozen-Liquid Formulation in MDVs 1 Month After Dose 2, measured over 1 Month after Dose 2.

The primary endpoint is the geometric mean concentration (GMC) of full-length S-binding IgG, reported one month after dose 2 for the lyophilized SDV and frozen-liquid MDV formulations. The unit is unit per milliliter. Although the outcome title refers to geometric mean titers, its description says the descriptive results are reported as GMCs. The registry also reports a geometric mean ratio (GMR) comparing the lyophilized formulation with the frozen-liquid formulation.

Antibody concentrations can vary across a wide range and are commonly analyzed on a logarithmic scale. A geometric mean summarizes values after taking logarithms, averaging those transformed values, and converting the result back to the original scale. This makes the summary suitable for multiplicative comparisons: the GMR describes the lyophilized group’s GMC relative to the frozen-liquid group’s GMC. This is a general explanation of the statistical measure, not a claim about the clinical meaning of a particular antibody level.

The registry says GMCs and confidence intervals were calculated by exponentiating the mean logarithm of concentrations and corresponding confidence limits based on the Student t distribution. Results below the assay’s lower limit of quantification (LLOQ) were set to 0.5 times the LLOQ. That substitution is part of the stated analysis procedure and affects how values below the assay’s quantification limit enter the summaries.

This endpoint measures an immune-response marker at a specified post-dose time. It does not directly measure infection, symptomatic disease, or duration of protection. The registry record does not report a clinical threshold that would translate a particular GMC or GMR into a clinical outcome.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a geometric mean summarize?

For a set of positive measurements, a geometric mean is obtained by taking the logarithm of each value, averaging those logarithms, and exponentiating the average. In general, this approach is useful when measurements are naturally interpreted on a multiplicative scale or are strongly right-skewed. It summarizes the center of the measurements in a way that is less dominated by very large values than an arithmetic mean can be. The registry states that the GMCs were calculated by exponentiating the mean logarithm of the concentrations.

The posted GMCs are group summaries, not individual-level results. They do not show how much participants’ measurements varied within each group. The registry’s stated confidence intervals describe uncertainty in the estimated group GMCs; they are not intervals containing a specified proportion of individual participants’ antibody concentrations.

How should the geometric mean ratio be read?

A ratio compares two quantities by dividing one by the other. For this analysis, the registry identifies the GMR as the ratio of GMCs for BNT162b2 30µg lyophilized SDV and frozen-liquid MDV. In general, a ratio of 1 represents equal group summaries; a ratio below 1 indicates a lower numerator summary, and a ratio above 1 indicates a higher numerator summary. The reported GMR is 0.68, with a 95% confidence interval from 0.60 to 0.77. The direction depends on the stated ordering of the groups, so the lyophilized formulation must be read as the numerator in this comparison.

A confidence interval gives a range of estimates compatible with the data and the statistical model at the stated confidence level. It is not a guarantee that the true value lies in the particular interval, nor does it describe the probability that the treatment effect is in that interval. Its width reflects statistical precision under the model and analysis. Interpretation of a noninferiority confidence interval also requires the prespecified noninferiority margin and decision rule; those are not reported in the registry record.

Why analyze logarithms?

Logarithmic analysis converts multiplicative differences into additive differences on the transformed scale. This can make statistical methods designed for means more appropriate for data whose variation is proportional to their magnitude. Exponentiating the estimated mean and interval returns the results to the original concentration scale. This is a general explanation of the method. The registry specifically says that the GMC and confidence interval were calculated by exponentiating mean logarithms and corresponding confidence limits based on the Student t distribution.

The Student t distribution is commonly used to account for uncertainty in estimating a mean when the population variability is estimated from sample data. The registry identifies this distribution for the confidence limits. It does not report additional modeling details, diagnostics, or adjustments for covariates in the primary outcome analysis.

How were results below quantification handled?

The lower limit of quantification is the assay boundary below which the measurement cannot be quantified reliably according to the stated laboratory procedure. The registry says results below the LLOQ were set to 0.5 times the LLOQ. This is a substitution rule: rather than excluding those values, the analysis assigns them a fixed value relative to the assay limit. In general, such a rule allows the analysis to retain below-limit observations, but the resulting estimates can depend on how many values fall below the limit and on the chosen substitution. The registry record does not report the number of below-LLOQ results or alternative analyses using other handling rules.

What does the evaluable population definition imply?

The primary outcome population was restricted to eligible and randomized participants who received two doses, received dose 2 within 19 to 42 days after dose 1, had at least one valid immunogenicity result within 28 to 42 days after dose 2, and were negative for both SARS-CoV-2 tests at both Day 1. These rules define which participants contribute to the primary immunogenicity comparison. Such restrictions can make the analysis target more specific than the full randomized population. They also mean that the result may not describe participants who lacked a valid sample, did not meet the dose timing condition, or did not meet the stated testing condition.

The record does not report how many participants were excluded for each reason, how missing immunogenicity measurements were handled beyond the evaluable-population definition, or whether a separate analysis used a broader population. It also does not report a protocol-based sample-size justification or details of interim analyses. Those omissions limit assessment of how the analysis plan was established and how robust the result is to alternative analytic choices.

What is the role of noninferiority analysis?

In general, a noninferiority analysis asks whether the data exclude a loss larger than a prespecified acceptable margin. The margin and the direction of the comparison are essential: a ratio and confidence interval cannot be judged against a criterion without knowing that boundary. A confidence interval wholly on one side of a boundary may support a conclusion under a specified rule, while an interval spanning the boundary may not. This is a general description; the registry labels the reported comparison as noninferiority but does not state the margin or the exact rule. Consequently, the reported estimate can be interpreted numerically, but the criterion cannot be independently checked from the registry record.

10. Analyses Reported by the Sponsor

The registry reports the sponsor’s analysis as a GMR of 0.68, with a 95% confidence interval from 0.60 to 0.77, and labels the analysis noninferiority. It also reports the two group GMCs and their confidence intervals. These are the posted estimates to use when describing the sponsor’s reported result.

The registry record does not report a separate recalculated GMR or confidence interval for this comparison. It therefore does not present two numerical analyses that can be compared for agreement. The record does not flag a discrepancy between the sponsor’s analysis and a recalculation. The sponsor’s reported analysis remains the available inferential result in the registry.

The registry does not state the noninferiority margin or the decision rule. That omission matters because the confidence interval’s relation to that margin determines how a noninferiority conclusion is assessed. The estimate and interval can be read as posted, but they do not independently establish whether the study met its prespecified criterion.

11. Multiplicity, Missing Data and Analysis Populations

The study includes several formulations and study components, and the registry lists 16 outcome measures and 2 analyses. The primary comparison concerns the lyophilized SDV and frozen-liquid MDV formulations. The secondary outcomes listed include baseline GMCs and geometric mean fold rises from baseline to one month after dose 2 for full-length S-binding IgG in Part 1.

When a study evaluates multiple outcomes, time points, groups, or analyses, the chance of finding at least one apparently notable result can increase if each analysis is interpreted independently. General approaches to multiplicity include prespecifying a hierarchy, adjusting significance thresholds, or defining a limited set of confirmatory comparisons. The registry record does not report whether any multiplicity procedure was planned or used, nor does it describe an outcome hierarchy beyond identifying the primary outcome. No adjustment should be assumed from the number of posted outcomes alone.

The primary outcome population is specifically defined by eligibility and randomization, receipt and timing of two doses, availability and timing of a valid immunogenicity result, and negative results for both SARS-CoV-2 tests at Day 1. The registry provides group counts for the primary immunogenicity results but does not provide a participant flow by exclusion reason. The handling of missing outcomes beyond defining the evaluable population is not reported in the registry record. The number of missing measurements, reasons for missingness, and any sensitivity analyses are also not reported in the registry record.

The record does not state whether the analysis was prespecified in a statistical analysis plan, whether interim analyses occurred, or what procedures were used to control type I error for the noninferiority question. The noninferiority margin is not reported in the registry record. These details are needed to assess the confirmatory status of the result and the effect of multiple analyses on its interpretation.

12. 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 non-serious adverse event terms

The registry reports no serious adverse events in the listed groups: 0 of 279 participants in the lyophilized SDV group, 0 of 280 in the frozen-liquid MDV group, 0 of 35 in the frozen-liquid group with lipid nanoparticle size at the upper end of specification, 0 of 35 in the ready-to-use group, and 0 of 114 in the frozen-liquid dose-3 group. The registry also reports zero deaths in each of these groups.

These are observed counts in the safety groups reported by the registry. A count of zero means that no serious adverse event was recorded in the listed group during the safety reporting represented in the record; it does not establish that the underlying risk is zero. Rare events can remain unobserved in a study, and the precision of a safety estimate depends on the amount and duration of observation. The registry’s listed counts alone do not quantify that uncertainty.

The record also gives counts for participants with other adverse events, but those are separate from serious adverse events. The registry’s serious adverse event terms list is empty. The detailed observation period and the definitions or ascertainment procedures for these safety summaries are not reported in the registry record. The group counts should therefore be read as descriptive safety information, not as a complete comparative assessment of risk across formulations.

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

The primary results show that, in the defined evaluable immunogenicity population, the reported GMC of full-length S-binding IgG one month after dose 2 was 4797.0 unit per milliliter for lyophilized SDV and 7032.0 unit per milliliter for frozen-liquid MDV. The reported GMR, ordered as lyophilized relative to frozen-liquid, was 0.68, with a 95% confidence interval from 0.60 to 0.77. These numbers describe the measured antibody endpoint and its statistical uncertainty at the specified assessment time.

The result does not, on its own, establish whether the noninferiority criterion was met, because the registry record does not report the noninferiority margin or decision rule. Nor does an antibody comparison directly establish comparative effects on COVID-19 or SARS-CoV-2 infection. The primary endpoint is an immunogenicity measure, not a clinical disease outcome.

The result also does not establish how antibody levels change after the reported assessment, how durable the response is, or whether the comparison applies to populations outside the healthy adults 18 through 55 years of age described in the study. The registry’s evaluable population criteria further limit interpretation to participants meeting the stated dose, timing, sample, and Day 1 test conditions.

The broader registry record includes safety information and additional formulation groups, including a dose-3 group. Those results are not substitutes for the primary two-formulation comparison. The absence of reported serious adverse events in the listed groups is descriptive and does not prove absence of risk. Taken together, the posted results provide a specific antibody comparison and safety counts, while important elements needed for broader clinical conclusions are not reported in the registry record.

14. Why This Trial Matters Statistically

This trial raises a statistical question that is different from asking whether two formulations have exactly the same measured antibody response. Its primary analysis is labeled noninferiority, where interpretation depends on a prespecified margin and decision rule. The registry reports a ratio and confidence interval but does not report that margin. This illustrates why an effect estimate alone is not enough to verify a noninferiority conclusion.

The outcome also illustrates the distinction between an immunogenicity measure and a clinical endpoint. GMCs and their ratio summarize measured antibody concentrations in a defined population at a specified time. They do not directly estimate differences in infection or disease outcomes. Statistical interpretation should remain aligned with the endpoint actually reported.

The geometric mean and ratio provide summaries suited to a multiplicative comparison, while the confidence intervals communicate uncertainty around those estimates. At the same time, the evaluable-population rules define a narrower analysis group than the overall enrollment. Readers need both pieces of information—the summary method and the population definition—to understand what the estimates represent.

Finally, the record includes multiple formulations and safety groups but limited detail about multiplicity, missing data, and the analysis plan. That combination makes clear the value of distinguishing a posted numerical result from the broader evidence needed to assess its confirmatory status, robustness, and clinical scope.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record does not report the noninferiority margin or the precise decision rule, so the noninferiority label cannot be independently evaluated from the posted ratio and confidence interval alone. It also does not report the statistical analysis plan, any interim analyses, or any procedures for multiplicity control. These omissions limit assessment of how the comparison and its interpretation were specified.

The primary outcome applies to an evaluable immunogenicity population with defined vaccination timing, sample timing, and Day 1 test requirements. Although the registry reports the outcome group counts, it does not report a breakdown of exclusions or missing samples. The effect of these restrictions, and the sensitivity of the estimates to alternative population definitions, is therefore not assessable from the registry record.

The registry states that below-LLOQ results were assigned 0.5 times the LLOQ but does not report how many measurements were below the limit or whether results were examined under alternative handling rules. It also does not report further model details or diagnostics. Such information would help readers assess the influence of assay limits and analytic assumptions.

The primary endpoint is a laboratory measure at one month after dose 2. The registry result does not establish longer-term antibody patterns or comparative clinical outcomes. The record describes healthy adults 18 through 55 years of age, so conclusions should not be extended automatically to other populations. Safety information is presented as event counts, but the detailed observation period and ascertainment procedures are not reported in the registry record.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study to Evaluate Safety, Tolerability, & Immunogenicity of Multiple Formulations of BNT162b2 Against COVID-19 in Healthy Adults — NCT04816669.
• Registry record last updated 2022-12-23; status last verified 2022-11.
• Results first posted 2022-12-23.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

NCT04816669 is a completed Phase 3 randomized, observer-blind, parallel study sponsored by BioNTech SE. It enrolled 629 participants and evaluated multiple BNT162b2 formulations in healthy adults 18 through 55 years of age. The primary immunogenicity comparison was between lyophilized SDV and frozen-liquid MDV formulations one month after dose 2.

The reported GMCs were 4797.0 unit per milliliter for lyophilized SDV and 7032.0 unit per milliliter for frozen-liquid MDV. The reported GMR was 0.68, with a 95% confidence interval from 0.60 to 0.77. The registry labels the analysis noninferiority but does not report the margin or decision rule.

The record also lists safety counts for five groups, with zero serious adverse events and zero deaths reported in each. The primary endpoint population is restricted by dose timing, valid sample timing, and Day 1 testing requirements. Details on multiplicity, missing-data handling, and the statistical analysis plan are not reported in the registry record.