A Randomized, Double-blind, Phase 3 Efficacy Trial of PROSTVAC-V/F +/- GM-CSF in Men With Asymptomatic or Minimally Symptomatic Metastatic Castrate-Resistant Prostate Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Randomized, Double-blind, Phase 3 Efficacy Trial of PROSTVAC-V/F +/- GM-CSF in Men With Asymptomatic or Minimally Symptomatic Metastatic Castrate-Resistant Prostate Cancer (NCT01322490), a randomized phase 3 study. The primary outcome is Overall Survival.

2. Clinical Question

This was a randomized, double-blind, phase 3 trial sponsored by Bavarian Nordic. The study purpose, as stated in the registry, was to determine whether PROSTVAC alone or with GM-CSF could prolong overall survival in men with few or no symptoms from metastatic castrate-resistant prostate cancer. The results address that question through an overall-survival endpoint, while the separate six-month secondary outcome captures whether participants were alive without an event at a specified early follow-up point. Neither endpoint alone describes every aspect of a participant’s experience or the balance of benefits and harms.

3. Trial Design

The trial is registered as NCT01322490 and is marked completed. It was a phase 3, randomized, parallel-group study with quadruple masking. The registry lists enrollment of 1297 participants. Study start was recorded as 2011-11-28, and primary completion as 2017-09-25. The registry says participants were followed for approximately 6 years from the first subject randomized to study completion.

Three arms are represented in the posted primary results: PROSTVAC-V/F-TRICOM with GM-CSF placebo, PROSTVAC-V/F-TRICOM with GM-CSF, and placebo control. The primary analysis included 432, 432, and 433 randomized subjects, respectively. Randomization supports comparison of assigned groups, while masking is intended in general to reduce the influence of treatment knowledge on care and outcome assessment. The exact procedures used to implement masking, including which roles were masked, are not reported in the registry record.

The primary outcome was overall survival, defined as time from randomization to death from any cause. The registry says that subjects who did not die and did not experience the competing events of definite loss to follow-up or withdrawal of consent were right-censored at their last contact. A secondary outcome counted subjects alive without an event at 6 months, with its timeframe described as randomization through Week 25 or the end-of-treatment visit.

The record lists two primary pairwise log-rank analyses, each comparing a PROSTVAC arm with placebo control. It does not report a direct comparison of the two PROSTVAC arms. Additional protocol-level details, including the randomization procedure, the full analysis plan, and the intended handling of interim looks, 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: Screening activities were completed within 28 days prior to the first dose of any medication and were completed prior to dosing. Subjects who required anti-androgen wash-out specifically for this protocol were consented prior to beginning withdrawal of therapy, however, no other screening procedures were performed the subject was deemed eligible.

Treatment Period

Reasons for not completing this period:

Long Term Follow-up Period

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent to treat, including all randomized subjects.

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, including all randomized subjects.

6. Primary Endpoint

The registry lists the primary outcome as Overall Survival, measured over Randomization through the date of death due to any cause. Subjects were followed up for approximately 6 years from the first subject randomized to the completion of the study..

Overall survival measures the time from randomization until death from any cause. It is a time-to-event endpoint: the outcome is not simply whether a participant died, but how long each participant was observed before death or censoring. The registry defines the calculation as the date of death, competing event, or censoring minus the date of randomization, plus 1. Subjects alive at last contact without the listed competing events were right-censored at that contact date.

In general, overall survival is used in trials because it captures a clearly defined event and allows follow-up times to contribute even when participants have not experienced the event by the end of observation. Censoring preserves those participants’ observed follow-up rather than treating them as if they died at the end of the study. Its interpretation depends on adequate follow-up and on censoring not distorting comparisons in a way the analysis fails to address. The registry’s detailed assessment of those assumptions is not reported.

The posted summaries are medians with 95% confidence intervals. A median is the estimated time by which half of the analyzed population has experienced the event under the time-to-event analysis. It does not describe every participant’s survival time and is not itself a comparison of treatment groups. The confidence interval conveys uncertainty around the estimated median; it does not show the range of individual survival times. The registry reports log-rank tests for the two active-versus-control comparisons, but no primary hazard-ratio estimate or confidence interval.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a time-to-event measure, and the registry reports medians with 95% confidence intervals together with log-rank tests for the two active-versus-control comparisons. These methods answer related but distinct questions. The medians describe the estimated time by which half the group has experienced death; the log-rank test assesses whether the event-time patterns differ between groups across follow-up. The registry does not report a hazard ratio for these primary comparisons.

What does a median survival estimate describe?

In general, the median is the point on an estimated survival distribution at which the estimated proportion remaining alive reaches one half. It is a summary of the distribution, not a prediction for a particular participant and not a statement that every participant lived for the median duration. Because it focuses on a single point, it can omit differences in the timing of events elsewhere in follow-up. Two groups can have similar medians while their survival patterns differ at other times, or different medians without a consistent separation throughout follow-up.

The confidence interval around a median reflects statistical uncertainty in that estimate. In general, a wider interval indicates less precision than a narrower interval, but the interval is not a range containing a fixed proportion of individual survival times. Here, the registry reports 95% confidence intervals for the three arm-specific medians. It does not report confidence intervals for differences between those medians, so overlap or non-overlap of arm-specific intervals is not a substitute for the reported pairwise test.

Why use a log-rank test?

In general, the log-rank test compares the observed timing of events between groups while accounting for participants who are censored during follow-up. It uses event and risk-set information across follow-up, rather than reducing each participant to a simple alive-or-dead status at one fixed time. The registry reports one log-rank test for each PROSTVAC arm against placebo control. It does not report a direct test between the two PROSTVAC arms, a primary hazard ratio, or a confidence interval for a relative treatment effect.

A test result should be interpreted alongside effect estimates and their precision. A p-value is not a measure of how large a difference is, and a result above a chosen significance threshold does not establish equivalence. The registry’s primary results support a description of the median estimates and the reported tests, but not a detailed estimate of the relative rate of death over time. The protocol’s specified significance threshold and any adjustment for the two primary comparisons are not reported in the registry record.

How does censoring enter the analysis?

The registry defines death from any cause as the event and says that subjects without death were right-censored at last contact unless they experienced definite loss to follow-up or withdrew consent, which are identified as competing events in the endpoint description. In general, right-censoring means that the participant contributed observed follow-up up to a known time, after which the event time is unknown. It does not mean that the participant was assumed to remain alive indefinitely. Time-to-event methods can use such partial follow-up, but their validity depends on assumptions about how censoring relates to future event risk.

The posted record does not give arm-specific censoring counts, follow-up distributions, reasons for loss to follow-up, or sensitivity analyses for censoring. It also does not report how missing or incomplete outcome information was handled beyond the endpoint’s stated censoring definition. The registry describes the primary analysis population as intent to treat, including all randomized subjects, but does not provide a participant-level accounting of exclusions or departures from assigned treatment.

What can the secondary re-calculations show?

For the six-month binary outcome, recalculated risk differences are simple unadjusted comparisons based on the posted counts. In general, a risk difference expresses the difference between the proportions meeting an outcome in two groups, with its confidence interval indicating uncertainty. The registry also reports odds ratios for the secondary comparisons. Odds ratios and risk differences use different scales and are not interchangeable; their estimates need not look numerically similar, especially when the event is not rare.

The re-calculated analyses are not replacements for the sponsor’s reported analyses and do not establish a prespecified treatment effect. The secondary endpoint is a count at a particular follow-up window, whereas overall survival incorporates event timing throughout follow-up. Sample-size illustrations based on an observed secondary effect are planning scenarios, not observed power and not proof that the original trial was adequately or inadequately sized. They rely on assumptions stated with the planning illustration; the trial’s own design justification is not reported in the registry record.

10. Analyses Reported by the Sponsor

For overall survival, the registry reports arm-specific median estimates and two sponsor-reported log-rank tests, one for each PROSTVAC regimen versus placebo control. Both reported tests are above the conventional 0.05 threshold. The registry does not give primary hazard ratios or confidence intervals, so the reported comparisons cannot be expressed as a relative death-rate estimate with a stated precision interval. No recalculated primary comparison is posted in the registry record, and there is no discrepancy between a sponsor result and a recalculation to resolve for this endpoint.

For the six-month secondary outcome, the sponsor reports odds ratios with 95% confidence intervals for the two comparisons against placebo. The posted simple recalculations instead express unadjusted risk differences from the arm counts, with corresponding confidence intervals and unadjusted p-values. These are different effect measures and analysis presentations, so their estimates should not be treated as duplicate versions of a single statistic. The sponsor’s intervals for both secondary odds ratios include the null value of 1; the recalculated risk-difference intervals include zero. In general, those intervals indicate that the posted data are compatible with no difference as well as with effects in either direction.

Neither analysis format alone resolves issues such as multiplicity, missing outcome information, or whether the analysis was prespecified. No discrepancy is flagged for the secondary recalculations. The registry’s reported analyses remain the record of the sponsor’s analyses, while unadjusted recalculations offer a transparent additional description of the posted summary counts.

11. Multiplicity, Missing Data and Analysis Populations

The registry identifies overall survival as the primary outcome and lists two primary log-rank comparisons, each between a PROSTVAC arm and placebo control. It also reports one secondary outcome with two comparisons. The record does not state whether the two primary tests were handled under a multiplicity procedure, whether a testing hierarchy was prespecified, or whether the secondary analyses were adjusted for multiple testing. The protocol’s statistical analysis plan and any interim-analysis procedures are not reported in the registry record.

Multiplicity matters in general when more than one hypothesis is tested: the chance of at least one nominally unusual result can rise as the number of tests increases. The appropriate interpretation depends on which outcomes and comparisons were planned, their order, and any adjustments. Those details are not available in the registry record, so the reported p-values should be described as the results of the listed tests rather than as a complete account of confirmatory testing.

The primary population is explicitly described as intent to treat and includes all randomized subjects. This supports interpreting the primary analysis by randomized assignment, but the registry does not provide a full accounting of participants included in each analysis or describe adherence to assigned treatment. The endpoint definition states that participants without an event were right-censored at last contact, with definite loss to follow-up and withdrawal of consent treated as competing events. Counts and reasons for these cases, arm-specific censoring patterns, and sensitivity analyses are not reported.

For the six-month outcome, the record provides arm counts but does not describe how missing status was handled, whether imputation was used, or whether the denominator included every randomized subject. Protocol-specific missing-data methods are not reported. These omissions limit assessment of whether missingness could affect the secondary comparison. The registry also does not report the sample-size justification, assumptions used in the original design, or any interim stopping rules. The planning illustrations associated with the secondary recalculations are not evidence of the trial’s original design rationale or observed power.

12. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: haematuria (1.6%), urinary retention (1.4%), hydronephrosis (1.2%), pulmonary embolism (1.2%), spinal cord compression (0.9%).

The posted safety counts show 56 participants with serious adverse events in each PROSTVAC arm and 53 in placebo control. The corresponding numbers at risk were 429, 429, and 428. For other adverse events, the reported affected counts were 351, 367, and 346, respectively, with the same arm-specific numbers at risk. These are counts, not adjusted comparisons; the registry does not report a statistical test or confidence interval for these differences in the safety summary.

The listed serious adverse-event terms include haematuria at 1.6%, urinary retention at 1.4%, hydronephrosis at 1.2%, pulmonary embolism at 1.2%, spinal cord compression at 0.9%, and atrial fibrillation at 0.7%. These are the most frequently listed serious terms in the posted summary, not a complete description of all events or their timing. The registry also lists deaths of 252, 254, and 248 across the three safety groups. Those counts should not be read as deaths caused by treatment; overall survival is a separate outcome, and the safety summary does not establish causation.

In general, safety counts need context about event definitions, exposure time, severity, relatedness assessment, and completeness of reporting. Such details are not reported in the registry record. The counts alone do not show whether differences are meaningful or attributable to a regimen.

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

The posted primary results describe median overall survival estimates of 34.4 months for PROSTVAC-V/F-TRICOM with GM-CSF placebo, 33.2 months for PROSTVAC-V/F-TRICOM with GM-CSF, and 34.3 months for placebo control. The registry’s two log-rank tests do not show statistically clear separation from placebo control at the conventional 0.05 threshold. The results therefore do not establish a difference in overall survival for either active-versus-control comparison in the reported analysis.

That conclusion is narrower than saying the treatments are identical or that no effect is possible. Failure to identify a statistically clear difference is not an equivalence finding. The registry does not report an equivalence or non-inferiority design, a primary hazard ratio, or confidence intervals for the relative primary effects. Precision about the size of any possible difference is therefore limited by the reported summary information.

The medians are group-level summaries. They do not predict how long a particular person will live, and they do not describe every feature of survival over time. The time-to-event analysis depends on follow-up and censoring, and the registry does not provide detailed censoring counts or sensitivity analyses. The secondary six-month measure is a distinct endpoint and should be treated as supplementary rather than as a substitute for the primary survival comparison.

These findings apply to the randomized trial population and the regimens studied: men described as having asymptomatic or minimally symptomatic metastatic castrate-resistant prostate cancer. They do not establish results for different populations, treatment schedules, combinations, or outcomes. The adverse-event counts add descriptive safety context but do not, by themselves, establish causation or comparative safety. Clinical decisions require evidence and individual assessment beyond the statistical summaries in this registry record.

14. Why This Trial Matters Statistically

This trial is statistically informative because it compares two related PROSTVAC regimens with a shared placebo control using a time-to-event primary outcome. The randomized, parallel-group structure and intent-to-treat primary population provide a clear framework for estimating between-arm differences under the study conditions. Reporting the median survival and confidence interval for each arm makes the absolute time scale visible, while the log-rank analyses address the primary pairwise questions.

It also illustrates why a set of similar arm medians should not be mistaken for a complete treatment-effect analysis. Median survival summarizes one point on each estimated survival distribution; the log-rank test uses event timing across follow-up. Without primary hazard ratios, their confidence intervals, or survival curves in the registry record, readers have less information about the magnitude, precision, and possible variation of relative effects over time. The reported tests do not identify statistically clear separation, but they do not demonstrate equivalence.

The trial also highlights the distinction between primary and secondary evidence. The six-month status measure gives a different, fixed-time view from overall survival. Its unadjusted recalculations and sponsor-reported odds ratios can offer context, but multiplicity and prespecification are not described. More generally, results should be interpreted in light of the analysis plan, missingness, censoring, and the number of comparisons, not just the most notable estimate.

Finally, the record shows the value and limits of registry results reporting. It includes core design features, arm-level outcomes, and selected safety counts, but does not report several details needed for a full statistical audit, including the sample-size justification, interim analyses, multiplicity procedure, and detailed missing-data assessment. Recognizing both the strengths and the reporting gaps supports a more calibrated reading of the evidence.

15. 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 limitations are specific to the available record. The registry reports the primary medians and their confidence intervals, plus log-rank p-values, but no primary hazard ratios or confidence intervals for the treatment comparisons. This leaves the relative size and precision of the primary effects less directly characterized. Arm-specific median intervals overlap, but that observation is not a formal test of the difference between arms.

The record gives an endpoint definition and censoring rule, but not detailed numbers censored, the reasons for censoring, follow-up distributions by arm, or sensitivity analyses. These omissions limit assessment of how incomplete follow-up may affect the time-to-event results. The intent-to-treat population is stated, but a detailed participant flow and any deviations from assigned treatment are not reported.

The statistical plan is not described in enough detail to determine whether the two primary comparisons were adjusted for multiplicity or how the secondary outcome fit into a prespecified testing hierarchy. Interim analyses, stopping rules, the trial’s sample-size justification, and the assumptions behind its original design are not reported in the registry record. Consequently, the separate planning illustrations based on an observed secondary effect should not be confused with the trial’s own design calculations.

For the secondary endpoint, the posted summary does not describe the event definition in detail or explain how missing six-month status was handled. Sponsor odds ratios and unadjusted recalculated risk differences are different measures and may not align numerically. The safety summary provides counts and selected serious-event terms, but does not provide detailed exposure-adjusted rates, relatedness assessments, or a complete account of event timing. These gaps constrain how far the posted statistics can support causal or comparative safety conclusions.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Randomized, Double-blind, Phase 3 Efficacy Trial of PROSTVAC-V/F +/- GM-CSF in Men With Asymptomatic or Minimally Symptomatic Metastatic Castrate-Resistant Prostate Cancer — NCT01322490.
• Registry record last updated 2019-09-04; status last verified 2019-08.
• Results first posted 2019-08-08.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

NCT01322490 is a completed randomized, double-blind, phase 3 parallel-group trial sponsored by Bavarian Nordic. It enrolled 1297 men with asymptomatic or minimally symptomatic metastatic castrate-resistant prostate cancer, comparing PROSTVAC-V/F-TRICOM with GM-CSF, PROSTVAC-V/F-TRICOM with GM-CSF placebo, and placebo control.

The primary endpoint was overall survival from randomization to death from any cause, analyzed in the intent-to-treat population. The registry reports arm-specific median estimates and 95% confidence intervals, with log-rank tests for each active regimen versus placebo. Neither reported primary test identifies statistically clear separation at the conventional 0.05 threshold. A secondary outcome counted participants alive without event at 6 months; sponsor-reported odds ratios and simple unadjusted recalculations provide supplementary evidence, not a replacement for the primary analysis.

The registry also posts selected safety counts and serious-event terms. Details about multiplicity, the original sample-size rationale, interim analyses, and several missing-data and censoring diagnostics are not reported in the registry record.