The ClinicalTrials.gov record lists CESAR as NOT_YET_RECRUITING. The registry reports a planned enrollment of 25000 and one registered primary endpoint. No statistical analysis results are posted in the registry.
1. Trial at a Glance
CESAR is a prospective cohort of acute cardiology referrals in an ambulatory day-hospital setting. The study concerns patients presenting for suspected acute or sub-acute cardiovascular disease and defines its primary outcome as the proportion with a confirmed cardiovascular diagnosis during six months of follow-up.
| Feature | CESAR |
|---|---|
| Acronym | CESAR |
| Official title | Prospective Cohort of Acute Cardiology Referrals in an Ambulatory Day-Hospital Setting |
| Design | Prospective cohort |
| Status | NOT_YET_RECRUITING |
| Planned enrollment | 25000 |
| Study start | 2026-08-01 |
| Primary completion date | 2027-02-01 |
| Lead sponsor | Assistance Publique - Hôpitaux de Paris |
| Sponsor type | OTHER |
| ClinicalTrials.gov | NCT07675434 |
2. Clinical Question
The central clinical question is descriptive: among consecutive patients consulting for suspicion of acute or sub-acute cardiovascular disease at an ambulatory day-hospital unit, what proportion have a confirmed cardiovascular diagnosis during the specified follow-up period?
Population
Consecutive patients consulting for suspicion of acute or sub-acute cardiovascular disease at an ambulatory day-hospital unit.
Clinical setting
An ambulatory day-hospital unit receiving acute cardiology referrals.
Conditions
Acute Coronary Syndromes; Heart Failure; Pulmonary Embolism, Deep Vein Thrombosis; HTN-Hypertension; Arrhythmia.
Primary question
What proportion of the eligible consecutive patients receive a confirmed cardiovascular diagnosis from enrollment through six months of follow-up?
This is fundamentally different from a randomized treatment comparison. The registry describes a prospective cohort rather than two randomized treatment groups. Consequently, the primary statistical quantity is a proportion, not a hazard ratio, treatment difference, or randomized risk ratio.
3. Trial Design
The prospective-cohort structure is statistically important because the study is intended to characterize the diagnostic yield of an acute cardiology referral pathway. There is no randomized allocation described in the registry information, so causal comparisons between interventions are not part of the registered primary question.
Consecutive patients and selection
The phrase consecutive patients has statistical importance. If enrollment genuinely captures patients sequentially as they present for the relevant clinical indication, the resulting cohort can provide a more direct description of the diagnostic composition of that referral population than a convenience sample would.
That advantage depends on how consistently the eligibility criteria are applied and how completely eligible referrals are captured. A proportion estimated from a selected subset of referrals can answer a narrower question than the same proportion estimated from a truly consecutive clinical stream.
4. Registered Primary Endpoint
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Primary endpoint | Proportion of patients with confirmed cardiovascular diagnosis among consecutive patients consulting for suspicion of acute or sub-acute cardiovascular disease at the ambulatory day-hospital unit. | From enrollement to six months of follow-up. |
The endpoint is a binary patient-level outcome summarized at the cohort level: each patient either contributes to the numerator if a confirmed cardiovascular diagnosis is established or contributes to the denominator without such a confirmed diagnosis. The registry's wording defines the population and follow-up period, while the reported statistical quantity is the resulting proportion.
The primary descriptive quantity is therefore a proportion p. Its interpretation depends on the denominator: it describes the observed diagnostic proportion among the defined consecutive referral population, rather than the probability of cardiovascular disease in an unrelated population.
5. Planned Analysis
The ClinicalTrials.gov record identifies the primary endpoint but does not report a posted statistical analysis for it. Results have not been posted on ClinicalTrials.gov.
Analysis of the primary proportion
For an endpoint defined as the proportion of patients with a confirmed diagnosis, the natural primary summary is the number of patients meeting the endpoint divided by the number of patients contributing to the defined cohort. The result can be expressed as a percentage for clinical readability.
A confidence interval around the proportion is also important. A point estimate alone does not communicate sampling uncertainty. For a binary outcome, an exact binomial interval or another prespecified binomial confidence-interval method can quantify uncertainty around the observed proportion.
Here, X represents the number of patients with a confirmed cardiovascular diagnosis, n represents the relevant number of patients in the cohort, and p represents the underlying proportion being estimated.
Why the denominator matters
For a proportion, the denominator is not a technical detail. If some patients do not complete the six-month follow-up or otherwise cannot be classified according to the endpoint definition, the analysis population must be defined carefully. The resulting percentage can change depending on whether the denominator consists of all enrolled patients, all evaluable patients, or another prespecified population.
The registry information identifies the endpoint and time frame but does not provide a detailed missing-data strategy. A final analysis would therefore need to make the relationship between enrollment, evaluability, follow-up, and endpoint ascertainment explicit.
Confidence intervals are more informative than a percentage alone
Suppose a future analysis reports a proportion p. The point estimate answers how much of the observed cohort met the endpoint. The confidence interval addresses how precisely that proportion was estimated. With a large cohort, the interval can be narrow even when the underlying endpoint is clinically heterogeneous.
That distinction is particularly useful for a study with a planned enrollment of 25000. Large sample size can provide high numerical precision for a cohort proportion, but precision should not be confused with generalizability or freedom from selection bias.
6. Statistical Methodology
Descriptive estimation of a proportion
The primary endpoint is naturally summarized as a binomial proportion. If x patients among n eligible patients have a confirmed cardiovascular diagnosis, the estimated proportion is:
Multiplying p̂ by 100 expresses the same quantity as a percentage. The estimate is descriptive of the defined cohort and its specified follow-up period.
Binomial confidence intervals
Because the endpoint is binary, confidence intervals for a proportion can be constructed using binomial methods. Exact binomial intervals are one option, while score-based methods such as the Wilson interval are another. The important statistical principle is that the interval should be specified consistently with the analysis plan rather than selected after seeing the observed result.
Six-month follow-up
The endpoint's time frame is from enrollment to six months of follow-up. This makes endpoint ascertainment a longitudinal process even though the final primary summary is a proportion. The analysis must therefore distinguish a patient who is observed throughout the specified period from a patient whose diagnostic status cannot be established within that period.
Time frame versus time-to-event analysis
A six-month endpoint does not automatically make a study a survival-analysis study. The registered endpoint is a proportion over a defined follow-up period, not time from enrollment to diagnosis. Unless the registry specifies a time-to-diagnosis endpoint, Kaplan-Meier estimation and Cox regression would answer a different question from the registered primary endpoint.
Descriptive association versus causal inference
A prospective cohort can establish temporal ordering between enrollment and subsequent diagnostic assessment, but prospective observation alone does not create randomization. If patient characteristics are later compared with diagnostic status, those analyses would describe associations unless a design and analytic strategy supports a causal interpretation.
7. Statistical Methods Explained
What exactly is being estimated?
The primary quantity is the proportion of consecutive patients in the defined referral population who have a confirmed cardiovascular diagnosis. It is not a treatment effect. There is no randomized intervention-versus-control contrast in the registered study description.
Why use a binomial model?
The primary endpoint can be represented as a binary outcome for each patient: the patient either meets the definition of having a confirmed cardiovascular diagnosis or does not. Summing these binary outcomes produces the numerator of the cohort proportion, making the binomial framework a natural statistical representation.
Why report a confidence interval?
A percentage such as 30% or 60% is a point estimate. It does not communicate how much statistical uncertainty surrounds the estimate. A confidence interval supplies an interval-based measure of precision under the chosen statistical framework.
Does a large sample guarantee an unbiased estimate?
No. A large sample can substantially improve precision while leaving systematic sources of bias intact. If the enrolled cohort differs systematically from the broader population of patients with suspected acute or sub-acute cardiovascular disease, increasing the sample size does not by itself remove that selection issue.
Why does "consecutive" matter?
Consecutive enrollment is intended to reduce the opportunity for investigators or referral processes to selectively include particular patients. If consecutive capture is implemented consistently, the resulting cohort can more closely describe the clinical referral stream from which it was assembled.
Why isn't a hazard ratio the primary measure?
The registered endpoint is a proportion at a specified follow-up horizon. A hazard ratio would instead compare instantaneous event rates over time and would require a time-to-event outcome. Applying a hazard ratio to the registered proportion would change the statistical question.
8. Interpretation of the Primary Endpoint
If the final analysis reports a proportion p, that value would describe the fraction of the defined consecutive referral cohort with a confirmed cardiovascular diagnosis during the registered follow-up period. It would be an estimate of diagnostic yield within that clinical pathway.
The proportion would not automatically represent the prevalence of cardiovascular disease in the general population, the probability that an arbitrary patient has cardiovascular disease, or a causal effect of the ambulatory day-hospital pathway. Those interpretations require populations and study designs appropriate to those questions.
A confidence interval would describe statistical uncertainty around the estimated cohort proportion under the selected binomial framework. It would not describe the range of diagnostic probabilities for individual patients, nor would it account automatically for systematic referral or ascertainment bias.
The endpoint is tied to the specified six-month follow-up. Extending or shortening the observation period can change which diagnoses are confirmed and therefore change the proportion. The time horizon is consequently part of the definition of the estimand rather than merely an administrative follow-up detail.
9. Conditions Represented in the Cohort
| Condition listed in the registry | Statistical relevance |
|---|---|
| Acute Coronary Syndromes | Part of the cardiovascular conditions represented by the study population. |
| Heart Failure | Part of the cardiovascular conditions represented by the study population. |
| Pulmonary Embolism, Deep Vein Thrombosis | Part of the conditions represented by the study population. |
| HTN-Hypertension | Part of the conditions represented by the study population. |
| Arrhythmia | Part of the cardiovascular conditions represented by the study population. |
The list of conditions provides clinical context for the referral population, but the registered primary endpoint is broader: it asks for the proportion with a confirmed cardiovascular diagnosis. The condition list should therefore not be treated as if it were a set of separate primary endpoints.
10. Enrollment and Study Timing
Study start
The registry lists 2026-08-01 as the study start date.
NOT_YET_RECRUITING
The ClinicalTrials.gov record currently lists the study as not yet recruiting.
Primary completion
The registry lists 2027-02-01 as the primary completion date.
The planned enrollment is 25000. For a proportion endpoint, this scale can provide substantial statistical precision if the planned cohort is successfully enrolled and the endpoint is ascertained consistently. The inferential value of the resulting estimate will still depend on how the enrolled patients relate to the target referral population.
11. No Posted Results
The absence of a posted statistical result is consistent with the registry's current NOT_YET_RECRUITING status. The primary endpoint remains a planned measurement: the proportion of patients with confirmed cardiovascular diagnosis among consecutive patients consulting for suspicion of acute or sub-acute cardiovascular disease at the ambulatory day-hospital unit, from enrollment to six months of follow-up.
12. What Would Constitute a Complete Statistical Result?
For this endpoint, a statistically informative result would contain more than a single percentage. A clear report would identify the number of patients included in the denominator, the number who met the confirmed-diagnosis definition, the resulting proportion, and an appropriate confidence interval.
Numerator
The number of patients meeting the registered definition of having a confirmed cardiovascular diagnosis.
Denominator
The patient population to which the primary endpoint proportion applies, with its eligibility and follow-up handling clearly defined.
Point estimate
The observed proportion, conventionally expressed as both a fraction and a percentage.
Precision
A confidence interval showing statistical uncertainty around the estimated proportion.
This structure makes the result auditable. A percentage without its numerator and denominator can be difficult to interpret, particularly when follow-up completeness or eligibility changes the analysis population.
13. Missing Data and Follow-Up Considerations
The registry defines the endpoint over six months, which makes follow-up completeness relevant to the interpretation of the final proportion. Patients for whom the diagnostic outcome cannot be established within the specified period create a distinction between enrollment and endpoint evaluability.
The ClinicalTrials.gov record does not report a detailed missing-data or imputation strategy. Statistically, several approaches can be considered in a completed cohort analysis, but their appropriateness depends on why endpoint information is missing and on the prespecified analysis population.
| Issue | Why it matters |
|---|---|
| Incomplete follow-up | A patient without sufficient follow-up may have an unknown diagnostic status at the registered endpoint horizon. |
| Endpoint ascertainment | The confirmed-diagnosis classification must be applied consistently across the cohort. |
| Denominator definition | The estimated proportion depends directly on which patients are included in the denominator. |
| Missingness mechanism | If loss to follow-up is related to clinical characteristics, a complete-case proportion can differ systematically from the underlying cohort proportion. |
For a binary endpoint, imputation is not automatically preferable to retaining missingness as an explicit analysis issue. The appropriate strategy depends on the extent and pattern of missing information and should be specified before the final outcome analysis when possible.
14. Why This Trial Matters Statistically
CESAR illustrates a useful clinical-trial principle: not every important study question requires a randomized comparison or a time-to-event model. A large prospective cohort can address a fundamentally descriptive question about the diagnostic yield of a defined clinical referral pathway.
| Concept | How it appears in CESAR |
|---|---|
| Prospective cohort | Patients are followed prospectively from enrollment in an ambulatory day-hospital referral setting. |
| Consecutive enrollment | The primary endpoint explicitly refers to consecutive patients consulting for suspected acute or sub-acute cardiovascular disease. |
| Binary outcome | The primary endpoint distinguishes patients with and without a confirmed cardiovascular diagnosis. |
| Proportion estimation | The principal statistical quantity is the fraction of the defined cohort meeting the diagnosis endpoint. |
| Confidence intervals | Precision around a cohort proportion can be quantified using a binomial confidence interval. |
| Follow-up horizon | The primary endpoint is assessed from enrollment through six months of follow-up. |
| Generalizability | The clinical setting and consecutive-referral population determine the population to which the estimate can reasonably be applied. |
| Missing data | Six-month endpoint ascertainment makes follow-up completeness relevant to the final denominator and interpretation. |
The statistical lesson is that estimand definition comes before model selection. Once the endpoint is understood as a six-month cohort proportion, the natural first analysis is a proportion with uncertainty, not a more elaborate model chosen simply because it is commonly used in clinical research.
15. Limitations
- No randomized comparator: the registered design is a prospective cohort, so the primary endpoint is descriptive rather than a randomized treatment-effect comparison.
- Referral population: the study concerns patients consulting for suspicion of acute or sub-acute cardiovascular disease at an ambulatory day-hospital unit. Its results would therefore be tied to that clinical pathway rather than automatically representing the general population.
- Consecutive enrollment is operational: the statistical value of a consecutive cohort depends on consistent capture of eligible patients and consistent application of the endpoint definition.
- Follow-up dependence: the endpoint is defined from enrollment to six months of follow-up, so incomplete follow-up can affect endpoint ascertainment and the analysis denominator.
- Broad diagnostic endpoint: the registered primary outcome concerns confirmed cardiovascular diagnosis, while the registry lists several individual conditions. A single overall proportion does not by itself describe the distribution of every listed condition.
- No posted statistical analysis: the registry currently provides the endpoint definition but no statistical-analysis results from which precision or observed diagnostic yield can be assessed.
- Precision versus validity: a large enrollment can make a proportion statistically precise without eliminating systematic referral, ascertainment, or generalizability limitations.
16. Clinical Interpretation vs Statistical Interpretation
Statistical interpretation
The primary endpoint is a proportion measured over a defined six-month period. Its natural statistical summary is a numerator, denominator, point estimate, and confidence interval using an appropriate binomial framework.
Clinical interpretation
The resulting proportion, once reported, would describe the diagnostic yield among the defined consecutive acute cardiology referral population. Its clinical meaning depends on the setting and population represented by the cohort.
Keeping these interpretations separate is important. Statistical precision describes how well the study estimates its cohort proportion. Clinical relevance asks what that proportion means for the patients and referral pathway represented by the study.
17. Population, Estimand, and Generalizability
A useful way to understand CESAR is to separate three questions that are sometimes conflated.
| Question | CESAR perspective |
|---|---|
| Who is being studied? | Consecutive patients consulting for suspicion of acute or sub-acute cardiovascular disease at the ambulatory day-hospital unit. |
| What is being estimated? | The proportion with a confirmed cardiovascular diagnosis. |
| Over what period? | From enrollment to six months of follow-up. |
This framing defines the estimand before the numerical analysis. If the study ultimately reports a proportion, the interpretation should remain attached to this population and time horizon rather than being generalized automatically to patients outside the referral setting.
18. Sponsor and Registry Information
| Registry field | Reported information |
|---|---|
| Lead sponsor | Assistance Publique - Hôpitaux de Paris |
| Sponsor type | OTHER |
| Study acronym | CESAR |
| NCT ID | NCT07675434 |
| Status | NOT_YET_RECRUITING |
| Start date | 2026-08-01 |
| Primary completion date | 2027-02-01 |
| Enrollment | 25000 |
These registry characteristics describe the study's administrative and operational context. They should be kept distinct from the statistical estimand itself: sponsorship, recruitment status, and planned enrollment do not constitute evidence about the eventual magnitude of the primary endpoint.
19. Statistical Concepts in This Trial
The statistical structure of CESAR centers on prospective cohort measurement, binary outcome definition, proportion estimation, confidence intervals, follow-up completeness, and interpretation of a clinically defined referral population.
20. Sources
- ClinicalTrials.gov: CESAR (NCT07675434).
Continue through Clinical Biostats
Explore statistical methods for clinical research, cohort studies, binary outcomes, confidence intervals, and clinical-trial analysis.
21. Record Summary
CESAR is a prospective cohort of acute cardiology referrals in an ambulatory day-hospital setting, with a planned enrollment of 25000. The registered primary endpoint is the proportion of patients with confirmed cardiovascular diagnosis among consecutive patients consulting for suspicion of acute or sub-acute cardiovascular disease at the ambulatory day-hospital unit, measured from enrollment to six months of follow-up.
Statistically, this endpoint is most naturally understood as a binomial proportion. The key quantities for a completed analysis would be the number of patients meeting the confirmed-diagnosis definition, the appropriate denominator, the resulting proportion, and a confidence interval describing statistical precision. The interpretation must remain tied to the defined consecutive referral population and its six-month follow-up period.
The ClinicalTrials.gov record currently lists CESAR as NOT_YET_RECRUITING and does not report posted statistical-analysis results. The study is listed with a start date of 2026-08-01 and a primary completion date of 2027-02-01, with Assistance Publique - Hôpitaux de Paris as the lead sponsor.