Safety and Efficacy Study of MDV3100 in Patients With Castration-Resistant Prostate Cancer Who Have Been Previously Treated With Docetaxel-based Chemotherapy: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Safety and Efficacy Study of MDV3100 in Patients With Castration-Resistant Prostate Cancer Who Have Been Previously Treated With Docetaxel-based Chemotherapy (NCT00974311), a randomized phase 3 study. The primary outcome is Overall Survival.

2. Clinical Question

The registry describes a phase 3, randomized, double-blind, placebo-controlled study. Its primary endpoint was overall survival during the study period, reported as extending up to 101 months. The comparison is based on the time-to-event analysis and the medians reported for each group. A median survival is the time by which half of the relevant survival distribution has experienced the event, subject to the analysis and censoring described in the record; it is not a prediction of an individual participant’s survival.

The central statistical question is whether the posted survival data distinguish the groups, and how precisely they estimate the relative event rate over follow-up. The registry also lists outcomes concerning disease progression, skeletal events, patient-reported measures, pain and response. Those outcomes can add context, but the primary question remains the comparison of overall survival. The results describe the randomized groups and do not, by themselves, settle every question about benefit, harm, or applicability to people outside the study population.

3. Trial Design

The registry record identifies NCT00974311 as a phase 3, randomized, parallel-group study comparing enzalutamide with placebo. It reports triple masking and an overall enrollment of 1199 participants. The arm totals in the posted primary analysis are 800 for enzalutamide and 399 for placebo. The study is described as multinational in its official title; the record names Pfizer as the sponsor and gives the status as completed.

The enrolled population was people with castration-resistant prostate cancer previously treated with docetaxel-based chemotherapy. The registry’s description of the primary analysis states that the intention-to-treat population included all randomized participants. Overall survival was defined as time from randomization to death from any cause. Follow-up time for participants who were lost to follow-up or not known to have died by the analysis cut-off was right-censored, including participants known to have died after that cut-off.

Random allocation aims to create groups that are comparable on average at the start of a trial. Masking is intended to reduce the influence of treatment knowledge on care, assessment, and reporting; the precise operational details of masking are not reported in the registry record. The parallel design means participants were allocated to one of the study arms rather than being described as crossing between arms during the randomized comparison.

The registry identifies a placebo group and reports results for the randomized arms. Its primary time frame is during the study period, up to 101 months. Some secondary outcomes are described for the double-blind phase, with a stated period up to 24 months. The trial’s original sample-size justification, planned allocation ratio, detailed protocol procedures, and further design assumptions are not reported in the registry record. Planning illustrations based on the observed results should therefore not be treated as documentation of the original design.

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: Participants were randomized 2:1 to receive either Enzalutamide or Placebo in DB phase. All participants who continued in OLE phase, received Enzalutamide.

DB Phase (up to 24 Months)

Reasons for not completing this period:

OLE Phase (up to 77 Months)

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT included all participants who were randomized into the study.

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: Intent-to-treat population (ITT) included all participants who were randomized into the study.

6. Primary Endpoint

The registry lists the primary outcome as Overall Survival, measured over During study period (up to 101 months).

Overall survival is the time from randomization until death from any cause. It is a time-to-event endpoint: each participant contributes information about how long they were followed and whether the event was observed. The registry reports median overall survival for each randomized group and a hazard ratio from a log-rank analysis.

Overall survival is used because it captures death regardless of its cause and does not require deciding whether a particular death was related to the cancer or study treatment. It is also an endpoint that can be interpreted directly as time until an event. That does not make its analysis free of assumptions: participants may have different lengths of follow-up, and some will have no recorded death by the data cut-off. The registry states that such observations were right-censored in specified circumstances.

General statistical explanation: the median describes the point in the estimated survival distribution at which the event has occurred for half of the population represented by that distribution. It is a summary of the group’s observed time-to-event pattern, not a guarantee about an individual’s course. The hazard ratio summarizes the relative event rates over the analysis period under the time-to-event model used for the reported comparison. It does not express a difference in median time, and it should be read alongside its confidence interval and the follow-up and censoring information.

The endpoint is evaluated in the intention-to-treat population, which the record defines as all participants randomized into the study. This preserves the randomized group comparison in the primary analysis description, even when the registry does not provide further detail about treatment received or adherence.

7. Results Posted in the Registry

• Enzalutamide vs Placebo: hazard ratio 0.63 (95% CI 0.53–0.75), reported p < 0.001 (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.

Radiographic Progression-free Survival

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor (Log Rank); it was not recalculated.
• Result:
  • Enzalutamide vs Placebo: hazard ratio 0.40 (95% CI 0.35–0.47), reported p < 0.001 (as reported in the registry).
• Interpretation:
  • The 95% confidence interval excludes no difference (a ratio of 1), so the hazard of the event is statistically lower for Enzalutamide vs Placebo than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Time to First Skeletal-related Event

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor (Log Rank); it was not recalculated.
• Result:
  • Enzalutamide vs Placebo: hazard ratio 0.69 (95% CI 0.57–0.83), reported p < 0.001 (as reported in the registry).
• Interpretation:
  • The 95% confidence interval excludes no difference (a ratio of 1), so the hazard of the event is statistically lower for Enzalutamide vs Placebo than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percentage of Participants Who Were Responders for Functional Assessment of Cancer Therapy-Prostate (FACT-P)

• 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:
  • Enzalutamide: 43.2% (281/651) versus 18.3% (47/257) on control; risk difference 24.9 percentage points (95% CI 18.8 to 30.9), risk ratio 2.36 (1.80–3.10); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Enzalutamide than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes cover several aspects of disease course and participant experience, including radiographic progression-free survival, time to first skeletal-related event, functional assessment, pain, prostate-specific antigen outcomes, soft-tissue response, and quality of life. The registry record lists nine posted secondary outcomes. Numerical results are available in the record for some of them, while detailed estimates for every listed endpoint are not reported in the registry record.

For the outcomes with numerical estimates described in the record, the direction is broadly consistent with the primary comparison: the reported time-to-event hazard ratios are below 1, and the functional assessment responder percentage is higher in the enzalutamide group. This pattern can provide supporting context because the outcomes address progression, skeletal events, and functioning rather than mortality alone. They remain separate measures with distinct definitions, time frames, and analysis populations; consistency in direction does not mean they measure the same effect.

Secondary outcomes are generally interpreted with more caution when many outcomes are examined. General statistical explanation: testing multiple outcomes increases the chance that at least one result appears statistically notable by chance, unless the analysis plan controls that risk. The registry record does not report a multiplicity-adjustment procedure, so the secondary p-values should not automatically be read as if each were the sole prespecified test. Their findings are best treated as supportive or exploratory in the context of the full outcome set.

The available secondary results do not establish that every listed outcome favored one group, and missing numerical details should not be inferred from the primary result. The record does not report enough about prespecification, hierarchy among secondary endpoints, or multiplicity control to assign the secondary findings the same evidential weight as the primary endpoint. Their interpretation depends on the posted outcome-specific estimates and the limits of the registry information.

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

• Radiographic Progression-free Survival
• Time to First Skeletal-related Event
• Percentage of Participants Who Were Responders for Functional Assessment of Cancer Therapy-Prostate (FACT-P)
• Time to Prostate-specific Antigen (PSA) Progression
• Percentage of Participants With Pain Palliation
• Percentage of Participants With Prostate Specific Antigen (PSA) Response
• Percentage of Participants With Soft-tissue Objective Response
• European Quality of Life Five-Domain (EQ-5D) Scale

9. Statistical Interpretation

10. Statistical Methods Explained

The primary result is a time-to-event comparison. The registry reports the medians by group and a hazard ratio from a log-rank analysis. General statistical explanation: these statistics describe related but distinct features of survival data. A median summarizes a point in the survival distribution; a hazard ratio summarizes relative event rates across follow-up. Neither should be interpreted as an individual-level forecast.

What does a hazard ratio represent?

General statistical explanation: a hazard is the event rate among participants who have not yet experienced the event, considered at a given point in follow-up. A hazard ratio compares the hazards between groups over time under the assumptions of the analysis. A value below 1 for enzalutamide versus placebo indicates that the estimated event hazard is lower in the enzalutamide group in this comparison. It is not a direct comparison of the number of deaths, a difference in median survival, or a percentage of participants who survive.

The registry reports a hazard ratio of 0.63, a 95% confidence interval from 0.53 to 0.75, and p < 0.001 for overall survival. The ratio and interval are reported from the sponsor’s log-rank analysis. The registry’s comparison information identifies the same estimate and interval. No newly calculated time-to-event estimate is needed to explain the posted result.

What does the log-rank analysis do?

General statistical explanation: a log-rank test compares the observed and expected pattern of events between groups over the follow-up period. It uses event timing and the number of participants still at risk, rather than reducing each participant’s experience to a simple yes-or-no outcome at a single time. The registry names the log-rank method for the primary comparison and reports its hazard ratio and p-value. More detailed information about model specifications, stratification, covariate adjustment, or proportional-hazards checks is not reported in the registry record.

A common interpretation of a reported p-value is how unusual the data or a more extreme result would be under a specified null hypothesis and the test’s assumptions. It is not the probability that the null hypothesis is true. The reported p < 0.001 indicates strong statistical evidence against the null comparison under the reported analysis; it does not, by itself, describe clinical importance or account for all possible sources of bias.

How should censoring be understood?

General statistical explanation: right-censoring occurs when the event time is not observed during the available follow-up, while the participant’s time under observation is still informative. The registry says that overall survival time was censored for participants lost to follow-up or not known to have died by the analysis cut-off, including participants known to have died after the cut-off. The record does not give further details on follow-up completeness, censoring patterns by group, or sensitivity analyses addressing departures from censoring assumptions.

Survival analyses use censored observations rather than treating them as if the event had occurred at the last contact. The validity of the resulting estimates depends in part on the relationship between censoring and the chance of the event after censoring. This is a general property of time-to-event methods; the registry record does not report enough detail to assess that assumption in this trial beyond its description of censoring.

What does the intention-to-treat population contribute?

The primary endpoint description defines the intention-to-treat population as all participants randomized into the study. General statistical explanation: analyzing participants according to their randomized groups maintains the comparison created by random assignment and avoids selecting participants based on events after randomization. The registry provides group sizes for the primary analysis and this population definition. It does not provide a detailed account of missing endpoint data beyond the censoring description or explain whether any additional analyses used different populations.

The arm counts are 800 for enzalutamide and 399 for placebo in the primary results. These are the analysis-group sizes reported for the endpoint. They should not be mistaken for numbers of deaths or numbers of participants with complete follow-up; those event counts are not reported in the registry record.

How do the secondary analyses differ?

The registry reports additional analyses using time-to-event and binary-outcome approaches. For radiographic progression-free survival and time to first skeletal-related event, it reports medians and hazard ratios from log-rank analyses. For the functional assessment responder outcome, the sponsor reports a difference in percentage of participants using a Cochran-Mantel-Haenszel analysis. General statistical explanation: a binary outcome classifies participants by whether a defined response occurred during the measurement period; a time-to-event outcome also uses when the event occurred and accommodates varying follow-up through censoring.

The reported secondary methods are not interchangeable. A hazard ratio addresses event timing, whereas a difference in percentages compares response proportions. Their estimates, confidence intervals, and p-values should be understood within their own endpoint definitions and analysis populations. The registry lists a broader set of secondary outcomes, but detailed results and analysis methods for every listed outcome are not reported in the registry record.

Methods not described in the record should not be inferred. In particular, the original sample-size justification, interim analyses, detailed prespecified statistical plan, and multiplicity procedure are not reported in the registry record. The planning illustrations based on observed effects are separate from the trial’s original design.

11. Confidence Intervals and Precision

The primary hazard-ratio estimate is 0.63, with a 95% confidence interval from 0.53 to 0.75. The interval gives a range of estimates compatible with the data under the analysis and its assumptions. Because the full interval is below 1, the reported estimate and its uncertainty are on the side of a lower hazard in the enzalutamide group. The interval is not a range containing 95% of individual participants’ outcomes, and it does not account for every possible source of uncertainty or bias.

The registry also reports median overall survival of 18.4 months for enzalutamide, with a lower confidence limit of 17.3 months and no upper confidence limit reported, and 13.6 months for placebo, with a 95% confidence interval from 11.3 to 15.8 months. The absent upper limit for the enzalutamide median is not reported in the registry record. It should not be filled in by extrapolation.

General statistical explanation: confidence intervals are useful because they show both the direction and the precision of an estimate. A narrower interval indicates less statistical uncertainty under the method; a wider interval indicates more. The interval depends on the observed data and analysis assumptions. It is not a guarantee that the true effect falls within its endpoints, and it does not establish that the estimated effect is clinically important.

The p-value, p < 0.001, addresses compatibility with a null comparison under the reported test; it is not a measure of precision by itself. The confidence interval is more useful for seeing the range of plausible hazard-ratio values. Precision for the medians and hazard ratio should be considered together with follow-up, censoring, and the fact that the registry does not report the primary event count or additional sensitivity analyses.

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 illustrations translate observed effects into planning quantities; they do not describe the trial’s original design assumptions. For time-to-event outcomes, the entries are total events required across the study, not participants per arm. For the binary functional assessment outcome, the entries are participants per arm. The units therefore differ by outcome, and event totals should not be read as enrollment targets.

The illustrations use a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. They consider effects equal to 100%, 75%, or 50% of the observed effect. For overall survival, planning for 100% of the observed effect gives 148 total events for 80% power and 197 total events for 90% power. Planning for 50% of the observed effect gives 589 total events for 80% power and 788 total events for 90% power. These examples illustrate the direction of the relationship, rather than a recommendation or description of the actual study plan.

The 80% power column means the planning calculation targets an 80% chance of detecting the stated effect under its assumptions if that effect were true. The 90% column uses a higher target chance of detection. For the same assumed effect, the higher target generally requires more information. Smaller effects are harder to distinguish from random variation, so the illustration requires more participants or events as the assumed effect becomes smaller.

For other endpoints, read the row using its stated unit: total events for the time-to-event outcomes and participants per arm for the binary outcome. The curve is another way of displaying how the planning requirement changes as the assumed effect changes. These figures are not observed power. The 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 = 0.630, so ln(HR) = -0.4620.
2. d = 4 × (1.96 + 0.842)² / [ln(0.630)]² = 147.1, rounded up to 148 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 <- 0.63                              # 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(974311)
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 = 148 (80% power), 197 (90% power)
 75% of observed effect: events = 262 (80% power), 351 (90% power)
 50% of observed effect: events = 589 (80% power), 788 (90% power)
Simulated power with 148 events: 0.813 (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. Group-Sequential Design Analysis

Many time-to-event trials, especially in oncology, stop or report early when interim data are overwhelming, so a single fixed-sample number understates the real design problem. The calculation below is an illustrative group-sequential design built from this trial's posted control median and hazard ratio. The registry record does not state the planned number of looks or the spending function, so the design is a planning illustration, not a reconstruction of the protocol. Assumed: one-sided α = 0.025, 90% power, three equally spaced looks on information, Lan-DeMets O'Brien-Fleming spending, 24 months of accrual, 60 months in total and a 1% monthly dropout hazard.

Under these assumptions the design needs about 275 participants in total, with the boundaries below.

Fixed design versus group-sequential alternatives

Group-sequential designs cost a small increase in maximum sample size and events compared with the fixed design, in exchange for the chance to stop early for overwhelming efficacy. O'Brien-Fleming spending keeps early boundaries very conservative and the final boundary close to the fixed-design value, whereas Pocock spending spends more alpha early.

Sensitivity to a smaller true effect

Planning for exactly the observed effect is optimistic, because observed effects in published trials tend to be somewhat larger than the true effect. The table repeats the main design (three looks, O'Brien-Fleming spending, 90% power) for smaller true hazard ratios.

For context, this trial enrolled 1,199 participants.

Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993. Agreement between the two packages confirms the boundary calculation.

R code and output

suppressWarnings(suppressMessages({ library(gsDesign); library(rpact) }))
med_control <- 13.6          # control-arm median, months (converted from the posted unit)
hr          <- 0.63          # hazard ratio used for planning
lambdaC     <- log(2) / med_control
eta         <- 0.01               # dropout hazard per month (assumption)
acc_months  <- 24                 # accrual duration (assumption)
study_months<- 60                 # total study duration (assumption)

design <- function(k, sfu, beta, h = hr) gsSurv(k = k, test.type = 1, alpha = 0.025, beta = beta, sfu = sfu,
  lambdaC = lambdaC, hr = h, hr0 = 1, eta = eta, gamma = 1, R = acc_months, T = study_months, minfup = study_months - acc_months)

# Main design: 3 looks, Lan-DeMets O'Brien-Fleming spending, one-sided alpha 0.025, 90% power
x <- design(3, sfLDOF, 0.10)
cat(sprintf("N_TOTAL|%s\n", format(ceiling(sum(x$gamma * x$R)), big.mark = ",")))
for (i in 1:3) cat(sprintf("GS|%d|%.1f|%.0f|%.3f|%.5f|%.5f\n", i, x$T[i], x$n.I[i], x$upper$bound[i], pnorm(-x$upper$bound[i]), cumsum(x$upper$spend)[i]))
cat("\n"); print(gsBoundSummary(x))

# Comparison of designs
cmp <- list(list("Fixed design, one analysis", 1, sfLDOF), list("Two looks, O'Brien-Fleming", 2, sfLDOF),
            list("Three looks, O'Brien-Fleming", 3, sfLDOF), list("Three looks, Pocock", 3, sfLDPocock))
cat("\n")
for (b in c(0.20, 0.10)) for (d in cmp) {
  y <- design(d[[2]], d[[3]], b)
  cat(sprintf("DES|%s|%d|%.0f|%s|%.1f\n", d[[1]], round(100 * (1 - b)), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ","), tail(y$T, 1)))
}

# Sensitivity: what if the true effect is smaller than the observed one? (3 looks, O'Brien-Fleming, 90% power)
cat("\n")
for (fr in c(1, 0.75, 0.5)) {
  y <- design(3, sfLDOF, 0.10, exp(log(hr) * fr))
  cat(sprintf("SENS|%.2f|%.3f|%.0f|%s\n", fr, exp(log(hr) * fr), tail(y$n.I, 1), format(ceiling(sum(y$gamma * y$R)), big.mark = ",")))
}

# Independent cross-check of the boundaries with rpact
r <- getDesignGroupSequential(kMax = 3, alpha = 0.025, beta = 0.10, sided = 1, typeOfDesign = "asOF", informationRates = x$timing)
cat(sprintf("\nCross-check, upper boundaries: gsDesign %s | rpact %s\n", paste(round(x$upper$bound, 3), collapse = ", "), paste(round(r$criticalValues, 3), collapse = ", ")))

Method: LachinFoulkes 
    Analysis               Value Efficacy
   IA 1: 33%                   Z   3.7103
      N: 224         p (1-sided)   0.0001
  Events: 66        ~HR at bound   0.4011
   Month: 20    P(Cross) if HR=1   0.0001
             P(Cross) if HR=0.63   0.0338
   IA 2: 67%                   Z   2.5114
      N: 276         p (1-sided)   0.0060
 Events: 132        ~HR at bound   0.6458
   Month: 31    P(Cross) if HR=1   0.0060
             P(Cross) if HR=0.63   0.5603
       Final                   Z   1.9930
      N: 276         p (1-sided)   0.0231
 Events: 198        ~HR at bound   0.7533
   Month: 60    P(Cross) if HR=1   0.0250
             P(Cross) if HR=0.63   0.9000



Cross-check, upper boundaries: gsDesign 3.71, 2.511, 1.993 | rpact 3.71, 2.511, 1.993

15. Analyses Reported by the Sponsor

The sponsor’s reported primary analysis is a log-rank comparison with a hazard ratio of 0.63, a 95% confidence interval from 0.53 to 0.75, and p < 0.001. The registry’s comparison details give the same estimate and confidence limits. The discrepancy indicator is false, so there is no flagged disagreement between the posted comparison and the sponsor’s reported analysis.

For time-to-event outcomes, the interpretation uses the sponsor-reported hazard ratios rather than an unadjusted reconstruction from group medians. A median alone does not contain enough information to recover a hazard ratio or the full timing of events and censoring. The sponsor’s log-rank analysis uses the time-to-event data described in the record, whereas a recalculation from summary values would not reproduce that analysis.

General statistical explanation: summary-statistic recalculations can be useful checks when they are explicitly identified as unadjusted, but they do not replace a trial’s reported analysis when the two differ. In this record, there is no indicated discrepancy to resolve. The reported analysis should be read with its stated method, confidence interval, p-value, endpoint definition, and censoring description. Further model details and sensitivity analyses are not reported in the registry record.

16. Multiplicity, Missing Data and Analysis Populations

The registry identifies overall survival as the primary outcome and lists several secondary outcomes, but it does not report whether or how the statistical plan controlled for multiplicity across endpoints. This matters because repeated statistical tests can increase the chance of at least one result appearing notable even when there is no underlying effect for that outcome. General statistical explanation: procedures such as a prespecified testing order or adjusted significance thresholds can address this issue, but no such procedure should be attributed to this trial unless it is documented.

Interim analyses and any associated stopping rules are not reported in the registry record. The record also does not provide the original statistical analysis plan, the rationale for endpoint hierarchy, or details of whether reported secondary analyses were prespecified. Consequently, the posted secondary p-values are most safely read as endpoint-specific results rather than as a fully documented family of confirmatory tests.

The primary endpoint description states that the intention-to-treat population included all randomized participants. It defines survival from randomization to death from any cause and describes right-censoring for participants lost to follow-up or not known to have died by the data cut-off, including participants known to have died after the cut-off. This is the missing-observation information explicitly described for the primary time-to-event endpoint.

Other missing-data details are not reported in the registry record. The record does not provide a complete accounting of missing endpoint observations, reasons for missingness by group, sensitivity analyses, or imputation methods for the outcomes. The primary arm sizes do not establish that every participant had complete outcome follow-up. The stated censoring approach allows incomplete event observation to contribute follow-up time, but does not remove the need to consider assumptions about censoring.

The registry defines the primary analysis population, but does not give the same level of analysis-population detail for every secondary outcome. It also does not report whether additional populations or per-protocol analyses were used. These gaps limit independent assessment of how multiplicity and incomplete outcome data were handled beyond the descriptions explicitly recorded.

17. 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: spinal cord compression (6.8%), general physical health deterioration (6.0%), metastatic pain (6.0%), anaemia (4.0%), cardiac failure (4.0%).

The registry reports serious adverse event counts of 319 among 800 participants in the enzalutamide double-blind plus open-label extension group, and 155 among 399 participants in the placebo double-blind group. It also lists a placebo double-blind to enzalutamide open-label extension group with 25 serious adverse events among 50 participants. The group labels describe different study phases and should be retained when reading these figures; they are not a single, uniform comparison across the randomized period.

Other adverse events are reported for 784 of 800 participants in the enzalutamide group, 385 of 399 in the placebo group, and 47 of 50 in the placebo-to-enzalutamide extension group. These are counts of participants with events in the registry’s adverse-event groupings, not counts of distinct events. The registry lists spinal cord compression at 6.8%, general physical health deterioration at 6.0%, metastatic pain at 6.0%, anaemia at 4.0%, cardiac failure at 4.0%, and convulsion at 4.0% among the top serious terms.

These figures do not establish that an event was caused by a study drug. The registry does not report death counts in the adverse-event results shown, and the different phase definitions and group sizes limit direct interpretation as a like-for-like randomized safety comparison. Further details on event timing, exposure duration, and attribution are not reported in the registry record.

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

The primary result describes a randomized comparison of overall survival in participants with castration-resistant prostate cancer previously treated with docetaxel-based chemotherapy. The posted median survival values and hazard-ratio estimate indicate a difference between the groups in the direction of lower mortality hazard in the enzalutamide group during the analyzed follow-up. The confidence interval quantifies statistical uncertainty around the reported relative estimate, while the median confidence limits describe uncertainty around the group medians.

This finding does not mean that every participant in one group had a better outcome than every participant in the other. Group summaries cannot determine an individual’s outcome. A hazard ratio is also not a direct statement about how many months any particular person may live, and it is not equivalent to a difference between the median values.

The result does not, by itself, answer every question about safety, quality of life, treatment burden, or outcomes beyond those measured and reported. The registry includes adverse-event counts and a set of secondary endpoints, but does not report complete numerical results for every listed secondary outcome or all details of their statistical handling. The serious-event groupings also span different phases, so they should not be treated as a single simple randomized comparison.

General statistical explanation: a statistically strong result and a clinically important result are related but distinct ideas. Statistical evidence concerns how compatible observed data are with a null comparison under a model and its assumptions; clinical importance requires interpreting the size and relevance of the effect in context. The registry statistics support an interpretation of a group-level difference in the recorded endpoint, but do not establish the magnitude of importance for every person or resolve questions not addressed in the record.

Applicability is also bounded by the population studied: participants had castration-resistant prostate cancer and prior docetaxel-based chemotherapy. The registry results should not automatically be extended to populations with different disease or treatment histories. The record does not report enough detail to assess every aspect of generalizability.

19. Why This Trial Matters Statistically

This trial is statistically notable because it reports a randomized, placebo-controlled comparison using overall survival as its primary endpoint. The endpoint is time to death from any cause, and the analysis accounts for differing follow-up through censoring as described in the registry. This makes the central result a time-to-event comparison rather than a simple comparison of event counts at one fixed time.

The report gives both group medians and a hazard ratio with a confidence interval. These summaries serve different purposes: the medians describe the survival distributions in terms of time, while the hazard ratio summarizes the relative event rate over follow-up. Reading both prevents a common mistake: treating a hazard ratio as if it were a direct ratio or difference of median survival.

The registry also illustrates why statistical interpretation requires more than a p-value. The confidence interval conveys the range of estimates compatible with the data under the analysis, while the endpoint definition and censoring description clarify what was measured. The stated intention-to-treat population links the primary analysis to randomized assignment. These features help readers evaluate the result without confusing statistical precision with individual prediction or clinical importance.

Finally, the record highlights limits that matter in a broad outcome program. It lists multiple secondary measures, but does not report a multiplicity procedure, interim analyses, or the original sample-size justification. The posted secondary results can therefore add context, while their evidential weight should be considered in light of those missing details. The sample-size illustrations based on observed effects serve a separate planning purpose and are not evidence of power achieved in the completed trial.

20. 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 registry record provides a useful account of the primary endpoint and its main result, but does not report the full statistical analysis plan. Details such as stratification, covariate adjustment, proportional-hazards assessment, interim analyses, and stopping rules are not reported in the registry record. Their absence limits an independent review of the complete analysis strategy and should not be filled by assumptions about common trial practice.

For overall survival, the record explains the censoring rule but does not provide event counts, follow-up completeness by group, patterns of loss to follow-up, or sensitivity analyses. These details would help assess the robustness of the time-to-event estimate. The upper confidence limit for the enzalutamide median is not reported in the registry record, while the placebo median has reported limits.

The registry lists multiple secondary outcomes and gives numerical details for some, but detailed outcome estimates and methods for every listed endpoint are not reported in the registry record. No multiplicity procedure is reported, which limits how strongly secondary p-values can be treated as confirmatory evidence. The record also does not report an original sample-size justification, so observed-effect planning illustrations should not be mistaken for the trial’s design calculations.

The safety groups combine or distinguish different study phases, including an open-label extension, and their denominators differ. The registry does not provide death counts in the adverse-event results shown or enough details about exposure time and event attribution to interpret these counts as a comprehensive comparative safety analysis.

Finally, the results pertain to the study population described in the record: participants with castration-resistant prostate cancer previously treated with docetaxel-based chemotherapy. The registry information alone does not establish how closely the results apply to people outside that population.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: Safety and Efficacy Study of MDV3100 in Patients With Castration-Resistant Prostate Cancer Who Have Been Previously Treated With Docetaxel-based Chemotherapy — NCT00974311.
• Registry record last updated 2018-12-11; status last verified 2018-11.
• Results first posted 2012-10-30.
• 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.

24. Record Summary

The registry describes a completed phase 3, randomized, parallel, triple-masked study comparing enzalutamide, formerly MDV3100, with placebo in participants with castration-resistant prostate cancer previously treated with docetaxel-based chemotherapy. Enrollment was 1199. The primary endpoint was overall survival, defined from randomization to death from any cause and analyzed in the intention-to-treat population.

Median overall survival was 18.4 months for enzalutamide and 13.6 months for placebo. The reported hazard ratio was 0.63, with a 95% confidence interval from 0.53 to 0.75 and p < 0.001. The registry describes right-censoring for specified participants without an observed death by the analysis cut-off.

The record lists secondary outcomes and serious adverse-event information, but does not report a multiplicity procedure, interim analyses, or the trial’s original sample-size justification. Planning illustrations based on observed effects are distinct from that original design information.