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
endTB-Q was a randomized, parallel, open-label phase 3 treatment trial enrolling 323 participants with multidrug-resistant tuberculosis and related tuberculosis conditions. The registry identifies two treatment arms and a primary efficacy endpoint based on the proportion of participants with a favorable outcome at Week 73 after randomization.
| Feature | endTB-Q |
|---|---|
| Trial acronym | endTB-Q |
| ClinicalTrials.gov identifier | NCT03896685 |
| Phase | Phase 3 |
| Status | Completed |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 323 |
| Number of arms | 2 |
| Start date | 2020-04-06 |
| Primary completion date | 2024-10-04 |
| Lead sponsor | Médecins Sans Frontières, France |
2. Clinical Question
The trial evaluates treatment regimens for multidrug-resistant tuberculosis, with the registry title specifically describing multidrug-resistant TB with fluoroquinolone resistance. The primary statistical question is whether the randomized treatment groups differ in the proportion of participants with a favorable outcome at Week 73 after randomization.
Population
The registered conditions include tuberculosis, multidrug-resistant tuberculosis, pulmonary tuberculosis, mycobacterial infections, bacterial infections, and gram-positive bacterial infections.
Intervention
Bedaquiline 100 MG, delamanid 50 MG oral tablet, clofazimine 100 MG oral capsule, and linezolid 600Mg Tab.
Comparator
Control arm MDR-TB regimen, designed according to latest WHO guidelines.
Primary question
What proportion of participants in each randomized treatment group has a favorable outcome at Week 73 after randomization?
3. Trial Design
Participants were assigned to one of two parallel treatment arms.
The registered design is open-label rather than blinded.
The primary purpose recorded in the registry is treatment.
The registry records 323 participants for the study.
Newly approved drug combination
- Bedaquiline 100 MG
- Delamanid 50 MG oral tablet
- Clofazimine 100 MG oral capsule
- Linezolid 600Mg Tab
MDR-TB regimen
- Control arm MDR-TB regimen
- Regimen designed according to latest WHO guidelines
The statistical structure is therefore a straightforward two-group randomized comparison at a prespecified follow-up time. The absence of masking is an important design feature because the primary endpoint incorporates bacteriological, radiological, and clinical evolution in addition to culture results.
4. Endpoints
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Week 73 Efficacy | Proportion of participants with favorable outcome at Week 73 | Week 73 after randomization |
Definition of a favorable outcome
The registry defines a participant's outcome as favorable at Week 73 if the outcome is not classified as unfavorable, and one of the specified favorable conditions is met.
| Favorable-outcome pathway | Registry definition |
|---|---|
| Pathway 1 | The last two culture results are negative. These two cultures must be taken from sputum samples collected on separate visits, with the latest between Week 65 and Week 73. |
| Pathway 2 | The last culture result from a sputum sample collected between Week 65 and Week 73 is negative, and either there is no other post-baseline culture result or the penultimate culture result is positive due to laboratory cross contamination; additionally, bacteriological, radiological and clinical evolution is favorable. |
| Pathway 3 | There is no culture result from a sputum sample collected between Week 65 and Week 73, or the result of that culture is positive due to laboratory cross contamination; the most recent culture result is negative; and bacteriological, radiological and clinical evolution is favorable. |
This endpoint is more informative than a simple culture-positive versus culture-negative comparison because the favorable classification integrates culture information with bacteriological, radiological, and clinical evolution. The statistical analysis therefore concerns a composite clinical classification, even though several of its components are microbiological.
5. Statistical Methodology
The registry identifies one primary endpoint: the proportion of participants with favorable outcome at Week 73 after randomization. No posted statistical analyses are listed in the ClinicalTrials.gov record.
Binary endpoint structure
The primary endpoint is binary at the participant level: at Week 73, each participant's outcome is classified as favorable or not favorable according to the registered definition. The primary descriptive quantity is therefore a proportion within each randomized treatment group.
A between-group analysis would compare the probability of a favorable Week 73 outcome under the two randomized treatment assignments.
Measures of treatment effect
For a binary endpoint such as this one, several effect measures can be informative. The risk difference expresses the absolute difference in favorable-outcome proportions, while the risk ratio compares the proportions multiplicatively. An odds ratio compares the odds of a favorable outcome between groups.
| Measure | Concept | Interpretation |
|---|---|---|
| Risk difference | p1 − p0 | Absolute difference in the probability of a favorable Week 73 outcome. |
| Risk ratio | p1 / p0 | Relative probability of a favorable outcome in the intervention arm versus the control arm. |
| Odds ratio | [p1/(1−p1)] / [p0/(1−p0)] | Relative odds of a favorable outcome between the randomized groups. |
The choice among these measures depends on the prespecified statistical analysis plan. The registry information available here identifies the endpoint but does not specify which effect measure, confidence-interval method, or hypothesis test was used.
Confidence intervals
A confidence interval around a treatment-effect estimate quantifies statistical uncertainty associated with the estimated difference, ratio, or odds ratio. For a binary endpoint, the interval should be constructed using a method appropriate to the endpoint and analysis population.
A point estimate alone describes the observed treatment-group contrast but does not show how precisely that contrast has been estimated. A confidence interval adds that information. Narrower intervals generally indicate greater statistical precision, while wider intervals indicate greater uncertainty.
Hypothesis testing
A conventional two-group analysis of a binary endpoint would test a null hypothesis corresponding to no difference between treatment groups, depending on the prespecified effect measure. A p-value from such a test would address compatibility with the null hypothesis; it would not itself describe the magnitude or clinical importance of the treatment effect.
6. Planned Analysis
The registry says the primary efficacy outcome will be measured as the proportion of participants with favorable outcome at Week 73 after randomization. Because this is a binary endpoint, a typical analysis would summarize the favorable-outcome proportion in each randomized arm and estimate a between-arm treatment effect with a corresponding confidence interval.
Primary descriptive analysis
Calculate the proportion of participants classified as having a favorable outcome at Week 73 in each treatment group.
Between-group comparison
Compare the two proportions using a prespecified binary-outcome method and report an appropriate treatment-effect estimate with uncertainty.
Analysis population
The choice of participants included in the primary analysis should follow the prespecified statistical analysis plan, particularly for participants without a Week 73 culture result or complete clinical information.
Endpoint timing
The primary assessment is fixed at Week 73 after randomization, with the favorable-outcome definition incorporating information from specified culture windows extending through Week 73.
7. Statistical Methods Explained
What makes the Week 73 endpoint a binary outcome?
Although the favorable-outcome definition contains several clinical and microbiological components, the final Week 73 classification has two possible states for statistical analysis: favorable or not favorable. That converts the endpoint into a binary participant-level outcome.
Why use a proportion for the primary endpoint?
The endpoint asks for the proportion of participants who meet the favorable-outcome definition at a specific time point. The proportion therefore directly describes the frequency of the desired outcome within each treatment arm.
Why might risk difference be useful?
Risk difference is directly interpretable on the probability scale. If two treatment groups had favorable-outcome proportions of p1 and p0, the risk difference would be p1 − p0. It answers how much the observed probability differs in absolute terms.
Why might risk ratio be useful?
Risk ratio places the two probabilities on a relative scale. A value above 1 would indicate a higher probability of a favorable outcome in the intervention group, while a value below 1 would indicate a lower probability, assuming the intervention group is placed in the numerator.
Why does the endpoint definition matter statistically?
The favorable classification is not determined solely by one culture result. It depends on specified culture timing, the possibility of laboratory cross contamination, and favorable bacteriological, radiological, and clinical evolution. Changing any of those classification rules could change the resulting proportion, so the endpoint definition is part of the statistical estimand rather than merely a clinical footnote.
Why is missing Week 73 information important?
A participant without sufficient information to establish the registered favorable-outcome criteria creates an analysis-population question. Whether such participants are classified as unfavorable, excluded from a particular analysis, or handled through another prespecified strategy can affect the estimated treatment effect. The appropriate rule should come from the statistical analysis plan rather than being chosen after observing the results.
Why is randomization important for this comparison?
Randomization creates the basis for comparing outcomes between treatment assignments without relying on observational adjustment for every possible prognostic factor. It does not guarantee identical groups in every characteristic, but it provides the core design framework for causal comparison between the randomized strategies.
8. Interpretation of the Primary Endpoint
The primary endpoint is the probability that a randomized participant meets the registered definition of a favorable outcome at Week 73. It is therefore a fixed-time treatment outcome, rather than a time-to-event endpoint such as overall survival.
A positive between-arm difference in favorable-outcome proportions would indicate that the intervention group had a higher observed probability of satisfying the Week 73 favorable-outcome definition. It would not mean that every component of the composite definition improved by the same amount.
If a formal hypothesis test were reported, its p-value would quantify evidence against the specified null hypothesis under the statistical model and testing procedure. It would not measure the size of the treatment effect, the probability that the treatment works, or the probability that the null hypothesis is true.
A confidence interval around a treatment-effect estimate would show the statistical precision of the estimated difference or ratio. It should be interpreted together with the point estimate and the clinical meaning of the Week 73 endpoint rather than as a substitute for either.
9. Endpoint Timing and Censoring Considerations
The primary endpoint differs from a conventional time-to-event endpoint because the registered outcome is evaluated at a specified follow-up point: Week 73 after randomization. The definition nevertheless uses longitudinal information, including cultures collected between Week 65 and Week 73 and the most recent culture result in specified circumstances.
| Timing component | Statistical relevance |
|---|---|
| Randomization | Defines the starting point for the primary endpoint's time frame. |
| Week 65 to Week 73 | Defines the principal late culture window used in two of the favorable-outcome pathways. |
| Week 73 | Defines the primary efficacy assessment point. |
| Separate sputum visits | The first favorable pathway requires the last two negative cultures to come from sputum samples collected on separate visits. |
| Laboratory cross contamination | The favorable definition explicitly addresses culture results classified as positive due to laboratory cross contamination. |
This structure illustrates why the endpoint cannot be reduced to a single culture result at one visit. The statistical classification depends on a sequence of observations and prespecified clinical and microbiological conditions.
10. Open-Label Design and Endpoint Assessment
The registry records no masking. In an open-label randomized trial, participants and study personnel know the assigned treatment. That design feature is particularly relevant when an endpoint incorporates radiological and clinical evolution, because some components of the classification may involve assessment rather than being determined solely by an automated laboratory measurement.
Strength of randomization
Random allocation remains the central mechanism for comparing treatment strategies and helps establish a causal framework for the randomized comparison.
Assessment considerations
Because masking is absent and the favorable endpoint incorporates clinical and radiological evolution, assessment procedures are relevant to interpretation.
Microbiological component
Culture results provide objective components of the endpoint, including explicit rules for negative cultures and laboratory cross contamination.
Composite classification
The final favorable classification integrates several conditions rather than representing a single isolated measurement.
11. Missing Data and Endpoint Classification
The registered definition explicitly addresses situations in which there is no culture result from a sputum sample collected between Week 65 and Week 73. It also specifies circumstances in which a culture result is considered positive because of laboratory cross contamination.
These provisions are statistically important because missingness and measurement-quality issues can otherwise create ambiguity in a binary endpoint. By defining specific circumstances under which a favorable outcome can still be established, the endpoint attempts to specify how available longitudinal information contributes to the Week 73 classification.
12. Limitations
- No posted statistical results: the ClinicalTrials.gov record described here does not contain a posted statistical analysis for the primary endpoint, so the registry information does not provide an arm-specific favorable-outcome estimate, confidence interval, or p-value.
- Open-label design: the trial is registered with no masking. This is relevant because the favorable-outcome definition includes bacteriological, radiological, and clinical evolution.
- Composite classification: the primary endpoint incorporates several conditions, so an overall favorable-outcome proportion does not by itself reveal the contribution of each component.
- Fixed-time endpoint: Week 73 provides a prespecified snapshot of treatment outcome. It does not describe the complete trajectory of disease response outside the registered endpoint definition.
- Culture timing: the endpoint relies on specified culture windows and separate visits in some pathways, making timing and completeness of microbiological assessment important to classification.
- Laboratory cross contamination: the endpoint explicitly accommodates culture results attributed to laboratory cross contamination, highlighting the importance of measurement quality in the outcome definition.
- Unreported analysis details: the available registry information does not specify the statistical test, treatment-effect measure, confidence-interval method, missing-data strategy, or multiplicity procedure for the primary endpoint.
- Generalizability: the registered population and treatment regimens define the population and intervention strategies represented by the randomized trial.
13. Why This Trial Matters Statistically
endTB-Q is a useful statistical teaching case because its primary endpoint illustrates a common but easily misunderstood type of clinical-trial outcome: a fixed-time binary classification built from multiple longitudinal clinical and microbiological observations.
| Concept | How it appears in endTB-Q |
|---|---|
| Randomization | Participants were randomized to one of two parallel treatment arms. |
| Binary endpoint | The primary outcome classifies participants as having a favorable outcome or not at Week 73. |
| Fixed follow-up time | The primary assessment occurs at Week 73 after randomization. |
| Composite clinical definition | Favorable status can depend on culture results together with bacteriological, radiological and clinical evolution. |
| Longitudinal information | The endpoint incorporates culture results from defined periods and, in one pathway, requires negative cultures from separate visits. |
| Measurement quality | The endpoint explicitly addresses positive culture results attributed to laboratory cross contamination. |
| Open-label design | The registry records no masking. |
| Treatment comparison | The intervention combines bedaquiline, delamanid, clofazimine and linezolid; the control is an MDR-TB regimen designed according to latest WHO guidelines. |
14. Statistical Questions for Readers
What is the estimand?
The primary outcome concerns the proportion of randomized participants who meet the favorable-outcome definition at Week 73 after randomization.
What is being compared?
The randomized treatment strategies are compared through their Week 73 favorable-outcome proportions.
What is the key scale?
The endpoint is naturally expressed as a proportion, with risk difference and risk ratio among the possible measures of treatment effect.
Where can bias enter?
The absence of masking is particularly relevant because clinical and radiological evolution contributes to the endpoint classification.
15. Trial Timeline
Trial start
The registered study start date is April 6, 2020.
Randomized treatment comparison
The study is registered as a phase 3, randomized, parallel, open-label treatment trial with two arms and an enrollment of 323 participants.
Primary efficacy assessment
The primary endpoint is the proportion of participants with favorable outcome at Week 73 after randomization.
Primary completion
The registered primary completion date is October 4, 2024.
16. Sources
- ClinicalTrials.gov: NCT03896685 — endTB-Q.
- PubMed: PubMed record 40683298.
- PubMed: PubMed record 38037119.
Continue through the Clinical Biostats knowledge graph
Clinical-trial endpoints become easier to interpret when their definitions are connected to the underlying statistical concepts: proportions, treatment effects, confidence intervals, randomization, and analysis populations.
17. Record Summary
endTB-Q is a randomized phase 3, parallel, open-label treatment trial with 323 participants and two treatment arms. Its primary endpoint is the proportion of participants with favorable outcome at Week 73 after randomization, with a detailed definition based on culture results, culture timing, laboratory cross contamination, and favorable bacteriological, radiological and clinical evolution.
From a statistical perspective, the endpoint is a fixed-time binary outcome embedded within a longitudinal clinical assessment. The most natural summary is therefore an arm-specific favorable-outcome proportion followed by a prespecified between-group treatment-effect measure and confidence interval. The absence of posted statistical analyses in the ClinicalTrials.gov record means that the registry identifies the outcome and its classification rules but does not provide the corresponding numerical treatment effect or formal hypothesis test.