Investigational Device Evaluation of the WATCHMAN FLX™ LAA Closure Technology: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Investigational Device Evaluation of the WATCHMAN FLX™ LAA Closure Technology (NCT02702271), a non-randomized na study. The primary outcome is Number of Participants With a Safety Event.

2. Clinical Question

3. Trial Design

This was a prospective, non-randomized, multi-center device investigation sponsored by Boston Scientific Corporation. Its stated purpose was to establish the safety and effectiveness of WATCHMAN FLX for people with non-valvular atrial fibrillation who were eligible for long-term anticoagulation and had a rationale to seek a non-pharmacologic alternative. The registry lists the phase as NA and the study status as completed.

The intervention listing names WATCHMAN FLX. The registry distinguishes a WATCHMAN FLX Main Cohort from a WATCHMAN FLX Roll-In Cohort. Allocation is recorded as non-randomized, masking as none, and the model as sequential. These design features do not create a randomized comparison: the two cohorts should not be treated as exchangeable treatment and control groups. The record does not report a control arm or enough detail to establish that differences between cohorts reflect any intervention effect.

Enrollment is listed as 458. The study start date is 2018-05-07, and primary completion is recorded as 2020-01-28. Those dates describe the registry timeline; they do not establish when individual participants were followed or when each analysis was conducted. The registry record does not report further details about cohort entry, site procedures, follow-up beyond the primary endpoint window, or the rationale for the sequential model.

The record lists two outcome measures and no analyses. The primary outcome is a participant count for a defined safety-event composite. A primary estimate, uncertainty interval, hypothesis test, or effectiveness result is not reported in the registry record. Consequently, the study's stated objective can be described, but its primary result cannot be assessed from the information recorded.

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

Reasons for not completing this period:

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 a Safety Event, measured over events occurring between the time of implant and 7 days post procedure or by discharge, up to 30 days, whichever is later.

The primary endpoint counts participants with at least one event in a specified safety composite. Events qualify if they are all-cause death, ischemic stroke, systemic embolism, or a device- or procedure-related event requiring open cardiac surgery or a major endovascular intervention. Examples of the latter include pseudoaneurysm repair, AV fistula repair, or another major endovascular repair.

The window starts at implant and extends to 7 days post procedure or by discharge, up to 30 days, whichever is later. This defines a short-term post-procedure safety question; it is not a measure of all possible outcomes over longer follow-up. The endpoint is expressed as a count of participants, so it concerns whether a participant experienced a qualifying event rather than how many separate events occurred. The registry does not report a primary endpoint count or analysis.

The description also distinguishes events that are captured as adverse events but excluded from this endpoint. These include percutaneous catheter drainage of pericardial effusions, snaring of an embolized device, thrombin injection for femoral pseudoaneurysm, and non-surgical treatment of access-site complications. The distinction matters: an event may be recorded as an adverse event without meeting the primary composite's criteria. The composite is used to focus on a defined set of serious outcomes and major interventions, but it does not represent every possible complication or safety concern.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

A binary participant-count endpoint records whether each participant experienced an event meeting the definition during the specified window. General statistical explanation: this type of outcome can be summarized with the number of participants with an event and the number assessed, and often with a proportion. A proportion describes frequency in the observed group; it does not by itself establish that the intervention caused or prevented an event. In this record, the primary endpoint is described, but its observed count and analysis are not reported in the registry record.

What does a composite endpoint combine?

The primary outcome combines several event types: all-cause death, ischemic stroke, systemic embolism, and selected device- or procedure-related events requiring major intervention. Combining outcomes can make a broad safety question measurable when several clinically important events are relevant. General statistical explanation: a composite can also conceal variation among its components. A total composite count does not reveal which event type contributes most unless component results are separately reported. Here, no component-level results for the primary endpoint are reported in the registry record.

Endpoint definitions also determine what is excluded. The record explicitly excludes some interventions and treatments from the primary endpoint while capturing them as adverse events. That distinction means that a primary composite count and a general adverse-event summary answer different questions. Readers should not assume that every adverse event appears in the primary endpoint, or that the adverse-event listings reproduce the endpoint count.

What can a confidence interval tell a reader?

General statistical explanation: a confidence interval expresses the statistical uncertainty around an estimate under the assumptions of the analysis. A narrower interval indicates greater precision for that estimate; a wider interval indicates less precision. The interval does not correct for selection bias, confounding, incomplete outcome ascertainment, or an unsuitable endpoint definition. No confidence interval for the primary endpoint is reported in the registry record, so its precision cannot be evaluated.

Likewise, a hypothesis test would address compatibility between data and a specified null model, conditional on the analysis assumptions. A p-value is not an estimate of the size or importance of an effect and does not measure the probability that a hypothesis is true. The registry record lists no analyses and gives no primary test result. No claim about statistical significance can therefore be made from this record.

Why does non-random allocation matter?

General statistical explanation: random allocation can balance measured and unmeasured baseline characteristics on average, whereas non-random allocation may leave group differences that affect outcomes. In a non-randomized study, an observed difference between cohorts may reflect those differences as well as any intervention-related effect. Statistical adjustment can address measured factors only to the extent they are recorded and modeled appropriately; it cannot guarantee removal of unmeasured confounding. The registry record does not report an adjusted analysis or the covariates that might be used for one.

The main and roll-in cohorts are not described as randomized arms or as a like-for-like control comparison. Their event counts are therefore best understood as descriptive cohort summaries. A direct comparison would require clarity about cohort definitions, timing, baseline characteristics, outcome ascertainment, and a prespecified analysis; these details are not reported in the registry record.

How should planning illustrations be read?

General statistical explanation: sample-size planning asks how much information a future study might need to detect a target effect under stated assumptions. The registry's planning illustration is described as using the observed effect, a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It considers detecting a stated fraction of the observed effect. These figures are planning illustrations, not observed power and not proof that the completed investigation was designed with those assumptions. The trial's own sample-size justification is not reported in the registry record.

Such calculations depend on the effect size, outcome variability or event frequency, analysis method, and assumptions about allocation and follow-up. General statistical explanation: if the underlying effect estimate is uncertain, a calculation based on it inherits that uncertainty. For clustered, repeated, or time-to-event outcomes, the appropriate model and information measure may differ from a simple binary participant count. The registry notes that time-to-event planning uses total events, while other planning results are expressed as participants per arm; that convention does not turn a planning illustration into an observed finding.

Finally, a complete analysis needs a defined population, outcome handling, and approach to missing observations. Those details, along with any multiplicity strategy, are not reported in the registry record. Without them, readers cannot determine whether the reported safety summaries represent all enrolled participants, a specific analysis population, or a particular missing-data procedure.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("WATCHMAN FLX - M", "WATCHMAN FLX - R")
n      <- c(400, 58)
events <- c(2, 0)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  WATCHMAN FLX - M  400  2  0.5  0.1  1.8  0.1  1.8  0.2  1.8
ARM  WATCHMAN FLX - R  58  0  0.0  0.0  6.2  0.0  6.2  0.0  6.1

11. Multiplicity, Missing Data and Analysis Populations

The registry lists two outcome measures, but the details needed to evaluate the other measure are not reported in the registry record. No analysis is listed. The record does not describe a hierarchy of outcomes, a multiplicity adjustment, or how the primary endpoint would be selected if several analyses were conducted. General statistical explanation: testing several outcomes can increase the chance of at least one apparently unusual result under a null model. A prespecified testing strategy helps readers understand how such results should be interpreted. No such strategy is reported.

Missing-data handling is also not reported in the registry record. The available information does not explain whether all enrolled participants were assessed for the primary endpoint, how incomplete follow-up was treated, or whether any participants were excluded from an analysis. The listed enrollment alone does not establish the analysis denominator for a particular outcome.

Analysis populations are not described in the registry record. It is therefore not possible to determine whether results, if available elsewhere, use all enrolled participants, only participants who received the device, or another defined group. These omissions limit reproducibility and make it difficult to assess potential bias from exclusions or missing outcomes. No assumption about the handling of missing observations or analysis populations should be made from the cohort labels alone.

12. 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

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: gastrointestinal haemorrhage (13.8%), cardiac failure congestive (10.3%), atrial fibrillation (9.0%), pneumonia (6.9%), syncope (6.9%).

The registry's adverse-event summary lists serious adverse events affecting 157 of 400 participants in the WATCHMAN FLX Main Cohort and 25 of 58 in the WATCHMAN FLX Roll-In Cohort. It separately lists other adverse events affecting 96 of 400 and 15 of 58, respectively. Deaths are listed as 36 in the main cohort and 7 in the roll-in cohort. These are descriptive counts; they are not the primary endpoint result.

The record also lists serious-event terms with percentages: gastrointestinal haemorrhage, 13.8%; cardiac failure congestive, 10.3%; atrial fibrillation, 9.0%; pneumonia, 6.9%; syncope, 6.9%; and myocardial infarction, 3.4%. The denominator and calculation basis for these percentages are not reported in the registry record, so they should not be interpreted as percentages of participants in a cohort. The record also does not establish whether individual people contributed to more than one event category. Counts across categories should not be added to infer a total.

Serious adverse events are not interchangeable with the defined primary composite. The adverse-event categories may include events outside that composite, and some events meeting the primary definition may be represented differently in the safety listing. These figures describe what is recorded, not whether the device caused an event or how risks compare with an alternative.

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

The record establishes that a prospective, non-randomized investigation of WATCHMAN FLX was completed and that its primary safety endpoint was defined as a participant count over a specified post-implant period. It also provides cohort-level adverse-event counts and a list of serious-event terms. These registry entries describe the study design and recorded safety summaries; they do not provide a reported estimate for the primary endpoint.

The result does not establish how frequently participants experienced the defined primary composite, how precise any estimate might be, or whether an observed outcome differed from a comparator. The registry record lists no analyses, and it does not report an effectiveness estimate. The study's stated goal to establish safety and effectiveness should not be mistaken for a conclusion that either was established.

Because allocation was non-randomized and the cohorts are not a like-for-like control comparison, the cohort summaries cannot by themselves identify causal effects. Any difference could reflect cohort composition or other factors, and the registry does not report an analysis addressing those possibilities. The adverse-event counts also do not show causation, and the listed percentages lack a reported denominator.

General statistical explanation: evidence from an uncontrolled or non-randomized investigation may contribute descriptive information about outcomes in the studied population, but causal conclusions require an appropriate comparison and assumptions that support it. Precision also matters: an estimate without an uncertainty interval can give an incomplete picture of what the data can support. Neither a primary estimate nor an interval is reported in the registry record.

Accordingly, the defensible interpretation is limited to the registry's design, endpoint definition, and recorded adverse-event summaries. It does not support a claim that the device does or does not work, nor a comparison of its outcomes with another treatment. Further clinical conclusions would require results and methodological details not reported in the registry record.

14. Why This Trial Matters Statistically

This study raises statistical questions that are common in device investigations: how to define a safety outcome, how to interpret a non-randomized cohort structure, and how to distinguish an endpoint from a broader adverse-event listing. The primary composite focuses on death, ischemic stroke, systemic embolism, and selected major device- or procedure-related events. Its explicit exclusions show why endpoint definitions matter: different summaries can count different kinds of events and answer different questions.

The main and roll-in cohorts are useful labels for describing the registry record, but they do not create a randomized comparison. This is an important design distinction for clinical researchers. Cohort-specific counts can describe what was recorded in each group; they cannot, without an appropriate comparison and analysis, separate intervention effects from differences between the people enrolled or other influences.

The record also illustrates the importance of transparent reporting. Although it defines the primary endpoint and lists adverse-event summaries, it lists no analyses and reports no primary endpoint estimate or uncertainty interval. Without those results, readers cannot judge the magnitude or precision of the principal safety finding. The missing details about analysis populations, missing outcomes, and multiplicity further constrain interpretation.

Finally, sample-size planning illustrations should be kept separate from observed evidence. A planning calculation based on an observed effect and stated assumptions is not an observed power result and does not establish that a completed study can answer its primary question. Clear separation of design assumptions, descriptive safety data, and inferential results is essential when assessing a clinical registry record.

15. 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 central limitation for interpretation is that the registry record does not report a primary endpoint result or analysis. It therefore does not permit assessment of the endpoint's observed frequency, uncertainty, or statistical comparison. The record lists two outcome measures, but information about the other measure is not reported in the registry record.

The non-randomized, sequential design and the main and roll-in cohorts do not offer a randomized control comparison. The record does not describe a control group, cohort comparability, baseline characteristics, or adjusted analyses. Any comparison between the cohorts would need to account for the possibility of systematic differences, but whether or how that was done is not reported in the registry record.

The registry does not report analysis populations, missing-data procedures, multiplicity handling, or a sample-size justification for the trial itself. The planning illustration's assumptions do not fill those gaps. The basis for the percentages attached to serious-event terms is also not reported in the registry record, limiting their interpretation as participant-level frequencies.

For safety, the recorded adverse-event summaries cannot be substituted for the primary composite because their definitions differ. The record does not provide enough information to determine overlap between event categories, whether all enrolled participants had complete follow-up, or how the listed deaths relate to the primary endpoint window. No totals across categories should be inferred.

These limitations describe gaps in the registry information, not proof that the underlying study did not collect or analyze the relevant information. The conclusions that can be drawn from the registry record remain limited to its stated design, endpoint definition, and adverse-event summaries.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Investigational Device Evaluation of the WATCHMAN FLX™ LAA Closure Technology — NCT02702271.
• Registry record last updated 2022-06-14; status last verified 2022-05.
• Results first posted 2022-06-14.
• 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.

19. Record Summary

The registry record describes a completed, prospective, non-randomized, multi-center investigation of the WATCHMAN FLX Left Atrial Appendage Closure Device in people with non-valvular atrial fibrillation. It lists a main cohort and a roll-in cohort and states that the study aimed to assess safety and effectiveness for people eligible for long-term anticoagulation who had a rationale to seek a non-pharmacologic alternative.

The primary endpoint is a count of participants with a defined safety event during the period from implant through 7 days post procedure or by discharge, up to 30 days, whichever was later. The registry lists two outcome measures and no analyses. A primary endpoint estimate and its uncertainty are not reported in the registry record.

Serious and other adverse-event counts, deaths, and several serious-event terms are listed. They are descriptive summaries and should not be treated as the primary endpoint or a randomized comparison. Missing-data methods, analysis populations, multiplicity handling, and the trial's own sample-size justification are not reported in the registry record.