A Study Evaluating the Efficacy and Safety of AG-348 in Regularly Transfused Adult Participants With Pyruvate Kinase Deficiency (PKD): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study Evaluating the Efficacy and Safety of AG-348 in Regularly Transfused Adult Participants With Pyruvate Kinase Deficiency (PKD) (NCT03559699), a na phase 3 study. The primary outcome is Percentage of Participants Achieving a Reduction in Transfusion Burden in Part 2.

2. Clinical Question

The clinical question is about a change in transfusion burden within a treated group, not a direct comparison of AG-348 with another treatment. The endpoint classifies participants according to whether their transfusion use crossed a prespecified reduction threshold. It therefore captures the share meeting that criterion, rather than the average change in units for the group.

3. Trial Design

This was a phase 3, open-label, single-group study of AG-348 in regularly transfused adults with pyruvate kinase deficiency. It was multicenter, had no allocation to separate treatment arms, and used no masking. The registry record lists enrollment of 27 and a completed status. The study start date was 2018-06-26, and primary completion was 2020-11-12.

The study had two parts. In Part 1, the dose-optimization period, participants started AG-348 at 5 mg twice daily. The description says that doses were increased sequentially to 20 mg twice daily and then 50 mg twice daily depending on tolerance. Part 2 was a fixed-dose period in which participants received the optimized dose established during Part 1. The intervention description also identifies a 16-week period for the Part 1 dose levels. The registry record does not establish that every participant reached every dose level.

Because the study was single-group and open-label, it did not estimate a randomized between-group treatment contrast. The primary endpoint instead compared transfusion burden during the fixed-dose period with a historical burden for the same participants. The registry record does not report details of how the historical period was selected or how potential differences between the historical and on-study periods were handled.

The study’s listed conditions were pyruvate kinase deficiency and hemolytic anemia. Its secondary outcomes include transfusion use and episodes, transfusion-free status, normal hemoglobin concentrations, and adverse events. Those outcome labels indicate the range of questions recorded, but do not supply their observed results. The primary endpoint result itself is not 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.

Before assignment: Screening was done for a period of 8 weeks after the participant provided the informed consent. Investigators determined if the participants met all the inclusion criteria and none of the exclusion criteria to enroll in Part 1: Dose Optimization Period to receive AG-348 to determine the optimized dose followed by Part 2: Fixed Dose Period.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full analysis set included all participants who received at least 1 dose of study drug.

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: Full analysis set included all participants who received at least 1 dose of study drug.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants Achieving a Reduction in Transfusion Burden in Part 2, measured over From Part 2, Day 1 to Part 2 Week 24.

The primary endpoint was the percentage of participants achieving a reduction in transfusion burden during Part 2. A participant met the endpoint definition by having at least a 33% reduction in red blood cell units transfused during the fixed-dose period, standardized to 24 weeks, compared with that participant’s historical transfusion burden, also standardized to 24 weeks. The assessment window ran from Part 2 Day 1 to Part 2 Week 24.

This binary endpoint expresses how many participants met a threshold, rather than the size of each participant’s change or the group’s average change. A threshold can be clinically interpretable as a responder definition, but it also discards information about the distance from that threshold: participants just below and far below it are both counted as not meeting the endpoint.

Standardizing both periods to 24 weeks is intended to put transfusion counts on a comparable duration basis. That comparison still depends on the historical period being an appropriate reference for the on-study period. The registry record does not report details of the historical-period selection or the endpoint result. In particular, it does not report the observed percentage or its 95% confidence interval, so the magnitude and precision of the primary finding cannot be assessed from the record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a percentage endpoint represent?

The primary outcome is a binary responder measure summarized as a percentage of participants. Each participant is classified according to whether their standardized transfusion burden met the defined reduction threshold. The group percentage describes the share meeting that rule; it does not describe the average number of units avoided or the size of the reduction among responders.

Binary endpoints can make a threshold easy to communicate, but the classification depends on the chosen cutoff. Participants on opposite sides of the cutoff may have similar underlying changes, while participants on the same side may have quite different changes. For that reason, the responder proportion is best read alongside continuous measures and other outcomes when those results are available. The registry record lists additional transfusion outcomes, but does not report their results.

Why standardize both periods?

The endpoint compares transfusion counts from the fixed-dose period with historical counts, with each period standardized to 24 weeks. In general, standardization to a common duration makes rates or counts from periods of different lengths more comparable. It does not make the periods identical in other respects, and it does not remove potential differences in clinical status, observation, or practice. The registry record does not report the detailed method used to construct the historical reference.

How should a confidence interval be read?

The primary endpoint specifies a 95% confidence interval. In general, a confidence interval presents a range of values compatible with the observed estimate under the method’s assumptions, and communicates statistical uncertainty. A narrower interval generally indicates greater precision than a wider interval, but neither width nor location can be assessed without the actual limits. The registry record does not report the primary percentage or its confidence interval.

A confidence interval is not a guarantee that the true effect lies in the reported interval, nor does it address every source of uncertainty. In this study, uncertainty from a small single-group experience, the historical comparison, and the completeness of endpoint data would need consideration alongside sampling uncertainty. The record does not provide enough endpoint information to examine the interval’s precision.

What does the single-group design imply?

In a randomized comparison, allocation can support a causal contrast between groups, subject to the trial’s conduct and analysis. This study had no randomized comparator arm. Its primary endpoint instead compares an on-study period with a historical period. Within-participant comparisons can control for stable characteristics of a participant, but do not control automatically for changes over time or differences in how historical and on-study information was collected.

The study was open-label and unmasked. In general, knowledge of treatment can influence behavior, care decisions, or outcome assessment. The importance of these possibilities depends on the endpoint and the procedures used. The registry record does not report masking, adjudication, or other safeguards beyond identifying the study as open-label with no masking.

What statistical details are not available?

The registry record lists 9 outcome measures and 1 analysis, but does not report a statistical analysis plan or the method used to calculate the primary confidence interval. It does not report whether the analysis included covariate adjustment, how missing endpoint data were handled, or whether any interim analyses occurred. These details should not be inferred from the endpoint definition or study design.

In general, analysis choices matter because they determine which observations contribute, how uncertainty is estimated, and how multiple outcomes are interpreted. If multiple outcomes or analyses are considered, multiplicity can increase the chance of a nominally notable result occurring by chance. Whether multiplicity adjustments were planned or used is not reported in the registry record. The count of outcomes alone does not show that a particular adjustment was necessary or applied.

What is the role of planning illustrations?

Sample-size calculations can illustrate how many participants per arm might be needed under specified assumptions in a future two-arm study. Such calculations are planning illustrations, not observed evidence from this single-group trial and not a measure of the probability that the study detected an effect. The record’s planning assumptions describe a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and effects defined as fractions of the observed effect. They do not turn this study into a randomized two-arm comparison.

In general, a planning result is conditional on its assumed effect, variability, statistical method, and design. If an assumed effect is based on an observed estimate, that estimate may be imprecise, particularly in a small study. Any resulting sample-size illustration should therefore be read as conditional planning information rather than a claim about the study’s observed power or an assurance of future results.

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("AG-348")
n      <- c(27)
events <- c(10)
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  AG-348  27  10  37.0  19.4  57.6  21.5  55.8  21.5  55.9

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists 9 outcome measures and 1 analysis. It names the primary endpoint and several secondary outcomes, including annualized red blood cell units transfused, transfusion episodes, transfusion-free status, normal hemoglobin concentrations, and adverse events. The record does not report whether the outcomes were tested as a family, whether any multiplicity adjustment was used, or how the primary endpoint was prioritized within a prespecified statistical plan. Those details are not reported in the registry record.

In general, when many outcomes or hypotheses are examined, the chance of at least one apparently notable result can rise unless the analysis accounts for multiplicity. This general concern does not establish that a particular result in this study is spurious, nor does it show that an adjustment was required. Without the statistical plan and outcome results, the multiplicity approach cannot be evaluated.

The stated analysis population was the full analysis set, defined as participants who received at least one dose of study drug. The registry record does not report the number included in the primary endpoint analysis, the endpoint denominator, or the amount and pattern of missing outcome data. It also does not report how missing data were handled, including whether any imputation or sensitivity analyses were used. Therefore, the completeness and robustness of the primary endpoint analysis cannot be assessed from the record.

General statistical practice distinguishes the population defined for analysis from the people with usable data for a specific endpoint. That distinction matters because exclusions or missing observations can affect both the estimated percentage and its uncertainty. The enrollment figure should not be substituted for an endpoint denominator when that denominator is not reported.

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: blood triglyceride increased (3.7%), ovarian cyst (3.7%), renal colic (3.7%).

The registry record reports 27 participants at risk in the AG-348 group for the serious adverse event summary, with 3 participants affected. It reports 27 participants at risk and 27 affected for the other adverse event summary, and 0 deaths. These are the posted safety counts; the record does not provide event-level timing or individual clinical details in the information summarized for this analysis.

The listed serious adverse event terms are Blood Triglyceride Increased, Ovarian Cyst, and Renal Colic, each reported at 3.7%. These term-level percentages should be read as descriptions of reported events, not as a comparison with another treatment group. Because this was a single-group study, there is no concurrent randomized safety comparator in the registry record.

In general, adverse event summaries describe what occurred during observation; they do not by themselves establish that study treatment caused an event. The registry record does not supply a causal attribution analysis or comparative risk estimate. The safety figures also do not establish that an event will or will not occur in other patients. They should be interpreted as observations from this study, with the study’s design and recorded follow-up in mind.

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

The registry record establishes that this phase 3 study evaluated AG-348 in regularly transfused adults with pyruvate kinase deficiency using a single-group, open-label design. It defines a primary responder endpoint based on a reduction in red blood cell transfusion burden during a fixed-dose period compared with a standardized historical burden. It also reports the enrollment and a limited set of safety counts.

The record does not report the primary endpoint percentage or its 95% confidence interval. As a result, it cannot show from the recorded information how many participants met the endpoint, how large the estimated responder proportion was, or how precisely it was estimated. The absence of these results is not evidence that no reduction occurred; it means the primary outcome result is not reported in the registry record.

Even if the responder result were available, a single-group historical comparison would answer a different question from a randomized comparison. It would describe the observed change relative to a prior period, but would not isolate treatment effects from all other changes over time. Open-label conduct and the lack of a concurrent comparison group also limit causal interpretation.

The endpoint is a threshold-based summary, not a complete account of transfusion burden. It does not convey the full distribution of changes or the number of units for each participant. The registry record lists other outcomes that could add context, but their results are not reported in the record. Safety counts likewise describe events recorded in this group and do not establish causation or comparative safety.

Accordingly, the most defensible reading is descriptive: the registry record documents the study’s design, endpoint definition, and selected safety figures, while leaving the main efficacy estimate and several analysis details unreported. It does not support a treatment recommendation or a conclusion about comparative effectiveness.

14. Why This Trial Matters Statistically

This trial illustrates how a clinical question can be framed as a within-participant responder outcome when there is no concurrent comparison group. The primary endpoint translates a change in transfusion burden into a threshold-based percentage. That is a recognizable statistical estimand, but its interpretation depends on both the threshold definition and the quality of the historical reference.

The study also highlights the distinction between an outcome definition and an outcome result. A detailed endpoint description can clarify what investigators intended to measure, yet it cannot establish the observed finding. Here, the registry record gives the threshold, time window, standardization basis, and confidence-interval type, but does not report the primary percentage or interval. Without those, readers cannot assess the estimate’s magnitude or precision.

For biostatistical interpretation, the design matters as much as the endpoint. A single-group comparison with historical burden can describe change over time but cannot reproduce the protection against confounding that random allocation provides in a concurrent controlled study. Open-label conduct adds further context for interpreting outcomes and care processes. These are features to account for, not reasons to assume a particular result.

The record also shows why analysis-population details and missing-data reporting are essential. Enrollment is not necessarily the denominator for every endpoint, and the analysis population definition does not reveal how many participants had evaluable endpoint data. The registry record does not report that denominator or its missing-data methods.

Finally, the safety summary should be separated from causal claims. Counts and event terms describe what was recorded in the study group; without a comparator and fuller context, they do not establish excess risk or treatment causation. The trial is statistically instructive because it makes these distinctions important while leaving key efficacy and analysis details unreported.

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 primary limitation for interpreting efficacy is that the registry record does not report the primary endpoint result or its 95% confidence interval. It therefore does not permit assessment of the responder proportion, its precision, or its compatibility with a range of possible values. The endpoint definition alone cannot substitute for those estimates.

The study had a single group, no allocation to a comparator arm, and no masking. The primary endpoint used a historical transfusion burden rather than a concurrent randomized control. A historical comparison may be affected by changes over time or differences in measurement and care; how these issues were addressed is not reported in the registry record. The design supports descriptive interpretation more directly than causal attribution.

The responder definition uses a threshold of at least a 33% reduction in standardized transfusion burden. As with any dichotomized endpoint, it does not retain the full variation in transfusion changes. The registry record does not report the continuous distribution of those changes alongside the primary endpoint result.

The record defines the full analysis set as participants who received at least one dose, but does not report the primary endpoint denominator, missing-data extent, or missing-data handling. It also does not report the statistical analysis plan, the method for constructing the confidence interval, or whether multiplicity adjustments were made. These omissions limit assessment of analytic choices and robustness.

The safety summary provides selected counts and serious event terms, but does not provide a randomized safety comparison or detailed causal assessment. The study’s open-label, single-group design and the unreported efficacy result should remain central when interpreting the available registry information.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study Evaluating the Efficacy and Safety of AG-348 in Regularly Transfused Adult Participants With Pyruvate Kinase Deficiency (PKD) — NCT03559699.
• Registry record last updated 2022-01-04; status last verified 2021-12.
• Results first posted 2021-12-10.
• 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 a completed, multicenter, phase 3, open-label, single-group study of AG-348 in regularly transfused adults with pyruvate kinase deficiency. Enrollment was 27. Participants began with dose optimization, followed by a fixed-dose period using the optimized dose.

The primary endpoint was the percentage achieving at least a 33% reduction in red blood cell units transfused during the fixed-dose period, standardized to 24 weeks, relative to historical burden standardized to 24 weeks. The endpoint’s 95% confidence interval is specified, but its observed percentage and interval are not reported in the registry record.

The safety summary reports 3 participants affected by serious adverse events, 27 affected in the other adverse event summary, and 0 deaths. Listed serious event terms each appear at 3.7%. The record does not report the primary endpoint denominator, missing-data methods, or a statistical analysis plan.