This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for A Safety Trial of Nivolumab in Patients With Advanced or Metastatic Non-Small Cell Lung Cancer Who Have Progressed During or After Receiving At Least One Prior Chemotherapy Regimen (NCT02066636), a randomized phase 3 study. The primary outcome is The Number of Participants With Treatment Related Select Adverse Events (Grade 3-4 and Grade 5).
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT02066636 |
| Phase | Phase 3 |
| Condition | Non Small Cell Lung Cancer (NSCLC) |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None (open label) |
| Enrollment | 1428 |
| Arms in the primary outcome | n/a |
| Lead sponsor | Bristol-Myers Squibb |
| Results posted | Yes |
2. Clinical Question
Population
People with advanced or metastatic non-small cell lung cancer who had progressed during or after at least one prior systemic regimen. The primary endpoint population is described as treated participants before randomization and treated participants randomized to Cohort A and Cohort B.
Intervention
Nivolumab was administered as an intravenous infusion. Both cohorts received nivolumab monotherapy, with different registry-described treatment limits: Cohort A continued until progression, unacceptable toxicity, or withdrawal of consent; Cohort B had a one-year treatment limit, with retreatment allowed at progression.
Comparator
No different drug or usual-care comparator is listed. The randomized cohorts both received nivolumab; they differed in their specified treatment duration. The registry record does not describe a randomized comparison of nivolumab against another treatment.
Primary question
The primary endpoint counts participants with treatment-related select adverse events graded 3, 4, or 5. The select categories include events used to describe pneumonitis, interstitial nephritis, diarrhea or colitis, hepatitis, rash, endocrinopathies, and hypersensitivity or infusion reactions. The registry states the aim was to estimate incidence and characterize outcomes of high-grade select events.
The clinical question is therefore principally a safety-description question: how many participants experienced the specified high-grade, treatment-related events during the defined observation period, and what were the outcomes of those events? The posted registry record gives an endpoint definition and separate adverse-event summaries, but it does not provide a single comparative treatment-effect estimate for the primary endpoint. In particular, the reported serious adverse-event counts should not be read as the primary endpoint counts: the endpoint is restricted to select events, specified grades, and treatment-related events. The registry record does not report a different-treatment control group or a primary endpoint contrast between treatment strategies.
3. Trial Design
The registry lists this as a Phase 3, randomized, parallel trial with no masking. Its official title describes it as a Phase IIIb/IV safety trial. The condition is non-small cell lung cancer, and the stated study purpose is to estimate the incidence and characterize outcomes of high-grade select adverse events in participants treated with nivolumab. The registry lists Bristol-Myers Squibb as sponsor and reports the study as completed.
The enrollment value is 1428. The intervention listed is nivolumab. Cohort A received nivolumab 3 mg/kg by intravenous infusion over 30 minutes every two weeks until disease progression, unacceptable toxicity, or withdrawal of informed consent. Cohort B received the same dose and schedule until one year, described as 52 weeks; treatment was to be discontinued, and retreatment was allowed at progression. Both cohorts are described as nivolumab monotherapy, so the cohort distinction concerns the stated treatment limit rather than a different listed drug.
The primary endpoint observation period is described as beginning at first dose and continuing until 100 days after the last dose, with a note that the last dose for Cohort B is up to randomization. The record gives the possible duration as up to approximately 88 months. The listed study start is 2014-04-09, and primary completion is 2021-10-06. These registry dates describe the study record and do not themselves establish how long any individual participant was followed.
The record describes the primary population as all treated participants before randomization and treated participants randomized to Cohorts A and B. The precise sequence and rationale for the pre-randomization and post-randomization groupings are not reported in the registry record. Nor is a different-treatment control arm listed. Consequently, randomization of participants between the two nivolumab cohorts should not be mistaken for a randomized comparison against another therapy.
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: After 1 year of treatment, participants who are still on treatment are randomized to Cohort A or Cohort B.
Overall Study
| Stage | Cohort A: Nivolumab Monotherapy | Cohort B: Nivolumab Monotherapy | Nivolumab Monotherapy (Not Randomized) |
|---|---|---|---|
| STARTED | 127 | 125 | 1176 |
| COMPLETED | 0 | 0 | 0 |
| NOT COMPLETED | 127 | 125 | 1176 |
Reasons for not completing this period:
| Reason | Cohort A: Nivolumab Monotherapy | Cohort B: Nivolumab Monotherapy | Nivolumab Monotherapy (Not Randomized) |
|---|---|---|---|
| Disease progression | 51 | 60 | 700 |
| Study drug toxicity | 13 | 2 | 58 |
| Death | 4 | 2 | 158 |
| Adverse event unrelated to study drug | 6 | 4 | 45 |
| Participant request to discontinue treatment | 12 | 12 | 70 |
| Participant withdrew consent | 10 | 18 | 61 |
| Lost to Follow-up | 0 | 2 | 2 |
| Maximum clinical benefit | 2 | 2 | 3 |
| Poor/non-compliance | 2 | 1 | 1 |
| Participant no loner meets study criteria | 3 | 2 | 28 |
| Administrative reasons by sponsor | 8 | 5 | 0 |
| Other reasons | 16 | 15 | 50 |
Analysis population for the primary outcome, as defined in the registry: All treated participants prior to randomization and treated participants randomized to Cohort A and B
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.
| Characteristic | Unit | Cohort A: Nivolumab Monotherapy | Cohort B: Nivolumab Monotherapy | Nivolumab Monotherapy (Not Randomized) | Total |
|---|---|---|---|---|---|
| Participants analysed | 127 | 125 | 1176 | 1428 | |
| Age, Continuous | Years | 66.2 (SD 10.27) | 66.6 (SD 7.97) | 66.1 (SD 9.82) | 66.1 (SD 9.71) |
| Sex: Female, Male: Female | Participants | 67 | 62 | 528 | 657 |
| Sex: Female, Male: Male | Participants | 60 | 63 | 648 | 771 |
| Ethnicity (NIH/OMB): Hispanic or Latino | Participants | 3 | 7 | 31 | 41 |
| Ethnicity (NIH/OMB): Not Hispanic or Latino | Participants | 124 | 118 | 1139 | 1381 |
| Ethnicity (NIH/OMB): Unknown or Not Reported | Participants | 0 | 0 | 6 | 6 |
| Race (NIH/OMB): American Indian or Alaska Native | Participants | 0 | 1 | 4 | 5 |
| Race (NIH/OMB): Asian | Participants | 3 | 1 | 38 | 42 |
| Race (NIH/OMB): Native Hawaiian or Other Pacific Islander | Participants | 0 | 0 | 2 | 2 |
| Race (NIH/OMB): Black or African American | Participants | 13 | 8 | 84 | 105 |
| Race (NIH/OMB): White | Participants | 111 | 114 | 1041 | 1266 |
| Race (NIH/OMB): More than one race | Participants | 0 | 0 | 0 | 0 |
| Race (NIH/OMB): Unknown or Not Reported | Participants | 0 | 1 | 7 | 8 |
6. Primary Endpoint
The registry lists the primary outcome as The Number of Participants With Treatment Related Select Adverse Events (Grade 3-4 and Grade 5), measured over From first dose and 100 days after last dose (last dose up to randomization for cohort B) (up to approximately 88 months).
| Element | Registry entry |
|---|---|
| Result type | Count Of Participants |
| Measure of spread | not applicable |
| Unit | Participants |
| Endpoint classification | Binary Endpoint |
The primary endpoint is the number of participants with treatment-related select adverse events graded 3, 4, or 5. The registry defines a treatment-related adverse event as a new untoward medical occurrence or worsening of a pre-existing medical condition with a causal relationship to the investigational treatment. The select event categories cover pneumonitis, interstitial nephritis, diarrhea or colitis, hepatitis, rash, endocrinopathies, and hypersensitivity or infusion reactions. The record defines Grade 3 as severe, Grade 4 as life-threatening, and Grade 5 as death, using NCI CTCAE Version 4.0.
As a count endpoint, this measure describes how many participants met the specified event definition over the stated observation period. It can help characterize the burden of serious, clinically important toxicity in a safety study. It does not by itself describe when events began, how long they lasted, or whether treatment caused them beyond the registry’s stated attribution process. The registry lists time to onset and time to resolution among secondary outcomes, alongside medication use for select events and counts of high-grade select events.
The posted adverse-event summaries are not interchangeable with this endpoint. They include serious adverse events and other adverse events, while the primary endpoint is limited to select event types, specified grades, and treatment-related events. The registry record does not report primary endpoint counts in a form that yields one comparison between the randomized cohorts.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The registry’s primary endpoint is a participant count, but the record does not report a single primary endpoint comparison or an inferential estimate for it. The registry notes that the primary outcome has several categories or time points and is not reduced to one comparison. No confidence interval, p-value, or adjusted treatment contrast for this endpoint is reported in the registry record. The available material therefore supports a description of the endpoint definition and safety summaries, not a claim about a comparative effect.
The serious adverse-event summary lists 57 affected participants among 127 at risk in post-randomization Cohort A, and 52 among 125 at risk in post-randomization Cohort B. For the pre-randomized nivolumab group, it lists 870 affected among 1428 at risk; the total nivolumab group lists 918 affected among 1428 at risk. These are serious adverse-event summary counts, not necessarily counts of treatment-related select events meeting the primary endpoint definition. The record’s separate other-adverse-event counts also have different event scope: 101 among 127 for Cohort A, 82 among 125 for Cohort B, 1287 among 1428 for the pre-randomized group, and 1306 among 1428 for the total group.
The figures should be interpreted within the registry’s stated group labels. The record does not explain the relationship between pre-randomized and post-randomization summaries in enough detail to reconstruct participant flow or derive a cohort contrast. It also does not report adjustment for follow-up duration, exposure time, or baseline characteristics in these counts. A direct numerical comparison between groups could therefore be misleading, especially because the cohorts have different stated treatment limits and the observation period can depend on treatment and follow-up.
The serious-event term list includes malignant neoplasm progression at 36.5%, pneumonia at 13.4%, dyspnoea at 6.4%, chronic obstructive pulmonary disease at 5.6%, colitis at 3.1%, and pulmonary embolism at 3.0%. The denominator and counting rules for these percentages are not reported in the registry record. They describe the listed serious-event terms, not the frequency of the primary endpoint categories, and should not be interpreted as treatment-attributable rates.
The registry also lists deaths in the adverse-event summaries: 63 for Cohort A, 77 for Cohort B, 1040 for the pre-randomized group, and 1180 for the total group. The record does not specify in this summary whether these counts represent deaths attributed to treatment, deaths from any cause, or another defined category. No causal conclusion should be drawn from these counts alone. Overall, the posted results are descriptive safety summaries, with key details needed for comparative or causal interpretation not reported in the registry record.
9. Statistical Methods Explained
The primary endpoint is a count of participants with a defined event, rather than a continuous measurement or a time-to-event estimate. For a count endpoint, the basic descriptive quantities are the number meeting the endpoint and the number in the population under observation. A proportion can be estimated from those quantities, but no new proportion is calculated because the posted serious adverse-event counts do not match the primary endpoint’s restricted definition. The registry record does not report a single primary endpoint numerator and denominator for each cohort in a way that supports a direct comparison.
What does randomization establish?
General statistical explanation: Random allocation can support a fair comparison when groups are assigned to meaningfully different strategies and analyzed according to a prespecified plan. It helps balance measured and unmeasured characteristics on average, but does not guarantee identical groups. In this trial, the registry identifies randomized parallel assignment, but both listed cohorts received nivolumab and differ in their stated treatment limits. No different-treatment comparator is listed. Thus, randomization should not be described as establishing a nivolumab-versus-control effect from the information in the registry record.
The record does not report the allocation ratio, randomization procedure, concealment process, or details of how the pre-randomized group relates to the randomized cohorts. These omissions matter for interpreting the group summaries and reconstructing the analysis population. They cannot be filled in by assuming a conventional design or a particular participant flow.
How should a confidence interval be understood?
General statistical explanation: A confidence interval expresses uncertainty around an estimated quantity under a specified statistical model and sampling framework. A narrow interval indicates greater precision for that estimate under those assumptions; a wide interval indicates less precision. It does not correct bias caused by missing observations, unclear outcome definitions, or non-comparable groups. The registry record does not report a confidence interval for the primary endpoint, so its precision cannot be assessed from an interval estimate.
Likewise, a p-value would address compatibility between data and a specified null hypothesis under an analysis plan; it would not establish that a treatment caused an outcome or measure the clinical importance of an effect. No primary endpoint p-value is reported in the registry record. The record lists zero analyses, and its primary outcome is not reduced to one comparison. There is therefore no reported hypothesis test to interpret for the primary endpoint.
Why do event definitions matter?
General statistical explanation: Safety outcomes depend on what counts as an event, its grade, attribution, observation period, and the population included. A broad all-event count and a count limited to select, treatment-related, high-grade events answer different questions. The primary endpoint uses a specified set of event categories and grades and requires treatment-related attribution. The adverse-event summary counts include serious adverse events and other adverse events. Those summaries cannot be substituted for the primary endpoint merely because they concern safety.
Attribution is a clinical classification recorded under the study’s procedures; a treatment-related label is not, by itself, proof of causation. Conversely, counts of events without treatment attribution do not isolate treatment effects. The registry definition explains the intended attribution concept, but the adjudication process and its reliability are not reported in the registry record.
What is the role of follow-up time?
General statistical explanation: Participants observed for longer may have more opportunity to experience an adverse event. Counts and proportions do not automatically account for unequal follow-up or differences in treatment exposure. Time-to-event methods can account for the timing of first events under specified assumptions, while exposure-adjusted rates use person-time denominators. The registry lists median time to onset and resolution as secondary outcomes, but the analysis methods, estimates, and censoring rules are not reported in the registry record. It is therefore not possible to determine from the posted counts how differences in exposure or follow-up were handled.
The primary observation window is defined from first dose through 100 days after last dose, with the Cohort B note referring to last dose up to randomization; the possible overall duration is described as up to approximately 88 months. The registry does not provide participant-level follow-up distributions, exposure time, or a detailed explanation of the Cohort B timing note. General concerns about opportunity for event detection therefore remain relevant to interpretation.
What can the posted summaries support?
The serious adverse-event summaries support a descriptive account of counts within the listed groups. The registry identifies 57 affected participants among 127 at risk in Cohort A, 52 among 125 in Cohort B, 870 among 1428 in the pre-randomized group, and 918 among 1428 in the total group. These values are not recalculated primary endpoint estimates. The record also gives other-adverse-event counts and death counts, but does not give enough detail to establish their relation to treatment attribution or the primary endpoint definition.
The registry lists several secondary outcomes, including time to onset, time to resolution, use and duration of immune-modulating medication, prednisone-equivalent medication use, and high-grade select adverse events. Listing an outcome does not establish that its results or analysis are available. The registry record does not report the statistical models, confidence intervals, missing-data rules, or multiplicity procedures for these outcomes. Accordingly, the sound statistical reading is descriptive and bounded: the available counts identify recorded safety events in named groups, but do not provide a complete comparative analysis of the primary endpoint.
10. Multiplicity, Missing Data and Analysis Populations
The registry lists a primary endpoint and several secondary outcomes. These include median time to onset and resolution of select high-grade adverse events, counts of participants receiving immune-modulating medication or hormonal replacement therapy, use of at least 40 mg prednisone equivalents, total duration of immune-modulating medication, and counts with high-grade select adverse events. Testing or estimating many outcomes can increase the chance of at least one apparently notable finding when multiple statistical tests are conducted. General statistical explanation: A multiplicity plan can specify which outcomes are confirmatory and how uncertainty is handled across analyses. The multiplicity approach is not reported in the registry record.
The record does not report a detailed statistical analysis plan, hypothesis hierarchy, interim analysis plan, or adjustment method. It lists zero analyses and notes that the primary outcome has several categories or time points and is not reduced to one comparison. No single primary effect estimate, confidence interval, or p-value is reported. Whether any analyses were conducted outside the registry record is not established by these registry entries.
Missing-data handling is also not reported in the registry record. The record does not describe how incomplete follow-up, treatment discontinuation, withdrawal, or missing event assessments were handled in the endpoint analyses. These details matter because participants may differ in the time they remain under observation, and a participant without a recorded event is not necessarily equivalent to a participant observed throughout the full endpoint window. The registry supplies an observation-period definition, but not enough information to assess completeness of follow-up or the assumptions used for missing outcomes.
The population description includes treated participants before randomization and treated participants randomized to Cohorts A and B. The handling of participants who were not treated, did not enter randomization, or lacked endpoint data is not reported in the registry record. The analysis population rules and any sensitivity analyses are likewise not reported. Without these details, the posted counts should be read as registry summaries rather than as a fully specified inferential analysis.
11. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Cohort A: Nivolumab Monotherapy (Post-Randomization) | 57 | 127 | 44.9% |
| Cohort B: Nivolumab Monotherapy (Post-Randomization) | 52 | 125 | 41.6% |
| Nivolumab Monotherapy (Pre-Randomized) | 870 | 1428 | 60.9% |
| Nivolumab Monotherapy (Total) | 918 | 1428 | 64.3% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Cohort A: Nivolumab Monotherapy (Post-Randomization) | 63 of 127 |
| Cohort B: Nivolumab Monotherapy (Post-Randomization) | 77 of 125 |
| Nivolumab Monotherapy (Pre-Randomized) | 1040 of 1428 |
| Nivolumab Monotherapy (Total) | 1180 of 1428 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 5% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Cohort A: Nivolumab Monotherapy (Post-Randomization) | 101 of 127 |
| Cohort B: Nivolumab Monotherapy (Post-Randomization) | 82 of 125 |
| Nivolumab Monotherapy (Pre-Randomized) | 1287 of 1428 |
| Nivolumab Monotherapy (Total) | 1306 of 1428 |
Most frequent serious adverse event terms
| Event term | Organ system | Cohort A: Nivolumab Monotherapy (Post-Randomization) | Cohort B: Nivolumab Monotherapy (Post-Randomization) | Nivolumab Monotherapy (Pre-Randomized) | Nivolumab Monotherapy (Total) |
|---|---|---|---|---|---|
| Malignant neoplasm progression | Neoplasms benign, malignant and unspecified (incl cysts and polyps) | 11 of 127 | 15 of 125 | 509 of 1428 | 521 of 1428 |
| Pneumonia | Infections and infestations | 17 of 127 | 7 of 125 | 118 of 1428 | 137 of 1428 |
| Dyspnoea | Respiratory, thoracic and mediastinal disorders | 1 of 127 | 1 of 125 | 91 of 1428 | 92 of 1428 |
| Pulmonary embolism | Respiratory, thoracic and mediastinal disorders | 2 of 127 | 1 of 125 | 40 of 1428 | 43 of 1428 |
| Pleural effusion | Respiratory, thoracic and mediastinal disorders | 1 of 127 | 2 of 125 | 37 of 1428 | 38 of 1428 |
| Dehydration | Metabolism and nutrition disorders | 2 of 127 | 0 of 125 | 33 of 1428 | 35 of 1428 |
| Chronic obstructive pulmonary disease | Respiratory, thoracic and mediastinal disorders | 5 of 127 | 7 of 125 | 27 of 1428 | 31 of 1428 |
| Sepsis | Infections and infestations | 3 of 127 | 2 of 125 | 22 of 1428 | 26 of 1428 |
| Pneumonitis | Respiratory, thoracic and mediastinal disorders | 1 of 127 | 0 of 125 | 23 of 1428 | 24 of 1428 |
| Respiratory failure | Respiratory, thoracic and mediastinal disorders | 1 of 127 | 1 of 125 | 22 of 1428 | 23 of 1428 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Cohort A: Nivolumab Monotherapy (Post-Randomization) | Cohort B: Nivolumab Monotherapy (Post-Randomization) | Nivolumab Monotherapy (Pre-Randomized) | Nivolumab Monotherapy (Total) |
|---|---|---|---|---|---|
| Fatigue | General disorders | 26 of 127 | 17 of 125 | 516 of 1428 | 544 of 1428 |
| Dyspnoea | Respiratory, thoracic and mediastinal disorders | 14 of 127 | 17 of 125 | 339 of 1428 | 359 of 1428 |
| Nausea | Gastrointestinal disorders | 22 of 127 | 9 of 125 | 332 of 1428 | 352 of 1428 |
| Cough | Respiratory, thoracic and mediastinal disorders | 27 of 127 | 22 of 125 | 302 of 1428 | 331 of 1428 |
| Decreased appetite | Metabolism and nutrition disorders | 18 of 127 | 13 of 125 | 313 of 1428 | 336 of 1428 |
| Arthralgia | Musculoskeletal and connective tissue disorders | 30 of 127 | 18 of 125 | 284 of 1428 | 314 of 1428 |
| Diarrhoea | Gastrointestinal disorders | 36 of 127 | 17 of 125 | 275 of 1428 | 292 of 1428 |
| Constipation | Gastrointestinal disorders | 9 of 127 | 6 of 125 | 274 of 1428 | 290 of 1428 |
| Anaemia | Blood and lymphatic system disorders | 11 of 127 | 7 of 125 | 208 of 1428 | 222 of 1428 |
| Back pain | Musculoskeletal and connective tissue disorders | 17 of 127 | 11 of 125 | 189 of 1428 | 211 of 1428 |
Serious events with the highest arm-level rate: malignant neoplasm progression (36.5%), pneumonia (13.4%), dyspnoea (6.4%), chronic obstructive pulmonary disease (5.6%), colitis (3.1%).
The serious adverse-event summaries distinguish participants at risk from participants affected. Cohort A lists 127 at risk and 57 affected; Cohort B lists 125 at risk and 52 affected. The pre-randomized nivolumab group lists 1428 at risk and 870 affected, while the total nivolumab group lists 1428 at risk and 918 affected. The separate other-adverse-event counts are 101 affected in Cohort A, 82 in Cohort B, 1287 in the pre-randomized group, and 1306 in the total group.
These are registry safety summaries, not the primary endpoint result. They do not establish that an event was treatment-related or that it was one of the select high-grade categories. The record lists serious-event terms and percentages, including malignant neoplasm progression at 36.5% and pneumonia at 13.4%, but does not report the percentage denominator or counting method. The term percentages should not be treated as participant-level risks or as rates of treatment-related select events.
Deaths are listed as 63 in Cohort A, 77 in Cohort B, 1040 in the pre-randomized group, and 1180 in the total group. The registry summary does not specify the attribution or definition attached to these death counts. They should not be interpreted as treatment-caused deaths. Differences between group counts also should not be treated as comparative effects without clear population definitions, follow-up information, and an analysis plan.
12. What the Result Does, and Does Not, Mean
The results establish that the registry reports safety-event counts for the listed nivolumab groups and defines a primary endpoint focused on treatment-related select adverse events graded 3, 4, or 5. The endpoint is relevant to describing severe safety outcomes because it restricts the event categories, severity grades, and attribution required. The record also lists secondary safety measures concerning onset, resolution, and medication use.
The results do not, in the form reported, establish a comparative estimate of the primary endpoint between the randomized cohorts. Both cohorts received nivolumab, and the registry does not list a different-treatment comparator. The record does not give one primary endpoint numerator and denominator per cohort, an effect estimate, a confidence interval, or a p-value. The serious adverse-event counts are broader summaries and should not be substituted for the restricted primary endpoint.
Nor do the counts prove that nivolumab caused every event categorized as treatment-related, or that events without that label were unrelated. Treatment attribution is a recorded assessment. The death counts do not specify attribution in the posted summary and therefore do not support a conclusion about treatment-caused mortality. The listed serious-event term percentages also lack a stated denominator in the registry record, limiting interpretation of their frequency.
General statistical explanation: Descriptive counts can be useful for identifying the range and burden of events recorded during a study, but comparative claims require clear group definitions, suitable denominators, adequate follow-up information, and an analysis that accounts for the study design. Precision requires interval estimates or another uncertainty measure; none is reported for the primary endpoint. The absence of a reported inferential result should not be confused with evidence of no difference, just as a reported count alone does not establish a treatment effect.
The most defensible reading is therefore limited: the registry describes a randomized safety study of nivolumab in advanced or metastatic non-small cell lung cancer and posts adverse-event summaries. It does not provide enough primary endpoint analysis detail to quantify a comparative effect or its precision. The findings should be understood as registry-reported safety information, not as a complete evaluation of benefit, harm, or comparative effectiveness.
13. Why This Trial Matters Statistically
Safety trials can address questions that are not captured by a single measure of tumor response. This study’s stated purpose was to estimate incidence and characterize outcomes of high-grade select adverse events among people with advanced or metastatic non-small cell lung cancer treated with nivolumab. Its primary endpoint focuses on treatment-related events in specified clinical categories and grades, while secondary outcomes include event timing and medication use. That structure reflects the statistical distinction between counting whether an event occurred and describing when it occurred or how it was managed.
The record also illustrates why endpoint definitions must be kept separate. Serious adverse events, other adverse events, and treatment-related select events are not interchangeable sets. A count can only answer the question attached to its definition and denominator. The posted serious-event summaries therefore contribute safety context, but they cannot be taken as the primary endpoint’s result without matching event criteria and analysis populations.
Randomized assignment is an important design feature, but its meaning depends on the contrast under study. Here, the two listed randomized cohorts received the same drug and differed in their treatment limits. The record does not list a different-treatment control group or a single comparative primary endpoint analysis. This is a useful reminder that the word randomized alone does not identify the treatment effect being estimated.
Finally, registry reporting makes it possible to identify major gaps relevant to statistical interpretation. The record does not report a detailed analysis plan, missing-data methods, multiplicity procedures, or uncertainty estimates for the primary endpoint. These are not merely technical details: they determine how counts are turned into estimates and how confidently results can be interpreted. The trial is statistically informative as a safety-focused registry record, while its posted summaries leave important questions about comparative analysis and precision unresolved.
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 main limitation for interpreting the primary endpoint is that the registry record does not report a single endpoint result for a clearly defined comparison. It identifies the endpoint, event categories, severity grades, attribution concept, and observation window, but the posted serious adverse-event counts do not match all those restrictions. The record itself notes that the primary outcome has several categories or time points and is not reduced to one comparison.
The study is described as randomized, but both listed cohorts received nivolumab. Their treatment limits differ, and no different-treatment comparator is listed. The registry does not report the allocation ratio, randomization procedure, participant flow, or the relationship between the pre-randomized and post-randomization summaries in enough detail to reconstruct the analysis populations. A comparison of raw counts across these groups would not account for differences in treatment duration, follow-up, or group construction.
Important statistical methods are not reported in the registry record: the primary analysis model, confidence intervals, hypothesis tests, missing-data handling, multiplicity approach, and any sensitivity analyses. No estimate of precision for the primary endpoint is available. The serious-event term percentages are listed without a reported denominator or counting rule, and the death summaries do not specify attribution in the posted information.
The record’s primary observation period extends from first dose to 100 days after last dose, with a Cohort B note referring to last dose up to randomization and an overall possible duration of up to approximately 88 months. Individual follow-up distributions and exposure time are not reported. These limitations prevent a complete assessment of event opportunity, censoring, or comparative safety. The posted information is therefore best treated as descriptive registry reporting rather than a fully specified comparative safety analysis.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: A Safety Trial of Nivolumab in Patients With Advanced or Metastatic Non-Small Cell Lung Cancer Who Have Progressed During or After Receiving At Least One Prior Chemotherapy Regimen — NCT02066636.
- Registry record last updated 2022-10-27; status last verified 2022-09.
- Results first posted 2022-10-27.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
18. Record Summary
The registry describes a completed, randomized, parallel, unmasked Phase 3 study of nivolumab in advanced or metastatic non-small cell lung cancer. The official title calls it a Phase IIIb/IV safety trial. The listed enrollment is 1428, and Bristol-Myers Squibb is named as sponsor. The study start is 2014-04-09 and primary completion is 2021-10-06.
The primary endpoint counts participants with treatment-related select adverse events graded 3, 4, or 5. Both listed cohorts received nivolumab, with different treatment limits. The registry posts serious and other adverse-event counts, along with serious-event terms, but these summaries are not equivalent to the restricted primary endpoint.
A single comparative primary endpoint estimate, its uncertainty, and details of missing-data and multiplicity methods are not reported in the registry record. The posted information supports descriptive safety reading, but not a quantified comparative effect.