Watchman Device Reports to FDA: What the MAUDE Data Reveals About Boston Scientific's Leading LAA Closure Device

The FDA's MAUDE database holds 10,998 adverse-event reports tied to Boston Scientific's Watchman device components since 2019. Of those, 1,210 are coded as deaths and 8,890 as injuries. These are reports, not rates. A report coded as a death does not establish that the device caused the death, and MAUDE carries no denominator for how many Watchman implants surgeons performed in a given year.

What the data does show is acceleration. Reports climbed from 996 in 2019 to 2,026 in 2024, and death-coded reports rose to 265 that year. The 2025 record already holds 2,086 reports and 268 deaths.

Our pipeline flagged the original Watchman with a critical signal for an increasing reporting trend, and the disproportionality detector fired on the original device at an EBGM of 126.0. The newer Watchman FLX Pro line carries multiple high-priority disproportionality flags ranging from EBGM 57.2 to 109.8. Those numbers measure relative excess reporting against the background of all FDA-regulated devices, not absolute danger, a distinction we argued in Signal Detection in Post-Market Surveillance Is Broken.

The signal reflects two forces at once: a device with deep market penetration as the first FDA-approved left atrial appendage closure device, and a set of recognized complications. You can run the same query on the public MAUDE lookup tool.

#Reports by year

Year	Reports	Deaths
2019	996	76
2020	935	67
2021	1,142	105
2022	1,263	120
2023	1,684	174
2024	2,026	265
2025	2,086	268
2026	866	135

#What the Watchman Device Is and Why It Matters

The original Watchman received FDA premarket approval in March 2015, the first left atrial appendage closure device approved in the United States to reduce stroke risk in patients with non-valvular atrial fibrillation. The appendage is a common origin point for the clots that drive AFib-related stroke, and sealing it offers an alternative to lifelong anticoagulation.

Boston Scientific followed with Watchman FLX, approved on July 21, 2020, which the company described as the first LAAC device that can be fully recaptured, repositioned, and redeployed for precise placement. The Watchman FLX Pro arrived in September 2023 with a polymer coating, new visualization markers, and a broader size matrix that includes a 40 mm option, all aimed at reducing device-related thrombus.

In July 2025, the FDA expanded labeling for FLX and FLX Pro to include patients following catheter ablation for non-valvular AFib, widening the addressable population. The full reporting history sits on the device detail page.

#Reports by Component: Which Watchman Models Drive the Signal

The original Watchman dominates by volume with 3,817 reports, a function of its decade on the market. The Watchman Access System follows at 1,037 reports. The FLX and FLX Pro generations spread across multiple size and variant entries.

The Access System's prominence reflects its role in delivery rather than the implant itself, a point underscored by the July 2025 Class I recall for air embolism risk and the 17 deaths and 120 serious injuries Boston Scientific later linked to it.

#Reports by component

Component	Reports
WATCHMAN®	3,817
WATCHMAN® Access System	1,037
WATCHMAN FLX™	811
WATCHMAN FLX™ Pro	688
WATCHMAN FLX™	661
WATCHMAN FLX™ Pro	543
WATCHMAN FLX™	515
WATCHMAN FLX™ Pro	510
WATCHMAN FLX™ Pro	362
WATCHMAN FLX™	288

#Failure-Mode Signature: Thrombosis, Pericardial Effusion, and Device Mechanics

The patient-problem coding clusters around the complications LAAC literature has documented for years. Pericardial effusion leads with 2,255 reports, thrombosis and thrombus together account for 2,094, and stroke or CVA appears in 1,168. Cardiac tamponade follows at 764 and cardiac perforation at 657. Pericardial effusion requiring intervention is a known and measurable hazard of these procedures, as Munir et al. (2021) found across 17,700 procedures in the United States.

The product-problem side captures the mechanical failures. Migration appears in 676 reports, device dislodgement or dislocation in 609, material deformation in 391, and failure to seal in 267. These point to a mix of procedural and design-related vulnerability.

The narrative clusters sharpen the picture. The largest is "Post-Implant Thrombosis with Unknown Delivery System" at 99 reports, followed by "Pericardial Effusion with Tamponade Post-Implant" at 98 and "Death from Acute Myocardial Infarction Post-Procedure" at 77. Both leading clusters describe life-threatening complications.

One number deserves a hard caveat. 8,050 reports, the large majority, carry the code "Adverse Event Without Identified Device or Use Problem." That code does not mean nothing happened. It reflects how hard it is to isolate device causation in a post-procedural patient whose presentation could stem from anatomy, technique, or underlying disease. We walk through reading these fields in How to Search the FDA MAUDE Database. The comparative work by Price et al. (2022) on first-generation Watchman versus FLX shows these complication profiles are category-level realities, not artifacts of one model.

#Most reported product problems

Product problem	Reports
Adverse Event Without Identified Device or Use Problem	8,050
Difficult to Open or Close	703
Migration	676
Device Dislodged or Dislocated	609
Material Deformation	391
Failure to Seal	267
Material Integrity Problem	261
Difficult to Remove	208
Detachment of Device or Device Component	147
Leak/Splash	42
Gas/Air Leak	42
Use of Device Problem	33

#Most reported patient problems

Patient problem	Reports
No Clinical Signs, Symptoms or Conditions	2,813
Pericardial Effusion	2,255
Thrombosis/Thrombus	2,094
Stroke/CVA	1,168
Cardiac Tamponade	764
Low Blood Pressure/ Hypotension	724
Cardiac Perforation	657
Thrombosis	429
Transient Ischemic Attack	349
No Consequences Or Impact To Patient	343

#FDA Recalls and Recent Regulatory Actions

Recalls and adverse-event reports are separate FDA systems, and the Watchman record carries three recalls on its product codes. All three are terminated. In May 2018, a Class I recall covered a 32 mm device whose delivery sheath tip could fold or buckle during recapture. Two Class II recalls in August 2015 addressed cross-threading of the hemostasis valve.

The serious current action sits with the delivery hardware. In July 2025, Boston Scientific issued a Class I recall for Watchman Access Systems, covering the FXD Curve, Truseal, and Trusteer models, after identifying an elevated risk of air embolism. The FDA classified it as its most serious recall type, warning the device may cause serious injury or death without the updated instructions.

The hazard is usage-dependent. At the time of its July 30, 2025 customer letter, Boston Scientific linked the issue to 120 serious injuries and 17 deaths, with risk roughly three times higher under conscious or deep sedation than under positive pressure-controlled ventilation.

#FDA recalls

Recall	Class	Date	Status	Reason
Z-2878-2018	Class I	2018-05-07	Terminated	Delivery sheath tip may fold or buckle when recapturing the 32-mm device
Z-2765-2015	Class II	2015-08-25	Terminated	Cross-threading of the hemostasis valve may prevent subsequent sealing
Z-2766-2015	Class II	2015-08-25	Terminated	Cross-threading of the hemostasis valve may prevent subsequent sealing

#Head-to-Head Evidence: Watchman FLX vs. Oral Anticoagulation and Competitors

MAUDE volume reads very differently against the controlled trial record. The OPTION trial, published in the New England Journal of Medicine in November 2024, enrolled 1,600 post-ablation AFib patients across 114 sites. Watchman FLX was non-inferior to oral anticoagulation for death, stroke, and embolism at 36 months (5.4% versus 5.8%) and superior for bleeding reduction (8.5% versus 18.1%).

The pivotal PINNACLE FLX trial, a single-arm study of 400 patients at 29 U.S. sites, met its primary safety endpoint with a 0.5% adverse-event rate at 7 days or discharge, below the prespecified 4.21% performance goal. It reported 100% effective LAA closure at one year and a 3.4% stroke or systemic embolism rate at two years. Long-term comparison against direct oral anticoagulants continues in the ongoing CHAMPION-AF trial.

Real-world practice tells a consistent story. The SURPASS analysis of the NCDR LAAO Registry, covering more than 97,000 patients and published by Kapadia et al. (2024), found a 0.45% major procedural adverse event rate within 7 days and 98% implant success.

Thrombosis and effusion are not unique to Watchman. The dual-center comparison by Galea et al. (2022) and the sealing-performance study of the Amplatzer Amulet by Korsholm et al. (2022) place these hazards at the category level. A 97,000-patient denominator is far more informative about risk than raw MAUDE counts.

#What MAUDE Data Can and Cannot Tell You About Watchman Safety

MAUDE captures event volume, not event probability. The figure of 10,998 reports is not 10,998 confirmed device failures. The database mixes device defects, procedural errors, patient anatomy, and natural disease progression into a single stream, a structural limit we describe in Post-Market Surveillance in the United States.

The missing denominator is the central problem. Without the total number of Watchman implants performed since 2019, raw counts cannot become rates. The SURPASS registry's 97,000-patient denominator does work that MAUDE volume cannot.

Disproportionality metrics carry their own caveat. The EBGM scores that flagged the original Watchman at 126.0 and FLX Pro between 57.2 and 109.8 measure relative excess reporting against a background, not absolute risk. A high EBGM for a newer device can reflect intense post-launch surveillance as much as elevated danger, which is why we separate flags from conclusions in Signal Detection in Post-Market Surveillance Is Broken.

Death coding compounds the uncertainty. The 1,210 death-coded reports span acute coronary syndrome, stroke, and tamponade, and not all are attributable to device malfunction. Confirming causation requires narrative review, one report at a time.

Temporal trends move for reasons that have nothing to do with device performance. Market share growth, FDA actions, and litigation publicity all drive reporting. The 2025 total of 2,086 reports overlaps with the July Access System recall, and the recall's publicity is a plausible contributor to the volume. Reading a spike as a pure safety signal misses how the system behaves.

#For Patients and Clinicians: Practical Guidance and Risk Mitigation

Watchman FLX and FLX Pro carry trial and registry support through PINNACLE, OPTION, and SURPASS, and remain a standard option for suitable candidates, especially patients at high bleeding risk on anticoagulation. No device achieves total safety. Thrombosis, pericardial effusion, and stroke remain recognized risks even in expert hands, though the SURPASS registry's 0.45% major adverse event rate is low and consistent across LAAC devices.

Proceduralists should follow the updated July 2025 instructions for Watchman Access Systems. Positive pressure-controlled ventilation is now the recommended approach, given the roughly threefold higher air embolism risk under conscious or deep sedation. Patients implanted before this change, particularly under sedation without strict pressure-controlled protocols, may want to discuss follow-up imaging with their cardiologist.

Litigation underscores that procedural execution matters. A federal judge declined to dismiss a wrongful death suit against Boston Scientific in 2021, allowing manufacturing and assembly defect claims to proceed. A separate malpractice case involving negligent device deployment settled for over $1,050,000. Selecting an experienced LAAC program is a concrete way to reduce the procedural component of risk.

#Methodology: How These Numbers Were Derived and What They Miss

Report counts come from the FDA's MAUDE database, a passive surveillance system. Manufacturers and healthcare facilities face mandatory reporting requirements, while voluntary consumer reports introduce variable quality and timeliness. The Claripulse query aggregated reports tied to Watchman product codes from 2019 forward.

Event-type coding for Death and Injury is assigned by the reporter, not independently verified by the FDA. A report coded Death may reflect correlation rather than causation. Component and product-problem tables aggregate by device variant and coded description. Multiple reports can describe one event, and a single report can cite several problems. Failure-mode clusters were extracted from narrative fields through thematic coding and represent dominant patterns rather than an exhaustive list.

The EBGM signals flag disproportionate reporting relative to the background across all FDA-regulated devices in the same period. The 2026 year-to-date figure of 866 reports and 135 deaths is incomplete and should not be annualized.

Trial and registry rates from PINNACLE, OPTION, and SURPASS support stronger causal inference but come from smaller, selected populations. MAUDE is broader and messier, closer to the full diversity of real-world practice, and that breadth is both its value and its limit. Our full reading guide lives in How to Search the FDA MAUDE Database.

This analysis is based on publicly available FDA MAUDE data and published literature. It does not constitute medical advice. MAUDE reports are submitted under FDA regulation, are de-identified, and may be incomplete or unverified. No attempt should be made to re-identify any patient.