← Clinical Trials
Chronic Myelogenous Leukemia Phase 3 Randomized NCT00471497

ENESTnd: Complete Statistical Analysis of Nilotinib in Chronic Myelogenous Leukemia

An independent statistical analysis of the randomized phase 3 ENESTnd study comparing imatinib with two nilotinib regimens in adults with newly diagnosed Philadelphia chromosome positive chronic myelogenous leukemia in chronic phase.

Trial status: COMPLETED  ·  Enrollment: 846  ·  Primary completion: September 2, 2009
Scope of this record

This page separates reported trial results from statistical interpretation. Trial-specific numerical results, endpoint definitions, analysis populations, and reported statistical methods are restricted to the ClinicalTrials.gov record for NCT00471497.

Registry context: This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.

1. Trial at a Glance

ENESTnd was a completed, randomized phase 3 study of adults with newly diagnosed Philadelphia chromosome positive (Ph+) chronic myelogenous leukemia in chronic phase (CML-CP). The registry reports 846 randomized patients and three study arms involving imatinib and nilotinib.

846
Enrolled
Full analysis set
3
Arms
Randomized
2
Primary endpoints
Binary outcomes
11
Statistical analyses
Posted on registry
FeatureENESTnd
TrialENESTnd
PhasePhase 3
ConditionMyelogenous Leukemia, Chronic
PopulationAdults with newly diagnosed Philadelphia chromosome positive (Ph+) chronic myelogenous leukemia in chronic phase (CML-CP)
AllocationRandomized
Registry design modelSINGLE_GROUP
MaskingNone
Primary purposeTreatment
Enrollment846
Arms3
InterventionsNilotinib; imatinib
Lead sponsorNovartis Pharmaceuticals
Sponsor typeIndustry
Study periodJuly 31, 2007 to September 2, 2009 primary completion
ClinicalTrials.govNCT00471497
Registry design-field note: the ClinicalTrials.gov record identifies the allocation as RANDOMIZED and report 3 arms, while the registry's normalized design-model field is SINGLE_GROUP. Those fields are internally inconsistent. This page reports the fields as reported in the registry rather than silently resolving the discrepancy.

2. Clinical Question

The statistical question is whether the percentage of participants achieving major molecular response (MMR) differs between the randomized treatment groups at specified time points, using the registry-defined MMR criterion.

Population

Adults with newly diagnosed Philadelphia chromosome positive (Ph+) chronic myelogenous leukemia in chronic phase.

Intervention groups

The registry reports imatinib 400 mg QD and two nilotinib regimens: 300 mg BID and 400 mg BID.

Comparator

For the posted primary comparisons, imatinib 400 mg QD is compared separately with nilotinib 300 mg BID and nilotinib 400 mg BID.

Primary question

Do the randomized groups differ in MMR rate at 12 months, including analyses by Sokal risk group with imputation?

3. Trial Design

01
Randomize846 patients
02
3 armsImatinib / nilotinib
03
AssessMMR by RQ-PCR
04
12 monthsPrimary endpoint
05
24 monthsDurable MMR endpoint
ARM 1 · Safety denominator 280

Imatinib 400 mg QD

  • Imatinib
  • Serious adverse events: 96/280
ARM 2 · Safety denominator 279

Nilotinib 300 mg BID

  • Nilotinib
  • Serious adverse events: 112/279
ARM 3 · Safety denominator 277

Nilotinib 400 mg BID

  • Nilotinib
  • Serious adverse events: 126/277

The three safety denominators sum to 836 rather than the 846 patients in the full analysis set. The ClinicalTrials.gov record therefore support reporting these as the posted safety denominators, but do not provide enough information to explain the difference.

4. Endpoints

EndpointTime frameRegistry definition
Major Molecular Response Rate (MMR) at 12 Months Between All 3 Arms - With Imputation Baseline, 12 months MMR is defined as the percentage of participants in MMR (reduction of ≥ 3 logs in BCR-ABL transcripts compared to the standardized baseline established in IRIS, or ≤ 0.1% BCR-ABL/ABL % by international scale and measured by real-time quantitative polymerase chain reaction (RQ-PCR)) at 12 months.
Percentage of Participants With MMR at 12 Months Between All 3 Arms by Sokal Risk Group With Imputation 12 months MMR is defined as the percentage of participants in MMR (reduction of ≥ 3 logs in BCR-ABL transcripts compared to the standardized baseline established in IRIS, or ≤ 0.1% BCR-ABL/ABL % by international scale and measured by real-time quantitative polymerase chain reaction (RQ-PCR)) at 12 months.
Rates of Durable MMR at 24 Months Between All 3 Arms 24 months Percentage of participants
Rate of Major Molecular Response (MMR) at 12 Months Between Two Nilotinib Arms 12 months Percentage of participants

The two registered primary endpoints are binary outcomes. The posted statistical analyses use the difference in response rate as the principal effect measure, with two-sided 95% confidence intervals for the reported comparisons.

5. Statistical Methodology

Full analysis set and randomized treatment assignment

The primary analyses were reported in the Full analysis Set (FAS), which contained all 846 randomized patients. Patients were analyzed according to the treatment to which they were randomized, regardless of actual treatment received.

Analysis principle
Analyze participants according to randomized assignment

This is an intention-to-treat principle. Its purpose is to preserve the treatment-group comparability created by randomization when estimating the effect of assignment to a treatment strategy.

Cochran-Mantel-Haenszel test

The registry explicitly reports the Cochran-Mantel-Haenszel method for the overall 12-month MMR comparison between imatinib and each nilotinib regimen, and for the comparison between the two nilotinib arms. The method belongs to the categorical-data family and is appropriate for comparing response rates while accounting for stratification when appropriate.

For a binary response, the underlying data can be represented as a two-by-two table: treatment group by MMR status. The analysis then evaluates whether the response distributions differ between treatment groups under the specified categorical-data framework.

Difference in response rate

The reported effect measure is a difference in response rate. Conceptually, a response-rate difference compares the percentage responding in one group with the percentage responding in another group:

Conceptual effect measure
Difference = Response rate in Group A − Response rate in Group B

The ClinicalTrials.gov record does not specify the coding direction used to construct each reported difference. Therefore, the numerical estimates below are reported exactly as posted without assigning an unreported direction to the contrast.

Confidence intervals

Each formal comparison with an estimate has a two-sided 95% confidence interval. The interval describes statistical uncertainty around the estimated response-rate difference under the analysis framework. It is not a prediction interval for individual patients and does not describe the range of responses that future patients must experience.

Imputation

The primary endpoint names explicitly include With Imputation. The ClinicalTrials.gov record does not identify the exact imputation algorithm or assumptions used. That distinction matters because a binary endpoint with missing assessments can change depending on how missing observations are classified or otherwise handled.

6. Statistical Methods Explained

Why use a Cochran-Mantel-Haenszel test?

The MMR endpoint is binary: a participant either meets the registry definition of MMR at the specified assessment or does not. The Cochran-Mantel-Haenszel framework is designed for categorical comparisons and can provide a treatment comparison while accounting for categorical strata when those strata are part of the analysis.

What does a response-rate difference measure?

A response-rate difference is an absolute measure rather than a ratio. For example, a difference of 20 percentage points would mean that the observed response percentages differ by 20 percentage points, subject to the direction in which the contrast was coded. It is different from a relative risk, odds ratio, or hazard ratio.

Why does the analysis use the full analysis set?

The FAS contains the 846 randomized patients, and patients were analyzed according to randomized treatment regardless of actual treatment received. This preserves the treatment assignment created by randomization and avoids redefining the comparison based on treatment exposure after randomization.

Why does imputation matter for a binary endpoint?

If the MMR status at 12 months is unavailable for some participants, the analysis needs a rule for handling those missing observations. Different rules can lead to different response-rate estimates. Because the ClinicalTrials.gov record identifies imputation but do not specify the algorithm, the exact sensitivity of the result to missing-data assumptions cannot be evaluated from the registry information alone.

What does a 95% confidence interval tell us?

The confidence interval communicates precision around the estimated response-rate difference. A narrower interval indicates greater statistical precision than a wider interval, all else equal. The interval should not be interpreted as saying that 95% of individual patient effects lie within its boundaries.

What does the p-value tell us?

A p-value evaluates the compatibility of the observed data with a specified null hypothesis under the statistical model and testing framework. It does not measure the size of the treatment effect, clinical importance, or probability that one treatment is effective. Those questions require the effect estimate, confidence interval, endpoint definition, and clinical context to be considered together.

7. Primary Results: MMR at 12 Months

The first registered primary endpoint evaluated MMR at 12 months in the full analysis set, with imputation. The registry reports two formal pairwise comparisons involving imatinib and the two nilotinib regimens.

Imatinib 400 mg QD vs Nilotinib 300 mg BID

Difference in response rate

22.1

95% CI: 14.5–29.6   ·   P < 0.0001

Two-sided 95% confidence interval; Cochran-Mantel-Haenszel test; FAS of 846 randomized patients.

Clinical Biostats interpretation

The registry reports a 22.1 difference in response rate for the comparison of imatinib 400 mg QD with nilotinib 300 mg BID. The numerical estimate describes an absolute difference in MMR response rates, not a relative effect such as a risk ratio or odds ratio.

The ClinicalTrials.gov record does not specify the coding direction of the difference. Therefore, the value 22.1 should not be rewritten as a claim that one named arm had a 22.1-percentage-point higher response rate than the other without the underlying response-rate values or contrast convention.

The 95% CI of 14.5 to 29.6 describes uncertainty around the estimated difference. The p-value of <0.0001 addresses statistical evidence under the testing framework; it does not measure the magnitude or clinical importance of the response difference.

Because the endpoint uses imputation and the ClinicalTrials.gov record does not specify the imputation procedure, the numerical result should also be understood in the context of the missing-data assumptions used by the registry analysis.

Imatinib 400 mg QD vs Nilotinib 400 mg BID

Difference in response rate

20.4

95% CI: 12.9–28.0   ·   P < 0.0001

Two-sided 95% confidence interval; Cochran-Mantel-Haenszel test; FAS of 846 randomized patients.

Clinical Biostats interpretation

The registry reports a 20.4 difference in response rate for the comparison of imatinib 400 mg QD with nilotinib 400 mg BID. This is an absolute response-rate contrast, not a statement about relative risk and not a time-to-event effect.

Again, the ClinicalTrials.gov record does not identify which group was subtracted from which. The estimate should therefore be preserved as reported rather than assigning a direction that is not documented in the ClinicalTrials.gov record.

The 95% CI of 12.9 to 28.0 quantifies uncertainty around the estimate. The p-value of <0.0001 indicates strong statistical evidence under the reported test, but a p-value does not tell us whether a particular numerical difference is clinically important.

The analysis is based on the randomized FAS and an endpoint specified with imputation. Those features are important to interpreting the estimate and to reproducing the analysis.

8. Primary Results by Sokal Risk Group

The second registered primary endpoint examined MMR at 12 months by Sokal risk group, with imputation. The ClinicalTrials.gov record provides six pairwise estimates: three risk categories for each of the two imatinib-versus-nilotinib comparisons.

Sokal risk group Comparison Difference 95% CI P-value
Low Imatinib 400 mg QD vs Nilotinib 300 mg BID 14.8 2.1–27.5 Not reported
Low Imatinib 400 mg QD vs Nilotinib 400 mg BID 27.4 14.6–40.2 Not reported
Intermediate Imatinib 400 mg QD vs Nilotinib 300 mg BID 27.7 15.0–40.4 Not reported
Intermediate Imatinib 400 mg QD vs Nilotinib 400 mg BID 17.2 4.6–29.8 Not reported
High Imatinib 400 mg QD vs Nilotinib 300 mg BID 24.4 10.7–38.1 Not reported
High Imatinib 400 mg QD vs Nilotinib 400 mg BID 15.4 2.1–28.6 Not reported

The registry's statistical-analysis records do not report a formal method for these Sokal risk-group estimates. They do identify the FAS and intention-to-treat principle and report two-sided 95% confidence intervals.

Subgroup interpretation: differences among the six point estimates do not by themselves demonstrate that the treatment effect differs across Sokal risk groups. A formal claim of effect modification requires an appropriate comparison of treatment effects across groups, such as an interaction analysis. The ClinicalTrials.gov record does not report such an analysis.

Low-risk group

300 mg BID comparison

Difference in response rate: 14.8; 95% CI 2.1–27.5.

400 mg BID comparison

Difference in response rate: 27.4; 95% CI 14.6–40.2.

Clinical Biostats interpretation

Both low-risk comparisons have confidence intervals that are above zero under the reported contrast coding. However, the registry does not provide a p-value for either subgroup analysis or a formal interaction test. A subgroup estimate is therefore best treated as an estimate within that risk category rather than as proof that the treatment contrast differs from other Sokal groups.

Intermediate-risk group

300 mg BID comparison

Difference in response rate: 27.7; 95% CI 15.0–40.4.

400 mg BID comparison

Difference in response rate: 17.2; 95% CI 4.6–29.8.

Clinical Biostats interpretation

The intermediate-risk estimates are both positive under the registry's reported contrast coding, with two-sided 95% confidence intervals of 15.0–40.4 and 4.6–29.8. Their differing point estimates should not be treated as evidence of a dose-specific difference without a formal comparison of the two treatment effects.

High-risk group

300 mg BID comparison

Difference in response rate: 24.4; 95% CI 10.7–38.1.

400 mg BID comparison

Difference in response rate: 15.4; 95% CI 2.1–28.6.

Clinical Biostats interpretation

The high-risk estimates are 24.4 and 15.4, respectively, with two-sided 95% confidence intervals that remain above zero under the reported contrast coding. The registry does not provide a formal heterogeneity test, so these subgroup results should be used to describe the observed estimates rather than to establish differences between Sokal risk categories.

9. Secondary Endpoint Results

Durable MMR at 24 Months

ComparisonEndpointDifference in response rate95% CIP-value
Imatinib 400 mg QD vs Nilotinib 300 mg BID Rates of Durable MMR at 24 Months Between All 3 Arms 21.3 13.9–28.8 Not reported
Imatinib 400 mg QD vs Nilotinib 400 mg BID Rates of Durable MMR at 24 Months Between All 3 Arms 18.7 11.3–26.0 Not reported

The secondary endpoint is also binary and is reported as a difference in response rate with two-sided 95% confidence intervals. The ClinicalTrials.gov record does not identify the statistical method used for these two comparisons.

Clinical Biostats interpretation

The 24-month durable-MMR estimates are 21.3 and 18.7 for the two imatinib-versus-nilotinib comparisons. These are absolute response-rate differences as reported by the registry. Because p-values and a formal analysis method are not posted on ClinicalTrials.gov for these secondary comparisons, the confidence intervals provide the principal quantitative information about their precision in the posted statistical-analysis record.

Comparison of the Two Nilotinib Arms

MMR at 12 months

−1.6

95% CI: −9.8–6.6   ·   P = 0.6987

Nilotinib 300 mg BID vs Nilotinib 400 mg BID; Cochran-Mantel-Haenszel test; FAS of 846 randomized patients.

Clinical Biostats interpretation

The reported absolute difference between the two nilotinib arms is −1.6, with a two-sided 95% CI of −9.8 to 6.6. The interval includes zero, meaning that the registry estimate is compatible with no response-rate difference under the stated statistical framework.

The p-value of 0.6987 is not a measure of the size of the difference. The point estimate itself is close to zero relative to the scale of the response-rate difference, while the confidence interval shows the range of values supported with the stated precision.

This comparison is particularly useful for distinguishing statistical significance from effect size: a large p-value does not prove that the two regimens are identical, just as a small p-value would not by itself establish clinical importance.

10. Statistical Analysis Summary

Endpoint / comparisonPopulationMethodEffect measureEstimate95% CIP-value
MMR at 12 months: imatinib vs nilotinib 300 mg BID FAS, n=846 Cochran-Mantel-Haenszel Difference in response rate 22.1 14.5–29.6 <0.0001
MMR at 12 months: imatinib vs nilotinib 400 mg BID FAS, n=846 Cochran-Mantel-Haenszel Difference in response rate 20.4 12.9–28.0 <0.0001
MMR at 12 months by Sokal risk group FAS, n=846 Not reported Difference in response rate 14.8 to 27.7 2.1–40.4 depending on comparison Not reported
Durable MMR at 24 months: imatinib vs nilotinib 300 mg BID FAS, n=846 Not reported Difference in response rate 21.3 13.9–28.8 Not reported
Durable MMR at 24 months: imatinib vs nilotinib 400 mg BID FAS, n=846 Not reported Difference in response rate 18.7 11.3–26.0 Not reported
MMR at 12 months: nilotinib 300 mg BID vs nilotinib 400 mg BID FAS, n=846 Cochran-Mantel-Haenszel Absolute difference −1.6 −9.8–6.6 0.6987

The registry reports 11 statistical analyses in total, of which 8 are primary-endpoint analyses and all 8 primary analyses include an estimate and confidence interval. The ClinicalTrials.gov record indicates that no primary analyses were omitted for length.

11. Safety Results

The ClinicalTrials.gov record reports serious adverse events by randomized arm using the following affected/at-risk counts.

ArmSerious adverse eventsAt risk
Imatinib 400 mg QD96280
Nilotinib 300 mg BID112279
Nilotinib 400 mg BID126277
All Patients334836

The registry therefore reports 334/836 serious adverse events across all patients in the registry-reported safety analysis. The three arm-specific denominators are 280, 279, and 277, which sum to 836.

Serious adverse events — affected / at risk
Imatinib 400 mg QD
96 / 280
Nilotinib 300 mg BID
112 / 279
Nilotinib 400 mg BID
126 / 277

The bar visualization is based only on the affected and at-risk counts reported in the ClinicalTrials.gov record. It is intended to display the relative numerator and denominator information, not to substitute an inferential safety analysis.

12. Safety Interpretation

Serious adverse events

The ClinicalTrials.gov record reports 96/280 for imatinib 400 mg QD, 112/279 for nilotinib 300 mg BID, and 126/277 for nilotinib 400 mg BID.

All-patient record

The registry reports 334/836 for all patients in the registry-reported serious-adverse-event analysis.

No derived inference

The ClinicalTrials.gov record does not include a formal statistical comparison of serious adverse-event rates, so this page does not assign a p-value or confidence interval to the safety comparison.

Different denominators

The safety denominator of 836 differs from the FAS of 846. The ClinicalTrials.gov record does not explain that difference.

13. Missing Data and Imputation

Imputation is explicitly part of both registered primary endpoints. That is statistically important because MMR at 12 months is a binary outcome assessed at a defined time point. If the MMR status is unavailable, the analysis must determine how that missing observation enters the response-rate calculation.

Why the assumption matters
Observed response rate depends on both observed outcomes and the treatment of missing outcomes

The ClinicalTrials.gov record identifies that imputation was used but do not specify the imputation rule. Consequently, an independent reproduction of the exact primary estimates would require the underlying analysis specification or patient-level information.

This limitation does not invalidate the posted results. It limits what can be learned from the registry record alone about the sensitivity of the estimate to the missing-data mechanism and the particular imputation procedure.

14. Multiplicity and Multiple Comparisons

The ClinicalTrials.gov record contains multiple related comparisons: two overall imatinib-versus-nilotinib primary comparisons, six Sokal-risk-group estimates, two 24-month durable-MMR comparisons, and a comparison between the two nilotinib arms. The ClinicalTrials.gov record does not specify an alpha-allocation strategy, multiplicity-adjustment procedure, hierarchical testing sequence, or familywise error-control plan.

Interpretation caution: the presence of multiple estimates means that the individual p-values and confidence intervals should be interpreted according to the prespecified testing framework, if one exists. Because that framework is not included in the ClinicalTrials.gov record, this page does not reinterpret the nominal results as if a particular multiplicity procedure had been used.
FeatureWhat the ClinicalTrials.gov record establishesWhat is not reported
Primary endpoints2 registered primary binary endpointsDetailed multiplicity hierarchy
Overall primary comparisonsImatinib vs each nilotinib regimenExplicit alpha allocation
Sokal analysesLow, intermediate, and high estimatesFormal interaction test
Secondary analysesDurable MMR at 24 months and two-nilotinib-arm comparisonMultiplicity adjustment

15. Randomization and Intention-to-Treat Analysis

Randomization is the core design feature that supports the causal comparison between treatment assignments. The registry-reported analysis population explicitly contains the 846 randomized patients, and patients were analyzed according to randomized treatment regardless of actual treatment received.

Why this matters statistically

Analyzing randomized patients according to assignment maintains the comparison created at randomization. If patients who discontinue, switch, or otherwise deviate from treatment were instead reassigned to groups based on what they actually received, the original randomized comparison could be altered.

The ClinicalTrials.gov record therefore provide a clear ITT-type principle for the efficacy analyses. They do not, however, provide enough detail to determine every population used across all 31 posted outcome measures.

16. Confidence Intervals and Effect Size

The primary overall MMR comparisons provide a useful demonstration of why confidence intervals should accompany p-values.

ComparisonEstimate95% CI widthP-value
Imatinib 400 mg QD vs Nilotinib 300 mg BID22.114.5–29.6<0.0001
Imatinib 400 mg QD vs Nilotinib 400 mg BID20.412.9–28.0<0.0001
Nilotinib 300 mg BID vs Nilotinib 400 mg BID−1.6−9.8–6.60.6987

The first two estimates are accompanied by p-values below 0.0001, whereas the nilotinib-versus-nilotinib comparison has a p-value of 0.6987 and a confidence interval spanning zero. This illustrates the distinction between evidence against a null hypothesis and the magnitude and precision of an estimated effect.

17. What the Response-Rate Difference Does — and Does Not — Mean

Statistical interpretation

A reported response-rate difference is an absolute comparison of two binary response probabilities. It is not a hazard ratio, does not describe time to response, and does not quantify the proportion of patients who personally benefit.

Why the direction must be preserved carefully

The ClinicalTrials.gov record identifies the two groups being compared and the numerical difference, but do not specify the subtraction convention used to generate the estimate. For that reason, the safest registry-faithful presentation is to report the estimate exactly as posted rather than infer which arm's response rate was subtracted from which.

Why the confidence interval matters

The 95% confidence interval communicates uncertainty around the response-rate difference. For example, the nilotinib 300 mg BID versus nilotinib 400 mg BID estimate of −1.6 has a 95% CI of −9.8 to 6.6, showing substantially more uncertainty about the direction and magnitude of the difference than the point estimate alone conveys.

18. Limitations

19. Why This Trial Matters Statistically

ENESTnd is a useful teaching case because it shows how a randomized multi-arm trial can generate several distinct statistical questions from the same binary molecular-response endpoint.

ConceptHow it appears in ENESTnd
Randomization846 patients were randomized across 3 arms.
Intention-to-treat analysisThe FAS contained all 846 randomized patients, analyzed according to randomized treatment.
Binary endpointMMR is reported as the percentage of participants meeting a defined molecular-response criterion.
Cochran-Mantel-Haenszel testUsed for the reported overall MMR comparisons and the comparison between the two nilotinib arms.
Effect measureDifference in response rate / absolute difference.
Confidence intervalsTwo-sided 95% intervals accompany all eight primary-endpoint estimates.
ImputationExplicitly incorporated into the two registered primary endpoints.
Subgroup analysisMMR at 12 months was reported by Sokal risk group.
MultiplicityMultiple primary and secondary comparisons are present, while the ClinicalTrials.gov record does not specify the adjustment strategy.
Safety analysisSerious adverse events are reported by arm using affected/at-risk counts.

20. Clinical Interpretation vs Statistical Interpretation

Statistical interpretation

The registry reports positive response-rate differences for the two overall imatinib-versus-nilotinib primary comparisons, each with a two-sided 95% confidence interval and P < 0.0001. The two nilotinib arms have a reported difference of −1.6 with a 95% CI of −9.8–6.6 and P = 0.6987.

Clinical interpretation

The statistical record describes differences in the percentage of participants meeting the MMR definition at specified time points. Clinical interpretation requires considering the endpoint definition, durability, treatment context, safety, uncertainty, and the prespecified analysis framework together.

21. Trial Timeline

July 31, 2007

Study start

The ClinicalTrials.gov record identifies July 31, 2007 as the study start date.

September 2, 2009

Primary completion

The ClinicalTrials.gov record identifies September 2, 2009 as the primary completion date.

Registry results

Statistical results posted

The registry reports 31 posted outcome measures and 11 posted statistical analyses, including 8 primary-endpoint analyses.

22. Statistical Concepts in This Trial

Learn more about the methods used in this trial:

23. Related Statistical Calculators

24. Sources

Continue with the underlying statistical methods

Explore the statistical concepts that appear in ENESTnd, from categorical-data testing and confidence intervals to randomized-trial analysis and interpretation of p-values.

25. Record Summary

ENESTnd provides a clear example of statistical analysis for a randomized three-arm trial with a binary molecular-response endpoint. The ClinicalTrials.gov record identifies 846 randomized patients, two registered primary MMR endpoints, an intention-to-treat full analysis set, Cochran-Mantel-Haenszel testing for the principal overall comparisons, two-sided 95% confidence intervals, imputation for the primary endpoints, Sokal risk-group analyses, durable MMR at 24 months, and a direct comparison of the two nilotinib regimens.

The principal overall estimates are 22.1 and 20.4 for the two imatinib-versus-nilotinib comparisons, each with a 95% confidence interval excluding zero and P < 0.0001. The two-nilotinib comparison is −1.6 with a 95% CI of −9.8–6.6 and P = 0.6987. The Sokal analyses provide additional estimates but no reported formal interaction test, while the ClinicalTrials.gov record does not specify the multiplicity strategy or imputation algorithm.

Clinical Biostats methodology: The most informative reading of a trial result combines the randomized analysis population, endpoint definition, effect measure, confidence interval, p-value, missing-data strategy, subgroup structure, multiplicity considerations, and safety population. A statistical result is strongest when each of those components is kept explicit rather than compressed into a single significance statement.