Evaluation Of New Onset Postoperative Atrial Fibrillation: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Evaluation Of New Onset Postoperative Atrial Fibrillation (NCT01247974), a randomized na study. The primary outcome is Primary Safety Endpoint.

2. Clinical Question

The trial focused on new-onset postoperative atrial fibrillation, but the posted primary endpoint is a safety composite rather than an atrial-fibrillation-only endpoint. Its components were death, myocardial infarction, stroke, mediastinal reoperation, and percutaneous coronary intervention. The endpoint description specifies procedure-related serious adverse events occurring within 30 days postprocedure.

The registry summary frames the comparison as an evaluation of pericardial closure by circumferentially reconstructing normal pericardial anatomy following isolated, first-time CABG. The control group did not undergo pericardial closure. The central statistical question is whether the observed proportion with at least one qualifying composite event differed between the groups, and how precisely that difference was estimated. The posted results also include a risk difference, relative risk, odds ratio, and an unadjusted p-value. These statistics describe the group comparison; they do not by themselves establish a causal mechanism or address outcomes outside the specified endpoint and population.

3. Trial Design

The registry describes a prospective, randomized, controlled, parallel-group trial. Its enrollment is listed as 439, and its status is completed. The listed phase is NA. The study was planned as a multicenter trial at up to 20 clinical sites in the U.S., according to the registry summary. The registry lists Elutia Inc. as sponsor.

Participants were assigned to pericardial closure with CorMatrix ECM or to control with no pericardial closure. The registered allocation is randomized and the masking is listed as single. The identity of the people who were masked is not reported in the registry record, so the masking label should not be given a more specific interpretation. The model is parallel, meaning the registered comparison uses separate assigned groups rather than a crossover arrangement.

The intervention is categorized as a device. The brief summary describes the intended setting as isolated, first-time CABG procedures and the intervention as circumferential reconstruction of normal pericardial anatomy following surgery. The comparison is against subjects who do not undergo pericardial closure.

The posted group summaries list 217.0 participants for CorMatrix ECM for Pericardial Closure and 222.0 for Control. The primary outcome timeframe is 30 days postprocedure. Its analysis population is described as the FAS population, with missing or incomplete values imputed using multiple imputation. The record specifies that FAS means subjects in the ITT population who had CABG surgery and received their assigned treatment. The registry record does not report the original sample-size justification or the detailed rationale for the selected design and analysis choices.

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.

Before assignment: One subject, Subject 011-006, was randomized in error to the ECM treatment group, but did not provide informed consent. This deviation was discovered prior to the surgery.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Performed with the FAS population using multiple imputation methods to impute values for missing or incomplete data. FAS population was defined as subjects in the ITT population who had CABG surgery and received their assigned treatment.

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 Primary Safety Endpoint, measured over 30 days postprocedure.

The primary safety endpoint is the percentage of participants with a composite of procedure-related serious adverse events occurring within 30 days postprocedure. The component events are death, myocardial infarction, stroke, mediastinal reoperation, and percutaneous coronary intervention. A composite combines multiple clinically relevant events into a single outcome for the registered primary comparison. A participant meeting any component definition can contribute to the composite; the record does not give a component-by-component breakdown for this primary endpoint.

The percentage-of-participants measure describes the proportion of the specified analysis population with a qualifying event, rather than the number of events accumulated over person-time. That distinction matters: the endpoint is participant-based and binary at the level of whether the composite occurred. The registered analysis population is the FAS population, defined as subjects in the ITT population who had CABG surgery and received their assigned treatment. Missing or incomplete endpoint values in that population were handled using multiple imputation.

A composite can capture a broader range of safety outcomes than a single event alone, but its interpretation depends on which events are combined and how often each occurs. Because the posted primary result does not separate its components, it does not show which component contributed most to the group percentages. The registry record does not report a separate primary endpoint result for new-onset atrial fibrillation.

7. Results Posted in the Registry

• CorMatrix ECM for Pericardial Closure: 3.7% (8/217) versus 3.6% (8/222) on control; risk difference 0.1 percentage points (95% CI -3.4 to 3.6), risk ratio 1.02 (0.39–2.68); unadjusted two-proportion z-test p = 0.963.

8. Statistical Interpretation

9. Statistical Methods Explained

The registry presents a binary participant-level endpoint as a percentage and reports a group comparison using a risk difference, relative risk, odds ratio, confidence intervals, and an unadjusted p-value. Each statistic frames the comparison differently. Reading them together helps distinguish the estimated direction, the scale of the estimate, and the uncertainty around it.

What does the risk difference describe?

A risk difference subtracts the event proportion in one group from the event proportion in the comparison group, following the specified group ordering. It is an absolute measure: it expresses how far apart the proportions are on the probability scale. The posted estimate is 0.001, with a confidence interval from -0.034 to 0.036. A value of zero corresponds to no difference in proportions. Since this interval spans zero, the data are compatible with no difference as well as with differences in either direction within the interval. The interval does not say that every value inside it is equally plausible, and its endpoints should not be mistaken for guaranteed limits on the true effect.

How do relative risk and odds ratio differ?

The relative risk compares the event proportion in one group with that in the other as a ratio. A relative risk at its no-association value indicates equal proportions; values on either side express relative increases or decreases. The registry reports 1.023, with a confidence interval from 0.391 to 2.677. The odds ratio compares event odds rather than event proportions. It is another relative measure, with its own interpretation and scale. The reported odds ratio is 1.024, with a confidence interval from 0.377 to 2.779. Both intervals include their respective no-association value and are wide, indicating substantial uncertainty in these relative estimates. Relative measures can sound large or small without conveying the absolute event proportions, so they are best read alongside the percentages and risk difference.

What does the p-value tell the reader?

The unadjusted p-value is 0.963. In general, a p-value quantifies how unusual data at least as discrepant as those observed would be under a specified null model and its assumptions. It does not give the probability that the null hypothesis is true, and it is not a measure of the size or clinical importance of an effect. An unadjusted result is not adjusted for other comparisons or covariates. The registry record does not report a multiplicity-adjusted result for this comparison.

What does multiple imputation contribute?

The primary endpoint description says that multiple imputation was used for missing or incomplete values in the FAS population. In general, multiple imputation represents uncertainty about missing values by creating multiple plausible completed datasets, analyzing each, and combining the estimates so that missing-data uncertainty is reflected in the inference. It is distinct from simply treating every missing value as an event or as no event. The record does not report the imputation model, variables included, assumptions about the missingness process, or sensitivity analyses, so the robustness of the result to alternative missing-data assumptions cannot be assessed from the registry record.

Why can a composite endpoint be useful, and what does it conceal?

In general, combining related outcomes can produce a single endpoint that captures a broader safety experience and can make an overall comparison more informative than focusing on a rare component alone. But components may differ in severity, frequency, or clinical meaning. A combined percentage can therefore conceal which events drove the result. The registry lists the composite components, but does not give separate primary-endpoint estimates for death, myocardial infarction, stroke, mediastinal reoperation, or PCI. The posted composite percentage should not be interpreted as if all component events were interchangeable.

10. Confidence Intervals and Precision

The risk-difference estimate is 0.001, with a confidence interval from -0.034 to 0.036. A confidence interval expresses sampling uncertainty around an estimate under the analysis method and its assumptions. In repeated use of a method under appropriate conditions, intervals constructed in this way would cover the underlying parameter at the method's stated confidence level at the corresponding frequency. The registry record does not state the confidence level alongside the posted interval, so it should not be inferred from the record.

The interval crosses zero, the value representing no risk difference. This means the posted data do not precisely determine whether the underlying difference favors either group or is near zero. Its breadth also means that the result is compatible with a range of possible differences, not only the point estimate. The point estimate is not itself a measure of precision: precision is conveyed by the uncertainty interval.

The relative risk interval, 0.391 to 2.677 around an estimate of 1.023, and the odds-ratio interval, 0.377 to 2.779 around an estimate of 1.024, likewise include their no-association values. Their breadth shows that the relative comparisons are imprecise. Relative measures can have wide intervals even when the absolute event percentages appear close, particularly when the event is uncommon. That is a general statistical consideration, not a separate finding about the cause of uncertainty in this trial.

The posted percentages are 3.9% and 3.6%, and the comparison details list 8 events in each group. These observed summaries should be read together with the interval estimates: similar point estimates do not remove uncertainty, and an interval that includes no difference does not prove the absence of a potentially meaningful difference. The record does not report a prespecified threshold for clinical importance, so the estimates cannot be classified against such a threshold from the registry information.

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 sample-size figures are planning illustrations based on the observed effect, not the trial's original design calculation. They assume a two-sided significance level of 0.05, equal allocation between the groups, and a normal approximation. Each row uses the stated fraction of the observed effect as the effect to detect. The original trial's sample-size assumptions and justification are not reported in the registry record.

The entries are participants per arm for the 80% power and 90% power targets; they are not total event counts. The 80% power column gives the planned number of participants in each group under the stated assumptions for a design that would have an 80% chance of detecting the specified effect if that effect were the true one. The 90% power column gives the corresponding per-arm planning number for a 90% chance. Power is a property of a planned study under assumed conditions, not a probability that the observed result is correct.

The rows describe planning for 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect. Smaller effects require more participants to distinguish them from random variation under otherwise unchanged assumptions. The figures are very large because the calculation uses the small observed effect as its planning basis. They should not be read as enrollment targets adopted by the trial, nor as a claim that the completed trial had a particular probability of success. This is not observed power.

The curve should be interpreted using the same planning context: it relates the assumed effect size to the number of participants required under the stated assumptions. It is an illustration of design sensitivity, not an additional outcome from the trial. If the curve and the row entries are considered together, the main point is how quickly required sample sizes rise when planning for effects smaller than the observed estimate.

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 = 0.036; treatment proportion p₁ = 0.037; average p̄ = 0.036.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.2650 = 0.5195.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.2650 = 0.2231.
4. n = (0.5195 + 0.2231)² / (0.0008)² = 799,701.6, rounded up to 799,702 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 <- 0.036036036036036                          # control proportion
p1 <- 0.0368663594470046                          # 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(1247974)
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 = 799,702 (80% power), 1,070,574 (90% power)
 75% of observed effect: n per arm = 1,417,798 (80% power), 1,898,028 (90% power)
 50% of observed effect: n per arm = 3,181,279 (80% power), 4,258,830 (90% power)
power.prop.test: 799701.6 
Simulated power at n = 799,702 per arm: 0.797 (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 primary endpoint description identifies the FAS population and says that multiple imputation was used for missing or incomplete data. It defines the FAS as subjects in the ITT population who had CABG surgery and received their assigned treatment. The registry therefore describes an analysis population that is restricted by surgery and receipt of assigned treatment; it is not described as including every randomized participant regardless of what happened after assignment.

In general, missing data can affect an estimate if the participants with missing values differ systematically from those whose outcomes are observed. Multiple imputation is a way to account for missing values while reflecting uncertainty, but its validity depends on assumptions and modeling decisions. The record does not report the imputation model, the amount or pattern of missing endpoint data, the assumptions used, or sensitivity analyses. It is therefore not possible from the registry record to evaluate how the primary result might change under alternative missing-data assumptions.

The posted comparison includes an unadjusted p-value of 0.963. The registry lists 2 outcome measures and 0 analyses, and no multiplicity-adjusted comparison is posted. The record does not report a testing hierarchy, adjustment procedure, or interim analysis plan. In general, when many outcomes or comparisons are tested, the chance of at least one apparently unusual result can increase; multiplicity methods address that issue by changing how evidence is evaluated. This is general context, not a claim that a particular adjustment was required or performed in this trial.

Beyond the FAS definition and multiple-imputation statement, the registry record does not report a detailed statistical analysis plan. It does not state how the imputation was combined with the group comparison, specify covariate adjustment, or describe sensitivity analyses. These omissions limit the ability to reproduce the analysis or assess its robustness. They should not be filled in by assuming that a conventional approach was used.

14. Safety Results

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

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 1% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: atrial fibrillation (5.8%), respiratory failure (5.3%), angina pectoris (4.8%), event led to death (4.5%), pleural effusion (3.8%).

The adverse-event summaries list serious adverse events affecting 76 of 208 at risk in the CorMatrix ECM group and 54 of 220 at risk in Control. For other adverse events, the summaries list 190 of 208 at risk in the CorMatrix ECM group and 189 of 220 at risk in Control. These are adverse-event summaries and should not be substituted for the primary composite endpoint, which has its own population, definition, and 30-day timeframe.

The registry lists the leading serious adverse-event terms as Atrial fibrillation, 5.8%; Respiratory failure, 5.3%; Angina pectoris, 4.8%; Event led to Death, 4.5%; Pleural effusion, 3.8%; and Cardiac failure congestive, 3.4%. These terms are reported as a list, without group-specific values in the record. In particular, the term “Event led to Death” is not a reported death count for either group. The death-count fields are not reported in the registry record.

The adverse-event timeframe and further event-level definitions are not reported in the registry record. The listed counts and percentages should therefore be understood as the registry's safety summaries, not as a direct estimate of the primary endpoint or as proof that an event was caused by the assigned intervention.

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

The posted result is an estimate of the difference between groups in a defined composite safety outcome among the stated analysis population. The percentages are close, the reported risk difference is 0.001, and the unadjusted p-value is 0.963. The risk-difference confidence interval includes zero and spans both directions. In statistical terms, the posted result does not provide clear evidence of a group difference in this endpoint, while the interval also prevents a confident conclusion that the true difference is exactly zero.

This distinction matters. Failure to demonstrate a difference is not the same as demonstrating equivalence or non-inferiority. The registry record does not report an equivalence or non-inferiority objective or margin. The interval is compatible with a range of effects, and the record does not identify a threshold at which a difference would be considered clinically important. The estimate should therefore be interpreted with its uncertainty, rather than reduced to a binary conclusion based only on the p-value.

The outcome is a composite of death, myocardial infarction, stroke, mediastinal reoperation, and PCI within 30 days postprocedure. The posted overall composite does not establish the result for each component separately. Nor does it provide the primary endpoint result for new-onset atrial fibrillation alone, despite that condition being central to the trial summary and title.

The comparison is specific to the registered population, intervention, control condition, endpoint, and follow-up timeframe. It does not answer questions about other procedures, other populations, longer-term outcomes, or outcomes not included in the composite. The statistical results describe observed data and uncertainty; they are not medical advice and do not establish whether a treatment should be used.

16. Why This Trial Matters Statistically

This trial illustrates why an endpoint's name and the endpoint actually analyzed both matter. The study focuses on new-onset postoperative atrial fibrillation, but the posted primary endpoint is a safety composite that includes death, myocardial infarction, stroke, mediastinal reoperation, and PCI. Readers must distinguish the motivating clinical context from the specific outcome behind the reported primary comparison.

It also demonstrates the value of reading absolute and relative measures together. The group percentages and risk difference convey the absolute comparison, while relative risk and odds ratio express the comparison on ratio scales. The confidence intervals show that the point estimates alone do not convey the substantial uncertainty in the comparison. A large p-value cannot replace that assessment or establish equivalence.

The population definition and missing-data approach are central to interpretation. The analysis used an FAS population defined by inclusion in the ITT population, having CABG surgery, and receiving assigned treatment, with missing or incomplete values handled using multiple imputation. Because the record does not provide the imputation model, missingness details, or sensitivity analyses, readers cannot fully assess the dependence of the result on those analytical choices.

Finally, the planning illustration shows the consequences of using a small observed effect as the basis for future sample-size calculations. It emphasizes that detecting smaller effects generally requires more participants under the same assumptions. Such calculations help frame precision and design questions, but do not replace the trial's original design rationale, which is not reported in the registry record.

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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The registry record leaves important details unavailable for evaluating design and analysis. It does not report the original sample-size justification or the assumptions used to design the completed trial. The planning figures associated with the observed effect are illustrations and cannot be treated as the trial's planned enrollment rationale.

The masking label is single, but who was masked is not reported in the registry record. The record also does not describe allocation concealment or other implementation details. These omissions mean that the masking and randomization labels alone do not provide a complete account of how assignment and outcome assessment were conducted.

The primary endpoint is a composite, and separate primary results for its components are not posted. This prevents assessment of whether the overall result was driven by a particular event type. The record does not post a separate primary result for new-onset atrial fibrillation. Although safety summaries list serious adverse-event terms, they are not a substitute for component-specific estimates of the primary composite.

For missing data, the record states that multiple imputation was used but does not report the imputation model, missingness pattern, or sensitivity analyses. It also does not provide a detailed analysis plan or multiplicity procedure. The adverse-event summary does not state a timeframe in the posted details, and deaths are not reported as group counts. These limitations constrain interpretation and reproducibility; they do not, by themselves, determine the direction of bias or the validity of the result.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Evaluation Of New Onset Postoperative Atrial Fibrillation — NCT01247974.
• Registry record last updated 2023-07-27; status last verified 2023-07.
• Results first posted 2015-09-16.
• 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 describes a completed, prospective, randomized, controlled, parallel-group study of CorMatrix ECM pericardial closure versus no pericardial closure following isolated, first-time CABG procedures. Enrollment is listed as 439, and the posted group summaries list 217.0 participants in the CorMatrix ECM arm and 222.0 in Control.

The primary safety endpoint is the percentage of participants with a composite procedure-related serious adverse event within 30 days postprocedure. The posted percentages are 3.9% and 3.6%; the risk difference is 0.001, with a confidence interval from -0.034 to 0.036 and an unadjusted p-value of 0.963. The registry states that missing or incomplete primary endpoint values were handled with multiple imputation in the FAS population.

The result does not show a clear group separation, but the confidence interval leaves uncertainty about the size and direction of the difference. The registry record does not report the original sample-size rationale, detailed imputation methods, or component-specific primary endpoint results.