Trial to Evaluate the Efficacy, Safety and Tolerability of BPaMZ in Drug-Sensitive (DS-TB) Adult Patients and Drug-Resistant (DR-TB) Adult Patients: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Trial to Evaluate the Efficacy, Safety and Tolerability of BPaMZ in Drug-Sensitive (DS-TB) Adult Patients and Drug-Resistant (DR-TB) Adult Patients (NCT03338621), a randomized phase 2, phase 3 study. The primary outcome is Number of Participants With Culture Negative Status by 8 Weeks.

2. Clinical Question

3. Trial Design

NCT03338621 is a completed, open-label, parallel-group trial identified as Phase 2 and Phase 3. The registry lists randomized allocation, while the official title calls the study partially randomized. Masking is listed as none. Enrollment was 455. The registry names the Global Alliance for TB Drug Development as sponsor.

The trial involved adults with drug-sensitive (DS-TB) and drug-resistant (DR-TB) smear-positive pulmonary tuberculosis. In the DS-TB comparison, participants received either BPaMZ for 17 weeks or standard HRZE/HR for 26 weeks. The drug-resistant group received BPaMZ for 26 weeks. The primary endpoint reported in the registry is confined to the DS-TB groups; it should not be read as a comparison involving the DR-TB group.

The primary endpoint timeframe was Days 0-56 (8 weeks) from treatment start. The registry describes additional follow-up outcomes at 52 weeks and 104 weeks in its trial summary. Listed secondary outcomes include unfavorable bacteriologic or clinical outcomes at 52 weeks and time to culture-negative status. The record lists three outcome measures and one analysis, though it does not report the full analysis plan for each measure.

The registry records a start date of 2018-07-30, a primary completion date of 2021-07-01, and completed status. These design details establish the scope of the comparison, but the record does not report the trial’s original sample-size justification or provide enough information to reconstruct every operational detail of randomization and analysis.

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.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Limited to Drug Sensitive TB groups. Participants were considered unassessable and therefore excluded from the Modified Intent to Treat (MITT) population if they were late exclusion or if they did not have culture confirmation of M.tb at baseline.

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 Number of Participants With Culture Negative Status by 8 Weeks, measured over Days 0-56 (8 weeks).

The primary endpoint counts participants with culture-negative status by 8 weeks, within Days 0-56. The registry defines this status as at least two negative culture results at separate visits, with those visits at least 7 days apart and no intervening positive culture result for M.tb. This is a count endpoint: it records whether participants met the stated culture criterion during the specified period.

Culture results are used in tuberculosis trials as a microbiologic measure of response. In general, requiring repeated negative results can reduce reliance on a single culture observation, while the no-intervening-positive rule makes the endpoint depend on a sustained pattern across visits. These are general features of such endpoint definitions; the registry does not explain why this particular endpoint or its timing was selected.

The endpoint is limited to the drug-sensitive treatment groups. The registry says late exclusions and participants without culture confirmation of M.tb at baseline were considered unassessable and excluded from the modified intention-to-treat population. The record does not fully describe how missing follow-up cultures were handled in the posted outcome counts. A culture-negative status at 8 weeks is an early microbiologic outcome; by itself, it does not describe later relapse, clinical failure, or the overall experience of every participant.

7. Results Posted in the Registry

• Drug Sensitive-TB 4BPaMZ: 84.1% (122/145) versus 47.3% (70/148) on control; risk difference 36.8 percentage points (95% CI 26.8 to 46.8), risk ratio 1.78 (1.48–2.14); unadjusted two-proportion z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

What does the binary endpoint comparison estimate?

The primary result is a binary outcome: each participant either met the culture-negative criterion by 8 weeks or did not. A count and denominator describe the observed endpoint frequency in each group. The posted counts are 122 of 145 for BPaMZ and 70 of 148 for standard treatment. The unadjusted recalculation compares these group summaries without adjustment for other participant characteristics.

A risk difference expresses the difference between the two groups’ outcome proportions. It keeps the comparison on an absolute scale, which can be useful when considering how far apart the observed group results are. A risk ratio expresses one group’s outcome proportion relative to the other’s. An odds ratio compares odds rather than proportions; when an outcome is common, it can look more extreme than a risk ratio. These are general explanations of the measures, not additional trial findings.

The recalculated risk-ratio estimate is 1.779 (95% confidence interval 1.48 to 2.139), and the recalculated odds-ratio estimate is 5.911 (95% confidence interval 3.41 to 10.25). The unadjusted risk-difference confidence interval lies above zero. The p-value for the unadjusted comparison is <0.001. A small p-value describes how incompatible the observed data are with a specified null model; it is not the probability that the null hypothesis is true, nor a measure of clinical importance.

What does a time-to-event method add?

A time-to-event method uses information about when an event occurs, rather than reducing the outcome to whether it occurred by one fixed cutoff. The registry reports a Log Rank analysis with a Cox Proportional Hazard method. In general, the Log Rank test compares event-time patterns between groups, while a Cox model estimates a relative hazard over follow-up under its modeling assumptions. A hazard ratio is not the same quantity as a risk ratio or an odds ratio.

The reported Cox value is 2.93, with a 95% confidence interval from 2.17 to 3.96 and p <0.001. Although the analysis is labeled as superiority, the registry entry does not make the endpoint, event coding, or reference-group direction explicit. The number therefore cannot be safely paraphrased as a particular group having faster culture conversion. The record also does not report checks of proportional-hazards assumptions or explain how censoring was handled.

What assumptions matter?

For an unadjusted two-group binary comparison, interpretation assumes that the group definitions and endpoint assessments are comparable and that the reported denominators correspond to the intended analysis populations. Randomized allocation can support a causal comparison when implementation, follow-up, and analysis preserve the randomized groups, but the registry record alone does not establish every one of those conditions. The trial was open-label, and the record does not describe whether culture assessment was blinded.

Missing outcome data can affect a binary result if participants without adequate cultures are excluded or classified differently from participants with complete assessments. The endpoint description identifies some participants as unassessable for the modified intention-to-treat population, but the handling of other missing cultures is not reported in the registry record. The posted counts should therefore be interpreted within the reported denominators rather than assumed to represent all enrolled participants.

The planning illustrations are based on a normal approximation, equal allocation, and a two-sided significance level of 0.05, using stated fractions of the observed effect. General sample-size calculations depend on assumptions about the effect to detect and the outcome distribution. These illustrations are not evidence about the trial’s original design decisions, which are not reported in the registry record.

10. Confidence Intervals and Precision

A confidence interval describes a range of values compatible with the data and the statistical method, under the method’s assumptions. It conveys precision as well as direction: a narrower interval indicates less sampling uncertainty on that measure, while a wider interval leaves more uncertainty about the size of the effect. It is not a range containing a fixed percentage of individual participant outcomes.

For the recalculated risk ratio, the estimate is 1.779 and the 95% confidence interval is 1.48 to 2.139. For the recalculated odds ratio, the estimate is 5.911 and the interval is 3.41 to 10.25. The recalculated risk-difference interval is entirely above zero. These intervals indicate that the observed comparison is not centered near the no-difference value on the respective scales, while still allowing a range of effect sizes.

The registry’s reported time-to-event estimate is 2.93, with a 95% confidence interval from 2.17 to 3.96. Its interpretation is limited because the registry analysis entry does not state event coding or reference direction. The interval’s numerical precision does not resolve that ambiguity.

Confidence intervals account for sampling variability under a statistical model, not every source of uncertainty. They do not correct for missing-data problems, measurement differences, unreported analysis choices, or limitations in applying the result beyond the studied groups. The record does not report whether any adjustment was made for multiple outcomes or analyses; that uncertainty is separate from the width of an individual interval.

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 sample-size illustrations relate the observed effect to hypothetical future planning under stated assumptions. The binary-outcome figures are participants per arm, not total events. The planning rows consider 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect. Under the stated assumptions, the listed participants per arm for 80% power are 25, 48, and 113, respectively; for 90% power they are 33, 64, and 151, respectively.

The assumptions 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. For time-to-event outcomes, the planning convention is total number of events using Schoenfeld; for all other outcomes, it is participants per arm. The primary endpoint is binary, so its figures are participants per arm.

In plain terms, 80% power means that, if the assumed effect and calculation assumptions were true, the design would have an 80% chance of detecting an effect at the specified significance level. The 90% power column uses a higher target chance of detection and therefore requires more participants per arm in each listed scenario. Smaller effects need more participants because they are harder to distinguish from random variation.

These are planning illustrations, not observed power. The trial’s own original sample-size assumptions and justification are not reported in the registry record, so the illustrations should not be read as a reconstruction of its design.

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 = 0.473; treatment proportion p₁ = 0.841; average p̄ = 0.657.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6713 = 1.3156.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6187 = 0.5207.
4. n = (1.3156 + 0.5207)² / (0.3684)² = 24.8, rounded up to 25 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 <- 0.472972972972973                          # control proportion
p1 <- 0.841379310344828                          # 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(3338621)
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 = 25 (80% power), 33 (90% power)
 75% of observed effect: n per arm = 48 (80% power), 64 (90% power)
 50% of observed effect: n per arm = 113 (80% power), 151 (90% power)
power.prop.test: 24.8 
Simulated power at n = 25 per arm: 0.825 (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. Analyses Reported by the Sponsor

The registry’s own reported analysis is a Log Rank analysis using a Cox Proportional Hazard method. It reports a value of 2.93, a 95% confidence interval from 2.17 to 3.96, and p <0.001, and labels the analysis as superiority. The entry does not specify the event coding or reference direction, so the direction of that time-to-event contrast is not clear from the analysis entry alone.

The recalculated results are simple, unadjusted comparisons of the posted binary endpoint summaries. They include a risk ratio of 1.779 (95% confidence interval 1.48 to 2.139), an odds ratio of 5.911 (95% confidence interval 3.41 to 10.25), and a risk-difference interval entirely above zero. These quantities are not substitutes for the sponsor’s time-to-event analysis: they use a different outcome summary and statistical scale.

The registry does not identify the endpoint associated with the Cox analysis item in enough detail to establish a direct numerical comparison with the 8-week count endpoint. The registry record reports no discrepancy between the posted summaries and the recalculations. Both sets of results should therefore be described according to their own methods, without treating their estimates as if they measured the same parameter.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary endpoint, two secondary outcomes, and one reported analysis. The secondary outcomes concern unfavorable bacteriologic or clinical outcomes at 52 weeks and time to culture-negative status. The record does not report a multiplicity adjustment, a hierarchy for testing outcomes, or whether the listed outcomes were intended to support confirmatory conclusions. When multiple outcomes or analyses are examined, the chance of at least one apparently unusual result can increase; adjustment or a prespecified testing strategy can address that issue. This is a general statistical concern, not evidence that a particular adjustment was required or omitted from the full trial conduct.

The record describes the primary-endpoint population as limited to the DS-TB groups. It says participants were considered unassessable and excluded from the modified intention-to-treat population if they were late exclusions or lacked culture confirmation of M.tb at baseline. The posted endpoint denominators are 145 for BPaMZ and 148 for standard treatment. The registry does not provide a full participant flow showing how these denominators relate to every enrolled participant or explain how all missing or unavailable follow-up cultures were handled.

Missing data matter because culture status cannot be determined by the endpoint rule without sufficient results at the required visits. Excluding participants without assessable cultures can alter the composition of the analyzed groups if the reasons for missingness are related to outcome or treatment. The record does not report sensitivity analyses for missing endpoint data, imputation procedures, or a detailed definition of analysis populations beyond the stated modified intention-to-treat exclusions.

The trial is described as randomized, open-label, and partially randomized in its official title. The record does not report the allocation sequence, concealment procedures, stratification factors, or whether outcome assessors were blinded. It also does not report a detailed statistical analysis plan, interim analyses, or the trial’s original sample-size justification. These omissions limit independent assessment of prespecification and analysis choices, but they do not by themselves establish that such procedures were absent.

15. 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: hepatotoxicity (5.3%), transaminases increased (2.0%), hepatic enzyme increased (1.3%), tuberculosis (1.3%), dyspnoea (1.3%).

The registry reports serious adverse events and other adverse events separately for three groups. In the drug-sensitive BPaMZ group, 17 of 150 participants at risk had a serious adverse event and 139 of 150 had other adverse events; 3 deaths are listed. In the drug-sensitive standard-treatment group, 7 of 153 had a serious adverse event and 144 of 153 had other adverse events; 2 deaths are listed. In the drug-resistant BPaMZ group, 16 of 149 had a serious adverse event and 142 of 149 had other adverse events; 4 deaths are listed.

The registry lists hepatotoxicity at 5.3% among the top serious-event terms, followed by transaminases increased at 2.0%, hepatic enzyme increased at 1.3%, tuberculosis at 1.3%, dyspnoea at 1.3%, and anaemia at 0.7%. These term percentages describe the registry’s listed serious-event terms; the record does not give the full event-level detail needed to interpret overlap or duration.

These safety counts are descriptive. The drug-sensitive groups are relevant to the primary comparison, while the drug-resistant group is a separate treatment group and should not be combined with them as if it were a randomized comparator. Event counts alone do not establish that treatment caused an event, and the registry entry does not provide adjusted comparative safety analyses or detailed severity and timing information.

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

The posted primary results show that more participants in the DS-TB BPaMZ group met the defined culture-negative criterion by 8 weeks than in the standard-treatment group: 122 of 145 compared with 70 of 148. The recalculated unadjusted confidence interval for the risk difference lies above zero, and the p-value is <0.001. Together, these summaries describe a clear difference in the recorded endpoint between these groups in this trial.

That finding concerns an early microbiologic outcome with a specific definition. It does not by itself establish what happened to participants at 52 weeks or 104 weeks, whether participants later experienced relapse, or whether clinical outcomes differed. The registry lists later outcomes, but the results needed to assess those questions are not reported in the registry record.

The result also does not directly answer the question for drug-resistant tuberculosis. The primary endpoint is limited to the DS-TB groups, and the DR-TB group received BPaMZ without the standard-treatment comparator described for DS-TB. Nor does the endpoint alone provide a complete balance of benefits and harms; safety counts must be considered separately and have their own reporting limitations.

Statistical significance and precision do not remove uncertainty about the analysis population, missing cultures, event coding in the reported time-to-event analysis, or the generalizability of the comparison. The record identifies the study as open-label and does not provide all details needed to evaluate implementation and analysis choices. The appropriate interpretation is therefore specific: the posted DS-TB counts differ on the 8-week culture endpoint, while broader conclusions require additional outcome and methods information.

17. Why This Trial Matters Statistically

This trial illustrates why clinical results should be read on the scale of the outcome that was actually recorded. A count endpoint, a risk difference, a risk ratio, an odds ratio, and a hazard ratio each summarize data differently. Their numerical values cannot be exchanged without regard to their definitions. The registry’s binary endpoint counts and its Cox analysis therefore need to be interpreted as distinct statistical summaries.

The endpoint definition also matters. Culture-negative status required repeated negative results separated by at least 7 days and no intervening positive result. This makes the outcome more specific than a single negative culture, but it also means that assessment depends on participants having sufficient culture observations. The stated exclusions and unreported details of missing follow-up cultures are relevant to understanding who contributed to the reported comparison.

The trial brings together DS-TB and DR-TB treatment groups, but the primary endpoint comparison is only between the DS-TB regimens. Keeping those populations separate prevents the mistaken impression that the 8-week result directly compares treatment strategies in drug-resistant disease. The reported adverse-event groups likewise include a separate DR-TB group.

Finally, the coexistence of a posted endpoint count, a registry-reported time-to-event analysis, and planning illustrations highlights three different purposes of statistics: describing observed outcomes, estimating uncertainty, and projecting requirements for a hypothetical future study. The illustrations are not evidence about observed power or the trial’s original design. Reading each result according to its purpose avoids making broader claims than the record supports.

18. 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.

The primary endpoint analysis is confined to DS-TB groups and does not establish a comparative result for DR-TB. The registry describes the modified intention-to-treat eligibility rule but does not provide a complete participant flow or detailed handling of all missing cultures. Since the endpoint requires multiple culture observations, the availability of those observations is particularly relevant.

The trial was open-label, and the registry record does not report whether laboratory or outcome assessors were blinded. It also does not detail allocation concealment, stratification, or the exact procedures behind the partially randomized description. These reporting gaps make it difficult to assess potential sources of bias from the registry record alone.

The reported Cox analysis lacks explicit event coding, reference direction, and a clear link to a particular endpoint in its analysis entry. Its numerical result should not be assigned a direction beyond what the registry states. The record also does not report proportional-hazards checks or censoring details.

Only one analysis is listed despite multiple outcome measures. A multiplicity strategy, prespecified hierarchy, and complete analysis plan are not reported in the registry record. The trial’s original sample-size justification and assumptions are also not reported; the available planning illustrations are based on the observed effect and are not the original design rationale.

Finally, the posted safety summaries do not provide detailed event timing, duration, overlap, or adjusted comparisons. The listed counts and terms are useful descriptive information but cannot alone establish attribution or provide a complete comparison of safety across regimens.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Trial to Evaluate the Efficacy, Safety and Tolerability of BPaMZ in Drug-Sensitive (DS-TB) Adult Patients and Drug-Resistant (DR-TB) Adult Patients — NCT03338621.
• Registry record last updated 2026-08-27; status last verified 2024-03.
• Results first posted 2023-11-08.
• 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.

22. Record Summary

NCT03338621 is a completed, open-label, parallel-group Phase 2 and Phase 3 trial with randomized allocation and a registry enrollment of 455. It studied BPaMZ in DS-TB and DR-TB adult participants, with standard HRZE/HR as the comparator for the DS-TB group. The primary endpoint was culture-negative status by 8 weeks in the DS-TB groups.

The posted endpoint counts were 122 of 145 for BPaMZ and 70 of 148 for standard treatment. The unadjusted recalculations show a risk-difference interval above zero, with risk ratio 1.779 and odds ratio 5.911; the reported unadjusted p-value is <0.001. The registry also reports a Cox/Log Rank result but does not specify its event coding or reference direction. Important details on missing data, multiplicity, and the original sample-size rationale are not reported in the registry record.