Mavrilimumab to Reduce Progression of Acute Respiratory Failure in COVID-19 Pneumonia and Systemic Hyper-inflammation: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Mavrilimumab to Reduce Progression of Acute Respiratory Failure in COVID-19 Pneumonia and Systemic Hyper-inflammation (NCT04463004), a randomized phase 2 study. The primary outcome is Proportion of Subjects Alive and Off of Oxygen at Day 14.

2. Clinical Question

The clinical question is whether early mavrilimumab treatment changes the chance of being alive and no longer requiring oxygen at the specified assessment time, compared with placebo, among people with severe COVID-19 pneumonia and hyper-inflammation. The endpoint combines survival and respiratory status: a participant must meet both conditions to be counted as having the outcome. It is therefore not a measure of oxygen use alone or survival alone.

The registry describes a prospective, multicenter, blinded, randomized, placebo-controlled Phase 2 study. Its brief summary states that the purpose was to demonstrate prevention of progression of respiratory failure. The posted primary endpoint is narrower and assessed at day 14. A secondary outcome is listed as the number of subjects alive and without respiratory failure at 28 days, but its result is not reported in the registry record.

3. Trial Design

The registry record identifies NCT04463004 as a Phase 2 randomized trial sponsored by Virginia Commonwealth University. It describes the study as prospective, multicenter, blinded and placebo-controlled. Masking is recorded as quadruple, and the design model is recorded as factorial. The record lists mavrilimumab as the intervention and placebo as the control.

The listed study start date is 2020-09-02, and the primary completion date is 2020-11-30. The study status is completed. The enrollment field lists 2. The posted primary outcome data list one participant in the intervention arm and one in the control arm. These sparse data do not support a reliable comparison of outcomes between treatments.

Although the model is labeled factorial, the registry record does not report the factorial factors, their levels, or how they were assigned. It lists treatment infusion and placebo infusion as the arm descriptions. Further details of the randomization procedure, allocation concealment, treatment schedule, sites, protocol deviations and follow-up procedures are not reported in the registry record.

The primary outcome is titled “Proportion of Subjects Alive and Off of Oxygen at Day 14,” with a timeframe of 14 days. The posted measure is a count of participants, not a comparative estimate. The secondary outcome listed is the number of subjects alive and without respiratory failure at 28 days; its numerical result is not reported in the registry record. There are no posted analyses in the record. The analysis population, analysis methods and handling of missing outcome data are not reported in the registry record.

A Phase 2 label describes the trial stage, but it does not establish how the study was designed to estimate an effect or how precise its estimates would be. The registry record does not report a protocol-based sample-size justification, a planned primary comparison, or a detailed statistical analysis plan. The posted information therefore supports a description of the design and recorded outcomes, but not a full reconstruction of the intended 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.

Overall Study

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 Proportion of Subjects Alive and Off of Oxygen at Day 14, measured over 14 days.

The primary endpoint is the proportion of subjects who are alive and off oxygen at day 14. It combines two clinically relevant conditions into a single binary outcome: a participant either meets both conditions at the assessment or does not. A count of participants is posted for each arm. The endpoint title uses the word “proportion,” while the reported measure is identified as a participant count; the registry does not post an accompanying percentage or comparative estimate.

An endpoint combining survival and oxygen status can summarize whether participants have reached a specified state by a fixed assessment time. It differs from an endpoint that measures only survival, only oxygen use, or the time until either event. It also does not describe how quickly a participant improved before day 14 or what happened after that assessment.

In general, a binary endpoint allows comparison of the frequency of an outcome between randomized groups. Such a comparison requires interpretable denominators and enough outcome information to estimate uncertainty. Here, the record lists a single participant in each arm and a value of 1.0 for each. A comparative effect and its precision are not estimable from the posted summary. The endpoint remains the registered primary measure, but the observed counts do not establish a treatment effect.

The record lists a separate 28-day secondary outcome concerning survival without respiratory failure. Its result is not reported in the registry record, so it cannot clarify the day-14 finding or provide a longer-term comparison.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a binary endpoint measure?

In general, a binary endpoint classifies each participant into one of two outcome categories at a defined assessment: the outcome is met or it is not met. For this trial, the primary outcome requires that a participant be alive and off oxygen at day 14. The registry record describes the measure as a count of participants. A count is not itself a treatment-effect estimate; interpretation also depends on the denominator and on a suitable comparison between randomized arms.

Common ways to express a binary comparison include differences in outcome frequencies or ratios of those frequencies. Each approach answers a somewhat different question. Any such comparison also needs an assessment of uncertainty, often expressed with a confidence interval. A confidence interval describes the range of effect sizes compatible with the observed data under the statistical method used. It is not a guarantee that the true effect lies in that range, and its usefulness depends on the quality and amount of information in the data. No comparative estimate or confidence interval is reported in this record.

Why is the two-arm comparison not estimable?

The record posts one participant in each arm and a value of 1.0 in each arm. It also explicitly indicates that the binary two-arm comparison is not estimable from the posted values. With so little information, a numerical difference could be driven entirely by the observation of an individual participant. An estimate calculated from sparse counts can be unstable, and standard methods may not yield an informative measure of uncertainty. The registry record does not report an alternative analysis or a reason for the non-estimability.

More generally, sparse binary data can create difficulties for statistical models and tests. Some methods rely on approximations that work better when information is less limited. Other methods can be designed for sparse outcomes, but their choice depends on the planned estimand, trial design and data structure. The appropriate method cannot be identified from this record: the statistical analysis plan and the specific comparative method are not reported in the registry record. No new calculation should be substituted for the sponsor's analysis; in this instance, no sponsor analysis is posted.

What does randomization contribute?

In general, randomization assigns participants to treatment groups by chance and can reduce systematic differences between groups at the start of a trial. Its value is strongest when the trial has enough outcome information to make a meaningful comparison. Randomization does not make an extremely limited set of observations precise, and it does not eliminate chance imbalances in small studies. The record identifies the allocation as randomized but does not report the randomization sequence, allocation concealment or stratification details.

The record also labels the model factorial. In general, a factorial design studies more than one intervention factor within the same trial structure and may allow estimates of separate effects and, where planned, interactions. The factors, their levels and the intended analysis of any interaction are not reported in the registry record. The posted arm descriptions identify an intervention infusion and a placebo infusion, but do not give enough detail to explain the factorial label. It would be inappropriate to infer the structure or analyses beyond what is registered.

How should planning calculations be read?

General sample-size planning estimates describe how many participants might be needed under stated assumptions to detect a target effect with a chosen statistical approach. They are planning illustrations, not observed results and not evidence about the probability that this trial detected an effect. The stated assumptions for the planning illustration are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect defined as a stated fraction of the observed effect. The registry record does not report numerical planning estimates or the underlying observed effect needed to interpret them.

In general, planning estimates can change substantially when the assumed effect size, outcome frequency, allocation, significance level or method changes. If an observed effect is very uncertain, using it as the basis for planning can also produce illustrations that are themselves uncertain. These calculations do not repair limitations in the trial's posted outcome data, and they should not be described as observed power. No numerical sample-size conclusion can be drawn from the information reported in the registry record.

What can be said about precision and analysis choices?

Precision concerns how narrowly an analysis estimates the treatment effect. A larger, informative dataset generally allows a more precise estimate than a dataset with very few observations, although design and outcome characteristics also matter. In this trial, the record does not provide a comparative estimate, standard error or confidence interval. Precision therefore cannot be quantified from the posted primary outcome.

The registry record lists no analyses. It does not report whether the intended analysis would be based on assigned treatment, treatment received, or another analysis population. It also does not describe adjustments, interim analyses, covariates, multiplicity procedures or missing-data methods. These omissions prevent readers from evaluating whether the posted results follow a prespecified analysis strategy. General statistical principles can explain what such methods are for, but they cannot establish which methods this trial intended to use.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Intervention", "Control")
n      <- c(1, 1)
events <- c(1, 1)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Intervention  1  1  100.0  2.5  100.0  20.7  100.0  15.8  98.7
ARM  Control  1  1  100.0  2.5  100.0  20.7  100.0  15.8  98.7

11. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates multiple outcomes, comparisons, time points or analyses. In general, examining many results can increase the chance of finding an apparently notable result by chance, so a prespecified analysis plan may explain how the primary question is distinguished from secondary or exploratory questions. This record lists a primary outcome at day 14 and a secondary outcome at 28 days. It does not report a multiplicity strategy, hierarchy of testing or other adjustment. The planned handling of multiple comparisons is therefore not reported in the registry record.

The record lists one secondary outcome: the number of subjects alive and without respiratory failure at 28 days. Its result is not reported in the registry record. It cannot be used to make a secondary comparison or to corroborate the primary endpoint.

Missing-data handling is also not described. The record does not state whether any endpoint values were missing, how missing status would be classified, or whether sensitivity analyses were planned. The posted primary values alone do not establish how missing observations were handled. The analysis population is not reported in the registry record, so it is unclear which participants were intended to contribute to each analysis.

These are reporting limitations, not proof that particular methods were or were not used. Without a posted analysis plan or analysis results, readers cannot determine whether decisions about outcomes, multiplicity, missing data or analysis populations were prespecified. The registry record lists zero analyses, and no primary or secondary comparative analysis is available to assess.

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 5% reporting threshold.

The adverse-event records list one participant at risk in each arm. In both arms, the count of participants with a serious adverse event is zero, the count with another adverse event is zero, and the death count is zero. These are descriptive counts from the registry record.

A count of zero in a very small group does not show that an adverse event cannot occur or establish that a treatment is safe. It also does not support a reliable comparison of adverse-event rates between mavrilimumab and placebo. The record does not provide detailed event descriptions, exposure duration, severity grades, causal assessments or a fuller safety analysis. The list of top serious adverse-event terms is empty, consistent with no serious events listed in these records, but does not add information about risks that could not be observed in this dataset.

Safety interpretation should remain separate from the primary efficacy endpoint. The primary outcome concerns being alive and off oxygen at day 14; adverse-event counts address a different aspect of the trial. Neither the sparse event counts nor the absence of a posted comparative analysis supports a general conclusion about treatment safety.

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

The registry record documents a randomized, quadruple-masked Phase 2 study of mavrilimumab versus placebo in severe COVID-19 pneumonia with hyper-inflammation. For the primary endpoint, it posts a participant count of 1.0 in each arm, with one participant listed in each arm. The record states that a two-arm comparison of the binary endpoint is not estimable. These data describe what was entered in the record; they do not quantify a treatment effect.

The result does not establish that mavrilimumab improved the chance of being alive and off oxygen at day 14, and it does not establish that the treatment had no effect. Equal posted counts do not demonstrate equivalence. Demonstrating equivalence or excluding important differences requires a suitable design and an informative estimate with adequate precision; neither is available in the posted results.

The result also does not answer the listed 28-day secondary question. That outcome's numerical result is not reported in the registry record. Similarly, the adverse-event counts do not establish the absence of risks or a comparative safety profile. They are limited observations with very little information about the range of possible outcomes.

In general, trial results should be interpreted in light of the design, outcome definition, analysis population, statistical method and uncertainty around estimates. Several of those details are not reported in this registry record, and no analyses are posted. The record does not report a confidence interval, p-value, adjusted estimate or unadjusted comparative estimate. Consequently, readers cannot assess the magnitude or precision of a treatment contrast.

The study is recorded as completed, but that status does not change what can be inferred from its posted data. The appropriate conclusion is that the registry lists sparse arm-level counts and no estimable two-arm comparison. These results are insufficient for a conclusion about efficacy or comparative safety and should not be treated as proof for or against a treatment effect.

14. Why This Trial Matters Statistically

This trial illustrates an important distinction between a registered clinical question and an answer that can be estimated from posted results. The question is clinically specific: whether participants are alive and off oxygen at day 14. But a clearly defined endpoint does not by itself ensure that the data can support a useful comparison. The registry posts only one participant in each arm and reports that the binary comparison is not estimable.

Statistically, this highlights how sparse outcome data limit both effect estimation and precision. In general, randomization helps make treatment groups comparable in expectation, but it does not create information when very few outcomes are recorded. An estimate based on an extremely small amount of information may be highly unstable, and a lack of a comparative result should not be mistaken for evidence of no difference.

The record also illustrates why registry entries matter. The design is labeled factorial and quadruple-masked, but the factorial structure and statistical analysis plan are not reported. The primary endpoint is described as a proportion while the posted measure is a participant count. No analyses are posted, and the 28-day secondary outcome has no reported result. These gaps limit independent assessment of how the trial's question was intended to be analyzed and what the results can support.

Finally, the safety counts demonstrate the difference between observing no events in a small dataset and establishing that risks are absent. The registry record is useful for identifying the trial question and the limits of the reported results. Its statistics do not provide a dependable estimate of treatment effects or comparative harms.

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 main limitation for interpretation is the small amount of posted outcome information: one participant is listed in each arm, and the primary binary comparison is not estimable. The record provides counts but no comparative effect estimate or measure of uncertainty. The title describes a proportion, whereas the posted measure is a participant count, and no percentage is reported.

Additional trial details needed to evaluate the analysis are not reported in the registry record. These include the eligibility criteria in detail, baseline participant characteristics, the factorial factors and their levels, the randomization procedure, the planned statistical methods, the analysis population and the handling of missing outcome data. The record does not state whether interim analyses or covariate adjustments were planned. Their use or absence cannot be inferred.

The listed 28-day secondary outcome has no numerical result reported in the registry record. No analyses are posted. The adverse-event counts are also very limited and do not include fuller descriptions or a comparative safety analysis. The completed status and multicenter description do not resolve these information gaps.

Finally, the registry record does not report a protocol-based sample-size justification or numerical planning estimates. Its planning assumptions refer to a normal approximation and a fraction of the observed effect, but the numerical illustrations and the effect used are not reported. These limitations restrict interpretation to the posted design descriptions and sparse outcome counts; they do not allow a reliable estimate of treatment benefit, lack of benefit or risk.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Mavrilimumab to Reduce Progression of Acute Respiratory Failure in COVID-19 Pneumonia and Systemic Hyper-inflammation — NCT04463004.
• Registry record last updated 2021-08-11; status last verified 2021-08.
• Results first posted 2021-08-11.
• 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

NCT04463004 is recorded as a completed Phase 2 randomized, quadruple-masked, factorial-model study sponsored by Virginia Commonwealth University. The registry describes a prospective, multicenter, placebo-controlled trial of mavrilimumab infusion in severe COVID-19 pneumonia with hyper-inflammation. The listed control received placebo infusion.

The primary endpoint was being alive and off oxygen at day 14. The registry posts a participant count of 1.0 in each arm, with one participant listed in each, and indicates that a binary two-arm comparison is not estimable. No analyses are posted. A 28-day secondary outcome is listed, but its result is not reported in the registry record. The adverse-event records list zero serious adverse events, zero other adverse events and zero deaths in each arm, with one participant at risk per arm. These sparse counts do not support reliable conclusions about comparative efficacy or safety.