A Study to Assess Pulsed Inhaled Nitric Oxide vs Placebo in Subjects With Mild or Moderate COVID-19: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Assess Pulsed Inhaled Nitric Oxide vs Placebo in Subjects With Mild or Moderate COVID-19 (NCT04421508), a randomized phase 3 study. The primary outcome is The Number of Subjects Who Died or Had Respiratory Failure.

2. Clinical Question

The clinical question concerns whether the observed frequency of the combined outcome differed between the randomized groups. Combining death and respiratory failure into one endpoint means that a participant meeting either condition contributes to the endpoint; the registry title does not present separate primary results for the two components. The outcome is specified through Day 28, so it concerns events during that stated follow-up window.

The record reports 95 participants in the iNO group and 96 in the placebo group. It gives 12 participants with the primary outcome in the iNO group and 10 in the placebo group. Those counts describe the posted results, but the counts alone do not establish that the groups differ beyond what could plausibly arise from sampling variation. The estimated comparison and its uncertainty are therefore important alongside the event counts.

The question is about the trial’s reported comparison, not a recommendation for individual care. The registry record does not report enough detail to establish how every analysis choice was made or how the findings apply to populations outside the trial’s stated description.

3. Trial Design

This was a randomized, quadruple-masked, parallel-group phase 3 trial comparing pulsed inhaled iNO with placebo in subjects described as having mild or moderate COVID-19. The sponsor was Bellerophon Pulse Technologies. The registry record reports enrollment of 191 participants, with 95 assigned to iNO and 96 to placebo for the primary outcome. The primary analysis population is named the All Randomized Analysis Set.

The intervention is described as pulsed inhaled iNO at 125 mcg/kg IBW/hour. The placebo is described as pulsed inhaled N2, 99.999% gas. The record characterizes the trial as randomized and quadruple-masked, but does not state which specific roles were masked. It also does not give further details about sequence generation, allocation concealment, or how masking was maintained.

The primary endpoint is the number of participants who died or had respiratory failure through Day 28. The record also lists secondary outcomes covering recovery, clinical status on the NIAID 8-point Ordinal Scale, discharge alive, hospitalization duration, all-cause mortality, and RT-PCR results at discharge. Some clinical-status entries are repeated in the list, and the record reports 10 outcome measures and 3 analyses; it does not explain the relationship between those counts and each listed outcome.

The trial status is TERMINATED. The registry record lists a start date of 2020-07-12 and a primary completion date of 2020-12-28. It does not state why the trial was terminated, whether termination affected follow-up, or how the decision relates to the reported results. The trial’s original sample-size justification, planned interim analyses, and full statistical analysis plan 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 Randomized Analysis Set

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: All Randomized Analysis Set

6. Primary Endpoint

The registry lists the primary outcome as The Number of Subjects Who Died or Had Respiratory Failure, measured over Through Day 28.

The primary endpoint counts participants who died or had respiratory failure through Day 28. It is a binary participant-level outcome: each person either met the combined endpoint during the stated period or did not. The registry record reports the number of participants with the outcome in each group, rather than the time of the event or separate counts for death and respiratory failure as primary components.

A combined endpoint can capture more than one clinically important event within a single outcome definition. In general, this can make an endpoint more inclusive than either component alone, but it also means the total count does not show which component contributed to the result. The registry record does not report the primary result broken down by component, so the relative contributions of death and respiratory failure cannot be assessed from these counts.

The endpoint is measured through Day 28, which defines the follow-up window relevant to this result. The primary analysis population is the All Randomized Analysis Set. The record does not explain how incomplete follow-up or unavailable outcome information was handled, so the effect of missing observations on the endpoint cannot be evaluated from the posted details.

For the primary comparison, the posted counts are 12 of 95 participants in the iNO group and 10 of 96 in the placebo group. The unadjusted comparison and its confidence interval help describe the size and precision of the difference, while the sponsor’s regression analyses are separate reported analyses. Neither the counts nor a single statistical test, by itself, explains the clinical importance of the endpoint.

7. Results Posted in the Registry

• Inhaled Nitric Oxide (iNO): 12.6% (12/95) versus 10.4% (10/96) on control; risk difference 2.2 percentage points (95% CI -6.8 to 11.3), risk ratio 1.21 (0.55–2.67); unadjusted two-proportion z-test p = 0.632.

8. Secondary Efficacy Endpoints

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

Clinical Status Using National Institute of Allergy and Infectious Diseases (NIAID) 8-point Ordinal Scale

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Pulsed Inhaled iNO 125 mcg/kg IBW/Hour: mean 5.28 versus 5.81 on control; difference -0.53 (95% CI -1.07 to 0.01), standardized effect d = -0.29; z-test p = 0.055.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Pulsed Inhaled iNO 125 mcg/kg IBW/Hour and the control. The interval gives the range of effects that remain compatible with the data.

Duration of Hospitalization

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Inhaled Nitric Oxide (iNO): mean 8.90 versus 8.10 on control; difference 0.80 (95% CI -1.04 to 2.64), standardized effect d = 0.12; z-test p = 0.394.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Inhaled Nitric Oxide (iNO) and the control. The interval gives the range of effects that remain compatible with the data.

All Cause Mortality

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Inhaled Nitric Oxide (iNO): 5.3% (5/95) versus 1.0% (1/96) on control; risk difference 4.2 percentage points (95% CI -0.7 to 9.1), risk ratio 5.05 (0.60–42.44); unadjusted two-proportion z-test p = 0.094.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Inhaled Nitric Oxide (iNO) and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes extend the trial’s assessment beyond the primary combined endpoint. They cover clinical status, recovery, discharge, hospitalization duration, mortality, and RT-PCR results at discharge. Collectively, these measures address different aspects of the course of illness and use different outcome scales. They therefore provide complementary descriptions, not interchangeable confirmations of one underlying treatment effect.

The posted quantitative results include a Day 7 clinical-status score, duration of hospitalization through Day 28, and all-cause mortality through Day 28. The registry also lists other secondary measures, but the record summarized for this analysis does not include numerical results for every listed outcome. It reports 10 outcome measures and 3 analyses, while listing repeated clinical-status entries; the relationship between those counts is not explained.

The secondary results do not overturn the primary finding or establish a consistent difference across outcomes. They should be considered alongside the primary result as additional, but less definitive, evidence. In general, examining many outcomes and analyses increases the chance that at least one result will appear notable by chance. The secondary p-values are unadjusted, and the registry does not report a multiplicity adjustment. Consequently, isolated secondary findings carry less weight than a prespecified primary result interpreted with its uncertainty.

Secondary endpoints can also differ in the number of participants analyzed and in the meaning of their measurement scales. The record reports different group sizes for some outcomes, but does not explain why those denominators differ. Without details about outcome-specific missingness, analysis populations, and planned comparisons, apparent patterns across outcomes should be treated cautiously rather than assembled into a single overall claim.

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

• Number of Subjects to Recover
• Clinical Status Using National Institute of Allergy and Infectious Diseases (NIAID) 8-point Ordinal Scale
• Clinical Status Using National Institute of Allergy and Infectious Diseases (NIAID) 8-point Ordinal Scale
• Clinical Status Using National Institute of Allergy and Infectious Diseases (NIAID) 8-point Ordinal Scale
• Number of Subjects Discharged Alive From Hospital
• Duration of Hospitalization
• All Cause Mortality
• Difference in Proportion in RT-PCR Results at Discharge by Treatment Group

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is a count of participants meeting a binary endpoint: death or respiratory failure through Day 28. For a binary outcome, a risk is the proportion of participants with the event in a group. A risk difference compares those proportions by subtraction. The posted recalculation uses the placebo group as the reference and reports a risk difference of 0.022 for iNO versus placebo. Because this value is reported to four decimal places in the registry-derived statistics, it is not repeated; the estimate’s direction and uncertainty are described instead.

In general, a risk difference is expressed on the absolute probability scale. A positive value under this comparison’s stated direction indicates a larger observed proportion in the intervention group, while a negative value indicates a smaller observed proportion. The confidence interval describes a range of values compatible with the estimate under the method’s assumptions. When an interval includes zero, the data do not rule out no difference at the confidence level used. This is not proof that there is no effect; it is an indication that the estimate is not sufficiently precise to distinguish the observed difference from zero on that basis.

The unadjusted p-value for the primary comparison is 0.632. A p-value describes how unusual data at least as discrepant as those observed would be under a specified null model, assuming that model and its assumptions apply. It is not the probability that the null hypothesis is true, nor a measure of the probability that the result will replicate. The registry’s sample-size illustration uses a two-sided significance level of 0.05, but the trial’s own original analysis plan and decision threshold are not reported in the registry record.

The sponsor also reports logistic-regression analyses with odds ratios. Logistic regression models the odds of a binary outcome, and an odds ratio compares those odds between groups under the model. Odds ratios and risk differences use different scales; their numerical values are not expected to match. Adjusted and unadjusted estimates can also differ when a model includes covariates, but the registry record does not identify the covariates or the distinct specifications for the two posted analyses. The odds-ratio results should therefore be reported as the registry lists them, without inferring an unreported model structure.

The registry’s secondary outcomes include continuous measures and further binary outcomes. For the Day 7 ordinal-scale outcome, the posted group values are least-squares means with standard errors. A standard error describes the sampling uncertainty of an estimated mean, not the spread of individual participants’ scores. The recalculated mean difference is -0.53, and its unadjusted p-value is 0.055. The source notes that the pooled standard deviation used for the planning illustration is approximate, based on the posted standard errors. This approximation affects the planning calculation, not the reported group means themselves.

Hospitalization duration is reported as a mean with standard deviation. A standard deviation describes variation among observations, whereas a standard error describes uncertainty in an estimated mean. The registry reports a mean of 8.9 days in the iNO group and 8.1 days in placebo, and a recalculated mean difference of 0.8 days. The unadjusted p-value is 0.394. In general, mean-based comparisons can be sensitive to skew and extreme observations; the registry entries summarized for this analysis do not state whether a different distributional method or transformation was used in the sponsor’s analysis.

All-cause mortality is a binary secondary outcome reported through Day 28. The registry lists 5 participants in the iNO group and 1 in placebo. The unadjusted risk-difference interval includes zero, and the unadjusted p-value is 0.094. With few events, binary-outcome estimates can be unstable and confidence intervals can be wide. In general, exact or other small-sample methods may be considered for sparse data, but the registry record does not report which method, if any, was used for the sponsor’s mortality analysis.

Planning illustrations based on observed effects are not a reconstruction of the trial design. They use the assumptions stated in the record: two-sided significance level 0.05, equal allocation between two arms, and a normal approximation, with the effect to detect set to a stated fraction of the observed effect. Because these calculations start from the observed estimates, they are illustrations of hypothetical future planning, not evidence that the completed trial had a particular planned sample size or statistical power.

Randomization is intended, in general, to support a fair comparison by balancing prognostic factors on average across assigned groups. It does not guarantee that every factor is balanced in a particular trial, especially with limited enrollment. Masking can, in general, reduce the influence of expectations on behavior, assessment, or reporting. The record names quadruple masking but does not describe who was masked, so the specific protections cannot be assessed from the registry entry alone.

11. Confidence Intervals and Precision

The primary risk-difference confidence interval crosses zero. That means the posted estimate is compatible with a range that includes no difference and also allows for effects in either direction. The interval is broad relative to the small observed counts, so the primary data do not pin down the size of a possible difference. The p-value of 0.632 is consistent with this lack of precision; it does not establish equivalence or prove that the treatments have the same effect.

Confidence intervals are useful because they show the uncertainty around an estimate rather than presenting a single point as definitive. In general, a confidence interval’s width reflects factors such as sample size, event frequency, and the statistical method. When events are uncommon and the groups are modest in size, the amount of information about a difference may be limited. The registry reports 12 primary events in one group and 10 in the other, so the point estimate rests on a small number of observed events.

Other posted comparisons are also imprecise. The reported hospitalization-duration interval spans negative and positive differences. The Day 7 score interval reaches slightly above zero while extending below it. The mortality interval includes zero, and the sponsor’s second logistic-regression interval is wide. These intervals do not all use the same outcome scale or analysis method, so their widths should not be compared as if they were interchangeable measures.

Precision is distinct from clinical importance. The registry record does not state a threshold for a clinically important difference on the primary endpoint, and the combined endpoint does not provide separate primary estimates for death and respiratory failure. The posted estimates therefore describe statistical uncertainty, but do not alone establish how large a difference would matter to patients or clinicians.

The record does not provide enough detail to assess the effects of missing data, alternative model specifications, or subgroup analyses on precision. It also does not give a primary confidence interval for the first reported logistic-regression odds ratio. Those omissions limit how fully the reported uncertainty can be evaluated.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning illustrations report participants per arm for 80% power and participants per arm for 90% power. They are not total event counts: the outcomes considered are not time-to-event outcomes in these calculations. The registry’s stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect set to the stated fraction of the observed effect.

The 80% power column gives the hypothetical number of participants needed in each arm for an 80% chance of detecting the specified effect under those assumptions, if that effect were the true one and the model assumptions held. The 90% power column gives the corresponding per-arm number for a 90% chance. These are planning illustrations based on observed effects; they do not describe the trial’s original sample-size justification, which is not reported in the registry record.

Each planning row represents a different fraction of the observed effect. For example, the primary-outcome illustration lists 3,262 participants per arm for 80% power when planning for 100% of the observed effect, and 12,500 participants per arm when planning for 50% of it. Smaller effects require more participants to detect, all else equal. These figures should be read as conditional planning scenarios, not as a recommended study size or a claim about what the trial should have enrolled.

This is not observed power. It is a hypothetical planning calculation using the observed estimate as its starting point. The trial’s actual design assumptions, including any assumptions used to choose enrollment, are not reported in the registry record. The sample-size figures consequently cannot be used to infer the original design rationale or to interpret the observed p-value as a probability that the trial was adequately powered.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.104; treatment proportion p₁ = 0.126; average p̄ = 0.115.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.4516 = 0.8851.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.4513 = 0.3798.
4. n = (0.8851 + 0.3798)² / (0.0221)² = 3,261.4, rounded up to 3,262 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.104166666666667                          # control proportion
p1 <- 0.126315789473684                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(4421508)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 3,262 (80% power), 4,366 (90% power)
 75% of observed effect: n per arm = 5,678 (80% power), 7,600 (90% power)
 50% of observed effect: n per arm = 12,500 (80% power), 16,734 (90% power)
power.prop.test: 3261.4 
Simulated power at n = 3,262 per arm: 0.797 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

The registry reports two sponsor logistic-regression analyses comparing the iNO and placebo groups for the primary outcome. One reports an odds ratio of 0.36 with a 95% confidence level indicated but no confidence limits listed. The other reports an odds ratio of 1.502 with a 95% confidence interval from 0.567 to 3.977. The corresponding reported p-values are 0.632 and 0.413; those values are not repeated because they exceed the permitted displayed precision.

The recalculation is an unadjusted comparison derived from posted group counts. It reports the risk difference, whereas the sponsor analyses report odds ratios from logistic regression. These are different effect measures and may reflect different model specifications; their estimates should not be compared as if they measured the same quantity. The registry record does not explain how the two sponsor analyses differ, which covariates were included, or why the first analysis has no confidence limits listed.

The record does not flag a discrepancy between the sponsor analysis and the recalculation. The sponsor’s reported analyses remain the trial’s own inferential results, while the recalculation is a straightforward unadjusted summary of the posted counts. The absence of a flagged discrepancy does not resolve the missing model details or explain the two distinct odds-ratio estimates. Interpretations should preserve that distinction and avoid choosing one sponsor estimate as definitive without further documentation.

15. Multiplicity, Missing Data and Analysis Populations

The registry record lists several secondary outcomes and reports multiple analyses, but it does not describe a procedure for controlling multiplicity. When several hypotheses are tested, the chance of at least one apparently notable result can rise. In general, multiplicity methods can adjust decision thresholds or otherwise account for the number and structure of tests. No such method is reported in the registry record, so the posted unadjusted secondary p-values should not be interpreted as if they had a documented family-wise error control or other correction.

The record identifies the primary analysis population as the All Randomized Analysis Set. It does not provide a fuller description of that population, such as how participants were handled if they did not receive assigned treatment or had incomplete follow-up. The record also does not explain the analysis populations used for each secondary outcome. Group denominators vary across outcomes, but the reasons are not reported in the registry record.

Missing-data handling is not described in the information available for this analysis. The record does not state how missing primary outcomes were treated, whether any imputation was used, or whether sensitivity analyses assessed alternative assumptions. This matters because the interpretation of a group comparison can depend on how unavailable outcomes are handled, especially if missingness differs between groups or relates to outcome status. Those possibilities cannot be assessed from the posted details.

The registry does not report whether interim analyses were planned or conducted, whether any stopping boundary was used, or how the terminated status relates to the analysis. It also does not provide the trial’s original statistical analysis plan or original sample-size justification. These are reporting gaps, not evidence that a particular method was or was not used.

Without prespecified analysis details, multiplicity control, and missing-data procedures, the primary counts and reported model results can still be described, but some aspects of confirmatory interpretation remain uncertain. A reader should distinguish the primary outcome from the broader set of secondary comparisons and avoid treating nominal secondary p-values as definitive evidence.

16. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: respiratory failure (7.7%), hypoxia (3.3%), acute respiratory distress syndrome (1.1%), pulmonary embolism (1.1%), covid-19 pneumonia (1.1%).

The adverse-event figures use safety-group denominators that differ from the randomized primary-outcome groups. Serious adverse events were reported for 13 participants among 91 at risk in the iNO group and 12 among 95 at risk in placebo. Other adverse events were reported for 16 among 91 at risk in the iNO group and 28 among 95 at risk in placebo. The record lists 5 deaths in the iNO group and 1 in placebo in its adverse-event information. These figures are descriptive counts; they are not adjusted comparisons.

The most frequent serious-event terms listed include respiratory failure at 7.7%, hypoxia at 3.3%, and acute respiratory distress syndrome, pulmonary embolism, COVID-19 pneumonia, and septic shock at 1.1% each. The registry record does not specify the denominator or detailed coding context for these term percentages in the information. They should therefore be read as registry-listed term frequencies, not converted into comparisons between groups.

Safety counts can include different event types and may not correspond one-for-one to participants, depending on the registry’s event reporting conventions; those conventions are not explained. The difference between counts at risk and randomized group sizes is also not explained. These data describe what was posted, but do not by themselves establish causation or provide a complete safety comparison.

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

The primary result describes a small numerical difference in the observed counts of death or respiratory failure through Day 28: 12 participants in the iNO group and 10 in placebo. The unadjusted risk-difference interval includes zero, and the unadjusted p-value is 0.632. The posted data therefore do not establish a clear difference between the groups on this endpoint. They also do not demonstrate that the groups are equivalent or exclude a potentially meaningful difference in either direction.

The distinction between absence of clear statistical evidence and evidence of no effect is important. In general, a wide confidence interval can include both little or no difference and larger differences. With the counts and uncertainty, the result should be interpreted as inconclusive about the size and direction of a possible effect, rather than as a precise estimate of no difference.

The secondary findings add context but have limitations. They include different outcomes, scales, and analysis populations, and their p-values are unadjusted. They cannot be combined into a single claim that the intervention did or did not affect all aspects of illness. The registry record does not report numerical results for every listed secondary outcome or enough detail about missing data and analysis specifications to resolve those limitations.

Safety data are also descriptive and use denominators that differ from those for the primary outcome. The posted event counts and serious-event terms should not be taken as proof that an event was caused by treatment, nor as a comprehensive account of comparative safety.

Finally, the trial is listed as terminated, but the reason is not reported in the registry record. Termination alone does not explain the observed results. The available data should be understood as the results posted for this randomized trial, with uncertainty about the primary comparison and gaps in the record’s statistical and follow-up details. They are not a basis for medical advice or a conclusion about treatment for an individual.

18. Why This Trial Matters Statistically

This trial illustrates why an event count and a statistical comparison answer different questions. The counts show how many participants met the primary endpoint in each group; the risk difference and confidence interval describe the estimated contrast and its uncertainty. Here, the primary event counts are close in number, but the central statistical issue is not simply which count is larger. It is how much uncertainty surrounds the between-group estimate.

It also illustrates the value of distinguishing effect measures. The recalculation uses a risk difference, while the sponsor reports odds ratios from logistic regression. These measures express comparisons on different scales. Without the model details and the explanation for the two sponsor estimates, a reader should not treat the numbers as directly interchangeable or infer an unreported analysis rationale.

The secondary outcomes show why multiple testing matters. A trial can assess several dimensions of illness, but unadjusted results across many endpoints carry more chance of an apparently notable result arising by chance. Secondary findings can still inform interpretation, but their evidential weight depends on how they were planned, analyzed, and interpreted. Those details are not fully reported in the registry record.

Finally, the sample-size illustrations demonstrate the difference between planning for a future study and describing a completed one. Their assumptions are stated, but the original design justification is not. They cannot establish the trial’s planned power or explain the termination. Together, these distinctions help separate observed results, statistical uncertainty, and information the registry does not report.

19. Limitations

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

Several features of the registry record limit interpretation beyond the uncertainty of the primary comparison. The primary endpoint combines death and respiratory failure, but separate primary counts or estimates for those components are not reported. This prevents assessment of whether one component contributed more strongly to the combined result. The record also does not state a threshold for a clinically important difference on the endpoint.

The reason for trial termination is not reported in the registry record. Nor does the record explain whether termination affected recruitment, follow-up, outcome collection, or the analysis. The start and primary completion dates are listed, but those dates do not establish why the trial stopped or whether all planned follow-up was completed.

The sponsor’s two logistic-regression odds ratios are not fully explained. One has no confidence limits listed, and the registry does not identify differences in model specification, covariate adjustment, or coding. These missing details make it difficult to relate the sponsor estimates to the unadjusted risk-difference calculation. The registry also does not report the original statistical analysis plan, the trial’s own sample-size justification, or whether interim analyses occurred.

The listed adverse-event denominators differ from the randomized group sizes, and the reason is not reported. The detailed denominator and coding context for serious-event term percentages are also not given in the information. Safety counts therefore have to be interpreted descriptively.

Finally, the record does not describe missing-data handling, outcome-specific analysis populations, or multiplicity control. Some secondary outcomes have different group sizes, and numerical results are not reported for every listed secondary measure. These gaps restrict assessment of bias, robustness, and the consistency of secondary findings. They should not be filled by assumptions about the protocol or analysis.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study to Assess Pulsed Inhaled Nitric Oxide vs Placebo in Subjects With Mild or Moderate COVID-19 — NCT04421508.
• Registry record last updated 2023-02-21; status last verified 2023-01.
• Results first posted 2023-02-01.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

23. Record Summary

The registry record describes a randomized, quadruple-masked, parallel-group phase 3 trial comparing pulsed inhaled iNO with placebo in subjects with mild or moderate COVID-19. Enrollment was 191, with 95 participants in the iNO group and 96 in placebo for the primary endpoint. The primary outcome was the number of participants who died or had respiratory failure through Day 28.

The posted primary counts were 12 in the iNO group and 10 in placebo. The unadjusted comparison is imprecise, with an interval that includes zero and an unadjusted p-value of 0.632. Sponsor logistic-regression results are also reported, but the registry does not explain why two odds ratios appear or provide full details for both analyses.

Secondary outcomes cover clinical status, recovery, discharge, hospitalization, mortality, and RT-PCR results. Their interpretation is limited by unadjusted testing and incomplete reporting of analysis details. The record lists the trial as terminated; the reason is not reported in the registry record. The original sample-size justification, missing-data procedures, multiplicity approach, and interim-analysis details are also not reported.