Watchman Device Migration and Embolization: A Report from the NCDR LAAO Registry

Abstract

Introduction

Incomplete anchoring of the Watchman left atrial appendage closure (LAAO) device can result in substantial device migration or device embolization (DME) requiring percutaneous or surgical retrieval.

Methods

We performed a retrospective analysis of Watchman procedures (January 2016 through March 2021) reported to the National Cardiovascular Data Registry LAAO Registry. We excluded patients with prior LAAO interventions, no device released, and missing device information. In-hospital events were assessed among all patients and post-discharge events were assessed among patients with 45-day follow-up.

Results

Of 120,278 Watchman procedures, the in-hospital DME rate was 0.07% (n=84) and surgery was commonly performed (n=39). In-hospital mortality rate was 14% among patients with DME and 20.5% among patients who underwent surgery. In-hospital DME was more common: at hospitals with a lower median annual procedure volume (24 vs. 41 procedures, p<0.0001), with Watchman 2.5 versus Watchman FLX devices (0.08% vs. 0.04%, p=0.0048), with larger LAA ostia (median 23mm vs. 21mm, p=0.004), and with a smaller difference between device and LAA ostial size (median difference 4mm vs. 5mm, p=0.04). Of 98,147 patients with 45-day follow-up, post-discharge DME occurred in 0.06% (n=54) patients and cardiac surgery was performed in 7.4% (n=4) of cases. The 45-day mortality rate was 3.7% (n=2) among patients with post-discharge DME. Post-discharge DME was more common among men (79.7% of events but 58.9% of all procedures, p=0.0019), taller patients (177.9cm vs. 172cm, p=0.0005), and those with greater body mass (99.9kg vs. 85.5kg, p=0.0055). The rhythm at implant was less frequently AF among patients with DME compared to those without (38.9% vs. 46.9%, p=0.0098)

Conclusion

While Watchman DME is rare, it is associated with high mortality and frequently requires surgical retrieval, and a substantial proportion of events occur after discharge. Due to the severity of DME events, risk mitigation strategies and on-site cardiac surgical back-up are of paramount importance.

INTRODUCTION

Incomplete anchoring of the Watchman left atrial appendage occlusion (LAAO) device can result in device migration or device embolization (DME) requiring percutaneous or surgical retrieval. While device embolism was reported in ~1/200 procedures in pivotal trials^{1, 2} and ~1/500 patients from the LAAO Registry who were matched to trial patients,³ the rates of this morbid and sometimes fatal complication in clinical practice, where the device is being used broadly, are unknown.

METHODS

We performed an analysis of Watchman procedures (1/2016–3/2021) from the National Cardiovascular Data Registry LAAO Registry⁴ to determine rates and characteristics of in-hospital and post-discharge Watchman DME events in the US (https://cvquality.acc.org/docs/default-source/ncdr/data-collection/laao_v1-3_data-collection-form.pdf). We excluded patients with prior LAAO interventions, no device released, and missing device information. In-hospital events were assessed among all patients and post-discharge events were assessed among patients with 45-day follow-up data. Differences among procedures with or without DME were assessed using Wilcoxon rank sum test for continuous variables and the χ² test for categorical variables. Due to the small number of outcomes, no statistical adjustment was performed. The Yale University Human Investigation Committee approved analysis of data from the LAAO Registry with a waiver of informed consent.

RESULTS

Of 120,278 Watchman procedures, DME occurred in 0.07% of patients (n=84; 51 were embolizations) during the index hospitalization and surgery was performed in 49 patients (Table). In-hospital DME was more common: at hospitals with a lower median annual procedure volume (24 vs. 41 procedures, p<0.0001), with first-generation Watchman versus Watchman FLX devices (0.08% vs. 0.04%, p=0.0048; Figure), with larger LAA ostia (median 23mm vs. 21mm, p=0.004), and with a smaller difference between device and LAA ostial size (median difference 4mm vs. 5mm, p=0.04). The Supplemental Table describes all characteristics that were significantly different among patients with versus without DME in the in-hospital setting. There were no differences by age, sex, hospital type, hospital size, or teaching versus non-teaching status. The in-hospital mortality rate was 14% among patients with DME and 20.5% among patients who underwent surgery.

Table.

Outcomes among patients with and without device embolization or migration during the index hospitalization or after discharge

Outcome	In-hospital Cohort (n=120,362)					Post-discharge cohort* (n=98,147)
	Embolization or migration		No embolization or migration		p-value	Embolization or migration		No embolization or migration		p-value
	N	%	N	%		N	%	N	%
Total	84/120,362	0.07	120,278/120,362	99.93		54/98,147	0.06	98,093/98,147	99.94
Death	12/84	14.29	197/120,278	0.16	<0.0001	2/54	3.70	961/98,093	0.98	0.0986
Cardiac surgery	49/84	58.33	281/120,278	0.23	<0.0001	4/54	7.41	48/98,093	0.05	<0.0001
Death after cardiac surgery	9/84	10.71	37/120,278	0.03	<0.0001	0/54	0.00	3/98,093	0.003	NA
Death without cardiac surgery	3/84	3.57	160/120,278	0.13	<0.0001	2/54	3.70	958/98,093	0.98	0.0862

^*between discharge and post-operative day 45 follow-up data

NA = not applicable

Figure.

Combined bar and line graph with bar shading reflecting the proportion of total devices that were first generation Watchman 12.5 devices versus Watchman FLX devices and the line depicting the rate of in-hospital (grey line) and post-discharge (yellow) device embolization or migration per 100,000 patients.

Of 98,147 patients with 45-day follow-up, DME after discharge occurred in 0.06%(n=54; 23 were embolizations) patients and cardiac surgery was performed in 7.4%(n=4) of cases (Table). The 45-day mortality rate was 3.7% (n=2) among patients with post-discharge DME. Post-discharge DME was more common among men (79.7% of events but 58.9% of all procedures, p=0.0019), taller patients (177.9cm vs. 172cm, p=0.0005), and those with greater body mass (99.9kg vs. 85.5kg, p=0.0055); in contrast to in-hospital events, there were no differences in hospital volume, device characteristics (Figure), or LAA characteristics. Patients with post discharge DME were less commonly in AF at the time of the implant (38.9% vs. 46.9%, p=0.0098) and had a lower platelet count (145,000/μL vs. 202,000/μL, p=0.0043). The Supplemental Table describes all characteristics that were significantly different among patients with versus without DME after discharge.

The mean device size was larger among patients with DME who required cardiac surgery in the in-hospital (28±3mm vs. 26±4mm, p=0.0082) but not post discharge time period (26±3mm vs. 26±4mm, p=0.9964). Device size was similar among patients who did and did not die after DME, regardless if DME was diagnosed in the in-hospital or post discharge setting.

DISCUSSION

In this analysis of commercial Watchman implantations in the US from January 2016 through March 2021, DME was rare but associated with high rates of surgical intervention and risk of mortality. Approximately half of all DME cases occur after discharge and different patient and procedure characteristics are associated with in-hospital versus post-discharge cases. Although we observed substantially less DME compared to the pivotal trials, given the morbidity and mortality associated with DME, maintaining risk mitigation strategies and on-site cardiac surgical back-up are of paramount importance.

A manuscript of in-hospital and post-discharge LAAO device embolization among procedures performed until November 2017 reported a rate of ~2%; this report was limited by inclusion of multiple device types and reliance on older published reports of embolizations for calculation of a numerator and an estimated number of procedures performed was used for the denominator.⁵ The PINNACLE FLX Trial reported no device embolizations among the 400 patients in the study;⁶ this is overall consistent with the 0.04% rate of events observed in the current study.

Characteristics associated with in-hospital DME include some modifiable factors including hospital volume, use of the Watchman FLX device, and device sizing. In contrast, post-discharge events were associated with male sex, increased height, and increased weight; the newer generation Watchman FLX device was not associated with post-discharge events. The smaller differences between LAA ostial size and among patients with DME suggests that device undersizing may increase the risk for DME. However, embolization of larger devices was associated with increased need for cardiac surgery, potentially related to increased risk of intracardiac entrapment which can cause abrupt hemodynamic collapse when entrapment results in acute valvular dysfunction. Patients with post-discharge DME were less likely to be in AF at the time of the implant procedure. It is possible that that device sizing in sinus rhythm results in a relatively smaller device size (due to more robust LAA contraction and smaller measured LAA orifice) compared to AF where the LAA may be more dilated. In this setting, relative under-sizing could make late DME more common among patients who were implanted in sinus rhythm and subsequently experienced AF. The concept of dynamic LAA size in AF versus sinus rhythm has also been used to explain the higher reported rate of 30-day pericardial effusion and tamponade among patients who underwent Watchman LAAO in AF.⁷

Limitations

This analysis has limitations including those inherent to any registry study, including the possibility of underreporting of adverse events. Additionally, the LAAO Registry has limited data on site of embolization and management of DME, including attempts at percutaneous retrieval prior to surgical intervention. The LAA orifice size reported to the LAAO Registry is a clinically obtained value and there is no required imaging modality, measurement approach, or core laboratory confirmation.

CONCLUSIONS

Watchman device migration or embolization is rare but associated with high mortality and frequently requires surgical retrieval. A substantial proportion of all device migration or embolization cases occur after discharge and different patient and procedure characteristics are associated with in-hospital versus post-discharge cases.