A Study of Two Different Doses of Cabozantinib (XL184) in Progressive, Metastatic Medullary Thyroid Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Two Different Doses of Cabozantinib (XL184) in Progressive, Metastatic Medullary Thyroid Cancer (NCT01896479), a randomized phase 4 study. The primary outcome is Progression Free Survival (PFS) Per Blinded Independent Radiology Committee (BIRC) Per Response Evaluation Criteria in Solid Tumors Version 1.1 (RECIST 1.1).

2. Clinical Question

3. Trial Design

This was a randomized, quadruple-masked, parallel-group Phase 4 study of cabozantinib in progressive, metastatic medullary thyroid cancer. Its official title describes a comparison of cabozantinib at 60 mg/day with 140 mg/day. The registry record lists Exelixis as sponsor and an enrollment of 247. The reported primary outcome groups include 123 participants in the 60 mg arm and 124 in the 140 mg arm.

Participants were assigned to one of two active-dose regimens. The 60 mg group received cabozantinib capsules and placebo tablets; the 140 mg group received cabozantinib tablets and placebo capsules. Both regimens were described as oral and once daily. Placebo forms were included alongside the active formulations, and the registry lists masking as quadruple. The exact individuals or roles masked, and the procedures used to maintain masking, are not reported in the registry record.

The primary endpoint was progression-free survival assessed by a Blinded Independent Radiology Committee under RECIST 1.1. The registry record describes measurement from randomization until documented disease progression or death from any cause, whichever occurred first. Assessment was specified for up to 31 months. The analysis was event-driven: the primary analysis was triggered by at least 150 events in the Intent-to-Treat population, and the cutoff for that analysis occurred when 155 events were reported.

The record gives a median follow-up of 30.2 months through the primary data cutoff on 15 July 2020. It also lists objective response rate as a secondary outcome. The registry status is active, not recruiting. Further details about eligibility criteria, treatment changes, adherence, and the original sample-size rationale are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: Cabozantinib-matched placebo capsules and tablets were administered along with the corresponding tablets and capsules to maintain the blinding of the study treatment. Per prespecified analysis, all participants who died during the study met the study requirements of study completion. Therefore, none of the deaths that occurred during the study were recorded as leading to study discontinuation.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Measured in the Intent-to-Treat (ITT) Population, which included all randomized participants.

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: Measured in the Safety Population, which included participants who received any amount of study drug.

6. Primary Endpoint

The registry lists the primary outcome as Progression Free Survival (PFS) Per Blinded Independent Radiology Committee (BIRC) Per Response Evaluation Criteria in Solid Tumors Version 1.1 (RECIST 1.1), measured over Median time of follow-up (from the date of randomization of the first participant through primary data cut off [15 July 2020]) was 30.2 months.

Progression-free survival measures the time from randomization to the first documented disease progression or death from any cause, whichever occurs first. In this trial, progression was assessed by a Blinded Independent Radiology Committee using RECIST 1.1. The registry record says the endpoint was assessed for up to 31 months and analyzed in the Intent-to-Treat population, which included all randomized participants.

This endpoint combines disease progression and death into a time-to-event outcome. A participant's time contributes information until an event is recorded or the observation is otherwise censored. The registry record does not describe censoring rules in detail. A Kaplan-Meier estimate is used to describe the probability of remaining free of the endpoint over time, and the median is the time at which that estimated probability reaches one half. This is a general explanation of the method.

Progression-free survival is useful in a trial that compares cancer treatments because it captures the timing of disease progression as well as death, rather than counting only whether an event occurred by a single fixed visit. Independent radiology review can help separate outcome assessment from treatment assignment, although the registry record does not provide operational details of that review. The endpoint does not, by itself, describe overall survival, symptoms, or quality of life; those outcomes are not reported as results in the primary outcome data.

7. Results Posted in the Registry

• Cabozantinib (XL184) 60 mg vs Cabozantinib (XL184) 140 mg: hazard ratio 1.24 (95% CI 0.90–1.70), reported p = 0.192 (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.

Objective Response Rate (ORR) Per BIRC Per RECIST 1.1

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Cabozantinib (XL184) 140 mg: 33.1% (41/124) versus 33.3% (41/123) on control; risk difference -0.3 percentage points (95% CI -12.0 to 11.5), risk ratio 0.99 (0.70–1.41); unadjusted two-proportion z-test p = 0.964.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Cabozantinib (XL184) 140 mg and the control. The interval gives the range of effects that remain compatible with the data.

Objective response rate is the secondary outcome with results posted in the record. Its group estimates are 33.0% in each arm, with confidence intervals that express uncertainty around those response proportions. The sponsor's reported comparison is an odds ratio of 1.020, with a 95% confidence interval from 0.6 to 1.7 and p = 0.944. The posted unadjusted recalculation uses the response counts to estimate a risk difference, which is a different measure of the contrast. These results should be understood as a secondary outcome and not as a substitute for the primary progression-free survival analysis.

Across the posted endpoints, the response-rate estimates have the same displayed percentage in both groups, while the primary time-to-event estimates and hazard ratio have their own pattern and uncertainty. The endpoints measure different aspects of outcome: objective response records a response classification, whereas progression-free survival concerns time until progression or death. Their results need not move in lockstep, and one cannot be used to confirm or invalidate the other without an appropriate prespecified framework.

Unadjusted secondary results generally carry less confirmatory weight when multiple outcomes or comparisons are considered, because testing many questions can increase the chance of at least one apparently notable result. This is a general statistical concern. The registry record does not report a multiplicity adjustment, a testing hierarchy, or a complete list of secondary outcome analyses. The secondary endpoint should therefore be read as supportive descriptive information, with attention to its confidence interval and analysis method, rather than as an independent definitive conclusion about the primary question.

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

• Objective Response Rate (ORR) Per BIRC Per RECIST 1.1

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is a time-to-event measure. Its sponsor-reported comparison uses a Cox proportional hazards model and a log-rank method. The report gives a hazard ratio, a confidence interval, and a p-value. These quantities describe different aspects of the evidence and should be read together rather than treated as interchangeable.

What does the hazard ratio describe?

In general, a hazard is the instantaneous event rate among participants who have not yet experienced the event, at a particular time. A hazard ratio compares those rates between groups. The reported value of 1.24 is for cabozantinib 60 mg versus cabozantinib 140 mg. Under the usual interpretation, a value of 1 represents equal hazards, while values above or below 1 indicate different relative rates in the direction determined by the comparison order. The registry record labels the result as a Cox proportional hazards analysis. The proportional-hazards model assumes that the relative hazards are sufficiently stable over time for one summary ratio to be meaningful; whether that assumption was assessed is not reported in the registry record.

How does the log-rank test fit in?

The log-rank test compares event experience across the follow-up period, taking account of when events occur and how long participants remain under observation. It is commonly used with time-to-event outcomes. The reported p-value is 0.192. In general, a p-value measures how unusual data at least this discrepant from a specified null model would be if that model were true, under the test's assumptions. It is not the probability that the null hypothesis is true, and a value that does not meet a chosen threshold does not prove the groups are equivalent. The registry record does not give a complete statistical analysis plan or state how the p-value was to be used in a non-inferiority decision.

What do the medians and confidence intervals add?

The Kaplan-Meier method estimates the event-free proportion over time while accounting for censored observations. The median is the estimated point at which the event-free proportion reaches one half. The reported medians are 11.0 months for 60 mg and 13.9 months for 140 mg. Their 95% confidence intervals quantify uncertainty around each median estimate. They are not intervals containing 95% of individual participants' event times, and overlap or non-overlap of separate intervals is not, by itself, the formal test of the between-group comparison.

A confidence interval is a range of estimates compatible with the observed data under the method's assumptions. A wider interval indicates less precision than a narrower interval. For the hazard ratio, the interval is 0.90 to 1.70. It includes 1, so the data do not rule out equal hazards under this interval-based reading. It also includes estimates on either side of 1, showing that the relative effect remains uncertain. The interval should not be turned into a claim of equivalence; equivalence or non-inferiority requires a prespecified margin and an appropriate analysis.

Why is event-driven timing relevant?

For time-to-event outcomes, the number of observed events often carries more direct information about a treatment comparison than enrollment alone. The registry says that the prespecified primary analysis was triggered by at least 150 events in the Intent-to-Treat population and that 155 events were reported at the data cutoff. This explains the event-based trigger described in the record. The exact rules for censoring, handling protocol deviations, and conducting sensitivity analyses are not reported in the registry record.

General statistical considerations also matter when reading a proportional-hazards summary. If treatment effects vary over time, a single hazard ratio can conceal that pattern; the underlying survival curves and a time-varying assessment can provide additional context. Those details are not reported in the registry record details. The listed follow-up is a median of 30.2 months through the cutoff, while the endpoint description says assessments were made for up to 31 months. Median follow-up is not the same as the duration of observation for every participant.

The randomized assignment supports a comparison between the assigned groups, while the Intent-to-Treat definition retains all randomized participants in the primary population. The registry record does not describe missing-data procedures, the extent of missing outcome data, or additional analysis populations. Consequently, the reported summaries can be explained on their stated basis, but details needed to independently evaluate all model assumptions and analysis choices are not reported in the registry record.

11. Confidence Intervals and Precision

Precision is reflected in the width of an estimate's confidence interval. For median progression-free survival, the 60 mg group has a 95% confidence interval of 8.3 to 13.6 months around the estimate of 11.0 months. The 140 mg group has a 95% confidence interval of 9.0 to 16.6 months around the estimate of 13.9 months. These intervals show uncertainty in the estimated medians; they do not describe the range of outcomes for individual participants.

The hazard ratio is 1.24, with a 95% confidence interval of 0.90 to 1.70. The interval includes 1 and covers a range of possible relative hazards in both directions. Thus, the point estimate alone gives an incomplete account: the interval indicates that the treatment comparison is not estimated with enough precision to exclude a range of effects. The log-rank p-value is 0.192, but the p-value does not quantify the size or precision of the effect as directly as the interval does.

The interval for each group's median and the interval for the hazard ratio concern different estimands, so they should not be compared as though they were the same measure. Nor does overlap between the separate median intervals serve as a substitute for the reported between-group test. Interpretation of the hazard-ratio interval also depends on the model and its assumptions. The registry record does not report an assessment of proportional hazards or a non-inferiority margin. The intervals therefore describe statistical uncertainty, but do not alone settle whether a clinically acceptable difference has been excluded.

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 sample-size calculations for the primary time-to-event outcome are expressed as total events, not participants per arm. The figures correspond to planning for 100%, 75%, and 50% of the observed effect, with separate columns for a design targeting 80% power and one targeting 90% power. For example, planning for 100% of the observed effect lists 679 total events for 80% power and 909 total events for 90% power. These are planning illustrations based on the observed effect, not the trial's original design calculations.

The stated calculation assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The time-to-event calculations use the Schoenfeld approach. The effect to detect is the stated fraction of the observed effect. For a given assumed effect, the 80% power column represents the event count used in a planning scenario designed to have an 80% chance of detecting that effect under the assumptions; the 90% column uses a higher target chance and therefore lists more events. Smaller assumed effects require more information, which is why planning for a smaller fraction of the observed effect requires more total events.

The separate objective-response calculations are expressed as participants per arm, rather than total events. Their assumptions are the same stated two-sided significance level, equal allocation, normal approximation, and effect fractions. Those figures should be read as planning illustrations for the response outcome, not as the number enrolled in this trial. Neither set of calculations is observed power. The registry record's own original sample-size justification and design assumptions are not reported in the registry record.

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 = 1.240, so ln(HR) = 0.2151.
2. d = 4 × (1.96 + 0.842)² / [ln(1.240)]² = 678.5, rounded up to 679 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 <- 1.24                              # 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(1896479)
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 = 679 (80% power), 909 (90% power)
 75% of observed effect: events = 1,207 (80% power), 1,615 (90% power)
 50% of observed effect: events = 2,714 (80% power), 3,634 (90% power)
Simulated power with 679 events: 0.790 (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

For progression-free survival, the sponsor-reported analysis is a Cox proportional hazards comparison using the log-rank method. It reports a hazard ratio of 1.24 for 60 mg versus 140 mg, a 95% confidence interval of 0.90 to 1.70, and p = 0.192. The primary interpretation should be anchored in that reported analysis, rather than replacing it with a calculation based on summary values.

The secondary objective-response result also has a sponsor-reported analysis: a Cochran-Mantel-Haenszel odds ratio of 1.020, with a 95% confidence interval of 0.6 to 1.7 and p = 0.944. A separate unadjusted recalculation based on the posted response counts uses a risk difference, so it is not the same estimand or analysis as the sponsor's odds ratio. These analyses should not be treated as competing versions of one identical statistic. The registry record does not flag a discrepancy between the sponsor analysis and the recalculations.

Recalculations from posted summaries can serve as a check or offer a different descriptive contrast, but they cannot reproduce adjustments, stratification, or other details that may be part of an analysis unless those details are known. The sponsor's analysis is the recorded inferential result. The registry record does not report a complete analysis specification for either endpoint, so interpretation should remain tied to the methods and estimates explicitly posted.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates more than one endpoint, comparison, subgroup, or analysis. If many hypotheses are tested, interpreting each result against the same significance threshold can make false-positive findings more likely across the set of tests. A prespecified hierarchy or adjustment can help control that issue. The registry record lists progression-free survival as the primary outcome and objective response rate as a secondary outcome, but multiplicity procedures and any hierarchy for interpreting results are not reported in the registry record.

The primary outcome is described as an event-driven Intent-to-Treat analysis. That population included all randomized participants. The record states that the primary analysis was triggered by at least 150 events and that 155 events were reported at the cutoff. These details identify the analysis population and event trigger, but do not provide all the rules used for the analysis. Stratification factors, censoring details, protocol-deviation handling, and sensitivity analyses are not reported in the registry record.

Missing data can affect time-to-event and response outcomes in different ways. For time-to-event analysis, participants without a recorded event may be censored, and the validity of the estimate depends on how censoring relates to the outcome and analysis assumptions. For response, missing assessments can affect which participants are counted as evaluable or classified as responders. The registry record does not report the amount or pattern of missing outcome data, the rules for handling missing response assessments, or any imputation method.

The available description also does not report whether additional analysis populations were used for the non-inferiority assessment or how those populations were defined. No non-inferiority margin is reported in the registry record. These omissions limit the ability to assess the full confirmatory framework, including how multiplicity, missingness, and analysis-population choices may affect interpretation. The reported estimates remain interpretable as posted, but they cannot answer methodological questions that require details not reported in the registry record.

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: medullary thyroid cancer (7.3%), pneumonia (4.1%), anaemia (2.4%), asthenia (2.4%), sepsis (2.4%).

The safety figures count participants with serious adverse events and participants with other adverse events in each group. Serious adverse events were reported for 55 of 123 participants in the 60 mg group and 62 of 124 participants in the 140 mg group. Other adverse events were reported for 122 of 123 and 123 of 124 participants, respectively. The record also lists 59 deaths in the 60 mg group and 51 in the 140 mg group. These counts are descriptive and do not establish that an event was caused by treatment.

The listed serious adverse event terms include medullary thyroid cancer at 7.3%, pneumonia at 4.1%, and anaemia, asthenia, sepsis, and nausea at 2.4% each. The registry record does not clarify in the registry record whether the listed terms represent distinct participants or how overlapping events were handled. It also does not provide event-level timing, severity beyond the serious classification, duration, or attribution. These figures should therefore be read as reported safety summaries, not as a complete comparison of risks or evidence of causality.

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

The results describe a randomized comparison of two cabozantinib doses in participants with progressive, metastatic medullary thyroid cancer. The primary endpoint was progression-free survival assessed by independent radiology review. The reported median estimates were 11.0 months for 60 mg and 13.9 months for 140 mg, while the sponsor-reported hazard ratio for 60 mg versus 140 mg was 1.24, with a 95% confidence interval of 0.90 to 1.70 and p = 0.192.

These results show uncertainty around the relative comparison. The hazard-ratio interval includes 1, and the registry record does not provide a non-inferiority margin. Therefore the posted p-value and interval do not establish that the doses are equivalent or that the lower dose meets a non-inferiority criterion. A non-significant test is not proof that there is no difference, just as a point estimate alone is not a complete description of the evidence.

The results also do not determine which dose is appropriate for an individual, and they are not medical advice. They do not, by themselves, establish effects on outcomes that are not reported in the primary results, such as quality of life. The secondary objective-response outcome is a different measure and should not be treated as a replacement for the primary endpoint.

Interpretation is bounded by what the registry record reports. The endpoint definition, analysis population, event trigger, estimates, and intervals are available, but important details such as the original non-inferiority margin, complete statistical analysis plan, handling of missing data, and multiplicity procedures are not reported in the registry record. The posted results therefore support a careful description of the observed estimates and their precision, but not a broader claim that goes beyond those data and methods.

18. Why This Trial Matters Statistically

This trial is statistically notable because it compares two active doses rather than comparing treatment with no active treatment. The central analytic question is consequently about the relative outcome under 60 mg and 140 mg, with the study objective describing an interest in whether the lower dose could yield similar progression-free survival and response with fewer adverse events. The posted primary result must be judged against the intended comparison framework, especially because a non-inferiority question requires a prespecified margin.

The primary endpoint is time-to-event, and the registry describes an event-driven analysis in the Intent-to-Treat population. That makes the event count and timing central to the evidence, not merely the number enrolled. Reporting the medians alongside a hazard ratio helps distinguish absolute time summaries from a relative event-rate summary. Their confidence intervals also make visible the uncertainty that a point estimate alone would hide.

The registry's non-inferiority label makes the missing margin particularly consequential: without it, readers cannot tell what degree of disadvantage was considered acceptable under the study's formal criterion. Similarly, a p-value that does not show a statistically detectable difference cannot establish similarity. The secondary response outcome contributes a separate perspective, but its different scale and multiplicity context limit its ability to settle the primary question.

The safety counts add descriptive context, but counts without complete event definitions, timing, and attribution do not establish comparative causality. Overall, the trial illustrates why trial design labels, endpoint definitions, model-based estimates, confidence intervals, and unreported analysis details all matter when interpreting a clinical comparison.

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.

The most consequential limitation for interpreting the intended non-inferiority question is that the non-inferiority margin is not reported in the registry record. Without that margin, the reported hazard ratio and confidence interval cannot be evaluated against the formal boundary for an acceptable loss of effect. The complete statistical analysis plan, including any decision rule for non-inferiority, is also not reported in the registry record.

The proportional-hazards analysis provides one summary hazard ratio, but the registry entries do not report whether proportional hazards were assessed or provide enough detail to inspect the survival curves over time. The median estimates add useful context, yet they do not show the full time pattern of outcomes. The analysis is described as event-driven, but censoring rules and sensitivity analyses are not reported in the registry record.

For objective response, the sponsor reports an odds ratio while the unadjusted recalculation is a risk difference. Because these are different effect measures and the full analysis details are not reported, their estimates should not be treated as interchangeable. Multiplicity handling across endpoints is also not reported. Missing outcome data, response assessment completeness, and the procedures for handling missingness are not reported in the registry record.

The safety summaries give participant counts and selected serious adverse event terms, but do not provide the detail needed to assess event timing, duration, recurrence, attribution, or overlap between terms. The record lists deaths by group, but the registry entries do not indicate cause or relationship to treatment. These limitations narrow what can be concluded from the posted results; they do not negate the randomized comparison, but they require restraint in interpreting its scope.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study of Two Different Doses of Cabozantinib (XL184) in Progressive, Metastatic Medullary Thyroid Cancer — NCT01896479.
• Registry record last updated 2026-10-20; status last verified 2026-09.
• Results first posted 2025-09-02.
• 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

NCT01896479 is a randomized, quadruple-masked, parallel-group Phase 4 study sponsored by Exelixis. It compares oral cabozantinib (XL184) at 60 mg/day with 140 mg/day in progressive, metastatic medullary thyroid cancer. The record lists 247 participants enrolled and a status of active, not recruiting.

The primary outcome is progression-free survival assessed by a Blinded Independent Radiology Committee under RECIST 1.1 in the Intent-to-Treat population. Median estimates are 11.0 months for 60 mg and 13.9 months for 140 mg. The sponsor-reported hazard ratio is 1.24, with a 95% confidence interval of 0.90 to 1.70 and p = 0.192. Objective response rate is also reported as a secondary outcome.

The posted comparison describes the observed estimates but does not, on its own, establish equivalence or non-inferiority. The non-inferiority margin, multiplicity procedures, and several analysis details are not reported in the registry record.