Study Assessed the Safety and Efficacy of Eltrombopag in Chinese Refractory or Relapsed Severe Aplastic Anemia (SAA) Subjects.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study Assessed the Safety and Efficacy of Eltrombopag in Chinese Refractory or Relapsed Severe Aplastic Anemia (SAA) Subjects. (NCT03988608), a na phase 2 study. The primary outcome is Hematologic Response Rate at 6 Months (Week 26) by Investigator.

2. Clinical Question

3. Trial Design

This was a Phase II, non-randomized, open-label, multi-center study of eltrombopag in Chinese subjects with refractory or relapsed severe aplastic anemia. The registry lists a single-group model and an enrollment of 20. The study is marked completed. Its recorded start was 2019-12-09 and its primary completion was 2021-07-16. These dates describe the registry record; they do not by themselves provide information about the timing of individual participants’ follow-up.

All participants were assigned to the eltrombopag group; there was no listed comparison group. The treatment began at 25 mg/day on Day 1. According to the study description, the dose could be increased by 25 mg/day every 2 weeks according to platelet count, up to 150 mg/day. The registry record does not state the dose actually taken by each participant, the number who had dose adjustments, or the reasons for any changes.

The primary endpoint was assessed at 6 months (Week 26). The study description also says hematologic response was assessed at 3 months, 6 months, and 1 year after treatment started, corresponding to Week 13, Week 26, and Week 52. Other listed outcomes include blood-count changes, time to and duration of investigator-assessed response, and platelet transfusion frequency and amount. The registry record lists 18 outcome measures but does not provide the primary result or analysis results for these secondary outcomes.

Because the study was open-label and lacked a randomized comparator, its design is descriptive rather than a controlled test of relative treatment effects. Investigator assessment is part of the primary endpoint definition, and knowledge of treatment assignment may be relevant when interpreting an investigator-assessed outcome. The registry record does not report blinding procedures, a protocol-based sample-size justification, or a detailed statistical analysis plan.

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: One of the main criteria for enrollment was for Chinese patients aged ≥ 18 years, previously diagnosed with SAA and who had insufficient response following at least one treatment course in the period time of \> 6 months of immunosuppression with a regimen containing anti-thymocyte globulin (ATG), anti-lymphocyte Immunoglobulin (ALG), and/or cyclophosphamide, or alemtuzumab.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The Full Analysis Set (FAS) comprised all patients to whom study treatment had been assigned and received one dose of study treatment. The patients who discontinued from the trial before Week 26 were treated as non-responders.

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 (FAS) comprised all patients to whom study treatment had been assigned and received one dose of study treatment.

6. Primary Endpoint

The registry lists the primary outcome as Hematologic Response Rate at 6 Months (Week 26) by Investigator, measured over 6 months (Week 26).

The primary endpoint is the percentage of participants with a hematologic response at 6 months (Week 26), assessed by an investigator. A participant counted as a responder if they met any one of the specified International Working Group criteria for platelet count, hemoglobin, or neutrophil count. The endpoint therefore combines distinct ways of meeting a response definition; it is not a measure of change in one blood count alone.

The criteria account for baseline transfusion status. For platelets, the definition distinguishes participants who were transfusion independent at baseline from those who were transfusion dependent. The hemoglobin criteria likewise distinguish baseline red blood cell transfusion status. The neutrophil criterion requires the absence of recent G-CSF, as specified in the endpoint description. The detailed thresholds are stated in the registry record and should be used when interpreting the endpoint; this page does not substitute a different response definition.

A percentage is useful for summarizing how many participants met a clinically defined response criterion within the analysis population. It is more interpretable alongside its denominator and a confidence interval, which conveys uncertainty around the estimated percentage. Here, the registry specifies a 90% confidence interval, but does not report the observed percentage or interval. Consequently, the endpoint definition is available, while its estimated result and precision are not reported in the registry record.

Participants who discontinued before Week 26 were treated as non-responders. This rule makes early discontinuation part of the endpoint classification rather than leaving those participants out of the primary response definition. The registry record does not state how many participants discontinued before that assessment.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a binary participant-level classification: each participant either met the response definition by Week 26 or did not. The reported measure is a percentage of participants. In general, a percentage summarizes the share of an analysis population with a specified outcome. It is important to read that percentage together with the population definition, the outcome rule, and an interval expressing uncertainty. In this record, the endpoint definition and planned confidence-interval level are stated, but the observed percentage and interval are not reported.

What does a confidence interval convey?

In general, a confidence interval describes uncertainty in an estimated quantity under the assumptions of the statistical method. A narrower interval indicates greater precision, while a wider interval indicates less precision. It does not mean that every value inside the interval is equally likely, and it does not remove limitations caused by study design or measurement. The registry specifies a 90% confidence interval for the primary percentage. Since neither endpoint of that interval nor the response estimate is present, there is no reported interval width to interpret.

The interval’s confidence level is not a measure of whether the treatment is effective. It is a property of the estimation procedure. Interpreting the endpoint would require the actual response estimate and interval, and ideally the number of responders and the analysis denominator. None of those primary result values is reported in the registry record. The record’s definition of the Full Analysis Set does identify who belonged in the primary population: patients assigned study treatment who received one dose.

How does the response definition work?

The response is determined from three hematologic domains: platelet count, hemoglobin, and neutrophil count. Meeting any one of the specified criteria qualifies a participant as a responder. For platelet count and hemoglobin, the exact rule depends on whether the participant was transfusion independent or transfusion dependent at baseline. The neutrophil rule also specifies a condition concerning recent G-CSF use. This structure matters because the same measured value may not mean the same thing under every baseline circumstance; the protocol-defined criteria determine classification.

Combining these domains into one endpoint can provide a summary of whether a participant achieved a response through at least one route. It can also conceal variation in which route produced that classification. This is a general feature of composite or “any criterion” endpoints. The registry record does not report how many participants met each individual component, so the overall endpoint cannot be decomposed from the available record.

What does the discontinuation rule imply?

The endpoint description says that participants who discontinued before Week 26 were treated as non-responders. This is a specified outcome-handling rule. In general, assigning non-response after early discontinuation avoids excluding those participants from the response classification, but it does not reveal why they discontinued or what their response would have been had they remained in follow-up. The registry record does not report the number who discontinued before Week 26 or the reasons for discontinuation.

That rule should not be confused with a statement that every missing measurement was handled in the same way. The record specifically describes early discontinuation before Week 26 as non-response. It does not give a broader account of missing data, missed assessments, or other analysis decisions. Those details are not reported in the registry record.

Why does the design affect interpretation?

In a single-group study, the observed response proportion is an estimate within the treated study population. It is not a randomized contrast between eltrombopag and another treatment. General statistical reasoning distinguishes an outcome observed after treatment from a treatment effect established by comparison: without a suitable comparison group, changes may have more than one possible explanation. The open-label design also means that participants and investigators were not masked to treatment assignment, according to the registry record. Because the primary outcome is investigator assessed, the lack of masking is relevant context, though the record does not establish that it altered any assessment.

The study is identified as Phase II and enrolled 20 participants. The registry record does not report a sample-size justification, target precision, hypothesis-testing plan, or interim analyses. These should not be inferred from the phase or enrollment. A sample-size calculation elsewhere on the page, if shown, is a planning illustration based on the stated assumptions, not the trial’s own design justification and not observed power. The assumptions described for such illustrations are a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect set as a stated fraction of the observed effect. Since this trial itself had one group, those two-arm illustrations do not describe its allocation.

In general, over-dispersion is relevant when observed variability exceeds what a statistical model assumes; ignoring it can make uncertainty appear too small. This is a general concept, not a method reported for this trial. The registry record does not identify a model for over-dispersion, or any model used for the primary endpoint. It also lists no analyses in the available record. Accordingly, the statistical method used to produce a primary estimate is not reported in the registry record.

Finally, secondary outcomes are listed, including changes in blood counts and transfusion measures, as well as time to and duration of response. Different outcome types call for different estimation methods in general. The registry record does not provide analysis results or a detailed account of the methods for these measures. The relevant interpretation is therefore limited to what is explicitly recorded, rather than assuming a particular test, model, or adjustment strategy.

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("Eltrombopag")
n      <- c(20)
events <- c(14)
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  Eltrombopag  20  14  70.0  45.7  88.1  48.1  85.5  47.8  85.4

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists a primary endpoint and multiple secondary outcomes, including response at other assessment times, changes in platelet count, hemoglobin and neutrophil count, time to and duration of response, and platelet transfusion measures. Testing or estimating many outcomes can raise a general multiplicity concern: when many statistical questions are examined, the chance of at least one apparently notable result may increase. Whether and how that concern was addressed depends on the prespecified analysis plan.

The registry record does not report a multiplicity adjustment, a hierarchy for testing outcomes, or a strategy for controlling the overall error rate. It also does not provide a detailed statistical analysis plan. The absence of this information in the record is not evidence that no such planning existed; the approach is not reported in the registry record.

For the primary endpoint, participants who discontinued before Week 26 were classified as non-responders. This is the only specific missing-outcome handling rule stated in the endpoint description. The registry record does not give the number of early discontinuations, the number of missing Week 26 assessments, or the handling of other missing observations. It does not describe sensitivity analyses for missing data.

The Full Analysis Set comprised patients assigned study treatment who received one dose. The record does not report further details about analysis populations for the listed secondary outcomes, nor does it provide a per-protocol or safety population definition beyond the group-specific at-risk counts. No primary analysis result is posted, so the effect of any population or missing-data choices on the numerical result cannot be assessed. These reporting gaps limit evaluation of robustness, but they should not be filled by assumptions about the protocol.

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: pneumonia (15.0%), upper respiratory tract infection (10.0%), intentional product misuse (10.0%), febrile neutropenia (5.0%), gastritis (5.0%).

The registry lists serious and other adverse-event counts for the on-treatment group. For serious adverse events, 8 participants were affected among 20 at risk. For other adverse events, 19 participants were affected among 20 at risk. These are participant counts in the posted record; they are not counts of all event episodes, and the record does not give the number of repeated events per participant in these summary figures.

The listed serious-event terms include pneumonia (15.0%), upper respiratory tract infection (10.0%), intentional product misuse (10.0%), febrile neutropenia (5.0%), gastritis (5.0%), and gastrointestinal haemorrhage (5.0%). The registry presents these as percentages, but does not provide event-level details. The percentages should be read as reported, not treated as evidence that the study treatment caused the events.

The on-treatment group lists zero deaths. A separate post-treatment group lists one death, with zero participants at risk in that group. Because the record gives no at-risk denominator for the post-treatment group, the death entry cannot be converted into a rate or interpreted as a comparison with the on-treatment group. The registry record does not give further context for that death.

These safety figures are descriptive. The study has no randomized control group, and the posted counts alone cannot estimate excess risk attributable to treatment. They also do not capture the full clinical context, event timing, or severity details beyond the categories stated in the record.

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

The study asks a focused question: what proportion of treated participants met a defined hematologic-response criterion at 6 months (Week 26)? The registry record clearly describes the response rules and says that early discontinuations before that time were treated as non-responders. However, it does not report the observed response percentage or its 90% confidence interval. Without those values, readers cannot assess the magnitude or precision of the primary result from this record.

The study design also sets an important boundary on interpretation. This was a non-randomized, open-label, single-group Phase II study. Its results can describe what was observed among participants in the eltrombopag group, but cannot establish how outcomes would compare with a concurrent control group. In general, an uncontrolled before-and-after or single-group observation cannot by itself separate a treatment effect from other influences on outcomes.

The endpoint combines platelet, hemoglobin, and neutrophil response criteria. A participant could qualify by meeting any one of them. Therefore, even a reported combined percentage would not tell readers which blood-count domain drove the result unless component results were also shown. Those component results are not reported.

The safety section gives counts for participants with serious and other adverse events during on-treatment reporting, alongside selected serious-event terms. These figures describe what was recorded; they do not establish causation. The post-treatment death entry has no participants at risk listed for that group, which prevents a denominator-based interpretation from the registry record.

In short, this record supports understanding of the population, intervention, endpoint definition, and selected safety observations. It does not supply the primary efficacy estimate, its uncertainty interval, or the analysis detail needed to assess robustness. The result should not be inferred from the study’s enrollment, endpoint schedule, or safety figures.

14. Why This Trial Matters Statistically

This trial is statistically informative as an example of how a response endpoint depends on its operational definition. The primary outcome is not simply a change in one laboratory measure: it combines criteria across platelet count, hemoglobin, and neutrophil count, with some thresholds conditioned on baseline transfusion status. Careful reading of those rules is essential to understanding what a response percentage would represent.

The study also illustrates the distinction between an outcome description and a comparative treatment effect. A single-group design can document outcomes in participants receiving the study treatment, but does not offer a randomized estimate of how those outcomes differ from a control condition. Open-label investigator assessment is another design feature that belongs in the interpretation, particularly because the primary outcome was assessed by an investigator.

Missing result reporting is itself important for statistical interpretation. The registry specifies a 90% confidence interval but does not report the primary percentage or interval. Without the estimate and its uncertainty, readers cannot judge either the observed response level or how precisely it was estimated. An endpoint definition alone is not a result.

Finally, the handling of early discontinuation is explicitly stated: participants discontinuing before Week 26 were classified as non-responders. That makes the rule visible, while the number of affected participants and broader missing-data procedures remain unreported. This trial therefore highlights why endpoint definitions, analysis populations, missing-data rules, and numerical results all need to be considered together.

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 from this registry record is that the primary response estimate and its 90% confidence interval are not reported. The record defines the endpoint but does not state the observed percentage or responder count. The listed secondary outcomes likewise lack posted analysis results in the information.

The design was non-randomized, open-label, and single group, with no like-for-like control comparison. Accordingly, the record cannot provide a randomized estimate of comparative efficacy. General limitations of uncontrolled studies include difficulty separating treatment-associated outcomes from other factors; this is a design limitation, not a finding about what caused any individual outcome.

The primary endpoint was investigator assessed. The open-label design means treatment assignment was not masked, but the registry record does not report whether outcome assessment procedures included safeguards against assessment bias. The response definition combines several hematologic criteria, and component-specific results are not reported, limiting interpretation of which criteria contributed to any overall response.

The primary analysis population and early-discontinuation rule are described, but the record does not report how many participants discontinued before Week 26 or how other missing assessments were handled. It also does not report a protocol-based sample-size justification, detailed statistical analysis plan, multiplicity strategy, or interim analyses. Those details are not reported in the registry record and should not be inferred.

For safety, the on-treatment summaries provide participant counts, but not a randomized comparator or enough event-level context to establish causation. The post-treatment group lists zero participants at risk alongside one death, so a denominator-based interpretation of that entry is not possible from the record. Together, these gaps limit what can be concluded from the posted information.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Study Assessed the Safety and Efficacy of Eltrombopag in Chinese Refractory or Relapsed Severe Aplastic Anemia (SAA) Subjects. — NCT03988608.
• Registry record last updated 2025-12-30; status last verified 2025-01.
• Results first posted 2025-02-03.
• 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, non-randomized, open-label, multi-center Phase II study of eltrombopag in Chinese subjects with refractory or relapsed severe aplastic anemia. Enrollment was 20, and the study used a single-group model. Treatment started at 25 mg/day, with described increases of 25 mg/day every 2 weeks according to platelet count, up to 150 mg/day.

The primary endpoint was investigator-assessed hematologic response at 6 months (Week 26), defined using platelet, hemoglobin, or neutrophil criteria. Participants who discontinued before Week 26 were treated as non-responders. The primary response percentage and 90% confidence interval are not reported in the registry record. Selected on-treatment safety counts are available, but the single-group design does not support a randomized comparison, and the posted safety figures do not establish causation.