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
MONOI was a completed randomized phase 3 trial in HIV infections. The study enrolled 225 patients and used a two-arm parallel design without masking. Its registered primary endpoint was the proportion of patients with virological success at W48, with virological failure defined using two consecutive plasma viral load measurements greater than or equal to 400 cp/ml within 2 weeks at W48.
| Feature | MONOI |
|---|---|
| Study acronym | MONOI |
| ClinicalTrials.gov identifier | NCT00421551 |
| Phase | Phase 3 |
| Status | Completed |
| Condition | HIV Infections |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None |
| Primary purpose | Treatment |
| Enrollment | 225 |
| Number of arms | 2 |
| Lead sponsor | French National Agency for Research on AIDS and Viral Hepatitis |
| Sponsor type | Other government |
| Start date | 2007-03 |
| Primary completion date | 2010-06 |
2. Clinical Question
The registry describes MONOI as a study comparing the efficacy and safety of darunavir boosted with ritonavir with HART containing 2 NRTI and darunavir boosted with ritonavir in HIV-1 infected patients. The registered primary efficacy question was expressed through virological success at W48.
Population
HIV-1 infected patients enrolled in a phase 3 treatment study of HIV infections.
Intervention
Darunavir boosted with ritonavir.
Comparator
HART with 2 NRTI and darunavir boosted with ritonavir.
Primary question
What proportion of patients achieved virological success at W48 under the registered definition of virological failure?
3. Trial Design
The design combines random allocation with a parallel two-arm structure. The registry identifies the study as unmasked and assigns a primary purpose of treatment. These features determine the basic framework for interpreting the primary endpoint.
Darunavir + ritonavir
- Darunavir
- Ritonavir
HART + 2 NRTI + darunavir/ritonavir
- HART
- 2 NRTI
- Darunavir boosted with ritonavir
What the parallel design contributes
In a parallel randomized trial, each patient is assigned to one treatment strategy and outcomes are subsequently compared between the randomized groups. Randomization is important because it establishes the treatment comparison before the outcome is observed, reducing the role of measured and unmeasured baseline differences in the causal comparison.
What the absence of masking means
The registry identifies the study as having no masking. For a virological endpoint based on plasma viral load measurements, the laboratory measurement itself can be relatively objective, but an unmasked design can still matter for treatment decisions, adherence, discontinuation, follow-up, and other aspects of trial conduct.
4. Endpoints
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Primary endpoint | Proportion of patients with virological success; virological failure is defined as 2 consecutive plasma viral load measurements greater or equal to 400 cp/ml within 2 weeks at W48. | W48 |
Understanding the primary endpoint
The endpoint is a proportion, rather than a time-to-event measure. The clinical outcome is therefore summarized by the fraction of patients meeting the definition of virological success at the specified assessment point.
The registry's definition of virological failure is based on repeated plasma viral load measurements. Requiring two consecutive measurements greater than or equal to 400 cp/ml within 2 weeks distinguishes the endpoint from a definition based on a single elevated measurement.
The exact analysis population and handling of patients without an evaluable W48 assessment are important statistical details for interpreting a proportion endpoint. Those details are not specified in the registry record summarized here.
5. Statistical Methodology
The registered primary endpoint is binary at W48: each patient can be classified according to whether the virological success criterion is met. The natural statistical comparison is therefore a comparison of proportions between the randomized treatment groups.
Analysis of a binary endpoint
For each randomized group, the primary outcome can be represented as the number of patients classified as successful and the number classified as unsuccessful. A treatment comparison can then be expressed using an absolute difference in proportions, a risk ratio, an odds ratio, or an appropriate hypothesis test for two independent proportions.
Here, p1 and p0 denote the proportions with virological success in the two randomized groups.
Why the endpoint is not analyzed like survival
The registered endpoint is assessed at W48 rather than defined as the time until virological failure. Consequently, a Kaplan-Meier estimator or Cox proportional-hazards model is not automatically required by the endpoint definition. Those methods are designed for time-to-event outcomes, whereas the registered primary outcome is a binary W48 status.
Confidence intervals
A confidence interval around a difference, ratio, or odds ratio quantifies statistical uncertainty around the estimated treatment contrast. The interval should be interpreted alongside the observed proportions and the analysis population because a statistical effect measure without its underlying event counts provides limited information about the absolute frequency of virological success.
Hypothesis testing
A formal hypothesis test for two independent proportions would typically evaluate a null hypothesis of no difference between the randomized groups. The p-value from such a test would describe the compatibility of the observed data with that null hypothesis under the specified statistical model; it would not quantify the size or clinical importance of the treatment effect.
6. Planned Analysis
The registry identifies the primary outcome as the proportion of patients with virological success at W48. For an endpoint of this type, the statistical analysis would normally compare the proportions of patients achieving virological success between the two randomized groups.
| Statistical component | Typical approach for this endpoint |
|---|---|
| Outcome type | Binary outcome at W48 |
| Descriptive analysis | Number and proportion of patients with virological success in each randomized group |
| Between-group comparison | Comparison of two independent proportions |
| Effect estimate | Risk difference and/or relative measure such as a risk ratio or odds ratio |
| Uncertainty | Confidence interval for the selected treatment-effect measure |
| Hypothesis test | Two-group test appropriate for a binary endpoint |
| Endpoint timing | W48 |
The definition of virological failure is particularly important because it determines which patients are classified as failures rather than successes. Two consecutive plasma viral load measurements greater than or equal to 400 cp/ml within 2 weeks at W48 constitute the registered failure definition.
The exact statistical analysis also depends on how the study protocol defines the analysis population and handles patients without the required W48 measurements. Those rules can materially affect the estimated proportion and should be specified before interpreting the treatment comparison.
7. Statistical Methods Explained
Why is a proportion appropriate for virological success?
The registered primary endpoint asks how many patients achieve virological success at a particular assessment point. Each patient therefore contributes a binary outcome: success or failure. Summarizing those outcomes as a proportion directly answers the endpoint question.
Why require two consecutive viral-load measurements?
The registry defines virological failure using two consecutive plasma viral load measurements rather than one measurement. Statistically, repeated measurements make the failure definition dependent on a sequence of observations rather than a single observation. This can distinguish the registered failure criterion from an endpoint triggered by one isolated measurement.
Why does W48 matter?
W48 establishes the assessment time for the registered primary endpoint. A proportion measured at W48 answers a different question from a proportion measured earlier or later. The timing is therefore part of the endpoint definition, not merely a reporting detail.
What is the difference between a risk difference and a risk ratio?
A risk difference describes the absolute separation between two proportions. A risk ratio describes the proportional relationship between them. For example, a risk difference can directly express how many additional successes occur per comparable number of patients, whereas a risk ratio describes the relative frequency of success between groups.
Why is the p-value not the treatment effect?
A p-value addresses a hypothesis-testing question. It does not tell the reader how large the difference in virological success is. The treatment effect should therefore be reported using the observed proportions and an interpretable effect measure, together with its confidence interval.
Why does missing W48 information matter?
If a patient does not have the information required to classify virological success at W48, the treatment group proportion depends on how that patient is handled. Different analysis-population rules can produce different estimates. For this reason, the protocol's missing-data and analysis-population definitions are an important part of a rigorous binary-endpoint analysis.
8. Randomization and Causal Interpretation
Randomization is the central design feature supporting a causal comparison between the two study strategies. Because treatment assignment occurs before the outcome is observed, systematic differences between randomized groups are less likely to explain a subsequent difference in virological success than they would be in an observational comparison.
What randomization supports
A comparison of outcomes between the randomized treatment groups under the conditions of the trial.
What randomization does not guarantee
Numerically identical baseline characteristics or identical treatment exposure and follow-up for every patient.
The study's parallel design means the treatment strategies are compared across separate groups rather than through a within-patient crossover comparison. The registry does not report a crossover design.
9. Analysis Population and Missing Data
The primary endpoint requires an assessment at W48 and applies a specific definition of virological failure based on plasma viral load measurements. Consequently, the statistical analysis needs an explicit rule for patients whose W48 status cannot be classified.
| Issue | Statistical importance |
|---|---|
| W48 assessment | Determines whether the registered primary endpoint can be evaluated. |
| Viral-load measurements | The failure definition depends on two consecutive measurements within 2 weeks. |
| Missing assessment | Can affect the denominator and classification of the binary endpoint. |
| Treatment discontinuation | May affect whether a patient has the measurements needed for endpoint classification. |
| Analysis population | Determines which randomized patients contribute to the primary comparison. |
The registry record does not report the specific missing-data or imputation strategy used for the primary endpoint. In a prespecified statistical analysis plan, this would ordinarily be defined before the outcome comparison is interpreted.
10. Safety and Efficacy Should Be Separated
The study title identifies both efficacy and safety as areas of comparison, but the registered primary endpoint is specifically a virological efficacy endpoint. These two dimensions should remain statistically distinct.
Efficacy
The registered primary efficacy endpoint is the proportion of patients with virological success at W48.
Safety
The registry record identifies safety in the study title but does not provide serious adverse-event counts by treatment arm in the information reported here.
A treatment comparison should therefore not be reduced to a single combined efficacy-and-safety statistic. Virological success and adverse events answer different questions and require their own definitions, denominators, and analysis populations.
11. Statistical Interpretation of the Primary Endpoint
The primary endpoint is the proportion of patients with virological success at W48. Its interpretation depends on the registered definition of virological failure: two consecutive plasma viral load measurements greater than or equal to 400 cp/ml within 2 weeks at W48.
If the randomized groups differed in their W48 virological-success proportions, that difference would describe the observed separation between the two treatment strategies under the study's endpoint definition. An absolute difference and a relative effect measure would communicate different aspects of the same comparison.
A confidence interval around the treatment contrast would describe statistical uncertainty associated with the estimated difference or ratio. It would not describe the range of outcomes that an individual patient could experience.
A p-value would not represent the probability that one treatment is effective or ineffective. It would instead quantify compatibility between the observed data and the specified null hypothesis under the statistical testing framework.
12. Multiplicity and Interim Analysis
The registry information summarized here identifies one registered primary endpoint: the proportion of patients with virological success at W48. It does not report a multiplicity strategy, interim efficacy analysis, alpha-spending approach, or other group-sequential procedure.
This distinction matters because repeated formal testing can change the probability of a false-positive conclusion. If interim analyses or multiple confirmatory endpoints are used, the statistical analysis plan normally specifies how the overall type I error is controlled.
13. Non-Inferiority and Superiority Logic
The registry information provided for MONOI does not report a non-inferiority margin or identify the primary hypothesis as a non-inferiority hypothesis. Therefore, the primary endpoint should not be interpreted using a non-inferiority margin that is not documented in the registry.
This is an important statistical distinction. In a superiority framework, the central question is whether the randomized groups differ. In a non-inferiority framework, the question is whether the new strategy is not worse than the comparator by more than a prespecified clinically acceptable margin. The latter requires an explicit margin and prespecified decision rule.
14. Stratification and Covariate Adjustment
The registry record identifies the trial as randomized and parallel but does not report stratification factors or a prespecified covariate-adjustment model for the primary endpoint.
For a binary endpoint, an adjusted analysis could use a regression model such as logistic regression, provided that the covariates and model were prespecified appropriately. An unadjusted analysis would compare the randomized groups directly. Neither approach should be attributed to MONOI without documentation in the statistical analysis plan or registry record.
In a logistic model, the coefficient for treatment can be transformed into an odds ratio. This is a general statistical framework for binary outcomes, not a reported MONOI analysis.
15. Why the Endpoint Definition Matters Statistically
The phrase "virological failure is defined as 2 consecutive plasma viral load measurements greater or equal to 400 cp/ml within 2 weeks at W48" contains several statistical decisions.
| Component | Statistical role |
|---|---|
| Two measurements | Failure depends on repeated measurements rather than a single measurement. |
| Consecutive | The measurements must occur in sequence according to the registered definition. |
| Greater or equal to 400 cp/ml | Defines the viral-load threshold for each qualifying measurement. |
| Within 2 weeks | Defines the temporal relationship between the two measurements. |
| W48 | Anchors the primary endpoint to the registered assessment time. |
Changing any of these components could change which patients are classified as failures. Endpoint definitions should therefore be treated as part of the statistical estimand rather than as interchangeable clinical wording.
16. A Statistical View of Virological Success
Suppose the two randomized groups have observed success proportions \(p_1\) and \(p_0\). Three common summaries illustrate why the choice of effect measure matters.
Absolute difference
The difference \(p_1-p_0\) describes the absolute separation in virological-success proportions.
Relative ratio
The ratio \(p_1/p_0\) describes how the probability of success in one group compares proportionally with the other.
Odds ratio
The odds ratio compares the odds of virological success rather than the probabilities themselves.
Confidence interval
The interval communicates the precision of the selected effect estimate and should accompany the point estimate.
These measures are not interchangeable. A relative effect can appear large even when the absolute difference is modest, while an absolute difference is often easier to interpret when the clinical question concerns how many more patients achieve success.
17. Clinical Question vs Statistical Question
| Level | Question |
|---|---|
| Clinical | Which randomized treatment strategy is associated with a different proportion of virological success at W48? |
| Endpoint | How is virological success defined at W48? |
| Descriptive | What proportion of patients in each randomized group meets the success definition? |
| Comparative | How large is the difference or ratio between the two proportions? |
| Uncertainty | How precisely is the treatment contrast estimated? |
| Hypothesis testing | How compatible are the observed data with a specified null hypothesis? |
Keeping these questions separate prevents a common statistical error: treating the p-value as though it were the magnitude of the treatment effect. A rigorous analysis reports the observed endpoint, the treatment contrast, its uncertainty, and the hypothesis-test result as distinct pieces of evidence.
18. Limitations
- No posted statistical analyses: The ClinicalTrials.gov record does not provide posted statistical analyses for the registered primary endpoint.
- No posted outcome estimates: The registry information summarized here does not report treatment-group proportions, effect estimates, confidence intervals, or p-values for virological success at W48.
- Analysis-population uncertainty: The registry record does not specify the analysis population used for the primary endpoint.
- Missing-data rules: The registry record does not report how patients without the measurements required to classify the W48 endpoint were handled.
- Multiplicity: No multiplicity-adjustment strategy is reported in the registry information summarized here.
- Interim monitoring: No interim-analysis or alpha-spending procedure is reported in the registry information summarized here.
- Stratification: No primary-analysis stratification factors are reported in the registry information summarized here.
- Safety detail: The registry information does not provide serious adverse-event counts by arm.
- Unmasked design: The registry identifies the study as having no masking, which can be relevant to treatment management and other aspects of trial conduct even when the primary endpoint is based on laboratory measurements.
19. Why This Trial Matters Statistically
MONOI provides a useful teaching example of a randomized clinical trial whose registered primary endpoint is a fixed-time binary outcome rather than a time-to-event endpoint. The statistical problem is therefore centered on proportions, treatment contrasts, endpoint classification, and the handling of missing W48 information.
| Concept | How it appears in MONOI |
|---|---|
| Randomization | The trial is randomized. |
| Parallel design | The study has 2 parallel arms. |
| Binary endpoint | Virological success is expressed as a proportion at W48. |
| Endpoint definition | Virological failure requires 2 consecutive plasma viral load measurements greater or equal to 400 cp/ml within 2 weeks at W48. |
| Effect measures | Binary treatment comparisons can be expressed through differences or ratios of proportions. |
| Confidence intervals | Uncertainty should accompany the treatment-effect estimate. |
| Hypothesis testing | A formal test can evaluate the null hypothesis for the between-group comparison. |
| Missing data | W48 classification requires explicit rules when required measurements are unavailable. |
| Unmasked treatment | The registry identifies the trial as having no masking. |
20. Sources
- ClinicalTrials.gov: MONOI (NCT00421551).
Continue through the Clinical Biostats trial library
Explore additional clinical trial analyses and statistical resources across the Clinical Biostats site.
21. Record Summary
MONOI was a completed randomized phase 3 trial enrolling 225 patients in a two-arm parallel design. The registered primary endpoint was the proportion of patients with virological success at W48, with virological failure defined as two consecutive plasma viral load measurements greater than or equal to 400 cp/ml within 2 weeks at W48. From a statistical perspective, this is a binary fixed-time endpoint, making comparison of proportions the central analytical framework. Interpretation depends not only on the treatment-group proportions but also on the effect measure, confidence interval, hypothesis-testing framework, and prespecified rules for endpoint classification and missing W48 information.