This page separates reported trial results from statistical interpretation. The numerical results presented here are limited to the statistical analysis posted in the ClinicalTrials.gov record and the ClinicalTrials.gov record.
1. Trial at a Glance
CLARINET was a randomized, double-blind, parallel phase 3 trial studying lanreotide Autogel versus placebo in endocrine tumors. The registered primary endpoint was progression-free survival, a time-to-event endpoint analyzed using a log-rank test with stratification according to progression status at baseline and prior therapy.
| Feature | CLARINET |
|---|---|
| Trial name | CLARINET |
| Brief title | Study of Lanreotide Autogel in Non-functioning Entero-pancreatic Endocrine Tumours |
| Phase | Phase 3 |
| Condition | Endocrine Tumors |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Double |
| Primary purpose | Treatment |
| Enrollment | 264 |
| Interventions | Lanreotide (Autogel formulation) and placebo |
| Primary endpoint | Progression-Free Survival (PFS) |
| Primary endpoint type | Time-to-event |
| Primary analysis method | Log-rank test |
| Hypothesis type | Superiority |
| Lead sponsor | Ipsen |
| Sponsor type | Industry |
| Trial status | Completed |
| ClinicalTrials.gov identifier | NCT00353496 |
2. Clinical Question
The central statistical question was whether treatment with lanreotide Autogel changed progression-free survival compared with placebo in the randomized trial population with endocrine tumors.
Population
The trial's brief title identifies the study population as patients with non-functioning entero-pancreatic endocrine tumours; the registry condition is Endocrine Tumors.
Intervention
Lanreotide in the Autogel formulation.
Comparator
Placebo.
Primary question
Does lanreotide Autogel produce a different progression-free survival experience from placebo under a prespecified superiority comparison?
3. Trial Design
Lanreotide
- Lanreotide, Autogel formulation
- Randomized treatment assignment
- Included in the primary ITT efficacy analysis
Placebo
- Placebo
- Randomized comparator assignment
- Included in the primary ITT efficacy analysis
The ClinicalTrials.gov record does not provide an allocation ratio, arm-level randomized sample sizes, treatment dosing schedule, crossover information, or a factorial component. Those design features therefore are not inferred here.
4. Endpoints
| Endpoint | Registry definition | Time frame | Type |
|---|---|---|---|
| Progression-Free Survival (PFS) | Time from randomization to first documentation of disease progression, or death. Disease progression centrally assessed using Response Evaluation Criteria in Solid Tumours (RECIST) v1.0. | From randomisation up to the last tumour assessment (scheduled at 96 weeks). Radiological scans were performed every 12 weeks during the first year and every 24 weeks during the second year | Time-to-event |
The time frame is reproduced from the ClinicalTrials.gov record. The text ends at “every 12” in the provided record extract, so no additional unit or interval is reported here.
Why PFS is a time-to-event endpoint
PFS combines two possible events: documented disease progression and death. A patient who has not experienced either event by the relevant end of follow-up contributes information up to that point and may be censored. This is fundamentally different from a simple binary outcome measured at one fixed time.
The endpoint is defined by the first qualifying event. Patients without a qualifying event by their available follow-up contribute censored survival information rather than being treated automatically as if they had experienced progression.
5. Primary Statistical Result
The posted statistical analysis evaluates progression-free survival in the intent-to-treat population comprising 204 randomized subjects. The comparison used a log-rank test, and the effect measure was a hazard ratio.
Progression-Free Survival
Hazard ratio for progression or death
95% CI: 0.30–0.73 · P < 0.001
Lanreotide (Autogel formulation) vs placebo · Two-sided 95% confidence interval
| Primary endpoint | Analysis population | Method | Effect measure | Estimate | 95% CI | P-value |
|---|---|---|---|---|---|---|
| Progression-Free Survival (PFS) | ITT; 204 randomized subjects | Log-rank test | Hazard ratio | 0.47 | 0.30–0.73 | <0.001 |
The reported hazard ratio of 0.47 means that the estimated instantaneous rate of experiencing the PFS event—first documented progression or death—was approximately 47% as high in the lanreotide group as in the placebo group under the time-to-event analysis. Equivalently, 1 − 0.47 = 0.53, so the estimate corresponds to an approximately 53% lower estimated hazard for the PFS event.
This does not mean that 53% of patients avoided progression, that 53% of patients were cured, or that every patient experienced exactly a 53% reduction in their individual probability of progression or death. A hazard ratio is a relative time-to-event measure, not an absolute risk difference.
The 95% confidence interval of 0.30–0.73 describes the statistical uncertainty surrounding the estimated hazard ratio under the analysis framework. It does not describe the range of outcomes that individual patients can experience. Its width also matters: the estimate is 0.47, but the plausible values represented by the interval extend from 0.30 to 0.73.
The P-value of <0.001 addresses evidence against the null hypothesis used for the superiority comparison; it does not measure the magnitude of the treatment effect. Effect size is communicated by the hazard ratio and its confidence interval. A very small P-value can coexist with a modest effect in a large study, while a clinically meaningful effect can fail to reach a conventional significance threshold in a small study.
The interpretation also depends on the analysis population and censoring rules. The posted analysis used the ITT population of 204 randomized subjects, and PFS is inherently affected by when observations become censored. In addition, a hazard ratio is most straightforward to interpret as a common relative hazard when the proportional-hazards assumption is reasonable; the registry analysis does not provide a diagnostic assessment of that assumption.
6. What the Primary Result Tells Us
The primary result contains three distinct pieces of statistical information: the direction of the estimated effect, its magnitude, and the uncertainty and evidence against the null hypothesis.
Direction
An HR below 1 favors the lanreotide group for an endpoint defined as progression or death. The reported HR of 0.47 is therefore on the lower-than-1 side of the null value.
Magnitude
The point estimate corresponds to an approximately 53% lower estimated hazard of the PFS event relative to placebo. This is a relative measure, not an absolute probability.
Precision
The 95% CI extends from 0.30 to 0.73. Confidence intervals communicate uncertainty around the estimated treatment effect and should be read alongside the point estimate.
Evidence
The reported P-value is <0.001. This is evidence against the null hypothesis under the stated testing framework, not a measure of how large or clinically important the effect is.
7. Statistical Methodology
Intention-to-treat analysis
The primary analysis was based on the intent-to-treat (ITT) population, which comprised 204 randomized subjects. The ITT principle analyzes randomized subjects according to their assigned treatment group rather than redefining treatment groups according to what happened after randomization.
This matters because randomization creates the foundation for a fair comparison between treatment assignments. Restricting or reassigning patients after randomization can undermine that foundation if the decision to remain in an analysis population is related to prognosis or treatment experience.
The ITT analysis is therefore particularly natural for the primary efficacy question: what happened under the treatment strategies to which participants were randomized?
Log-rank test
The registered and posted primary method was the log-rank test. This is a standard method for comparing survival distributions between randomized groups when the outcome is time to an event.
Unlike a simple comparison of the proportion of patients who progressed, the log-rank framework uses the timing of events and the number of patients at risk over follow-up. At each event time, the observed events are compared with what would be expected under the null hypothesis of no difference between groups.
For CLARINET, the analysis was stratified according to progression status at baseline and prior therapy. Stratification allows the comparison to account for these prespecified factors when constructing the log-rank test rather than treating all patients as belonging to one unstratified risk set.
Hazard ratio
The effect measure reported for PFS was a hazard ratio. For this endpoint, the event is first documentation of disease progression or death.
The reported HR of 0.47 is a relative measure of the event hazard. The null value is 1.00: values below 1 indicate a lower estimated hazard in the numerator group, while values above 1 indicate a higher estimated hazard.
Kaplan-Meier estimation
A time-to-event endpoint such as PFS is commonly displayed with a Kaplan-Meier curve, which estimates the probability of remaining event-free over time while accounting for right censoring. The registry-reported statistical analysis identifies the endpoint as time-to-event and reports a log-rank comparison, but it does not provide the underlying event and censoring data needed to reconstruct a Kaplan-Meier curve independently.
Stratified analysis
The primary log-rank test was stratified by progression status at baseline and prior therapy. Stratification is different from performing separate treatment tests inside each subgroup. The purpose here is to incorporate the specified strata into the primary comparison while retaining one overall treatment-effect analysis.
Superiority testing
The trial's hypothesis type was superiority. This differs conceptually from a non-inferiority design: the objective is to establish evidence that the treatment groups differ, rather than to demonstrate that a new treatment is not unacceptably worse than a comparator by a prespecified margin.
8. Statistical Methods Explained
Why was a log-rank test appropriate for PFS?
PFS records when progression or death occurs, not merely whether it occurred. Some participants can remain event-free at the end of observation and therefore contribute censored follow-up. The log-rank test is designed for comparing such time-to-event distributions between groups. It makes use of event timing and risk sets rather than reducing the endpoint to a single binary status.
What does an HR of 0.47 mean?
The HR of 0.47 is a relative measure of the estimated hazard of the PFS event in the lanreotide group compared with placebo. In simple terms, the estimated hazard was 47% of the placebo-group hazard under the reported analysis. The complementary interpretation, 1 − 0.47 = 0.53, is an approximately 53% lower estimated hazard. It should not be translated directly into a 53-percentage-point improvement in PFS.
Why is the confidence interval important?
The point estimate alone suggests an HR of 0.47, but the 95% CI of 0.30–0.73 shows that the estimate is uncertain. The interval gives a statistical range around the estimated effect under the analysis framework. It is not a prediction interval for individual patients and does not mean that every patient's true treatment effect lies between 0.30 and 0.73.
Why doesn't the P-value measure treatment effect size?
The P-value of <0.001 measures the compatibility of the observed data with the null hypothesis under the specified statistical test. It does not say that the treatment has a “<0.1% chance” of working, nor does it quantify the size of the benefit. The HR and confidence interval are the appropriate reported quantities for describing the estimated relative effect and its precision.
Why was the ITT population used?
The posted primary analysis was based on 204 randomized subjects. An ITT approach keeps participants associated with their randomized assignment for the primary efficacy comparison, helping preserve the benefit of randomization. It also avoids selectively excluding patients after randomization simply because their treatment course differed from expectations.
Why does stratification matter?
The primary log-rank test was stratified by progression status at baseline and prior therapy. If these characteristics are prognostically relevant, incorporating them into the analysis can make the treatment comparison more closely reflect the randomized design. Importantly, stratification does not mean that the treatment effect must be reported separately within each stratum.
What does a time-to-event result not tell us?
An HR does not by itself provide the absolute probability of progression-free survival at a particular time, the median PFS, or the proportion of patients who benefit. Those quantities require additional information. The registry-reported CLARINET data do not report them, so they are not added to this analysis.
9. Analysis Population and What Is Actually Reported
| Item | Reported information | Statistical significance |
|---|---|---|
| Total enrollment | 264 | Not applicable |
| Primary efficacy analysis population | ITT; 204 randomized subjects | Not applicable |
| Primary endpoint | Progression-Free Survival | Formal analysis posted |
| Analysis method | Log-rank test | Stratified by baseline progression status and prior therapy |
| Effect measure | Hazard ratio | HR 0.47 |
| Confidence interval | Two-sided 95% CI | 0.30–0.73 |
| P-value | <0.001 | Reported for superiority comparison |
The difference between 264 enrolled and 204 subjects in the posted ITT analysis is important. The primary statistical result should not be described as though all 264 enrolled participants necessarily contributed to the reported analysis. The registry analysis explicitly identifies 204 randomized subjects as the analysis population.
10. Safety
The ClinicalTrials.gov record reports serious adverse events by treatment arm. These figures should be interpreted as a separate safety summary rather than as part of the PFS hypothesis test.
| Treatment group | Serious adverse events affected | At risk |
|---|---|---|
| Lanreotide (Autogel formulation) | 25 | 101 |
| Placebo | 32 | 103 |
The ClinicalTrials.gov record gives the number affected and the number at risk for serious adverse events, but they do not provide a formal statistical comparison, confidence interval, or P-value for this safety measure. Accordingly, the safety figures are presented descriptively rather than as evidence from a separate inferential test.
11. Statistical Interpretation of the Serious Adverse Event Data
There were 25 affected patients among 101 at risk in the lanreotide group and 32 among 103 at risk in the placebo group. Because the ClinicalTrials.gov record does not report an inferential comparison for this safety outcome, these counts should remain descriptive.
What can be described
The number of patients affected and the corresponding arm-specific denominators are reported directly in the ClinicalTrials.gov record.
What is not reported
No confidence interval, hypothesis test, risk ratio, risk difference, or odds ratio is posted on ClinicalTrials.gov for serious adverse events.
Why this distinction matters
Descriptive differences do not automatically establish a statistically significant treatment difference.
Different question from PFS
Safety counts and time-to-event efficacy analysis answer different questions and use different statistical structures.
12. What the Hazard Ratio Does — and Does Not — Mean
The reported PFS HR of 0.47 is below the null value of 1.00. For an endpoint in which progression or death is the event, this indicates a lower estimated event hazard in the lanreotide group than in the placebo group.
The estimate can be expressed as an approximately 53% lower estimated hazard because 1 − 0.47 = 0.53. This is a relative hazard statement, not a statement that 53% of participants avoided progression.
The two-sided 95% CI of 0.30–0.73 communicates uncertainty around the estimated hazard ratio. The interval is not a statement about the range of individual patient outcomes.
The reported P < 0.001 provides evidence against the null hypothesis under the specified superiority test. It does not measure effect size, clinical importance, or the probability that the null hypothesis is true.
Hazard-ratio interpretation should also be aware of the proportional-hazards assumption. The ClinicalTrials.gov record does not report a formal assessment of that assumption, so this page does not claim that it was demonstrated.
13. Why the Analysis Is Stratified
The primary log-rank test was stratified according to progression status at baseline and prior therapy. These variables therefore belong to the statistical structure of the primary comparison.
Stratification can be useful when the risk of the event differs across clinically relevant baseline categories. Rather than pooling every patient into a single unstructured comparison, a stratified log-rank test evaluates treatment differences within the risk-set structure defined by the strata and then combines the evidence.
This is different from a collection of subgroup tests. The primary result remains one overall treatment comparison, while the specified stratification factors are incorporated into how that comparison is calculated.
14. Randomization and Blinding
CLARINET is registered as a randomized, double-blind, parallel phase 3 trial. These design features are closely connected to the validity of the subsequent statistical analysis.
Randomization
Random allocation is intended to create comparable treatment groups in expectation, reducing systematic confounding by baseline factors.
Double masking
Double masking reduces the potential for knowledge of treatment assignment to influence behavior, assessment, or other aspects of trial conduct.
Parallel design
Participants are evaluated within their randomized treatment groups rather than sequentially receiving both randomized interventions as part of a crossover design.
Superiority framework
The statistical objective was a superiority comparison rather than a non-inferiority demonstration.
15. Missing Data, Censoring, and What the Registry Does Not Specify
Because PFS is a time-to-event endpoint, missingness and censoring are not interchangeable concepts. A patient who has not yet experienced progression or death at the end of observed follow-up can be censored, contributing information up to the relevant time. Missing tumor assessments can introduce additional complexities depending on the prespecified rules for determining progression and censoring.
The registry-reported CLARINET data provide the endpoint definition and its assessment schedule but do not report a detailed missing-data or imputation strategy. Therefore, this page does not attribute a specific imputation method, censoring convention beyond the registered endpoint definition, or sensitivity analysis to the trial.
16. Interim Analysis, Multiplicity, and Other Design Topics
The ClinicalTrials.gov record identifies the primary endpoint, superiority hypothesis, log-rank method, stratification factors, and ITT population. They do not provide details supporting an interim-analysis procedure, alpha-spending strategy, multiplicity adjustment, crossover plan, non-inferiority margin, or Bayesian analysis.
| Design topic | Information supported by the ClinicalTrials.gov record |
|---|---|
| Superiority | Yes. Hypothesis type is superiority. |
| Non-inferiority margin | Not reported in the ClinicalTrials.gov record. No non-inferiority design is identified. |
| Crossover | Not reported. |
| Factorial design | Not reported. Design model is parallel. |
| Multiplicity adjustment | Not reported. |
| Interim analysis | Not reported. |
| Missing-data imputation | Not reported. |
| Bayesian methods | Not reported. |
| Stratified analysis | Yes. Log-rank test stratified by baseline progression status and prior therapy. |
This distinction is important because absence of information in the ClinicalTrials.gov record should not be converted into an assumption that a particular method was or was not used in the complete protocol or statistical analysis plan.
17. Limitations
- Analysis population differs from enrollment: the trial enrolled 264 participants, while the posted primary efficacy analysis is explicitly based on 204 randomized subjects in the ITT population.
- Limited numerical outcome reporting: the registry-reported statistical analysis reports an HR, confidence interval, and P-value but does not provide a median PFS or time-specific PFS estimates.
- No Kaplan-Meier curve: the ClinicalTrials.gov record does not include the event and censoring information required to independently reproduce the PFS curve.
- Hazard-ratio assumptions: the reported HR is a relative time-to-event measure and should not automatically be interpreted as a constant relative risk at every time point without considering proportional-hazards assumptions.
- Safety inference is limited: serious adverse-event counts are reported by arm, but no formal inferential comparison is reported.
- Incomplete design details: the registry extract does not report interim monitoring, multiplicity procedures, missing-data methods, crossover, or Bayesian methods.
- No subgroup effects are reported: the ClinicalTrials.gov record identifies the stratification factors but does not provide subgroup-specific treatment-effect estimates or interaction tests.
18. Why This Trial Matters Statistically
CLARINET is a useful statistical teaching example because the trial's primary analysis connects several core ideas in clinical-trial methodology without requiring an unusually complicated model. A randomized, double-blind parallel design leads to an ITT efficacy population; the endpoint is time-to-event; the primary comparison uses a stratified log-rank test; and the treatment effect is summarized with a hazard ratio and two-sided confidence interval.
| Concept | How it appears in CLARINET |
|---|---|
| Randomization | The trial is registered as randomized. |
| Blinding | The trial is double masked. |
| Parallel design | The design model is parallel. |
| Intention-to-treat analysis | The primary efficacy analysis uses the ITT population of 204 randomized subjects. |
| Time-to-event endpoint | PFS is defined from randomization to progression or death. |
| Kaplan-Meier framework | PFS is a time-to-event outcome for which survival-distribution estimation is relevant. |
| Log-rank test | Posted primary analysis method. |
| Stratified analysis | Log-rank test stratified by baseline progression status and prior therapy. |
| Hazard ratio | Primary effect measure; HR 0.47. |
| Confidence interval | Two-sided 95% CI of 0.30–0.73. |
| P-value | Reported as <0.001. |
| Superiority | Registered hypothesis type. |
| Safety analysis | Serious adverse events reported descriptively by treatment arm. |
19. Primary Statistical Story
The statistical story of CLARINET can be summarized without reducing it to the P-value alone.
Randomized and double-blind
The trial used a randomized, double-blind, parallel phase 3 design with lanreotide Autogel and placebo as the two interventions.
Progression-Free Survival
PFS was defined as time from randomization to first documented progression or death, with central assessment using RECIST v1.0.
ITT analysis
The posted primary analysis used the intent-to-treat population comprising 204 randomized subjects.
Stratified log-rank test
The primary comparison used a log-rank test stratified by progression status at baseline and prior therapy.
Hazard ratio 0.47
The reported HR was 0.47, with a two-sided 95% CI of 0.30–0.73 and P < 0.001.
20. Interpreting the Evidence Without Overinterpreting It
A statistically disciplined reading of the result starts with the endpoint itself. The event was not “any change in tumor status”; it was the first documented disease progression or death. Therefore, the HR describes the relative hazard of that composite PFS event.
The next step is to separate relative effect from absolute outcome. The HR of 0.47 indicates a relative reduction in estimated event hazard, but the ClinicalTrials.gov record does not give median PFS or PFS probabilities at specified time points.
The confidence interval provides the next layer of information. The interval from 0.30 to 0.73 indicates uncertainty around the point estimate. A reader should therefore retain both pieces of information: the estimated HR is 0.47, and the statistical uncertainty around that estimate is represented by 0.30–0.73.
Finally, the P-value should be kept in its proper role. P < 0.001 indicates strong evidence against the null hypothesis under the reported superiority test. It is not a measure of effect magnitude, a probability that the treatment works, or a substitute for the confidence interval.
21. Related Statistical Concepts
Learn more about the methods used in this trial:
22. Related Statistical Calculators
23. Sources
- ClinicalTrials.gov: CLARINET — NCT00353496. Official registry record containing the trial design, registered endpoint, and posted statistical analysis.
- Linked publication: PubMed record — PMID 25014687.
- Linked publication: PubMed record — PMID 33052555.
- Linked publication: PubMed record — PMID 25408662.
Continue through the Clinical Biostats statistical library
Use the related tutorials and calculators to explore the survival-analysis concepts that underlie the CLARINET primary endpoint and statistical comparison.
24. Record Summary
CLARINET is a phase 3 randomized, double-blind, parallel trial of lanreotide Autogel versus placebo in endocrine tumors, with progression-free survival registered as the primary time-to-event endpoint. The posted primary efficacy analysis used the ITT population of 204 randomized subjects and a stratified log-rank test, with stratification according to progression status at baseline and prior therapy.
The reported treatment effect was a hazard ratio of 0.47, with a two-sided 95% confidence interval of 0.30–0.73 and P < 0.001. The HR corresponds to an approximately 53% lower estimated hazard of progression or death relative to placebo, but it does not represent an absolute percentage-point improvement or a probability that an individual patient will benefit.
The ClinicalTrials.gov record also report serious adverse events affecting 25 of 101 patients at risk in the lanreotide group and 32 of 103 in the placebo group. These safety results are descriptive because no formal inferential comparison is reported.
Statistically, the most important lesson is the connection between randomization, ITT analysis, time-to-event endpoints, stratified log-rank testing, hazard ratios, confidence intervals, and P-values. Reading the trial correctly requires keeping each of these quantities in its proper role rather than treating the P-value or hazard ratio as a complete description of the treatment effect.