Carotid artery dissections from transcarotid artery revascularization as reported by the Food and Drug Administration

Introduction

Carotid endarterectomy (CEA) remains to be the standard method of treatment for carotid artery stenosis. The development of the less invasive transfemoral carotid artery stenting (TFCAS) procedure has introduced an alternative for patients who are at high surgical risk for CEA. However, due to concerns over its risk for perioperative stroke, transcarotid artery revascularization (TCAR) was developed to address these shortcomings.^1,2

TCAR is a hybrid procedure that combines a flow reversal neuroprotection system with carotid artery stenting through direct, surgical access of the carotid artery. TCAR has demonstrated its safety in the Safety and Efficacy Study for Reverse Flow Used During Carotid Artery Stenting Procedure (ROADSTER) trial, reporting the lowest rate of stroke at 30 days and 1 year (1.4% and 0.6%) among any prospective trials of carotid artery stenting to date.^3,4 Even though TCAR appears to be a safe alternative to those that meet its indications, there remains a need to better characterize intraoperative injuries and their management.

ROADSTER trial revealed that the most common local complication related to the TCAR procedure was a carotid artery dissection (CD), which was observed in 5.7% of the patients. Rare complications, such as CD, seen in the ROADSTER trial can be studied on a national level by analyzing the Manufacturer and User Facility Device Experience (MAUDE) database maintained by the U.S. Food and Drug Administration (FDA) for surveillance of all FDA-approved devices. ⁵ The MAUDE database collects adverse event reports from a wide variety of mandatory reporters (manufacturers, importers, and device user facilities) and voluntary reporters (healthcare professionals, patients, and consumers), and thus is able to capture real-world use of medical devices and provide a larger patient population than the original clinical trials. This study aims to review reports submitted to the MAUDE database related to the Silk Road Medical’s ENROUTE Transcarotid Neuroprotection System used in the TCAR procedure in the first 4 years of FDA approval to better understand how CDs are managed intraoperatively. We anticipate that this analysis will add qualitative insight in further characterizing adverse outcomes of this emerging technology.

Material and methods

The MAUDE database was queried for all adverse events involving the Silk Road Medical’s ENROUTE Transcarotid Neuroprotection System from September 2016 to October 2020. The event types, “patient injury,” “device malfunction,” and “patient death” were selected. Duplicated reports were consolidated into single report for review. Event descriptions were individually analyzed to determine type of injury (carotid artery dissection, stroke, etc.), and time of injury (intraoperative, recovery, and post-discharge follow-up). CD reporting was further analyzed for associated procedural event at the time of injury, number of access attempts to CD repair, and type of CD repair. CD repairs were further categorized into endovascular repair (e.g., stent over dissection and conversion to transfemoral carotid artery stenting) and open surgical repair (e.g., conversion to CEA). Fisher’s exact was used for statistical analysis of categorical variables, where a p-value of <0.05 was determined to be statistically significant. No institutional review board approval was required as the MAUDE database is publicly available and de-identified.

Results

The MAUDE database query returned a total of 115 unique medical adverse reports. CDs accounted for 58 of these. Sheath placement was the most commonly associated procedural event (58.6% N = 34) followed by wire manipulation (8.6%, N = 5). Most of the CDs were identified intraoperatively (94.8%, N = 55) and three were incidentally identified postoperatively. Of the 55 CDs that were identified intraoperatively, 52 of them had adequate information about intraoperative CD repair. 40 of them did not require any additional access established whereas 12 of them involved two or more access attempts. Of the 3 CDs identified intraoperatively but did not undergo intraoperative repair, one resulted in ischemic stroke secondary to stent-thrombosis whereas the other two recovered fully with no complications.

Among CDs that underwent intraoperative repair, the proportion of endovascular repair was significantly higher (62.5%) compared to the proportion of open surgical repair (37.5%) in cases that did not require additional access (p = .044). However, the proportion of open surgical repair (75%) was significantly higher than the proportion of endovascular repair (25%) in CDs with persistent failure to engage the true lumen despite two or more additional access attempts (p = .039) as seen in Table 1.

OPEN IN VIEWER

Table 1.

Type of carotid artery dissection repair based on number of access attempts.

Intraoperative CD repair	Endovascular repair	Open surgical repair	p-value*
	N = 28	N = 24
Repair without additional access	25/40	15/40	0.044
Repair that required two or more access attempts	3/12	9/12	0.039

Two strokes were associated with intraoperative CD repair. One incident of stroke occurred during a conversion to CEA for CD repair. One incident of stroke was associated with a fall from a hypotensive event 7 hours after an endovascular CD repair. Both incidents did not require additional access to engage the true lumen.

Of the three CDs identified postoperatively, two of them were found during routine postoperative surveillance. Both of these underwent endarterectomy without complication. The third was identified in a patient that returned with a contralateral stroke 4 days after TCAR procedure.

Discussion

The present study aimed to characterize CDs that occur from TCAR procedure and its intraoperative management on a large scale by analyzing all incidents of CD reported on the MAUDE database. This study revealed findings that are consistent with the ROADSTER trial in that, despite the overall safety of the procedure, CD was the most commonly reported complication from TCAR and the majority of events were associated with sheath placement.

Arterial dissection is one of the known iatrogenic complications of invasive vascular procedures. An arterial dissection can lead to ischemic injury secondary to stenosis, occlusion, or thromboembolism. Dissection of the carotid artery specifically poses a risk for stroke if not properly managed. CD secondary to iatrogenic injury from cerebral angiograms have been discussed. Iatrogenic CD from transfemoral cerebral angiogram are rare, generally known to have benign clinical course and can be managed conservatively.^6–8 Among iatrogenic CDs that require intervention, studies have shown that stenting is safe and reliable.^9–12 However, the management of iatrogenic CD from a hybrid procedure such as TCAR is less well studied.

Unlike iatrogenic CDs from cerebral angiograms, the current study showed that most CDs identified as a complication from a TCAR procedure underwent open surgical repair (N = 54/58). This is also in contrast to the ROADSTER study in which only three of eight patients underwent invasive repair. This may suggest that the mechanism of injury involved in placing the TCAR sheath with 10F outer diameter may be associated with greater degree of flow-limiting/hemodynamically significant CD. Device use in its early years by inexperienced operators may have contributed to this as well. Additionally, a considerable number of interventions were converted to open surgical repair (N = 24/52). This likely reflects the hybrid nature of the procedure where the common carotid is already surgically exposed. This likely allowed vascular surgeons, who are trained on both endovascular and open surgical approach to have lower threshold for conversion to CEA. This is also consistent with CD repairs seen in the ROADSTER trial where 1 was repaired endovascularly and 2 were repaired with open surgical approach.

Despite its rarity, precautions need to be taken to prevent CDs in the TCAR procedure, as it can lead to thromboembolic events. There are several steps a vascular surgeon can take to minimize CD occurrence. First, computed tomography angiography and duplex ultrasound should be utilized prior to the procedure to ensure that the patient anatomy is suitable according to the manufacturer’s instructions for use (IFU). Second, the point of arterial access should be relatively free of disease, as calcified lesions at the access site can increase the risk for CDs.^13,14 Thirdly, there should be adequate purchase for the stiff wire in the common carotid artery when advancing the Uber-Flex^TM arterial sheath. Lastly, it is critical for the assistant to provide adequate countertraction on the common carotid artery as the primary operator advances the arterial sheath.

While the MAUDE database provides descriptions of adverse events from a medical device use that can be utilized to investigate specific adverse events, it is not without limitations. Since it is a database that consists of reports submitted to the FDA by both the mandatory reporters (manufacturers, importers, and device user facilities) and voluntary reporters (healthcare professionals, patients, and consumers), the frequency of CD seen in the MAUDE database may not reflect its incidence in the real world. Additionally, non-standardized reporting of event descriptions and procedural details does not allow for deriving risk factors based on patient anatomy, demographics, comorbidities, and surgical technique. For example, reports detailing dissections secondary to sheath insertion do not specify whether it was due to the microsheath or the Uber-Flex^TM arterial sheath. Similarly, reports describing dissections from wire manipulations do not further detail type of wire and its locations (e.g., ECA and CCA). Lastly, whether the procedures were done within the manufacturer’s IFU were not included in these reports. Despite these limitations, the MAUDE database is a useful tool in understanding the management of a rare complication, such as CD, in the TCAR procedure that would otherwise be difficult to investigate from the original ROADSTER trial.

Conclusion

Carotid artery dissection is the most common operative complication related to TCAR as reported on the MAUDE database. The most common procedural event associated with CD was sheath placement. The proportion of intraoperative endovascular and open surgical CD repair was associated with whether the access to the true lumen required additional access attempts or not. This should add qualitative insight among vascular surgeons regarding when they might encounter a carotid artery dissection and how they can be managed. Finally, the MAUDE database was utilized to help investigate this rare complication, but due to the limitations of this database, results of this study may not be in line with real-world experience.