Colchicine For The Prevention Of Perioperative Atrial Fibrillation In Patients Undergoing Thoracic Surgery (COP-AF): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Colchicine For The Prevention Of Perioperative Atrial Fibrillation In Patients Undergoing Thoracic Surgery (COP-AF) (NCT03310125), a randomized phase 3 study. The primary outcome is Number of Participants With Clinically Important Perioperative Atrial Fibrillation.

2. Clinical Question

This is a comparison of assigned treatment groups over a specified postoperative observation period. The primary outcome is not every detected episode of atrial fibrillation: it is limited to events meeting the registry definition of clinical importance. The result therefore addresses that defined event, rather than all rhythm abnormalities or all possible consequences of surgery.

3. Trial Design

The registry describes COP-AF as a phase 3, randomized, quadruple-masked, parallel-group trial. Its stated enrollment is 3209. The two listed groups are Active Colchicine and Active Placebo. The registry record gives a start date of 2018-02-14, a primary completion date of 2023-06-01, and a status of completed.

The active regimen was oral colchicine 0·5 mg twice daily, beginning within 4 h before surgery and continuing for 10 days. The placebo regimen was also oral, described as 0·5 mg twice daily over the same stated timing. The record identifies patients undergoing major thoracic surgery as the intended population. More detailed eligibility criteria, such as age limits or specific procedure types, are not reported in the registry record.

Random allocation is intended to make the treatment groups comparable on average, while masking is intended to limit the influence of treatment knowledge on care, outcome assessment, and other trial processes. The registry calls the masking quadruple, but does not identify which roles were masked. The parallel-group structure means participants were assigned to one of the two groups rather than described as crossing between treatments.

The primary outcome was the number of participants with clinically important perioperative atrial fibrillation, assessed over 14 days after randomisation. The registry also lists secondary outcomes concerning composites of death, myocardial injury, myocardial infarction, and stroke, as well as hospital and intensive-care duration measures. The record lists 14 outcome measures and 14 analyses, but the detailed analysis methods and the intended hierarchy across outcomes are not reported in the registry record.

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 Number of Participants With Clinically Important Perioperative Atrial Fibrillation, measured over 14 days after randomisation.

The primary endpoint counts participants with clinically important perioperative atrial fibrillation during 14 days after randomisation. The registry defines this as atrial fibrillation or atrial flutter that results in angina, heart failure, or symptomatic hypotension, or that requires treatment with a rate-controlling drug, an anti-arrhythmic drug, or electrical cardioversion. This definition focuses on events with specified clinical consequences or treatment requirements, rather than counting every detected rhythm episode.

Using a participant count makes the endpoint a binary outcome at the individual level: a participant either meets the definition during the stated timeframe or does not. The event counts are 103 in the Active Colchicine group and 120 in the Active Placebo group. The posted group risks are 0.064 and 0.075, respectively. These descriptive values show the observed frequency in each assigned group, but do not by themselves establish whether the difference is distinguishable from chance variation.

As a clinical trial endpoint, the measure connects the comparison to a postoperative complication that the registry identifies as an important target of prevention. The treatment question is limited to this operational definition and timeframe. It does not directly measure all atrial fibrillation, mortality, stroke, symptoms outside the definition, or longer-term outcomes. The reasons for choosing this particular endpoint and the endpoint-adjudication procedures are not reported in the registry record.

7. Results Posted in the Registry

• Active Colchicine: 6.4% (103/1608) versus 7.5% (120/1601) on control; risk difference -1.1 percentage points (95% CI -2.8 to 0.7), risk ratio 0.85 (0.66–1.10); unadjusted two-proportion z-test p = 0.225.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Number of Participants With Composite of All-cause Death, Non-fatal MINS, and Non-fatal Stroke

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Active Colchicine: 18.7% (300/1608) versus 20.9% (335/1601) on control; risk difference -2.3 percentage points (95% CI -5.0 to 0.5), risk ratio 0.89 (0.78–1.03); unadjusted two-proportion z-test p = 0.107.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Active Colchicine and the control. The interval gives the range of effects that remain compatible with the data.

Number of Participants With Composite of All-cause Death, Non-fatal Myocardial Infarction, and Non-fatal Stroke

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Active Colchicine: 1.3% (21/1608) versus 1.9% (31/1601) on control; risk difference -0.6 percentage points (95% CI -1.5 to 0.2), risk ratio 0.67 (0.39–1.17); unadjusted two-proportion z-test p = 0.157.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Active Colchicine and the control. The interval gives the range of effects that remain compatible with the data.

Number of Participants With MINS Not Fulfilling the Fourth Universal Definition of Myocardial Infarction

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Active Colchicine: 17.7% (285/1608) versus 19.4% (311/1601) on control; risk difference -1.7 percentage points (95% CI -4.4 to 1.0), risk ratio 0.91 (0.79–1.05); unadjusted two-proportion z-test p = 0.215.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Active Colchicine and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes cover several related but distinct postoperative events, including composites involving death, myocardial injury, myocardial infarction, and stroke, as well as measures of hospital and critical-care duration. The posted estimates for the three secondary binary outcomes with detailed results generally point toward fewer events in the colchicine group, but their uncertainty intervals include the relevant no-difference value. The registry lists additional secondary outcomes, but detailed results for all listed outcomes are not included in the available result details.

This broad pattern does not overturn or independently confirm the primary result. Secondary outcomes answer different questions, may have different event frequencies, and can have different precision. A finding for a composite also cannot identify which component accounts for the observed contrast unless component results are separately available. The primary endpoint remains the prespecified central question only if the trial's protocol made it so; the registry identifies it as primary, while the full hierarchy and analysis plan are not reported in the registry record.

Unadjusted secondary p-values carry less weight when many outcomes are examined because the chance of at least one apparently unusual result increases as more tests are considered. In general, multiplicity procedures can address this issue, but the registry record does not report whether any such procedure was used. The absence of a reported adjustment does not establish that none was performed; it means the record does not document it. Secondary findings are therefore best treated as supportive or exploratory unless the full analysis plan establishes a confirmatory role.

The secondary outcomes add context about other events around surgery, but they should not be combined informally into a single overall treatment conclusion. They also do not supply evidence about outcomes or populations that were not measured or reported. Their interpretation depends on their definitions, follow-up, missing data, and analysis populations, several of which are not described in the registry record.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Number of Participants With Composite of All-cause Death, Non-fatal MINS, and Non-fatal Stroke
• Number of Participants With Composite of All-cause Death, Non-fatal Myocardial Infarction, and Non-fatal Stroke
• Number of Participants With MINS Not Fulfilling the Fourth Universal Definition of Myocardial Infarction
• Number of Participants With Myocardial Infarction
• Time to Chest Tube Removal, Days
• Days in Hospital
• Nights in Step-down Unit
• Nights in Intensive Care Unit

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is binary because it records whether each participant met the clinically important atrial fibrillation definition during the stated period. A straightforward summary is the risk in each group: the proportion of participants who experienced the event. The posted group risks are 0.064 for Active Colchicine and 0.075 for Active Placebo. These summaries retain the group-specific event frequencies, but do not by themselves account for the timing of events or any analysis model.

What does a risk difference represent?

The risk difference subtracts the comparator-group risk from the intervention-group risk. The unadjusted recalculation is -0.011, with a 95% confidence interval from -0.029 to 0.007. A negative value indicates a lower observed risk in the colchicine group under this coding. A value of zero would mean equal risks in the two groups. Because the interval includes zero, the data are compatible with no risk difference as well as with differences in either direction. This recalculation is a simple comparison of the posted summary counts, not the registry-reported time-to-event model.

What does the hazard ratio describe?

The registry reports a regression, Cox analysis with a hazard ratio of 0.85 and a 95% confidence interval from 0.65 to 1.1. In general, a hazard ratio compares the instantaneous event rates between groups over follow-up, conditional on the model structure. It uses event timing, whereas a risk difference compares the proportions with an event by the end of the stated period. These measures answer related but distinct questions, so their numeric values should not be expected to match. The registry record does not provide further detail about the model specification, covariates, or treatment of censoring.

Why are confidence intervals useful?

In general, a confidence interval expresses the uncertainty associated with an estimate under the method used. Its width conveys precision: narrower intervals indicate more precise estimates, while wider intervals indicate greater uncertainty. For a risk difference, zero is the no-difference value; for a hazard ratio, one is the no-difference value. The primary interval includes zero for the recalculated risk difference and includes one for the reported hazard ratio. That pattern is consistent with uncertainty about whether the groups differ, rather than proof that the groups are identical.

What does the p-value add?

The unadjusted recalculation reports p = 0.225; the registry reports p = 0.22 for its Cox analysis. In general, a p-value measures the compatibility of data with a specified null model, assuming the analysis model and its conditions. It does not measure the size or clinical importance of an effect, and it is not the probability that the null hypothesis is true. Different analyses can yield slightly different p-values because they use different estimands or model structures. The registry record does not report a full statistical analysis plan that would explain every analytic choice.

Why do raw counts and model estimates differ in emphasis?

Counts and risks are transparent descriptions of what was observed in each group. A model-based hazard ratio summarizes the relative event experience over time. The simple recalculation is unadjusted and based on posted summary statistics; it should not be read as a replacement for the sponsor-reported analysis. Since the discrepancy flag is false, the record does not indicate a conflict requiring one result to supersede another. The sponsor analysis remains the registry's own reported analysis, while the recalculation offers a complementary, less model-dependent view.

These distinctions matter because statistical interpretation depends on the estimand: the exact quantity being compared. Here, a fixed-period participant risk difference and a time-to-event hazard ratio are not interchangeable. Neither should be interpreted without its uncertainty interval and the outcome definition. The registry does not report whether the primary analysis used an intention-to-treat population, how missing outcome data were handled, whether there were interim analyses, or how secondary outcomes were incorporated into an overall testing strategy.

General statistical considerations also include the assumptions underlying each method. A Cox regression analysis generally relies on assumptions about the relationship of hazards over time and how censoring operates. The available registry entries do not establish whether those assumptions were checked or what procedures were used if they were not met. For the binary risk comparison, the posted summary statistics allow an unadjusted contrast, but do not account for covariates or event timing. Such limitations affect what each estimate can say, not the arithmetic of the posted counts.

11. Confidence Intervals and Precision

The primary risk-difference estimate is -0.011, with a 95% confidence interval from -0.029 to 0.007. The interval spans zero, the value corresponding to equal group risks. It therefore does not isolate a single direction of effect: values below zero and above zero remain compatible with the interval under the calculation method. The estimate itself is a point summary, not a guarantee that the true difference is exactly that value.

The registry's time-to-event analysis reports a hazard ratio of 0.85, with a 95% confidence interval from 0.65 to 1.1. For a hazard ratio, one is the no-difference value. This interval includes one. The two intervals describe different measures and are based on different analyses; they should not be treated as competing versions of the same quantity. Both indicate uncertainty around the estimated comparison.

In general, a 95% confidence interval is a range generated by a procedure that, under its assumptions, would cover the target parameter in repeated use at the stated rate. It is not a range containing a known proportion of individual participant outcomes, nor does it assign a probability to the parameter after the data are observed. Precision is reflected in the interval's breadth, while interpretation also depends on which values are clinically important. The registry record does not report a threshold for a clinically important difference, so the interval cannot be classified against such a threshold using the available information.

The p-values, 0.225 for the unadjusted recalculation and 0.22 for the sponsor-reported analysis, are consistent with the intervals crossing their respective no-difference values. They should not replace the effect estimates and intervals. The observed event counts provide important context, but precision is best judged from the interval for the comparison rather than from whether the group counts look different.

12. 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 illustrations translate an assumed effect into a planning quantity, rather than describing the size of this completed trial. For this binary primary outcome, the entries are participants per arm, not total participants and not total events. The table presents planning for 80% power and 90% power at three fractions of the observed effect: 100%, 75%, and 50%. At each assumed effect, the 90% power column requires more participants per arm than the 80% power column.

The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is the stated fraction of the observed effect. These assumptions describe the illustration only. The trial's own original sample-size assumptions and justification are not reported in the registry record.

For example, planning for 100% of the observed effect corresponds to 8,547 participants per arm for 80% power and 11,441 participants per arm for 90% power. Planning for 75% of the observed effect corresponds to 15,470 participants per arm for 80% power and 20,709 participants per arm for 90% power. Planning for 50% of the observed effect corresponds to 35,425 participants per arm for 80% power and 47,424 participants per arm for 90% power. The pattern illustrates the general principle that smaller effects are harder to distinguish from random variation and require more participants.

In general, power is the probability that a planned test will reject its null hypothesis if a particular effect is real and the planning assumptions hold. The 80% and 90% columns describe different target probabilities for a future design, not the probability that this trial's result is correct. These calculations are not observed power; they are planning illustrations based on the observed effect and stated assumptions. They should not be interpreted as evidence that the completed trial was designed around these figures or assumptions.

13. 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 = 0.075; treatment proportion p₁ = 0.064; average p̄ = 0.070.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.3596 = 0.7049.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.3596 = 0.3026.
4. n = (0.7049 + 0.3026)² / (-0.0109)² = 8,546.2, rounded up to 8,547 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 <- 0.0749531542785759                          # control proportion
p1 <- 0.0640547263681592                          # 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(3310125)
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 = 8,547 (80% power), 11,441 (90% power)
 75% of observed effect: n per arm = 15,470 (80% power), 20,709 (90% power)
 50% of observed effect: n per arm = 35,425 (80% power), 47,424 (90% power)
power.prop.test: 8546.2 
Simulated power at n = 8,547 per arm: 0.798 (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.

14. Analyses Reported by the Sponsor

The registry-reported primary analysis is a regression, Cox analysis with a hazard ratio of 0.85, a 95% confidence interval from 0.65 to 1.1, and p = 0.22. It is a time-to-event analysis. The recalculated comparison uses the posted group counts to produce an unadjusted risk difference of -0.011, a 95% confidence interval from -0.029 to 0.007, and p = 0.225. That calculation compares the observed risks by the end of the stated timeframe rather than summarizing relative event rates over time.

Because these are different measures and methods, their point estimates are not expected to be numerically identical. Both estimates indicate a comparison with uncertainty that includes the no-difference value for the measure used. The registry discrepancy flag is false, so the record does not identify a discrepancy between the recalculation and sponsor result. The sponsor's reported analysis is the trial's own analysis; the unadjusted recalculation is a supplementary way to inspect the posted summary counts, not a replacement for that analysis.

The registry identifies the sponsor method as Regression, Cox, but does not report further model details in the record. In particular, the available information does not specify covariates or other implementation choices. The comparison should therefore retain the distinction between the sponsor-reported model result and the simpler recalculation, without assuming details that are not recorded.

15. Multiplicity, Missing Data and Analysis Populations

The registry record lists 14 outcome measures and 14 analyses, including a primary endpoint and multiple secondary outcomes. Considering multiple outcomes raises a multiplicity question: when many statistical tests are conducted, the chance of observing at least one result that appears notable by chance can increase. In general, a prespecified testing hierarchy or adjustment procedure can help control error rates or clarify which analyses are confirmatory. Whether this trial used such a procedure, and how it was implemented, is not reported in the registry record.

The record also does not report whether the secondary outcomes were prespecified in a hierarchy, whether all analyses were planned before results were known, or whether any interim analyses took place. These omissions do not show that such procedures were absent; they limit what can be established from the registry record. The reported secondary p-values should consequently be read with caution, particularly when considered across a set of outcomes.

Missing-data methods are not reported in the registry record. It is not stated how missing endpoint information, loss to follow-up, or censoring was handled, whether imputation was used, or whether sensitivity analyses assessed the impact of missingness. The record also does not identify the analysis population for the reported primary result, such as whether it followed an intention-to-treat definition or another population. These details matter because exclusions and missing outcomes can affect which participants contribute to an estimate.

Other analytic details are also not reported, including the full statistical analysis plan, model covariates, and procedures for checking model assumptions. The registry identifies a regression, Cox method for the primary analysis, but that label alone does not resolve these questions. The safest reading is to distinguish what the record explicitly reports—the outcome definition, group counts, effect estimates, intervals, and p-values—from the analysis decisions that remain undocumented.

16. 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 3% reporting threshold.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: pulmonary air leakage (0.7%), subcutaneous emphysema (0.4%), pneumothorax (0.4%), respiratory failure (0.4%), acute respiratory failure (0.2%).

The registry reports serious adverse events affecting 67 participants in the Active Colchicine group and 66 in the Active Placebo group. It reports deaths of 7 and 11, respectively. The other-adverse-event field is recorded as 0 for each group. The record's serious-event categories include pulmonary air leakage at 0.7%, subcutaneous emphysema at 0.4%, pneumothorax at 0.4%, respiratory failure at 0.4%, acute respiratory failure at 0.2%, and chylothorax at 0.2%.

These figures are descriptive safety results from the registry record. They do not, on their own, establish whether an event was caused by the assigned intervention or whether the groups differ reliably in safety. The record does not provide detailed event definitions, attribution assessments, severity grading, exposure-adjusted rates, or the timing of each event. It also does not explain the other-adverse-event field beyond the recorded values. Interpretation should therefore remain limited to the posted counts and event terms, rather than treating them as a complete account of all possible harms.

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

The primary result describes the frequency of a narrowly defined, clinically important atrial fibrillation or flutter outcome during 14 days after randomisation. The observed counts were 103 in the colchicine group and 120 in the placebo group. The unadjusted risk-difference estimate is -0.011, and the sponsor-reported hazard ratio is 0.85. Both comparisons have uncertainty intervals that include their respective no-difference values. The data therefore do not provide a precise estimate establishing the direction of the underlying difference.

This does not mean the groups are proven identical. A confidence interval that includes no difference can still be compatible with effects in either direction, and the primary interval includes a range of possible contrasts. Nor does the point estimate alone demonstrate that colchicine prevented the outcome. Interpretation requires the magnitude and uncertainty of the estimate, the endpoint definition, and the analysis method together.

The result applies to the recorded trial comparison: the stated oral regimen versus placebo among patients undergoing major thoracic surgery, for the outcome and timeframe specified. It does not establish what would happen with other regimens, other durations, other surgical settings, or populations not described in the registry. The record does not report longer-term primary-outcome results, detailed eligibility criteria, or a clinical-importance threshold for the risk difference.

The result also should not be generalized to every form of atrial fibrillation, all postoperative complications, or every secondary outcome. The primary endpoint excludes episodes that do not meet the registry's clinical consequence or treatment criteria. Secondary outcomes are separate measures, and their interpretation is limited by uncertainty and by the unreported multiplicity strategy. The most defensible conclusion is descriptive: the observed primary event frequency was lower in the colchicine group, while the reported analyses leave substantial uncertainty about the size and direction of the group difference.

18. Why This Trial Matters Statistically

This trial poses a clinically focused statistical question: whether assignment to an anti-inflammatory drug regimen changes the risk of a defined postoperative rhythm complication in patients undergoing major thoracic surgery. Its randomized, masked, parallel-group design is relevant because it aims to support a comparison between assigned interventions while limiting some forms of bias. The endpoint's definition is also consequential: it counts atrial fibrillation or flutter only when specified clinical consequences occur or treatment is required.

Statistically, the results illustrate why effect estimates should be read alongside their uncertainty. The observed event counts and the negative risk-difference estimate suggest a direction in the sample, but the confidence interval spans the no-difference value. The sponsor's time-to-event analysis likewise reports an interval including its no-difference value. Neither a directionally favorable point estimate nor a p-value alone captures this uncertainty.

The trial also illustrates how different valid summaries can address different estimands. The risk difference summarizes the contrast in observed event risks over the stated period, while a Cox-model hazard ratio summarizes relative event rates over time. Keeping those measures distinct helps prevent overinterpretation. Finally, the range of secondary outcomes makes multiplicity and analysis transparency important: the registry record lists multiple outcomes, but does not document a testing hierarchy or adjustment method. These features make careful reporting of estimates, intervals, and analysis limitations central to understanding the evidence.

19. 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 does not report detailed eligibility criteria or a fuller description of the enrolled participants. That limits assessment of how closely the study population represents particular clinical settings or procedure types. The record identifies major thoracic surgery as the target context, but does not specify the range of operations or relevant baseline characteristics.

Further details about outcome assessment are not reported, including the process for detecting and adjudicating atrial fibrillation or flutter, the use of monitoring, and whether assessors were masked. The registry describes quadruple masking but does not identify the masked roles. Such details would help clarify how consistently the primary endpoint was ascertained.

The statistical analysis plan is not reported in detail. The registry names a regression, Cox method but does not state covariates, model diagnostics, censoring procedures, the primary analysis population, or missing-data handling. It also does not report the trial's original sample-size justification, any interim analyses, or a multiplicity strategy. These gaps limit assessment of whether all analytic decisions were prespecified and how robust the findings are to alternative assumptions.

Finally, the posted primary outcome concerns 14 days after randomisation, and the record's secondary outcomes cover selected perioperative events and care-duration measures. Longer-term outcomes and detailed results for all listed secondary outcomes are not reported in the registry record for this analysis. Safety figures are limited to the recorded serious-event and death counts and selected event terms, without detailed attribution or timing. These limitations do not invalidate the reported estimates, but define the scope of what can be concluded from the registry record.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Colchicine For The Prevention Of Perioperative Atrial Fibrillation In Patients Undergoing Thoracic Surgery (COP-AF) — NCT03310125.
• Registry record last updated 2026-10-16; status last verified 2025-06.
• Results first posted 2026-10-16.
• 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.

23. Record Summary

COP-AF is listed as a completed, phase 3 randomized, quadruple-masked, parallel-group trial with enrollment of 3209. It compared oral colchicine 0·5 mg twice daily, starting within 4 h before surgery and continuing for 10 days, with an oral placebo regimen described on the same schedule in patients undergoing major thoracic surgery.

The primary outcome was clinically important perioperative atrial fibrillation during 14 days after randomisation. The unadjusted risk-difference estimate was -0.011, with a 95% confidence interval from -0.029 to 0.007. The registry-reported Cox analysis gave a hazard ratio of 0.85, with a 95% confidence interval from 0.65 to 1.1. Both analyses leave uncertainty around the group difference. Important details about missing data, analysis populations, multiplicity, and the original sample-size plan are not reported in the registry record.