Decitabine and Clofarabine in Higher Risk Myelodysplastic Syndromes (MDS): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Decitabine and Clofarabine in Higher Risk Myelodysplastic Syndromes (MDS) (NCT00903760), a randomized phase 2 study. The primary outcome is Event Free Survival (EFS) at 1 Year.

2. Clinical Question

3. Trial Design

This was a randomized, parallel-group phase 2 study. The registry identifies the study as unmasked, meaning masking was not used. Enrollment was 42, and the study status is recorded as completed. The listed sponsor was M.D. Anderson Cancer Center. The recorded study start was 2010-01 and primary completion was 2015-06.

The two strategies differed in their use of clofarabine. In the Decitabine + Clofarabine arm, participants received decitabine intravenously at 20 mg/m^2 for 5 days during the first 3 cycles, then clofarabine intravenously at 10 mg/m^2 for 5 days during the next 3 cycles. That alternating pattern could continue for up to 24 cycles. In the Decitabine arm, decitabine was given intravenously at 20 mg/m^2 daily for 5 days, for a total of 24 courses.

The posted primary outcome is event-free survival (EFS) at 1 year, expressed as a percentage of participants. The outcome description defines an event as either death or transformation to AML, with marrow and/or blood blasts >= 20%. The registry also lists Participant Response as a secondary outcome. It records two outcome measures and one posted secondary outcome.

The registry record does not report the randomization procedure, allocation concealment, the intended sample-size justification, or the detailed timing and conduct of outcome assessment beyond the stated 12-month/1-year timeframe. Those details should not be inferred from the randomized label or from the observed group sizes. The primary outcome population description specifically states that one of 18 participants in the Decitabine arm was not treated due to early death and was excluded from that analysis. The exact procedures used to address missing outcomes 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: Of the 42 participant enrolled, five were not randomized or treated and are therefore excluded from the trial details. One participant included in group assignment was randomized but not treated as a result of early death.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: One of 18 participants in the Decitabine arm was not treated due to early death thus is excluded from analysis.

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 Event Free Survival (EFS) at 1 Year, measured over Assessed at 12 months/1 year.

The primary endpoint is the percentage of participants with event-free survival at 1 year. An event was defined as death or transformation to acute myeloid leukemia (AML), with marrow and/or blood blasts >= 20%. The assessment timeframe is recorded as 12 months/1 year. The reported group values are 26.0% for Decitabine + Clofarabine and 65.0% for Decitabine.

As a general statistical concept, an event-free survival endpoint combines more than one type of unfavorable event into a single outcome. It can therefore summarize whether participants remained alive without the specified AML transformation by the assessment time. The definition matters: changing which events count, or how events are confirmed, would change the endpoint being measured. Here, the registry’s stated definition includes death and the specified AML transformation threshold.

A percentage at a fixed time point is an easily interpreted summary of the proportion meeting the endpoint definition at that time. It is not the same as a full time-to-event analysis, which uses the timing of events and may account for participants with incomplete follow-up. The registry record presents the primary result as a percentage at 1 year, and no time-to-event analysis is posted. Details about censoring, follow-up completeness, and handling of participants without an observed 1-year status are not reported in the registry record.

The outcome description also identifies a specific exclusion in the Decitabine arm: one of 18 participants was not treated due to early death and was excluded from analysis. This is relevant when interpreting the denominator and the estimand, meaning the precise treatment contrast the analysis is intended to represent. The registry does not report enough detail to determine whether other exclusions or missing outcomes occurred.

7. Results Posted in the Registry

• Decitabine: 64.7% (11/17) versus 26.3% (5/19) on control; risk difference 38.4 percentage points (95% CI 8.3 to 68.5), risk ratio 2.46 (1.07–5.64); unadjusted two-proportion z-test p = 0.021.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is binary at the stated assessment time: each participant either met the event-free survival definition at 1 year or did not. The reported percentages are therefore proportions, not average survival times. The outcome definition counts death or transformation to AML, with marrow and/or blood blasts >= 20%, as an event. The registry gives 26.0% for Decitabine + Clofarabine and 65.0% for Decitabine.

What does the risk difference express?

The recalculated risk difference is 0.384 for Decitabine relative to Decitabine + Clofarabine. A risk difference expresses an absolute contrast between two proportions. Its scale is direct: it describes how far apart the estimated event-free proportions are, rather than comparing them as a ratio. The 95% confidence interval is 0.083 to 0.685. Because this is an unadjusted recalculation from the posted summary counts, it should be distinguished from a sponsor-reported adjusted analysis. No registry analysis is posted, and the registry record does not report an adjusted model for this endpoint.

How do relative risk and odds ratio differ?

The recalculated relative risk is 2.459, with a 95% confidence interval from 1.072 to 5.64. It compares the two estimated probabilities. The recalculated odds ratio is 5.133, with a 95% confidence interval from 1.234 to 21.36. It compares the odds rather than the probabilities. Odds and probabilities are related but are not interchangeable, so the odds ratio should not be read as a direct multiplication of the percentage. Both measures are unadjusted summaries from the posted counts.

What does the p-value tell the reader?

The unadjusted p-value is 0.021. In general, a p-value measures how unusual data at least as discrepant as those observed would be under a specified null hypothesis and statistical model. It is not the probability that the null hypothesis is true, nor does it measure the size or practical importance of an effect. The registry record does not report the original statistical test or a broader analysis plan. The recalculated p-value therefore describes an unadjusted comparison, rather than a documented result from a fully specified registry analysis.

Why does the endpoint format matter?

The registry reports a percentage at 1 year rather than a survival curve or a hazard ratio. A fixed-time binary analysis summarizes endpoint status at one time, whereas time-to-event methods use event timing and can incorporate incomplete follow-up under specified assumptions. Those methods are not equivalent. The registry record does not report a time-to-event analysis, censoring rules, or how any participants without 1-year status were handled. The results should therefore be read in the form actually posted: group percentages and an unadjusted contrast calculated from the posted counts.

What can randomization support?

Random allocation can, in general, help balance measured and unmeasured participant characteristics between treatment groups on average. It does not guarantee balance in a small realized study, and it cannot remedy missing data or departures from the assigned treatment strategy. The registry identifies this study as randomized, but details about sequence generation, allocation concealment, stratification, and adherence are not reported in the registry record. Those methods cannot be assumed.

The primary outcome population description says one of 18 participants in the Decitabine arm was not treated due to early death and was excluded from analysis. That exclusion is part of the posted endpoint information and affects interpretation of the analysis population. The registry does not describe whether the analysis was intention-to-treat, per-protocol, or another defined population. It also does not give a full participant flow or reasons for any other exclusions. The analysis population should not be reconstructed beyond the stated counts and exclusion.

As a general statistical concept, over-dispersion is extra variability beyond what a chosen probability model expects. It can make uncertainty estimates too small if the analysis assumes less variability than is present. The registry record does not report a model involving over-dispersion, and the binary endpoint comparison described by the posted counts does not establish that over-dispersion was assessed. More broadly, model assumptions and diagnostics are not reported in the registry record.

All recalculated effects are unadjusted. No adjustment variables, subgroup analyses, sensitivity analyses, or multiplicity procedure are reported in the registry record. Because the registry also lists a secondary outcome, the primary result should not be treated as if it were the only outcome considered in the research process. The number and nature of analyses actually conducted cannot be inferred from the posted primary comparison. The registry lists no analyses, so details of inferential methods beyond the posted result and the simple recalculations are not available.

10. Confidence Intervals and Precision

The confidence intervals around the recalculated comparison show that the data do not pin down one exact value for the difference between groups. The risk difference for Decitabine relative to Decitabine + Clofarabine is 0.384, with a 95% confidence interval from 0.083 to 0.685. The interval includes a range of possible contrasts, from a more modest difference to a substantially larger one. It is therefore important to distinguish the point estimate from the uncertainty around it.

The recalculated relative risk is 2.459, with a 95% confidence interval from 1.072 to 5.64. The recalculated odds ratio is 5.133, with a 95% confidence interval from 1.234 to 21.36. These intervals are also broad, particularly for the odds ratio. Their width reflects limited precision in these estimates; the small endpoint groups of 19 and 17 participants are relevant context. A small number of outcomes can have a notable influence when group sizes are limited.

As a general statistical explanation, a confidence interval is a range produced by a method that, under its assumptions, would capture the target value at its stated coverage rate over repeated studies. It is not a range containing a fixed probability that the true value lies inside this particular interval. Nor does it account automatically for every uncertainty, such as missing outcomes, bias, or differences between this study and another population.

The registry record does not report the confidence-interval method, adjustment strategy, or a time-to-event analysis. The recalculated intervals are unadjusted, based on the posted summary counts. They should be read as statistical uncertainty conditional on that calculation, not as a complete accounting of design or data limitations. The stated exclusion of one Decitabine participant due to early death also matters to the interpretation of the population being analyzed.

11. 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 figures are illustrations based on the observed effect, not the trial’s original design assumptions. The original design assumptions and sample-size justification are not reported in the registry record. The planning calculation assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It asks how many participants per arm would be needed if the effect to detect were the stated fraction of the observed effect. These figures are participants per arm, not total participants and not total events.

The 80% power column means the planned sample size is calculated to give an 80% chance of detecting the specified effect under the stated assumptions, if that effect is real and the model assumptions hold. The 90% power column uses a higher target chance of detection, so it requires more participants for a given effect size. For planning at 100% of the observed effect, the figures are 26 participants per arm for 80% power and 34 participants per arm for 90% power. For planning at 75% of the observed effect, they are 45 participants per arm for 80% power and 60 participants per arm for 90% power. For planning at 50% of the observed effect, they are 97 participants per arm for 80% power and 130 participants per arm for 90% power.

The pattern illustrates a general principle: smaller effects are harder to distinguish from random variation, so more participants are needed to achieve a chosen power target. These planning figures are not observed power, and they do not show that the original trial was designed around the observed contrast. They also depend on the assumptions stated above; different endpoint definitions, analysis methods, or design assumptions could produce different planning requirements.

12. 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 = 0.263; treatment proportion p₁ = 0.647; average p̄ = 0.455.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.7043 = 1.3803.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6498 = 0.5469.
4. n = (1.3803 + 0.5469)² / (0.3839)² = 25.2, rounded up to 26 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 <- 0.263157894736842                          # control proportion
p1 <- 0.647058823529412                          # 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(903760)
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 = 26 (80% power), 34 (90% power)
 75% of observed effect: n per arm = 45 (80% power), 60 (90% power)
 50% of observed effect: n per arm = 97 (80% power), 130 (90% power)
power.prop.test: 25.2 
Simulated power at n = 26 per arm: 0.804 (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.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists Participant Response as a secondary outcome and records two outcome measures, with one secondary outcome posted. Multiple outcomes or repeated analyses can raise a multiplicity concern: when many comparisons are examined, the chance of finding at least one apparently unusual result can increase. Statistical approaches can address this, but the registry record does not report a multiplicity adjustment, a hierarchy among outcomes, or a prespecified testing strategy. It is therefore not possible to determine from the record how the primary comparison was positioned relative to all analyses that may have been planned or conducted.

The registry records zero posted analyses. That absence means the registry entry does not document a formal analysis description for the outcome beyond the posted summary and the recalculated unadjusted comparison. Details about the statistical test selected in advance, covariate adjustment, interim analyses, stopping rules, and sensitivity analyses are not reported in the registry record. None should be inferred from the p-value alone.

Missing-data handling is another important uncertainty. The primary outcome description specifically states that one of 18 participants in the Decitabine arm was not treated due to early death and was excluded from analysis. The record does not explain the rationale for the analysis population beyond that statement, nor does it describe how any other unavailable endpoint information was handled. It does not report whether participants were analyzed according to randomized assignment, treatment received, or another population definition.

As a general concept, analyses based on the randomized groups can preserve the benefits of randomization, while excluding participants after assignment may change the comparison and its interpretation. The available record is not detailed enough to classify the posted result as an intention-to-treat or per-protocol analysis. It also does not report imputation, censoring, or missingness assumptions. These gaps limit assessment of whether the endpoint comparison is robust to incomplete data or analysis-population choices.

14. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

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: lung infection (5.6%), bacteremia - escherichia coli (e. coli) (5.6%), death (5.6%), sepsis (5.3%).

The serious adverse-event figures report 1 affected participant among 19 at risk in the Decitabine + Clofarabine group and 1 among 18 at risk in the Decitabine group. For other adverse events, the figures are 19 affected among 19 at risk and 18 affected among 18 at risk, respectively. These are counts of participants with at least one event in the reported category, not counts of all individual event episodes. The registry does not provide a numerical deaths total in these group summaries.

The listed serious adverse-event terms include Lung Infection at 5.6%, Bacteremia - Escherichia coli (E. coli) at 5.6%, Death at 5.6%, and Sepsis at 5.3%. The record does not specify in this list how terms overlap within participants, so the percentages should not be added or interpreted as separate groups of people. Nor does the list establish a comparative rate for each term by treatment arm.

Safety counts require context about the population at risk, the event definition, observation period, severity grading, and how events were collected. The registry record does not report all of those details, and it does not provide enough information to make a statistical comparison of these adverse events between arms. The posted figures describe reported events in this study; they do not by themselves establish whether an event was caused by a treatment.

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

The registry reports that 26.0% of participants in the Decitabine + Clofarabine arm and 65.0% in the Decitabine arm had event-free survival at 1 year under the stated definition. The unadjusted recalculation from the posted counts gives a risk difference of 0.384, with a 95% confidence interval from 0.083 to 0.685, and an unadjusted p-value of 0.021. These are descriptions of the results observed in this randomized phase 2 study.

The result does not establish that the same contrast would occur in a larger or different population. The outcome groups are small, the confidence intervals are broad, and the record does not report an adjusted analysis or full details of the analysis population. One participant in the Decitabine arm was excluded from the primary outcome analysis after not being treated due to early death. Details about missing endpoint data and the original statistical plan are not reported in the registry record.

The result also does not answer every clinically relevant question. The posted primary outcome is a fixed-time percentage, not a reported analysis of event timing across follow-up. The registry lists Participant Response as a secondary outcome, but the detailed secondary result is not included in the registry record. Safety information is limited to the posted counts and serious-event terms, without enough detail to establish comparative event rates or causal attribution.

As a general statistical principle, a trial estimate is conditional on its endpoint definition, design, participant population, and analysis. The event definition in this study includes death or AML transformation with marrow and/or blood blasts >= 20%. A different endpoint or follow-up approach could yield a different summary. The findings should therefore be read as a record of what was observed under this study’s stated definitions, rather than as a universal conclusion or a treatment recommendation.

16. Why This Trial Matters Statistically

This study offers a useful statistical example of how a striking contrast in percentages can coexist with substantial uncertainty. The posted one-year event-free survival percentages are 26.0% and 65.0%, and the recalculated risk difference is 0.384. Yet the 95% confidence interval extends from 0.083 to 0.685. The interval reminds readers that the point estimate is not known with high precision, particularly when the compared groups include 19 and 17 participants.

The study also illustrates why endpoint definitions and analysis populations matter. Event-free survival at 1 year is a binary summary with a specific composite event definition: death or AML transformation with marrow and/or blood blasts >= 20%. The description also documents exclusion of one Decitabine participant who was not treated due to early death. These details help determine what the percentages actually represent and who contributes to them.

Finally, the record distinguishes observed results from planning assumptions. It provides no report of the original sample-size justification or detailed statistical plan, while illustrative planning based on the observed effect gives different participant requirements for different effect fractions and power targets. This is a reminder that planning from an observed effect is not evidence that the original trial was adequately designed to detect a smaller effect. The registry’s limited analysis detail also shows why readers should separate an unadjusted recalculation from a documented sponsor analysis and assess uncertainty alongside the estimate.

17. 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 trial was a phase 2 study with enrollment of 42, and the primary outcome comparison uses groups of 19 and 17. This limited size constrains precision and makes the estimates sensitive to individual outcomes. The confidence intervals around the recalculated risk difference, relative risk, and odds ratio are broad. The registry record does not report an original sample-size justification, so the intended detectable effect and planned power cannot be assessed.

The primary outcome is stated as a percentage at 12 months/1 year. A time-to-event analysis, details of follow-up duration beyond that assessment, and methods for censoring are not reported in the registry record. The endpoint definition is specified, but the record does not describe how AML transformation was adjudicated or how endpoint assessments were conducted. The primary outcome description reports that one of 18 participants in the Decitabine arm was not treated due to early death and was excluded from analysis; additional details about analysis populations and any other missing outcomes are not reported.

The registry identifies randomization and lack of masking, but does not report the randomization procedure, allocation concealment, stratification, adherence, or treatment discontinuation details. It does not provide an analysis plan, adjusted results, sensitivity analyses, or a multiplicity strategy. The record lists zero analyses, so the posted comparison should not be mistaken for a fully documented inferential analysis. The unadjusted recalculations are based on the reported summary counts and do not resolve these design and reporting gaps.

Safety reporting is also limited. Group summaries give numbers at risk and affected for serious and other adverse events, and the registry lists selected serious-event terms with percentages. The available information does not establish term-specific comparisons by arm, overlap among terms, event attribution, or the full observation and grading procedures. These limitations restrict interpretation of both effectiveness-related and safety-related results.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Decitabine and Clofarabine in Higher Risk Myelodysplastic Syndromes (MDS) — NCT00903760.
• Registry record last updated 2018-11-15; status last verified 2018-10.
• Results first posted 2018-11-15.
• 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.

21. Record Summary

The registry record describes a completed, randomized, unmasked, parallel-group phase 2 study of decitabine alternating with clofarabine versus decitabine alone in higher risk myelodysplastic syndromes. Enrollment was 42. The primary outcome was the percentage with event-free survival at 1 year, defined as no death or transformation to AML with marrow and/or blood blasts >= 20%.

The posted primary percentages were 26.0% for Decitabine + Clofarabine and 65.0% for Decitabine. An unadjusted recalculation from the posted counts estimates a risk difference of 0.384, with a 95% confidence interval from 0.083 to 0.685, and an unadjusted p-value of 0.021. The registry lists Participant Response as a secondary outcome. It also reports adverse-event group counts and selected serious-event terms.

Detailed information about the original sample-size justification, statistical analysis plan, missing-data procedures, and several design and safety methods is not reported in the registry record. The estimates should be interpreted alongside those reporting limitations and their uncertainty.