A Study of Oral Sapacitabine in Elderly Patients With Newly Diagnosed Acute Myeloid Leukemia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Oral Sapacitabine in Elderly Patients With Newly Diagnosed Acute Myeloid Leukemia (NCT01303796), a randomized phase 3 study. The primary outcome is Overall Survival.

2. Clinical Question

The question applies to a selected group: older patients with newly diagnosed AML who were not receiving standard intensive treatment, either because the investigator assessed low-intensity therapy as appropriate or because the patient declined standard induction chemotherapy after discussion with a doctor. The registry record does not establish that its results apply to people outside this eligibility description. Nor does an overall-survival comparison, by itself, describe every aspect of treatment experience. The record lists remission, hematological improvement, transfusions, hospital days, and 1-year survival among secondary outcomes, but those outcomes address different questions from the primary survival comparison.

3. Trial Design

This was a Phase III randomized, parallel-group study with no masking. It enrolled 482 participants and is recorded as completed. The study compared two initial-treatment regimens for patients at least 70 years old with newly diagnosed AML: sapacitabine in alternating cycles with decitabine, and decitabine alone. The record lists Cyclacel Pharmaceuticals, Inc. as sponsor. The study began on 2011-10-01, and primary completion is recorded as 2016-12-15.

The primary endpoint was overall survival, measured in months and described by the median and a 95% confidence interval. The planned analysis description says overall survival was estimated using the Kaplan-Meier method and that a log-rank analysis stratified by randomization stratification factors was used to compare the arms. Cox proportional hazards models were described as a way to identify predictive factors for overall survival. Details of the stratification factors and the specific covariates used in those models are not reported in the registry record.

The registry lists 226 participants in the alternating-regimen arm and 218 in the decitabine arm for the posted primary outcome. These outcome counts do not match the 241 participants listed per arm for the secondary 1-year survival measure; the reason for this difference is not reported in the registry record. The recorded analysis does not give a full account of analysis populations, handling of missing observations, or participant flow. Those omissions matter when interpreting how the randomized groups relate to the estimates displayed for each endpoint.

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: The study comprised a single-arm Lead-in phase to confirm the safety and tolerability of administering sapacitabine in alternating cycles with decitabine prior to patients being enrolled in the parallel group, randomized Phase III phase to compare the two treatment arms. 21 patients were enrolled in the Lead-in phase.

Lead In Phase

Randomized Phase III

Analysis population for the primary outcome, as defined in the registry: Patients \>= 70 years with histologically or pathologically confirmed AML; not treated by systemic therapy; on low-intensity therapy by investigator assessment or the patient has refused standard induction chemotherapy; no prior treatment with HMA or other anti-cancer agents; ECOG status 0-2; adequate hepatic and renal function

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: Elderly patients with newly diagnosed acute myeloid leukemia (AML) for whom the treatment of choice was low-intensity therapy by investigator assessment, or who had refused intensive induction therapy recommended by the investigator.

6. Primary Endpoint

The registry lists the primary outcome as Overall Survival, measured over up to 43 months.

Overall survival measures the time from the study’s specified starting point until death, with participants who have not had the event handled as censored observations in a time-to-event analysis. The registry describes the endpoint as the distribution of overall survival and gives a timeframe of up to 43 months. The exact operational definition of the starting point and censoring rules are not reported in the registry record, although the secondary 1-year survival description refers to time after randomization.

The median is the point in the estimated survival distribution at which half the population is estimated to have experienced the event and half has not. It provides a readable summary of a time-to-event distribution without requiring every participant to have the event observed. Kaplan-Meier estimation is commonly used for this purpose because it incorporates follow-up information from participants whose observation ends before the event is recorded.

Overall survival is used because it is a direct time-to-event outcome that captures survival regardless of the particular pathway leading to death. It differs from remission or disease-response measures, which describe clinical states rather than the duration of survival. The registry reports median overall survival of 5.9 months in the alternating-regimen arm, with a 95% confidence interval from 4.7 to 8.0, and 5.7 months in the decitabine arm, with a 95% confidence interval from 4.9 to 8.2. These are group-level estimates, not predictions for an individual participant.

7. Results Posted in the Registry

• Sapacitabine-decitabine Alternating vs Decitabine: hazard ratio 1.01 (95% CI 0.84–1.23), reported p < 0.025 (as reported in the registry).
• Sapacitabine-decitabine Alternating vs Decitabine: hazard ratio 1.08 (95% CI 0.86–1.35), reported p < 0.025 (as reported in the registry).
• Sapacitabine-decitabine Alternating vs Decitabine: hazard ratio 1.57 (95% CI 1.12–2.19), reported p < 0.025 (as reported in the registry).
• Sapacitabine-decitabine Alternating vs Decitabine: hazard ratio 1.01 (95% CI 0.77–1.32), reported p < 0.025 (as reported in the registry).

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.

1-year Survival

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Decitabine: 34.4% (83/241) versus 33.6% (81/241) on control; risk difference 0.8 percentage points (95% CI -7.6 to 9.3), risk ratio 1.02 (0.80–1.31); unadjusted two-proportion z-test p = 0.848.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Decitabine and the control. The interval gives the range of effects that remain compatible with the data.

The registry names a range of secondary outcomes, including remission categories, hematological improvement, stable disease, blood products transfused, hospitalized days, and 1-year survival. These outcomes address response, supportive care, health-service use, and survival at a specified time. They are related to the primary overall-survival question but are not substitutes for it: changes in disease response or care use do not automatically imply a corresponding change in overall survival.

The detailed secondary result described in the record is 1-year survival. It lists 81 participants in the alternating-regimen arm and 83 in the decitabine arm as alive at that time, with 241 participants listed in each arm for the measure. The record also contains a secondary analysis entry with a hazard ratio, confidence interval, and p-value, but it does not explain how that analysis relates to the binary 1-year survival summary. The reason for the differing group counts between the primary and secondary outcome entries is not reported in the registry record.

Outcome-specific estimates and interpretations for the other named endpoints are not reported in the registry record available for this analysis, so a reliable pattern across all secondary outcomes cannot be established. The existence of multiple secondary outcomes also raises the general possibility of chance findings when many comparisons are made. Without a stated multiplicity strategy, individual unadjusted secondary results carry less weight than a prespecified primary comparison. They should be viewed as supportive or exploratory rather than treated as independent confirmation of the primary result.

Because the secondary outcomes span different measurement types and timeframes, they should be interpreted on their own terms. The record does not give sufficient analysis details to combine them into a single overall conclusion or to determine whether any observed secondary pattern was prespecified.

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

• Complete Remission (CR)
• Complete Remission With Incomplete Platelet Count Recovery (CRp)
• Partial Remission (PR)
• Hematological Improvement
• Stable Disease (SD)
• Blood Products Transfused
• Hospitalized Days
• 1-year Survival

9. Statistical Interpretation

10. Statistical Methods Explained

The registry describes a time-to-event analysis for overall survival. Such methods use both the timing of observed events and information from participants whose follow-up ends before an event is observed. This is different from comparing only the number of events at a single time point, because the length and timing of observation contribute to the analysis.

How does Kaplan-Meier estimation work?

Kaplan-Meier estimation constructs an estimated survival curve from the observed event times. At each event time, the estimate reflects the proportion of participants still under observation who experience the event. Participants who leave observation without a recorded event are censored: their known follow-up contributes information up to that point, but their later event status is unknown. This general explanation does not establish the exact censoring rules used in this trial; those rules are not reported in the registry record.

The median survival is read from the estimated curve at the point where the estimated probability of remaining alive reaches one half. A median is useful as a compact summary, but it does not describe the entire curve. Two survival curves may have similar medians yet differ at other follow-up times, or they may cross. The registry record does not provide enough curve-specific detail to characterize such features beyond the reported analyses.

What does the log-rank comparison assess?

A log-rank test compares the observed event experience between groups over follow-up, taking account of the timing of events and the participants still at risk at those times. The study description specifies stratification by randomization stratification factors. In general, stratification allows comparisons to be made within defined groups and then combined, which can account for those factors in the analysis. The factors themselves and the precise implementation are not reported in the registry record.

A p-value summarizes how compatible the observed data are with a specified null model, subject to the test’s assumptions. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. The repeated p <0.025 entries require particular caution: they accompany different estimates, and the record does not clarify which test, model, or analysis each value belongs to. The primary hazard-ratio interval includes 1 even though its reported p-value is below 0.025. Without more detail, the discrepancy between these summaries cannot be resolved.

What does a hazard ratio represent?

A hazard is the instantaneous event rate among participants who have not yet experienced the event, conditional on the model and the observed follow-up. A hazard ratio compares those rates between groups. A value of 1 is the reference for equal hazards; values above or below 1 indicate a higher or lower estimated hazard for the group in the numerator, respectively, if the comparison direction and model are as specified. A hazard ratio is not a ratio of median survival times and should not be read as a direct percentage change in survival time.

The registry says Cox proportional hazards models were used to identify predictive factors for overall survival. Cox models relate the hazard to group assignment and potentially other covariates. A proportional-hazards interpretation assumes the relative hazard is sufficiently stable over time for the model’s summary to be meaningful. Whether that assumption was assessed, which factors were included, and whether the reported estimates are adjusted are not reported in the registry record.

How should the confidence intervals be read?

A confidence interval conveys uncertainty in an estimate under the statistical procedure used. A 95% confidence interval is not a range expected to contain 95% of individual participant outcomes. In repeated applications under the procedure’s assumptions, intervals constructed in this way have a stated long-run coverage property. Wider intervals indicate less precision than narrower intervals, all else equal.

For a hazard ratio, the value 1 is the no-difference reference. An interval containing 1 includes estimates corresponding to either side of that reference, so it does not by itself establish a direction of association. The posted primary interval, 0.837 to 1.226, contains 1. The registry also reports intervals for several other estimates, but does not identify the full model or analysis population for each. Confidence intervals should therefore be interpreted alongside their corresponding estimates and the record’s incomplete descriptions, not as stand-alone proof.

These explanations describe general statistical concepts. They do not fill in unreported trial-specific choices such as censoring, model diagnostics, adjustment variables, missing-data handling, or the analysis population.

11. Confidence Intervals and Precision

Precision concerns how much uncertainty surrounds an estimate. The primary median estimates are 5.9 months and 5.7 months, with respective 95% confidence intervals of 4.7 to 8.0 and 4.9 to 8.2 months. The intervals overlap substantially. This indicates that each median is estimated with uncertainty; the overlap alone is not a formal test of the difference between the arms.

The reported hazard ratio of 1.013 has a 95% confidence interval from 0.837 to 1.226. Because this interval includes 1, the values compatible with the interval include estimates on either side of the equal-hazard reference. The span should not be translated into a range of individual survival experiences. It describes uncertainty about a group-level model estimate, conditional on the analysis assumptions.

The registry gives several other hazard-ratio estimates. The reported intervals range from relatively narrow to broad, and some estimates have no interval at all. The entries without intervals cannot be judged for precision from the estimate alone. Moreover, the registry does not state what distinguishes these analyses, making direct comparison among them difficult.

The reported p <0.025 accompanying the primary hazard ratio is not readily aligned with an interval that includes 1 under a conventional two-sided test of the same estimate. The registry does not explain this apparent tension, so neither the p-value nor the confidence interval should be silently substituted for the other. The record’s incomplete descriptions also limit assessment of model uncertainty and the relevance of the estimates to the randomized population.

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 figures are illustrations based on the observed effect, not the original trial design. For the overall-survival time-to-event calculations, figures refer to the total number of events required, not the number of participants per arm. For the binary 1-year survival calculation, figures refer to participants per arm. The assumptions listed for these illustrations are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; the effect to detect is the stated fraction of the observed effect. The original trial’s own sample-size assumptions are not reported in the registry record.

The 80% power column represents the event or participant count that the planning calculation estimates would give an 80% chance of detecting the specified effect under its assumptions, if that effect and the statistical model were correct. The 90% power column uses a higher target chance of detection and therefore generally requires a larger planned count. Rows for 100%, 75%, and 50% of the observed effect illustrate progressively smaller target effects; smaller effects need more participants or events to distinguish them from random variation.

These figures are not observed power. They are hypothetical planning illustrations, not evidence about how likely the completed study was to detect an effect after the results were known. They also do not reconstruct the design choices used by the investigators. Read the units carefully: total events for the primary time-to-event planning figures, and participants per arm for the 1-year survival illustration.

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

d is the total number of events needed (Schoenfeld's formula) with 1:1 allocation and proportional hazards.

Worked calculation for this trial

1. Hazard ratio HR = 1.013, so ln(HR) = 0.0129.
2. d = 4 × (1.96 + 0.842)² / [ln(1.013)]² = 188,190.0, rounded up to 188,190 events in total 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
hr <- 1.013                              # observed hazard ratio
events_needed <- function(power, frac = 1) ceiling(round(4 * (z_a + qnorm(power))^2 / log(hr^frac)^2, 6))   # Schoenfeld, 1:1 allocation
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: events = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(events_needed(0.80, f), big.mark = ","), format(events_needed(0.90, f), big.mark = ",")))

# Simulation: given d total events under proportional hazards, events in the treatment arm are Binomial(d, hr / (1 + hr)).
# The score (log-rank type) statistic is (D1 - d / 2) / sqrt(d / 4).
set.seed(1303796)
d    <- events_needed(0.80)
reps <- 20000
D1 <- rbinom(reps, d, hr / (1 + hr))
z  <- (D1 - d / 2) / sqrt(d / 4)
cat(sprintf("Simulated power with %s events: %.3f (%d replicates)\n", format(d, big.mark = ","), mean(abs(z) > z_a), reps))

Output from running the code

100% of observed effect: events = 188,190 (80% power), 251,933 (90% power)
 75% of observed effect: events = 334,560 (80% power), 447,881 (90% power)
 50% of observed effect: events = 752,760 (80% power), 1,007,732 (90% power)
Simulated power with 188,190 events: 0.800 (20000 replicates)

The simulation uses the fact that, with d events in total and proportional hazards, the number of events in the treatment arm is binomial. It is an approximation that conditions on the total number of events, and it should land close to the target power.

14. Analyses Reported by the Sponsor

The primary endpoint description specifies Kaplan-Meier estimation and a log-rank comparison stratified by randomization factors, with Cox models used to identify predictive factors. The registry’s listed reported analyses include a hazard ratio of 1.013 with a 95% confidence interval from 0.837 to 1.226 and p <0.025. It also lists other hazard-ratio estimates, some with intervals and some without, but does not explain the population, adjustment, or purpose associated with each entry.

The recalculated summaries are simple unadjusted recalculations from posted summary statistics. For time-to-event outcomes, the site’s stated approach uses the sponsor’s reported hazard ratios rather than deriving a new comparative estimate from the median survival values. The discrepancy indicator is false, so no discrepancy is flagged between the posted and recalculated results. This does not resolve the unexplained relationship between the reported p-value and the primary interval, nor does it identify which of the multiple listed analyses should be treated as the definitive model.

When a registry analysis and a recalculation disagree, the sponsor’s own analysis takes precedence; that situation is not flagged for this record. The sponsor’s reported analysis should still be interpreted with the incomplete descriptions and the apparent p-value and interval tension in mind. No unreported model details can be inferred from the numerical entries alone.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists 14 outcome measures and 14 analyses, as well as multiple secondary outcomes. Testing several outcomes or analyses creates a multiplicity issue: even when there is no underlying effect, the chance of at least one apparently unusual result can increase as the number of tests grows. Statistical approaches can address this by controlling error rates or by distinguishing confirmatory outcomes from exploratory ones. Whether any such approach was planned or used is not reported in the registry record.

The record identifies overall survival as the primary endpoint and describes a stratified log-rank comparison and Cox modeling. It does not report a full statistical analysis plan, a hierarchy for secondary endpoints, a multiplicity adjustment, or a clear mapping from each of the multiple listed hazard-ratio estimates to an analysis question. It also repeats p <0.025 across entries with different estimates, leaving the intended interpretation of those values unclear. These gaps make it difficult to assess which comparisons were confirmatory and how to weigh the collection of reported results.

Missing-data and analysis-population details are also limited. The registry does not report how missing outcome measurements, loss to follow-up, or censoring were handled beyond naming Kaplan-Meier estimation. It does not identify whether the primary analysis followed an intention-to-treat population, a modified intention-to-treat population, or another defined set. The sample sizes attached to the primary outcome and 1-year survival outcome differ, but the reason is not reported in the registry record.

These omissions do not demonstrate that the trial lacked appropriate procedures; they mean that those procedures cannot be evaluated from the registry record alone. A full report or protocol would be needed to clarify endpoint hierarchy, analysis populations, missing-data rules, model specifications, and multiplicity handling.

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 (24.0%), pneumonia (23.7%), bacteraemia (14.3%), disease progression (13.1%), sepsis (10.2%).

The serious adverse event summary lists 199 affected participants among 236 at risk in the sapacitabine-decitabine alternating group, and 188 affected among 233 at risk in the decitabine group. It also lists a lead-in group, with 17 affected among 21 at risk. These are counts from the registry’s safety summary; the reason for the lead-in group and its relation to the randomized comparison are not reported in the registry record. The safety denominators also differ from the primary outcome group counts.

The listed serious adverse event terms include febrile neutropenia, pneumonia, bacteraemia, disease progression, sepsis, and anaemia. The registry gives percentages for these terms, but does not clearly state the denominator basis in the information. The groups also have death counts listed, but the record does not establish that these deaths were caused by treatment or provide enough context to interpret them as treatment-related.

Serious adverse event counts describe events recorded under the registry’s safety reporting, not necessarily the number of distinct event episodes or the causal effect of a regimen. Differences in denominators, exposure, follow-up, and reporting definitions can affect comparisons. The summary should therefore be read as a description of recorded safety events, not as a causal comparison or a complete safety assessment.

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

The registry’s primary outcome reports median overall survival of 5.9 months for sapacitabine alternating with decitabine and 5.7 months for decitabine alone. The reported hazard ratio is 1.013, with a 95% confidence interval from 0.837 to 1.226. These figures describe estimated group-level outcomes in the trial population; they do not determine what will happen to an individual participant.

The hazard-ratio interval includes 1, and the record’s reported p-value appears difficult to reconcile with that interval under a conventional two-sided analysis. The registry does not explain this issue or fully distinguish the several other listed estimates. Accordingly, the figures do not support a simple, unqualified claim about a difference between regimens. They also do not prove that the regimens are equivalent: demonstrating equivalence or noninferiority requires a defined margin and an analysis designed for that question, which are not reported in the registry record.

Overall survival is only one dimension of a treatment comparison. The record names secondary outcomes involving remission, blood counts, transfusions, hospitalization, and 1-year survival, but it does not provide enough detailed results and analysis context to integrate all these outcomes into a comprehensive account of benefit and harm. The serious adverse event summary likewise describes recorded events but does not establish causality.

These findings should be understood within the enrolled population: older adults with newly diagnosed AML for whom standard intensive treatment was not recommended or was declined. The registry record alone cannot establish how well the estimates apply to other populations, resolve unreported design details, or replace a full clinical and statistical assessment. No treatment recommendation follows from these numerical summaries.

18. Why This Trial Matters Statistically

This trial raises several important statistical issues common in clinical research. First, overall survival is a time-to-event outcome, so the analysis depends on event timing, follow-up, censoring, and the structure of the survival curves—not only on the two reported medians. Kaplan-Meier estimation and a log-rank comparison are recognizable approaches for such data, but full interpretation still depends on the details of their implementation.

Second, the principal hazard-ratio estimate is accompanied by an interval that includes 1 and a reported p-value below 0.025. The registry does not explain this apparent tension. A reader should not treat a single p-value as a complete account of evidence when the corresponding estimate, interval, and analysis description do not align transparently. Statistical reporting is most useful when the estimate, uncertainty interval, test, and analysis population clearly refer to the same comparison.

Third, the record lists multiple additional hazard-ratio estimates without enough context to identify the questions or models they represent. It also names many secondary outcomes. Without a clear analysis hierarchy and multiplicity strategy, isolated secondary findings can be overinterpreted. Their interpretation depends on prespecification, analysis choices, and whether findings are consistent across related outcomes.

Finally, the record illustrates why denominators and outcome populations matter. Primary and secondary entries list different group counts, and the reason is not described. A registry can communicate important design and outcome information, but complete interpretation also requires transparent reporting of participant flow, analysis populations, missing data, and model details. These are statistical reporting considerations, not conclusions about the quality or conduct of the trial.

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.

Several limitations are specific to the information recorded for this trial. The primary endpoint description identifies Kaplan-Meier estimation, a stratified log-rank comparison, and Cox modeling, but the registry does not identify the stratification factors or explain which covariates and populations correspond to the several reported hazard-ratio estimates. Some estimates have no confidence intervals, and multiple entries carry the same reported p-value despite differing estimates. The record does not clarify these analysis entries.

The primary hazard ratio’s 95% confidence interval includes 1, while its reported p-value is p <0.025. The registry provides no explanation for this apparent tension. The record’s discrepancy indicator is false, but that does not resolve the interpretive question or supply missing model details.

The outcome counts differ across the primary overall-survival and secondary 1-year survival entries. The record does not explain the difference or provide participant-flow details that would clarify it. It also does not report the complete operational definition of the survival endpoint, censoring rules, missing-data procedures, or the analysis population used for each estimate. The eligibility description contains additional criteria but does not provide a complete account in the available record.

Finally, the registry names many secondary outcomes but does not give enough outcome-specific results to assess their overall pattern. It does not state a multiplicity strategy or establish how the additional analyses relate to the primary question. These limitations describe gaps in the registry record; they do not establish that the underlying trial documents or analyses lack those details.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study of Oral Sapacitabine in Elderly Patients With Newly Diagnosed Acute Myeloid Leukemia — NCT01303796.
• Registry record last updated 2022-06-22; status last verified 2022-05.
• Results first posted 2022-06-22.
• 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 III randomized, parallel-group study of sapacitabine administered in alternating cycles with decitabine versus decitabine alone for older patients with newly diagnosed AML who were not receiving standard intensive treatment. Enrollment was 482, and masking was none. The primary endpoint was overall survival, estimated by Kaplan-Meier methods and compared using a stratified log-rank analysis as described in the record.

Reported median overall survival was 5.9 months in the alternating-regimen arm and 5.7 months in the decitabine arm. The reported hazard ratio was 1.013, with a 95% confidence interval from 0.837 to 1.226 and p <0.025. The record also lists additional estimates but does not fully explain their analysis context. The primary interval includes 1, and the relationship between that interval and the reported p-value is not explained.

The record names secondary efficacy and care-use outcomes and includes serious adverse event summaries. Full analysis-population, missing-data, multiplicity, and model details are not reported in the registry record. The posted summaries therefore require cautious interpretation.