Study Of Gemcitabine Plus AG-013736 Versus Gemcitabine For Advanced Pancreatic Cancer.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study Of Gemcitabine Plus AG-013736 Versus Gemcitabine For Advanced Pancreatic Cancer. (NCT00471146), a randomized phase 3 study. The primary outcome is Overall Survival (OS).

2. Clinical Question

3. Trial Design

This was a randomized, double-blind, parallel-group phase 3 study. It compared axitinib, also called AG-013736, plus gemcitabine with placebo plus gemcitabine for first-line treatment of locally advanced, unresectable or metastatic pancreatic cancer. The registry lists an enrollment of 630. The reported primary analysis included 314 participants assigned to axitinib plus gemcitabine and 316 assigned to placebo plus gemcitabine.

Randomization assigns participants to treatment groups to support a comparison that is less vulnerable to systematic differences between groups. Double masking means the trial was recorded as double-blind, although the registry record does not describe which people were masked or how masking was maintained. A parallel design means participants were assigned to a group rather than being described as switching between study treatments.

The primary endpoint was overall survival in the intent-to-treat population. Participants remained analyzed according to their initial randomization, regardless of whether they received study drug or a different drug. This preserves the randomized comparison as defined in the registry. Overall survival was assessed from randomization until death due to any cause, with the stated timeframe extending until death or at least 1 year after the randomization of the last participant.

The registry lists progression-free survival, objective response, duration of response, quality-of-life measures, pain, and health-state measures among the secondary outcomes. Results are posted for progression-free survival and objective response. Numerical results for the other listed secondary outcomes are not reported in the registry record. The study is marked completed. The registry gives a start of 2007-07 and a primary completion of 2009-01; these are registry dates, not a description of follow-up for each participant.

The trial's original sample-size justification, treatment schedules, and detailed procedures for follow-up are not reported in the registry record. The sample-size illustrations use the observed effects and should not be mistaken for the trial's original design calculations.

4. Participant Flow and Accounting

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

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent-to-treat (ITT) population included all participants who were randomized, with study drug assignment designated according to initial randomization, regardless of whether participants received study drug, or received a different drug from that to which they were randomized.

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 Survival (OS), measured over Baseline until death or at least 1 year after the randomization of last participant.

Overall survival measures the time from randomization to death from any cause. In this study, the registry defines the time in weeks as the death date minus the randomization date plus 1, divided by 7. Death status was determined from adverse-event data when the outcome was death, or from follow-up contact data when the participant's current status was death.

Time to death is used as a clinical trial endpoint because it records an event that is clearly defined and does not depend on a judgment that a tumor has progressed or responded. Measuring from randomization also gives the groups a common starting point for comparison. Because deaths can occur at different times and some participants may not have died by the end of observation, survival analysis accounts for the timing of events and for incomplete follow-up under its analysis assumptions.

The registry reports the median survival time for each group, with a 95% confidence interval, and a hazard ratio comparing the groups. The medians describe the time by which the estimated survival distribution reaches its midpoint. The hazard ratio summarizes the relative event rate over follow-up under the time-to-event analysis; it is not a direct difference in median weeks. The primary analysis used the intent-to-treat population, which keeps participants in the group to which they were randomized.

Overall survival was the primary endpoint, so it anchors interpretation of the trial. A secondary result, including one based on tumor response or progression, addresses a different outcome and cannot replace the primary survival comparison.

7. Results Posted in the Registry

• Axitinib + Gemcitabine vs Placebo + Gemcitabine: hazard ratio 1.01 (95% CI 0.79–1.31), reported p = 0.544 (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.

Progression Free Survival (PFS)

• 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:
  • Axitinib + Gemcitabine vs Placebo + Gemcitabine: hazard ratio 1.01 (95% CI 0.78–1.30), reported p = 0.520 (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 Axitinib + Gemcitabine vs Placebo + Gemcitabine and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Participants With Objective Response (OR)

• 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:
  • Axitinib + Gemcitabine: 4.8% (15/314) versus 1.6% (5/316) on control; risk difference 3.2 percentage points (95% CI 0.5 to 5.9), risk ratio 3.02 (1.11–8.21); unadjusted two-proportion z-test p = 0.022.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Axitinib + Gemcitabine than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The posted secondary results cover progression-free survival and objective response. The pattern is not uniform across endpoint types: the time-to-event comparison for progression-free survival is close to the primary overall-survival comparison, while the objective-response summaries point in a different direction. These endpoints measure different features of disease and treatment experience, so a difference in their estimates is not inherently inconsistent. Response does not measure how long participants live, and progression-free survival is not the same outcome as overall survival.

Neither secondary result changes the interpretation of the primary endpoint. Overall survival was designated as primary, and its reported hazard-ratio interval includes the no-difference reference. A secondary finding cannot establish a primary survival effect merely because its estimate or p-value appears more favorable. The sponsor's objective-response result is also uncertain, as reflected in its confidence interval.

Many secondary outcomes are listed, including duration of response, quality-of-life scores, pain, and health-state measures. Numerical findings for those listed measures are not reported in the registry record. The record lists 10 outcome measures and nine posted secondary outcomes, but the numerical results discussed in the available record do not cover every named outcome.

Multiple outcomes create a multiplicity concern: the more comparisons examined, the greater the chance that at least one result may appear notable by chance alone. This is a general statistical issue, not evidence that any specific result is false. Adjustment procedures or a prespecified hierarchy can help control interpretation across multiple tests, but the multiplicity strategy is not reported in the registry record. Accordingly, unadjusted secondary results carry less weight than the prespecified primary analysis and should be treated as supportive or exploratory rather than conclusive.

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

• Progression Free Survival (PFS)
• Percentage of Participants With Objective Response (OR)
• Duration of Response (DR)
• Change From Baseline in European Organization for Research and Treatment of Cancer, Quality of Life Questionnaire Core-30 (EORTC QLQ- C30) Score
• Change From Baseline in European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-Pancreatic 26 (EORTC QLQ- PAN26) Score
• Change From Baseline Brief Pain Inventory-short Form (BPI-sf) Score
• Change From Baseline in Euro QoL Questionnaire- 5 Dimension (EQ-5D) Health State Profile
• Change From Baseline in Euro QoL Questionnaire- 5 Dimension (EQ-5D) VAS Score

9. Statistical Interpretation

10. Statistical Methods Explained

Overall survival and progression-free survival are time-to-event outcomes: the analysis uses both whether an event occurred and when it occurred. For overall survival, the event is death from any cause. The registry describes progression-free survival as measured until disease progression or the stated follow-up boundary. Time-to-event methods are useful when participants have different lengths of observation or when some observations end before an event is recorded.

What does the hazard ratio summarize?

A hazard ratio compares the event hazards of the two groups over follow-up. In the reported comparison, the order is axitinib plus gemcitabine versus placebo plus gemcitabine. A value of 1 is the reference point for equal hazards; values on either side indicate different estimated relative hazards, with direction interpreted in light of the group order. The hazard ratio is not a ratio of the median survival times, and it is not the probability that a participant will experience the event.

The registry reports hazard ratios for overall survival and progression-free survival. It does not report a different model-based time-to-event estimate in the posted results. The median time and the hazard ratio answer related but distinct questions: the median locates the midpoint of an estimated event-time distribution, while the hazard ratio compares event rates across follow-up under the analysis method. They should be read together rather than treated as interchangeable summaries.

What does the log-rank test do?

The log-rank method compares the observed event pattern between groups across follow-up, taking account of the timing of events and the participants still at risk at each event time. The registry identifies the log-rank method for both time-to-event comparisons. The associated p-value evaluates the evidence against a null hypothesis of no group difference under the test's assumptions. It is not the probability that the null hypothesis is true, nor does it measure the size or importance of a difference.

Time-to-event analysis depends on the handling of censoring, which occurs when the event time is not observed during a participant's recorded follow-up. General statistical interpretation assumes censoring is handled in a way appropriate to the analysis. The registry describes the endpoint and timeframe, but detailed censoring rules are not reported in the registry record.

What does a confidence interval add?

A confidence interval describes the uncertainty around an estimated quantity under the statistical method. In repeated use of a method under its assumptions, intervals at the stated confidence level have a defined long-run coverage property. For an individual trial, the interval marks a range of estimates compatible with the data and method; it does not assign a probability to each possible true value. For a hazard ratio, an interval spanning 1 includes the no-difference reference value, while its width indicates how much uncertainty remains.

The overall-survival hazard-ratio interval, 0.786 to 1.309, includes the reference value and extends to estimates on both sides. The progression-free-survival interval, 0.779 to 1.298, has the same broad feature. This is why the point estimates alone should not be read as definitive evidence about direction. The medians also have their own confidence intervals, which describe uncertainty in those estimates rather than directly testing the hazard-ratio comparison.

How is objective response analyzed?

Objective response is reported as a percentage in each group. The sponsor's analysis used the Cochran-Mantel-Haenszel method and reported a risk ratio. This method can account for stratification factors when such factors are part of the analysis; the registry record does not identify the factors used. A risk ratio compares response probabilities between groups, whereas a risk difference expresses their absolute separation. The unadjusted recalculation uses the posted summary counts and a different measure, so its estimate and p-value need not match the sponsor's analysis.

For response, the sponsor reported a risk ratio of 3.2, with a 95% confidence interval from 1.0 to 10.1 and p = 0.038. The unadjusted risk-difference calculation has p = 0.022. Those p-values arise from different analytical formulations and are not contradictory simply because they differ. The registry does not report enough detail to reconstruct every aspect of the sponsor's stratified analysis.

What is the role of the intent-to-treat population?

The primary endpoint definition states that the intent-to-treat population includes all randomized participants, with assignment based on initial randomization regardless of treatment received. This approach maintains the randomized group comparison and avoids excluding participants solely because their treatment exposure differed from assignment. The reported group sizes for the primary endpoint are 314 and 316. The registry record does not describe alternative primary analysis populations or sensitivity analyses.

These explanations describe general statistical concepts. The trial-specific methods that are stated in the registry are the log-rank method for the time-to-event outcomes, the Cochran-Mantel-Haenszel method for the sponsor's response analysis, the reported confidence intervals and p-values, and the intent-to-treat definition for overall survival. Other modeling choices, assumption checks, and analysis details are not reported in the registry record.

11. Confidence Intervals and Precision

The overall-survival medians are 36.9 weeks and 35.8 weeks, but each estimate has uncertainty. The 95% confidence interval is 30.1 to 41.1 weeks for axitinib plus gemcitabine and 30.0 to 44.8 weeks for placebo plus gemcitabine. The intervals overlap, and the reported medians are close. Overlap alone is not a formal test of the difference between groups; the hazard-ratio analysis is the registry's reported comparative analysis.

The hazard ratio of 1.014 has a 95% confidence interval from 0.786 to 1.309. This interval includes 1, the reference for equal hazards, and spans both lower and higher relative-hazard estimates. Its width signals uncertainty about both direction and magnitude. The p-value of 0.544 is consistent with the interval's lack of a clear separation from the reference, but a p-value and confidence interval communicate different aspects of the analysis.

For progression-free survival, the hazard ratio is 1.006, with a 95% confidence interval from 0.779 to 1.298. It too includes 1 and leaves uncertainty about the direction and size of a relative difference. For objective response, the sponsor's risk-ratio interval runs from 1.0 to 10.1 around an estimate of 3.2. That interval is broad, so the response comparison is not estimated precisely, even though its lower endpoint is at the reference value as rounded in the registry.

Confidence intervals are not guarantees that the true effect lies within the particular interval, and they do not account for every possible source of bias or analysis choice. They describe sampling uncertainty under the methods and assumptions used. The registry record does not report additional sensitivity analyses that would show how robust these intervals are to alternative assumptions.

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 planning illustrations are based on the observed effect, not on the original trial design. For overall survival and progression-free survival, the figures are total event counts, not participants per arm. For objective response, the figures are participants per arm. These are planning quantities for hypothetical future studies, not a count of the participants or events observed in this trial.

The stated 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 the listed fraction of the observed effect. The time-to-event calculations use the Schoenfeld approach. For example, the overall-survival illustration at 100% of the observed effect lists 162,427 total events for 80% power and 217,443 total events for 90% power. The objective-response illustration at 100% of the observed effect lists 473 participants per arm for 80% power and 632 participants per arm for 90% power.

The 80% power column describes a planning scenario designed to have an 80% chance of detecting the specified effect under the stated assumptions if that effect is the true one. The 90% column describes the corresponding scenario with a higher target chance of detection. Neither percentage is a probability that the treatment effect is true. These figures are not observed power.

The rows use smaller fractions of the observed effect as progressively more modest effects to detect. Smaller effects are harder to distinguish from random variation, so they require more participants or, for time-to-event outcomes, more events under these assumptions. The original study's own design assumptions and sample-size justification are not reported in the registry record.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

d is the total number of events needed (Schoenfeld's formula) with 1:1 allocation and proportional hazards.

Worked calculation for this trial

1. Hazard ratio HR = 1.014, so ln(HR) = 0.0139.
2. d = 4 × (1.96 + 0.842)² / [ln(1.014)]² = 162,426.4, rounded up to 162,427 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 <- 1.014                              # 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(471146)
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 = 162,427 (80% power), 217,443 (90% power)
 75% of observed effect: events = 288,758 (80% power), 386,566 (90% power)
 50% of observed effect: events = 649,706 (80% power), 869,772 (90% power)
Simulated power with 162,427 events: 0.799 (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 overall-survival analysis as a log-rank hazard ratio of 1.014, with a 95% confidence interval from 0.786 to 1.309 and p = 0.544. The interpretation of the primary endpoint should be anchored to that reported analysis. The posted results also give group medians and confidence intervals, which provide useful descriptive context but are not substitutes for the reported hazard-ratio comparison.

For progression-free survival, the sponsor-reported hazard ratio is 1.006, with a 95% confidence interval from 0.779 to 1.298 and p = 0.520. These are the reported time-to-event comparisons; no separate recalculated hazard ratio is used. Simple recalculations from summary data are not a replacement for the sponsor's time-to-event analysis, which uses event-time information that cannot be recovered from group medians alone.

For objective response, the sponsor used the Cochran-Mantel-Haenszel method and reported a risk ratio of 3.2, with a 95% confidence interval from 1.0 to 10.1 and p = 0.038. The unadjusted recalculation is a risk-difference analysis and gives p = 0.022. Different measures and analytical methods can produce different estimates and p-values. The registry does not flag a discrepancy between the posted sponsor result and the recalculation. The sponsor's result remains the reported trial analysis; the unadjusted result is a supplementary calculation based on summary data.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary outcome and a broad set of secondary outcomes. The listed secondary measures include progression-free survival, objective response, duration of response, several quality-of-life scores, pain, and health-state measures. Testing many outcomes can increase the chance of at least one apparently notable result arising by chance. Whether and how that concern was addressed depends on the analysis plan.

The multiplicity strategy is not reported in the registry record. It does not state whether secondary tests were adjusted, arranged in a testing hierarchy, or interpreted without adjustment. Therefore, the p-value for objective response should be read as the reported result for that comparison, not as proof that the full family of outcome tests has controlled error. The number of listed outcomes makes this qualification relevant, but does not by itself determine the validity of any individual result.

The overall-survival population is explicitly defined as intent-to-treat, including all randomized participants analyzed according to initial assignment. The registry does not report detailed missing-data procedures for secondary outcomes, the amount of missing outcome information, or sensitivity analyses addressing missingness. Handling incomplete quality-of-life questionnaires, response assessments, or follow-up can affect analyses, but the record does not provide the details needed to assess those effects.

For overall survival, the endpoint definition identifies adverse-event and follow-up contact data as sources for death status. Further details about censoring rules and the handling of participants whose status could not be determined are not reported in the registry record. Interim analyses, stopping rules, subgroup analyses, and a full statistical analysis plan are also not reported in the registry record. These omissions limit assessment of how the posted analyses relate to all planned comparisons and operational decisions.

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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: disease progression (4.9%), abdominal pain (4.6%), pyrexia (4.3%), vomiting (3.6%), asthenia (2.6%).

The registry lists serious adverse-event figures for 304 participants at risk in the axitinib plus gemcitabine group and 309 at risk in the placebo plus gemcitabine group. It records 108 affected participants in the first group and 102 in the second. For other adverse events, the figures are 276 affected participants in each group, with the same respective at-risk counts of 304 and 309. These are descriptive counts; they should not be interpreted as a randomized comparison of risk without considering definitions, follow-up, and exposure.

The most frequently listed serious terms include disease progression at 4.9%, abdominal pain at 4.6%, pyrexia at 4.3%, vomiting at 3.6%, asthenia at 2.6%, and nausea at 2.3%. These percentages describe the reported serious-event terms, not necessarily mutually exclusive groups of participants. A participant may contribute to more than one term, so the percentages should not be added.

Deaths as an adverse-event category are not reported in the registry record. The safety figures alone do not establish comparative safety or causation, and the record does not give detailed event duration, severity beyond the serious classification, or exposure-adjusted analyses.

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

The primary result describes the estimated overall-survival comparison between axitinib plus gemcitabine and placebo plus gemcitabine in the randomized intent-to-treat population. The reported hazard ratio is close to 1, its 95% confidence interval includes 1, and its log-rank p-value is 0.544. The data do not precisely establish the direction or magnitude of a difference in overall survival.

This is not the same as proving that the groups are identical. A confidence interval that includes the no-difference value can reflect uncertainty about a range of plausible effects. The interval also includes relative-hazard values on either side of 1. The trial's posted estimate is therefore best read as inconclusive about a survival difference, rather than as evidence of equivalence.

The secondary progression-free-survival comparison is similarly centered close to 1, with an interval that includes 1. Objective response has a different reported pattern, but it is a secondary endpoint, uses a different measure, and has a broad sponsor-reported confidence interval. It does not establish that overall survival differed. Results for several other listed secondary outcomes are not reported in the registry record.

The summaries do not establish what would happen for every participant, nor do they supply a complete account of benefit, harm, or quality of life. The registry record omits some analysis and follow-up details that would matter for a fuller appraisal. These statistical findings describe the posted trial results; they are not medical advice and do not determine an appropriate treatment choice.

18. Why This Trial Matters Statistically

This trial illustrates why a clinical trial result should be read through its prespecified primary endpoint and its uncertainty, rather than through a single point estimate or p-value. The overall-survival hazard ratio is close to the no-difference reference, and its confidence interval spans values on both sides. That combination makes the precision and range of compatible estimates central to interpretation.

It also demonstrates that related outcomes are not interchangeable. Overall survival, progression-free survival, and objective response describe different events or measures. A response percentage that differs between groups does not directly answer whether overall survival differs. Keeping the endpoint definitions distinct prevents a secondary result from being treated as a substitute for the primary question.

The registry's randomized, double-blind design supports a structured comparison, while the intent-to-treat definition maintains participants in their originally assigned groups. At the same time, public registry reporting has limits: the record does not provide the full statistical plan, multiplicity strategy, or detailed missing-data procedures. Those gaps matter when interpreting secondary outcomes and assessing the robustness of the analyses.

Finally, planning illustrations based on observed effects are not retrospective evidence of trial power. They help show how demanding it can be to plan for smaller effects under stated assumptions, but they do not explain the original design of this trial. Distinguishing planning calculations, sponsor analyses, and unadjusted recalculations is essential to a transparent reading of clinical trial statistics.

19. Limitations

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

The registry record does not report the trial's original sample-size justification or design assumptions. The planning illustrations are based on observed effects and do not reveal what assumptions were used to plan this study. Interim analyses and stopping rules are also not reported in the registry record.

Although the trial is described as double-blind, the registry does not explain which groups of people were masked or how masking was maintained. Arm descriptions are empty, so dosing, schedule, treatment duration, and treatment modifications are not described in the record. The detailed protocol and statistical analysis plan are not reported in the registry record.

The primary endpoint definition identifies how time to death was calculated and how death was determined, but detailed censoring procedures and sensitivity analyses are not reported. The record also does not give information needed to assess missing outcome data across the secondary measures. This is particularly relevant for questionnaire-based outcomes, for which numerical results are not reported in the registry record.

The multiplicity strategy for the many listed outcomes is not reported. The available secondary results therefore do not establish how the analyses were prioritized or whether statistical error across multiple comparisons was controlled. Detailed subgroup results and the numerical results for duration of response, quality of life, pain, and health-state measures are not reported in the registry record.

Safety reporting is also limited in the available summary. It gives at-risk and affected counts and selected serious adverse-event terms, but deaths as an adverse-event category, event timing, and exposure-adjusted comparisons are not reported in the registry record. These limitations restrict interpretation of completeness and robustness; they do not, by themselves, imply that the trial analyses are biased.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Study Of Gemcitabine Plus AG-013736 Versus Gemcitabine For Advanced Pancreatic Cancer. — NCT00471146.
• Registry record last updated 2012-07-16; status last verified 2012-06.
• Results first posted 2012-07-16.
• 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, double-blind phase 3 parallel-group study of axitinib, also called AG-013736, plus gemcitabine versus placebo plus gemcitabine for first-line treatment of locally advanced, unresectable or metastatic pancreatic cancer. Enrollment is listed as 630. The primary outcome was overall survival in the intent-to-treat population.

The reported overall-survival hazard ratio was 1.014, with a 95% confidence interval from 0.786 to 1.309 and p = 0.544. The result does not precisely establish the direction or magnitude of a survival difference. Progression-free survival showed a similarly uncertain time-to-event comparison. Objective response had a different secondary result, but it does not replace the primary survival analysis.

The registry lists additional secondary outcomes, but numerical results for several are not reported in the registry record. The original design assumptions, multiplicity strategy, and detailed missing-data procedures are also not reported in the registry record.