This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
EVEREST was a randomized, parallel-group, open-label phase 3 treatment trial in de novo renal transplant recipients. The study enrolled 223 participants and evaluated two everolimus/cyclosporine strategies, with primary endpoints covering biopsy-proven acute rejection through Month 24 and renal function at Months 12, 18, and 24.
| Feature | EVEREST |
|---|---|
| Phase | Phase 3 |
| Status | Completed |
| Population | De novo renal transplant recipients |
| Design | Randomized, parallel, open-label treatment trial |
| Allocation | Randomized |
| Enrollment | 223 |
| Number of arms | 2 |
| Primary endpoints | Number of participants with biopsy-proven acute rejection; renal function assessed by creatinine clearance |
| Study start | May 2006 |
| Primary completion | February 2009 |
| Lead sponsor | Novartis Pharmaceuticals |
| Sponsor type | Industry |
2. Clinical Question
The statistical question is whether different combinations of everolimus and cyclosporine microemulsion in de novo renal transplant recipients are associated with differences in acute rejection and renal function during the study period.
Population
De novo renal transplant recipients enrolled in the phase 3 EVEREST study.
Intervention strategy
Everolimus 0.25 and 0.75 mg tablets combined with cyclosporine microemulsion at a very low dose, with a target of 150–300 ng/mL.
Comparator strategy
Everolimus 0.25 and 0.75 mg tablets combined with cyclosporine microemulsion at a low dose, with a target of 350–500 ng/mL.
Primary questions
How many participants experience biopsy-proven acute rejection by Month 24, and how does creatinine clearance characterize renal function at Months 12, 18, and 24?
3. Trial Design
Treatment strategies
Upper Everolimus Blood Target + Very Low Cyclosporine
- Everolimus 0.25 and 0.75 mg tablets
- Cyclosporine very low-dose microemulsion
- Cyclosporine target: 150–300 ng/mL
Standard Everolimus Blood Target + Low Cyclosporine
- Everolimus 0.25 and 0.75 mg tablets
- Cyclosporine low-dose microemulsion
- Cyclosporine target: 350–500 ng/mL
The key design contrast is therefore the cyclosporine exposure strategy accompanying everolimus. Because the study was open-label, treatment assignment was not masked to participants or investigators. Randomization remains important statistically because it establishes the treatment comparison before outcomes are observed.
4. Endpoints
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Number of Participants With Biopsy-proven Acute Rejection | A graft core biopsy was performed on all suspected acute rejection episodes within 48 hours. Biopsies were read by the local pathologist according to the 1997 Banff criteria. A biopsy-proven acute rejection was defined as a biopsy graded IA, IB, IIA, IIB, or III. | Baseline to end of study (Month 24) |
| Renal Function Assessed by Creatinine Clearance | Renal function was assessed by measuring serum creatinine and by computing creatinine clearance using the formula of Cockcroft-Gault. | Month 12, Month 18, and Month 24 |
Why the two primary endpoints are statistically different
Biopsy-proven acute rejection is an event count: each participant can be classified according to whether a qualifying rejection episode occurred during the specified period. Creatinine clearance is a quantitative measure of renal function assessed at specified follow-up times. These endpoint types require different statistical thinking.
For the rejection endpoint, the basic comparison concerns the proportion of participants experiencing a qualifying event. Depending on the prespecified statistical analysis plan, an event proportion can be compared with an absolute risk difference, a risk ratio, an odds ratio, or a time-to-event method when timing is incorporated.
For creatinine clearance, the central question is instead whether the distributions or means of renal function differ between randomized groups at the specified assessment times. A longitudinal analysis can also account for the repeated measurements within the same participant when the statistical plan is designed to evaluate change over time.
5. Statistical Methodology
The registry identifies the endpoints and their assessment procedures but does not post formal statistical analyses for these endpoints. The appropriate analysis therefore depends on the endpoint structure and the prespecified statistical analysis plan.
A risk difference expresses the absolute difference in the probability of biopsy-proven acute rejection between the randomized treatment strategies over the specified follow-up period.
For a binary endpoint, an analysis can also express the treatment comparison as a risk ratio or odds ratio. The choice matters because these measures have different interpretations. A risk ratio compares probabilities directly, whereas an odds ratio compares odds rather than probabilities.
A mean difference describes the average separation in creatinine clearance between the two randomized groups at a specified assessment time.
Because creatinine clearance is assessed at Month 12, Month 18, and Month 24, repeated-measures methods may be useful when the objective is to evaluate the trajectory of renal function rather than treating every assessment as an unrelated comparison. Such models can account for correlation among measurements from the same participant.
Creatinine clearance and the Cockcroft-Gault formula
The registry specifies that creatinine clearance was computed using the Cockcroft-Gault formula after measuring serum creatinine. Conceptually, the calculation transforms serum creatinine and participant characteristics into an estimate of creatinine clearance. The statistical endpoint is therefore a derived quantitative measure rather than a directly observed laboratory concentration.
That distinction matters because uncertainty in the underlying measurements and the formula-based derivation can contribute to variability in the resulting renal-function endpoint.
Biopsy classification
The acute-rejection endpoint uses a clinically defined classification rule. Suspected episodes triggered a graft core biopsy within 48 hours, and the biopsy was read according to the 1997 Banff criteria. Only grades IA, IB, IIA, IIB, or III qualified as biopsy-proven acute rejection under the registry definition.
Statistically, this creates a prespecified binary outcome from a clinical and pathological assessment process. The definition is important because changing the threshold for what constitutes an event can change the observed event rate even when the underlying clinical experience is unchanged.
6. Planned Analysis
No formal statistical analyses were posted to ClinicalTrials.gov. The registry does, however, define two primary endpoints whose statistical treatment follows established methods for randomized clinical trials.
Analysis of biopsy-proven acute rejection
The primary rejection endpoint is evaluated from baseline through Month 24. A natural analysis begins by calculating, within each randomized treatment arm, the proportion of participants who experience a biopsy-proven acute rejection according to the specified Banff-based definition.
The between-group comparison can then be expressed using an absolute risk difference and a corresponding confidence interval. A risk ratio or odds ratio can provide a relative measure. If the exact timing of rejection events and censoring is incorporated into the prespecified analysis, a time-to-event approach could instead or additionally be used.
The important distinction is between what happened and how it was summarized. The clinical endpoint identifies participants who experienced a qualifying rejection event; the statistical analysis quantifies how the event experience differs between randomized groups.
Analysis of creatinine clearance
Creatinine clearance is assessed at three specified time points: Month 12, Month 18, and Month 24. A simple analysis could compare the two randomized groups separately at each time point, while a longitudinal model could analyze all measurements together and account for within-participant correlation.
For repeated measurements, a model may estimate an overall treatment difference, a time effect, and potentially a treatment-by-time interaction. The interaction is relevant when the scientific question concerns whether the separation between treatment strategies changes over follow-up.
A single creatinine-clearance comparison at one time point answers a different question from a longitudinal treatment comparison. The former describes a cross-sectional difference at that assessment; the latter evaluates the pattern of renal-function measurements across follow-up.
7. Safety Results
The registry reports serious adverse events by treatment arm. These counts describe participants affected by serious adverse events relative to the number at risk in each arm.
| Safety measure | Participants affected | Participants at risk |
|---|---|---|
| Upper Everolimus Blood Target + Very Low | 85 | 142 |
| Standard Everolimus Blood Target + Low D | 90 | 143 |
These are serious-adverse-event counts by arm rather than a formal comparative effect estimate. The numerator identifies participants affected, while the denominator identifies participants at risk. A formal statistical comparison would require specifying the estimand, analysis population, handling of multiple events within a participant, and whether the endpoint is being treated as a binary participant-level outcome or a recurrent-event outcome.
8. Statistical Methods Explained
Why was randomization important?
Randomization assigns participants to treatment strategies without using the eventual outcome to determine assignment. This creates the framework for comparing outcomes between groups while reducing systematic differences in treatment assignment. With 223 participants randomized across two parallel arms, the primary comparison is between the outcomes observed under the two assigned strategies.
Why is biopsy-proven acute rejection a binary endpoint?
The registry definition establishes whether a participant experienced an episode meeting the specified biopsy criteria during baseline through Month 24. That naturally produces an event indicator: qualifying rejection versus no qualifying rejection during the defined period. Binary-event methods can then summarize and compare the resulting proportions.
Why does the 48-hour biopsy window matter?
The registry specifies that a graft core biopsy was performed on suspected acute rejection episodes within 48 hours. This creates a standardized relationship between the clinical suspicion of rejection and the pathological assessment used to classify the endpoint. Consistent assessment rules reduce ambiguity in determining whether an episode qualifies as the primary endpoint.
Why are Months 12, 18, and 24 different from the rejection endpoint?
The creatinine-clearance endpoint is measured at three scheduled assessment times rather than being defined simply by whether an event occurred during follow-up. This makes the endpoint longitudinal. Statistical analysis must therefore distinguish differences at a particular time from differences in the overall trajectory of renal function.
What does a mean difference in creatinine clearance tell us?
A mean difference describes the average separation between the two treatment groups in creatinine clearance. A positive or negative value must be interpreted according to the direction in which the groups are defined. It does not describe the response of every individual participant, and it does not by itself establish whether the difference is statistically distinguishable from random variation.
Why are confidence intervals useful?
A confidence interval places the estimated treatment contrast in the context of statistical uncertainty. A narrow interval indicates greater precision under the specified model and sampling framework; a wide interval indicates less precision. The interval does not describe the range of outcomes that individual participants will experience.
Why should the serious-adverse-event counts not be treated as a complete safety analysis?
The registry reports participants affected and participants at risk for serious adverse events by arm. That is useful descriptive information, but a complete statistical safety analysis would also depend on the endpoint definition, timing, exposure, severity, recurrent events, treatment discontinuation, and the chosen comparison measure.
9. Interpreting the Two Primary Endpoints Together
The EVEREST endpoint structure illustrates why a clinical trial can require more than one statistical lens. Biopsy-proven acute rejection and creatinine clearance measure different dimensions of treatment experience.
| Endpoint | Data type | Primary statistical question |
|---|---|---|
| Biopsy-proven acute rejection | Binary clinical/pathological event | Does the proportion experiencing a qualifying rejection differ between randomized groups? |
| Creatinine clearance | Quantitative renal-function measure | Does renal function differ between randomized groups at Months 12, 18, and 24? |
A treatment strategy could show a difference in one endpoint without showing the same magnitude of difference in the other. That is not a statistical contradiction. Acute rejection is an event-based endpoint, whereas creatinine clearance provides a continuous measure of renal function.
The most informative interpretation therefore keeps the endpoint definitions separate before considering them together. The rejection analysis addresses occurrence of a clinically defined event; the creatinine-clearance analysis addresses quantitative renal function at specified times.
10. Precision, Effect Size, and Statistical Significance
For the EVEREST endpoints, an effect estimate would answer how much the groups differ, while a confidence interval would address how precisely that difference has been estimated. A p-value, if part of a formal analysis, would address the compatibility of the observed data with a specified null hypothesis under the statistical model.
Effect size
A risk difference, risk ratio, odds ratio, or mean difference describes the magnitude of a treatment contrast. The choice of measure changes the way the result is interpreted.
Confidence interval
The interval communicates statistical precision around the estimated contrast. It should be considered alongside the point estimate rather than treated as a substitute for it.
P-value
A p-value is not a measure of effect size. A very small p-value can accompany a small effect when the sample is sufficiently informative, while a clinically important estimate can have substantial uncertainty.
Clinical meaning
Statistical evidence and clinical importance are related but distinct. A numerical treatment difference must be interpreted in the context of the endpoint, its measurement, and the transplant setting.
11. Repeated Measures and Renal Function
The creatinine-clearance endpoint is unusual from a simple two-group comparison perspective because the registry specifies measurements at Month 12, Month 18, and Month 24. The same participant can therefore contribute multiple renal-function observations.
A repeated-measures model can represent the relationship between treatment, follow-up time, and the treatment-by-time pattern while accounting for correlation among measurements from the same participant.
The treatment-by-time term is especially useful when the scientific question concerns whether the difference between treatment strategies changes from Month 12 to Month 18 to Month 24. If the treatment contrast is stable, the interaction may contribute little; if the trajectories diverge or converge, the interaction becomes more informative.
Any particular covariance structure, estimation method, missing-data strategy, or model specification would need to come from the prespecified statistical analysis plan. The registry identifies the measurement schedule and the Cockcroft-Gault calculation but does not post those additional modeling details.
12. Acute Rejection as an Event Endpoint
The biopsy-proven acute rejection endpoint has a clear clinical definition but several statistical choices remain possible. If the endpoint is analyzed simply as occurrence by Month 24, every participant contributes a binary indicator. If timing of rejection is incorporated, participants can instead contribute follow-up time until rejection or censoring.
These approaches answer related but different questions. A Month 24 event proportion asks about cumulative occurrence by a fixed point. A time-to-event analysis asks how the event process differs over follow-up while accounting for when events occur and when observations are censored.
13. Serious Adverse Events as a Safety Endpoint
The reported serious-adverse-event figures are participant-level affected counts with denominators. The two arms have 142 and 143 participants at risk, respectively, according to the registry.
The bars above are visual representations of the reported affected and at-risk counts. They should not be mistaken for a formally estimated treatment effect. The displayed widths correspond to the affected-to-at-risk proportions calculated directly from the reported counts.
Safety interpretation also differs from efficacy interpretation. A participant-level serious-adverse-event indicator summarizes whether an event occurred, but serious adverse events can vary considerably in clinical characteristics, timing, duration, and relationship to treatment. A single aggregate count therefore does not capture the full safety profile.
14. Limitations
- Open-label design: masking was listed as none. Knowledge of treatment assignment can influence some clinical assessments and behaviors, although the biopsy-based definition provides a structured endpoint classification.
- Endpoint-specific methods: the registry defines the primary endpoints but does not post formal statistical analyses for them. The appropriate model depends on the prespecified statistical analysis plan.
- Continuous endpoint structure: creatinine clearance is measured repeatedly at Months 12, 18, and 24, making correlation among observations within participants an important consideration.
- Derived renal-function measure: creatinine clearance was computed using the Cockcroft-Gault formula rather than being a directly observed clinical measurement.
- Binary endpoint definition: biopsy-proven acute rejection depends on a specified biopsy classification threshold. Changes to the endpoint definition could change the event count.
- Safety interpretation: the reported serious-adverse-event counts describe affected participants but do not by themselves provide a complete characterization of event severity, timing, recurrence, or treatment relationship.
- Sample size: 223 participants were enrolled across two randomized arms. Precision for treatment contrasts depends on the number of participants contributing information to each endpoint and the variability of the measurements.
15. Why This Trial Matters Statistically
EVEREST is a useful teaching case because its primary endpoints illustrate two fundamentally different statistical problems within one randomized trial: a clinically defined binary event and a repeated quantitative measure.
| Concept | How it appears in EVEREST |
|---|---|
| Randomization | Participants were randomized to two parallel treatment strategies. |
| Open-label design | Masking was listed as none. |
| Binary endpoint | Biopsy-proven acute rejection is defined by whether a qualifying event occurred during baseline through Month 24. |
| Pathology-based endpoint | Suspected rejection episodes were assessed using graft core biopsy and the 1997 Banff criteria. |
| Continuous endpoint | Creatinine clearance provides a quantitative assessment of renal function. |
| Repeated measurements | Creatinine clearance was assessed at Month 12, Month 18, and Month 24. |
| Derived variable | Creatinine clearance was computed using the Cockcroft-Gault formula. |
| Safety analysis | Serious adverse events were reported as participants affected relative to participants at risk by arm. |
| Endpoint-specific analysis | The two primary endpoints require different statistical frameworks because they have different data structures. |
16. Statistical Interpretation vs Clinical Interpretation
Statistical interpretation
The trial uses randomization to establish a comparison between two everolimus/cyclosporine strategies. Its primary endpoints require analysis of both event occurrence and quantitative renal-function measurements.
Clinical interpretation
The clinical questions concern acute rejection and renal function in de novo renal transplant recipients. These outcomes should be considered separately because they capture different aspects of treatment experience.
The distinction is important. Statistical methodology describes how evidence is quantified; clinical interpretation determines what that evidence means in the context of transplant care. Neither endpoint should automatically be treated as a surrogate for the other.
17. A Practical Analysis Framework
A rigorous analysis of the EVEREST endpoints can be organized into a sequence that keeps the clinical definition, statistical estimand, and uncertainty assessment aligned.
Define the population
Start with the randomized participants and specify the analysis population appropriate to the endpoint and statistical estimand.
Define the outcome
For rejection, apply the biopsy-based Banff definition. For renal function, use creatinine clearance as specified by the registry.
Choose the estimand
Decide whether the primary contrast is an event proportion, risk difference, risk ratio, odds ratio, mean difference, or longitudinal treatment effect.
Quantify uncertainty
Report confidence intervals with the point estimates so that magnitude and precision are considered together.
Interpret in context
Separate statistical evidence from clinical interpretation and consider the endpoint's definition, measurement process, and follow-up structure.
18. What the Registry Establishes
The registry establishes a phase 3 randomized parallel-group trial with no masking, 223 enrolled participants, two treatment arms, and two registered primary endpoints. The acute-rejection endpoint extends from baseline through Month 24, while renal function is assessed at Months 12, 18, and 24.
The registry also provides the operational definitions needed to understand the endpoints: suspected acute rejection was evaluated by graft core biopsy within 48 hours and classified according to the 1997 Banff criteria, while renal function was assessed using serum creatinine and creatinine clearance calculated by the Cockcroft-Gault formula.
For safety, the registry reports serious adverse events affecting 85 of 142 participants in the Upper Everolimus Blood Target + Very Low group and 90 of 143 participants in the Standard Everolimus Blood Target + Low D group.
19. Sources
- ClinicalTrials.gov: EVEREST — NCT01276457.
Continue through Clinical Biostats
Explore additional clinical trial results, biostatistics tutorials, and statistical analysis tools.
20. Record Summary
EVEREST provides a compact example of how randomized clinical-trial statistics can involve fundamentally different endpoint structures. The trial randomized 223 de novo renal transplant recipients across two parallel treatment strategies involving everolimus and different cyclosporine exposure targets. Its primary endpoints were biopsy-proven acute rejection from baseline through Month 24 and renal function assessed by creatinine clearance at Months 12, 18, and 24.
The rejection endpoint is based on a clearly specified pathological classification, with suspected episodes evaluated by graft core biopsy within 48 hours and classified using the 1997 Banff criteria. The renal-function endpoint is quantitative and repeated, with creatinine clearance calculated using the Cockcroft-Gault formula. These features lead naturally to different statistical considerations: event-proportion or time-to-event methods for rejection, and cross-sectional or longitudinal methods for creatinine clearance.
The reported serious-adverse-event counts provide an additional safety dimension, with 85 of 142 participants affected in the Upper Everolimus Blood Target + Very Low group and 90 of 143 in the Standard Everolimus Blood Target + Low D group. Together, the endpoint definitions and safety data illustrate why clinical-trial analysis must distinguish outcome type, estimand, uncertainty, and interpretation rather than reducing the study to a single numerical comparison.