A Study of ABI-H2158-containing Regimens in Participants With Chronic Hepatitis B Virus Infection: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of ABI-H2158-containing Regimens in Participants With Chronic Hepatitis B Virus Infection (NCT04398134), a randomized phase 2 study. The primary outcome is Percentage of Participants With Adverse Events.

2. Clinical Question

The clinical question is framed around safety: the primary outcome records whether each participant had at least one adverse event during treatment, up to 72 weeks. The results include four arms: ABI-H2158 plus ETV and placebo plus ETV within each HBeAg population. The posted counts therefore describe several groups rather than a single pooled treatment-versus-control comparison.

The registry also lists secondary outcomes concerning drug concentrations, changes in HBV pgRNA and HBsAg, and reductions in HBV DNA or HBV pgRNA below assay limits. These outcomes address pharmacokinetics and antiviral activity, but the primary result is the adverse-event count. The registry record does not report the full statistical plan for how these outcomes were to be analyzed or how comparisons across outcomes were to be handled.

3. Trial Design

This was a randomized, parallel-group Phase 2a study of ABI-H2158-containing regimens in chronic hepatitis B. The official title describes a multicenter, placebo-controlled, multiple-cohort study. The record lists enrollment of 88 participants and identifies Assembly Biosciences as the sponsor. The listed start date is 2020-08-28, the primary completion date is 2021-10-14, and the status is terminated.

The treatment groups were organized by HBeAg status. In the HBeAg-positive population, 32 participants were assigned to ABI-H2158 plus ETV and 11 to placebo plus ETV. In the HBeAg-negative population, the corresponding group sizes were 33 and 12. The arm descriptions state that ABI-H2158 or matching placebo was taken once daily for 72 weeks, with ETV once daily for 96 weeks. The study summary describes ABI-H2158 administration for up to 72 weeks in combination with ETV.

The registry record lists double masking, while the official title describes the study as single-blind. These labels are not fully consistent, and the record does not clarify the difference. Allocation is listed as randomized and the model as parallel. More detailed information about sequence generation, concealment, stratification, and the roles of masked participants or staff is not reported in the registry record.

The registry record does not report the reason for termination, the intended sample-size justification, or whether the planned enrollment was reached before termination. Those omissions limit interpretation of whether the posted group sizes reflect the original design or a shortened study. The posted counts can describe the observed groups, but the design details needed to assess the conduct and intended scope of the trial are incomplete.

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:

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 Percentage of Participants With Adverse Events, measured over Up to 72 weeks.

The primary endpoint is the number of participants with one or more adverse events while on treatment with study drug, with a timeframe of up to 72 weeks. It is a participant-level count: each participant is classified according to whether at least one adverse event occurred. The endpoint does not count every event separately, and it does not by itself describe event severity, duration, recurrence, or causal relation to treatment.

This type of endpoint is used to summarize how many people in each group experienced an adverse event during the stated observation period. Reporting the count alongside the number of participants at risk lets readers see the scale of the observed data in each arm. The registry reports 18 of 32 participants in the ABI-H2158 plus ETV HBeAg-positive arm and 6 of 11 in the corresponding placebo arm. For the HBeAg-negative population, it reports 19 of 33 and 6 of 12, respectively.

A participant-level any-event endpoint is useful as a broad safety summary, but it collapses different kinds of events into the same outcome category. A participant with one event and a participant with multiple events both count as participants with an adverse event. The endpoint therefore needs to be read alongside the separate serious-adverse-event information and any detailed event listings. It does not establish that an event was caused by the assigned regimen.

The registry record does not report further endpoint definitions such as rules for coding or attribution. The phrase “while they were on treatment” supplies the stated context, but the detailed analysis rules for this endpoint are not reported in the registry record.

7. Results Posted in the Registry

• ABI-H2158 Plus ETV (HBeAg Positive Population): 56.2% (18/32) versus 54.5% (6/11) on control; risk difference 1.7 percentage points (95% CI -32.4 to 35.8), risk ratio 1.03 (0.55–1.92); unadjusted two-proportion z-test p = 0.922.
• ABI-H2158 Plus ETV (HBeAg Negative Population): 57.6% (19/33) versus 54.5% (6/11) on control; risk difference 3.0 percentage points (95% CI -30.9 to 36.9), risk ratio 1.06 (0.57–1.95); unadjusted two-proportion z-test p = 0.861.
• Placebo Plus ETV (HBeAg Negative Population): 50.0% (6/12) versus 54.5% (6/11) on control; risk difference -4.5 percentage points (95% CI -45.4 to 36.3), risk ratio 0.92 (0.42–2.00); unadjusted two-proportion z-test p = 0.827.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is binary at the participant level: a person either had at least one adverse event during the stated on-treatment period or did not. The registry presents the number with events in each arm. This is different from an event-rate outcome, which would count repeated events or relate events to person-time. The endpoint description does not indicate that repeated events were counted separately.

What does a risk difference describe?

In general, a risk difference compares the event proportions in two groups by subtracting one proportion from the other. A value of zero corresponds to equal observed proportions; positive and negative values indicate the direction of the contrast according to the order of the groups. The posted risk differences are 0.017 for the HBeAg-positive ABI-H2158 group compared with the HBeAg-positive placebo group, 0.030 for the HBeAg-negative ABI-H2158 group compared with the HBeAg-positive placebo group, and -0.045 for the HBeAg-negative placebo group compared with the HBeAg-positive placebo group. These are estimates from posted counts, not adjusted estimates from a reported model.

The reference group matters. The first comparison uses the placebo plus ETV HBeAg-positive arm as its control. The second and third comparisons also use that same reference arm in the posted calculation inputs. Consequently, the HBeAg-negative ABI-H2158 comparison is not an ABI-H2158-versus-placebo comparison within the HBeAg-negative population. Comparing the two HBeAg-negative arms directly would be a different contrast, and it is not among the listed comparisons.

What do confidence intervals convey?

In general, a confidence interval expresses uncertainty around an estimate under the assumptions of the method used to construct it. A narrow interval indicates greater precision than a wide interval, all else equal. The reported intervals for the risk differences are broad and include zero. Their span indicates that the posted counts do not pin down the direction or size of the differences precisely. The registry record does not identify the interval construction method or its confidence level.

The posted results also include risk ratios and odds ratios with intervals. In general, a risk ratio compares the proportions between groups, while an odds ratio compares their odds. These measures use different scales and can give a different numerical impression from a risk difference. Their interpretation depends on the group order and event definition. They do not remove the underlying uncertainty from the small group sizes. The values should be read as accompanying measures of the same posted counts, not as independent evidence.

Why are the p-values unadjusted?

A p-value summarizes how incompatible observed data are with a specified null model, under that model's assumptions. It is not the probability that a treatment works, nor does it measure the magnitude or clinical importance of an effect. The reported values are identified as unadjusted. The registry record does not report whether an adjustment for multiple comparisons was planned or whether any of these comparisons were primary inferential tests. Interpretation should therefore remain cautious, especially because the record lists multiple populations and outcomes.

In general, small samples make estimates sensitive to the outcome of a small number of participants. Sparse binary outcomes can also produce unstable ratio estimates. The posted arms contain 11, 12, 32, and 33 participants. The confidence intervals reflect substantial uncertainty in these comparisons. A non-small p-value does not demonstrate equivalence or prove that the groups have the same event experience; establishing equivalence would require a prespecified margin and an analysis designed for that question, neither of which is reported in the registry record.

What can be said about the analysis plan?

The registry record reports no analysis entries. It lists 13 secondary outcomes as posted, including pharmacokinetic measures, changes in viral markers, and reductions below assay limits. However, the statistical plan, analysis populations, missing-data procedures, and multiplicity strategy are not reported in the registry record. It is therefore not possible to determine from the record whether the posted calculations match a prespecified analysis, how participants with incomplete follow-up were handled, or whether the outcomes were analyzed in an intention-to-treat or other population.

General statistical considerations help explain why these details matter. Missing outcome data can alter the group proportions if participants with missing status differ from those observed. Multiple comparisons can increase the chance of apparently notable results when many hypotheses are examined. Prespecified analysis populations and multiplicity procedures help define how results should be interpreted. Without those details, the posted counts and recalculated comparisons can still be described, but their relationship to the original analysis plan remains uncertain.

The trial's own original sample-size assumptions and justification are not reported in the registry record. The planning illustrations associated with the posted data use a stated set of assumptions, described separately, and should not be treated as evidence that the trial was designed around those assumptions.

10. Confidence Intervals and Precision

The precision of the adverse-event comparisons is limited by the number of participants in each arm. The HBeAg-positive comparison includes 32 participants in the ABI-H2158 group and 11 in the placebo group; the HBeAg-negative ABI-H2158 group has 33 participants, while its placebo group has 12. With counts of this size, each participant represents a substantial part of the observed group total, so a small change in event classification could shift the estimated proportions.

The risk-difference intervals are broad. For the HBeAg-positive ABI-H2158 comparison, the interval runs from -0.324 to 0.358. For the HBeAg-negative ABI-H2158 comparison against the HBeAg-positive placebo reference, it runs from -0.309 to 0.369. The placebo HBeAg-negative versus placebo HBeAg-positive comparison has an interval from -0.454 to 0.363. Each interval includes zero and spans values on both sides of it. Thus, the posted estimates are not precise enough to establish a clear direction for these differences.

In general, an interval describes uncertainty in an estimate, not the range of individual participant outcomes and not a guarantee that the true value lies within the reported bounds. The registry record does not state the confidence level or the method used to calculate these intervals. The intervals are based on posted summary counts and should be interpreted with that limitation in mind.

Precision is distinct from whether an estimate is clinically important. A narrow interval can still center on a small difference, while a broad interval can include both small and larger differences. The posted intervals leave considerable uncertainty about the size and direction of the adverse-event differences. They do not establish that the regimens are equivalent, and the registry record does not report an equivalence or non-inferiority framework.

11. 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 sample-size illustrations use the observed effect as a starting point and show planning for 100% of that effect, 75% of the observed effect, and 50% of the observed effect. For the binary adverse-event outcome, the figures are participants per arm, not total events. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; the effect to detect is the stated fraction of the observed effect. The original trial's own design assumptions are not reported in the registry record.

The 80% power column describes the number of participants per arm in a hypothetical design that would have an 80% chance of detecting the specified effect under the planning assumptions, if that effect and the assumptions were correct. The 90% power column gives the corresponding planning count for a higher target chance of detecting that effect. These are design illustrations, not counts of participants who were observed in the trial. For example, the planning row for 75% of the observed effect in the HBeAg-positive ABI-H2158 comparison gives 23,754 participants per arm for 80% power and 31,799 participants per arm for 90% power.

Smaller effects require more participants to detect under otherwise unchanged assumptions. This explains why the planning counts increase from the 100% row through the 75% and 50% rows. The estimates are especially sensitive to the observed effect used as their starting point; the posted comparisons are uncertain, and the record does not report that these planning values informed the original trial design. This is not observed power. It is a hypothetical planning exercise and should not be read as a measure of whether the completed trial had a meaningful chance of detecting its actual effect.

The curve should be read in the same planning context: it illustrates how the required sample size changes with the effect size under the stated assumptions. It does not describe an additional observed outcome. Time-to-event planning tables, if present, would report total events under the stated calculation description, but the adverse-event endpoint is binary and its planning values are participants per arm.

12. 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 = 0.545; treatment proportion p₁ = 0.562; average p̄ = 0.554.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.7030 = 1.3778.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.7029 = 0.5916.
4. n = (1.3778 + 0.5916)² / (0.0170)² = 13,348.5, rounded up to 13,349 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 <- 0.545454545454545                          # control proportion
p1 <- 0.5625                          # 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(4398134)
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 = 13,349 (80% power), 17,870 (90% power)
 75% of observed effect: n per arm = 23,754 (80% power), 31,799 (90% power)
 50% of observed effect: n per arm = 53,493 (80% power), 71,612 (90% power)
power.prop.test: 13348.5 
Simulated power at n = 13,349 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.

13. Multiplicity, Missing Data and Analysis Populations

The registry record lists a primary adverse-event outcome and multiple secondary outcomes. The secondary topics include trough plasma concentrations and trough-to-peak concentration ratios for ABI-H2158 and ETV, changes in HBV pgRNA, reductions in HBV DNA or HBV pgRNA below assay limits, and change in HBsAg. The record reports 13 secondary outcomes as posted. Considering several outcomes and group comparisons raises a general multiplicity issue: when many hypotheses are examined, the chance of obtaining an apparently notable result by chance can increase.

The registry record does not report a multiplicity procedure, such as a hierarchy of tests or an adjustment to significance levels. It also does not report whether the secondary outcomes were confirmatory or exploratory, how the two HBeAg populations were incorporated into a testing strategy, or whether any subgroup comparisons were prespecified. The unadjusted p-values accompanying the recalculated adverse-event comparisons should therefore not be interpreted as adjusted evidence across the full set of outcomes and contrasts.

Missing-data rules are also not reported in the registry record. The outcome timeframe is stated as up to 72 weeks, but the record does not explain how participants with incomplete follow-up, interrupted treatment, or missing adverse-event status were handled in the analysis. Nor does it report whether analyses used all randomized participants, treated participants, or another analysis population. These details affect which participants contribute to an estimate and how missing observations might influence a result.

The registry reports zero analysis entries, while the posted summary counts permit simple unadjusted comparisons. The absence of analysis entries does not show that no analysis was conducted; it means the record does not provide those details. Without the prespecified statistical plan and analysis-population definitions, the posted calculations cannot be verified against the original analytic strategy, and the extent of any adjustments or sensitivity analyses is not reported in the registry record.

14. 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: depression (9.1%), tibia fracture (8.3%), altered state of consciousness (3.1%).

The serious-adverse-event figures are counts of participants affected in each arm, rather than counts of all serious events. The registry reports 1 affected participant among 32 in the ABI-H2158 plus ETV HBeAg-positive group, 1 among 11 in the corresponding placebo group, 0 among 33 in the ABI-H2158 plus ETV HBeAg-negative group, and 1 among 12 in the corresponding placebo group. Deaths are reported as 0 in each arm.

The record also lists depression at 9.1%, tibia fracture at 8.3%, and altered state of consciousness at 3.1% among the top serious adverse-event terms. Those percentages are presented as registry values; the record excerpt does not give the denominator or additional detail needed to interpret them in relation to each treatment group. They should not be assigned to a particular arm without that information.

These serious-event summaries are distinct from the primary endpoint, which counts participants with any adverse event, whether or not serious. The two summaries cannot be directly compared as if they measured the same outcome. In general, serious adverse events are a specific category within safety reporting, while the primary any-event endpoint is broader. The small numbers and limited detail mean these figures describe the posted record but do not establish causation or provide a complete safety profile.

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

The posted results describe adverse-event counts in four groups and show that the number of participants with at least one event was 18 of 32 and 6 of 11 in the HBeAg-positive ABI-H2158 and placebo groups, respectively, and 19 of 33 and 6 of 12 in the HBeAg-negative groups. The recalculated risk-difference intervals are broad and include zero. They do not define a precise direction or size of difference in these data.

These results do not establish that ABI-H2158 caused, prevented, or had no effect on adverse events. The observed outcome is an all-event participant count, and the registry does not provide detailed causal attribution or a complete analysis plan. The comparisons are unadjusted recalculations from summary counts, and the record does not report a sponsor analysis entry for them. The results should not be treated as evidence of equivalence; no equivalence margin or corresponding design is reported.

The counts also do not show whether particular event types, repeated events, timing, severity, or treatment exposure differed between groups. The serious-adverse-event counts and death counts provide additional descriptive information, but they are sparse and do not resolve the uncertainty in the broader endpoint. The listed serious-event terms have percentages but lack sufficient context in the record to assign them to treatment groups or infer comparative risks.

The study's terminated status is a registry fact, but the reason for termination and its implications for follow-up are not reported in the registry record. It would be inappropriate to infer a cause from the status alone. Likewise, the posted findings do not answer the secondary questions about antiviral activity or pharmacokinetics simply because those outcomes are listed; their full results and analyses are not included in the information described in the primary endpoint comparison.

In general, a clinical trial's interpretation depends on its design, execution, analysis plan, and completeness of follow-up, as well as the observed estimates. The available record supports a description of the posted counts and their uncertainty, but several details needed for a fuller assessment are not reported in the registry record.

16. Why This Trial Matters Statistically

This trial is statistically instructive because its primary outcome is simple to state but easy to overinterpret. A count of participants with any adverse event compresses a broad range of event experiences into a yes-or-no classification. It gives a useful high-level summary, while leaving important distinctions such as event type, severity, recurrence, and timing outside the endpoint itself.

The four-arm structure also makes the comparison question important. The HBeAg-positive ABI-H2158 comparison uses the HBeAg-positive placebo group as a reference. The posted HBeAg-negative ABI-H2158 comparison uses that same reference rather than the HBeAg-negative placebo group. A reader who assumes that every treatment arm is compared only with the matching population's placebo arm could misread the calculated contrast. Stating the comparator is essential to understanding what a risk difference means.

The small group sizes and broad intervals illustrate why a point estimate alone is not enough. The posted risk differences are close to zero in the reported scale, but their intervals extend meaningfully in both directions. The p-values do not establish equivalence, and a lack of a small p-value does not resolve uncertainty about the size of a possible difference. Precision and design purpose must be considered together.

The planning illustrations further show how sample-size requirements depend on the effect chosen for detection. They are separate from the trial's original design rationale, which is not reported in the registry record. The illustrations should not be confused with observed power or treated as evidence that the study was planned to detect the displayed effects.

Finally, the record's missing details about analysis populations, missing data, multiplicity, and the termination rationale demonstrate how registry summaries can support useful descriptive reading while leaving important analytic questions unanswered. Statistical interpretation is strongest when estimates, uncertainty, comparison definitions, and reporting limitations are read together.

17. 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 arm sizes are small, and the event counts are correspondingly limited. This makes the estimates sensitive to the outcomes of individual participants and contributes to the broad intervals. The registry record does not report the original sample-size justification, so it is not possible to assess whether the planned enrollment was chosen to estimate safety outcomes with a specified precision.

The comparisons are unadjusted calculations based on posted summary counts. The registry record does not report a sponsor-reported analysis for these contrasts, an interval method, or a confidence level. The comparison structure also uses the HBeAg-positive placebo group as the reference for the listed contrasts, including those involving HBeAg-negative arms. This limits direct interpretation as within-population treatment comparisons.

The record does not report the analysis populations, missing-data rules, exposure-adjusted event summaries, or detailed follow-up completeness. Although the primary endpoint is described as events while on treatment up to 72 weeks, the record does not provide further rules for treatment interruptions, event attribution, or participants with incomplete observation. The terminated status is listed, but the reason is not reported in the registry record.

The masking descriptions are not fully consistent: the record lists double masking, while the official title calls the study single-blind. The record does not clarify that discrepancy. Also, the serious-adverse-event term percentages are reported without enough contextual detail in this record excerpt to identify their denominators or group allocation. These limitations do not invalidate the posted counts, but they narrow what can be concluded from them.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: A Study of ABI-H2158-containing Regimens in Participants With Chronic Hepatitis B Virus Infection — NCT04398134.
• Registry record last updated 2022-09-15; status last verified 2022-08.
• Results first posted 2022-08-26.
• 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.

21. Record Summary

The registry record describes a randomized, parallel-group Phase 2a study of ABI-H2158 plus ETV versus placebo plus ETV in participants with chronic hepatitis B. It reports four arms divided by HBeAg status and lists 88 participants enrolled. The status is terminated, with primary completion dated 2021-10-14.

The primary outcome is the number of participants with one or more adverse events while on treatment, up to 72 weeks. Posted arm counts range from 6 to 19 participants with an event, with arm sizes from 11 to 33. Recalculated risk-difference intervals are broad and include zero. The registry record does not report the original sample-size rationale, analysis populations, missing-data procedures, multiplicity strategy, or reason for termination. The results therefore support descriptive reading of the posted counts, with substantial uncertainty and important analytic details not reported in the registry record.