Risk of Stroke and Silent Cerebrovascular Thromboembolism After Cardioversion of Atrial Fibrillation: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Risk of Stroke and Silent Cerebrovascular Thromboembolism After Cardioversion of Atrial Fibrillation (NCT01924065), a randomized phase 3 study. The primary outcome is Combined Clinical Thromboembolic or Hemorrhagic Event, Death or Silent Cerebral Thromboemboli.

2. Clinical Question

3. Trial Design

This was a Phase 3, randomized, double-masked, parallel-group trial with enrollment of 112 participants. The registry lists 56 participants in the Transesophageal Echocardiography Group and 56 in the Oral Anticoagulant Group. Suleyman Demirel University is listed as the sponsor. The record gives a start date of 2013-08, primary completion of 2017-08, and status of completed.

The two groups followed different pre-cardioversion strategies. Participants assigned to transesophageal echocardiography-guided care could proceed to electrical cardioversion if no thrombus was detected in the left atrium or left atrial appendage. Participants assigned to the oral anticoagulant strategy took warfarin for 3 weeks before cardioversion, or an approved new oral anticoagulant agent if they did not want warfarin. The registry states that post-procedure oral anticoagulation was used in both groups for at least 4 Weeks.

Neurological examination and diffusion magnetic resonance imaging were planned at baseline and at the postprocedural 7th day. The primary outcome description combines clinical events during 2 years of follow-up with silent cerebral thromboemboli detected by imaging at 1 week post-cardioversion. The record labels the outcome as a count of participants with an event.

The registry identifies the design as double-masked but does not report who was masked or how masking was maintained. Details of sequence generation, allocation concealment, site number, participant flow, and adherence are not reported in the registry record. The registry lists cardioversion as a procedure and oral anticoagulant as a drug intervention.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Combined Clinical Thromboembolic or Hemorrhagic Event, Death or Silent Cerebral Thromboemboli, measured over 2 years.

The primary endpoint is the combined count of participants with a clinical thromboembolic or hemorrhagic event, death, or silent cerebral thromboemboli. The registry describes clinical events during 2 years of follow-up and silent cerebral thromboemboli detected by diffusion magnetic resonance imaging at 1 week post-cardioversion. It is recorded as a participant count rather than a count of every event.

A composite endpoint can capture several kinds of outcome within a single measure. As a general statistical consideration, this can provide a broader summary when the trial is concerned with more than one event type. The trade-off is that the components may differ in clinical meaning, severity, and how they are detected. A combined result does not reveal which component contributes most unless component-specific results are also reported.

The registry describes neurological examination and imaging at baseline and after cardioversion, supporting assessment of both clinical events and silent imaging findings. The primary result is not a stroke-only measure: it also includes hemorrhagic events, death, and silent cerebral thromboemboli. Component-specific counts for this primary composite are not reported in the registry record, so the overall group counts cannot identify which types of event account for the observed totals.

7. Results Posted in the Registry

• Oral Anticoagulant Group: 19.6% (11/56) versus 16.1% (9/56) on control; risk difference 3.6 percentage points (95% CI -10.6 to 17.7), risk ratio 1.22 (0.55–2.72); unadjusted two-proportion z-test p = 0.622.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is binary at the participant level: each participant either contributes to the event count or does not. The registry presents counts of participants with an event and the number in each group. For this type of outcome, a group risk is the event count relative to the group size. The comparison reports a risk difference, a risk ratio, and an odds ratio. These statistics describe the same group comparison using different scales.

What does the risk difference describe?

The risk difference expresses the contrast in event risk between the Oral Anticoagulant Group and the Transesophageal Echocardiography Group. The reported estimate is 0.036, with a confidence interval from -0.106 to 0.177. A positive estimate points toward a greater event risk in the oral anticoagulant group under this comparison’s direction. The interval expresses uncertainty around that estimate and spans both directions of effect. No conversion to a percentage-point scale is needed to interpret the reported estimate.

As a general statistical explanation, an absolute difference is often useful because it keeps the comparison on the outcome’s original risk scale. It can be easier to relate to observed group event proportions than a relative measure. Its size, however, depends on the risks in both groups and on the outcome definition. Here the endpoint combines several kinds of events, so the difference applies to that combined outcome, not to each component separately.

Why are relative measures also reported?

The risk ratio compares the group risks as a ratio, and the odds ratio compares the corresponding odds. The reported risk ratio is 1.222, with a confidence interval from 0.550 to 2.718. The reported odds ratio is 1.277, with a confidence interval from 0.483 to 3.371. Both point estimates indicate a higher observed risk or odds in the oral anticoagulant group, but their intervals are broad and include relative effects in either direction. Risk ratios and odds ratios have different interpretations; they should not be treated as interchangeable quantities.

As a general concept, relative measures can convey proportional contrasts, while absolute differences convey the change on the risk scale. Neither scale alone describes the clinical importance of the result. Interpretation also depends on the outcome’s meaning, the precision of the estimate, and the clinical context. That context and a component-level account of the primary composite are not reported in the registry record.

How should the p-value be read?

The registry lists an unadjusted p-value of 0.622. A p-value is a measure of how unusual data at least as discrepant as those observed would be under a specified null model and the test assumptions. It is not the probability that the null hypothesis is true, and it does not quantify the chance that the finding is due to chance. The exact test, its assumptions, and any covariate adjustment are not reported in the registry record. The label unadjusted indicates that the listed p-value is not reported as adjusted for other comparisons.

As a general statistical principle, a p-value does not replace the estimate and its confidence interval. The interval gives information about the range of effects compatible with the data under its method, whereas the p-value compresses a test result into a single measure. Here, the interval for the risk difference includes no difference and also allows effects in either direction. The result therefore does not establish equivalence or rule out a meaningful difference.

What can be inferred from the analysis record?

The record lists 7 outcome measures and 0 posted analyses. It includes a primary comparison with group counts and unadjusted summary measures, but it does not identify the specific statistical procedures used to calculate them. The analysis population, handling of missing outcomes, covariate adjustment, and treatment of participants who did not undergo cardioversion are not reported in the registry record. There is also no posted breakdown of primary endpoint components. These omissions limit how far the statistical result can be audited or extended beyond the reported comparison.

Random assignment is a design feature, not a substitute for knowing how outcomes were analyzed. General statistical interpretation distinguishes the effect of assignment to a strategy from the effect of actually receiving every part of that strategy. The registry does not report whether the analysis followed an intention-to-treat approach, a per-protocol approach, or another population definition. Consequently, the posted estimates are best read as the registry’s reported comparison, without assuming an unreported analysis framework.

10. Confidence Intervals and Precision

The risk-difference estimate is 0.036, and its confidence interval extends from -0.106 to 0.177. The interval is wide enough to include a higher event risk in either group, as well as no difference. Thus, the point estimate alone gives an incomplete picture: it has a direction, but the interval shows that the comparison is not estimated precisely.

The relative estimates convey similar uncertainty. The risk ratio is 1.222, with a confidence interval from 0.550 to 2.718. The odds ratio is 1.277, with a confidence interval from 0.483 to 3.371. Each interval includes the no-association value for its scale, while also allowing effects in either direction. The interval bounds should not be read as a list of equally likely effects or as proof that every value within them is plausible in a clinical sense.

As a general explanation, a confidence interval reflects sampling uncertainty under the statistical method used. Its width depends on factors such as the amount of information in the data and the variability of the outcome. The registry record does not report the interval construction method or confidence level, so these should not be inferred. The group counts are 56 in each arm, with 9 and 11 primary endpoint participants, respectively; those observed counts form the basis of the reported comparison.

Precision is distinct from whether an estimate is statistically significant. The broad intervals and unadjusted p-value of 0.622 leave substantial uncertainty about the size and direction of the difference. They do not demonstrate equivalence, and they do not identify which component of the composite accounts for the observed event counts.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning display gives participants per arm, not total events. Its figures are illustrations based on fractions of the observed effect; they are not the trial’s original design targets. The original sample-size assumptions and justification are not reported in the registry record.

The listed planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. For each stated fraction of the observed effect, the display gives a separate participant requirement for 80% power and for 90% power. In plain language, the 80% column describes the planned number of participants per arm for a design with an 80% chance of detecting the specified effect under those assumptions, if that effect is real. The 90% column describes the corresponding target for a design with a 90% chance of detection.

The effect rows use 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect. The display’s requirements increase as the assumed effect becomes smaller: detecting a smaller difference generally requires more participants, all else held constant. The plotted planning curve represents that relationship between the effect used for planning and the participant requirement; it is not an observed outcome curve.

These numbers are planning illustrations, not observed power for the completed trial. They do not show the probability that the trial’s result is true, and they do not change the uncertainty in the posted outcome comparison. They are calculated from the observed effect rather than from the trial’s original design assumptions, which are not reported in the registry record.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.161; treatment proportion p₁ = 0.196; average p̄ = 0.179.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.5416 = 1.0616.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5410 = 0.4554.
4. n = (1.0616 + 0.4554)² / (0.0357)² = 1,804.1, rounded up to 1,805 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.160714285714286                          # control proportion
p1 <- 0.196428571428571                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(1924065)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 1,805 (80% power), 2,415 (90% power)
 75% of observed effect: n per arm = 3,145 (80% power), 4,210 (90% power)
 50% of observed effect: n per arm = 6,934 (80% power), 9,282 (90% power)
power.prop.test: 1804.1 
Simulated power at n = 1,805 per arm: 0.806 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists 7 outcome measures but no posted secondary outcome results. It also lists 0 posted analyses. The primary comparison is reported, but there is no registry analysis record describing a broader testing strategy or how the outcome measures were prioritized beyond the outcome labelled primary.

As a general statistical issue, testing several outcomes or several contrasts can increase the chance of at least one apparently notable result even when there is no underlying effect. Methods for handling this include prespecifying a hierarchy or adjusting inference across tests. Whether such a strategy was used in this trial is not reported in the registry record. The listed p-value is unadjusted; no multiplicity adjustment is described.

Information about missing outcome data is not reported in the registry record. The record does not give the number of participants with completed imaging, neurological examination, or clinical follow-up, nor does it explain how unavailable measurements were handled. It also does not report whether any participants withdrew or were excluded from the primary analysis.

The analysis population is not reported in the registry record. There is no stated intention-to-treat, per-protocol, or safety population definition, and no account of participants who were assigned to a strategy but did not proceed to cardioversion. The registry’s randomized allocation label is useful design information, but it does not by itself establish which participants entered each analysis.

These gaps matter because missingness, exclusions, and analysis choices can affect the interpretation of a randomized comparison. Without the analysis population and outcome-completion information, the relation between the randomized groups and the reported event counts cannot be fully assessed. The absence of posted secondary analyses also means that no secondary outcome comparison can be described from the registry record.

14. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: ischemic stroke-transient ischemic attack (12.5%), gastrointestinal bleeding (5.4%), hemorrhagic stroke (3.6%).

The registry’s serious adverse event figures list 56 participants at risk in each group. Serious adverse events affected 9 participants in the Transesophageal Echocardiography Group and 10 in the Oral Anticoagulation Group. Other adverse events affected 0 participants in each group. Deaths are listed as 3 in the transesophageal echocardiography group and 2 in the oral anticoagulation group.

The registry also identifies Ischemic stroke-transient ischemic attack at 12.5%, Gastrointestinal bleeding at 5.4%, and Hemorrhagic stroke at 3.6% among the top serious terms. These percentages are reported as terms in the safety record; the record does not provide a group-specific breakdown for those listed percentages. They should not be treated as additional group-specific counts.

These safety figures describe reported adverse-event information, not a causal comparison of harms. The registry does not report timing, duration, relatedness, severity details beyond the serious designation, or whether event terms overlap within participants. The death counts are listed separately, and the record does not state their causes or relation to trial procedures or assigned strategies.

15. What the Result Does, and Does Not, Mean

The posted primary comparison concerns a composite endpoint among participants assigned to two cardioversion strategies. The Transesophageal Echocardiography Group had 9 endpoint participants among 56, while the Oral Anticoagulant Group had 11 among 56. The estimated risk difference is 0.036, with a confidence interval from -0.106 to 0.177, and the unadjusted p-value is 0.622. These results show an uncertain difference in the combined endpoint, rather than a precise estimate favoring either strategy.

The endpoint combines clinical thromboembolic or hemorrhagic events, death, and silent cerebral thromboemboli detected by imaging. The group counts therefore cannot be interpreted as stroke counts alone. Since a component-specific breakdown is not reported in the registry record, the result does not reveal whether the observed totals were driven by clinical events, deaths, silent imaging findings, or another combination of components.

The comparison also does not answer whether cardioversion itself causes events compared with no cardioversion. The trial compares pre-cardioversion approaches among people considered for cardioversion, with cardioversion withheld if a thrombus was detected under the transesophageal echocardiography strategy. Both groups are described as receiving oral anticoagulation after the procedure.

As a general interpretation, a non-significant p-value does not prove the strategies are equivalent, just as a directional point estimate does not establish a reliable difference. The confidence intervals allow a range of effects, including effects in either direction. The registry record does not report enough analysis detail to assess missing outcomes, analysis population, or the method used for the test.

The reported data therefore support a cautious description: the observed group counts are close, and the estimated difference is imprecise. They do not establish that either strategy improves or worsens outcomes, and they should not be used to make treatment decisions without broader clinical evidence and context.

16. Why This Trial Matters Statistically

This trial illustrates a statistical challenge in comparative clinical research: a randomized design can produce a direct group comparison, yet the estimate may remain uncertain when event counts are limited. Enrollment was 112, with 56 participants in each group. The primary comparison recorded 9 and 11 participants with the composite endpoint. The confidence interval for the risk difference spans effects in both directions, emphasizing why event counts and the point estimate should be read together with precision.

The composite outcome adds another interpretive challenge. It combines clinical thromboembolic or hemorrhagic events, death, and silent cerebral thromboemboli detected by imaging. Such a composite can represent a broad safety and cerebrovascular question, but it does not tell readers whether the components have similar importance or contribute similarly. Component-specific primary results are not reported in the registry record.

The planning illustrations also demonstrate the distinction between an observed effect and a design assumption. Their assumptions use fractions of the observed effect and specify equal allocation, a two-sided significance level of 0.05, and a normal approximation. Those illustrations are not the trial’s original planning record, which is not reported. Their role is to show how a target effect and chosen power relate to participant requirements, not to retroactively certify the completed study.

For biostatistical interpretation, the useful lesson is to keep the estimand, outcome definition, effect scale, uncertainty, and analysis details connected. The risk difference, risk ratio, and odds ratio are all summaries of the same comparison, while the registry leaves important analysis and missing-data details unavailable. The result is most informative when treated as a bounded, uncertain estimate rather than a binary declaration.

17. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record leaves several trial-specific details unreported. It does not describe eligibility criteria beyond the atrial fibrillation and cardioversion context, the method used to generate the random allocation sequence, or allocation concealment. Although the design is labelled double-masked, who was masked and how masking was maintained are not reported.

Participant flow is not reported. The record does not show how many participants completed planned imaging, neurological examinations, or clinical follow-up, or whether any participants were lost to follow-up. Missing-data handling, analysis population, adherence to assigned strategies, and treatment of participants who did not undergo cardioversion are also not reported. These details are important for understanding how the randomized assignment relates to the posted comparison.

The primary endpoint combines clinical thromboembolic or hemorrhagic events, death, and silent cerebral thromboemboli. The registry does not provide primary component-specific counts or a description of how each clinical event was adjudicated. The outcome description includes clinical events during 2 years of follow-up and imaging findings at 1 week post-cardioversion, so the composite combines observations from different assessment windows.

The original sample-size justification, intended effect, and planned power are not reported in the registry record. Interim analyses and stopping rules are not reported either. Statistical test methods, confidence-interval construction, covariate adjustment, and any multiplicity plan are also not described. These omissions limit assessment of the design and analysis beyond the posted group counts and comparison measures.

The registry lists safety counts but does not report their timing, relatedness, or group-specific results for the listed serious-event terms. These limits should be kept in view when interpreting the comparison; they do not establish that unreported procedures or analyses did or did not occur.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Risk of Stroke and Silent Cerebrovascular Thromboembolism After Cardioversion of Atrial Fibrillation — NCT01924065.
• Registry record last updated 2024-11-15; status last verified 2024-10.
• Results first posted 2024-10-08.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

21. Record Summary

The registry identifies NCT01924065 as Risk of Stroke and Silent Cerebrovascular Thromboembolism After Cardioversion of Atrial Fibrillation. It is listed as a completed Phase 3, randomized, double-masked, parallel-group trial sponsored by Suleyman Demirel University. Enrollment was 112, with 56 participants in each group. The record gives a start date of 2013-08 and primary completion of 2017-08.

The trial compared transesophageal echocardiography-guided cardioversion with a strategy using oral anticoagulation before cardioversion. The primary outcome was a composite count of participants with clinical thromboembolic or hemorrhagic events, death, or silent cerebral thromboemboli. The registry reports 9 primary endpoint participants in the transesophageal echocardiography group and 11 in the oral anticoagulant group. The risk-difference estimate is 0.036, with a confidence interval from -0.106 to 0.177 and an unadjusted p-value of 0.622. The registry lists 7 outcome measures and 0 posted analyses.