A Study of Avelumab With Axitinib Versus Sunitinib In Advanced Renal Cell Cancer (JAVELIN Renal 101): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Avelumab With Axitinib Versus Sunitinib In Advanced Renal Cell Cancer (JAVELIN Renal 101) (NCT02684006), a randomized phase 3 study. The primary outcome is Progression Free Survival (PFS) as Assessed by Blinded Independent Central Review (BICR) in Programmed Death-Ligand 1 (PD-L1) Positive Participants.

2. Clinical Question

3. Trial Design

This was a phase 3, multinational, randomized, open-label, parallel-arm study of first-line treatment for advanced renal cell carcinoma. The registry lists Pfizer as the sponsor, records enrollment of 886, and marks the study as completed. The study began on 2016-03-23, and its primary completion date is recorded as 2023-08-31.

Participants were assigned to avelumab plus axitinib or sunitinib monotherapy. The combination arm used avelumab 10 mg/kg intravenously every two weeks with axitinib 5 mg orally twice daily. Sunitinib was given at 50 mg orally once daily on a 4/2 schedule. No masking was used for the treatment assignment. However, the primary progression-free survival endpoint was assessed by blinded independent central review, a distinction between the open-label treatment conduct and the endpoint assessment method.

The primary outcome was median progression-free survival in PD-L1-positive participants, measured in months. The analysis population was a subset of randomized participants; the registry describes the full analysis set as all randomized participants. The primary outcome timeframe extended from randomization to progressive disease, death, or censoring, with the registry stating a maximum of approximately 26 months for that timeframe. The registry also lists outcomes in participants irrespective of PD-L1 expression, including progression-free survival and overall survival, along with response and disease-control outcomes assessed by central review or the investigator.

The registry record does not report the original sample-size justification, the full allocation procedures, or the detailed protocol for treatment changes. These details should not be inferred from the enrollment, the reported outcomes, or the planning illustrations.

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: A total of 886 participants were enrolled and randomized in the study.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: FAS included all participants who were randomized. Analysis was performed on subset of randomized participants, who were PD-L1 positive.

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: Full analysis set (FAS) included all randomized participants.

6. Primary Endpoint

The registry lists the primary outcome as Progression Free Survival (PFS) as Assessed by Blinded Independent Central Review (BICR) in Programmed Death-Ligand 1 (PD-L1) Positive Participants, measured over From date of randomization to the first documentation of PD or death due to any cause or censoring date, whichever occurred first (maximum up to approximately 26 months).

Progression-free survival measures the time from randomization until the first documented progressive disease or death from any cause, whichever occurs first. In this trial, progression was assessed using RECIST version 1.1 and reviewed by a blinded independent central review. The primary analysis was limited to randomized participants who were PD-L1 positive.

This endpoint combines disease progression and death into an event. It therefore describes how long participants remained alive without documented progression under the trial’s endpoint definitions. Median progression-free survival is the time by which half of the analyzed group has experienced an event, accounting for censoring in the time-to-event analysis; it is not an average duration for every participant.

Central review can provide an assessment that is less directly dependent on the knowledge of treatment assignment than an unblinded investigator assessment. That feature is relevant because the trial itself was open-label. The registry defines censoring for participants without an event, those who began new anti-cancer therapy before an event, and those with an event after at least two missing tumor assessments. Censoring means that an event time is not counted as observed at that point; the time-to-event methods use the follow-up information available before censoring.

Progression was defined by specified changes in target lesions or the appearance of new lesions. The record describes a relative increase of at least 20% in the sum of target-lesion longest diameters, together with an absolute increase of more than 5 mm, or the appearance of one or more new lesions. The endpoint is useful for comparing the timing of progression or death, but it does not by itself describe overall survival, quality of life, or every aspect of disease burden.

7. Results Posted in the Registry

• Avelumab + Axitinib vs Sunitinib: hazard ratio 0.61 (95% CI 0.47–0.79), 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.

PFS as Assessed by BICR in Participants Irrespective of PD-L1 Expression

• 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:
  • Avelumab + Axitinib vs Sunitinib: hazard ratio 0.69 (95% CI 0.56–0.84), 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 Avelumab + Axitinib vs Sunitinib than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

OS in Participants Irrespective of PD-L1 Expression

• 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:
  • Avelumab + Axitinib vs Sunitinib: hazard ratio 0.88 (95% CI 0.75–1.04), reported p = 0.134 (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 Avelumab + Axitinib vs Sunitinib and the control. The interval gives the range of effects that remain compatible with the data.

Percentage of Participants With Objective Response (OR) as Assessed by BICR Irrespective of PD-L1 Expression

• 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:
  • Sunitinib: 25.7% (114/444) versus 51.4% (227/442) on control; risk difference -25.7 percentage points (95% CI -31.9 to -19.5), risk ratio 0.50 (0.42–0.60); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Sunitinib than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes broaden the assessment beyond the primary question in PD-L1-positive participants. They include progression-free survival and overall survival regardless of PD-L1 expression, central-review and investigator-assessed response and disease control, and time to response. This wider set of outcomes allows the record to address different aspects of disease course and different assessment methods, but it also means that the results do not all answer the same question.

The reported secondary progression-free survival comparison is directionally consistent with the primary result. The overall-survival estimate is less conclusive by its reported uncertainty interval and p value. The response outcome has a separate unadjusted re-calculation, rather than a sponsor-reported analysis in the record. These results should be read as supporting context, not as interchangeable confirmations of the primary endpoint.

When many outcomes are examined, the chance of observing an apparently notable result by chance can increase. Statistical adjustment or a prespecified testing sequence can address that issue, but the multiplicity strategy is not reported in the registry record. Accordingly, secondary p values and unadjusted calculations carry less weight than the primary analysis unless the trial’s full analysis plan establishes how they were prioritized and tested. The registry also lists secondary outcomes for which results are not included in the registry record, so their findings should not be inferred from the outcomes that are reported.

Differences in population, assessment, and endpoint definition matter as well. An analysis regardless of PD-L1 expression is not the same population as the primary PD-L1-positive subset. A central review and an investigator assessment may use the same broad response concept but are distinct assessments. Secondary findings can help describe the range of evidence, but they do not change what the primary analysis estimates.

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

• PFS as Assessed by BICR in Participants Irrespective of PD-L1 Expression
• OS in Participants Irrespective of PD-L1 Expression
• Percentage of Participants With Objective Response (OR) as Assessed by BICR Irrespective of PD-L1 Expression
• Percentage of Participants With OR as Assessed by Investigator Irrespective of PD-L1 Expression
• Percentage of Participants With Disease Control (DC) as Assessed by BICR Irrespective of PD-L1 Expression
• Percentage of Participants With DC as Assessed by Investigator Irrespective of PD-L1 Expression
• Time to Tumor Response (TTR) as Assessed by BICR in Participants Irrespective of PD-L1 Expression
• TTR as Assessed by Investigator in Participants Irrespective of PD-L1 Expression

9. Statistical Interpretation

10. Statistical Methods Explained

What does a hazard ratio describe?

Progression-free survival is a time-to-event outcome, so participants can have different lengths of follow-up and some may be censored before progression or death is documented. A hazard ratio summarizes the relative event hazard between treatment arms over the analyzed follow-up. It is not the ratio of the reported medians and does not describe an individual participant’s probability of progression or death. Interpretation also relies on the time-to-event analysis and its assumptions; the full diagnostic checks are not reported in the registry record.

For the primary comparison, the reported hazard ratio is 0.61. The estimate is below 1, indicating a lower estimated event hazard in the combination arm than in the sunitinib arm under the analysis. The estimate is accompanied by a confidence interval and a Log Rank test. The Log Rank method compares the observed and expected event experience across groups over time; it is not a test of the difference between the two medians alone.

How should the p value be read?

A p value describes how incompatible the observed data are with a specified null model, assuming that model and the analysis conditions hold. It is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of an effect. The primary reported p value is < 0.001. The registry identifies the analysis as a superiority test, but does not report the full statistical analysis plan or the decision thresholds used for every outcome. A p value should therefore be considered alongside the estimated effect, its confidence interval, the endpoint, and the analysis population.

What do medians and confidence intervals add?

The median is a time-based summary of the event distribution. It offers a direct description in months, while the hazard ratio uses information across the observed event times. The two measures answer related but different questions. A 95% confidence interval describes the uncertainty of an estimate under the statistical procedure; it does not guarantee that the true value lies within the particular interval. In the primary result, a lower limit is reported for the combination arm’s median, but an upper limit is not reported in the registry record. The hazard ratio interval is reported in full.

Confidence intervals can also help distinguish results that are more or less compatible with a no-difference value. For a hazard ratio, that reference value is 1. The overall-survival interval extends across 1, unlike the intervals reported for the progression-free survival comparisons. This is an uncertainty statement, not proof that there is no difference in overall survival.

Why is assessment method relevant?

The trial was open-label, so treatment assignment was not masked. The primary endpoint was nevertheless assessed by blinded independent central review. A separate review can reduce the influence of treatment knowledge on the assessment of radiographic progression. It does not remove every possible source of bias, including differences in follow-up, missing assessments, or decisions that occur before review. Investigator-assessed outcomes are also listed among the secondary outcomes, but their results are not included in the registry record described for this analysis.

How should the binary re-calculation be distinguished?

Objective response is a binary outcome: participants are classified according to whether the defined response was observed. The posted percentages are based on the central-review outcome in participants irrespective of PD-L1 expression. The unadjusted re-calculation uses the posted group proportions and does not represent the sponsor’s reported analysis. It is a simple comparison of the available summary statistics, rather than a substitute for a full participant-level model or adjustment for other factors. Its p value should be read as an unadjusted secondary result, not as a replacement for the primary time-to-event analysis.

For binary outcomes, uncertainty may be represented with confidence intervals for group proportions or their comparison. The registry lists confidence intervals for the response percentages. A risk difference expresses the difference between the two proportions on an absolute scale, while a relative measure compares them on a relative scale. These summaries have different interpretations and can convey different aspects of the same data. No adjustment for baseline characteristics is reported for the re-calculation.

What assumptions underlie time-to-event interpretation?

Censoring is a normal feature of time-to-event analysis, but it needs careful interpretation. A censored participant contributes observed follow-up up to the censoring point without being counted as having experienced the endpoint at that time. The registry specifies circumstances for censoring in the primary endpoint, including no event by the relevant assessment, starting new anti-cancer therapy before an event, and an event following at least two missing tumor assessments. The record does not report sensitivity analyses that would show how alternative handling of these cases affects the estimates.

Time-to-event methods also depend on how event times and follow-up are handled. The registry identifies the Log Rank method for the reported hazard-ratio comparisons, but the full model details, diagnostic checks, and any stratification factors are not reported in the registry record. These omissions limit how fully the analysis can be independently assessed from summary results alone.

11. Confidence Intervals and Precision

The primary median progression-free survival estimates are 13.8 months for avelumab plus axitinib and 7.2 months for sunitinib. The registry reports a lower 95% confidence limit of 11.1 months for the combination arm; an upper limit for that median is not reported in the registry record. For sunitinib, the reported 95% confidence interval is 5.7 to 9.7 months. The hazard ratio is 0.61, with a 95% confidence interval from 0.475 to 0.79.

The hazard-ratio interval is the more direct measure of precision for the relative time-to-event comparison. Its limits are both below 1, while the interval for the overall-survival hazard ratio extends from 0.749 to 1.039 and includes 1. The estimates therefore differ in how tightly they locate the relative comparison with respect to the no-difference value. This does not make a confidence interval a guarantee about the true effect; it describes uncertainty under the analysis procedure and assumptions.

For the secondary response percentages, the registry reports confidence intervals of 46.6 to 56.1 for avelumab plus axitinib and 21.7 to 30.0 for sunitinib. These intervals describe uncertainty around each arm’s response proportion, not the adjusted effect of treatment. The unadjusted re-calculation is a separate comparison of the posted proportions.

Precision depends on the amount and distribution of event or response information, follow-up, and the analysis method. The registry record does not report all details needed to assess those influences or to reconstruct every interval. In particular, a missing confidence limit should not be filled in by inference from another reported statistic.

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 use the observed effect as a starting point and consider detecting a stated fraction of that effect. They assume a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. For time-to-event outcomes, the figures are total events required; they are not participants per arm. For other outcomes, the figures are participants per arm.

For the primary time-to-event comparison, planning at 100% of the observed effect gives 129 total events for 80% power and 173 total events for 90% power. At 75% of the observed effect, the corresponding figures are 229 and 306 total events. At 50% of the observed effect, they are 514 and 689 total events. The secondary response comparison is a binary outcome and its planning figures are per arm: at 100% of the observed effect, 56 participants per arm for 80% power and 74 participants per arm for 90% power; at 75%, 102 and 136; and at 50%, 235 and 314.

In plain terms, the 80% power column describes the planning number associated with an 80% chance of detecting the specified effect under the stated assumptions if that effect is present. The 90% power column describes the corresponding planning number for a higher chance of detection. Smaller effects are harder to distinguish from random variation, so the illustrations require more events or participants as the target effect becomes smaller. These figures are not observed power; they are planning illustrations derived from the observed effects, not the trial’s original design calculations.

The registry record does not report the original sample-size justification or the assumptions used to design the trial. The planning illustrations should not be treated as evidence that the trial was designed for those exact targets, nor as a measure of the certainty of the results already observed.

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 = 0.610, so ln(HR) = -0.4943.
2. d = 4 × (1.96 + 0.842)² / [ln(0.610)]² = 128.5, rounded up to 129 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 <- 0.61                              # 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(2684006)
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 = 129 (80% power), 173 (90% power)
 75% of observed effect: events = 229 (80% power), 306 (90% power)
 50% of observed effect: events = 514 (80% power), 689 (90% power)
Simulated power with 129 events: 0.806 (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 202 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 886 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 <- 7.2          # control-arm median, months (converted from the posted unit)
hr          <- 0.61          # 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: 140         p (1-sided)   0.0001
  Events: 58        ~HR at bound   0.3762
   Month: 17    P(Cross) if HR=1   0.0001
             P(Cross) if HR=0.61   0.0338
   IA 2: 67%                   Z   2.5114
      N: 202         p (1-sided)   0.0060
 Events: 116        ~HR at bound   0.6263
   Month: 26    P(Cross) if HR=1   0.0060
             P(Cross) if HR=0.61   0.5603
       Final                   Z   1.9930
      N: 202         p (1-sided)   0.0231
 Events: 173        ~HR at bound   0.7385
   Month: 60    P(Cross) if HR=1   0.0250
             P(Cross) if HR=0.61   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

For the primary progression-free survival comparison, the sponsor-reported analysis uses the Log Rank method and reports a hazard ratio of 0.61, a 95% confidence interval from 0.475 to 0.790, and p< 0.001. The reported analysis is classified as superiority. The posted primary comparison gives the same estimate and p value, with the upper limit also expressed as 0.79. The registry record does not flag a discrepancy between these values.

The sponsor analysis is the primary reference for the trial’s reported time-to-event result. Simple re-calculations from summary statistics, where available, are not a replacement for that analysis. In particular, the objective-response comparison has an unadjusted re-calculated p value of < 0.001, but no sponsor-reported analysis is listed for that outcome in the record. That re-calculation is limited to the posted summary proportions and should not be presented as the sponsor’s inferential result.

Other sponsor-reported time-to-event analyses are listed for secondary outcomes, including progression-free survival and overall survival in participants irrespective of PD-L1 expression. Their interpretation remains subject to the outcome definitions, analysis populations, and uncertainty intervals reported for each result. Details of the complete sponsor statistical analysis plan are not reported in the registry record.

16. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary endpoint and a broad set of secondary outcomes, but the multiplicity plan is not reported in the registry record. It is therefore not possible to determine from this record whether secondary tests were adjusted, tested in a fixed sequence, or treated as descriptive. When several outcomes are examined, unadjusted p values can overstate the strength of evidence if interpreted as though only one hypothesis had been tested. This is a reason to give the designated primary analysis greater weight and to treat secondary findings as context unless the full analysis plan clarifies their status.

The primary analysis population is described as a subset of the full analysis set: the full analysis set included all randomized participants, and the primary analysis included randomized participants who were PD-L1 positive. The record does not report the number classified as PD-L1 positive beyond the analysis group counts recorded for the endpoint. It also does not provide enough detail to assess the handling of every participant who was randomized but did not contribute to the primary subset analysis.

The primary endpoint description specifies censoring rules for participants without an event, participants who started new anti-cancer therapy before an event, and participants with an event after at least two missing tumor assessments. Those rules define how some incomplete follow-up is represented in the time-to-event analysis. However, the registry record does not report sensitivity analyses, the amount or pattern of missing tumor assessments, or how missing information was handled across the secondary outcomes.

The record also does not report the full statistical analysis plan, the original sample-size justification, detailed interim-analysis procedures, or all modeling and diagnostic details. None of these should be inferred from the reported p values or planning illustrations. The available results can be described on their stated terms, while these gaps limit assessment of the complete inferential framework.

17. 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: diarrhoea (2.8%), abdominal pain (2.7%), acute myocardial infarction (2.5%), disease progression (2.3%), anaemia (2.3%).

The registry’s safety summary lists 434 participants at risk in the avelumab-plus-axitinib group and 439 in the sunitinib group. Serious adverse events affected 231 and 166 participants, respectively. The summary also lists other adverse events affecting 429 and 433 participants, and deaths of 284 and 296. These are posted counts; they should not be converted into rates by comparing them with other group sizes or treated as a direct causal comparison without further context.

The most frequently listed serious terms include diarrhoea at 2.8%, abdominal pain at 2.7%, acute myocardial infarction at 2.5%, disease progression at 2.3%, anaemia at 2.3%, and acute kidney injury at 2.1%. The registry entries do not specify an arm for each listed term, so those percentages should not be attributed to either treatment group. The record does not provide additional detail about event timing, severity grading, duration, or relatedness to treatment.

Safety counts describe events recorded during the study under the registry’s reporting conventions. They do not, on their own, establish that a treatment caused an event or permit a complete comparison of safety profiles. The meaning of the death counts is not further specified in the record, so they should not be interpreted as deaths caused by treatment.

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

The primary result is an estimate of progression-free survival for randomized participants who were PD-L1 positive, using central-review assessment and the registry’s event and censoring definitions. The reported median times and hazard ratio describe different features of that outcome: the medians summarize event timing in months, while the hazard ratio compares event hazards over follow-up. The confidence interval communicates uncertainty around the relative estimate.

The result does not establish that every participant experienced the same outcome, nor does it predict an individual participant’s course. It does not, by itself, answer whether overall survival differs, whether all secondary outcomes favor one arm, or how the findings apply to people outside the trial population. The overall-survival comparison has its own estimate and uncertainty, which should be interpreted separately from progression-free survival.

The findings also do not establish a treatment effect for every subgroup. The primary analysis was restricted to participants who were PD-L1 positive, while some secondary outcomes covered participants irrespective of PD-L1 expression. Those populations are distinct. The registry’s open-label design and central review for the primary endpoint are both relevant context, but central review cannot resolve every potential limitation of trial conduct or missing follow-up.

Statistical evidence is not the same as clinical importance. The registry results describe estimated differences and their statistical uncertainty; the record does not supply a threshold for a clinically meaningful difference. Nor does a p value establish the probability that a treatment is beneficial or harmful. Interpretation should remain tied to the measured outcomes, analysis populations, and the information reported in the registry.

19. Why This Trial Matters Statistically

This trial illustrates why the outcome definition and analysis population are central to reading a clinical result. Its primary question concerns progression-free survival assessed by blinded independent central review in PD-L1-positive participants. A different endpoint, such as overall survival, or a broader population can produce a different estimate and addresses a different question. Keeping those distinctions clear prevents a result for one analysis from being generalized to all trial outcomes.

It also shows how time-to-event summaries complement one another. Median times offer an interpretable description in months, while the hazard ratio summarizes the relative event experience over follow-up. Confidence intervals help convey how precisely the relative comparison is estimated. A point estimate alone cannot show the range of uncertainty, and a p value alone cannot convey effect size.

The open-label design alongside blinded central review demonstrates that different sources of bias can be addressed differently. The treatment assignment was unmasked, but the primary progression assessment used an independent blinded review. That distinction is important when considering the strength and limits of the endpoint assessment.

Finally, the registry record demonstrates the need to separate primary results, sponsor-reported analyses, unadjusted re-calculations, and planning illustrations. These are not interchangeable forms of evidence. Multiplicity, missing-data handling, and the original design assumptions also matter, and several of those details are not reported in the registry record. A statistically careful reading keeps the reported result informative without claiming more than its endpoint, population, and analysis support.

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 primary endpoint analysis covers only randomized participants who were PD-L1 positive, rather than all randomized participants. The registry identifies the full analysis set as all randomized participants but does not provide the broader context needed to assess all differences between the full set and the primary subset. Results for participants regardless of PD-L1 expression are listed as secondary analyses and should not be substituted for the primary result.

The study was open-label. Blinded independent central review was used for the primary progression-free survival assessment, which is relevant mitigation for assessment bias, but open-label conduct can still affect aspects of care, follow-up, and decisions made before central assessment. The registry record does not report detailed procedures that would allow those possibilities to be evaluated fully.

The primary median confidence interval for the combination arm has a reported lower limit but no upper limit in the registry record. The record also does not provide the full statistical analysis plan, multiplicity strategy, original sample-size justification, detailed interim-analysis procedures, or all missing-data sensitivity analyses. The primary endpoint description includes censoring rules, but the amount and pattern of missing assessments are not reported in the registry record.

Many secondary outcomes are listed, and not all have results included in the record. Their multiplicity handling is not reported. The objective-response comparison is an unadjusted re-calculation from posted summary statistics, not a listed sponsor-reported analysis. Safety information is also limited in the available summary: the listed serious-event terms are not assigned to an arm, and details such as event timing and relatedness are not reported. These limitations constrain interpretation without nullifying the reported primary estimate.

21. Related Tutorials

22. Related Statistical Calculators

23. Sources

• ClinicalTrials.gov: A Study of Avelumab With Axitinib Versus Sunitinib In Advanced Renal Cell Cancer (JAVELIN Renal 101) — NCT02684006.
• Registry record last updated 2025-07-30; status last verified 2025-07.
• Results first posted 2024-10-29.
• 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

NCT02684006 is a completed phase 3, multinational, randomized, open-label, parallel-arm trial of first-line treatment for advanced renal cell carcinoma. The study enrolled 886 participants and compared avelumab plus axitinib with sunitinib monotherapy.

The primary outcome was progression-free survival assessed by blinded independent central review in randomized participants who were PD-L1 positive. The reported hazard ratio was 0.61, with a 95% confidence interval from 0.475 to 0.79 and p< 0.001. Secondary outcomes included progression-free survival and overall survival irrespective of PD-L1 expression, response, disease control, and time to response.

The record reports several sponsor analyses and an unadjusted re-calculation for objective response. The multiplicity plan, original sample-size justification, and several details of missing-data and analysis procedures are not reported in the registry record.