A Multi-dose Study of ARC-520 in Patients With Hepatitis B 'e' Antigen (HBeAg) Positive, Chronic Hepatitis B Virus Infection: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Multi-dose Study of ARC-520 in Patients With Hepatitis B 'e' Antigen (HBeAg) Positive, Chronic Hepatitis B Virus Infection (NCT02604212), a randomized phase 2 study. The primary outcome is Change From Baseline in Quantitative Hepatitis B Surface Antigen (Log qHBsAg) at Day 113.

2. Clinical Question

3. Trial Design

This was a randomized, parallel-group, phase 2 study of ARC-520 in patients with HBeAg-positive chronic hepatitis B virus infection. The registry describes the study as multicenter, double-blind, and placebo-controlled, and labels masking as quadruple. It does not identify which participant, clinician, or assessment roles were included in that masking designation. The study was sponsored by Arrowhead Pharmaceuticals.

The record lists enrollment of 32 participants. The study started in 2015-11 and had a primary completion date of 2016-12. Its status is terminated. The reason for termination and any consequences for planned recruitment or follow-up are not reported in the registry record, so the status alone does not establish why the study ended or how termination affected the analysis.

The four described groups were placebo low-dose comparator, placebo high-dose comparator, ARC-520 1.0 mg/kg, and ARC-520 2.0 mg/kg. Both active doses were administered once every 4 weeks for 4 doses. Placebo was administered on the same stated schedule. All groups were paired with entecavir or tenofovir at the listed regimens. The record does not report how participants were assigned between the two background antiviral regimens, whether allocation was balanced across those regimens, or whether results were analyzed separately by background treatment.

The primary outcome was change from baseline in log quantitative hepatitis B surface antigen at Day 113. Although the broader study description refers to safety and efficacy, the registry record also lists 18 posted secondary outcomes, including changes over time and category-based changes at multiple visits. The primary analysis results are model-based estimates, not simple summaries of every randomized participant. The stated analysis population required dosing and valid outcome measurements at baseline and at least one post-baseline time point on or after Day 15.

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: Intent to treat population: all participants who received at least 1 dose of study drug and had valid qHBsAg values at baseline and at least one time point on or after the Day 15 visit.

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 Change From Baseline in Quantitative Hepatitis B Surface Antigen (Log qHBsAg) at Day 113, measured over Baseline, Day 113.

The primary endpoint measures the change from baseline to Day 113 in log quantitative hepatitis B surface antigen, reported in log IU/mL. Quantitative hepatitis B surface antigen is a laboratory measure, and expressing its change on a logarithmic scale describes change in the log-transformed measurement rather than a change in the original units. The registry gives least-squares mean changes, which are model-based estimates adjusted according to the stated repeated-measures model.

Serial measurement is useful when an outcome is assessed repeatedly during follow-up: it can describe the pattern of measurements across visits and use the available observations within the specified model. The primary analysis included baseline as a continuous covariate and treatment and visit as fixed factors, together with treatment-by-visit and baseline-by-visit interactions. The reported contrast at Day 113 therefore reflects the model’s estimated difference at that visit, rather than an unadjusted subtraction of the posted arm means.

The endpoint is a surrogate laboratory measure in this trial record. The posted Day 113 results do not by themselves describe symptoms, liver outcomes, or other clinical endpoints, and they do not establish how long any measured change persisted beyond the specified follow-up. The clinical interpretation of a particular magnitude of log change is not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the least-squares mean represent?

A least-squares mean is an estimate from a statistical model rather than a simple arithmetic mean of the observed values. In this study, the endpoint description says the model included baseline as a continuous covariate, treatment and visit as fixed factors, and interactions of treatment with visit and baseline with visit. The fitted estimates therefore account for the covariates and repeated-visit structure specified in that model. Such adjustment can make the estimated treatment contrast differ from the subtraction of the displayed arm estimates, particularly when the model accounts for baseline and observations across visits.

The reported primary endpoint is a continuous outcome: the measurement can take values along a scale, rather than representing an event count or a binary outcome. A mean difference is therefore an interpretable summary of the modeled separation between groups on the endpoint scale. Here the scale is log IU/mL, and the reported contrasts are least-squares mean differences at Day 113. The registry’s arm estimates and pairwise contrasts should be read as related outputs from the stated analysis, not as independently observed quantities.

Why use a mixed effect model for repeated measures?

In general, repeated-measures models are designed for outcomes collected at multiple visits from the same participants. Measurements from one participant are usually related to one another, so treating every visit as if it came from a different person can misrepresent uncertainty. A mixed effect model for repeated measures can model the visit pattern while accounting for within-participant dependence, according to its specified covariance and estimation assumptions.

The registry states that treatment and visit were fixed factors and that treatment interacted with visit. This allows the estimated treatment contrast to vary by visit, rather than forcing a single common contrast over the entire follow-up period. The baseline-by-visit interaction allows the relationship between baseline value and follow-up measurement to vary across visits. The record does not state the covariance structure, estimation method, degrees-of-freedom procedure, or other implementation details. Those details are not reported in the registry record, so the model cannot be fully reproduced from the description alone.

How should the standard errors and confidence intervals be read?

A standard error describes the estimated sampling uncertainty of an estimate under the fitted model. It is not the spread of individual participant outcomes. The arm estimates have standard errors of 0.073 for placebo, 0.065 for ARC-520 1.0 mg/kg, and 0.061 for ARC-520 2.0 mg/kg. The pairwise analyses instead report model-based differences and 95% confidence intervals. A confidence interval gives a range of effect values reasonably compatible with the data and model under the method’s assumptions; it is not a guarantee that the true effect falls in that particular interval.

For ARC-520 1.0 mg/kg versus placebo, the confidence interval runs from -0.243 to 0.159. Since that interval includes zero, it includes no difference as well as differences in either direction. The reported 95% interval for ARC-520 2.0 mg/kg versus placebo, -0.638 to -0.237, lies below zero. The corresponding interval for ARC-520 1.0 mg/kg versus ARC-520 2.0 mg/kg, -0.569 to -0.223, also lies below zero. These interval positions describe statistical uncertainty on the model’s outcome scale; clinical importance cannot be determined from interval position alone.

What do the p-values say?

In general, a p-value measures how incompatible the observed data are with a specified null hypothesis under the analysis assumptions. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. The reported p = 0.681 for the 1.0 mg/kg versus placebo comparison indicates that the data do not provide strong evidence against the no-difference hypothesis in that comparison. The p-values of < .0001 for the other two contrasts indicate stronger evidence against their respective no-difference hypotheses under the reported MMRM analysis. The record does not report an adjustment for testing multiple contrasts or outcomes.

What does the analysis population mean?

The stated intent-to-treat population was restricted to participants who received at least 1 dose and had a valid baseline value and at least one valid qHBsAg value on or after Day 15. This is the registry’s operational definition for this analysis. It is not simply a statement that every randomized participant contributed to the primary outcome analysis. The record does not report how many randomized participants were excluded from this population, the reasons for any exclusions, or the number with a Day 113 measurement specifically.

In general, repeated-measures models can use available post-baseline data without requiring every participant to have a measurement at every visit, but validity depends on the model and missing-data assumptions. The registry does not state how missing values were handled, whether missingness was assessed, or whether sensitivity analyses were performed. The estimated contrast should therefore be understood as conditional on the analysis population, model specification, and unreported missing-data details.

What can be inferred about planning?

The registry describes a planning illustration based on the observed effect, a two-sided significance level of 0.05, equal allocation, and a normal approximation. The illustration is not the trial’s own design justification and is not observed power. Specific planning counts are not reported in the registry record. The record also does not report the trial’s original sample-size justification, assumptions used in that justification, or any interim-analysis plan. Planning illustrations based on observed effects are retrospective scenarios and should not be mistaken for evidence about the probability that this completed study detected an effect.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Placebo", "ARC-520 1.0 mg/kg", "ARC-520 2.0 mg/kg")
n   <- c(6, 6, 8)
m   <- c(-0.104, -0.146, -0.542)
sp  <- c(0.073, 0.065, 0.061)
sd  <- sp * sqrt(n)      # standard errors converted to SDs
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Placebo  6  -0.104  0.179  -0.292  0.084
CARM  ARC-520 1.0 mg/kg  6  -0.146  0.159  -0.313  0.021
CARM  ARC-520 2.0 mg/kg  8  -0.542  0.173  -0.686  -0.398

11. Analyses Reported by the Sponsor

The registry reports the primary comparisons using a mixed effect model for repeated measures. The posted results are ARC-520 1.0 mg/kg versus placebo, ARC-520 2.0 mg/kg versus placebo, and ARC-520 1.0 mg/kg versus ARC-520 2.0 mg/kg, each with a model-based least-squares mean difference, 95% confidence interval, and p-value. These are the results to use when describing the trial’s reported primary analysis.

A separate unadjusted recalculation from posted summary statistics is not reported in the registry record, so there is no recalculated estimate to compare numerically with the model-based results. An unadjusted calculation would not necessarily reproduce the MMRM contrasts: the posted model includes baseline and repeated visits, while a simple calculation from arm summaries would not reproduce all those adjustments and covariance assumptions. The record does not flag a discrepancy between sponsor-reported and recalculated results.

The comparisons should therefore be described as the registry-posted MMRM analyses, not as independently verified recalculations. The reported intervals and p-values characterize those specific model contrasts. They do not resolve questions that the registry does not document, including the full model implementation, multiplicity handling, missing-data sensitivity analyses, or the clinical importance of the measured changes.

12. Multiplicity, Missing Data and Analysis Populations

The registry lists 19 outcome measures and 48 analyses, including the primary Day 113 endpoint, secondary measurements over time, and category-based changes at several visits. This creates multiple opportunities for comparisons and statistical tests. The record does not report a multiplicity-control procedure, a hierarchy of hypotheses, or which analyses were intended to be confirmatory. The posted p-values should therefore be interpreted as results for the individual reported contrasts, not as evidence that the overall family of outcome and treatment comparisons was controlled at a stated error rate.

The primary analysis population included participants who received at least 1 dose and had valid qHBsAg values at baseline and at least one time point on or after Day 15. Although it is called the intent-to-treat population in the registry, that operational description includes post-randomization requirements for dosing and outcome availability. The record does not report the number randomized, the number meeting each population criterion, participant flow by arm, or reasons for missing or invalid measurements. The relationship between total enrollment of 32 and the arm counts shown for the primary endpoint is not explained in the registry record.

The MMRM description indicates repeated observations and specifies baseline, treatment, visit, and interactions, but does not describe the missing-data assumptions or method used to handle incomplete follow-up. The record does not report sensitivity analyses testing alternative assumptions about missing outcomes, nor does it state whether missingness was related to treatment or observed outcomes. Consequently, the reader cannot assess from the registry alone how robust the Day 113 estimates are to missing-data assumptions.

The trial status is terminated, but the reason for termination and whether it affected recruitment, follow-up, or planned analyses are not reported in the registry record. No interim-analysis plan or stopping rule is reported. Those omissions limit interpretation of the analysis history and the extent to which the posted comparisons reflect a prespecified testing plan.

13. Safety Results

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

Other (non-serious) adverse events

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

Most frequent non-serious adverse event terms

The safety record reports serious adverse event counts of 0 in each of the four safety groups. The numbers at risk were 6 for the placebo low-dose comparator, 5 for the placebo high-dose comparator, 10 for ARC-520 1.0 mg/kg, and 11 for ARC-520 2.0 mg/kg. These figures describe the recorded serious adverse event counts in those groups; they do not establish that an event could not occur or that the groups have equivalent safety profiles.

The record also gives counts for other adverse events: 4 of 6 in the placebo low-dose comparator, 0 of 5 in the placebo high-dose comparator, 5 of 10 in the ARC-520 1.0 mg/kg group, and 3 of 11 in the ARC-520 2.0 mg/kg group. These are counts as recorded, not estimates of comparative risk. The registry does not provide event-term details for those counts in the information, so their nature, severity, timing, and relationship to study treatment cannot be characterized from these figures alone.

The death values are not reported in the registry record. Empty serious-event term listings should not be used to infer details beyond the posted serious-event counts. Safety groups are defined differently from the primary endpoint arms, including separate low-dose and high-dose placebo comparator groups. The safety counts should therefore be read using their own group definitions rather than assumed to match the primary outcome analysis population.

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

The posted results describe the modeled change in log quantitative hepatitis B surface antigen at Day 113 among participants meeting the stated analysis-population criteria. The 1.0 mg/kg comparison with placebo had an estimated difference of -0.042 and a 95% confidence interval that included zero. The 2.0 mg/kg comparison with placebo had an estimated difference of -0.438 and a confidence interval entirely below zero. The direct comparison of the two ARC-520 dose groups also had a confidence interval entirely below zero. These results characterize statistical differences on the specified laboratory scale under the reported MMRM.

They do not establish the clinical importance of a change in qHBsAg, because the registry does not state a clinically important threshold. They do not establish an effect on symptoms, liver complications, viral transmission, or other clinical outcomes. Those outcomes are not reported as results in the primary endpoint information. Nor do the Day 113 estimates establish the persistence of the measured changes after the stated assessment time.

The findings also do not establish equivalence between ARC-520 1.0 mg/kg and placebo. A confidence interval spanning zero can reflect uncertainty about direction and magnitude; it is not proof of identical outcomes. Similarly, a small p-value does not itself quantify clinical relevance or the chance that a finding will replicate.

The registry does not report multiplicity procedures, missing-data sensitivity analyses, a detailed model covariance structure, or the reason the study was terminated. The primary endpoint counts are smaller than total enrollment, but the record does not explain the participant flow underlying that difference. These gaps constrain how broadly the posted estimates can be interpreted. They should be read as results from a phase 2 randomized study with the stated design and analysis population, not as a complete account of all potential benefits, harms, or long-term outcomes.

15. Why This Trial Matters Statistically

This trial offers a useful example of why the estimated treatment contrast, its confidence interval, and its analysis method all matter. The primary outcome is continuous and measured repeatedly, while the posted results include model-based estimates adjusted for baseline and visit. Comparing only the arm-level means would not capture the full repeated-measures analysis or its model assumptions. The registry’s side-by-side dose comparisons also make clear that an estimate for one dose should not be taken as a summary of all doses.

The contrasts illustrate how evidence can differ across comparisons within the same study. The reported 1.0 mg/kg versus placebo interval includes zero, while the 2.0 mg/kg versus placebo interval lies below zero. The direct dose comparison is also reported. These are distinct inferential questions, and their interpretation depends on the contrast, scale, uncertainty, and testing context—not merely on whether a p-value is small.

The study also highlights the importance of analysis-population definitions. The registry’s intent-to-treat description requires dosing and valid outcome data at baseline and after baseline. That requirement matters when reading the primary endpoint counts alongside total enrollment. Without participant flow and missing-data details, readers cannot fully trace which participants contributed to the estimates or assess the implications of unavailable measurements.

Finally, the registry gives a terminated status and multiple outcomes and analyses, but does not report the reason for termination or a multiplicity plan. These omissions do not invalidate the posted estimates, but they limit what can be inferred about the study’s planned testing framework and completeness. Careful statistical reading separates the reported laboratory endpoint findings from questions the record does not answer.

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 available record leaves important design and analysis details unspecified. It does not report the reason for termination, the original sample-size justification, or an interim-analysis plan. It also does not state the allocation ratio, the allocation procedure beyond randomization, or which roles were covered by quadruple masking. Those omissions restrict assessment of how the trial was conducted and whether the posted analysis followed a fully documented prespecified plan.

The primary analysis population required treatment exposure and valid outcome measurements. Although the record labels it intent to treat, it does not provide participant flow or explain how many randomized participants met those criteria. Enrollment is listed as 32, while the primary endpoint arms report counts of 6, 6, and 8. The reasons those analysis counts do not account for all enrollment are not reported in the registry record. The number of participants with Day 113 values specifically is also not given.

The model description provides its principal covariates and interactions, but not the repeated-measures covariance structure or other computational details needed for full reproduction. The missing-data assumptions and sensitivity analyses are not reported. Multiplicity control is likewise not reported despite numerous listed outcomes and analyses. The status of terminated does not by itself explain whether planned follow-up or analyses were changed.

The endpoint is a log-transformed laboratory measure, and the registry does not define a clinically meaningful change. The primary result therefore cannot, by itself, answer broader clinical questions or establish persistence beyond Day 113. Safety reporting provides group-level counts but not the detailed adverse-event terms, timing, severity, or treatment attribution needed for a comprehensive safety interpretation. Death values are not reported in the registry record.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: A Multi-dose Study of ARC-520 in Patients With Hepatitis B 'e' Antigen (HBeAg) Positive, Chronic Hepatitis B Virus Infection — NCT02604212.
• Registry record last updated 2026-01-13; status last verified 2025-12.
• Results first posted 2019-04-16.
• 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

NCT02604212 is a randomized, parallel-group, phase 2, placebo-controlled study of ARC-520 in patients with HBeAg-positive chronic hepatitis B virus infection. The study used placebo or ARC-520 at 1.0 mg/kg or 2.0 mg/kg with entecavir or tenofovir. Total enrollment was 32; the registry lists the study as terminated, with a start date of 2015-11 and primary completion in 2016-12.

The primary outcome was change from baseline in log qHBsAg at Day 113, analyzed using MMRM. The posted contrast for ARC-520 1.0 mg/kg versus placebo was -0.042, with a 95% confidence interval from -0.243 to 0.159 and p = 0.681. The contrast for ARC-520 2.0 mg/kg versus placebo was -0.438, with a 95% confidence interval from -0.638 to -0.237 and p < .0001. The record does not report the reason for termination, multiplicity handling, or detailed missing-data analyses.