"Retrospective Study to Identify Clinical Factors Related to a High Benefit of Axitinib in mRCC": Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for "Retrospective Study to Identify Clinical Factors Related to a High Benefit of Axitinib in mRCC" (NCT03538717), a registered phase not stated study. The primary outcome is Number of Participants With Age Less Than or Equal to (<=) 65 Years and Greater Than (>) 65 Years at Initiation of Axitinib Treatment: Long Responders Versus Refractory Participants.

2. Clinical Question

The broader study objective was to describe the clinical profile of patients with a long response to axitinib and to explore clinical factors that might be related to long response, by comparing long responder participants with refractory participants. That objective is exploratory in character: a difference in a retrospective comparison would describe an association in the recorded groups, not establish that age or another factor caused a longer response.

3. Trial Design

The registry identifies this as a retrospective study of axitinib use in advanced or metastatic renal cell carcinoma. Data were collected from patients who had previously received axitinib under standard clinical practice. The stated purpose was to describe patients with a long response and to try to identify clinical factors associated with that response through comparison with refractory patients. The study is marked completed, with enrollment of 157.

The record gives a study start of 2018-04-19 and a primary completion date of 2019-03-01. It describes the timeframe for the primary outcome as assessment at axitinib initiation, within axitinib therapy of maximum 5.4 years approximately, based on data collected and observed retrospectively during 1 year. The record does not further explain how that observation period was organized or how the long-responder and refractory categories were operationally defined.

This design differs from a randomized comparison. Participants were grouped according to their observed response history, and allocation is not stated. The record does not describe a concurrent untreated group or a randomized treatment assignment. As a result, the comparison can identify patterns among the recorded groups but cannot by itself distinguish a response-related factor from differences in patient characteristics, treatment history, follow-up, or other influences.

The primary outcome is a count of participants by age category, comparing long responders with refractory participants. The record also lists secondary outcomes concerning progression-free survival, time to progression, overall survival, best response, objective response rate, and clinical benefit rate among long responders, including some outcomes for axitinib as second-line and subsequent treatment. It lists 37 outcome measures and 31 analyses, but the registry record does not state the detailed analysis plan or give numerical results for the primary age comparison.

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

Analysis population for the primary outcome, as defined in the registry: FAS included all evaluable participants, i.e. all participants who met all the inclusion criteria and none of the exclusion criteria and have received at least 1 dose of axitinib prior to inclusion in the study. Here, "Overall Number of Participants" signifies participants evaluable for this outcome measure.

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 evaluable participants, i.e. all participants who met all the inclusion criteria and none of the exclusion criteria and had received at least 1 dose of axitinib prior to inclusion in the study.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants With Age Less Than or Equal to (<=) 65 Years and Greater Than (>) 65 Years at Initiation of Axitinib Treatment: Long Responders Versus Refractory Participants, measured over At initiation of axitinib treatment, within axitinib therapy of maximum 5.4 years approximately (from the data collected and observed retrospectively during 1 year).

The primary endpoint counts participants aged less than or equal to 65 years and participants aged greater than 65 years at the initiation of axitinib treatment, separately for long responders and refractory participants. It is a categorical description of age at treatment initiation, not a direct measurement of response duration, survival, or treatment effect. Its purpose in this study is to examine whether the age profile differs between the two response-defined groups.

Age is represented as two categories in the endpoint rather than as a continuous value. In general, categorization can make a group comparison easy to describe, but it also removes detail about variation within each category. The registry record does not explain why this threshold was selected, whether age was also analyzed continuously, or whether the categorization was specified before the data were examined.

The Full Analysis Set for this outcome includes evaluable participants who met all inclusion criteria, met none of the exclusion criteria, and received at least one dose of axitinib before study inclusion. The record defines the outcome population this way but does not report how many participants in each response group were evaluable for the age endpoint.

The numerical counts for the age categories, any statistical comparison, and any uncertainty interval are not reported in the registry record. Thus, the endpoint identifies the question being asked, but the available registry information does not establish whether the age distribution differed between long responders and refractory participants or how precisely such a difference was estimated.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical interpretation begins by distinguishing what the endpoint measures from what a statistical comparison might estimate. The primary endpoint is a count: participants are classified by age at axitinib initiation, then considered in the long responder and refractory groups. A count describes the observed data. A comparison would require a defined contrast between groups and, for interpretation, enough information about the group-specific denominators and uncertainty. The registry record does not report those primary counts or a comparison estimate.

What does a count tell the reader?

A count states how many observed participants fall in a category. Counts are useful for showing the composition of a study group, but counts alone can be hard to compare when group sizes differ. A proportion can express a count relative to its group denominator, but it cannot be reconstructed responsibly when the needed group-specific counts and denominators are not reported. No percentages or contrasts for the primary age categories can therefore be stated from this record.

The outcome’s Full Analysis Set definition is important because it specifies who is intended to contribute: participants who met the inclusion criteria, met none of the exclusion criteria, and received at least one dose of axitinib before study inclusion. It does not mean every enrolled participant necessarily contributed to every outcome. The number evaluable for the primary endpoint, and any reasons for exclusion from that outcome, are not reported in the registry record.

Why distinguish association from causation?

In general, randomization can help balance measured and unmeasured characteristics across treatment groups, supporting causal comparisons under appropriate conditions. This study is described as retrospective, and the comparison is between response-defined groups. Allocation is not stated, and the groups are not a like-for-like randomized control comparison. A statistical association between a characteristic and response group, if reported, would therefore be descriptive or exploratory rather than proof that the characteristic caused the response pattern.

In general, confounding occurs when a factor is related both to the characteristic being studied and to the outcome or group classification, distorting an observed association. Treatment history, disease features, patient characteristics, and follow-up could be relevant in a retrospective clinical dataset, but the registry record does not say how these or other factors were handled. It does not report a detailed adjusted model for the primary age comparison.

What do uncertainty measures mean?

In general, a confidence interval describes the range of estimates reasonably compatible with the observed data under a statistical model and its assumptions. A narrower interval indicates greater statistical precision, while a wider interval indicates less precision. A p-value, in general, summarizes how unusual data at least as extreme as those observed would be under a specified null model; it does not measure the probability that the null hypothesis is true, nor does it convey clinical importance by itself. Neither an interval nor a p-value for the primary age comparison is reported in the registry record.

For a binary outcome such as membership in an age category, analysts may use methods for categorical data, but the suitable method depends on the study design, group sizes, analysis objective, and whether adjustment is intended. The registry record does not state the test, model, or estimation approach used for the primary endpoint. The endpoint title alone is not enough to infer those details.

How should multiple outcomes be considered?

Multiplicity arises when several outcomes, groups, or statistical tests are examined. In general, conducting many tests can increase the chance of finding an apparently notable result by chance if each test is interpreted independently. The registry lists 37 outcome measures and 31 analyses. These counts signal that the record contains a broad set of planned or reported measures, but do not identify which tests formed a single inferential family or whether multiplicity adjustments were used. The statistical plan for multiplicity is not reported in the registry record.

The secondary outcomes include several time-to-event measures. In general, time-to-event analysis concerns both how long it takes for an event to occur and whether observations are censored because follow-up ends before the event. Interpretation depends on event definitions, follow-up, censoring, and the analysis method. The registry lists progression-free survival, time to progression, and overall survival, but their detailed definitions, censoring rules, and numerical estimates are not reported in the information.

How should planning illustrations be read?

In general, sample-size calculations ask how many participants may be needed to detect a specified effect under stated assumptions. The registry describes planning illustrations based on the observed effect, using a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is a stated fraction of the observed effect. For time-to-event illustrations, the total number of events is used, while other illustrations are expressed as participants per arm.

These calculations are planning illustrations, not observed power and not evidence that the retrospective study was designed with those assumptions. The record does not state that those were the original design assumptions. The actual study is retrospective, and allocation is not stated. Without the numerical illustration outputs or a documented original design plan, no sample-size requirement or achieved precision should be inferred.

Dispersion is relevant to count outcomes when variability exceeds what a basic statistical model expects. In general, over-dispersion can make standard errors too small if ignored, leading to overly narrow uncertainty intervals or misleading test results. The primary endpoint is described as a participant count in categories, but the registry record does not report a dispersion parameter or a count-regression model; it would be inappropriate to infer that such a method was used.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 37 outcome measures and 31 analyses. It identifies an age-category count comparing long responder and refractory participants as the primary outcome, while the secondary outcome list includes several measures of survival, progression, response, and subsequent treatment. A broad outcome set can create multiplicity concerns because many analyses may produce apparently notable findings by chance, particularly when results are interpreted without a prespecified testing strategy.

The registry record does not report whether statistical testing was planned for all listed outcomes, how a family of tests was defined, whether any multiplicity adjustment was used, or how the primary outcome was distinguished inferentially from exploratory analyses. These details matter when interpreting p-values across multiple outcomes. The number of listed outcomes alone does not show that every outcome was tested or that any particular adjustment was applied.

Missing-data handling is not reported in the registry record. The Full Analysis Set is defined as all evaluable participants who met the eligibility criteria and received at least one dose before study inclusion, but the record does not state how many participants were evaluable for each outcome, whether any data were missing, how missing values were handled, or whether sensitivity analyses were conducted. The definition of the analysis population does not answer those questions.

The record also does not report whether the primary analysis was adjusted for other clinical characteristics, or how any such characteristics were selected. Because the study is retrospective and compares response-defined groups, adjustment decisions would affect interpretation of observed associations. Without a reported statistical plan and outcome-specific data completeness information, the primary finding cannot be assessed for robustness to missingness, multiple testing, or confounding.

11. 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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: diarrhea (2.5%), anemia (0.6%), pancreatitis (0.6%), vomiting (0.6%), asthenia (0.6%).

The registry lists 157 participants at risk for serious adverse events and reports 15 participants affected. It also lists 157 at risk for other adverse events and 116 affected, along with 2 deaths. These figures are counts in the registry record; they should be read as descriptive reporting, not as estimates from a randomized comparison. The record does not provide a comparison group for safety or establish that the events or deaths were caused by axitinib.

The serious adverse-event terms listed are diarrhea at 2.5%, and anemia, pancreatitis, vomiting, asthenia, and pyrexia at 0.6% each. These percentages are reported as written in the registry record. They do not by themselves describe severity beyond the serious-event classification, timing, duration, recurrence, or relationship to treatment. Those details are not reported in the registry information.

Serious adverse events and other adverse events are separate registry categories; the reported counts should not be combined or treated as mutually exclusive without further information. The registry record does not state whether a participant could contribute to more than one event term, the event-collection procedures, or the criteria used to judge treatment relatedness. The figures therefore give a limited description of reported safety events, rather than a complete causal or comparative safety assessment.

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

The record’s primary question is whether the age profile at axitinib initiation differs between long responder participants and refractory participants. The endpoint is explicitly defined as counts for age less than or equal to 65 years and age greater than 65 years. However, the registry information does not report those counts, denominators by group, or a statistical contrast. It is therefore not possible to state from this record whether one group was older or younger on this measure, or how uncertain a comparison would be.

The study has value as a retrospective description of clinical experience with axitinib and as an attempt to identify factors associated with long response. It also records a broad set of secondary outcomes relevant to treatment course, including progression, survival, response, and later treatment. But a list of outcomes is not the same as numerical evidence for those outcomes; their results are not reported in the registry information.

Even a clear association in a retrospective comparison would not establish causation. The long responder and refractory groups are defined by observed outcomes, allocation is not stated, and the comparison is not randomized. Potential differences between the groups may be linked to their response classification. The record does not report the analysis details needed to assess whether those differences were measured or adjusted for.

The safety counts also have a descriptive scope. The registry reports participants affected by serious and other adverse events and lists serious-event terms, but it does not provide a safety control group or establish event causality. These data cannot support a comparative claim about safety from the registry record alone.

Planning illustrations, if considered, should be understood as calculations under stated assumptions rather than evidence of observed power or proof that the original study had a particular design. In sum, the record documents a completed retrospective study and its questions, but the primary result and many analysis details needed for a stronger statistical interpretation are not reported.

13. Why This Trial Matters Statistically

This study illustrates a recurring challenge in clinical biostatistics: translating a clinically meaningful question into a comparison that supports a careful interpretation. The investigators aimed to describe patients with a long response to axitinib and explore factors associated with that response by comparing long responder and refractory participants. The primary age endpoint makes one part of that question concrete by defining two age categories at treatment initiation.

It also demonstrates why outcome definitions and analysis populations matter. The record specifies a Full Analysis Set based on eligibility and receipt of at least one axitinib dose, but does not give the number evaluable for the primary endpoint. Knowing the population definition helps readers understand who is intended to contribute; knowing group-specific denominators and data completeness would be needed to judge the precision and representativeness of the comparison.

The retrospective design is central to statistical interpretation. A comparison of groups defined by their treatment response can identify patterns in clinical records, but it does not have the protection against confounding that random assignment can provide. This makes transparent reporting of group definitions, adjustment methods, missing-data handling, and uncertainty especially important. These details are not reported in the registry record.

The broad outcome inventory also makes multiplicity relevant. With 37 outcome measures and 31 analyses listed, readers need to know which analyses were confirmatory, which were exploratory, and how multiple testing was addressed. The registry record does not report that strategy. Finally, the contrast between a defined primary question and unavailable numerical results shows why a registry’s study description and its posted outcome results serve different purposes: design information frames what was asked, while estimates and uncertainty are needed to answer it.

14. 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 limitation for interpreting this specific study from the registry record is the absence of the age-category counts for long responders and refractory participants. Group-specific denominators, an estimated contrast, and an uncertainty measure are also not reported. Without them, the primary outcome cannot be quantified or judged for precision.

The record describes the study as retrospective and compares participants based on long responder or refractory status. Allocation is not stated, and the comparison is not a like-for-like randomized control comparison. The operational definition of long response and the criteria defining refractory participants are not reported in the information. This limits assessment of how consistently participants were classified and how comparable the groups were.

The primary endpoint is age grouped at a threshold of 65 years. The reason for that threshold and whether age was also considered on a continuous scale are not reported. Categorization can conceal differences within a category, so the endpoint answers a narrower question than a full analysis of age variation.

The record gives a Full Analysis Set definition but does not report the number of evaluable participants for the primary outcome, missing-data amounts, missing-data handling, or sensitivity analyses. It also does not report a detailed statistical analysis plan, adjustment for potential confounders, multiplicity procedures, or the detailed definitions and censoring rules for listed time-to-event outcomes.

Safety information is also limited for comparative interpretation: no control group is described, and the record does not state event relatedness, collection procedures, or details such as timing and recurrence. These gaps restrict causal interpretation of both efficacy-related associations and adverse-event figures.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: "Retrospective Study to Identify Clinical Factors Related to a High Benefit of Axitinib in mRCC" — NCT03538717.
• Registry record last updated 2022-02-17; status last verified 2022-02.
• Results first posted 2020-03-13.
• 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.

18. Record Summary

The registry record describes a completed retrospective study of axitinib in advanced or metastatic renal cell carcinoma, sponsored by Pfizer. Enrollment was 157. The study start date is 2018-04-19, and the primary completion date is 2019-03-01.

The stated objective was to describe patients with a long response to axitinib and explore clinical factors associated with that response by comparing long responder participants with refractory participants. The primary outcome counts participants aged less than than or equal to 65 years and greater than 65 years at axitinib initiation. The outcome population is defined as evaluable participants meeting the eligibility criteria and having received at least one dose of axitinib.

The registry lists secondary survival, progression, response, and treatment outcomes, as well as adverse-event counts. The primary age-category counts and detailed statistical plan are not reported in the registry record. The design is retrospective, allocation is not stated, and the comparison should not be interpreted as randomized.