A Study of Intravenous Zanamivir Versus Oral Oseltamivir in Adults and Adolescents Hospitalized With Influenza: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Intravenous Zanamivir Versus Oral Oseltamivir in Adults and Adolescents Hospitalized With Influenza (NCT01231620), a randomized phase 3 study. The primary outcome is Time to Clinical Response (TTCR) in Participants With Confirmed Influenza.

2. Clinical Question

The registry describes a randomized, quadruple-masked, parallel-group phase III study. Its clinical question combines an assessment of two intravenous zanamivir doses with comparison against an oral antiviral active comparator. The posted primary outcome is a time-to-event measure: it records the time until a participant met the specified clinical-response definition. A shorter time would mean the response criterion was met sooner, but the registry results should be read as comparisons of recorded outcomes rather than as standalone evidence about treatment choice.

The primary endpoint description includes analyses in the influenza-positive population and in specified clinical subgroups. The registry record does not give the subgroup counts or enough detail to establish how subgroup analyses were planned or interpreted. The posted values therefore need to be understood in the context of the stated endpoint and the particular populations described, not generalized automatically to all hospitalized patients or other settings.

3. Trial Design

NCT01231620 is listed as a completed phase III randomized study with a parallel-group design and quadruple masking. The registry describes an international, double-blind, double-dummy comparison of intravenous zanamivir at 300 mg or 600 mg twice daily against oral oseltamivir 75 mg twice daily. The double-dummy arrangement paired each active treatment with a matching placebo delivered by the other route. That design helps maintain masking when the active regimens differ in route of administration. The record lists 626 participants as the enrollment and identifies human influenza as the condition.

The three treatment groups are intravenous zanamivir 300 mg, intravenous zanamivir 600 mg, and oral oseltamivir 75 mg. The primary outcome is time to clinical response in participants with confirmed influenza, with a stated timeframe of up to 42 days. The endpoint definition allows hospital discharge to count as the event if it occurs before the specified vital-sign resolution criterion is met. Participants who did not achieve the endpoint were assigned a censored time of 99 days, according to the outcome description.

The registry lists 32 outcome measures and 4 analyses. It identifies respiratory improvement, mortality, activities of daily living, return to premorbid function or activity, and influenza symptoms among secondary outcomes. The posted results also include counts for severe or grade 3/4 adverse events. An outcome being listed does not, by itself, show that it was the basis for the trial’s design or a confirmatory claim.

The record gives the study status as completed and lists its start and primary-completion dates as 2011-01-15 and 2015-03-18. It does not report a sample-size justification, the allocation ratio, details of the randomization procedure, or a full protocol-level statistical analysis plan. Those details should not be inferred from the enrollment or from the posted comparisons.

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: Male and female adult and adolescent participants \>=16 years of age hospitalized with documented influenza or suspected influenza were eligible for enrollment. A total of 626 participants were randomized, and 615 participants were included in the Intent-to-Treat Exposed (ITT-E) Population (pop)

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT-E population comprised of all randomized participants who received at least one dose of investigational product. The Influenza positive population (IPP) is comprised of all participants in the ITT-E population with proven influenza infection.

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 Time to Clinical Response (TTCR) in Participants With Confirmed Influenza, measured over Up to 42 days.

The primary endpoint is time to clinical response, measured in days, among participants with confirmed influenza. The registry defines clinical response as resolution of at least 4 of 5 vital signs—temperature, oxygen saturation, respiratory status, heart rate, and systolic blood pressure—within the relevant resolution criteria and maintained for at least 24 hours, or hospital discharge, whichever occurred first. The outcome description states that analyses were conducted for the influenza-positive population and for specified clinical subgroups.

A time-to-event endpoint is used when both whether and when an event occurs are important. Here, the event is meeting the clinical-response definition or reaching discharge under the stated rule. Comparing time to response can retain information about timing that a simple yes-or-no count at one selected visit would omit. This is a general statistical explanation, not a claim about the study’s rationale, which is not reported in the registry record.

The registry reports medians and full ranges for the treatment groups. A median describes the middle of the observed time distribution; it does not show the entire distribution or, by itself, the uncertainty in a between-group difference. The endpoint description says that 99 days was used as censored time for participants who did not achieve clinical response. Censoring indicates that the endpoint was not observed during the relevant observation period; it is not evidence that a response occurred at that time. The registry record does not give additional detail on censoring or missing-outcome handling.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a time-to-response measure, but the sponsor-reported comparisons use the Wilcoxon rank sum method and report median differences in final values. This rank-based method compares the ordering of observations across two groups rather than relying on a normal-distribution model for the outcome itself. In general, rank methods can be useful when outcome distributions are skewed or include extreme values. The registry record does not report the full analysis specification, including how censoring was handled in the reported rank analysis.

What does the median describe?

A median is the value at the midpoint of an ordered set of observations. It gives a concise description of a group’s central outcome, but it does not convey the shape of the full distribution, how many observations are clustered near the center, or how results vary over time. The registry reports full ranges alongside the medians. A range gives the lowest and highest reported values, but is sensitive to extreme observations and is not a confidence interval for the median or for a treatment contrast.

The sponsor’s listed primary contrasts are differences between medians, with 95% confidence intervals and p-values. A confidence interval expresses uncertainty about an estimated contrast under the assumptions of the analysis. When an interval includes zero for a difference, the data do not distinguish the estimated contrast from no difference at the stated confidence level. This does not demonstrate that the groups are identical, and it does not establish equivalence. The p-value measures how compatible the observed data are with a specified null model; it is not the probability that the null hypothesis is true or the probability that a result occurred by chance.

How should censoring be understood?

For participants who did not achieve time to clinical response, the registry says 99 days was used as censored time. In general, censoring means that the event time is not observed in the same way as an event that occurred during follow-up. It should not be interpreted as an observed response at the censoring time. Time-to-event analyses often use methods that account for this distinction. The registry’s primary-outcome results are described as medians and full ranges, with Wilcoxon rank sum comparisons; additional detail about censoring in those analyses is not reported in the registry record. The posted summaries alone therefore do not allow a reader to reconstruct a complete time-to-event analysis.

What are the secondary binary recalculations?

For selected secondary outcomes recorded as counts or percentages, the recalculated comparisons are unadjusted risk differences with confidence intervals. A risk difference expresses the contrast between the observed proportions in two groups on an absolute scale. The recalculations use the posted group counts and denominators and are not adjusted for other participant characteristics. Confidence intervals that include zero indicate that the recalculated data do not clearly distinguish the contrast from no difference under that method. These recalculations are not the sponsor’s analyses unless explicitly identified as such.

For respiratory improvement, the registry also lists sponsor-reported Wei-Johnson analyses with p-values for two comparisons. Those p-values are not the same analysis as the simple unadjusted risk-difference recalculations. The different methods may answer related but not identical statistical questions, so their estimates and inferential results should not be treated as interchangeable. For the posted severe or grade 3/4 adverse-event outcomes, no sponsor-reported comparison is listed in the record.

What do planning illustrations mean?

Planning estimates based on observed effects are illustrations of how sample size can depend on the effect size targeted and the chosen design assumptions. The stated assumptions for these illustrations are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The calculations use the stated fraction of the observed effect as the effect to detect. For time-to-event calculations, the record describes a total-event basis; for other outcomes, it describes participants per arm. These calculations are not the trial’s own design justification, and they are not observed power. They should not be read as evidence that the trial was adequately or inadequately powered for a particular claim.

More generally, power is a design property tied to assumptions about the effect, variability, allocation, and analysis. Calculating power after results are known by using the observed effect does not add independent evidence about whether a result is real. A confidence interval is generally more informative for interpreting the precision of a treatment-effect estimate. The registry record does not report the original sample-size assumptions or a prespecified power calculation, so the planning illustrations cannot fill that gap.

Why does over-dispersion matter in general?

For count outcomes, over-dispersion means that the observed variability exceeds what a basic model assumes. If present, it can make standard errors too small and confidence intervals too narrow when the model does not account for it. This is a general consideration for count-data analyses; the registry record does not identify an over-dispersion assessment or model for the outcomes. It should not be assumed that over-dispersion occurred in this trial.

Across methods, the interpretation depends on the endpoint definition, the population analyzed, handling of censoring or missing observations, and prespecified analysis choices. Some of these details are not reported in the registry record. The posted estimates can be described as recorded, but missing methodological detail limits independent verification and makes it important not to overstate what the comparisons establish.

10. Analyses Reported by the Sponsor

The sponsor-reported primary analyses use Wilcoxon rank sum tests and give median differences in final values for two comparisons: 300 mg versus 600 mg intravenous zanamivir, and 600 mg intravenous zanamivir versus oral oseltamivir. The reported confidence intervals include zero for both contrasts, and the reported p-values are 0.25 and 0.39, respectively. Those results are the registry’s own inferential analyses for the listed primary comparisons.

A simple recalculation is not available for this median-time endpoint in the same way it is for the posted binary outcomes. The record identifies median-time comparisons as not estimable for the two-arm recalculation approach. Accordingly, no recalculated primary effect estimate should be substituted for the sponsor’s analysis. The sponsor’s reported estimates and methods should be used when describing those comparisons, while recognizing that the registry does not post a sponsor-reported comparison of 300 mg zanamivir against oseltamivir.

For respiratory improvement, sponsor-reported Wei-Johnson p-values are listed for the 300 mg versus 600 mg zanamivir comparison and the 600 mg zanamivir versus oseltamivir comparison. The separate unadjusted recalculations use risk differences and their confidence intervals. They are a complementary descriptive check, not a correction to the sponsor’s analysis. There is no discrepancy flagged between the available recalculations and sponsor results. The registry does not report sponsor analyses for the posted severe or grade 3/4 adverse-event comparisons.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 32 outcome measures and 4 analyses, with 31 secondary outcomes posted. It does not report a multiplicity adjustment, a hierarchy of outcomes, or a plan specifying which comparisons were confirmatory. When several outcomes or contrasts are tested, the chance of at least one apparently unusual result can increase. This is a general statistical issue; without the trial’s stated approach, it is not possible to determine whether or how the sponsor controlled that risk. Secondary p-values should therefore be read as individual results rather than as a single confirmatory family of tests.

The primary outcome description names the ITT-E population and the influenza-positive population, and says the analysis was performed for participants with symptom onset at most 4 days earlier and for those on mechanical ventilation or in an intensive care unit. The record defines the ITT-E population as randomized participants who received at least one dose of investigational product, and the influenza-positive population as ITT-E participants with proven influenza infection. Subgroup-specific counts and a detailed account of which population underlies each posted arm value are not reported in the registry record.

The endpoint description notes that participants who did not achieve clinical response were assigned censored time of 99 days. It does not provide further detail about how censoring, withdrawal, missing vital-sign measurements, or other missing outcome data were handled in each analysis. The registry record also does not report imputation methods, sensitivity analyses for missing data, or a full protocol-level analysis plan. Those omissions limit assessment of whether missingness could have affected the estimates.

These limitations do not invalidate the posted results, but they constrain how confidently they can be interpreted and independently reproduced. A reader cannot infer unreported multiplicity control, missing-data assumptions, subgroup sample sizes, or interim monitoring procedures. Such details are not reported in the registry record.

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.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pneumonia (3.5%), respiratory failure (2.5%), acute respiratory distress syndrome (2.0%), acute kidney injury (2.0%), septic shock (2.0%).

The registry reports serious adverse events for the treatment groups as counts among participants at risk: 38 in the 300 mg intravenous zanamivir group, 33 in the 600 mg group, and 38 in the oral oseltamivir group. The associated at-risk group counts are 201, 209, and 205, respectively. Death counts are listed as 15, 15, and 11, respectively. These are descriptive registry counts; they do not establish that treatment caused an event or death.

The most frequently listed serious-event terms are pneumonia at 3.5%, respiratory failure at 2.5%, acute respiratory distress syndrome at 2.0%, acute kidney injury at 2.0%, septic shock at 2.0%, and multi-organ failure at 1.0%. The registry also posts counts for any severe or grade 3/4 adverse event and for severe or grade 3/4 treatment-related adverse events. These are distinct categories from the serious adverse-event counts and should not be conflated.

Comparisons of adverse-event counts are difficult to interpret without event timing, exposure duration, definitions, and statistical methods. Those details are not reported in the registry record. The serious-event figures are therefore best read as a description of what was recorded in the groups, not as a definitive estimate of comparative safety or proof of causation.

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

The posted primary results describe median times to clinical response that are close in scale across the three groups, with sponsor-reported comparisons for two pairings. For both reported contrasts, the confidence interval includes zero and the p-value does not indicate clear statistical separation under the reported analysis. The results are compatible with a range of possible group differences, so the point estimates alone should not be treated as precise estimates of a treatment advantage.

These findings do not establish that the treatments are equivalent, that they have identical effects, or that any one treatment is superior. Equivalence and non-inferiority require an appropriate design, a prespecified margin, and an analysis tied to that objective. Such a margin or design claim is not reported in the registry record. Nor should a lack of statistical evidence for a difference be translated into evidence that no difference exists.

The endpoint is specifically defined around vital-sign resolution or hospital discharge, and the primary analysis population is described as influenza-positive. The result does not automatically apply to people outside that population, to other illness settings, or to outcomes that were not measured. The listed subgroup analyses do not include subgroup-specific counts or estimates in the registry information, so no subgroup conclusion is warranted.

The secondary outcomes and safety counts add context but do not resolve the uncertainty in the primary comparison. Multiple outcomes were listed, and the registry record does not report a multiplicity plan. Some endpoint methods and missing-data details are also not reported. These are reasons to keep conclusions anchored to the specific posted estimates and their uncertainty rather than generalizing beyond the information recorded.

In practical statistical terms, the result is an account of observed group summaries and sponsor-reported comparisons, not a recommendation about treatment. It does not establish a causal explanation for any adverse event, nor does it show that all relevant benefits or harms were captured. Interpretation is limited to the study’s defined groups, outcomes, follow-up, and analyses as described in the registry record.

14. Why This Trial Matters Statistically

This trial is statistically informative because it compares two doses of an intravenous treatment with an active oral comparator in a randomized, masked, parallel-group design. The double-dummy approach is relevant to interpretation because the treatments differ by route; matching placebos are listed for both routes. The design addresses a direct comparative question rather than comparing treatment with no active therapy.

The primary endpoint also illustrates a common analytical challenge: a time-to-event question may be reported with medians and rank-based comparisons, while censoring affects how the underlying event times should be understood. The registry states that participants who did not achieve clinical response were assigned a censored time of 99 days, but it does not provide all the analysis details needed to reconstruct the time-to-event methods. This makes careful separation of the posted summaries, the sponsor’s inferential analyses, and any simple recalculation especially important.

The confidence intervals around the sponsor-reported median differences are central to interpretation. They express uncertainty that a single point estimate or p-value cannot convey. Since the intervals include zero, the available comparisons do not clearly separate the groups, but the intervals do not prove no difference. This distinction is essential when interpreting randomized trial results.

The record also demonstrates why a large list of secondary outcomes does not automatically strengthen a primary conclusion. Different outcomes capture different aspects of illness and recovery, and repeated testing can make isolated nominal findings harder to interpret when multiplicity handling is not reported. The value of the secondary results is therefore contextual and descriptive unless the analysis plan establishes a confirmatory role.

Finally, planning illustrations based on observed effects should remain separate from trial design claims. They are not observed power and do not reveal the original rationale for enrollment. Keeping design assumptions, sponsor analyses, recalculated comparisons, and descriptive safety counts distinct makes the statistical evidence easier to assess without overstating what the registry can support.

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.

Several limitations are specific to the posted record. The primary outcome is described as a median time with a full range, and the sponsor reports rank-based median-difference comparisons. The registry does not provide a full time-to-event analysis specification or additional detail on how censoring at 99 days was incorporated. Medians and ranges alone do not describe the complete event-time distributions.

The listed sponsor analyses cover 300 mg versus 600 mg intravenous zanamivir and 600 mg intravenous zanamivir versus oral oseltamivir. A sponsor-reported primary comparison of 300 mg zanamivir with oseltamivir is not included in the record. The available comparisons should not be used to reconstruct that missing contrast. The registry also does not state a non-inferiority or equivalence objective or provide a margin.

The primary description refers to an influenza-positive population and specific subgroups, but subgroup sizes and subgroup-specific estimates are not reported in the registry record. It is not possible to evaluate the precision or representativeness of those subgroup analyses from the information recorded. The handling of missing outcomes, withdrawals, and sensitivity analyses is also not reported.

There are numerous listed outcomes, while multiplicity procedures and a hierarchy of testing are not reported. This limits the confirmatory interpretation of secondary findings. For safety, event counts are descriptive; the record does not provide enough information about exposure time, event timing, or detailed adjudication to support a full comparative safety analysis from these counts alone.

Finally, the enrollment is reported, but the original sample-size justification and design assumptions are not. Planning illustrations computed from observed effects cannot replace those missing design details or be interpreted as observed power.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study of Intravenous Zanamivir Versus Oral Oseltamivir in Adults and Adolescents Hospitalized With Influenza — NCT01231620.
• Registry record last updated 2018-10-15; status last verified 2018-03.
• Results first posted 2017-11-20.
• 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

NCT01231620 is a completed phase III randomized, quadruple-masked, parallel-group study of intravenous zanamivir versus oral oseltamivir in hospitalized adults and adolescents with influenza. The registry lists 626 enrolled participants and three regimens: intravenous zanamivir 300 mg twice daily, intravenous zanamivir 600 mg twice daily, and oral oseltamivir 75 mg twice daily, each paired with matching placebo as appropriate.

The primary outcome was time to clinical response in participants with confirmed influenza, measured in days and defined by vital-sign resolution or hospital discharge. The sponsor reported Wilcoxon rank sum comparisons for two treatment pairings; both confidence intervals included zero. The registry also lists respiratory, functional, symptom, mortality, and safety outcomes. Multiplicity handling, detailed missing-data methods, and the original sample-size justification are not reported in the registry record, limiting interpretation beyond the posted estimates.