A Study of Luspatercept (ACE-536) to Treat Anemia Due to Very Low, Low, or Intermediate Risk Myelodysplastic Syndromes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Luspatercept (ACE-536) to Treat Anemia Due to Very Low, Low, or Intermediate Risk Myelodysplastic Syndromes (NCT02631070), a randomized phase 3 study. The primary outcome is Percentage Of Participants Who Achieved Red Blood Cell Transfusion Independence (RBC-TI) ≥ 8 Weeks From Week 1 to Week 24.

2. Clinical Question

This question focuses on a defined transfusion-free interval within a specific treatment period. It does not, by itself, describe the length of any response beyond the qualifying interval, changes in other measures of anemia, or outcomes after the study-treatment period. Those are distinct questions, reflected in the registry’s secondary outcomes. The primary endpoint is a participant-level binary outcome: each treated participant either met the response definition or did not.

3. Trial Design

The registry describes a phase 3, randomized, placebo-controlled, multicenter study with a parallel-group model. The study compared luspatercept with placebo in participants with anemia due to IPSS-R very low, low, or intermediate risk myelodysplastic syndromes with ring sideroblasts who required red blood cell transfusions. The listed enrollment is 229. The primary result is reported for all treated participants, with 153 participants in the luspatercept arm and 76 in the placebo arm.

Luspatercept was administered as a subcutaneous injection, starting at 1.0 mg/kg every 3 weeks. Placebo was also administered by subcutaneous injection every 3 weeks. The registry lists quadruple masking; the brief study summary describes the design as double-blind. The registry does not explain the masking terminology further. Allocation concealment procedures and the randomization ratio are not reported in the registry record.

The record lists the study start as 2016-02-09, primary completion as 2019-06-18, and status as completed. The primary outcome timeframe runs from Week 1 through Week 24 of study treatment. Its definition uses any consecutive 56-day period without a red blood cell transfusion. Participants who failed to meet the required period by the Week 24 cutoff were classified as non-responders.

The registry lists 28 outcome measures and 23 analyses. It also lists 27 secondary outcome measures, covering additional transfusion-independence definitions, changes in transfused units, hematologic response, hemoglobin change, and response duration. The trial’s original sample-size justification, allocation ratio rationale, and detailed statistical plan are 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: 229 participants were randomized and treated.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All treated participants

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 Who Achieved Red Blood Cell Transfusion Independence (RBC-TI) ≥ 8 Weeks From Week 1 to Week 24, measured over From Week 1 through Week 24 of study treatment.

The primary endpoint is the percentage of treated participants who achieved red blood cell transfusion independence for at least 8 weeks during the first 24 weeks of study treatment. Transfusion independence is defined as no red blood cell transfusion during any consecutive 56-day period. The definition allows a qualifying period to begin on different days within the assessment window; the record gives examples beginning on Days 1, 2, or 3.

A participant had to complete at least 56 days without transfusion before or on the Week 24 cutoff to count as a responder. Participants who did not meet that condition by the cutoff were counted as non-responders. This makes the endpoint a binary measure, summarized as a percentage of participants, rather than a measure of the total number of transfusions or the duration of each participant’s response.

This endpoint is used to capture whether a participant experienced a clinically interpretable interval without transfusion within a specified period. It is distinct from longer transfusion-free intervals, from outcomes assessed through Week 48, and from the duration of transfusion independence. Those measures are listed separately as secondary outcomes. The primary outcome does not establish how long a response lasted after its qualifying period, nor does it describe other aspects of anemia or overall health.

The endpoint description identifies all treated participants as the analysis population. Details about the handling of missing assessments, treatment discontinuation, or other intercurrent events are not reported in the registry record.

7. Results Posted in the Registry

• Luspatercept: 37.9% (58/153) versus 13.2% (10/76) on control; risk difference 24.8 percentage points (95% CI 13.9 to 35.6), risk ratio 2.88 (1.56–5.31); unadjusted two-proportion z-test p = <0.001.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Percentage of Participants Who Achieved Red Blood Cell Transfusion Independence (RBC-TI) ≥ 12 Weeks From Week 1 to Week 24

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Luspatercept: 28.1% (43/153) versus 7.9% (6/76) on control; risk difference 20.2 percentage points (95% CI 10.9 to 29.6), risk ratio 3.56 (1.59–7.99); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Luspatercept than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percentage of Participants Who Achieved Red Blood Cell Transfusion Independence (RBC-TI) ≥ 12 Weeks From Week 1 to Week 48

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Luspatercept: 33.3% (51/153) versus 11.8% (9/76) on control; risk difference 21.5 percentage points (95% CI 11.1 to 31.9), risk ratio 2.81 (1.46–5.41); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Luspatercept than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percentage of Participants Who Achieved Red Blood Cell Transfusion Independence (RBC-TI) ≥ 8 Weeks From Week 1 Through Week 48

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Luspatercept: 45.1% (69/153) versus 15.8% (12/76) on control; risk difference 29.3 percentage points (95% CI 17.9 to 40.7), risk ratio 2.86 (1.65–4.94); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Luspatercept than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The registry lists 27 secondary outcome measures. These include other durations and assessment windows for transfusion independence, changes in red blood cell units transfused, modified hematologic erythroid response, hemoglobin increase, and duration of transfusion independence. Several measures ask whether transfusion independence occurred under a different minimum-duration rule or over a longer observation window than the primary endpoint.

Across the three secondary transfusion-independence comparisons with posted estimates, the direction of the arm-level results is consistent with the primary comparison: a greater proportion in the luspatercept group met the specified response definition than in the placebo group. The reported Cochran-Mantel-Haenszel analyses for these measures are also labeled as superiority analyses. This broad consistency supports the view that the primary result is not isolated to a single stated transfusion-independence definition, while each secondary endpoint remains a distinct outcome with its own time window and response threshold.

Secondary analyses can help assess whether a primary finding is accompanied by a coherent pattern across related outcomes. They do not replace the primary endpoint, and they should not be treated as multiple independent confirmations. The registry record does not report whether these analyses were prespecified in a hierarchy or whether their p-values were adjusted for multiplicity. Without that information, the nominal secondary p-values carry less inferential weight than a single prespecified primary analysis because testing many outcomes increases the opportunity for chance findings.

Other secondary measures answer different questions and cannot be inferred from the transfusion-independence pattern alone. The registry’s outcome list includes changes in transfused units, hematologic response, hemoglobin increase, and response duration, but the results described in this statistical summary should not be generalized to those outcomes without their own estimates and analysis details.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Percentage of Participants Who Achieved Red Blood Cell Transfusion Independence (RBC-TI) ≥ 12 Weeks From Week 1 to Week 24
• Percentage of Participants Who Achieved Red Blood Cell Transfusion Independence (RBC-TI) ≥ 12 Weeks From Week 1 to Week 48
• Percentage of Participants Who Achieved Red Blood Cell Transfusion Independence (RBC-TI) ≥ 8 Weeks From Week 1 Through Week 48
• Change From Baseline in RBC Units Transfused Over Fixed 16-Week Period
• Percentage of Participants Who Achieved a Modified Hematologic Erythroid Response (mHI-E) Over Any Consecutive 56-Day Period
• Percentage of Participants Who Achieved a Mean Hemoglobin (Hgb) Increase of at Least 1.0 g/dL Over Any Consecutive 56-Day Period in Absence of Red Blood Cells (RBC) Transfusions
• Duration of Red Blood Cell Transfusion Independence (RBC-TI) - Week 1 Through Week 24
• Duration of Red Blood Cell Transfusion Independence (RBC-TI) - Week 1 Through Week 48

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is binary: each treated participant is classified as having achieved the defined transfusion-independent response or not. A binary outcome can be summarized by the number or percentage of participants meeting the criterion in each group. The registry reports percentages and 95% confidence intervals for the two arms, as well as comparative analyses.

What does a risk difference express?

General statistical explanation: a risk difference is the difference between the response probabilities in two groups. It is commonly expressed in percentage points when the probabilities are presented as percentages. It describes absolute separation between groups for the outcome as defined. It is not the same as a relative risk or an odds ratio, and it does not describe how long a response lasted.

The registry’s reported common risk difference is 24.56 percentage points, with a 95% confidence interval from 14.48 to 34.64. The analysis method is Cochran-Mantel-Haenszel. General statistical explanation: this method can combine comparisons across strata to form a common estimate, accounting for the stratification used in the analysis. The registry record does not identify the strata or provide further details about how this method was implemented. The reported common estimate therefore should be distinguished from a simple unadjusted comparison of the posted arm-level percentages.

How does an odds ratio differ?

General statistical explanation: odds compare the probability of an event with the probability of no event. An odds ratio compares those odds between groups. It is not a percentage-point difference, and it should not be interpreted as a direct multiplier of the response percentage. The reported primary odds ratio is 5.065, with a 95% confidence interval from 2.278 to 11.259. This is a separate description of the relative contrast, not a replacement for the response percentages or the risk difference.

What does the p-value contribute?

The reported primary p-value is p < 0.001. General statistical explanation: a p-value evaluates how compatible the observed data are with a specified null hypothesis, assuming the analysis model and its assumptions. It does not give the probability that the null hypothesis is true, measure the importance of the result, or indicate the probability that a participant will respond. The registry labels the reported analyses as superiority analyses; the p-value alone does not establish the scope or practical meaning of that conclusion.

What is recalculated, and what is the sponsor analysis?

Recalculated comparisons are simple, unadjusted re-calculations from the posted summary statistics. They provide a way to examine the arm-level results without reproducing the sponsor’s named analysis. The sponsor’s reported primary comparison instead uses Cochran-Mantel-Haenszel estimates for a common risk difference and an odds ratio. The two approaches are not necessarily numerically identical because they target the comparison using different analysis procedures. The registry record does not flag a discrepancy between the recalculation and the reported analysis.

General statistical explanation: an unadjusted estimate uses the available group summaries without incorporating stratification or covariate adjustment. An adjusted or stratified estimate may account for specified design features, but its interpretation depends on what was included and how. Since the registry does not report the relevant strata or the complete statistical plan, the degree and purpose of any such adjustment cannot be determined from the record.

The analyses concern a fixed binary response definition and a specified treatment window. No time-to-event analysis is described for the primary outcome. The registry does list duration of transfusion independence among secondary outcomes, which is a different statistical question. Analysis populations other than the primary outcome’s all-treated population, missing-data procedures, and model diagnostics are not reported in the registry record.

11. Confidence Intervals and Precision

The primary arm-level estimates are 37.91% for luspatercept and 13.16% for placebo. Their respective 95% confidence intervals are 30.2 to 46.1 and 6.49 to 22.87. These intervals describe uncertainty around each group’s estimated response percentage. They do not directly give the confidence interval for the between-group comparison.

For the reported common risk difference, the 95% confidence interval is 14.48 to 34.64 percentage points around the estimate of 24.56. The interval lies on the positive side of zero, and the reported p-value is p < 0.001. The common odds ratio estimate is 5.065, with a 95% confidence interval from 2.278 to 11.259. The odds ratio interval is on the side above one. Each interval applies to its own statistical measure and should not be interchanged with the arm-specific intervals.

General statistical explanation: a confidence interval gives a range of values compatible with the data under the method’s assumptions and reflects sampling uncertainty. A narrower interval indicates greater precision for the estimate, while a wider interval indicates more uncertainty. The interval is not a guarantee that the true value lies within its endpoints, and the stated confidence level does not mean that a particular participant’s response is known with that probability.

Precision depends in part on the amount of information in the data, including the number of participants and how frequently the outcome occurs. The registry reports 153 treated participants in the luspatercept group and 76 in the placebo group for this outcome. The placebo response estimate is based on the smaller group. The registry does not provide a precision target or original sample-size justification, so the intended precision cannot be compared with a prespecified target.

Confidence intervals also do not address all uncertainty. They do not resolve possible limitations from missing outcome data, analysis choices, outcome definition, or generalizability. Those issues require information about the conduct and analysis of the trial beyond the intervals themselves.

12. 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 planning values are participants per arm for this binary outcome. They are not total event counts. The illustrations use a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. Each planning scenario asks how many participants per arm would be needed to detect a stated fraction of the observed effect. The values are based on the observed effect, not on the trial’s original design assumptions.

For planning for 100% of the observed effect, the illustration gives 48 participants per arm for 80% power and 64 participants per arm for 90% power. For planning for 75% of the observed effect, it gives 79 participants per arm for 80% power and 105 participants per arm for 90% power. For planning for 50% of the observed effect, it gives 159 participants per arm for 80% power and 213 participants per arm for 90% power.

General statistical explanation: a power target describes the probability of detecting a specified effect under the assumed model if that effect is present. The 80% power column represents a lower probability target than the 90% power column; achieving the higher target generally requires a larger planned sample under otherwise identical assumptions. Smaller effects are harder to distinguish from random variation, so planning for a smaller fraction of the observed effect requires more participants.

These figures are planning illustrations, not a reconstruction of the trial’s sample-size calculation. The original design assumptions and sample-size justification are not reported in the registry record. This is not observed power: it does not measure the probability that the completed trial detected its actual effect, and it should not be used to interpret the result as having passed or failed a post hoc power threshold.

13. 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 = 0.132; treatment proportion p₁ = 0.379; average p̄ = 0.255.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6167 = 1.2086.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5913 = 0.4977.
4. n = (1.2086 + 0.4977)² / (0.2475)² = 47.5, rounded up to 48 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 <- 0.131578947368421                          # control proportion
p1 <- 0.379084967320261                          # 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(2631070)
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 = 48 (80% power), 64 (90% power)
 75% of observed effect: n per arm = 79 (80% power), 105 (90% power)
 50% of observed effect: n per arm = 159 (80% power), 213 (90% power)
power.prop.test: 47.5 
Simulated power at n = 48 per arm: 0.812 (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.

14. Analyses Reported by the Sponsor

The sponsor’s primary analysis reports a Cochran-Mantel-Haenszel common risk difference of 24.56 percentage points, with a 95% confidence interval from 14.48 to 34.64 and p < 0.001. Its reported odds ratio is 5.065, with a 95% confidence interval from 2.278 to 11.259 and p < 0.001. These are the registry’s analyses of the primary comparison.

The recalculated comparison is an unadjusted calculation from the posted arm-level summary statistics. It is a useful cross-check, but it is not a reproduction of the Cochran-Mantel-Haenszel analysis. The sponsor method is reported as a common-effect analysis, while the simple recalculation does not incorporate the same stratified method. The registry does not report the analysis strata or full details needed to recreate the sponsor’s procedure.

The registry does not flag a discrepancy between the recalculated and reported results. The sponsor’s reported analysis therefore remains the principal reported comparison, while the unadjusted calculation serves as a complementary descriptive check. Neither method changes the endpoint definition, the population, or the limits of inference associated with this trial.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome and numerous secondary outcome measures, with 28 outcome measures and 23 analyses overall. Multiple outcomes create an interpretive issue: when many hypotheses are tested, the chance of at least one apparently notable result arising by chance can increase. General statistical explanation: multiplicity procedures, such as a prespecified testing hierarchy or adjustment of significance thresholds, are used in some trials to manage this issue. The registry record does not report whether a hierarchy, family-wise error control, false-discovery procedure, or other multiplicity adjustment was used.

The primary analysis is identified as a Cochran-Mantel-Haenszel analysis of the common risk difference and odds ratio. The registry does not report the complete statistical analysis plan, the stratification variables, or the timing and rules for any interim analyses. Interim analyses are not reported in the registry record. These omissions limit the ability to evaluate whether the posted analysis fully matches a prespecified plan or how all planned comparisons were handled.

The primary population is described as all treated participants. The record does not specify in this outcome description how missing assessments, treatment discontinuation, protocol deviations, or other intercurrent events were handled. The rule that participants who failed to achieve the required transfusion-independent interval by the cutoff were counted as non-responders is reported, but that rule does not explain the handling of participants with missing outcome data.

General statistical explanation: missing-data assumptions matter because the outcome status of participants without complete assessment may not be known. Different approaches can produce different estimates, especially if missingness is related to response or treatment. The registry record does not report sensitivity analyses for missing outcomes or the assumptions used to classify incomplete observations.

These gaps do not invalidate the reported estimates, but they constrain independent appraisal of the full analysis strategy. In particular, the primary result can be interpreted as reported, while confidence in the broader set of secondary claims requires caution because their multiplicity handling and missing-data methods are not described.

16. 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 (5.2%), fall (4.6%), hip fracture (3.9%), femur fracture (3.9%), pyrexia (2.6%).

The registry reports serious adverse events affecting 66 of 153 participants in the luspatercept group and 23 of 76 in the placebo group. Other adverse events affected 145 of 153 and 63 of 76, respectively. Deaths are reported for 45 participants in the luspatercept group and 24 in the placebo group. These are counts of participants, not estimates of treatment-attributable risk, and the registry excerpt does not provide a causal assessment for the events or deaths.

The serious adverse event terms listed most prominently include pneumonia at 5.2%, fall at 4.6%, hip fracture at 3.9%, femur fracture at 3.9%, pyrexia at 2.6%, and sepsis at 2.6%. These term-specific percentages should be read as registry-reported event information; the record excerpt does not give additional detail on timing, recurrence, severity grading, relatedness, or how overlapping events were counted.

Comparisons of safety counts need context, including exposure time, event definitions, follow-up, and participant characteristics. Those details are not reported in the registry record. The figures alone cannot establish that an event was caused by study treatment or support an individualized safety judgment.

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

The primary result describes the proportion of treated participants who achieved at least 8 consecutive weeks without a red blood cell transfusion during the first 24 weeks of study treatment. That proportion was higher in the luspatercept arm than in the placebo arm in the posted results. The reported common risk difference and its confidence interval quantify the comparative result under the registry’s named analysis.

The outcome is a defined trial endpoint, not a complete account of a participant’s health or treatment experience. Meeting the endpoint means that the participant had a qualifying transfusion-free interval. It does not establish that transfusion independence continued after that interval, that anemia resolved, or that other clinical outcomes improved. The endpoint does not measure quality of life or outcomes after the specified study window.

The comparison applies to the enrolled population described in the registry: people with anemia associated with selected IPSS-R risk groups of myelodysplastic syndromes, ring sideroblasts, and a need for red blood cell transfusions. The findings should not automatically be extended to other populations, risk categories, or treatment comparisons. The study compared luspatercept with placebo, not with another active therapy.

A statistically notable result is not equivalent to a guarantee of benefit for every participant, nor does it establish the balance of all possible benefits and harms. The reported safety counts are a separate part of the trial record and require their own contextual interpretation. Individual outcomes can differ from group averages, and the registry results do not provide a personalized prediction.

The registry record also leaves important questions unanswered about missing data, multiplicity, the full statistical plan, and the original sample-size justification. Those limitations affect how completely the analysis can be independently evaluated. They do not change what the posted endpoint and estimates state, but they define the boundaries of what can be concluded from this record alone.

18. Why This Trial Matters Statistically

This trial is statistically informative because it defines a patient-level binary outcome around a concrete transfusion-free interval and compares that outcome between randomized groups. The endpoint is not simply a vague improvement measure: it specifies a minimum duration without transfusion and a fixed assessment window. Clear operational definitions make it possible to understand what counted as a response and what did not.

The result is presented in several complementary forms. Arm-level percentages show the response frequency in each group; the common risk difference expresses absolute separation; and the odds ratio expresses a relative contrast in odds. Confidence intervals convey uncertainty around those estimates. Reading these measures together is more informative than relying on a p-value alone, because statistical evidence, magnitude, and precision answer different questions.

The registry also makes visible the distinction between a sponsor-reported stratified analysis and a simple unadjusted recalculation. That distinction matters in clinical trial interpretation: analysis methods can target related but not identical estimands, and a transparent account of the method helps readers avoid treating all estimates as interchangeable. The record names the Cochran-Mantel-Haenszel method but leaves the strata and detailed plan unreported, illustrating why registry summaries can be useful while still incomplete.

Finally, the range of secondary outcomes demonstrates how a single trial can address related but distinct questions, including longer transfusion-free windows, changes in transfusion burden, and response duration. Those results need to be judged in light of multiplicity and reporting detail. The trial is therefore a useful example of how endpoint definition, effect measure, confidence interval, analysis method, missing-data information, and safety context jointly shape statistical interpretation.

19. 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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The primary outcome is tied to a specific definition: at least 56 days without red blood cell transfusion during the first 24 weeks. A participant who did not complete the qualifying interval by the cutoff was counted as a non-responder. The registry does not explain how participants with incomplete follow-up or missing assessments were classified, which limits assessment of possible bias from missing outcome information.

The record reports a common risk difference and odds ratio using the Cochran-Mantel-Haenszel method but does not identify the strata used or provide the full statistical analysis plan. It also does not report the original sample-size justification, whether interim analyses were planned or conducted, or how multiplicity across the many outcome measures was handled. These gaps make it difficult to verify the prespecified analysis strategy from the registry entry alone.

The trial population is restricted to participants with the specified myelodysplastic syndrome risk groups, ring sideroblasts, anemia, and transfusion requirements. The results may not be directly applicable to people outside those criteria. The comparator was placebo, so the record does not establish how the results compare with other active treatment options.

The safety record gives counts of serious adverse events, other adverse events, and deaths, along with several serious event terms. It does not give all the context needed to interpret comparative safety, such as exposure-adjusted rates, event timing, attribution, or detailed follow-up. The counts therefore should not be treated as causal comparisons on their own.

The registry lists both quadruple masking and a brief summary describing the study as double-blind, without explaining the terminology. The record also reports different group sizes for the primary outcome but does not state the allocation ratio rationale. These reporting limitations are reasons for careful interpretation rather than grounds for inferring undocumented design or analysis choices.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study of Luspatercept (ACE-536) to Treat Anemia Due to Very Low, Low, or Intermediate Risk Myelodysplastic Syndromes — NCT02631070.
• Registry record last updated 2021-12-17; status last verified 2021-11.
• Results first posted 2020-05-22.
• 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.

23. Record Summary

The registry describes a completed phase 3 randomized, placebo-controlled, parallel-group study of luspatercept for anemia associated with selected risk groups of myelodysplastic syndromes with ring sideroblasts in participants requiring red blood cell transfusions. Luspatercept was administered by subcutaneous injection, starting at 1.0 mg/kg every 3 weeks; placebo was administered by the same route and schedule.

The primary endpoint was the percentage of treated participants achieving at least 8 weeks without red blood cell transfusion from Week 1 through Week 24. The posted percentages were 37.91% for luspatercept and 13.16% for placebo. The reported common risk difference was 24.56 percentage points, with a 95% confidence interval from 14.48 to 34.64 and p < 0.001.

The registry also lists multiple secondary outcomes and reports safety counts. Important details, including the original sample-size justification, full analysis plan, missing-data procedures, and multiplicity handling, are not reported in the registry record. Those limits should inform interpretation of the posted estimates.