Immunotherapy With APC8015 (Sipuleucel-T, Provenge) for Asymptomatic, Metastatic, Hormone-Refractory Prostate Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Immunotherapy With APC8015 (Sipuleucel-T, Provenge) for Asymptomatic, Metastatic, Hormone-Refractory Prostate Cancer (NCT01133704), a randomized phase 3 study. The primary outcome is Overall Time to Disease Progression.

2. Clinical Question

3. Trial Design

This was a phase 3, randomized, parallel-group trial comparing sipuleucel-T (APC8015) with placebo. The registry describes the trial as double blind and records quadruple masking. These labels indicate that masking was intended to extend beyond participants alone, but the identities of the masked roles and the practical details of masking are not reported in the registry record.

The registry lists enrollment as 98. The primary outcome results are reported for 65 participants in the sipuleucel-T arm and 33 in the placebo arm. The record does not explain the relationship between enrollment and the outcome-specific analysis populations, or whether everyone enrolled contributed to each analysis. The reasons for any difference in analysis populations are not reported in the registry record.

The trial record lists Dendreon as sponsor and the status as completed. The recorded start was 2000-05 and primary completion was 2003-03. These dates describe the registry record; they do not establish when individual participants entered or completed follow-up.

The primary outcome was overall time to disease progression. Overall survival was listed as a secondary outcome. The registry gives time-to-event estimates and log-rank hazard ratios for both outcomes. It does not report the detailed protocol, the endpoint assessment procedures, allocation concealment, the treatment schedule, or the trial's original sample-size justification. Those details 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: Participants were screened for evaluation of subject eligibility and performance of baseline tests/procedures.

Overall Study

Reasons for not completing this period:

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 Overall Time to Disease Progression, measured over from randomization to 36 months.

The primary endpoint was overall time to disease progression, measured from randomization to disease progression, with the stated timeframe extending to 36 months. A time-to-event endpoint records not only whether an event occurred, but also how long participants remained without that event during follow-up. This can use information about event timing rather than reducing follow-up to a simple event count.

The registry reports the median time to progression for each arm and a 95% confidence interval for each median. A median is the point by which half of the analyzed group has experienced the event and half has not, under the time-to-event estimation. It is a summary of the distribution, not a description of every participant's experience. The record does not state the estimator used for the arm-specific medians or explain how censoring was handled.

The registry also reports a log-rank hazard ratio for the comparison. In general, a hazard ratio summarizes the relative event rates over follow-up under the assumptions of the analysis; it is not a difference between the two medians. The endpoint is useful for comparing event timing, but interpretation depends on the analysis population, censoring, event definition, and modeling assumptions. The detailed procedures for these matters are not reported in the registry record.

7. Results Posted in the Registry

• Sipuleucel-T (APC8015) vs Placebo: hazard ratio 1.09 (95% CI 0.69–1.70), reported p = 0.719 (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.

Overall 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:
  • Sipuleucel-T (APC8015) vs Placebo: hazard ratio 1.27 (95% CI 0.78–2.07), reported p = 0.331 (as reported in the registry).
• Interpretation:
  • The 95% confidence interval includes no difference (a ratio of 1), so these data do not show a statistically significant difference between Sipuleucel-T (APC8015) vs Placebo and the control. The interval gives the range of effects that remain compatible with the data.

Overall survival is the secondary endpoint reported alongside the primary disease-progression outcome. It adds a different time-to-event perspective: it concerns survival time rather than time to disease progression. In broad terms, the registry's estimates have the same interpretive feature as the primary result: their hazard-ratio confidence intervals include 1 and leave substantial uncertainty about the relative event pattern. The secondary endpoint does not resolve the uncertainty in the primary comparison.

Secondary outcomes can provide useful context, but they are generally less conclusive than a prespecified primary outcome when many endpoints or analyses are considered. Testing multiple outcomes increases the chance that at least one result will appear notable by chance if no adjustment is made. The record does not report a multiplicity adjustment or a plan for interpreting secondary tests in relation to the primary endpoint. It is therefore not possible to determine from the registry how the sponsor controlled the overall false-positive risk.

The secondary result should not be treated as a separate confirmation simply because it concerns a related clinical outcome. Outcomes within the same trial may be correlated, and their estimates may rely on overlapping participants and follow-up. The registry record does not report the correlation between endpoints or an analysis accounting for it. Overall survival is informative as a registered secondary measure, but it carries less weight for a definitive claim than the trial's primary endpoint and should be interpreted with its uncertainty and the absent multiplicity details in mind.

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

• Overall Survival

9. Statistical Interpretation

10. Statistical Methods Explained

Both posted outcome comparisons are time-to-event analyses. The registry reports arm-specific medians with 95% confidence intervals and a log-rank hazard ratio for the comparison. The log-rank method is a general approach to comparing event-time distributions between groups. It uses the timing of events and the risk sets over follow-up, including censored observations according to the analysis procedure. The registry does not report details of censoring rules or the number and timing of events.

What does a median time-to-event estimate describe?

In general, a median event time is the time at which the estimated event-free proportion reaches the halfway point. It offers an interpretable summary of when events occur, but it does not describe the full shape of the event-time distribution. Two groups can have similar medians while differing elsewhere in follow-up, or have different medians while their uncertainty remains substantial. A confidence interval around a median communicates uncertainty about its location. The interval is not a range containing a stated proportion of individual participants' event times.

What does the hazard ratio add?

A hazard is an event rate at a particular time among participants still at risk at that time. In general, a hazard ratio compares the hazards between groups across follow-up, according to the model or test used. A value of 1 is the usual reference for equal hazards. The ratio is not a ratio of median times, and it should not be translated into an absolute difference in weeks or months. Interpretation of which arm's hazard is in the numerator depends on the coding and reference convention. Although the registry names the comparison as sipuleucel-T versus placebo, the detailed coding convention is not reported in the registry record.

The log-rank test evaluates evidence for a difference in the event-time patterns under its assumptions. Its p-value is not the probability that the null hypothesis is true, nor does it quantify the size or clinical importance of an effect. A large p-value can arise when estimates are close to the null, when uncertainty is wide, or both. A small p-value would not by itself establish the size or practical importance of a difference.

How should confidence intervals and p-values be read together?

A confidence interval gives a range of estimates compatible with the data and the method at the stated confidence level. In general, a 95% confidence procedure has a long-run coverage interpretation across repeated studies under its assumptions; it is not a 95% probability statement about the fixed true value after observing this particular interval. The reported p-value and confidence interval are related summaries of uncertainty, but they answer different questions. Neither replaces consideration of the estimate's magnitude or the study's design and conduct.

The primary comparison reports a hazard ratio of 1.09, a 95% confidence interval from 0.69 to 1.70, and p = 0.719. The secondary comparison reports a hazard ratio of 1.27, a 95% confidence interval from 0.78 to 2.07, and p = 0.331. Both intervals include 1, and both extend across values that allow a range of possible relative hazard patterns. The record gives no adjusted analysis or alternative model for either endpoint.

Time-to-event methods are designed to make use of differing follow-up times, but their validity depends on how follow-up and censoring are handled. If participants leave observation or reach the end of follow-up without the event, their event times are censored rather than treated as if the event occurred at that point. In general, informative loss to follow-up can bias an analysis if censoring is related to event risk in ways not handled by the method. The registry does not report censoring counts, reasons for withdrawal, or analyses of sensitivity to missing outcome data.

The log-rank comparison and the reported hazard ratio are sponsor-reported analyses in the registry. The record does not state whether proportional hazards were assessed or whether the hazard ratio was intended as a constant relative effect over time. If hazards change relative to one another during follow-up, a single hazard ratio can be a limited summary. Whether that issue arose in this trial is not reported in the registry record.

Finally, the posted estimates are based on the analysis populations represented in the outcome records. The registry lists group sizes for the outcomes, but detailed analysis-population definitions and participant flow are not reported in the registry record. Such information matters because exclusions, missing outcomes, and deviations from assigned treatment can affect how a randomized comparison should be interpreted.

11. Confidence Intervals and Precision

Precision concerns how narrowly the data locate an estimate. The primary median estimates have 95% confidence intervals from 9.3 to 17.7 weeks for sipuleucel-T and from 8.4 to 18.0 weeks for placebo. Both intervals cover a range of plausible median times, rather than pinpointing a single value. The considerable overlap is consistent with uncertainty in comparing the arm-specific medians; overlap alone is not a formal test of the treatment contrast.

The primary hazard ratio is 1.09, with a 95% confidence interval from 0.69 to 1.70. Because this interval includes 1 and extends to values on either side, the result does not precisely determine the relative hazard pattern. The p-value of 0.719 is also not evidence that the groups are identical. It indicates that the observed comparison is not unusual under the null model and the test procedure, while the interval shows that a broad range of relative effects remains compatible with the data.

For overall survival, the median confidence intervals are 13.6 to 31.9 months for sipuleucel-T and 12.8 to 25.4 months for placebo. The hazard ratio confidence interval, from 0.78 to 2.07, is also broad and includes 1. These intervals signal uncertainty, not a direct probability that one treatment is better or worse.

Confidence intervals address sampling uncertainty under the analysis assumptions. They do not account automatically for biases from missing data, outcome measurement, protocol deviations, or selective reporting. The registry record does not provide enough detail to evaluate those sources of uncertainty. The estimates are therefore best read alongside the design and reporting limitations, rather than as exact descriptions of the effects.

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 time-to-event outcomes, the entries are the total number of events needed, not participants per arm. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to a stated fraction of the observed effect.

The 80% power column describes the event total that the illustration associates with an 80% chance of detecting the specified effect under the stated assumptions, if that effect and the assumptions were true. The 90% power column uses a higher target chance of detection and therefore requires a larger event total. These are planning targets, not guarantees about what a future study would find.

Rows describe planning for 100%, 75%, or 50% of the observed effect. A smaller effect is harder to distinguish from random variation and requires more events for the same target power. The corresponding entries should be read as hypothetical planning scenarios, not as observed outcomes from this trial. They are not observed power.

The registry record does not report the original sample-size calculation or its assumptions. The planning illustration cannot establish that the trial was designed to detect the effect ultimately observed, nor can it determine whether follow-up yielded enough events for a particular target. For time-to-event designs, event accrual is central: the needed event total is distinct from a participant count because participants may be followed for different durations and may not all experience the event during observation.

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 = 1.090, so ln(HR) = 0.0862.
2. d = 4 × (1.96 + 0.842)² / [ln(1.090)]² = 4,227.4, rounded up to 4,228 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 <- 1.09                              # 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(1133704)
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 = 4,228 (80% power), 5,660 (90% power)
 75% of observed effect: events = 7,516 (80% power), 10,062 (90% power)
 50% of observed effect: events = 16,910 (80% power), 22,638 (90% power)
Simulated power with 4,228 events: 0.807 (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

The registry reports the sponsor's log-rank analyses for the primary and secondary time-to-event outcomes. For disease progression, the sponsor-reported hazard ratio is 1.09, with a 95% confidence interval from 0.69 to 1.70 and p = 0.719. For overall survival, it is 1.27, with a 95% confidence interval from 0.78 to 2.07 and p = 0.331. These are the reported comparison estimates, not newly reconstructed estimates.

The record indicates no discrepancy between the reported analysis and the posted results. No separate recalculated hazard ratio is identified for these outcomes, and no alternative estimate should be inferred from the arm-specific medians. A hazard ratio cannot be recovered by simply comparing those medians.

The registry does not report enough detail to reproduce the sponsor's log-rank analyses, including full event data, censoring information, and the exact analysis-population definitions. Accordingly, the sponsor's reported estimates are the appropriate analysis results to describe from this record. Their interpretation remains subject to the confidence intervals and limitations; the fact that an analysis is sponsor-reported does not remove statistical uncertainty.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates more than one outcome, comparison, subgroup, or analysis. In general, as the number of statistical tests increases, the chance of at least one apparently unusual result can increase even if there is no underlying effect. Researchers may address this by prespecifying a testing hierarchy, adjusting significance thresholds, or treating secondary analyses as exploratory. The registry lists overall time to disease progression as primary and overall survival as secondary, but a multiplicity strategy is not reported in the registry record.

The registry provides outcome-specific group sizes and results, but it does not describe a complete analysis population, participant flow, or rules for excluding observations. Whether the analyses followed an intention-to-treat principle, a per-protocol approach, or another population definition is not reported in the registry record. The record also does not report missing-data methods, censoring counts, reasons for loss to follow-up, or sensitivity analyses. These omissions limit assessment of how robust the estimates are to incomplete follow-up or analysis choices.

Interim monitoring and stopping rules are not reported in the registry record. It is therefore not possible to assess whether repeated looks at accumulating outcome data affected the interpretation or the nominal p-values. The protocol, statistical analysis plan, and detailed endpoint adjudication procedures are also not reported in the registry record.

These are reporting limitations, not evidence that a particular analysis choice was made incorrectly. They mean that the posted estimates cannot be fully evaluated for multiplicity control, missing-data assumptions, or population selection from the registry information alone. Interpretation should remain anchored to the results that are reported, with appropriate caution about what the record does not establish.

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: deep vein thrombosis (6.5%), chills (4.6%), anaemia (3.2%), cardiac failure congestive (3.2%), supraventricular tachycardia (3.2%).

The registry reports serious adverse events for groups with 65 participants at risk in the sipuleucel-T arm and 31 at risk in the placebo arm. Serious adverse events affected 13 in the sipuleucel-T group and 9 in the placebo group. Other adverse events affected 63 and 29 participants, respectively, among the listed groups at risk. These counts describe the registry's safety reporting and should not be compared as if they were efficacy outcomes.

The listed serious adverse event terms include deep vein thrombosis at 6.5%, chills at 4.6%, and anaemia, congestive cardiac failure, supraventricular tachycardia, and abdominal pain at 3.2% each. The record does not specify the group attribution for these term percentages, so they should not be assigned to an arm. The registry also gives no death counts in the safety record.

Adverse-event counts do not by themselves establish causation. The record does not report event timing, severity grading beyond the serious-event category, attribution to treatment, or exposure-adjusted rates. Nor does it explain why the safety group at risk differs from the placebo efficacy group size. These safety figures are descriptive, and their interpretation is limited by the detail available in the registry.

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

The posted results describe a randomized comparison of sipuleucel-T and placebo for asymptomatic, metastatic, hormone-refractory prostate cancer. The primary endpoint was time to disease progression, and the registry also reports overall survival. For both outcomes, the hazard-ratio confidence interval includes 1, and the reported p-values do not provide strong evidence against the null comparison under the stated log-rank analyses.

That pattern is not proof that the treatments have identical effects. The confidence intervals retain a range of possible relative hazard values, and the record does not report enough detail to evaluate missing data, censoring, analysis populations, or multiplicity. Conversely, the arm-specific medians should not be read alone as proof of a treatment effect. The medians summarize each group's event-time distribution, while the log-rank comparison uses information across follow-up.

The results apply to the population and trial conduct represented in this registry record. Eligibility details beyond the population description, treatment procedures, event assessment, and follow-up completeness are not reported in the registry record. The record therefore cannot establish how well the results extend to people outside the described population or to settings with different care and follow-up.

Statistical nonsignificance is not equivalent to equivalence or noninferiority. The registry does not report an equivalence or noninferiority objective, a margin, or a corresponding analysis. Nor do these estimates provide an individual prediction. They are group-level trial summaries with uncertainty. Any clinical conclusions require consideration of the full evidence and context beyond this registry record; these statistics alone are not medical advice.

18. Why This Trial Matters Statistically

This trial is statistically instructive because it reports both arm-specific median times and comparative hazard ratios for time-to-event outcomes. Those summaries answer related but distinct questions. Medians locate the midpoint of each event-time distribution, whereas the hazard ratio and log-rank test compare event patterns across follow-up. Reading both helps avoid treating a difference between medians as if it were the full comparative analysis.

The result also illustrates why uncertainty matters as much as the point estimate. A hazard ratio close to 1 does not establish equality, particularly when the confidence interval spans values below and above 1. A p-value above a conventional threshold does not show that no effect exists; it indicates that the analysis did not produce strong evidence against its null comparison. The confidence intervals communicate the remaining uncertainty more directly than a binary significant/nonsignificant label.

The secondary survival outcome adds context but raises the importance of multiplicity and prespecification. Without a reported testing hierarchy or adjustment, secondary findings should not be treated as independent confirmation. Likewise, the planning illustrations based on observed effects are not substitutes for the trial's original design assumptions, which are not reported in the registry record.

Finally, a registry result is only as interpretable as its reporting detail allows. Analysis populations, censoring, missing data, and interim monitoring can influence time-to-event interpretation, yet these details are not reported in the registry record. Recognizing both the reported estimates and these gaps is essential to a careful statistical reading.

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 fuller assessment are not reported in the registry record. The arm descriptions are blank, so treatment schedules, dose or administration details, and placebo procedures cannot be characterized. The record names quadruple masking but does not identify which roles were masked or how masking was maintained. Eligibility criteria and the disease-progression assessment procedures are also not reported.

The outcome records list group sizes, but the analysis-population definitions and participant disposition are not reported. The registry does not describe the number of events, censoring, follow-up completeness, withdrawal reasons, or missing-data handling. These omissions matter for time-to-event results because incomplete follow-up and censoring assumptions can affect the estimates.

The statistical analysis plan, original sample-size justification, proportional-hazards assessment, multiplicity approach, interim analyses, and stopping rules are not reported in the registry record. The planning figures based on observed effects should not be mistaken for the study's original design calculation. There is also no detailed account of the safety-event definitions, attribution, timing, or arm-specific breakdown of the listed serious-event terms.

These reporting gaps do not demonstrate that the trial lacked such procedures; they prevent assessment from this record. They limit reproducibility and make it difficult to determine how sensitive the results may be to analysis decisions or incomplete observations. The estimates should therefore be interpreted as registry-reported results rather than as a complete account of the trial's methods and conduct.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Immunotherapy With APC8015 (Sipuleucel-T, Provenge) for Asymptomatic, Metastatic, Hormone-Refractory Prostate Cancer — NCT01133704.
• Registry record last updated 2010-09-08; status last verified 2010-09.
• Results first posted 2010-09-02.
• 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

The registry describes a completed, randomized, parallel, placebo-controlled phase 3 trial of sipuleucel-T (APC8015) for asymptomatic, metastatic, hormone-refractory prostate cancer. It records double blinding and quadruple masking. Enrollment is listed as 98, with a primary endpoint of overall time to disease progression and overall survival as a secondary endpoint.

The primary result is a log-rank hazard ratio of 1.09, with a 95% confidence interval from 0.69 to 1.70 and p = 0.719. The secondary survival analysis also has a confidence interval that includes 1. Important details about analysis populations, missing data, censoring, multiplicity, and the original design assumptions are not reported in the registry record.