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
TNT was a randomized, double-masked, parallel phase 4 trial with a primary purpose of prevention. The study enrolled 8600 participants and evaluated atorvastatin in the setting of cardiovascular disease, cerebrovascular accident, and coronary heart disease.
| Feature | TNT |
|---|---|
| Official title | A Study to Determine the Degree of Additional Reduction in CV Risk in Lowering LDL Below Minimum Target Levels |
| Phase | Phase 4 |
| Status | Completed |
| Enrollment | 8600 |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Double |
| Primary purpose | Prevention |
| Lead sponsor | Pfizer |
| Sponsor type | Industry |
| Intervention | Atorvastatin |
| ClinicalTrials.gov | NCT00327691 |
2. Clinical Question
The registry describes TNT as a study designed to determine the degree of additional reduction in cardiovascular risk associated with lowering LDL below minimum target levels. The primary outcome is defined as the time to occurrence of a major cardiovascular event.
Population
The registry lists cardiovascular disease, cerebrovascular accident, and coronary heart disease as the study conditions. Enrollment was 8600.
Intervention
Atorvastatin was the registered intervention and was classified as a drug.
Comparator
The registry record identifies atorvastatin as the intervention, but the available registry information does not specify the comparator regimen in the registered intervention field.
Primary question
Does the study intervention affect the time to occurrence of a major cardiovascular event, as defined by the registry?
3. Trial Design
Randomization is the central design feature for causal comparison because treatment assignment is determined by the study rather than by the participant's or investigator's choice. A parallel design means participants remain assigned to their randomized study group rather than being deliberately exposed to multiple randomized treatment sequences.
Double masking is intended to reduce the potential influence of knowledge of treatment assignment on participant behavior, clinical assessment, treatment decisions, or outcome assessment. The registry record identifies the study as double-masked but does not provide the operational details of who was masked.
4. Endpoints
The registry's primary endpoint is a time-to-event outcome. Its definition is presented below using the registered wording.
| Endpoint | Registry definition | Statistical structure |
|---|---|---|
| Primary outcome | The primary outcome is the time to occurrence of a major cardiovascular event, defined as the composite outcome of the following clinical endpoints: CHD death; Non-fatal/Non-procedure related MI; Resuscitated cardiac arrest or fatal/non-fatal stroke | Time to first occurrence of a component of a composite event |
Understanding the composite endpoint
The primary outcome combines several clinically distinct events into one composite endpoint. A participant contributes a primary endpoint event when one of the specified components occurs, with the analysis concerned with the time to occurrence of a major cardiovascular event.
Composite endpoints can increase the number of observed events and therefore provide more statistical information than a single component alone. But the components can differ substantially in clinical meaning and frequency. Interpretation should therefore distinguish the composite result from the effects on its individual components.
5. Primary Endpoint Components
| Component | Registered definition | Statistical implication |
|---|---|---|
| CHD death | CHD death | Time to death from coronary heart disease |
| Non-fatal / non-procedure related MI | Non-fatal/Non-procedure related MI | Time to a specified non-fatal myocardial infarction event |
| Resuscitated cardiac arrest | Resuscitated cardiac arrest | Time to a resuscitated cardiac arrest event |
| Fatal / non-fatal stroke | fatal/non-fatal stroke | Time to a fatal or non-fatal stroke event |
The composite is therefore not a simple measure of mortality. It incorporates CHD death, a specified myocardial infarction endpoint, resuscitated cardiac arrest, and fatal or non-fatal stroke. A statistical effect on the composite would need to be interpreted in the context of the contribution made by each component.
6. Planned Analysis
The registered primary outcome is a time-to-event composite endpoint. For an endpoint of this type, the conventional analysis would compare the distributions of time to the first qualifying event between randomized groups, while accounting for participants who have not experienced the event by the end of their available follow-up.
A Kaplan-Meier analysis would typically estimate the probability of remaining free of the composite event over time. A log-rank test would commonly be used to compare the event-time distributions between randomized groups. A Cox proportional-hazards model would commonly summarize the relative treatment effect using a hazard ratio and confidence interval.
A hazard ratio below 1 would indicate a lower estimated instantaneous event rate in the treatment group during the modeled follow-up. A hazard ratio above 1 would indicate a higher estimated instantaneous event rate.
These methods are appropriate to the structure of the registered primary endpoint because the endpoint records not only whether a participant experiences an event, but also when the first qualifying event occurs. Participants who remain event-free at the end of observation are generally handled through censoring rather than being treated as if they had experienced no event at an arbitrary common time.
7. Statistical Methods Explained
Why is a time-to-event analysis appropriate?
The primary endpoint is explicitly defined as the time to occurrence of a major cardiovascular event. Two participants can both experience the endpoint but do so at different times. A binary analysis that simply classified everyone as event or no event would discard this timing information. Survival-analysis methods retain it.
What does Kaplan-Meier estimation contribute?
The Kaplan-Meier estimator describes the event-free probability over follow-up while accommodating right-censored observations. At each observed event time, the estimated event-free probability is updated using the number of participants experiencing the event and the number still at risk immediately beforehand.
Here, di is the number of events at time ti and ni is the number at risk immediately before that time.
What does a hazard ratio represent?
A hazard ratio is a relative measure of the instantaneous event rate under a specified statistical model. For example, an HR of 0.70 would be interpreted as an estimated hazard 30% lower in the treatment group relative to the comparator, under the model. It would not mean that exactly 30% fewer participants experience the endpoint, nor would it directly provide an absolute risk reduction.
Why does censoring matter?
Not every participant will necessarily experience a major cardiovascular event during the period in which their outcome is observed. A participant who remains event-free through the last available assessment contributes information through that point. The statistical analysis must distinguish this from an observed event.
Why does the composite endpoint require component-level interpretation?
The primary outcome combines CHD death, non-fatal/non-procedure related MI, resuscitated cardiac arrest, and fatal/non-fatal stroke. A composite can be statistically efficient when events are sufficiently frequent, but the components may differ in frequency and clinical importance. An overall composite result therefore does not automatically imply that every component is affected to the same degree.
Why should randomized groups be compared according to assignment?
Randomization creates the basis for a treatment comparison by assigning participants independently of investigators' treatment preferences and participants' subsequent experiences. An intention-to-treat analysis generally preserves that randomized comparison by analyzing participants according to their assigned group, regardless of later treatment changes. The available registry information does not specify the analysis population used for any posted efficacy analysis.
What can a confidence interval tell us?
For a hazard ratio, a confidence interval expresses statistical uncertainty around the estimated relative effect under the model and sampling framework. A narrow interval indicates greater statistical precision than a wide interval. The interval does not describe the range of individual patient responses.
8. Statistical Interpretation of the Primary Endpoint
The primary endpoint is not simply the proportion of participants with a cardiovascular event. Its registered definition incorporates the time to occurrence of a major cardiovascular event. This makes both event occurrence and follow-up duration part of the statistical structure.
The primary composite combines four categories of clinical outcomes: CHD death, non-fatal/non-procedure related MI, resuscitated cardiac arrest, and fatal/non-fatal stroke. An overall composite effect, if reported, would represent the combined endpoint rather than a claim that all four components have identical treatment effects.
A hazard ratio is a relative measure. Clinical interpretation is strengthened when it is considered alongside absolute event probabilities, cumulative incidence, numbers of events, and clinically relevant follow-up times. Those comparative results are not reported in the available ClinicalTrials.gov record.
A p-value, when available, addresses evidence against a specified null hypothesis under the statistical model. It does not measure the magnitude or clinical importance of an effect. Effect estimates and confidence intervals are needed to describe magnitude and precision.
9. Analysis Population and Censoring
The registry identifies the randomized allocation and enrollment but does not provide a posted statistical analysis specifying the primary efficacy analysis population, censoring rules, or handling of protocol deviations.
For a randomized time-to-event analysis, an intention-to-treat framework is commonly used because it maintains the comparison created by randomization. Censoring rules are equally important because survival methods depend on when participants leave observation without experiencing the endpoint.
| Statistical issue | Why it matters for TNT | Registry information available here |
|---|---|---|
| Analysis population | Determines which randomized participants contribute to the efficacy comparison | Not specified in the available registered statistical analysis information |
| Censoring | Determines how event-free participants with incomplete follow-up contribute information | Not specified |
| Missing data | Can affect outcome ascertainment and covariate-based analyses | Not specified |
| Protocol deviations | Can affect the distinction between randomized and analyzed populations | Not specified |
10. Non-Inferiority, Multiplicity, and Interim Analysis
The registered information available for TNT does not specify a non-inferiority margin, a factorial design, an interim-analysis boundary, a multiplicity-adjustment strategy, or a Bayesian analysis.
This distinction matters because these design features change how statistical evidence should be interpreted. A non-inferiority trial, for example, requires a prespecified margin and a confidence-interval-based comparison against that margin. A superiority trial with multiple primary endpoints or repeated interim looks requires a prespecified approach to type I error control.
Non-inferiority margin
No non-inferiority margin is identified in the registered information available here.
Interim analysis
No interim-analysis procedure or efficacy boundary is identified in the registered information available here.
Multiplicity
No multiplicity-adjustment strategy is identified in the registered information available here.
Bayesian methods
No Bayesian analysis is identified in the registered information available here.
11. Stratification and Covariate Adjustment
The registry information available here identifies randomization but does not provide the stratification factors, prespecified covariates, or adjustment variables used for any statistical analysis.
This is an important distinction between the design description and the analysis model. Randomization establishes the treatment comparison, while stratification and covariate adjustment can be used to improve efficiency, account for prespecified prognostic factors, or align the analysis with the randomization scheme.
12. Safety and Adverse Events
The available registry information identifies the trial intervention and enrollment but does not provide posted serious adverse-event counts by randomized treatment arm.
Safety is statistically distinct from the primary efficacy endpoint. The primary outcome is a time-to-event cardiovascular composite, whereas adverse-event summaries generally describe treatment-emergent events by category, severity, seriousness, or relationship to study treatment. Those summaries require the underlying safety dataset or posted registry results.
| Safety item | Available registry information |
|---|---|
| Serious adverse events by arm | Not reported in the available registered statistical analysis information |
| Adverse-event counts by arm | Not reported |
| Adverse-event rates | Not reported |
| Safety analysis population | Not specified |
13. What a Complete Time-to-Event Analysis Would Report
For a randomized study with this primary endpoint, a statistically informative results presentation would normally include both the relative treatment effect and absolute event-time information.
| Result | Why it matters |
|---|---|
| Number of primary endpoint events | Describes the information actually contributing to the event-time comparison |
| Kaplan-Meier event-free curves | Shows how event-free probability changes over follow-up |
| Hazard ratio | Summarizes the relative event rate under a Cox model |
| 95% confidence interval | Describes uncertainty around the estimated hazard ratio |
| Log-rank p-value | Tests the specified null hypothesis for the event-time comparison |
| Absolute event probabilities | Places the relative treatment effect on an absolute scale |
| Component-specific outcomes | Shows how the individual elements contributed to the composite endpoint |
The registry record does not post these comparative statistical results for the primary endpoint. Consequently, the available record supports a detailed description of the endpoint and its appropriate statistical framework, but not a numerical treatment-effect estimate.
14. Interpreting Composite Cardiovascular Endpoints
Composite endpoints are common in cardiovascular trials because several clinically meaningful events may occur during follow-up. Combining them can increase the total number of observed endpoint events and potentially improve statistical efficiency.
Statistical advantage
Combining multiple qualifying events can increase the number of events available for a time-to-event comparison, potentially improving precision for the composite endpoint.
Clinical challenge
The components can differ in severity, frequency, and treatment responsiveness. A composite estimate therefore requires component-level context.
First-event structure
When the analysis is defined as time to the first qualifying event, later events generally do not replace the initial endpoint event in that primary time-to-event analysis.
Interpretation
The composite endpoint should be interpreted as its registered definition rather than as a synonym for mortality or myocardial infarction alone.
15. Hazard Ratios: A Practical Interpretation
For a randomized comparison, a hazard ratio is a relative treatment-effect measure. It does not directly state the percentage of participants who avoid the event, the absolute number of events prevented, or the duration of benefit for an individual participant.
What an HR does mean
If an analysis produced an HR of 0.75, the estimated instantaneous event rate in the treatment group would be 25% lower than in the comparator group under the fitted model.
What an HR does not mean
An HR of 0.75 would not mean that 25% of participants were prevented from having the event. It would also not mean that every participant experienced the same proportional reduction. Absolute event probabilities are needed to communicate those quantities.
Why the proportional-hazards assumption matters
A conventional Cox hazard ratio is often interpreted as a common relative hazard over follow-up. If the hazards cross or the relative effect changes substantially over time, a single hazard ratio can conceal important temporal structure. A complete analysis should therefore consider the survival curves and clinically relevant absolute event probabilities in addition to the HR.
16. Limitations
- No posted statistical analyses: the ClinicalTrials.gov record does not provide formal comparative statistical analyses for the registered primary endpoint.
- No treatment-arm sample sizes: total enrollment is 8600, but the registered information available here does not specify the number randomized to each treatment group.
- No comparative estimates: the available registry information does not provide a hazard ratio, confidence interval, p-value, median event time, or absolute event probability for the primary endpoint.
- No event counts: the available record does not provide the number of participants experiencing the composite endpoint or its individual components by treatment group.
- Composite endpoint: the primary outcome combines events with potentially different clinical significance and frequency.
- Analysis specification: the available registered information does not specify the exact efficacy analysis population, censoring rules, covariate adjustment, or stratification factors.
- Safety information: serious adverse events and treatment-specific safety summaries are not reported in the available registered statistical analysis information.
- Time frame: the primary endpoint definition identifies time to occurrence but does not provide a numerical follow-up duration in the available registry information.
17. Why This Trial Matters Statistically
TNT is a useful statistical teaching case because its primary endpoint combines a clinically defined cardiovascular composite with a time-to-event framework. The design also illustrates why the statistical analysis of a randomized trial cannot be reduced to a single p-value.
| Concept | How it appears in TNT |
|---|---|
| Randomization | The study uses randomized allocation across a parallel design. |
| Double masking | The registry identifies the study as double-masked. |
| Time-to-event analysis | The primary outcome is the time to occurrence of a major cardiovascular event. |
| Composite endpoint | The primary outcome combines CHD death, non-fatal/non-procedure related MI, resuscitated cardiac arrest, and fatal/non-fatal stroke. |
| Kaplan-Meier estimation | A standard method for describing event-free survival over time for this endpoint structure. |
| Hazard ratio | A conventional relative measure for comparing event rates between randomized groups. |
| Censoring | Required to appropriately analyze participants who remain event-free through the end of available follow-up. |
| Component analysis | Important for understanding which events contribute to a composite treatment effect. |
18. Statistical Reasoning: From Trial Design to Estimand
The most important statistical step is to identify exactly what is being compared. TNT's registered primary outcome is not simply "cardiovascular risk." It is the time to occurrence of a major cardiovascular event, where a major cardiovascular event is defined through a specific composite of clinical endpoints.
That definition determines the natural estimand. The comparison concerns the distribution of time until the first qualifying event between randomized groups. A treatment effect can then be summarized using an absolute measure such as an event-free probability at a specified time or a relative measure such as a hazard ratio.
The estimand also clarifies why follow-up and censoring are essential. A participant who has not experienced the endpoint by the end of observation has not necessarily demonstrated indefinite freedom from the event. Their contribution is limited to the period during which their outcome status is observed.
19. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The registered primary outcome is a time-to-event composite. The natural analytical framework is therefore survival analysis, with treatment groups compared according to their time to the first qualifying cardiovascular event.
Clinical interpretation
The composite outcome includes CHD death, non-fatal/non-procedure related MI, resuscitated cardiac arrest, and fatal/non-fatal stroke. Its clinical interpretation should therefore consider both the overall composite and the individual components.
The distinction is important because statistical significance and clinical importance are not interchangeable. A statistically precise treatment effect can still require clinical context, while a clinically meaningful difference can be estimated imprecisely when relatively few events occur.
20. What the Registry Record Establishes
| Question | Registered information |
|---|---|
| What is the study? | TNT, NCT00327691 |
| What phase is it? | Phase 4 |
| How many participants were enrolled? | 8600 |
| How were participants allocated? | Randomized |
| What is the design? | Parallel |
| Was the study masked? | Double |
| What was the intervention? | Atorvastatin |
| What was the primary purpose? | Prevention |
| What was the primary outcome? | Time to occurrence of a major cardiovascular event |
| What constitutes the primary event? | CHD death; non-fatal/non-procedure related MI; resuscitated cardiac arrest; or fatal/non-fatal stroke |
| Are comparative statistical results posted? | No formal statistical analyses are posted to ClinicalTrials.gov for the registered primary endpoint. |
21. Sources
- ClinicalTrials.gov: TNT — NCT00327691.
- PubMed: PubMed record associated with TNT.
Continue through Clinical Biostats
Explore additional clinical trial analyses, statistical tutorials, and quantitative tools for understanding randomized clinical-trial evidence.
22. Record Summary
TNT was a randomized, double-masked, parallel phase 4 study enrolling 8600 participants, with atorvastatin identified as the intervention and prevention as the primary purpose. Its registered primary outcome was the time to occurrence of a major cardiovascular event, defined as a composite of CHD death, non-fatal/non-procedure related MI, resuscitated cardiac arrest, and fatal/non-fatal stroke.
Statistically, the endpoint calls for a time-to-event framework rather than a simple end-of-study proportion. Kaplan-Meier estimation can describe event-free survival, a log-rank comparison can assess differences between randomized groups, and a Cox model can provide a relative hazard estimate with a confidence interval. Interpretation of any eventual treatment effect would also require attention to censoring, the analysis population, proportional-hazards assumptions, and the contribution of individual components to the composite endpoint.