This page separates reported trial information from statistical interpretation. The registry record reports the trial design, registered primary endpoint, posted outcome measures, and safety counts by arm. It does not provide a formal statistical analysis in the ClinicalTrials.gov record, so no treatment-effect estimate, confidence interval, or p-value is presented here.
1. Trial at a Glance
PRESERVED-HF was a randomized, parallel-group, quadruple-masked phase 4 trial evaluating dapagliflozin 10Mg Oral Tablet against dapagliflozin matching placebo in chronic heart failure with preserved systolic function. The registered primary endpoint evaluated heart-failure-related health status using the Kansas City Cardiomyopathy Questionnaire Clinical Summary Score from baseline to Week 12.
| Feature | PRESERVED-HF |
|---|---|
| Trial name | PRESERVED-HF |
| Brief title | Dapagliflozin in PRESERVED Ejection Fraction Heart Failure |
| Phase | Phase 4 |
| Status | COMPLETED |
| Condition | Chronic Heart Failure With Preserved Systolic Function |
| Allocation | RANDOMIZED |
| Design model | PARALLEL |
| Masking | QUADRUPLE |
| Primary purpose | TREATMENT |
| Enrollment | 324.0 |
| Arms | 2 |
| Lead sponsor | Saint Luke's Health System |
| Sponsor type | OTHER |
| ClinicalTrials.gov | NCT03030235 |
2. Clinical Question
The central question registered for PRESERVED-HF was whether dapagliflozin, compared with placebo, affects heart-failure-related health status as measured by the Kansas City Cardiomyopathy Questionnaire Clinical Summary Score (KCCQ-CS) from baseline to Week 12.
Population
Patients enrolled in the trial with chronic heart failure with preserved systolic function.
Intervention
Dapagliflozin 10Mg Oral Tablet.
Comparator
Dapagliflozin matching placebo.
Primary question
What is the effect of dapagliflozin, as compared with placebo, on heart-failure-related health status using the KCCQ-CS from baseline to Week 12?
3. Trial Design
Dapagliflozin
- Dapagliflozin 10Mg Oral Tablet
- Randomized parallel-group assignment
- Quadruple-masked trial design
Matching placebo
- Dapagliflozin matching placebo
- Randomized parallel-group assignment
- Quadruple-masked trial design
The ClinicalTrials.gov record identifies 324 participants and 2 arms. The serious-adverse-event denominator is 162 in each arm. No additional treatment schedule, duration of exposure, crossover procedure, stratification factors, or analysis-population definitions are provided in the ClinicalTrials.gov record.
4. Endpoints
| Endpoint | Registered definition | Time frame |
|---|---|---|
| Primary endpoint | Effect of Dapagliflozin, as Compared With Placebo, on Heart Failure Related Health Status Using the Kansas City Cardiomyopathy Questionnaire Clinical Summary Summary Score (KCCQ-CS) | Baseline to Week 12 |
What the KCCQ-CS measures
The registry defines the KCCQ as a disease-specific health status instrument composed of 23 items that quantifies the domains of physical limitations, symptoms, self-efficacy, social limitation and quality of life from heart failure.
The KCCQ-CS is scored from 0-100, with higher scores reflecting better health status. The registry definition states that the primary endpoint is the KCCQ-CS at Week 12.
The registered time frame is explicitly defined as baseline to Week 12. The endpoint is continuous and bounded on a 0-100 scale, with higher values representing better health status.
5. Primary Endpoint and Planned Analysis
The ClinicalTrials.gov record contains a posted outcome measure for the primary endpoint but contain no formal statistical analyses. Therefore, there is no registry-reported treatment-effect estimate, confidence interval, or p-value to reproduce for the KCCQ-CS endpoint.
How an endpoint of this type would typically be analysed
Because KCCQ-CS is a continuous score measured at baseline and again at Week 12, a prespecified analysis would typically compare the Week 12 outcome between randomized treatment groups while accounting for baseline KCCQ-CS. An analysis of covariance (ANCOVA) is a common approach for this structure because it estimates the treatment-group difference at follow-up while adjusting for baseline measurement.
An alternative longitudinal model could incorporate repeated KCCQ measurements if multiple post-baseline assessments were prespecified. The appropriate method depends on the trial's statistical analysis plan, including the exact estimand, covariance structure, missing-data assumptions, and analysis population. Those details are not reported in the ClinicalTrials.gov record.
The scientifically relevant quantity for a randomized KCCQ-CS comparison would ordinarily be a between-group difference in health-status score, ideally accompanied by a confidence interval. A positive difference would indicate a higher estimated Week 12 KCCQ-CS in the dapagliflozin group under the specified analysis.
That interpretation should not be confused with a claim about mortality, hospitalization, or another clinical outcome. KCCQ-CS specifically measures heart-failure-related health status. The ClinicalTrials.gov record does not provide the numerical treatment-group comparison, so no magnitude of effect can be inferred from this record alone.
6. Statistical Methodology
Randomization
Randomization is the fundamental design feature that permits a treatment-group comparison to be interpreted causally under the assumptions of the randomized trial. Before outcomes are observed, assignment to the two parallel treatment arms is intended to distribute measured and unmeasured prognostic factors without systematic treatment-selection bias.
For a continuous endpoint such as KCCQ-CS, randomization means that a difference in follow-up scores can be attributed to the assigned intervention more directly than in an uncontrolled observational comparison, provided that the prespecified analysis and trial conduct preserve the randomized comparison.
Baseline adjustment
For a baseline and Week 12 continuous outcome, adjusting for the baseline score can improve precision because participants begin the trial with different health-status measurements. ANCOVA is one standard framework for estimating the treatment-group difference at follow-up after accounting for baseline outcome level.
The treatment coefficient represents the adjusted difference between randomized treatment groups under this conceptual model. The actual model used for PRESERVED-HF cannot be established from the ClinicalTrials.gov record because no formal statistical analysis is posted.
Continuous outcome interpretation
A KCCQ-CS score is measured on a 0-100 scale, with higher scores indicating better health status. The natural effect measure is therefore a difference in score rather than a hazard ratio or odds ratio. This distinction matters: methods designed for time-to-event outcomes are not automatically appropriate for a continuous health-status endpoint.
Confidence intervals
A confidence interval around a treatment-group difference would quantify statistical uncertainty under the specified model and sampling framework. It would not describe the range of individual patient responses. The width of the interval would also provide information about the precision of the estimated treatment effect.
Analysis population
The ClinicalTrials.gov record does not identify an intention-to-treat population, per-protocol population, modified intention-to-treat population, or specific missing-data analysis set. Consequently, the ClinicalTrials.gov record does not establish which participants were included in a formal efficacy analysis.
7. Statistical Methods Explained
Why would ANCOVA be useful for KCCQ-CS?
Because the primary outcome is a continuous score measured at baseline and Week 12, baseline adjustment can account for differences in starting health-status scores and potentially improve precision. ANCOVA does this by modelling the Week 12 score while including baseline score as a covariate.
Why not use a hazard ratio for KCCQ-CS?
A hazard ratio is designed for a time-to-event outcome such as death or time to hospitalization. KCCQ-CS is a continuous health-status score assessed at a specified time point. A between-group mean or adjusted mean difference is therefore a more natural effect measure for this endpoint.
What does a positive KCCQ-CS difference mean?
Because higher KCCQ-CS scores represent better health status, a positive estimated difference for dapagliflozin versus placebo would indicate higher estimated health-status scores in the dapagliflozin group at the analysed follow-up, after accounting for whatever baseline adjustment the prespecified model uses.
Why does randomization matter statistically?
Randomization establishes the treatment assignment mechanism before outcome information is available. This reduces the risk that systematic differences in prognosis explain the observed treatment-group comparison. Randomization does not guarantee identical groups in every measured characteristic, particularly in finite samples, but it provides the foundation for unbiased treatment-effect estimation under the trial's assumptions.
What does quadruple masking add?
Quadruple masking means that masking was incorporated into the trial design at four levels as specified by the registry's masking classification. Masking can reduce the influence of treatment knowledge on participant behavior, clinical management, outcome assessment, or other trial processes, depending on which parties are masked. The ClinicalTrials.gov record does not identify the four masked parties individually.
Why is the absence of a posted formal analysis important?
The distinction between measurement and analysis is fundamental. Knowing that KCCQ-CS was assessed at Week 12 does not reveal the treatment-group difference, its uncertainty, the model used, or the hypothesis-test result. Those quantities must come from a reported statistical analysis rather than being reconstructed from the endpoint definition alone.
8. Safety Results
The ClinicalTrials.gov record reports serious adverse events by randomized arm using affected participants over participants at risk. These are safety counts, not efficacy comparisons, and the registry data do not provide a formal statistical comparison for serious adverse events.
| Safety measure | Dapagliflozin | Matching placebo |
|---|---|---|
| Serious adverse events, affected / at risk | 31/162 | 22/162 |
The reported serious-adverse-event counts should be interpreted with their denominators: 31 of 162 participants in the dapagliflozin arm and 22 of 162 in the placebo arm were affected. The ClinicalTrials.gov record does not report a risk ratio, risk difference, confidence interval, p-value, exposure-adjusted rate, seriousness-category breakdown, or formal safety hypothesis test.
9. Design Topics the Registry Does Not Specify
Several statistical design features that can materially affect interpretation are not provided in the ClinicalTrials.gov record. They should therefore not be inferred from the randomized, quadruple-masked design alone.
| Design topic | What the ClinicalTrials.gov record establishes | What is not established |
|---|---|---|
| Stratification | The trial was randomized. | No stratification factors are provided. |
| Multiplicity | One registered primary endpoint is identified. | No alpha-allocation or multiplicity procedure is provided. |
| Interim analysis | No interim-analysis information is provided. | No interim stopping boundary or alpha-spending approach can be established. |
| Missing data | The endpoint is measured from baseline to Week 12. | No imputation or missing-data strategy is provided. |
| Crossover | The trial is described as a parallel design. | No crossover procedure is reported in the ClinicalTrials.gov record. |
| Bayesian methods | No Bayesian analysis is identified. | No prior distribution, posterior analysis, or Bayesian decision rule is provided. |
| Analysis populations | Enrollment and treatment arms are reported. | No formal ITT, per-protocol, or other analysis-population definition is reported. |
10. Planned Analysis: What Would Be Needed for a Complete Statistical Result?
A complete statistical interpretation of the registered KCCQ-CS endpoint would require more than the endpoint definition. At minimum, the analysis record would need to establish the treatment-group estimates at Week 12, the effect measure used for the comparison, its confidence interval, and the corresponding hypothesis-test result if one was prespecified.
Effect estimate
The numerical between-group treatment effect would show the magnitude and direction of the KCCQ-CS difference.
Confidence interval
A confidence interval would show the statistical precision surrounding the estimated treatment effect.
P-value
If a formal hypothesis test was prespecified, its p-value would describe evidence against the null hypothesis under the specified testing framework.
Analysis population
The analysis set would clarify which randomized participants contributed to the formal efficacy comparison.
11. Limitations
- No formal statistical analysis is reported: the trial data contain a registered primary endpoint and posted outcome measures but no formal statistical analyses.
- No treatment-effect estimate: the ClinicalTrials.gov record does not provide a numerical between-group KCCQ-CS difference or another formal effect measure.
- No confidence interval or p-value: statistical uncertainty and hypothesis-test results cannot be established from the ClinicalTrials.gov record.
- No analysis-population definition: the ClinicalTrials.gov record does not identify whether the posted endpoint was analysed using an intention-to-treat, modified intention-to-treat, per-protocol, or another population.
- No missing-data strategy: the ClinicalTrials.gov record does not identify imputation, likelihood-based handling, or another missing-data approach.
- No stratification information: randomization is documented, but the ClinicalTrials.gov record does not identify stratification factors.
- Safety information is limited: serious adverse events are reported as affected/at-risk counts, without a formal comparative analysis in the ClinicalTrials.gov record.
- Endpoint-specific interpretation: KCCQ-CS is a health-status measure and should not be interpreted as a direct measure of mortality, hospitalization, or another clinical endpoint.
12. Why This Trial Matters Statistically
PRESERVED-HF is a useful teaching case because it illustrates a different statistical problem from a typical time-to-event cardiovascular trial. Its registered primary endpoint is a patient-reported health-status measure represented by a bounded continuous score rather than a survival endpoint. The design also combines randomization with quadruple masking, creating a useful framework for discussing how trial design and endpoint structure interact.
| Concept | How it appears in PRESERVED-HF |
|---|---|
| Randomization | Participants were assigned in a randomized parallel-group design. |
| Parallel-group design | The trial has 2 arms evaluated in parallel. |
| Quadruple masking | The registry classifies the trial as QUADRUPLE masked. |
| Continuous outcome | KCCQ-CS is a 0-100 health-status score. |
| Baseline-to-follow-up analysis | The registered primary time frame is baseline to Week 12. |
| ANCOVA | A common approach for a continuous follow-up endpoint with a baseline measurement, although the actual registry-reported analysis is not specified. |
| Confidence intervals | Would quantify uncertainty around a reported treatment-effect estimate if a formal analysis were posted. |
| Missing data | The ClinicalTrials.gov record does not identify the method used to handle missing Week 12 scores. |
| Safety analysis | Serious adverse events are reported by arm as affected participants over participants at risk. |
13. Blinding and Randomization as Statistical Concepts
Two design concepts identified in the trial's learning pathway are particularly important for understanding PRESERVED-HF: randomization and blinding.
Randomization
Randomization determines how treatment assignment is generated rather than allowing investigators or participants to choose the assigned intervention. In a parallel-group trial, the treatment groups can then be compared on the registered endpoint while preserving the treatment-assignment framework.
Blinding
The registry classifies PRESERVED-HF as quadruple masked. Masking can reduce opportunities for knowledge of treatment assignment to influence trial behavior or assessment. This is particularly relevant for a patient-reported health-status endpoint because treatment knowledge can potentially affect perceptions, reporting, or interactions surrounding outcome measurement.
Why these concepts matter together
Randomization addresses the allocation mechanism, while masking addresses knowledge of that allocation. They solve different methodological problems. A randomized trial can therefore benefit from considering both the validity of treatment assignment and the possibility that knowledge of assignment could influence trial conduct or outcome measurement.
14. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The registry establishes a randomized comparison of dapagliflozin with matching placebo and specifies KCCQ-CS at Week 12 as the primary endpoint. It does not supply a formal statistical comparison for that endpoint in the data provided.
Clinical interpretation
The primary endpoint concerns heart-failure-related health status. A complete clinical interpretation would require the actual treatment-group results and their uncertainty rather than relying on the endpoint definition or trial design alone.
Trial design tells us how the comparison was constructed; endpoint definition tells us what was measured; statistical analysis tells us what treatment effect was estimated and how uncertain that estimate is. For PRESERVED-HF, the ClinicalTrials.gov record provides the first two elements in detail but not the formal statistical comparison.
15. Trial Timeline
PRESERVED-HF begins
The ClinicalTrials.gov record lists 2017-03-01 as the trial start date.
Primary completion
The ClinicalTrials.gov record lists 2021-08-13 as the primary completion date.
Trial status
The registry status in the ClinicalTrials.gov record is COMPLETED, with 324 participants enrolled across 2 arms.
16. Secondary Endpoint Results
The ClinicalTrials.gov record reports 11 outcome measures in total, but the ClinicalTrials.gov record do not include numerical results or formal statistical analyses for secondary endpoints. Accordingly, this page does not assign treatment effects to secondary outcomes or infer their direction.
17. Interpreting the Reported Safety Counts
The numerator is the number of affected participants and the denominator is the number at risk reported for each arm.
Because the denominators are equal, the raw counts can be read directly as the number of participants affected in each randomized arm. However, the ClinicalTrials.gov record does not provide exposure duration or a formal inferential analysis, so the counts should remain descriptive.
A formal comparative safety analysis could use measures such as a risk difference or risk ratio when the event window and analysis population are appropriately defined. The choice of measure would depend on the statistical analysis plan and the structure of the adverse-event data.
18. What a Complete KCCQ-CS Result Would Tell Us
Magnitude
The estimated between-group difference would quantify how much the Week 12 KCCQ-CS differed between randomized treatment groups.
Direction
Because higher KCCQ-CS scores represent better health status, the sign of the treatment difference would indicate which group had the higher estimated score.
Precision
A confidence interval would indicate how precisely the treatment effect was estimated under the selected analysis framework.
Evidence against the null
If a p-value were reported from a prespecified test, it would describe compatibility with the null hypothesis under that testing framework, not the clinical magnitude of the effect.
19. Related Tutorials
Learn more about the methods used in this trial:
20. Related Statistical Calculators
21. Sources
- ClinicalTrials.gov: NCT03030235 — PRESERVED-HF.
- Linked publication: PubMed PMID 34711976.
- Linked publication: PubMed PMID 39421941.
- Linked publication: PubMed PMID 38639697.
- Linked publication: PubMed PMID 37869881.
- Linked publication: PubMed PMID 37203441.
Continue through the Clinical Biostats statistical library
Explore trial-design concepts, statistical tutorials, and calculators that can be used to understand randomized clinical-trial methodology.
22. Record Summary
PRESERVED-HF is a completed phase 4, randomized, parallel-group, quadruple-masked trial with 324 enrolled participants and 2 arms evaluating dapagliflozin 10Mg Oral Tablet versus dapagliflozin matching placebo in chronic heart failure with preserved systolic function. Its registered primary endpoint is the effect of dapagliflozin compared with placebo on heart-failure-related health status using the KCCQ-CS from baseline to Week 12.
The statistical teaching value of the trial lies in the distinction between trial design, endpoint definition, and formal statistical evidence. The ClinicalTrials.gov record establishes randomization, parallel allocation, quadruple masking, the KCCQ-CS endpoint, and serious-adverse-event counts by arm. It does not supply formal statistical analyses for the primary endpoint, so a responsible statistical review must stop short of assigning an effect estimate, confidence interval, or p-value that is not present in the record.