Prophylactic Antiviral Therapy in Patients With Current or Past Hepatitis B Virus Infection Receiving Anti-Cancer Therapy for Solid Tumors: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Prophylactic Antiviral Therapy in Patients With Current or Past Hepatitis B Virus Infection Receiving Anti-Cancer Therapy for Solid Tumors (NCT03887702), a randomized phase 3 study. The primary outcome is Number of Participants With Adverse Liver Outcome.

2. Clinical Question

The clinical rationale described in the registry is that people with chronic or past HBV undergoing cancer therapy may have liver changes, and antiviral treatment may help prevent HBV worsening or returning. This page concerns what the trial record documents, rather than whether one strategy should be chosen in clinical practice. It is important to keep the two cohort-specific questions separate: prophylaxis before reactivation was assigned to a chronic-HBV cohort, while the populated results are shown for a cohort in which the listed strategies began treatment upon indication or followed usual care.

3. Trial Design

This was a phase III randomized, single-masked, parallel trial sponsored by SWOG Cancer Research Network. Its title describes prophylactic antiviral therapy for patients with current or past HBV infection receiving anti-cancer therapy for solid tumors. The registry lists four treatment arms and an enrollment of 4. The record gives a start date of 2020-01-17, a primary completion date of 2022-12-20, and a terminated status. The reason for termination is not reported in the registry record.

The design separated treatment strategies by cohort. Cohort 1 was intended for chronic HBV infection. Its prophylactic arm used entecavir, tenofovir alafenamide, or tenofovir disoproxil fumarate beginning immediately or within 42 days after the initial chemotherapy dose. Its other arm used those antivirals after HBV reactivation during chemotherapy. Cohort 2 had an arm treating after reactivation and an arm in which the physician selected anti-HBV therapy under usual care. The arm descriptions state that treatment could continue for up to 6 months after the last chemotherapy dose or a maximum of 24 months, subject to the conditions described there.

The posted primary outcome population notes that no participants were registered to Cohort 1, Arms 1 and 2. The outcome table lists no participants in those arms and lists 2 participants in each Cohort 2 arm. Thus the recorded results do not populate all intended treatment groups. Randomization, masking, and the parallel structure describe the planned design, but the registry record does not report allocation procedures, the masking procedure, or how treatment assignments were implemented. No protocol-specific explanation of the design or sample-size justification is 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

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: There were no participants registered to the Chronic HBV infection cohort \[Cohort 1, Arms 1 and 2\]

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: There were no participants registered to the Chronic HBV infection cohort \[Cohort 1, Arms 1 and 2\]

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants With Adverse Liver Outcome, measured over From randomization to 24 months.

The primary endpoint counts participants with an adverse liver outcome from randomization to 24 months. The registry defines an adverse outcome as liver-related death or liver failure not due to disease progression in the liver. The specified liver failure events include ascites, hepatic encephalopathy, or impaired hepatic synthetic function, defined by total bilirubin of at least 5 mg/dL or INR of at least 2.0. Sites were to assess the endpoint, followed by adjudication by the research oversight team through blinded review.

This endpoint focuses on clinically consequential liver outcomes rather than on a laboratory marker or HBV reactivation alone. In principle, counting affected participants gives a direct summary of how many people in each group experienced the defined event during the stated follow-up period. The endpoint's composite definition also means that different events can qualify: liver-related death and specified forms of liver failure are counted under the same outcome. The registry record does not report separate primary-endpoint counts for each component.

For the populated Cohort 2 arms, the posted outcome count is 0 among 2 participants in each arm. No participants are listed for the Cohort 1 arms. The record does not provide a comparative estimate or analysis of the endpoint. The count therefore describes the small observed groups, while leaving considerable uncertainty about the occurrence of these outcomes in the broader populations of interest.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The posted primary endpoint is a binary participant-level outcome: each participant is either counted as having an adverse liver outcome or is not counted as having one during the specified follow-up. The registry gives the outcome as a count of participants, not as a continuous measurement. For this kind of result, a table commonly shows the number of participants with an event and the number assessed in each group. Those counts describe the observed data; they do not, by themselves, quantify how much an intervention changes the outcome relative to a comparator.

What does a comparative analysis add?

A comparative analysis estimates a contrast between groups. For a binary outcome, analysts may report an absolute difference in event probabilities or a relative measure, alongside a confidence interval. A confidence interval conveys the range of effect sizes compatible with the data under the method's assumptions. Its width reflects precision: less information generally produces less precise estimates. A hypothesis-test result, if reported, addresses compatibility with a specified null hypothesis; it does not replace the effect estimate or show the clinical importance of an effect.

In this record, no comparative analysis is posted, and the registry notes that values are missing or not estimable for a two-arm comparison. The posted information therefore does not support interpretation of an effect estimate, confidence interval, or test result. The count of zero events in each populated arm is a descriptive observation, not a direct estimate of equivalence. When events are rare and group sizes are small, uncertainty can remain substantial even when the observed count is zero.

Why are the cohort comparisons distinct?

The trial lists strategies within two cohorts. In Cohort 1, the intended comparison was prophylactic antiviral treatment versus antiviral treatment upon indication; the record states that no participants were registered to that cohort. In Cohort 2, the populated arms were treatment upon indication and usual care. Because the treatment contrasts and cohorts differ, combining them into a single comparison would require a justified analysis plan and appropriate data. Such a plan is not reported in the registry record. The displayed arm counts should therefore be read within their stated cohort and strategy rather than treated as evidence for a pooled effect.

How should the planning illustration be read?

The sample-size assumptions describe a planning illustration based on an observed effect, with a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is described as a stated fraction of the observed effect. For non-time-to-event outcomes, the illustration is expressed as participants per arm for a target power. These figures are design illustrations, not observed power and not evidence that the trial had a particular probability of detecting an effect. The registry record does not report the trial's own sample-size justification. Given the missing comparative results and the very small populated groups, the illustration should not be mistaken for a measure of the evidence obtained in this trial.

What does the analysis status imply?

The registry reports 0 analyses. That means no analysis results are posted in the record for the listed outcome, not that a particular statistical method was used and found no effect. The statistical model, adjustment variables, handling of repeated observations, and approach to interim analyses are not reported in the registry record. Because these details are absent, readers cannot assess whether an analysis addressed potential differences in participant characteristics or deviations from assigned treatment.

More generally, analysis choices should match the outcome, design, and question. A binary endpoint can be summarized by counts and proportions, while comparative inference depends on the chosen estimand and analysis method. An estimand clarifies which participants, treatment conditions, outcome, and follow-up period define the effect being sought. Here, the registry supplies an endpoint definition and timeframe, but a detailed estimand and a statistical analysis plan are not reported in the registry record. The available counts support a limited descriptive summary only.

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("Arm 3 (TAF, TDF, Entecavir: Upon Indication)", "Arm 4 (Anti-Viral Therapy: Usual Care)")
n      <- c(2, 2)
events <- c(0, 0)
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  Arm 3 (TAF, TDF, Entecavir: Upon Indication)  2  0  0.0  0.0  84.2  0.0  65.8  0.8  70.8
ARM  Arm 4 (Anti-Viral Therapy: Usual Care)  2  0  0.0  0.0  84.2  0.0  65.8  0.8  70.8

11. Multiplicity, Missing Data and Analysis Populations

The record lists a primary endpoint and three secondary outcomes: HBV reactivation, HBV flare, and a combined endpoint of adverse liver outcomes and HBV reactivation. Multiple outcomes can raise a multiplicity issue: if investigators test many hypotheses, the chance of at least one apparently notable result can increase. Methods such as prespecifying a hierarchy or adjusting inference may address that concern, but the registry record does not report a multiplicity strategy. The three secondary outcomes are listed, but the posted record does not include comparative analyses for them.

The handling of missing outcome data is also not reported in the registry record. The available primary outcome table lists 0 participants in each Cohort 1 arm and 2 participants in each Cohort 2 arm. The record explains that no participants were registered to the chronic HBV cohort. It does not state whether the Cohort 2 participants completed the full follow-up period, whether any assessments were unavailable, or how incomplete follow-up would have been treated in an analysis. A count of no recorded events should not be taken to establish complete outcome ascertainment unless that information is available.

The analysis population is not described beyond the stated endpoint population and the arm counts. The registry record does not report whether analyses were intended to follow randomized assignment, include only participants who received treatment, or use another population definition. Nor does it report how participants who changed treatment or did not follow the assigned strategy would be handled. These details matter because different analysis populations can answer different questions.

With no analyses posted, there is no basis to assess the impact of multiplicity, missingness, or analysis-population choices on a comparative estimate. These omissions limit reproducibility and interpretation; they do not establish that a specific analysis decision was made.

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 0% reporting threshold.

The safety table lists serious and other adverse-event counts by arm. For Arms 3 and 4, 2 participants were at risk for each category, and 0 were recorded as affected by serious adverse events or other adverse events. The table shows 0 deaths in each of these arms. Arms 1 and 2 have no participants listed as at risk, consistent with the absence of registered participants in the chronic HBV cohort.

These figures are descriptive counts, not comparative safety estimates. The small numbers mean that observing no reported events does not establish that the risk is absent. The registry record does not provide event rates with confidence intervals, detailed event descriptions, or a comparative safety analysis. It lists no top serious adverse-event terms. The record also does not explain the safety observation period or report whether adverse-event collection was complete for every participant.

In particular, the zero values for Arms 1 and 2 should not be read as evidence that no adverse events occurred among treated participants: no participants were listed at risk in those arms. For Arms 3 and 4, the displayed zeros mean no affected participants were recorded in the listed categories among the participants counted at risk. This is a description of the posted safety data, not a conclusion about safety in a broader population.

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

The posted results support a limited factual statement: among the two participants listed in each populated Cohort 2 arm, the primary endpoint count was 0, and the safety table lists no serious adverse events, other adverse events, or deaths in those arms. The registry record lists no participants in the Cohort 1 arms and explicitly states that no participants were registered to the chronic HBV cohort. These data do not answer the planned prophylaxis comparison in that cohort.

The results do not show that antiviral strategies are equivalent, that one strategy is better than another, or that adverse liver outcomes are impossible. Such conclusions require comparative information with adequate precision, and the registry posts no comparative analyses or effect estimates. A zero count in a small group is compatible with uncertainty about the event frequency beyond the observed participants. Likewise, absent event counts in unpopulated arms do not represent observed zero-risk groups.

The record's status is terminated, but the reason for termination is not reported in the registry record. The record does not explain whether enrollment ended early, whether recruitment challenges occurred, or whether other circumstances led to termination. Those possibilities should not be inferred. Nor does the posted record explain whether participants had complete follow-up through the primary endpoint's 24-month timeframe.

Accordingly, the clearest interpretation is descriptive rather than causal. The trial record documents its intended strategies and endpoint, a very small set of populated-arm observations, and no posted comparative analyses. It does not establish an estimate of treatment benefit or harm. The data can help readers understand what was recorded, while leaving the central comparative question unresolved.

14. Why This Trial Matters Statistically

This trial is statistically instructive because it distinguishes a carefully defined clinical question from the evidence available to answer it. The registry describes randomized treatment strategies, a primary endpoint with a defined follow-up period, and a plan to have site assessments reviewed by a blinded research oversight team. Yet the record lists only 4 participants, no participants in the chronic HBV cohort, and no comparative analyses. A design's intended rigor cannot compensate for a lack of outcome information in the groups needed to answer its principal question.

The record also illustrates why enrollment and outcome counts need context. A table with zero events can look conclusive at a glance, but the number of people observed determines how much precision is possible. Here, each populated arm has 2 participants, and the prophylaxis arms have none. The absence of an event in the populated arms is therefore a narrow observation, not a stable estimate of comparative risk. Confidence intervals and effect estimates would ordinarily help show uncertainty, but none are posted.

Finally, the trial demonstrates the value of transparent reporting about analysis populations, missing data, multiplicity, and termination. These details allow readers to judge how closely results answer the intended question. The registry record does not report those statistical-plan details or the reason for termination. Identifying these gaps is not a judgment about the trial's conduct; it clarifies which conclusions the posted record can and cannot support. For biostatisticians and clinical researchers, the essential lesson is to keep the planned design, observed data, and inferential claims distinct.

15. Limitations

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

The most important trial-specific limitation is the small enrollment. The record lists 4 participants in total, with 2 in each of the populated Cohort 2 arms. No participants were registered to Cohort 1, the chronic HBV cohort that included the prophylactic strategy and its treatment-upon-indication comparator. Consequently, the principal prophylaxis question described by the trial title is not represented in the posted participant outcomes.

The registry status is terminated, and the reason is not reported in the registry record. It is also not reported whether the listed participants completed the endpoint follow-up, whether all outcome assessments were available, or whether follow-up differed by arm. The endpoint timeframe is from randomization to 24 months, but the record does not provide participant-level follow-up summaries. These omissions make it difficult to assess outcome ascertainment.

No comparative analyses are posted. The record does not report effect estimates, confidence intervals, hypothesis tests, analysis populations, or a detailed statistical analysis plan. It also does not report a protocol-specific sample-size justification, a multiplicity approach, or methods for missing data. Without these details, readers cannot determine how the intended randomized comparison would have been analyzed or how precise any treatment contrast might be.

The populated Cohort 2 strategies are also not identical: one describes antiviral therapy after reactivation, and the other leaves treatment to physician discretion. The posted counts alone cannot separate the effect of those strategies from other influences or establish a causal difference. The safety information is similarly limited by the small numbers and lack of detailed event terms. These limitations constrain interpretation of this registry record; they do not establish why the trial ended or what unreported analyses might have shown.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Prophylactic Antiviral Therapy in Patients With Current or Past Hepatitis B Virus Infection Receiving Anti-Cancer Therapy for Solid Tumors — NCT03887702.
• Registry record last updated 2024-05-20; status last verified 2024-05.
• Results first posted 2024-05-20.
• 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

NCT03887702 is a phase III randomized, single-masked, parallel trial sponsored by SWOG Cancer Research Network. It evaluated antiviral strategies for people with current or past HBV infection receiving anti-cancer therapy for solid tumors. The record lists an enrollment of 4 and a terminated status. Its primary endpoint was the number of participants with an adverse liver outcome from randomization to 24 months.

No participants were registered to the chronic HBV cohort, which contained the prophylactic and treatment-upon-indication arms. The posted outcome table lists 2 participants in each of the two Cohort 2 arms and 0 primary-endpoint events in each. The safety table lists 0 affected participants in the serious and other adverse-event categories, and 0 deaths, in those populated arms. No comparative analyses are posted. The reason for termination, analysis plan, missing-data approach, and multiplicity strategy are not reported in the registry record. The findings should therefore be read as small descriptive counts, not as a comparative estimate of treatment effect.