The Efficacy and Safety of Oral Azacitidine Plus Best Supportive Care Versus Placebo and Best Supportive Care in Subjects With Red Blood Cell (RBC) Transfusion-Dependent Anemia and Thrombocytopenia Due to International Prognostic Scoring System (IPSS) Low Risk Myelodysplastic Syndrome (MDS): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for The Efficacy and Safety of Oral Azacitidine Plus Best Supportive Care Versus Placebo and Best Supportive Care in Subjects With Red Blood Cell (RBC) Transfusion-Dependent Anemia and Thrombocytopenia Due to International Prognostic Scoring System (IPSS) Low Risk Myelodysplastic Syndrome (MDS) (NCT01566695), a randomized phase 3 study. The primary outcome is Percentage of Participants Who Achieved Red Blood Cell (RBC) Transfusion Independence for ≥ 56 Days.

2. Clinical Question

3. Trial Design

This was a randomized, parallel-group Phase 3 study sponsored by Celgene. The registry record lists 216 participants enrolled and identifies the study status as completed. The study start date is 2013-04-26, and the primary completion date is 2019-01-25. The condition listed is myelodysplastic syndrome. The official title describes a multicenter, randomized, double-blind comparison; the registry's masking field specifies quadruple masking. The record does not explain the difference in those descriptions or identify which roles were masked.

One group received oral azacitidine with best supportive care, while the other received matching placebo with best supportive care. The active regimen was 300 mg daily on days 1 through 21 of each 28-day treatment cycle. The placebo schedule was also days 1 to 21 of each 28-day cycle. Because both groups received best supportive care, the contrast is the effect of assignment to oral azacitidine rather than placebo within that shared-care context.

The primary outcome used the intent-to-treat population, which the registry defines as all randomized participants regardless of whether they received treatment. The group sizes reported for the primary endpoint were 107 and 109. The registry reports the primary endpoint over the treatment period, with assessment up to the data cut-off date of 25 January 2019; its timeframe description also says approximately 5 months. The record lists 50 outcome measures and 38 analyses. It does not report the original sample-size justification, planned interim analyses, the detailed randomization procedure, or the full statistical analysis plan. Those design details should not be inferred from the observed sample or results.

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: Participants were stratified by: average baseline (BL) Red Blood Cell (RBC) transfusion requirement (≤ 4 units versus \> 4 units of RBC per 28 days), BL platelet transfusion status (dependent or independent), country of enrollment and Eastern Cooperative Oncology Group (ECOG) Performance Status (PS) (0 to 1 versus 2).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The intent-to-treat (ITT) population included all participants who were randomized, regardless of whether they received treatment or not.

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: The intent-to-treat (ITT) population included all participants who were randomized, regardless of whether they received treatment or not.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants Who Achieved Red Blood Cell (RBC) Transfusion Independence for ≥ 56 Days, measured over Each participant was assessed for at least 56 days or more; from the date of randomization of study drug up to the data cut-off date of 25 January 2019, approximately 5 months..

The primary endpoint was the percentage of participants who achieved red blood cell transfusion independence for at least 56 days. A participant qualified by having no red blood cell transfusion during any consecutive rolling 56-day interval within the treatment period. For example, the qualifying interval could run from day 1 to day 56 or from day 2 to day 57. The registry's definition therefore allows the qualifying window to move across the treatment period rather than fixing a single starting day.

This endpoint focuses on a clinically relevant treatment burden: whether a participant experienced a sustained interval without red blood cell transfusion. Requiring a consecutive interval distinguishes a sustained response from a shorter interruption in transfusion need. The registry defines the endpoint as a binary response for each participant—achieved or not achieved—and reports the percentage in each randomized group.

As with any binary outcome, the percentage communicates how common the defined response was in each group, but it does not describe how long every participant was followed or the timing of responses in detail. The endpoint description states that participants were assessed for at least 56 days or more and that the assessment ran from randomization up to the data cut-off date. The record does not report additional details about handling participants whose observation period was too short to assess the endpoint, or how any missing endpoint data were classified.

7. Results Posted in the Registry

• Placebo Plus Best Supportive Care: 11.9% (13/109) versus 30.8% (33/107) on control; risk difference -18.9 percentage points (95% CI -29.6 to -8.3), risk ratio 0.39 (0.22–0.69); 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 RBC Transfusion Independence for ≥ 84 Days

• 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:
  • Placebo Plus Best Supportive Care: 6.4% (7/109) versus 28.0% (30/107) on control; risk difference -21.6 percentage points (95% CI -31.3 to -11.9), risk ratio 0.23 (0.11–0.50); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Placebo Plus Best Supportive Care than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percentage of Participants With a Hematological Improvement Response in Platelets (HI-P) According to 2006 IWG Criteria

• 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:
  • Placebo Plus Best Supportive Care: 7.3% (8/109) versus 24.3% (26/107) on control; risk difference -17.0 percentage points (95% CI -26.4 to -7.5), risk ratio 0.30 (0.14–0.64); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Placebo Plus Best Supportive Care than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percentage of Participants Who Achieved Platelet Transfusion Independence With a Duration of ≥ 8 Weeks (56 Days)

• 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:
  • Placebo Plus Best Supportive Care: 14.3% (5/35) versus 16.7% (5/30) on control; risk difference -2.4 percentage points (95% CI -20.1 to 15.3), risk ratio 0.86 (0.27–2.68); unadjusted two-proportion z-test p = 0.791.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Placebo Plus Best Supportive Care and the control. The interval gives the range of effects that remain compatible with the data.

Across the secondary outcomes with posted results, the broad pattern is that several measures of red blood cell transfusion response and platelet response have higher observed percentages in the oral-azacitidine group, while the platelet transfusion-independence outcome does not show a clear separation in the posted estimates. This broad pattern is directionally compatible with the primary result, but secondary endpoints answer related rather than identical questions. They use different response definitions, durations, or analysis populations, and should not be treated as interchangeable replications of the primary endpoint.

Some secondary outcomes assess longer-duration transfusion independence or specific hematological responses. Others concern transfusion reduction, duration, or time to response. Those distinctions matter: a percentage achieving a threshold does not describe the duration of response among responders, and a duration measure among responders does not have the same interpretation as the proportion responding in all randomized participants. The registry lists 49 secondary outcomes posted among 50 outcome measures, and the listed secondary outcomes include a variety of these definitions.

Unadjusted secondary results carry less weight because examining many outcomes increases the chance that some will appear persuasive by chance alone. Multiplicity refers generally to the family of statistical tests and outcomes considered together. The registry record does not report a multiplicity adjustment or a hierarchy for interpreting the secondary results. Therefore, secondary p-values and intervals should be considered exploratory unless a prespecified testing strategy is documented elsewhere; such a strategy is not reported in the registry record. The overall pattern may provide context for the primary finding, but it does not independently establish each secondary effect or resolve the uncertainty associated with multiple comparisons.

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

• Duration of RBC Transfusion Independence Among Participants Who Achieved RBC Transfusion Independence for at Least 56 Days
• Time to RBC Transfusion Independence for at Least 56 Days Among Participants Who Achieved RBC Transfusion Independence for at Least 56 Days
• Duration of RBC Transfusion Reduction for Participants Who Achieved RBC Transfusion Reduction of at Least 4 Units of RBCs for at Least 8 Weeks
• Percentage of Participants Who Achieved RBC Transfusion Independence for ≥ 84 Days
• Duration of RBC Transfusion Independence Among Participants Who Achieved RBC Transfusion Independence for at Least 84 Days
• Time to RBC Transfusion Independence for at Least 84 Days Among Participants Who Achieved RBC Transfusion Independence for at Least 84 Days
• Percentage of Participants With an Erythroid Hematological Improvement (HI-E) Response According to 2006 IWG Criteria
• Percentage of Participants With a Hematological Improvement Response in Platelets (HI-P) According to 2006 IWG Criteria

9. Statistical Interpretation

10. Statistical Methods Explained

The main analysis concerns a binary outcome: each participant either met the defined transfusion-independence criterion or did not. General statistical explanation: for a binary outcome, a group percentage is the observed proportion with the event, expressed on a percentage scale. The confidence interval around that percentage conveys uncertainty about the underlying response proportion, conditional on the method and assumptions used to calculate it. The registry reports group-specific percentages and 95% confidence intervals for the primary result.

What does the risk difference describe?

General explanation: a risk difference subtracts one group's event proportion from the other group's event proportion. It is an absolute contrast, so it expresses how many percentage points separate the groups rather than expressing a relative multiplier. The primary recalculation defines the contrast as placebo minus oral azacitidine. Its negative estimate means that the observed response proportion was lower in the placebo group under that subtraction order. The registry's own analysis reports the rate difference with the opposite sign, placing the oral-azacitidine group above the placebo group. Sign conventions matter: reversing the subtraction order reverses the sign but does not change the underlying comparison.

Why are the analysis methods not identical?

The registry reports a stratified Mantel-Haenszel chi-squared analysis for the primary endpoint. A stratified analysis combines information across strata while accounting for the grouping used in the analysis. The registry record does not state which stratification factors were used, so their identities should not be inferred. The recalculation is a simple unadjusted comparison based on the posted group summary statistics, rather than a reproduction of the stratified analysis. It can serve as a check on the broad direction and scale of the contrast, but it is not a substitute for the sponsor's prespecified analysis.

The reported p-value evaluates how unusual a result at least as extreme as the observed one would be under the null hypothesis and the test's assumptions. It is not the probability that the null hypothesis is true, nor does it measure the clinical importance of the result. The confidence interval adds information about the range of effect sizes compatible with the data and is therefore useful alongside the p-value. Neither statistic resolves questions about endpoint relevance, bias, or missing observations by itself.

Why does over-dispersion matter for some outcomes?

General explanation: over-dispersion arises when the observed variability in data is greater than a model's assumptions allow. It can make standard errors too small and confidence intervals too narrow if a model does not account for the extra variability. The primary endpoint is binary and the registry reports confidence intervals and a stratified test; the record does not report an over-dispersion model for this endpoint. Over-dispersion is a general consideration for outcomes such as counts or rates, and should not be claimed as a feature of this primary analysis without evidence in the registry record.

The endpoint definition also affects interpretation. A rolling window allows any qualifying consecutive interval during treatment to count, and the analysis classifies participants based on whether they achieved that criterion. This is different from a time-to-event analysis, which would focus on the timing of an event and could account for varying follow-up through a specified method. The primary outcome is reported as a percentage, not as a time-to-event estimate.

The registry labels the primary analysis population as intent-to-treat. General explanation: this approach compares participants according to randomized assignment, rather than regrouping them according to treatment received. It aims to preserve the benefits of random allocation. Whether missing outcome data, treatment discontinuation, or other protocol events were handled in a particular way cannot be determined from the registry record, because those details are not reported there.

11. Confidence Intervals and Precision

The primary group estimates have 95% confidence intervals of 22.1% to 39.6% for oral azacitidine plus best supportive care and 5.8% to 18.0% for placebo plus best supportive care. These intervals indicate uncertainty around each estimated response proportion. Their width reflects that the estimates are based on a finite number of participants, not a census of everyone with the condition. The intervals should not be read as containing 95% of individual participant responses.

For the between-group contrast, the sponsor's reported rate difference is 18.9 percentage points, with a 95% confidence interval from 8.3 to 29.6 percentage points. The interval remains on the positive side of no difference when expressed as oral azacitidine minus placebo. The recalculation uses the opposite subtraction order, so its interval runs from -0.296 to -0.083. These are different sign conventions for the contrast; they do not represent separate underlying treatment comparisons.

Confidence intervals help distinguish a fairly precise estimate from one with substantial uncertainty. They do not guarantee that the true effect lies within the interval, and they depend on the statistical method and assumptions. In general, a wider interval means the data are less precise about the effect size, while a narrower interval indicates greater precision under the same framework. The primary interval supports a positive group contrast in the registry's reported direction, but it still allows a range of magnitudes. It does not specify the response expected for any individual participant, and it cannot address aspects of uncertainty not included in the analysis, such as unreported missing-data procedures.

The registry also reports confidence intervals for secondary outcomes, but their interpretation must account for each endpoint's definition and the number of outcomes examined. In particular, a result with a wide interval can be compatible with materially different effects even if its point estimate appears close to one group's estimate. Precision is not equivalent to importance, and statistical evidence for a group difference is not by itself a complete assessment of the treatment's overall effects.

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 sample-size figures are planning illustrations based on the observed effect, not the original design assumptions for this trial. For binary outcomes, the displayed figures are participants per arm; time-to-event planning displays instead use total events, but the calculations described for these posted endpoints are per arm. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each scenario targets the stated fraction of the observed effect.

Read a row such as planning for 75% of the observed effect as asking how many participants per arm a future study would need to detect an effect smaller than the one observed, under those planning assumptions. The 80% power column means the design would have an 80% chance of detecting the specified effect if that effect were the true one and the assumptions held. The 90% power column uses a higher chance of detection, which generally requires more participants. These are design targets, not guarantees about results.

Smaller effects are harder to distinguish from random variation, so the planning illustrations require more participants per arm when the target is a smaller fraction of the observed effect. This relationship is visible across the scenarios. The calculations are illustrations only: they do not establish that the trial was originally designed using these assumptions, because the original sample-size justification and design assumptions are not reported in the registry record. This is not observed power.

The observed effect used as a planning reference may be an unstable estimate, especially when the result comes from a single study. A future design would need to choose its target effect and assumptions for its own purpose rather than treating these illustrations as a definitive enrollment recommendation. The figures also do not account for any assumptions beyond those expressly stated in the registry-derived planning description.

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 = 0.308; treatment proportion p₁ = 0.119; average p̄ = 0.214.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.5798 = 1.1365.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5642 = 0.4749.
4. n = (1.1365 + 0.4749)² / (-0.1891)² = 72.6, rounded up to 73 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 <- 0.308411214953271                          # control proportion
p1 <- 0.119266055045872                          # 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(1566695)
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 = 73 (80% power), 97 (90% power)
 75% of observed effect: n per arm = 141 (80% power), 188 (90% power)
 50% of observed effect: n per arm = 338 (80% power), 452 (90% power)
power.prop.test: 72.6 
Simulated power at n = 73 per arm: 0.811 (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 is a stratified Mantel-Haenszel chi-squared analysis. It reports a rate difference of 18.9 percentage points, a 95% confidence interval from 8.3 to 29.6, and a p-value of 0.001. The registry identifies this analysis as a superiority analysis. The recalculation is an unadjusted comparison from the posted summary statistics, expressed as placebo minus oral azacitidine; that is why its estimate and confidence interval are negative. Its p-value is <0.001.

The results are consistent in direction: both indicate a higher observed primary-endpoint proportion in the oral-azacitidine group than in the placebo group. The estimates are not identical because the recalculation is unadjusted, whereas the sponsor's analysis is stratified. The registry does not state the stratification factors in the record, so the adjusted analysis cannot be reconstructed from the reported information alone. The record does not flag a discrepancy between the posted result and the recalculation.

The unadjusted recalculation is best treated as a transparent check based on group summaries, not as a replacement for the sponsor's analysis. The sponsor's reported method and result describe the registry's own inferential analysis. Differences in method can affect the precise estimate, interval, or test statistic even when the overall interpretation is aligned. Neither analysis removes the need to consider endpoint definition, missing-data handling, multiplicity, and the limitations of the registry record.

15. Multiplicity, Missing Data and Analysis Populations

The registry record lists a large set of outcome measures, including one primary endpoint and numerous secondary outcomes. When several endpoints are tested, the chance of at least one apparently unusual result can increase, even if no true differences exist for the outcomes being tested. General statistical practice may address this through a prespecified hierarchy, adjustment of significance thresholds, or an explicitly exploratory interpretation. The registry record does not report whether any such multiplicity strategy was planned or applied. The secondary findings therefore require more caution than a single prespecified primary test.

The record identifies the primary analysis population as intent-to-treat and defines it as all randomized participants, regardless of whether they received treatment. This is useful for understanding the intended comparison. However, detailed missing-data rules are not reported in the registry record. In particular, it does not state how participants without enough observation time to establish a 56-day interval were handled, how missing transfusion information was treated, or whether any imputation or sensitivity analyses were performed. No method should be assumed from the endpoint percentages alone.

The registry also does not report the detailed analysis plan, interim analyses, the original sample-size justification, or the factors used in the stratified primary analysis. These omissions limit the ability to evaluate whether all analyses followed prespecified procedures and whether the reported inferential strategy accounts for all design features. They do not negate the reported estimates, but they constrain how completely those estimates can be independently assessed from the registry record.

General explanation: randomized assignment helps make groups comparable at baseline on average, but it does not ensure complete follow-up or remove the consequences of missing outcome data. The effect of missingness depends on why data are missing and how the analysis handles them. Since those details are not reported in the registry record, the primary estimates should be interpreted as the posted analysis rather than as evidence that missingness could not have influenced the results.

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: febrile neutropenia (27.1%), pneumonia (12.1%), pyrexia (7.5%), sepsis (7.5%), anaemia (6.5%).

The registry reports serious adverse events in 83 of 107 participants assigned to oral azacitidine plus best supportive care and in 69 of 109 assigned to placebo plus best supportive care. It also reports other adverse events in 107 of 107 participants in the oral-azacitidine group and 104 of 109 participants in the placebo group. These counts describe the posted safety summaries; they are not a comparison adjusted for follow-up time or other factors.

The most frequently listed serious adverse-event terms include febrile neutropenia, pneumonia, pyrexia, sepsis, anaemia, and septic shock. The listed percentages are event-term summaries and should not be combined or added: participants may have experienced more than one event, and the registry excerpt does not provide the participant-level overlap. The record also lists death counts of 83 and 86 for the respective groups, but does not provide accompanying definitions or context in the available summary. Those counts should not be interpreted as treatment-related deaths.

Serious adverse-event counts do not by themselves establish causation, severity for every participant, or comparative risk over time. The registry record does not report additional details needed for a full safety assessment, such as event timing, duration, attribution, or the handling of exposure differences. These summaries provide a descriptive account of posted events, not medical advice or a complete benefit-harm judgment.

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

The primary result shows a higher observed proportion achieving the defined 56-day red blood cell transfusion-independence outcome among participants assigned to oral azacitidine plus best supportive care than among those assigned to placebo plus best supportive care. The sponsor's stratified analysis reports a positive rate difference, and its confidence interval excludes no difference. The unadjusted recalculation has the same direction once its reverse subtraction order is taken into account. The posted results therefore provide statistical evidence of a group difference for this endpoint in this trial.

The result does not mean that every participant receiving oral azacitidine achieved transfusion independence, nor that no participant receiving placebo did so. It does not predict an individual participant's outcome. The endpoint records whether a participant achieved at least one qualifying rolling interval during the treatment period; it does not establish how long the response continued beyond that interval or describe every aspect of transfusion needs.

The result also does not resolve all questions about safety, durability, generalizability, or the effect of treatment outside this study's population and care context. Both groups received best supportive care, so the comparison is about the addition of oral azacitidine against placebo in that setting. The registry record does not provide all details of missing-data methods, multiplicity handling, or the original design justification. Those gaps limit the conclusions that can be drawn from the posted information.

Statistical significance and clinical importance are related but distinct. A p-value assesses compatibility with a null hypothesis under a statistical model; a confidence interval describes uncertainty about the effect estimate. Neither alone determines whether an observed difference is important to patients or how it should influence care. The posted data characterize a randomized group comparison on a specified endpoint. They should not be interpreted as medical advice or as a complete assessment of treatment effects.

18. Why This Trial Matters Statistically

This trial illustrates how a registry result can be read at several levels: the response proportion in each randomized group, the absolute difference between groups, and the uncertainty around that difference. Reporting the group percentages makes the event definition tangible, while the rate difference summarizes the contrast on an absolute scale. Keeping the sign convention explicit also prevents a negative recalculated estimate from being mistaken for a result in the opposite direction.

The endpoint's rolling-window definition is statistically important. A response is not tied to a single calendar window; it can be established by any qualifying consecutive period during treatment. That flexibility is part of the stated outcome and should guide interpretation of what the percentage represents. The outcome is a binary response measure, not a full description of the timing or duration of each participant's transfusion needs.

The comparison also demonstrates why an unadjusted recalculation and a sponsor-reported stratified analysis should be distinguished. Both can be informative, but they answer the same broad contrast using different analytical approaches. The stratified result is the registry's own inferential analysis, while a summary-statistic recalculation is a transparent check, not a reconstruction of undisclosed stratification details.

Finally, the breadth of secondary outcomes emphasizes the need to separate a primary question from supportive or exploratory questions. A coherent pattern across related endpoints can add context, but multiplicity and incomplete reporting affect the weight that should be placed on each result. Careful statistical reading preserves what the reported estimates show while avoiding claims that go beyond the endpoint, analysis, and information recorded.

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.

Several limitations are specific to interpreting this registry record. The primary outcome is defined as any rolling 56-day period without a red blood cell transfusion during treatment. That definition does not, by itself, describe whether the response persisted after the qualifying interval, or how response timing varied among participants. The endpoint is a participant-level threshold, so it compresses potentially different clinical trajectories into a yes-or-no classification.

The record does not report the detailed missing-data rules, the number of participants with incomplete endpoint assessment, or sensitivity analyses for missing observations. This matters for a threshold that requires observation across a consecutive interval. The record also does not identify the stratification factors in the sponsor's primary analysis, making it impossible to reproduce that adjusted analysis from the registry summary alone.

The trial has many posted outcome measures, but the registry record does not state a multiplicity plan or identify a hierarchy for the secondary analyses. Individual secondary findings should therefore not be treated as confirmatory solely on the basis of their reported estimates or p-values. The original sample-size justification and the trial's own design assumptions are also not reported, so planning illustrations based on the observed result should not be mistaken for the design rationale.

The safety summaries are descriptive and do not supply full event-level context, such as timing, duration, or attribution. The record's masking descriptions differ between the official title and the masking field, and the record does not explain that difference. Finally, the enrolled population and shared best supportive care context define the scope of the comparison; the registry information alone does not establish how directly the findings apply beyond that setting.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: The Efficacy and Safety of Oral Azacitidine Plus Best Supportive Care Versus Placebo and Best Supportive Care in Subjects With Red Blood Cell (RBC) Transfusion-Dependent Anemia and Thrombocytopenia Due to International Prognostic Scoring System (IPSS) Low Risk Myelodysplastic Syndrome (MDS) — NCT01566695.
• Registry record last updated 2025-01-07; status last verified 2024-12.
• Results first posted 2020-02-26.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

23. Record Summary

NCT01566695 is a completed randomized Phase 3 parallel-group study comparing oral azacitidine plus best supportive care with placebo plus best supportive care in participants with red blood cell transfusion-dependent anemia and thrombocytopenia due to lower-risk myelodysplastic syndromes. The registry lists 216 enrolled participants and identifies the primary population as all randomized participants.

The primary endpoint was red blood cell transfusion independence for at least 56 days, defined as no transfusion during any consecutive rolling 56-day interval within the treatment period. The posted proportions were 30.8% with oral azacitidine and 11.9% with placebo. The registry's stratified analysis reports a rate difference of 18.9 percentage points, with a 95% confidence interval from 8.3 to 29.6 and a p-value of 0.001.

Interpretation should remain specific to this endpoint and the reported randomized comparison. The registry record does not report the original sample-size justification, detailed missing-data procedures, or a multiplicity strategy for the secondary outcomes.