A Study to Evaluate Denosumab in the Treatment of Postmenopausal Osteoporosis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Evaluate Denosumab in the Treatment of Postmenopausal Osteoporosis (NCT00089791), a randomized phase 3 study. The primary outcome is Number of Participants With New Vertebral Fractures.

2. Clinical Question

The primary endpoint uses a radiographic definition: a new vertebral fracture was identified by lateral spine X-ray using the Genant semiquantitative scoring method. The registry also reports two secondary fracture outcomes over 36 months: nonvertebral fractures and hip fractures.

This is a comparison of assigned treatment groups in a randomized, double-blind, parallel-group phase 3 trial. The central statistical question is the between-group contrast in the recorded fracture outcomes, alongside the uncertainty around that contrast. The registry record describes the endpoint and analysis population, but does not report the trial’s original sample-size justification or a complete statistical analysis plan.

3. Trial Design

The registry identifies this as a phase 3, randomized, double-blind, parallel-group study sponsored by Amgen. The enrollment was 7808. The study is marked completed. Its stated purpose was to evaluate the effectiveness and safety of denosumab in treating women with postmenopausal osteoporosis.

Participants were assigned to placebo or denosumab 60 mg, both administered subcutaneously once every 6 months for 3 years. The registry lists the primary outcome as the number of participants with new vertebral fractures at 36 months. It also lists nonvertebral fractures and hip fractures as secondary outcomes, each with a 36-month timeframe.

The analysis counts differ by outcome. For the primary endpoint, the placebo group included 3691 participants and the denosumab group included 3702 participants. The primary efficacy analysis set was defined as randomized participants with a baseline and at least one postbaseline vertebral-fracture evaluation at or before 3 years. The registry states that Last Observation Carried Forward was used for that analysis. The secondary endpoint groups included 3906 participants in the placebo group and 3902 in the denosumab group.

The registry record does not report the original sample-size justification, allocation ratio details beyond the randomized parallel-group design, stratification factors, interim analyses, or a complete account of the statistical analysis plan. Those details should not be inferred from the reported outcome counts or analyses.

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

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Primary Efficacy Analysis Set, which includes all randomized participants who have a baseline and ≥ 1 postbaseline evaluation of vertebral fracture at or before 3 years. Last Observation Carried Forward was used.

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 New Vertebral Fractures, measured over 36 months.

The primary endpoint is the number of participants with a new vertebral fracture during the 36-month timeframe. A new fracture was identified on lateral spine X-ray using the Genant semiquantitative scoring method. The definition required at least a one-grade increase from a baseline grade of 0 in any vertebra from T4 to L4.

The endpoint included morphometric vertebral fractures, assessed at scheduled visits and not associated with signs or symptoms indicative of fracture, as well as clinical vertebral fractures. Clinical fractures could be assessed at scheduled or unscheduled visits and were associated with signs or symptoms. Fractures associated with high trauma severity or a pathologic fracture were excluded from the clinical-fracture definition.

This endpoint combines radiographic findings with clinically apparent events under the stated definitions. It is used to compare the number of participants meeting the fracture criteria in each assigned group. The registry’s primary efficacy analysis set required a baseline evaluation and at least one postbaseline evaluation by 3 years; Last Observation Carried Forward was used. The registry record does not explain the rationale for selecting this endpoint or describe further details of how missing assessments were handled.

7. Results Posted in the Registry

• Denosumab 60 mg Q6M: 2.3% (86/3702) versus 7.2% (264/3691) on control; risk difference -4.8 percentage points (95% CI -5.8 to -3.9), risk ratio 0.32 (0.26–0.41); unadjusted two-proportion z-test p = <0.001.

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.

Number of Participants With Nonvertebral Fractures

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Denosumab 60 mg Q6M: 6.1% (238/3902) versus 7.5% (293/3906) on control; risk difference -1.4 percentage points (95% CI -2.5 to -0.3), risk ratio 0.81 (0.69–0.96); unadjusted two-proportion z-test p = 0.014.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Denosumab 60 mg Q6M than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Number of Participants With a Hip Fracture

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Denosumab 60 mg Q6M: 0.7% (26/3902) versus 1.1% (43/3906) on control; risk difference -0.4 percentage points (95% CI -0.8 to -0.0), risk ratio 0.61 (0.37–0.98); unadjusted two-proportion z-test p = 0.040.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Denosumab 60 mg Q6M than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Both posted secondary fracture outcomes have event counts in the same direction as the primary outcome: fewer participants in the denosumab group had recorded events than in the placebo group. The sponsor-reported Cox analyses also give hazard ratios below 1 for both secondary endpoints, with confidence intervals that remain below 1. This directional consistency is supportive context for the primary result, but it does not make the secondary outcomes equivalent to the primary endpoint or independently confirm the same claim.

Secondary analyses generally carry less confirmatory weight when several outcomes are examined and the approach to controlling false-positive findings is unknown. The registry record lists two secondary outcomes and gives p-values, but a multiplicity adjustment strategy is not reported. Without knowing which comparisons were prespecified, whether a testing hierarchy was used, or how multiplicity was handled, the secondary p-values should be read as nominal results rather than as evidence with a known family-wise error rate.

The endpoints also differ in event frequency and in the reported analysis method. The registry reports Cox regression for the secondary results, whereas the simple recalculations use the posted counts. Time-to-event analyses can use event timing and censoring; counts do not retain that information. Differences between those measures and methods mean that close numerical agreement is not expected or required.

The overall pattern is therefore coherent in direction, while the strength of inference depends on the primary endpoint, the analysis plan, and the handling of multiple comparisons. The registry record does not report a complete statistical analysis plan, so it is not possible to establish from the record whether these secondary analyses were intended as confirmatory.

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

• Number of Participants With Nonvertebral Fractures
• Number of Participants With a Hip Fracture

9. Statistical Interpretation

10. Statistical Methods Explained

General statistical explanation: a binary endpoint records whether each participant did or did not experience an event. The posted counts are participants with the event, rather than a measure of how many fractures each participant experienced. Comparing the groups can be expressed in different ways, including a risk difference, a risk ratio, an odds ratio, or a time-to-event measure. Each measure answers a related but distinct question and uses a different scale.

What does a risk difference describe?

General: a risk difference subtracts the event risk in one group from the event risk in the other. A value below zero for the denosumab-minus-placebo contrast means the recorded event risk was lower in the denosumab group. A confidence interval around this contrast describes the range of effect values compatible with the data and the calculation’s assumptions. If the interval includes zero, the data are compatible with no difference on that scale; if it excludes zero, they are not compatible with zero at the confidence level used.

The unadjusted primary comparison in the registry gives a risk difference, while the sponsor’s reported primary analysis gives a risk ratio. The recalculated comparison uses posted summary counts, so it is not a substitute for the sponsor’s own analysis. It should be read as a simple unadjusted calculation rather than an attempt to reproduce any unreported adjustments or modeling choices.

Why are the reported analysis methods different?

The registry reports a Mantel-Haenszel method for the primary endpoint and reports the result as a risk ratio. It reports Cox regression for the secondary endpoints and gives hazard ratios. General: a risk ratio compares event risks over a specified period, while a hazard ratio compares event rates over time among participants who have not yet experienced the event. The Cox model is commonly used for time-to-event data because it incorporates the timing of events and censoring, rather than reducing the information to a simple end-of-period event count.

The registry does not describe the covariates, strata, or other model details used in these reported analyses. It would therefore be inappropriate to infer what adjustments were made or to treat the posted unadjusted calculations as equivalent to the sponsor’s model-based results.

What does a p-value contribute?

General: a p-value measures how surprising data at least as incompatible with a null hypothesis as the observed data would be, if the null hypothesis and the analysis assumptions were true. It does not measure the size or clinical importance of an effect, nor does it give the probability that a finding is due to chance. Interpretation should consider the effect estimate, its confidence interval, the analysis population, and the number of outcomes being tested.

The registry reports p-values for the primary and secondary analyses. Its posted unadjusted recalculations also give p-values for comparisons based on the counts. These are not interchangeable: methods, estimands, and assumptions may differ. The registry does not report a multiplicity adjustment strategy, so the secondary p-values should not be treated as if they were adjusted for testing several endpoints.

How should missing observations and the analysis set be considered?

The primary efficacy analysis set included randomized participants who had a baseline and at least one postbaseline vertebral-fracture evaluation at or before 3 years. The registry says Last Observation Carried Forward was used. General: this approach carries a participant’s last observed value forward to a later assessment when a later value is missing. It relies on assumptions about how the unobserved outcome relates to the last recorded one; those assumptions can be problematic if outcomes change after the last assessment or if missingness differs across groups.

The registry does not report the extent or reasons for missing assessments, sensitivity analyses for missing data, or the precise implementation of the carried-forward procedure. It also does not report a complete statistical analysis plan or interim-analysis details. Consequently, readers can describe the stated analysis population and method, but cannot assess all robustness checks or analytic decisions from the registry record alone.

11. Confidence Intervals and Precision

The primary reported analysis gives a risk ratio of 0.32 with a 95% confidence interval from 0.26 to 0.41. The interval lies below 1, the value corresponding to equal risks on a ratio scale. The posted unadjusted risk-difference interval also excludes no difference. Together, these intervals indicate that the posted estimates are not especially close to the respective null values, while still expressing uncertainty about the size of the contrast.

General: a confidence interval is a range produced by a statistical procedure that, under its assumptions, would contain the target effect at the stated rate across repeated applications. It is not a guarantee that the true effect lies inside a particular interval, and it does not include uncertainty from every possible design or measurement limitation. Interval width reflects statistical precision: narrower intervals indicate less sampling uncertainty on that scale, all else equal.

The secondary reported intervals are wider in relation to their estimates than the primary interval, particularly for the hip-fracture hazard ratio. These secondary results therefore have more uncertainty about the magnitude of the association. The event counts are also smaller for the hip-fracture outcome than for the other posted outcomes, so the estimate should be read with its interval rather than as a precise measure.

Precision is not the same as importance. The registry results quantify uncertainty in the statistical contrasts, but do not establish the relevance of a particular effect size for an individual participant. The registry record does not report a prespecified threshold for a clinically important difference.

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 figures are planning illustrations based on the observed effect, not the trial’s original design justification. For these binary fracture outcomes, the figures are participants per arm. The stated assumptions are two-sided significance level 0.05, equal allocation between the two arms, and a normal approximation; the effect to detect is the stated fraction of the observed effect. The registry record does not report the trial’s original sample-size assumptions.

The 80% power column gives an illustrative number of participants per arm for a design with an 80% chance of detecting the specified effect under those assumptions, if that effect and the other planning assumptions were correct. The 90% power column describes the corresponding planning target for a 90% chance. These are design calculations, not guarantees about a future study’s result. This is not observed power.

Rows labeled as planning for 100%, 75%, or 50% of the observed effect show how the assumed effect size changes the required enrollment. A smaller effect is harder to distinguish from random variation, so it requires more participants per arm to reach the same planned power and significance level. The observed effect is only an estimate and may differ from the effect in another study.

For time-to-event outcomes, the stated planning convention would give total events rather than participants per arm; the outcomes are binary, so the figures are participants per arm. The registry record does not report whether these illustrations correspond to the assumptions used when the trial was originally designed.

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

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 = 0.072; treatment proportion p₁ = 0.023; average p̄ = 0.047.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.3004 = 0.5889.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.2985 = 0.2512.
4. n = (0.5889 + 0.2512)² / (-0.0483)² = 302.6, rounded up to 303 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 <- 0.0715253318883771                          # control proportion
p1 <- 0.0232306861156132                          # 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(89791)
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 = 303 (80% power), 405 (90% power)
 75% of observed effect: n per arm = 604 (80% power), 808 (90% power)
 50% of observed effect: n per arm = 1,505 (80% power), 2,014 (90% power)
power.prop.test: 302.6 
Simulated power at n = 303 per arm: 0.826 (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.

14. Analyses Reported by the Sponsor

The sponsor-reported primary analysis used a Mantel-Haenszel method and reported a risk ratio of 0.32, with a 95% confidence interval from 0.26 to 0.41 and p < 0.001. The unadjusted recalculation uses the posted group counts and reports a risk difference, with p <0.001. Both approaches indicate the same direction of contrast, but the effect measures are on different scales and should not be compared as though they were identical estimates.

The registry record marks no discrepancy between the posted analysis and recalculation. The sponsor’s reported analysis remains the trial’s own reported analysis; the recalculation is a simple, unadjusted check based on summary counts. The registry does not report enough detail about the Mantel-Haenszel analysis to reproduce all its choices independently.

For the secondary outcomes, the sponsor-reported analyses use Cox regression and report hazard ratios. The unadjusted recalculations are based on event counts and risk differences. These answer related questions, but one incorporates time-to-event information while the other compares event proportions over the stated timeframe. The distinction matters when interpreting estimates and intervals.

15. Multiplicity, Missing Data and Analysis Populations

The registry record identifies one primary endpoint and two secondary fracture outcomes. It reports analyses and p-values for these outcomes, but a multiplicity strategy is not reported in the registry record. It is therefore unknown whether testing was organized hierarchically, whether an adjustment was planned, or how the family of hypotheses was defined. This matters because assessing several outcomes can increase the chance of at least one apparently notable result even when no effect exists for any outcome.

The primary analysis population is described as randomized participants with a baseline and at least one postbaseline vertebral-fracture evaluation at or before 3 years. Last Observation Carried Forward was used. The registry record does not report the number of participants excluded from that analysis set, the amount of missing outcome data, reasons for missingness, or sensitivity analyses examining alternative missing-data assumptions. It also does not specify how the carried-forward procedure was applied to different types of missing assessment.

The secondary outcomes list group sizes that differ from those for the primary efficacy analysis. The registry record does not explain the analysis-population rules behind this difference. It also does not report a complete statistical analysis plan, details of any interim analyses, or the original sample-size justification. Those omissions limit what can be established about prespecification and the robustness of reported inferences.

General: missing data can affect both the estimated effect and its precision when the probability of missingness is related to outcomes or differs between groups. Multiplicity can affect how strongly nominal p-values support conclusions across several endpoints. Neither issue can be resolved simply by comparing event counts. The results should be interpreted in light of these unreported details rather than assuming that a particular strategy was used.

16. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

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: osteoarthritis (2.0%), pneumonia (0.9%), atrial fibrillation (0.9%), breast cancer (0.9%), angina pectoris (0.8%).

The registry reports serious adverse events for 972 of 3876 participants at risk in the placebo group and 1004 of 3886 participants at risk in the denosumab group. It also reports other, non-serious adverse events for 3107 of 3876 participants at risk in the placebo group and 3073 of 3886 participants at risk in the denosumab group. These are counts of participants affected, as described in the record.

The most frequent listed serious adverse-event terms include osteoarthritis (2.0%), pneumonia (0.9%), atrial fibrillation (0.9%), breast cancer (0.9%), angina pectoris (0.8%), and cerebrovascular accident (0.8%). The registry record does not state in this summary how these terms were distributed by treatment group. Death counts are not reported in the registry record.

These figures describe adverse events recorded during the study; they do not by themselves establish causation or compare event severity in a way that accounts for all clinical context. Differences in the number of participants at risk and event counts should be interpreted with the registry’s reporting definitions and limitations in mind.

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

The posted results show fewer participants with recorded new vertebral fractures in the denosumab group than in the placebo group over the 36-month timeframe. The primary sponsor-reported risk ratio and its confidence interval indicate a lower event measure for the denosumab group under the reported analysis. The secondary outcomes show the same direction of group contrast, though they carry less confirmatory weight because the multiplicity strategy is not reported.

These results do not establish that every participant’s outcome will match the group-level pattern. A group comparison describes average event occurrence under the trial’s definitions and analysis population; it does not predict an individual outcome. The results also do not settle questions beyond the outcomes, timeframe, and population reported in the registry.

The registry record does not report the original sample-size justification, a complete statistical analysis plan, an account of interim analyses, or a multiplicity strategy. It also does not report the extent of missing primary-endpoint assessments or sensitivity analyses around the stated Last Observation Carried Forward method. Those omissions limit assessment of all analytic choices and the robustness of the findings.

Statistical evidence is not the same as a complete clinical judgment. The p-values do not quantify the probability that an effect is real, and the confidence intervals do not remove uncertainty about design, measurement, or generalizability. These descriptions should be read as an account of the posted registry data and analyses, not as medical advice or a recommendation about treatment.

18. Why This Trial Matters Statistically

This trial illustrates how a large randomized comparison can report both an outcome count and a formal effect estimate. For the primary endpoint, the registry presents an unadjusted comparison based on group counts and a sponsor-reported Mantel-Haenszel risk ratio. Reading both requires attention to the measure used: a risk difference and a risk ratio convey related but non-identical contrasts.

It also demonstrates why uncertainty and analysis populations matter. The primary result is based on a specifically defined efficacy set, and the record states that Last Observation Carried Forward was used. The group sizes for secondary outcomes differ from those for the primary analysis, while the record does not explain all population rules. Clear definitions help readers understand what the counts represent, but missing methodological detail limits independent evaluation.

The secondary outcomes provide a consistent directional pattern, yet their interpretation depends on multiplicity and prespecification. Without a reported strategy for handling multiple comparisons, nominal p-values should not be treated as if each were the sole planned test. The trial is therefore a useful example of reading a primary finding alongside secondary analyses without overstating their evidential strength.

Finally, the registry’s sample-size illustrations are distinct from the trial’s original design rationale. They show how assumptions about the effect to detect relate to planning enrollment, but they cannot explain why the original study enrolled its participants. Separating observed results, sponsor analyses, and planning illustrations is essential for accurate statistical interpretation.

19. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

Several trial-specific details needed for a full appraisal are not reported in the registry record. These include the original sample-size justification, a complete statistical analysis plan, interim-analysis details, and the strategy for handling multiplicity across the primary and secondary outcomes. The record also does not explain the difference between the primary and secondary analysis group sizes.

The primary efficacy set required at least one postbaseline vertebral-fracture evaluation and used Last Observation Carried Forward. The registry does not report how many participants lacked such assessments, why evaluations were missing, or whether sensitivity analyses were conducted. These omissions make it difficult to assess how missing observations may have influenced the primary estimate.

The posted summary does not give all details of the reported Mantel-Haenszel or Cox analyses, including any covariates or stratification choices. The unadjusted recalculations use summary counts and cannot reproduce unreported model details. The registry record also does not provide the distribution of each listed serious adverse-event term by group in the summary, and deaths are not reported in the registry record.

These limitations concern what can be evaluated from the registry record; they do not establish that a particular procedure was or was not used. Conclusions should remain tied to the reported population, endpoints, timeframe, and analyses.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study to Evaluate Denosumab in the Treatment of Postmenopausal Osteoporosis — NCT00089791.
• Registry record last updated 2022-11-08; status last verified 2022-11.
• Results first posted 2014-01-10.
• 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

NCT00089791 is a completed phase 3 randomized, double-blind, parallel-group study sponsored by Amgen. It enrolled 7808 participants and compared denosumab 60 mg administered subcutaneously once every 6 months for 3 years with placebo administered on the same schedule.

The primary outcome was the number of participants with new vertebral fractures at 36 months. The registry also reports nonvertebral fractures and hip fractures as secondary outcomes. The sponsor-reported primary analysis used the Mantel-Haenszel method and gave a risk ratio of 0.32, with a 95% confidence interval from 0.26 to 0.41 and p < 0.001. The registry states that Last Observation Carried Forward was used for the primary efficacy analysis.

The posted secondary analyses used Cox regression. The original sample-size justification, multiplicity strategy, interim analyses, and complete statistical analysis plan are not reported in the registry record.