Convalescent Plasma vs. Standard Plasma for COVID-19: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Convalescent Plasma vs. Standard Plasma for COVID-19 (NCT04344535), a randomized phase 1, phase 2 study. The primary outcome is 28 Day Ventilator Free Days.

2. Clinical Question

The trial asked whether transfusing plasma from people who had recovered from COVID-19, containing anti-SARS-CoV-2 antibodies, was safe and could treat hospitalized patients with COVID-19. The registry’s description proposes that antibodies transferred in plasma might neutralize virus or reduce the chance of worsening illness; that is the study rationale, not a finding established by the posted results.

The primary outcome is a patient-centered measure that combines survival and time without invasive ventilation during the stated period. The registry also lists 90-day all-cause mortality as a secondary outcome. The posted record therefore addresses a short-term ventilation-related measure and a later mortality measure, but it does not report a complete account of the trial’s analysis plan or a sponsor-reported comparative analysis for the primary outcome.

3. Trial Design

This was a randomized, quadruple-masked, parallel-group trial in the registry’s Phase 1 and Phase 2 categories. It compared convalescent donor plasma with standard donor plasma among hospitalized patients with COVID-19. The listed sponsor is Stony Brook University. The study began on 2020-04-08, and the primary completion date is listed as 2021-01-01. Its current registry status is terminated.

Overall enrollment is listed as 82. The posted primary outcome summaries show 59 participants in the convalescent plasma arm and 15 in the standard plasma arm. The registry does not explain in the results how those arm counts relate to overall enrollment, or provide a participant flow account explaining who was randomized, treated, or included in each analysis. Although the primary outcome population is described as all randomized, the arm summaries alone do not resolve that accounting.

Quadruple masking indicates that the record classifies the study as masked at four levels, but the identities of the masked groups are not reported in the registry record. The allocation ratio, the sequence-generation procedure, how treatment assignments were concealed, and details of plasma preparation or administration are also not reported in the registry record. The arm descriptions themselves are blank. These omissions limit how fully the trial procedures can be reconstructed from the record.

The record lists two outcome measures and one posted secondary outcome, but no sponsor-reported analyses. It does not report the reason the study was terminated, whether recruitment ended as planned, or whether any protocol changes occurred. Those points should remain open rather than being inferred from the status or dates. The design establishes that the investigators intended a randomized comparison; it does not, by itself, establish how completely the enrolled participants contributed to the posted summaries.

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: 8 screen failures prior to randomization, thus 74 patients were randomized.

Overall Study

Analysis population for the primary outcome, as defined in the registry: All randomized

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 28 Day Ventilator Free Days, measured over 28 days post randomization.

The primary endpoint is 28-day ventilator-free days, measured through 28 days after randomization. It counts days a participant was not receiving invasive mechanical ventilation during that period. Under the registry definition, a participant who died during the period is assigned zero ventilator-free days. This makes the endpoint a composite summary of survival and freedom from invasive ventilation rather than a measure of ventilation duration among survivors alone.

The outcome is reported as a median with an inter-quartile range. In the convalescent plasma arm, the median is 28 days, with an inter-quartile range of 2 to 28 days. In the standard plasma arm, the median is also 28 days, with an inter-quartile range of 0 to 28 days. These summaries show the center and spread of the posted distributions, but they do not show individual participant trajectories or the full distribution of outcomes.

This type of endpoint can be useful when both death and ventilation are important clinical events: assigning zero to deaths prevents deaths from being treated as ordinary periods without ventilation. However, the resulting summary does not separately describe mortality and ventilation, and the median alone can conceal differences elsewhere in the distribution. A comparison of medians, even if both are reported, is not a complete account of treatment effect. The registry record does not provide a sponsor-reported inferential comparison for this primary endpoint.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry posts descriptive summaries for the primary endpoint and an unadjusted recalculation for the secondary mortality comparison. It does not report a sponsor-conducted statistical analysis for either comparison. The distinction is important: descriptive values summarize the participants represented in the record, while an inferential comparison estimates how compatible those data are with a range of possible between-group effects.

What does a median and inter-quartile range show?

A median is the midpoint of an ordered set of observations: half of the observations are at or below it and half are at or above it, allowing for ties. The inter-quartile range describes the middle portion of the observations, from the lower quartile to the upper quartile. It gives a sense of spread without assuming that the data follow a symmetric, bell-shaped distribution. Here, the reported median is 28 days in both groups, while the reported inter-quartile ranges are 2 to 28 days and 0 to 28 days.

General statistical point: a median and inter-quartile range do not fully describe the distribution. They do not show how many observations occur at each value, how the tails differ, or whether a formal comparison found evidence of a group difference. For an endpoint that assigns zero to participants who die, the distribution may include a concentration at zero as well as many values near the maximum. The registry does not provide the participant-level data or a fuller distribution.

Why is there no primary effect estimate?

The primary outcome is recorded as a median-time type of result, and the posted results include medians and inter-quartile ranges by arm. The registry’s listed reason for skipping a two-arm comparison is that values were missing or not estimable for that comparison. Consequently, there is no posted primary treatment-effect estimate with a confidence interval or p-value to interpret. One cannot derive a valid comparison from the two medians alone. General statistical point: an inferential method for medians or distributions requires information about the observations, or a suitable analysis result, beyond the group medians and quartiles.

The label “all randomized” describes the intended primary outcome population, but the arm summaries do not supply a complete participant flow or clarify how missing outcome data were handled. The approach to missing data, including any assumptions or sensitivity analyses, is not reported in the registry record. Without those details, it is not possible to determine whether the posted summaries include all randomized participants or how missing outcomes may have affected them.

How should the mortality comparison be read?

The secondary endpoint is a count of participants who died from any cause by 90 days. The posted counts and arm denominators were used for a simple, unadjusted recalculation of the risk difference. A risk difference expresses the difference between the two groups’ observed event risks. The recalculated estimate is -0.062, with a confidence interval from -0.326 to 0.202 and p = 0.634. The posted table also gives other effect measures; readers should use the table rather than infer additional measures from the counts.

A confidence interval describes the uncertainty around an estimate under the assumptions of the analysis. A wide interval indicates that the data are compatible with a broad range of effect sizes. Here, the interval spans values on both sides of no difference, so the posted result does not distinguish clearly between a reduction and an increase in risk. The p-value is a measure of how unusual data at least this incompatible with a no-difference model would be under the analysis assumptions; it is not the probability that either treatment is effective, nor a measure of clinical importance. General statistical point: a p-value does not replace the estimate and its interval.

The calculation is unadjusted and based on the posted summary counts. No covariate adjustment, stratification, or sponsor-reported analysis is listed. The discrepancy flag is false, so there is no registry-identified conflict between the posted calculation and a sponsor analysis. However, the absence of a discrepancy does not turn a recalculation into a prespecified analysis or establish that it was part of the trial’s plan.

What does over-dispersion mean in general?

General statistical point: over-dispersion occurs when outcome variability is greater than a model expects. If a model assumes less variability than the data contain, standard errors may be too small and intervals too narrow. Methods that account for over-dispersion can produce more appropriate uncertainty estimates. No over-dispersion model or adjustment is reported for this trial, and the posted summaries do not provide enough information to assess whether it would be relevant to a particular model.

What does the planning illustration mean?

The planning illustration uses the observed secondary effect as a starting point and shows participant counts per arm for detecting specified fractions of that effect. Its stated assumptions are a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation. It is a planning illustration, not an account of the trial’s actual design assumptions and not observed power. The sample-size justification used by the investigators is not reported in the registry record. Planning from an observed effect is especially uncertain when the underlying estimate is imprecise; a smaller assumed effect requires more participants to detect it under otherwise similar assumptions. The values in the planning section should therefore be read as conditional illustrations, not as evidence about what the trial was capable of detecting.

Randomization, when implemented as intended, is used to make treatment groups comparable on average at the start of a study. Masking can reduce some forms of expectation or assessment bias, but the registry’s quadruple-masked label does not specify which people were masked or how masking was maintained. The randomization procedure, allocation concealment, analysis population details, and statistical plan are not reported in the registry record. These methods cannot be assumed from the design labels alone.

10. Multiplicity, Missing Data and Analysis Populations

The registry record does not report a multiplicity plan specifying how the primary and secondary outcomes, or any additional analyses, were to be interpreted together. It also does not report an adjustment for multiple testing. The posted secondary comparison is unadjusted. General statistical point: when many outcomes or analyses are examined, the chance of finding an apparently unusual result can rise, so unadjusted secondary p-values require caution. The record does not indicate how many analyses were planned or conducted beyond the outcome and analysis information posted.

Missing-data methods are not reported in the registry record. The primary outcome population is described as all randomized, but the reported arm counts are 59 and 15, while overall enrollment is 82. The record does not give participant flow details that reconcile those figures, nor does it explain whether participants were excluded from an outcome summary, withdrew, or had unavailable measurements. Those possibilities should not be guessed. The handling of missing outcome values, assumptions about their causes, and sensitivity analyses are not reported in the registry record.

The registry also does not identify the analysis population used for the mortality recalculation beyond the denominators shown in the result. It does not state whether the analysis followed an intention-to-treat principle, used a per-protocol population, or applied another rule. The primary population label is useful, but it does not answer how the analyses treated participants who did not receive assigned plasma or lacked follow-up data.

These omissions affect interpretation because missingness and selective analysis can alter both the estimate and its uncertainty. General statistical point: an analysis restricted to participants with observed outcomes can be biased if the availability of those outcomes is related to prognosis or treatment. The registry record does not provide enough detail to assess that possibility. The reason for termination, any interim analyses, protocol changes, and the trial’s full statistical analysis plan are also not reported in the registry record. Accordingly, readers can describe what is posted, but cannot verify the intended analysis framework or fully assess whether all planned results are represented.

11. 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

Serious events with the highest arm-level rate: renal failure (8.5%), hematologic failure (6.7%), liver failure (5.1%), pulmonary failure (3.4%), infusion reaction (1.7%).

The safety record reports serious adverse events for 8 participants among 59 at risk in the convalescent donor plasma arm and 1 among 15 at risk in the standard donor plasma arm. It reports zero other adverse events in each arm. The registry also lists serious adverse event terms and percentages, including renal failure, hematologic failure, liver failure, pulmonary failure, and infusion reaction. These entries describe what is recorded; they are not, by themselves, evidence that plasma caused an event.

The registry does not provide a comparative statistical analysis of these safety figures, or enough detail to establish whether events were judged related to treatment, how severity was classified, or how the listed percentages were calculated. The arm sizes are limited, and the counts should be interpreted as descriptive rather than as a precise comparison of safety risks. The death counts are also posted, but deaths are separately represented in the 90-day mortality outcome and should not be conflated with the serious-event terms.

Safety summaries can identify events that occurred during a trial, but determining whether an intervention caused an event requires context, including timing, background illness, clinical assessment, and an appropriate comparison. That context is not reported in the registry record. These figures therefore summarize recorded events without establishing causation or a definitive difference between groups.

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

The posted record shows a randomized comparison of convalescent donor plasma with standard donor plasma, with a primary endpoint based on ventilator-free days through 28 days. Both arms have a reported median of 28 days, while their inter-quartile ranges differ. The registry does not provide an inferential primary comparison, so these descriptive summaries alone cannot establish whether the groups differed or whether any difference was meaningful.

For 90-day mortality, the posted counts are accompanied by an unadjusted recalculation with a broad confidence interval and p = 0.634. That interval is compatible with effects in either direction, and the p-value does not establish equivalence or absence of an effect. The secondary comparison is not a sponsor-reported analysis, and no sponsor-reported analyses are listed. It should be treated as an uncertain summary of the posted counts, not a definitive conclusion.

The results do not establish that convalescent plasma works, does not work, is safe, or is unsafe. Those claims would require a fuller understanding of the trial procedures, analysis plan, participant flow, missing data, and complete results. The registry record does not report the reason for termination, the protocol’s sample-size justification, interim analyses, or the statistical plan. It also does not explain the relationship between overall enrollment and the outcome denominators.

General statistical point: evidence is not reduced to whether a p-value crosses a threshold. Estimates, uncertainty, design, outcome definitions, completeness of follow-up, and the prespecified analysis all matter. In this record, the primary endpoint is described but lacks a posted comparative estimate, and the secondary mortality estimate is imprecise and unadjusted. The most defensible reading is therefore limited: the page presents the posted summaries and a recalculation, while important information needed for a stronger interpretation is not reported in the registry record.

13. Why This Trial Matters Statistically

This trial illustrates why a registry entry should be read as a record of design and posted results, not automatically as a complete statistical report. Its primary endpoint combines survival and freedom from invasive ventilation, and the definition explicitly assigns zero ventilator-free days to participants who die. That construction is important because it affects what the endpoint represents and why a median should not be interpreted as ventilation duration among survivors alone.

The record also shows the limits of descriptive summaries. Equal medians do not provide a formal comparison, and the registry does not post a primary effect estimate or uncertainty interval. For the secondary mortality outcome, an unadjusted recalculation is available, but its interval is broad and crosses no difference. Reading estimates with their uncertainty gives a more faithful picture than treating one summary number or p-value as a verdict.

Several practical reporting questions remain unresolved: the relationship between overall enrollment and outcome denominators, the handling of missing data, the analysis populations, the reason for termination, and the planned statistical methods. These omissions are not evidence of a flaw or of a particular outcome; they identify what the registry record cannot answer. The distinction between a sponsor-reported analysis and a recalculation from posted counts is also consequential. The latter can help describe the numbers, but it does not establish what was prespecified.

For biostatisticians and clinical researchers, the example reinforces the value of transparent participant flow, outcome definitions, analysis plans, and complete reporting of estimates with uncertainty. It also shows why planning illustrations based on observed effects must not be confused with observed power or with the trial’s actual design justification. The statistical importance of the record lies not in a definitive verdict, but in the careful separation of what is reported, what can be recalculated, and what remains unknown.

14. 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.

A central limitation is that the posted primary outcome result is descriptive. The registry supplies arm-specific medians and inter-quartile ranges but no inferential comparison, so readers cannot assess the primary endpoint through a reported effect estimate and uncertainty interval. Group medians alone do not show the complete outcome distribution or support a formal claim of difference or similarity.

The secondary mortality comparison is based on posted counts and is an unadjusted recalculation. Its confidence interval is broad, and the record lists no sponsor-reported analyses. The registry does not report whether the recalculation reflects the planned analysis, what covariates or stratification factors were intended, or whether any sensitivity analyses were conducted. The absence of a discrepancy flag does not resolve these questions.

There is also an accounting limitation: overall enrollment is 82, while the arm summaries list 59 and 15 participants. The registry does not provide a participant flow diagram or explain this relationship. The primary population is labelled all randomized, but treatment receipt, follow-up completeness, and the handling of missing outcomes are not described. This prevents a full assessment of the populations represented in the posted outcomes.

The record identifies the trial as terminated, but does not report why. It does not provide the protocol’s sample-size justification, interim analyses, full statistical analysis plan, or details of allocation concealment and masking roles. Arm descriptions are blank, and the standard donor plasma characteristics are not further described. The record also does not show whether event attribution was assessed for safety outcomes. These are limitations of what can be established from the registry record, not grounds to infer why the trial ended or what unreported procedures occurred.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Convalescent Plasma vs. Standard Plasma for COVID-19 — NCT04344535.
• Registry record last updated 2021-12-06; status last verified 2021-12.
• Results first posted 2021-12-06.
• 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.

18. Record Summary

NCT04344535 is a randomized, quadruple-masked, parallel-group Phase 1 and Phase 2 trial comparing convalescent donor plasma with standard donor plasma in hospitalized patients with COVID-19. Stony Brook University is listed as the sponsor. Enrollment is 82; the study began on 2020-04-08, its primary completion date is 2021-01-01, and its status is terminated.

The primary outcome is 28-day ventilator-free days, with deaths during the period assigned zero days. Both arms have a posted median of 28 days, but no inferential primary comparison is reported. The secondary outcome is 90-day all-cause mortality; its posted counts have an unadjusted recalculated comparison with substantial uncertainty. The registry record does not report the termination reason, missing-data methods, sample-size justification, interim analyses, or full statistical plan. Overall enrollment and outcome arm counts are not reconciled in the posted record.