Study of TJ003234 (Anti-GM-CSF Monoclonal Antibody) in Subjects With Severe Coronavirus Disease 2019 (COVID-19): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of TJ003234 (Anti-GM-CSF Monoclonal Antibody) in Subjects With Severe Coronavirus Disease 2019 (COVID-19) (NCT04341116), a randomized phase 2, phase 3 study. The primary outcome is Percentage of Subjects Alive and Free of Mechanical Ventilation Among Subjects Who Are Free of Mechanical Ventilation at Baseline.

2. Clinical Question

The trial asked whether intravenous TJ003234, given with supportive care, affected outcomes in subjects with severe COVID-19. It also aimed to assess the drug’s effect on cytokine levels. The primary endpoint concerned subjects who were free of mechanical ventilation at baseline: whether they remained alive and free of mechanical ventilation during Day 1 through Day 30.

The endpoint reflects both survival and respiratory support status, rather than survival alone. The registry’s outcome population and the groups named in its reported inferential analysis are not identical descriptions of a placebo-controlled comparison. That distinction matters when interpreting the reported estimate: it should not be read as an estimate of TJ003234 versus placebo unless the contrast is explicitly established. The contrast details are not reported in the registry record.

3. Trial Design

This was a Phase 2/3, randomized, double-blind, placebo-controlled, multicenter, parallel-group trial. The registry lists an enrollment of 149 and a completed status. The recorded start date is 2020-04-11, and primary completion is dated 2022-02-07. The study evaluated TJ003234 administered as an intravenous infusion to subjects with severe COVID-19 receiving supportive care.

The registered arms included TJ003234 and placebo, with dose groups distributed across trial parts. The arm descriptions identify a low-dose TJ003234 group as Part 1 only and a high-dose group as Part 2 Phase 3 only. The outcome data list TJ003234 6mg/kg and placebo groups for Part 1 with zero participants for this endpoint, and Part 2 groups of TJ003234 6mg/kg, TJ003234 10mg/kg, and placebo. The 10mg/kg group is also called TJ0032034 in the outcome listing; the registry record does not explain this spelling difference.

For the primary endpoint, the Part 2 groups had 67 participants and 56 meeting the endpoint in the TJ003234 6mg/kg group, 6 participants and 5 meeting it in the TJ003234 10mg/kg group, and 36 participants and 27 meeting it in the placebo group. Part 1 primary-endpoint data were not collected. These counts describe the posted endpoint results; they do not by themselves establish the comparison tested in the registry’s reported analysis.

The registry also lists secondary outcomes concerning recovery by Day 14, recovery on Day 30, all-cause mortality by Day 30, time to recovery among subjects alive by Day 30, and length of hospitalization. The protocol’s detailed allocation procedures, the rationale for the planned enrollment, and any changes to the study design 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.

Overall Study

Analysis population for the primary outcome, as defined in the registry: All subjects randomized to treatment who were mechanical ventilation free at baseline was collected only for Part 2 arms/groups. Data was not collected from the Part 1 Arms/Groups.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Only the first 6 of 7 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Subjects Alive and Free of Mechanical Ventilation Among Subjects Who Are Free of Mechanical Ventilation at Baseline, measured over Day 1 through Day 30.

The primary endpoint is the percentage of subjects alive and free of mechanical ventilation among those who were free of mechanical ventilation at baseline. It covers Day 1 through Day 30. The registry defines free of mechanical ventilation as a score from 1 to 5 on an eight-category ordinal scale. The scale runs from death at score 8, through ventilation with specified additional support at score 7, intubation and mechanical ventilation at score 6, and non-invasive ventilation or high-flow oxygen at score 5. Lower scores describe hospitalization with oxygen, hospitalization without oxygen, limitation of activities with discharge, and no limitation of activities with discharge.

This endpoint combines survival with a clinically relevant measure of respiratory support and recovery status. It can therefore distinguish subjects who remain alive without mechanical ventilation from those who die or require invasive mechanical ventilation during the observation period. It does not describe every aspect of recovery: subjects may still require other forms of oxygen or remain hospitalized while meeting the registry’s definition. The endpoint also applies only to the baseline-free population specified by the registry, and primary-outcome data were collected only for Part 2.

The outcome is named as a percentage, while the posted group results are presented as participant counts. The record does not provide uncertainty intervals for those group counts. Nor does it report a placebo-based primary comparison in the listed inferential analysis. These features constrain what can be concluded from the endpoint figures alone.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is binary at the participant level: each eligible subject either meets the definition of alive and free of mechanical ventilation over the stated period or does not. The registry reports counts, which preserve the number of participants in each category. A percentage is a natural way to summarize a binary outcome, but the record’s listed group results are counts and do not include uncertainty intervals for each group. A count alone does not show how precisely the underlying event probability is estimated.

What does the reported comparison describe?

The reported analysis lists the Part 2 TJ003234 6mg/kg and 10mg/kg groups. Its method is identified as Fisher Exact, and its result is labelled median difference (final values). These descriptions do not make the comparison’s estimand fully clear. The analysis does not list placebo among its groups, despite placebo being an arm in the trial and having posted endpoint counts. The unit of the reported difference and the precise null hypothesis are not reported in the registry record. Statistical interpretation should therefore retain the registry’s labels rather than assume a placebo contrast or translate the estimate into a percentage-point difference.

Fisher’s exact test is a general method for evaluating an association between two categorical variables in a contingency table. It is often used when group sizes or event counts are limited, because its reference distribution does not rely on the large-sample approximation used by some other tests. The test does not itself solve questions about which groups should be compared, what population is targeted, or how missing outcomes are handled. Those features depend on the analysis specification.

How should the interval and p value be read?

A confidence interval expresses uncertainty around an estimated effect. A narrower interval indicates greater precision than a wider interval, given the same type of estimate and assumptions. The reported 95% confidence interval runs from -6.5 to 27, so it includes zero and values on both sides of zero. This indicates uncertainty about the direction and size of the contrast as recorded. The p value of .28 is a measure of how incompatible the observed data are with the test’s null hypothesis, under the assumptions of that test. It is not the probability that the null hypothesis is true, and it does not measure the probability that the result arose by chance in a general sense.

The interval and p value should be considered together with the trial’s limited endpoint population and group sizes. In particular, the 10mg/kg Part 2 group contains 6 participants. Estimates based on small groups can be imprecise, and a non-small p value cannot establish that groups have equivalent outcomes. The record does not report a prespecified equivalence or non-inferiority framework.

Why is no recalculated comparison shown?

A simple unadjusted recalculation can sometimes be made from posted counts for a defined two-arm binary comparison. In this record, the recalculated values are missing or not estimable for that comparison. It would be misleading to construct a new estimate by selecting a contrast not specified in the reported analysis. The available counts remain useful descriptively, but they do not replace the analysis plan or establish which contrast the sponsor intended.

The record also describes planning illustrations based on a stated fraction of the observed effect, a two-sided significance level of 0.05, equal allocation, and a normal approximation. Such illustrations are prospective planning scenarios, not evidence about the probability that this completed study detected an effect. They are not observed power, and they do not document the trial’s own sample-size justification. The actual sample-size rationale is not reported in the registry record.

For binary outcomes, the event probability and the allocation between groups affect the precision of an estimated contrast. A planning calculation depends on assumptions about the effect to detect and the statistical method. If those assumptions do not match the final design or the outcome analysis, the planning result may not describe the achieved precision. No additional assumptions about the protocol or its analysis should be inferred from the planning illustration.

What does the outcome definition add?

The ordinal scale orders participants by clinical status, from death and invasive ventilation through different levels of respiratory support and discharge status. The primary endpoint reduces that ordered information to a threshold: scores 1 to 5 count as free of mechanical ventilation, while higher scores do not meet that definition. This makes a binary count easier to interpret, but it discards distinctions within each side of the threshold. For example, the endpoint groups together participants on non-invasive ventilation and participants discharged without activity limitations. The registry does not report a primary analysis of the full ordinal scale.

These methods explain the information contained in the posted results, not the analysis choices that may have appeared in the full protocol. The protocol’s detailed statistical methods, handling of missing endpoint values, and any adjusted models are not reported in the registry record.

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("TJ003234 6mg/kg (Part 2)", "TJ0032034 10mg/kg (Part 2)", "Placebo (Part 2)")
n      <- c(67, 6, 36)
events <- c(56, 5, 27)
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  TJ003234 6mg/kg (Part 2)  67  56  83.6  72.5  91.5  72.9  90.6  72.9  90.5
ARM  TJ0032034 10mg/kg (Part 2)  6  5  83.3  35.9  99.6  43.6  97.0  42.1  96.3
ARM  Placebo (Part 2)  36  27  75.0  57.8  87.9  58.9  86.2  58.8  86.2

11. Analyses Reported by the Sponsor

The registry records a Fisher Exact analysis for the Part 2 TJ003234 6mg/kg and 10mg/kg groups. It reports a median difference (final values) of 9, a 95% confidence interval from -6.5 to 27, and p = .28. The analysis does not list the placebo group, so the result should not be described as a TJ003234-versus-placebo estimate without clarification of the intended contrast. The record does not explain the estimate’s unit or the relationship between the median-difference label and the stated test method.

No recalculated two-arm estimate is available from the posted binary endpoint data; the registry record identifies the values as missing or not estimable for that comparison. There is therefore no recalculated result to reconcile with the recorded analysis. The sponsor-recorded result is the available inferential result, but its interpretation remains limited by the unspecified contrast and measure. The endpoint counts for placebo can be reported descriptively, but they should not be combined with the two listed active-treatment groups to create an unreported comparison.

12. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary endpoint and five secondary outcomes: recovery by Day 14, recovery on Day 30, all-cause mortality by Day 30, time to recovery among subjects alive by Day 30, and length of hospitalization. Testing multiple outcomes can increase the chance of obtaining at least one apparently notable result even when no true effect exists. Statistical approaches can address this issue by defining a hierarchy, adjusting significance thresholds, or treating some analyses as exploratory. The multiplicity plan is not reported in the registry record.

The record reports one inferential analysis for the primary outcome. It does not state whether other analyses were planned, whether the reported test was prespecified, or whether any interim analyses were performed. These details are not reported in the registry record and should not be inferred from the existence of the Phase 2/3 design or from the outcome list.

The endpoint population is described as all randomized subjects who were free of mechanical ventilation at baseline, but the registry says this information was collected only for Part 2 and not for Part 1. Further details of the analysis population, including whether all eligible randomized participants were analyzed in their assigned groups, are not reported. The record also does not explain how deaths, withdrawals, missing assessments, or other incomplete endpoint data were handled. These missing-data methods are not reported in the registry record.

Missing-data handling matters because the endpoint requires status over a defined period. Different rules for incomplete follow-up can lead to different counts and estimates. Without those rules and a clear analysis-population definition, the posted results cannot establish how all randomized participants were incorporated. No imputation method, sensitivity analysis, or protocol deviation rule is reported.

These omissions do not invalidate the posted counts, but they limit assessment of the statistical plan and the robustness of the analysis. They also make it harder to determine how the reported analysis relates to the full randomized, placebo-controlled design.

13. 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: (worsening) hypoxic respiratory failure (16.9%), (worsening) pneumonia (due to covid-19) (16.7%), cardiac arrest (12.5%), septic shock (12.5%), acute kidney injury (12.5%).

The registry lists serious adverse events by treatment group and reports deaths separately. In Part 2, serious adverse events affected 22 of 77 participants in the TJ003234 6 mg/kg group, 1 of 6 in the TJ003234 10mg/kg group, and 16 of 42 in the placebo group. The death counts listed for those groups are 11, 1, and 10, respectively. Part 1 serious adverse event counts are also reported: 2 of 8 for TJ003234 3 mg/kg, 4 of 8 for TJ003234 6mg/kg, and 2 of 8 for placebo. The corresponding death counts are 1, 1, and 2.

These are descriptive safety figures. They do not establish that treatment caused or prevented an event, and the registry record does not provide a comparative statistical analysis of the serious adverse event counts. The record lists zero other adverse events in each group, but it does not explain the event-collection procedures or whether these figures cover all events of interest. The serious-event terms listed include worsening hypoxic respiratory failure at 16.9%, worsening pneumonia due to COVID-19 at 16.7%, cardiac arrest, septic shock, acute kidney injury, and COVID-19 at 12.5% each. The basis for those term percentages is not reported in the registry record.

Safety group sizes differ from the primary-outcome group sizes, and the Part 1 safety data should not be mistaken for Part 1 primary-endpoint data. These figures should be read as counts recorded for the named groups, not as a complete causal assessment of harms.

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

The posted results establish that the registry contains primary-endpoint counts for three Part 2 groups, along with one reported inferential analysis for the two active-treatment groups. They also show that the endpoint was defined around survival and freedom from mechanical ventilation among participants who were free of mechanical ventilation at baseline. The counts are descriptive, and the reported estimate has a confidence interval extending on both sides of zero.

The results do not establish a clear treatment-versus-placebo effect, because the listed inferential analysis does not name placebo as a comparison group. The record does not clarify the contrast or the unit of the reported estimate. The absence of a recalculated comparison also means there is no independent unadjusted estimate available from the posted binary data to resolve that ambiguity. A result with p = .28 is not proof of no effect, and an interval containing zero does not demonstrate that groups are equivalent.

The outcome captures an important dimension of status but compresses the full ordinal scale into a binary classification. It does not distinguish among the several scale states counted as free of mechanical ventilation, nor does it describe recovery beyond the stated period. The secondary outcomes address other aspects, but their results and analysis details are not reported in the registry record presented for this evaluation.

The group counts should also be interpreted in light of the restricted endpoint population, the absence of Part 1 primary data, and the small Part 2 10mg/kg group. Missing-data procedures, multiplicity handling, and the detailed statistical plan are not reported. Consequently, the posted information allows a careful account of what was counted and what analysis was recorded, but not a complete appraisal of the trial’s prespecified statistical strategy or a definitive causal conclusion.

15. Why This Trial Matters Statistically

This trial is statistically informative because its design and outcome reporting do not line up perfectly in the available record. The study is described as randomized and placebo-controlled, and placebo endpoint counts are posted. Yet the listed primary analysis names the two active-treatment groups rather than an active treatment and placebo. A reader must keep the descriptive group results separate from the inferential contrast instead of assuming they answer the same question.

The endpoint also illustrates how an ordinal clinical scale can be converted into a binary outcome. That choice creates a straightforward count of participants meeting a threshold, while combining different states on the favorable side of that threshold. The selected definition is therefore central to interpretation: being free of mechanical ventilation does not necessarily mean being free of oxygen support, discharged, or fully recovered.

Precision is another key issue. The 10mg/kg group contains 6 participants, and the reported confidence interval ranges from -6.5 to 27. The interval includes values on either side of zero, so the estimate is uncertain about both direction and size. The posted p value should not be used as a substitute for considering that uncertainty or the missing contrast details.

Finally, the absence of reported methods for missing outcomes, multiplicity, sample-size justification, and analysis populations shows why registry results need methodological context. Those omissions do not erase the observed counts, but they limit the strength and scope of conclusions that can be drawn from them.

16. Limitations

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

The primary endpoint was collected only for Part 2; the registry explicitly says that primary-outcome data were not collected from Part 1. This means the listed primary results do not cover all trial parts, despite the broader Phase 2/3 description and total enrollment of 149. The record does not explain why the endpoint was unavailable for Part 1.

The inferential analysis lists the two active-treatment groups and not placebo. Although Part 2 placebo counts are posted, the registry does not specify a placebo-based analysis for those counts. It also does not clarify the reported estimate’s unit or the exact contrast. This is a substantial interpretive limitation for a placebo-controlled trial because the central treatment comparison cannot be assumed from the group listings alone.

The Part 2 TJ003234 10mg/kg group includes 6 participants, limiting the information that group can contribute to a comparative estimate. The reported confidence interval, from -6.5 to 27, also reflects uncertainty across both directions. The available record does not provide group-specific uncertainty intervals for the endpoint counts.

Further limitations include the absence of reported procedures for missing data, a detailed statistical analysis plan, multiplicity handling, interim analyses, and the rationale for the planned sample size. The registry lists five secondary outcomes but does not report how their analyses relate to the primary analysis. The spelling of the Part 2 10mg/kg group differs between the arm description and outcome listing, and this is not explained in the registry record. These limits constrain interpretation of the posted results, rather than proving that the underlying trial lacked such procedures.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: Study of TJ003234 (Anti-GM-CSF Monoclonal Antibody) in Subjects With Severe Coronavirus Disease 2019 (COVID-19) — NCT04341116.
• Registry record last updated 2023-05-06; status last verified 2023-05.
• Results first posted 2023-05-06.
• 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.

20. Record Summary

NCT04341116 is a completed Phase 2/3 randomized, double-blind, placebo-controlled, multicenter, parallel-group study of intravenous TJ003234 for severe COVID-19 under supportive care. The registry lists enrollment of 149, with a start date of 2020-04-11 and primary completion on 2022-02-07.

The primary endpoint assessed being alive and free of mechanical ventilation through Day 30 among subjects free of mechanical ventilation at baseline. Primary-outcome data were collected only for Part 2. The posted counts are 56 of 67 for TJ003234 6mg/kg, 5 of 6 for TJ003234 10mg/kg, and 27 of 36 for placebo.

The reported analysis names the two active-treatment groups and gives a median difference (final values) of 9, a 95% confidence interval from -6.5 to 27, and p = .28. The contrast and estimate unit are not clarified. A recalculated two-arm comparison is not estimable from the posted binary outcome data. The registry lists five secondary outcomes and safety counts, while detailed methods for multiplicity, missing data, and sample-size justification are not reported in the registry record.