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
COMPASS AF is a randomized, parallel-group clinical trial comparing two procedural approaches to ablation in patients with atrial fibrillation. The registry identifies recurrence of atrial tachyarrhythmia from 3 months to 12 months after the index procedure as the primary endpoint.
| Feature | COMPASS AF |
|---|---|
| Trial acronym | COMPASS AF |
| ClinicalTrials.gov identifier | NCT07726901 |
| Condition | Atrial Fibrillation (AF) |
| Phase | NA |
| Status | RECRUITING |
| Allocation | RANDOMIZED |
| Design model | PARALLEL |
| Masking | NONE |
| Primary purpose | TREATMENT |
| Planned enrollment | 360 |
| Number of arms | 2 |
| Lead sponsor | Samsung Medical Center |
2. Clinical Question
The central statistical question is whether two different ablation strategies can be compared with respect to recurrence of atrial tachyarrhythmia during the period from 3 months to 12 months after the index procedure.
Population
Patients enrolled in a clinical trial for atrial fibrillation (AF), with the trial title identifying persistent atrial fibrillation as the disease setting.
Intervention
Anatomy-based ablation.
Comparator
Emphasize map-guided ablation.
Primary question
How do the two randomized ablation strategies compare with respect to recurrence of atrial tachyarrhythmia after the index procedure?
3. Trial Design
Anatomy-based ablation
- Intervention type: procedure
- Randomized treatment strategy
- Primary comparison is recurrence of atrial tachyarrhythmia
Emphasize map-guided ablation
- Intervention type: procedure
- Randomized comparator strategy
- Primary comparison is recurrence of atrial tachyarrhythmia
The registry classifies COMPASS AF as a randomized parallel-group trial with no masking. The planned enrollment is 360 participants across two arms. The start date is May 19, 2026, and the primary completion date is December 31, 2030.
4. Endpoint Framework
The registry identifies one primary endpoint: recurrence of atrial tachyarrhythmia. The endpoint is binary, meaning each participant is classified according to whether recurrence is present or absent during the specified assessment period.
| Endpoint | Time frame | Registry definition | Outcome type |
|---|---|---|---|
| Recurrence of atrial tachyarrhythmia | From 3 months to 12 months after the index procedure | Any documented atrial tachyarrhythmias excluding the initial 3-month blanking period. Documented atrial tachyarrhythmias include atrial fibrillation, atrial flutter, and atrial tachycardia, defined by atrial tachyarrhythmias documented in 12 lead ECG or atrial tachyarrhythmias of more than 30 seconds in Holter monitoring. | Binary: presence or absence of recurrence |
5. Understanding the Primary Endpoint
The primary endpoint is fundamentally different from a time-to-event endpoint such as overall survival. It is recorded as a binary outcome: each participant is ultimately classified as having recurrence or not having recurrence according to the specified monitoring definition and time window.
The qualifying recurrence must satisfy the registry's documentation definition and occur after the initial 3-month blanking period and through 12 months after the index procedure.
This structure creates a straightforward treatment-group comparison. If the recurrence proportion differs between the two randomized groups, the magnitude of that difference can be summarized using an absolute risk difference, a risk ratio, or an odds ratio. The choice of effect measure determines how the result should be interpreted.
Why the measurement window matters
A binary recurrence rate is meaningful only when its observation period is specified. Here, the registry defines the relevant period as beginning at 3 months and ending at 12 months after the index procedure. A recurrence during the initial blanking period does not count toward the primary endpoint.
Why documentation criteria matter
The endpoint is not simply a subjective report of symptoms. The registry specifies documented atrial tachyarrhythmias identified by 12 lead ECG or atrial tachyarrhythmias lasting more than 30 seconds on Holter monitoring. This operational definition determines which observations enter the binary endpoint.
6. Planned Analysis
ClinicalTrials.gov has not posted statistical results for the primary endpoint. The registry therefore currently describes the endpoint that will be measured rather than reporting an observed treatment effect.
How a binary recurrence endpoint is typically analysed
For a randomized two-arm comparison with a binary endpoint, the primary descriptive analysis would normally begin by reporting the number and proportion of participants with recurrence in each treatment group. The between-group treatment effect can then be expressed using an absolute risk difference, a relative risk, or an odds ratio.
| Measure | Concept | Interpretation |
|---|---|---|
| Risk difference | p1 − p0 |
Absolute difference in recurrence probability between the two groups. |
| Risk ratio | p1 / p0 |
Relative recurrence probability in one group compared with the other. |
| Odds ratio | [p1/(1−p1)] / [p0/(1−p0)] |
Ratio of recurrence odds between the two groups. |
For an adequately sized randomized trial, a chi-square test is a common approach to comparing two proportions when its assumptions are appropriate. Fisher's exact test can be used when cell counts are sufficiently small that the chi-square approximation is unsuitable. A regression model for a binary outcome can also be used when adjustment for prespecified covariates is appropriate.
What the primary analysis should preserve
Because treatment assignment is randomized, the principal efficacy comparison should retain participants according to their randomized treatment assignment. This protects the interpretability of the treatment comparison established by randomization. Any departures from the randomized comparison, such as analyses restricted to participants with particular levels of treatment exposure or follow-up, require separate interpretation.
7. Statistical Methodology
Binary proportion comparison
The primary endpoint produces one binary observation per participant. The simplest analysis estimates a recurrence proportion for each randomized group and compares those proportions.
A negative risk difference would indicate a lower observed recurrence proportion in the first group, while a positive risk difference would indicate a higher observed recurrence proportion. The numerical direction must always be interpreted using the prespecified treatment-group ordering.
Risk ratio
The risk ratio compares recurrence probabilities directly. A risk ratio of 1 would correspond to equal recurrence probabilities. Values below 1 indicate lower recurrence probability in the numerator group, while values above 1 indicate higher recurrence probability.
Unlike an odds ratio, the risk ratio compares probabilities directly. It is often intuitive when the outcome proportion is not extremely close to 0 or 1.
Odds ratio
An odds ratio compares recurrence odds rather than recurrence probabilities. Odds and probabilities are related but are not interchangeable, particularly when the outcome is common. An odds ratio should therefore not automatically be described as a risk ratio.
Confidence intervals
A treatment-effect estimate should be accompanied by a confidence interval. The interval describes statistical uncertainty around the estimated effect under the specified analysis framework. It does not describe the range of individual patient outcomes.
Hypothesis testing
A formal comparison of the two recurrence proportions can generate a p-value. The p-value addresses the compatibility of the observed data with a specified null hypothesis; it is not a measure of the size or clinical importance of the treatment effect.
8. Statistical Methods Explained
Why is the recurrence endpoint binary?
The registry explicitly records recurrence as presence or absence. Once the defined assessment period is complete, each participant contributes one binary outcome to the primary endpoint. This allows direct comparison of recurrence proportions between randomized groups.
Why exclude the first 3 months?
The registry defines an initial 3-month blanking period and excludes it from the primary recurrence endpoint. Consequently, an event during that period is not classified as a primary endpoint recurrence. The endpoint is evaluated from 3 months through 12 months after the index procedure.
What does a risk difference tell us?
The risk difference expresses the treatment contrast in absolute percentage-point terms. For example, if one group had a recurrence probability of 30% and another had a recurrence probability of 20%, the risk difference would be 10 percentage points. This example is illustrative only and is not a COMPASS AF result.
What does a risk ratio tell us?
The risk ratio expresses one recurrence probability relative to another. A value of 0.80 would mean that the recurrence probability in the numerator group was 80% of the recurrence probability in the denominator group. It would not mean that 20% of patients avoided recurrence.
Why is the odds ratio different from the risk ratio?
The odds ratio compares odds, whereas the risk ratio compares probabilities. When recurrence is relatively common, the numerical values can differ substantially. Reporting an odds ratio as though it were a risk ratio can therefore distort the clinical interpretation.
Why are confidence intervals important?
A point estimate alone gives no indication of statistical precision. A confidence interval places the estimate within a range reflecting sampling uncertainty under the selected model and confidence level. Wider intervals indicate less precision than narrower intervals, all else being equal.
Why does the p-value not measure effect size?
A p-value is affected by both the observed data and the amount of information available. A small p-value does not tell the reader whether the treatment difference is large, while a larger p-value does not prove that the treatments are equivalent. Effect estimates and confidence intervals are needed to understand magnitude and precision.
9. Trial Timing and Current Registry Status
Trial start
The registry lists May 19, 2026 as the study start date.
Recruiting
The ClinicalTrials.gov record currently identifies COMPASS AF as recruiting.
Primary completion
The registry lists December 31, 2030 as the primary completion date.
These dates describe the registry's planned study timeline. They should not be interpreted as evidence that the primary endpoint has already been observed or analysed.
10. Interpreting a Future Treatment Effect
Once recurrence data are available, the most informative presentation would distinguish absolute recurrence proportions from relative measures. For example, a recurrence difference can be communicated in percentage points, while a risk ratio communicates the relative probability of recurrence.
Absolute effect
The difference in recurrence proportions directly describes how many percentage points separate the randomized groups during the specified endpoint window.
Relative effect
A risk ratio describes the recurrence probability in one group relative to the other. It should be interpreted using the corresponding absolute recurrence proportions.
Precision
The confidence interval indicates uncertainty around the treatment-effect estimate and helps distinguish a precise estimate from one with substantial sampling uncertainty.
Statistical evidence
A hypothesis test can address a null hypothesis about the treatment comparison, but the p-value should not substitute for the estimated effect and its confidence interval.
The importance of endpoint timing
The recurrence estimate is tied specifically to the period from 3 months to 12 months after the index procedure. An estimate based on a different follow-up window would answer a different question. Comparisons across studies therefore require careful attention to their endpoint definitions and observation periods.
11. What This Endpoint Does — and Does Not — Mean
The primary endpoint measures whether a qualifying atrial tachyarrhythmia recurrence was documented during the specified period after the initial 3-month blanking period. It incorporates atrial fibrillation, atrial flutter, and atrial tachycardia under the registry's documentation criteria.
A binary recurrence endpoint does not describe the number of recurrence episodes, time to first recurrence, duration of every recurrence, symptom severity, or quality of life unless those are separately defined and analysed as endpoints.
The registry specifies 12 lead ECG documentation and Holter-documented atrial tachyarrhythmias lasting more than 30 seconds. Differences in monitoring intensity or ascertainment can affect how frequently an otherwise similar biological phenomenon is detected.
The initial 3 months are explicitly excluded. Therefore, the primary endpoint is not a measure of every post-procedure arrhythmia occurring from the day of the index procedure onward.
12. Limitations
- No posted efficacy results: the ClinicalTrials.gov record does not currently report recurrence proportions, treatment-effect estimates, confidence intervals, or p-values for the primary endpoint.
- Planned enrollment is not an observed sample: the registry lists enrollment of 360, but the current record does not provide an analysed participant count for the primary endpoint.
- Binary endpoint compression: reducing recurrence to presence or absence simplifies analysis but does not capture the complete temporal pattern of arrhythmia recurrence.
- Monitoring-dependent ascertainment: recurrence must be documented according to specified ECG or Holter criteria. Detection depends on the observation and monitoring process.
- Blanking-period definition: events during the initial 3 months are excluded from the primary endpoint, so the endpoint does not represent all arrhythmias occurring after the index procedure.
- No masking: the registry classifies the trial as having no masking. Knowledge of treatment assignment can potentially affect aspects of treatment delivery, follow-up, or outcome assessment, depending on how assessments are conducted.
- No posted analysis specification: the registry currently provides the endpoint definition but does not post a statistical-analysis result for the endpoint.
- Generalizability: interpretation beyond the enrolled study population depends on the eligibility criteria and clinical characteristics of the participants, which are not detailed in the registry information presented here.
13. Why This Trial Matters Statistically
COMPASS AF is a useful example of a randomized procedural trial in which the primary outcome is not a survival time but a binary recurrence status defined over a specific post-procedure window.
| Statistical concept | How it appears in COMPASS AF |
|---|---|
| Randomization | Participants are allocated randomly between two procedural strategies. |
| Parallel-group design | The registry identifies a two-arm parallel design. |
| Binary endpoint | Recurrence is recorded as presence or absence. |
| Risk difference | Can express the absolute difference in recurrence probability between treatment groups. |
| Risk ratio | Can express the relative recurrence probability between treatment groups. |
| Odds ratio | Can compare recurrence odds between treatment groups. |
| Confidence interval | Provides a measure of statistical uncertainty around an estimated treatment effect. |
| Hypothesis testing | A formal test can assess a specified null hypothesis concerning the binary endpoint. |
| Endpoint definition | Only qualifying recurrences from 3 months through 12 months after the index procedure are included. |
| Outcome ascertainment | Recurrence requires specified ECG or Holter documentation. |
14. Statistical Interpretation of the Trial Design
The strongest feature of a randomized comparison is that treatment assignment is determined before the outcome is observed. If randomization is successfully implemented and follow-up is appropriately conducted, differences in outcome between the groups can be interpreted as differences associated with the assigned treatment strategies rather than simple differences between self-selected treatment groups.
The binary structure of the endpoint also makes the absolute treatment effect particularly transparent. Suppose a future analysis found recurrence in 25% of one randomized group and 20% of the other. The absolute difference would be 5 percentage points, while the relative risk would be 1.25 if the first group were the numerator. Those are mathematically related descriptions of the same hypothetical data, but they communicate different aspects of the treatment contrast.
For COMPASS AF, the endpoint definition is especially important because the trial excludes the initial 3-month blanking period. A treatment comparison based on recurrence from the day of the procedure would therefore not be directly equivalent to the registered primary endpoint.
Statistical significance versus clinical magnitude
A sufficiently large randomized study can detect relatively small differences, while a smaller study can fail to demonstrate statistical evidence for a meaningful difference because its confidence interval is wide. Consequently, interpretation should focus on the estimated recurrence proportions, the absolute and relative treatment effects, and their confidence intervals rather than on a p-value alone.
Why the endpoint should be reported with both groups
A relative measure such as a risk ratio is difficult to interpret without the underlying event probabilities. Reporting both recurrence proportions allows readers to see the absolute frequency of recurrence as well as the relative contrast between the two strategies.
15. Results Status
Primary endpoint results
The ClinicalTrials.gov record contains no posted statistical analyses for recurrence of atrial tachyarrhythmia.
The registry currently provides the endpoint definition and its assessment window, but no treatment-group recurrence estimates, confidence intervals, or p-values.
The absence of posted statistical analyses means that the current registry record supports interpretation of the trial design and endpoint methodology, but not an observed treatment-effect conclusion.
16. Sources
- ClinicalTrials.gov: COMPASS AF, NCT07726901.
Continue through Clinical Biostats
Explore additional clinical trial analyses, statistical tutorials, and statistical calculators covering the methods used across clinical research.
17. Record Summary
COMPASS AF is a randomized, parallel-group trial of 360 planned participants comparing anatomy-based ablation with Emphasize map-guided ablation in atrial fibrillation. Its primary endpoint is recurrence of atrial tachyarrhythmia from 3 months to 12 months after the index procedure, with recurrence defined through specified 12 lead ECG and Holter monitoring criteria and recorded as a binary outcome.
From a statistical perspective, the endpoint naturally supports comparison of recurrence proportions between the two randomized groups. Absolute risk difference, risk ratio, and odds ratio provide complementary measures of treatment effect, while confidence intervals quantify uncertainty around those estimates. The initial 3-month blanking period is an essential part of the endpoint definition and prevents the primary analysis from being interpreted as a measure of all post-procedure arrhythmias.
The ClinicalTrials.gov record currently identifies the study as recruiting, with a start date of May 19, 2026 and a primary completion date of December 31, 2030. No statistical results have been posted for the primary endpoint, so the current evidence available in the registry concerns the trial's design and planned outcome measurement rather than an observed comparative treatment effect.