← Clinical Trials
Cardiology Phase 3 Completed NCT02104817

STRENGTH: Complete Statistical Analysis of Epanova in High Cardiovascular Risk Patients With Hypertriglyceridemia

An independent statistical review of the randomized phase 3 STRENGTH trial evaluating Epanova (omega-3 carboxylic acids) versus a corn oil control in eligible men and women considered high risk for atherosclerotic cardiovascular disease.

Trial start: 2014-10-30  ·  Primary completion: 2020-05-27  ·  Enrollment: 13078
Scope of this record

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

STRENGTH was a randomized, parallel-group, triple-masked phase 3 treatment trial enrolling 13078 participants. The registered primary endpoint was the composite of major adverse cardiovascular events (MACE), analyzed as a time-to-event outcome using a Cox proportional-hazards model in the Full Analysis Set.

13078
Enrollment
2 treatment arms
3
Masking
Triple
0.99
Primary MACE HR
95% CI 0.90–1.09
0.837
Primary P-value
Superiority analysis
FeatureSTRENGTH
Trial nameSTRENGTH
Brief titleOutcomes Study to Assess STatin Residual Risk Reduction With EpaNova in HiGh CV Risk PatienTs With Hypertriglyceridemia
PhasePhase 3
StatusCompleted
PopulationEligible men or women considered high risk for atherosclerotic cardiovascular disease (CVD)
AllocationRandomized
Design modelParallel
MaskingTriple
Primary purposeTreatment
Enrollment13078
Arms2
Lead sponsorAstraZeneca
Sponsor typeIndustry
ClinicalTrials.govNCT02104817

2. Clinical Question

The statistical question was whether Epanova, compared with the corn oil control, would alter the time to the composite of major adverse cardiovascular events among eligible men and women considered high risk for atherosclerotic cardiovascular disease.

Population

Eligible men or women considered high risk for atherosclerotic cardiovascular disease (CVD). The trial's brief title identifies the study population as high cardiovascular risk patients with hypertriglyceridemia.

Intervention

Epanova® (omega-3 carboxylic acids).

Comparator

Corn oil control. The posted statistical analyses refer to the comparator as Placebo.

Primary question

Does Epanova reduce the hazard of the registered composite MACE endpoint relative to the control under a superiority hypothesis?

3. Trial Design

01
Randomize13078 participants
02
Two armsEpanova vs control
03
FollowFrom randomization
04
AssessMACE and other outcomes
05
AnalyzeCox proportional-hazards model
ARM A

Epanova

  • Epanova® (omega-3 carboxylic acids)
  • Randomized treatment assignment
  • Included in the Full Analysis Set efficacy analyses
ARM B

Corn oil control

  • Corn oil control
  • Randomized control assignment
  • Referred to as Placebo in the posted statistical analyses
Allocation
Randomized allocation to two parallel treatment groups.
Masking
Triple masking was reported in the registry profile.
Primary purpose
Treatment.
Hypothesis
Superiority for the primary MACE comparison.

4. Trial Dates and Registry Results

Registry featureReported value
Start2014-10-30
Primary completion2020-05-27
StatusCOMPLETED
Results postedYes
Outcome measures posted14
Statistical analyses posted14
Primary endpoint analyses1
Primary analyses with estimate + confidence interval1

The ClinicalTrials.gov record contains statistical analyses for all 14 posted outcome measures. The primary endpoint has one formal analysis with an effect estimate and confidence interval; the remaining posted analyses are secondary or other prespecified analyses.

5. Primary Endpoint

EndpointRegistered definition / time frameAnalysis
The Composite of Major Adverse Cardiovascular Events (MACE) From the date of randomization and up to completion of the end-of-treatment visit (Month 60) or at study closure.

MACE components include: cardiovascular (CV) death, nonfatal myocardial infarction (MI), nonfatal stroke, emergent/elective coronary revascularization, or hospitalization for unstable angina. Participants with no observed events are censored at the earliest of withdrawal of consent date and last study contact (defined as the latest of the dates of assessments contributing to an opportunity to assess as to whether the participant has had every component of the endpoint being analyzed).
Cox proportional-hazards model; Full Analysis Set; Epanova vs Placebo; superiority

6. Primary Result: MACE

The primary analysis compared Epanova with the control for time to the composite MACE endpoint from randomization through completion of the end-of-treatment visit at Month 60 or study closure. The analysis population was the Full Analysis Set.

Hazard ratio for MACE

0.99

95% CI: 0.90–1.09   ·   P = 0.837

Cox proportional-hazards model  ·  Superiority hypothesis

Primary endpointEpanova vs Placebo
EndpointThe Composite of Major Adverse Cardiovascular Events (MACE)
Analysis populationFull Analysis Set
MethodCox proportional-hazards model
Effect measureHazard ratio
Estimate0.99
95% CI0.90–1.09
P-value0.837
Hypothesis typeSuperiority
Clinical Biostats interpretation

An HR of 0.99 means that the fitted Cox model estimated an instantaneous MACE hazard approximately 0.99 times that of the comparator group over the analyzed follow-up. Expressed descriptively, the estimate is very close to 1.

The HR does not mean that 99% of participants experienced an event, nor does it represent an absolute risk difference, a probability of benefit for an individual participant, or a statement that every participant had the same relative hazard.

The 95% confidence interval of 0.90–1.09 describes statistical uncertainty around the estimated hazard ratio under the model and sampling framework. It includes 1, so the interval is compatible with both a modestly lower and a modestly higher estimated hazard relative to the comparator.

The P-value of 0.837 addresses the evidence against the null hypothesis used for the superiority comparison; it is not a measure of effect size. A P-value does not tell us that the treatment effect is “83.7%” or quantify clinical importance.

Because this is a Cox-model result, interpretation also depends on the model's assumptions, including the proportional-hazards assumption. Censoring is part of the time-to-event framework, and the registry defines censoring for participants without observed events in relation to withdrawal of consent and last study contact.

7. Secondary Endpoint Results

The registry also reports nine secondary statistical analyses. All use the Full Analysis Set and compare Epanova with Placebo using Cox proportional-hazards regression. The analyses are superiority analyses and use two-sided 95% confidence intervals.

Secondary endpointHR95% CIP-value
The Composite of MACE in the Subgroup of Participants With Established CV Disease(CVD) at Baseline 0.940.84–1.050.269
The Composite of CV Events 1.050.93–1.190.402
The Composite of CV Events in the Subgroup of Participants With Established CVD at Baseline 1.010.87–1.160.940
The Composite of Coronary Events 0.910.81–1.020.092
The Composite of Coronary Events in the Subgroup of Participants With Established CVD at Baseline 0.850.75–0.970.016
CV Death 1.090.90–1.310.372
CV Death in the Subgroup of Participants With Established CVD at Baseline 1.120.89–1.410.336
All-cause Death 1.130.97–1.310.112
All-cause Death in the Subgroup of Participants With Established CVD at Baseline 1.180.97–1.420.091

The secondary estimates span both sides of the null value of 1. Several confidence intervals include 1, while the subgroup analysis of the composite of coronary events has an estimated HR of 0.85 with a 95% CI of 0.75–0.97 and P = 0.016.

Interpretation caution: the presence of a P-value below 0.05 for one secondary endpoint does not, by itself, establish that this endpoint has the same confirmatory status as the registered primary endpoint. The ClinicalTrials.gov record identifies the primary MACE analysis separately from the secondary analyses but do not provide a multiplicity-adjustment scheme for these secondary comparisons. Accordingly, the secondary results should be interpreted as the registry-reported analyses rather than reordered into a new hierarchy.

8. Other Prespecified Endpoint Results

Four additional analyses are identified as Other_Pre_Specified in the ClinicalTrials.gov record. Each was analyzed as a time-to-event outcome using Cox proportional-hazards regression in the Full Analysis Set.

Other prespecified endpointHR95% CIP-value
Emergent/Elective Coronary Revascularization 0.940.83–1.080.408
Hospitalization for Unstable Angina 0.840.63–1.120.233
Non-fatal Myocardial Infarction 0.970.81–1.170.770
Non-fatal Stroke 1.140.90–1.450.278

These component-level results help show how the composite endpoint can be decomposed into clinically distinct event types. A composite can be influenced by the relative frequency and treatment effect of each component, so an overall composite HR should not automatically be interpreted as though every component had the same estimated effect.

9. Statistical Methodology

Cox proportional-hazards model

The registry identifies the Cox proportional-hazards model as the statistical method for the primary and posted secondary time-to-event analyses. The primary endpoint was analyzed from randomization through the end-of-treatment visit at Month 60 or study closure.

Conceptual model
h(t | X) = h0(t) exp(βX)

The model relates the instantaneous event hazard to treatment and other model terms. For a binary treatment indicator, the exponentiated treatment coefficient, exp(β), is interpreted as a hazard ratio.

The key feature is that Cox regression analyzes when an event occurs rather than simply classifying participants as having or not having an event. Participants who remain event-free contribute follow-up information until their defined censoring time.

Hazard ratio

The hazard ratio is the effect measure reported for every statistical analysis reported in the registry extract. An HR below 1 represents a lower estimated instantaneous event hazard in the Epanova group relative to Placebo; an HR above 1 represents a higher estimated instantaneous event hazard.

Interpretation of the hazard ratio
HR = exp(β)

An HR of 1 corresponds to equal modeled hazards. The HR is a relative time-to-event measure and is not equivalent to an absolute risk difference or a risk ratio.

Full Analysis Set

The primary and secondary analyses in the ClinicalTrials.gov record uses the Full Analysis Set. The analysis text also identifies the intention-to-treat concept. This is important because randomized treatment assignment remains the basis for the efficacy comparison rather than redefining groups according to treatment exposure after randomization.

Time-to-event censoring

The primary endpoint is assessed from the date of randomization through Month 60 or study closure. The registry definition states that participants with no observed events are censored at the earliest of withdrawal of consent and last study contact, with the last study contact defined using the dates of assessments contributing to the registry's stated opportunity-to-assess framework.

This structure allows participants with different lengths of observed follow-up to contribute information without requiring every participant to have the same observation time.

Two-sided confidence intervals

The statistical analyses report two-sided 95% confidence intervals. For the primary endpoint, the interval from 0.90 to 1.09 surrounds the estimated HR of 0.99. The interval therefore communicates more information than the point estimate alone: it shows the range of effect estimates supported by the specified statistical framework at the stated confidence level.

Superiority testing

The primary and posted secondary analyses are labeled as superiority hypotheses. The central inferential question is therefore whether the randomized treatment comparison provides evidence that the event-time distribution differs in the hypothesized direction, rather than whether Epanova satisfies a non-inferiority margin.

10. Statistical Methods Explained

Why was a Cox proportional-hazards model used?

The MACE endpoint is a time-to-event outcome: the analysis follows participants from randomization and records when the first relevant event occurs, while accounting for participants who do not have an observed event during their available follow-up. Cox regression is designed for this structure and produces the hazard ratio reported by the registry.

What does an HR of 0.99 mean?

An HR of 0.99 means that the estimated instantaneous MACE hazard under the fitted Cox model was 0.99 times the hazard in the comparator group. It does not mean that the probability of MACE was exactly 0.99 times the probability for every participant, because a hazard ratio is a model-based time-to-event measure.

Why is the confidence interval important?

The point estimate is only one summary of the treatment comparison. The 95% CI of 0.90–1.09 shows the statistical uncertainty around the primary HR estimate. Because the interval crosses 1, the ClinicalTrials.gov record is compatible with a range of relative hazards around the null value.

Why does the P-value not measure effect size?

The P-value describes the compatibility of the observed data with the null hypothesis under the specified statistical test. It does not quantify the magnitude of the treatment effect. The HR provides the relative effect estimate, while the confidence interval communicates its precision.

Why analyze a composite endpoint?

MACE combines several clinically important cardiovascular outcomes into one time-to-event endpoint: CV death, nonfatal MI, nonfatal stroke, emergent/elective coronary revascularization, and hospitalization for unstable angina. A composite can increase the number of observed endpoint events, but its interpretation depends on the contributions of its individual components. The ClinicalTrials.gov record therefore also provide separate analyses for several components.

Why should subgroup estimates be interpreted cautiously?

The registry reports several analyses restricted to participants with established CVD at baseline. These are subgroup analyses rather than separate randomized trials. Even when a subgroup estimate has a particular direction or P-value, a formal conclusion that treatment effects differ between subgroups requires an appropriate comparison of effects between subgroups, such as an interaction analysis. The ClinicalTrials.gov record does not report such an interaction analysis.

Why does censoring matter?

Not every participant necessarily has an observed event during the analysis period. Censoring allows the Cox model to use the participant's observed follow-up while recognizing that the event time is not observed beyond the censoring point. The validity of the resulting inference depends on the assumptions underlying the censoring and time-to-event analysis.

11. Reading the Primary Result Correctly

Point estimate

The primary MACE HR was 0.99. Relative to the null value of 1, this point estimate is close to no difference in the modeled instantaneous event hazard between the randomized groups.

Confidence interval

The 95% CI of 0.90–1.09 indicates that the point estimate should not be interpreted in isolation. The interval includes values below and above 1, representing uncertainty in the direction and magnitude of the relative hazard.

P-value

The primary P = 0.837 is the reported result for the superiority analysis. It should not be converted into a probability that the treatment works or does not work, and it should not be used as a substitute for the HR and confidence interval.

Absolute effects

The ClinicalTrials.gov record does not provide event counts for the primary MACE endpoint or time-specific event probabilities by randomized group. Consequently, the primary result should be presented using the reported hazard ratio and confidence interval rather than constructing an absolute risk difference from unavailable information.

12. Secondary Results in Statistical Context

The secondary analyses illustrate why a clinical-trial results page should not reduce every endpoint to a binary “significant” or “not significant” label.

Effect direction

An HR below 1 indicates a lower estimated hazard for Epanova; an HR above 1 indicates a higher estimated hazard. This directional description is separate from whether the corresponding confidence interval excludes 1.

Precision

The width of the confidence interval provides information about precision. For example, the coronary-events subgroup estimate has a 95% CI of 0.75–0.97, while the CV-death subgroup estimate has a 95% CI of 0.89–1.41.

Composite vs component

The composite MACE endpoint should not be assumed to have the same treatment effect as CV death, MI, stroke, revascularization, or unstable-angina hospitalization individually.

Subgroup analysis

Analyses restricted to participants with established CVD at baseline address a more specific population and should not automatically be treated as interchangeable with the overall Full Analysis Set analysis.

13. Serious Adverse Events

The ClinicalTrials.gov record reports serious adverse events by randomized arm using affected participants over the corresponding participants at risk.

Safety measureEpanovaPlacebo / control
Serious adverse events 2259 / 6532 affected / at risk 2195 / 6535 affected / at risk

These are the serious-adverse-event counts reported in the ClinicalTrials.gov record. They should be kept distinct from efficacy-event analyses: a serious adverse event is a safety classification, whereas MACE and its components are the registered cardiovascular efficacy outcomes analyzed as time-to-event endpoints.

Safety interpretation: the ClinicalTrials.gov record does not provide a formal statistical comparison, confidence interval, or P-value for the serious adverse-event counts. Accordingly, this page reports the affected/at-risk values without constructing an unreported inferential comparison.

14. What the Secondary Hazard Ratios Do — and Do Not — Mean

The secondary analyses demonstrate the importance of looking at the entire statistical estimate rather than only its P-value.

EndpointReported HRStatistical reading
The Composite of CV Events 1.05 The point estimate is above 1; the 95% CI of 0.93–1.19 includes 1.
The Composite of Coronary Events 0.91 The point estimate is below 1; the 95% CI of 0.81–1.02 includes 1.
CV Death 1.09 The point estimate is above 1; the 95% CI of 0.90–1.31 includes 1.
All-cause Death 1.13 The point estimate is above 1; the 95% CI of 0.97–1.31 includes 1.
Non-fatal Myocardial Infarction 0.97 The point estimate is below 1; the 95% CI of 0.81–1.17 includes 1.
Non-fatal Stroke 1.14 The point estimate is above 1; the 95% CI of 0.90–1.45 includes 1.

These examples show why the HR, confidence interval, endpoint definition, analysis population, and inferential role should be read together. A point estimate below 1 does not automatically establish a treatment effect, and a point estimate above 1 does not by itself establish harm.

15. Multiplicity and the Endpoint Hierarchy

The ClinicalTrials.gov record identifies one registered primary endpoint and a collection of secondary and other prespecified outcomes. The primary endpoint is the composite MACE outcome. The additional outcomes provide information about related cardiovascular event categories and subgroup populations.

Endpoint roleExamples in the ClinicalTrials.gov recordStatistical interpretation
Primary The Composite of Major Adverse Cardiovascular Events (MACE) Primary superiority analysis; Cox proportional-hazards model; HR 0.99; 95% CI 0.90–1.09; P = 0.837.
Secondary MACE subgroup, CV events, coronary events, CV death, all-cause death and established-CVD subgroup analyses Additional prespecified analyses; all registry-reported analyses use Cox regression and a superiority hypothesis.
Other prespecified Coronary revascularization, unstable-angina hospitalization, nonfatal MI, nonfatal stroke Component-level or related outcomes; reported as Cox-model hazard ratios.

The registry extract does not specify a multiplicity-adjustment procedure for the 14 posted statistical analyses. Therefore, the P-values should not be treated as though all 14 analyses formed a single unadjusted set of independent confirmatory tests, nor should a single secondary result be promoted above the registered primary endpoint solely because its P-value is smaller.

16. Intention-to-Treat Analysis and Randomization

The statistical-analysis text explicitly identifies the intention-to-treat concept, while the analysis population is recorded as the Full Analysis Set. The randomized design is important because randomization establishes the treatment comparison before subsequent events, treatment exposure, and follow-up occur.

Why randomization matters

Randomization creates the foundation for comparing treatment groups without assigning treatment based on post-randomization outcomes.

Why the analysis population matters

The registry identifies the Full Analysis Set for the reported efficacy analyses. This defines which participants contribute to the formal treatment comparison.

Why time zero matters

The primary endpoint starts at the date of randomization. This provides a common time origin for the time-to-event comparison.

Why post-randomization events matter

Follow-up, censoring, and observed cardiovascular events occur after randomization and are handled within the time-to-event analysis rather than redefining the initial treatment assignment.

17. Blinding and Bias Control

The registry identifies the study as triple masked. Masking is a design feature intended to reduce the influence of knowledge of treatment assignment on aspects of trial conduct and assessment.

For a time-to-event endpoint such as MACE, masking can be relevant to how events are identified, reported, assessed, and managed. The ClinicalTrials.gov record does not provide further detail about which specific trial personnel or participants were included in the triple-masking designation, so this page does not infer a more detailed masking structure.

18. Composite Endpoint Analysis

The primary MACE endpoint combines five types of cardiovascular outcomes: CV death, nonfatal MI, nonfatal stroke, emergent/elective coronary revascularization, and hospitalization for unstable angina.

MACE componentSeparate analysis reported?Reported HR
CV death Yes 1.09
Nonfatal myocardial infarction Yes 0.97
Nonfatal stroke Yes 1.14
Emergent/Elective Coronary Revascularization Yes 0.94
Hospitalization for Unstable Angina Yes 0.84

The availability of component analyses is statistically useful because a composite endpoint is a single outcome assembled from multiple event types. A composite result can reflect the combined occurrence of all specified components, while individual components can have different event frequencies, clinical importance, and estimated treatment effects.

Composite-endpoint caution: an HR for MACE should not be described as though it were an HR for CV death alone, MI alone, stroke alone, or any other individual component. The registry separately reports these component-level analyses, which should be interpreted on their own definitions.

19. Subgroup Analysis: Established CVD at Baseline

Several analyses were restricted to the subgroup of participants with established CVD at baseline. These include MACE, the composite of CV events, the composite of coronary events, CV death, and all-cause death.

Established CVD subgroup endpointHR95% CIP-value
MACE 0.94 0.84–1.05 0.269
Composite of CV Events 1.01 0.87–1.16 0.940
Composite of Coronary Events 0.85 0.75–0.97 0.016
CV Death 1.12 0.89–1.41 0.336
All-cause Death 1.18 0.97–1.42 0.091

These estimates answer narrower questions than the overall primary analysis. A subgroup HR describes the treatment comparison within the specified subgroup; it does not establish that the treatment effect differs from the treatment effect in participants without established CVD.

To establish effect modification, the relevant statistical question is whether the treatment-by-subgroup interaction is sufficiently supported. The ClinicalTrials.gov record does not report an interaction estimate or interaction P-value, so no formal heterogeneity conclusion is made here.

20. Censoring and Follow-up

The primary endpoint is evaluated from randomization through completion of the end-of-treatment visit at Month 60 or study closure. The registry definition specifies censoring for participants without observed events at the earliest of withdrawal of consent and last study contact.

Time-to-event structure
Randomization → observed MACE or defined censoring → Cox analysis

This framework means that participants can contribute different amounts of observed follow-up. The analysis uses both event occurrence and event timing rather than treating all participants as though they had identical observation periods.

Censoring is therefore not equivalent to an event-free outcome. A participant who is censored has known follow-up up to the censoring point, but the subsequent event time is not observed within the analysis framework.

21. What the Registry Does Not Report in the Supplied Data

The ClinicalTrials.gov record is sufficient for a full statistical interpretation of the posted hazard-ratio analyses, but they do not contain several quantities that would support additional types of presentation.

Potential resultStatus in the ClinicalTrials.gov record
Primary MACE event counts by armNot provided
Median time to MACENot provided
Kaplan-Meier survival probabilitiesNot provided
Baseline characteristic tableNot provided
Formal interaction testsNot provided
Multiplicity-adjustment procedure for the posted analysesNot provided
Bayesian analysisNot reported in the registry-reported statistical methods
Non-inferiority marginNot reported; the analyses are identified as superiority
Imputation methodNot reported in the ClinicalTrials.gov record
Interim-analysis methodNot reported in the ClinicalTrials.gov record

These omissions are important because they limit what can responsibly be reconstructed from the registry extract. In particular, a hazard ratio and confidence interval are not enough information to recreate a Kaplan-Meier curve or derive arm-specific event probabilities.

22. Limitations

23. Why This Trial Matters Statistically

STRENGTH is a useful teaching case because it combines a large randomized phase 3 design with a composite cardiovascular time-to-event endpoint and a broad set of related secondary and component analyses. The registry results illustrate how the same Cox-model framework can be applied to an overall endpoint, clinically defined subgroups, individual cardiovascular outcomes, and component events.

Statistical conceptHow it appears in STRENGTH
RandomizationRandomized allocation in a two-arm parallel phase 3 trial.
BlindingTriple masking.
Intention-to-treat conceptIdentified in the analysis text; efficacy analyses use the Full Analysis Set.
Time-to-event endpointPrimary MACE outcome assessed from randomization through Month 60 or study closure.
Kaplan-Meier conceptRelevant to time-to-event estimation, although Kaplan-Meier estimates are not reported in the ClinicalTrials.gov record.
Hazard ratioReported as the effect measure for the primary and posted secondary analyses.
Cox regressionRegistry-reported method for the posted statistical analyses.
Confidence intervalTwo-sided 95% intervals accompany the posted hazard-ratio estimates.
Superiority testingPrimary and posted secondary analyses are labeled as superiority hypotheses.
Composite endpointMACE combines CV death, nonfatal MI, nonfatal stroke, coronary revascularization, and unstable-angina hospitalization.
Subgroup analysisSeveral outcomes are analyzed in participants with established CVD at baseline.
MultiplicityMultiple primary, secondary, and other prespecified analyses require attention to endpoint hierarchy and interpretation of P-values.

24. Overall Statistical Reading

The primary STRENGTH analysis produced a MACE hazard ratio of 0.99, with a two-sided 95% confidence interval of 0.90–1.09 and P = 0.837. The point estimate is close to the null value of 1, and the confidence interval spans both sides of that value.

The secondary analyses show a range of estimates rather than a uniform directional pattern. For example, the composite of coronary events had an HR of 0.91 in the overall analysis, while CV death had an HR of 1.09, all-cause death had an HR of 1.13, and nonfatal stroke had an HR of 1.14. The established-CVD subgroup analysis for coronary events reported an HR of 0.85 with a 95% CI of 0.75–0.97 and P = 0.016.

Those individual estimates should be read in the context of their endpoint role, analysis population, confidence interval, and multiplicity. The statistically appropriate summary is therefore not a single label applied to the entire trial, but a structured reading of the primary endpoint followed by the prespecified secondary and component analyses.

25. Related Tutorials

Learn more about the methods used in this trial:

26. Related Statistical Calculators

27. Sources

Continue through the Clinical Biostats statistical pathway

Connect the STRENGTH trial's randomized design, time-to-event endpoint, hazard ratio, confidence interval, and Cox model to deeper statistical tutorials and analysis tools.

28. Record Summary

STRENGTH provides a detailed example of randomized cardiovascular outcomes analysis centered on a composite time-to-event endpoint. The primary MACE analysis used a Cox proportional-hazards model in the Full Analysis Set and reported an HR of 0.99 with a two-sided 95% CI of 0.90–1.09 and P = 0.837. The registry also reports Cox-model analyses for secondary composite outcomes, established-CVD subgroups, CV death, all-cause death, coronary revascularization, unstable-angina hospitalization, nonfatal MI, and nonfatal stroke.

The most informative statistical reading therefore combines the hazard ratio, confidence interval, P-value, endpoint definition, analysis population, time-to-event structure, and endpoint hierarchy. The ClinicalTrials.gov record does not provide the underlying event counts or Kaplan-Meier estimates needed for a reconstruction of absolute MACE risks, so those quantities are intentionally not inferred.

Clinical Biostats methodology: A trial-results page should distinguish reported evidence from statistical interpretation. For STRENGTH, that means preserving the registry's endpoint definitions and numerical results while explaining how randomization, masking, time-to-event analysis, Cox regression, hazard ratios, confidence intervals, censoring, subgroup analysis, composite endpoints, and multiplicity affect interpretation.