Efficacy and Safety of AG10 in Subjects With Transthyretin Amyloid Cardiomyopathy: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Efficacy and Safety of AG10 in Subjects With Transthyretin Amyloid Cardiomyopathy (NCT03860935), a randomized phase 3 study. The primary outcome is A Hierarchical Combination of All-Cause Mortality, Cumulative Frequency of CV-related Hospitalization, Change From Baseline in NT-proBNP and Change From Baseline in 6MWT at the Last Available Visit Where Both Subjects Had Non-missing Assessments..

2. Clinical Question

The registry describes a phase 3 randomized study in people with symptomatic transthyretin amyloid cardiomyopathy. Its primary question is therefore about the comparative ordering of participants across several clinical and functional outcomes, not about a single isolated measure. Death and cardiovascular hospitalization are considered before changes in NT-proBNP and six-minute walk distance. Heart transplantation and implantation of a cardiac mechanical assist device were handled in the same manner as death for this analysis.

The posted primary results report 63.7% of wins for acoramidis and 35.9% for placebo, with a reported superiority test of p < 0.001. These percentages describe the win shares generated by the hierarchical comparisons; they are not percentages of participants who survived, avoided hospitalization, or improved on every component. The registry does not report a confidence interval or a single effect estimate for this endpoint.

3. Trial Design

This was a phase 3, randomized, double-blind, placebo-controlled, parallel-group study. The registry also describes masking as quadruple. Acoramidis HCl 800 mg was administered twice daily, and placebo was intended to match it. The study is titled the ATTRibute-CM Trial and evaluates efficacy and safety in symptomatic transthyretin amyloid cardiomyopathy.

The registry lists enrollment as 632 and status as completed. The recorded study start date is 2019-03-19, and the primary completion date is 2023-05-11. The sponsor is Eidos Therapeutics, a BridgeBio company. The primary endpoint analysis lists 409 participants in the acoramidis group and 202 in the placebo group. These endpoint analysis counts are not interchangeable with the overall enrollment figure; the registry record does not explain the difference in this summary.

The primary outcome is a hierarchical combination of mortality, cardiovascular-related hospitalizations, NT-proBNP change, and six-minute walk distance change. Its stated timeframe is baseline up to Month 30, using the last available visit at which both participants had non-missing assessments. The arm descriptions also state that the six-minute walk test primary outcome would be assessed at the end of 12 months. The registry record does not clarify how that arm-description timing relates to the primary endpoint timeframe extending up to Month 30.

Secondary outcomes listed in the registry include six-minute walk distance, Kansas City Cardiomyopathy Questionnaire Overall Score, serum TTR level, and all-cause mortality by Month 30, with heart transplant or cardiac mechanical assist device included in that mortality outcome. Numerical results are posted for three of these outcomes. A numerical result for the listed all-cause mortality secondary outcome is not reported in the registry record.

Further design details—including the randomization ratio, the full protocol, the sample-size justification, and any interim analyses—are not reported in the registry record. The randomized, masked, parallel-group description supports a comparison between the assigned groups, but it does not by itself establish the details of the analysis plan or how every participant was handled after randomization.

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: The mITT population includes subjects who meet the definition of ITT which includes all randomized subjects who received at least one dose of IMP and have at least one post baseline efficacy assessment as well as an additional requirement which is to have a baseline eGFR \>= 30 mL/min/1.73 m\^2 . Subjects in this population were analyzed according to their assigned randomized treatment.

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: Safety Population

6. Primary Endpoint

The registry lists the primary outcome as A Hierarchical Combination of All-Cause Mortality, Cumulative Frequency of CV-related Hospitalization, Change From Baseline in NT-proBNP and Change From Baseline in 6MWT at the Last Available Visit Where Both Subjects Had Non-missing Assessments., measured over Baseline up to Month 30.

The primary endpoint combines several outcomes through the Finkelstein-Schoenfeld method. For each pairwise comparison within strata, participants are ordered first by all-cause mortality, then by cumulative frequency of cardiovascular-related hospitalizations, then by change from baseline in NT-proBNP, and finally by change from baseline in six-minute walk distance. A participant judged to do better receives a win, a participant judged to do worse receives a loss, and a tie contributes neither a win nor a loss. Heart transplantation and implantation of a cardiac mechanical assist device were treated in the same manner as death.

This approach is used when a trial seeks to compare groups across outcomes that differ in clinical importance and measurement scale. The hierarchy gives priority to earlier components: a difference on a higher-priority outcome can determine the pairwise ordering before lower-priority measures are considered. This is different from averaging all components into one score or interpreting the endpoint as a single continuous measurement.

The reported unit is percent of wins from the win-ratio procedure. Acoramidis had 63.7% of wins and placebo had 35.9%; the reported test was p < 0.001. These figures summarize the results of the pairwise process. They do not reveal how much of the result came from each component, and the registry record does not provide a confidence interval or a separate component-level breakdown for the primary endpoint. The endpoint timeframe is baseline up to Month 30, at the last available visit where both participants had non-missing assessments.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary analysis uses the Finkelstein-Schoenfeld method, a procedure for comparing groups across a prioritized set of outcomes. Rather than combining different measurements into a common numerical scale, it compares participants pairwise. Within strata, each participant is compared with each other participant; the hierarchy determines which outcome is considered first. A better outcome produces a win, a worse outcome produces a loss, and a tie produces neither. The registry specifies mortality first, followed by cumulative cardiovascular-related hospitalizations, NT-proBNP change, and six-minute walk distance change. Heart transplantation and cardiac mechanical assist device implantation are treated as death.

What does a win share represent?

The primary endpoint unit is percent of wins from the win-ratio procedure. The posted shares are 63.7% for acoramidis and 35.9% for placebo. These are summaries of the hierarchical pairwise comparisons. They are not event rates, probabilities of an individual participant benefiting, or percentages of participants who improved on all measures. The registry reports a superiority test using the Finkelstein-Schoenfeld method, with p < 0.001, but gives no primary confidence interval or single effect estimate. The win shares therefore describe the direction and distribution of pairwise wins, while the reported test addresses evidence against its null hypothesis under the method.

How should the secondary continuous outcomes be read?

For the posted secondary continuous outcomes, the registry reports group least-squares means and standard errors, alongside sponsor analyses using mixed models. A least-squares mean is a model-based group estimate, not necessarily the raw arithmetic average. A standard error describes the uncertainty in an estimated mean; it is not the spread of individual participant values. The sponsor-reported analyses give least-squares mean differences, confidence intervals, and p-values. These estimates arise from the reported mixed-model analyses and should not be assumed to be identical to a simple contrast calculated from the group summaries.

The recalculated comparisons are simple, unadjusted re-calculations from posted summary statistics. Their confidence intervals use an approximate standard deviation reconstructed from the posted standard errors, and the associated p-values are unadjusted. This is a different calculation from the sponsor’s mixed-model analysis. Differences between the resulting intervals can reflect the distinct statistical procedures and assumptions; the re-calculation does not replace the sponsor’s reported analysis. The record does not provide the full model specification needed to reproduce the sponsor analyses in detail.

What does a confidence interval add?

In general, a confidence interval gives a range of values compatible with the data and the statistical model at the stated confidence level. Its width reflects uncertainty: narrower intervals indicate greater precision under those assumptions, while wider intervals indicate less precision. The sponsor’s secondary analyses report 96% confidence intervals. The registry does not report a confidence interval for the primary win-share result. A confidence interval is not a guarantee that the true value lies within the displayed range, and it does not account for sources of bias or model misspecification that are not represented in the calculation.

What does a p-value say?

In general, a p-value measures how unusual data at least as incompatible with a null hypothesis as the observed data would be, assuming the null hypothesis and the test assumptions. It is not the probability that the null hypothesis is true, and it does not express clinical importance. The primary test is reported as p < 0.001. The sponsor’s secondary tests are also reported as p < 0.001. Recalculated secondary p-values are reported as p <0.001 and are unadjusted. Those values should be understood in the context of their respective methods and the missing information about multiplicity.

How should the planning illustrations be understood?

The planning illustrations use observed effects as inputs and are not the trial’s sample-size justification. They assume a two-sided significance level of 0.05, equal allocation, and a normal approximation, with the target effect set to the stated fraction of the observed effect. The resulting figures are participants per arm for the stated power target. They are illustrative planning calculations, not observed power, and they do not show that the trial was designed using those assumptions. The trial’s own sample-size justification is not reported in the registry record.

Over-dispersion is a general concern when count outcomes vary more than a simple statistical model expects. In that circumstance, a model that assumes less variability can understate uncertainty and make evidence appear more precise than it is. The primary outcome includes cumulative hospitalization frequency, but the registry’s posted primary result is a hierarchical comparison, not a stand-alone count model. The record does not provide enough information to assess any separate over-dispersion modeling for that component.

10. Analyses Reported by the Sponsor

The sponsor’s reported primary analysis uses the Finkelstein-Schoenfeld method and reports a superiority test of p < 0.001. The primary result is a hierarchical win-share endpoint, so the posted record does not provide a single mean difference or a confidence interval that could be directly checked against a simple re-calculation. The primary comparison is therefore best described using the registry’s own reported method and result.

For the secondary continuous outcomes with numerical results, the sponsor reports mixed-model analyses and least-squares mean differences with 96% confidence intervals. The re-calculated contrasts are simple, unadjusted calculations from the posted group summaries; they use an approximate standard deviation reconstructed from the standard errors. They are not a reproduction of the sponsor’s full model. The numerical estimates and intervals may differ because the calculations and assumptions differ. The record does not include sufficient model details to reconstruct every element of the sponsor’s analysis.

The registry’s sponsor-reported secondary results give p < 0.001 for each of the three numerical comparisons. The recalculated p-values are p <0.001. These are not competing versions of the same fully specified analysis: the sponsor’s model-based analyses take precedence as the trial’s reported analyses, while the re-calculations provide an unadjusted check based on summary statistics. The record does not report a discrepancy flag for these results.

11. Multiplicity, Missing Data and Analysis Populations

The primary endpoint uses a hierarchy among its components: mortality is considered before cardiovascular-related hospitalization, followed by NT-proBNP change and six-minute walk distance change. That hierarchy specifies how outcomes determine pairwise wins and losses. It is not, by itself, a description of how statistical testing across the primary and secondary endpoints was controlled.

The registry lists four secondary outcomes, and numerical results are available for three. It does not report whether secondary hypotheses were tested in a prespecified sequence, whether p-values were adjusted for multiple comparisons, or whether any secondary results were designated as confirmatory. The recalculated secondary p-values are unadjusted. Without a reported multiplicity plan, nominal p-values should not be interpreted as if they necessarily controlled the chance of false-positive findings across all outcomes.

For the primary endpoint, the stated timeframe is baseline up to Month 30, using the last available visit where both participants had non-missing assessments. The registry describes the mITT population as including randomized participants who received at least one dose and had at least one post-baseline efficacy assessment, together with a baseline eGFR requirement of at least 30 mL/min/1.73 m^2. The remainder of the population description is incomplete in the registry record. The handling of participants who did not meet these criteria is not fully detailed in the posted summary.

Using the last available visit with non-missing assessments identifies how the primary pairwise comparisons drew on available observations. It does not, on its own, explain why data were missing, how missingness was handled in each component, or whether sensitivity analyses assessed the effect of missing data. Those details are not reported in the registry record. The sponsor’s mixed-model approach for secondary outcomes is identified, but its model specification and missing-data assumptions are not reported in this summary.

Details about interim analyses, stopping rules, the complete statistical analysis plan, and the sample-size justification are also not reported in the registry record. Their absence from this record does not show that such procedures were or were not used. It limits the ability to assess how the reported analyses relate to a complete prespecified plan.

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: cardiac failure (18.5%), atrial fibrillation (7.1%), cardiac failure acute (6.6%), acute kidney injury (5.0%), covid-19 pneumonia (3.8%).

The registry posts serious adverse event counts for 230 of 421 participants in the acoramidis group and 137 of 211 participants in the placebo group. It also posts other adverse event counts for 410 of 421 and 199 of 211, respectively. Death counts are 84 in the acoramidis group and 55 in the placebo group. These are group counts, not adjusted comparisons, and the record does not provide a statistical analysis of these safety figures in this summary.

The registry lists cardiac failure at 18.5%, atrial fibrillation at 7.1%, cardiac failure acute at 6.6%, acute kidney injury at 5.0%, COVID-19 pneumonia at 3.8%, and fall at 3.1% among the leading serious adverse event terms. These listed percentages should be read as registry-reported event information; the summary does not provide the exposure-adjusted rates or further detail needed to compare risks over time.

Raw counts alone do not establish that an event was caused by study treatment or that the groups had comparable exposure time. The posted summary does not give event timing, severity breakdowns beyond the serious-event category, participant-level overlap among event terms, or an adjusted safety comparison. The figures describe what was recorded, not a causal assessment.

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

The posted results support a statistical difference in the primary hierarchical comparison between the randomized groups, with the win-share result favoring acoramidis and a reported p-value of p < 0.001. The result pertains to the combined ordering of mortality, cardiovascular-related hospitalizations, NT-proBNP change, and six-minute walk distance change, according to the stated hierarchy. It does not identify which component contributed most to the result or give a single estimate of the magnitude of benefit across the composite.

The secondary results with numerical data have a direction consistent with that primary comparison across several distinct measures. They remain secondary analyses, and the registry record does not state whether multiplicity was controlled. A consistent direction does not remove uncertainty about missing-data handling, model assumptions, or the contribution of separate endpoint components.

These registry results do not establish that every participant experienced improvement, that all components moved in the same direction, or that the observed comparisons apply to people outside the analyzed population. The posted primary analysis population has eligibility conditions beyond randomization, and the record does not provide a complete account of exclusions or missing observations. The results also do not provide a complete account of longer-term safety or a causal interpretation of individual adverse events.

Statistical significance is not the same as clinical importance. The primary p-value indicates evidence against the test’s null hypothesis under the analysis assumptions; it does not show the probability that the treatment is effective or quantify the practical importance of the observed comparison. The absence of a primary confidence interval and a single primary effect estimate limits assessment of the precision and magnitude of that comparison.

The trial record therefore supports a clear description of the reported statistical findings, but not broader claims beyond the outcomes and analyses posted. The full protocol, complete statistical analysis plan, missing-data sensitivity analyses, and additional details of the sponsor’s models are not reported in the registry record.

14. Why This Trial Matters Statistically

This trial illustrates why a composite endpoint’s structure matters as much as its headline result. A hierarchical endpoint can prioritize outcomes by ordering pairwise comparisons, allowing mortality and hospitalization to take precedence over biomarker and functional measures. That makes the endpoint different from a conventional average and changes what its win shares mean. Readers should understand the ordering rule before treating the reported shares as a single measure of patient benefit.

It also shows why a statistically small p-value does not answer every interpretive question. The primary test is reported as p < 0.001, but the registry does not provide a primary confidence interval or a single effect estimate. Without those quantities, the evidence for a group difference is clear under the reported test, while the precision and magnitude of the composite comparison remain harder to judge.

The secondary results illustrate another common issue: model-based estimates and simple re-calculations are not interchangeable. The sponsor used mixed models for the secondary continuous outcomes, while the re-calculations use posted summary statistics and approximate dispersion. The difference is methodological, not necessarily evidence of a contradiction. A re-calculation can be useful as a transparent check, but it cannot substitute for the trial’s reported model.

Finally, outcome multiplicity and analysis populations affect the strength of conclusions. Several secondary measures are listed, but the registry record does not describe a multiplicity strategy, complete missing-data methods, or all details of the mITT population. These omissions do not invalidate the posted results, but they define what can be evaluated from the registry summary and where conclusions should remain cautious.

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 registry provides an incomplete description of the primary analysis population: the text states that the mITT population includes randomized participants who received at least one dose, had at least one post-baseline efficacy assessment, and met a baseline eGFR threshold of at least 30 mL/min/1.73 m^2, but the description ends before its full details are available. The primary analysis counts are 409 for acoramidis and 202 for placebo, while overall enrollment is listed as 632. The registry summary does not explain the difference between enrollment and these endpoint counts.

The primary endpoint is a composite hierarchical comparison, but the posted summary does not provide separate primary results for mortality, cardiovascular-related hospitalizations, NT-proBNP, or six-minute walk distance. It also does not provide a primary confidence interval or a single win-ratio estimate. This limits assessment of component contributions and the precision of the aggregate result.

The registry gives different timing descriptions: the primary endpoint timeframe is baseline up to Month 30, while the arm descriptions state that the 6MWT primary outcome will be assessed at the end of 12 months. The record does not explain how those statements align. The secondary results list Month 30 outcomes, but a numerical result for the all-cause mortality secondary outcome is not reported.

Further limitations include the absence of a reported multiplicity plan, full missing-data procedures, sensitivity analyses, detailed sponsor model specifications, sample-size justification, and interim-analysis information. The safety summary gives counts and selected serious adverse event terms but does not provide exposure-adjusted comparisons or enough detail for causal assessment. These are limitations of what is available in the registry record, not evidence that the omitted methods or analyses were not performed.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Efficacy and Safety of AG10 in Subjects With Transthyretin Amyloid Cardiomyopathy — NCT03860935.
• Registry record last updated 2024-06-27; status last verified 2024-06.
• Results first posted 2024-06-27.
• 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

The registry record describes NCT03860935, a completed phase 3 randomized, double-blind, placebo-controlled, parallel-group study of acoramidis for symptomatic transthyretin amyloid cardiomyopathy. The sponsor is Eidos Therapeutics, a BridgeBio company. Enrollment is listed as 632; the study start date is 2019-03-19, and primary completion is dated 2023-05-11.

The primary endpoint is a Finkelstein-Schoenfeld hierarchical combination of all-cause mortality, cardiovascular-related hospitalization frequency, NT-proBNP change, and six-minute walk distance change. The registry reports win shares of 63.7% for acoramidis and 35.9% for placebo, with p < 0.001. Numerical secondary results are posted for three of the four listed secondary outcomes. The record does not provide a numerical result for the listed Month 30 all-cause mortality secondary outcome, or a primary confidence interval. Details of multiplicity control, full missing-data handling, and the complete analysis plan are not reported in the registry record.