Introduction

Cardiac surgery in patients deemed at high surgical risk is associated with substantial rates of morbidity and mortality. With a prevalence ranging between 1.1% up to 17%, perioperative stroke presents as a common and potentially devastating complication1,2. Pertinent risk factors include increased age, atrial fibrillation, calcification of the aortic valve and ascending aorta, type of surgery, and prolonged cardiopulmonary bypass (CPB) time3. Mitral annular calcification (MAC), which occurs in 8 to 15% of patients, is also associated with an increased risk of both stroke and mortality (up to 5% rate of stroke and 11.1% in-hospital mortality for transatrial transcatheter mitral valve replacement [TMVR])4. As most of these factors are unmodifiable, procedural strategies for the prevention of perioperative stroke are needed.

Cerebral embolic protection (CEP) devices are approved for use in transcatheter aortic valve replacement (TAVR) and may be a tool to mitigate the risk of perioperative stroke during cardiac surgery5. Randomized controlled trials (RCT) have clearly demonstrated the safety of the use of the Sentinel Cerebral Protection system during TAVR and its ability to capture embolic debris in the vast majority of cases, albeit with unclear efficacy in preventing clinical stroke5,6. Studies investigating the use of CEP in cardiac surgery are limited. Thus far, only the Embol-X intra-aortic filtration device and the CardioGard device, both integrated into the aortic cannula for cardiopulmonary bypass (CPB), have been analyzed in larger numbers in the setting of cardiac surgery with promising but inconclusive results7,8. Therefore, the aim of this study was to analyze the feasibility, safety, and efficacy of the transcatheter Sentinel CEP system in ten high-risk candidates undergoing open heart surgery.

Patients and methods

Patient selection

Ten high-risk surgical candidates with native valvular heart disease (severe MAC and mitral valve disease [regurgitant, stenotic, or mixed lesions]) or failed aortic bioprosthesis were consecutively enrolled between March 2023 and April 2024 (Fig. 1). All participants underwent surgical intervention (Valve in Mitral Annular Calcification [ViMAC] with direct, transatrial SAPIEN 3 Ultra device implantation; mitral valve replacement [MVR]; surgical aortic valve replacement [AVR]) per the local standard of care with the use of Sentinel CEP system (Boston Scientific, Maple Grove, MN). Appropriate vascular anatomy for the Sentinel CEP was confirmed by pre-operative computed tomography (CT) angiography of the chest and neck. In patients with MAC, valve sizing was performed based on annular measurements derived from cardiac CT using a dedicated protocol and then confirmed intraoperatively, as previously described9. Data regarding patient demographics, comorbidities, pre- and postprocedural imaging, and operative details were abstracted from the electronic medical record.

Fig. 1
figure 1

Preprocedural CT images and corresponding surgical views of three sample patients showing extensive MAC (#2), extensive aortic bioprosthetic calcification (#7), and mitral bioprosthetic infective endocarditis (#10).

Full size image

This analysis was performed retrospectively and approved by the Institutional Review Board of Columbia University Medical Center (CUMC) (number: AAAV2111; approval date: 17APR2024). Patients’ informed consent was waived by the Institutional Review Board of Columbia University Medical Center. The study was conducted in accordance with the Declaration of Helsinki.

Pre-procedural imaging

All patients underwent comprehensive transthoracic and/or transesophageal echocardiography imaging prior to surgery. Examinations were performed and analyzed according to published guidelines10.

Pre-operatively, a mitral- or aortic-protocol multi-detector CT scan with intravenous contrast was performed (ECG-gated, multi-phase 0–95% with slice thickness < 1.25 mm at 5–10 mm increments) in each case. Reconstruction of the mitral and aortic annulus was performed using 3Mensio software (3Mensio Structural Heart Mitral Workflow version 8.1 Pie Medical Imaging, Maastricht, the Netherlands).

Candidacy for Sentinel filter use were as previously described5. In brief, the CT angiography was reviewed for tortuosity, arch calcium, and innominate (9–15 mm) and left common carotid artery size (6.5–10 mm) for Sentinel filter placement. In addition, a palpable right radial pulse was confirmed in all patients pre-operatively.

Surgical procedure

Median sternotomy was performed. ACT-guided heparinization was instituted for an ACT > 400. A 6 F sheath was placed in the right radial artery. A clamp was placed in the chest for marking the takeoff of the innominate artery from the aorta for Sentinel placement. Next, the Sentinel filters were deployed in the typical fashion in the innominate artery and the left carotid artery using fluoroscopy (Fig. 2). Cannulation for CPB was performed, and the surgical procedure undertaken.

Fig. 2
figure 2

Sentinel cerebral embolic protection device implantation steps. (A) Deploying filter cone in innominate artery, (B) Making a “U-turn” with device tip to facilitate left common carotid artery access, (C) Deploying filter cone in left common carotid artery.

Full size image

The detailed course of ViMAC procedure was described previously (Supplementary Video 1)4,9.

The Sentinel device was removed under fluoroscopy after the removal of the cross-clamp and confirmation of acceptable valve and ventricular function, prior to protamine administration. The patient was then weaned from bypass, decannulated and hemostasis was achieved.

Study definitions

High-risk patients were defined as either having a high STS PROM score (> 8%) or presenting with concomitant conditions that significantly increase the actual perioperative risk of stroke (i.e. severe MAC, presence of degenerated bioprosthesis, porcelain aorta, bioprosthetic endocarditis)(Supplementary Video 2). The neurological events were classified according to the VARC-3 criteria11. Additionally, the MAC severity was classified according to the CT-based MAC score12.

Statistical analysis

Data are reported as means +/− standard deviation or medians with interquartile ranges (IQRs) for continuous variables and frequencies with corresponding percentages for categorial variables.

Results

The baseline clinical characteristics of 10 patients are presented in Table 1 and Supplementary Table (1) All individuals were symptomatic, high-surgical-risk patients, suffering from severe valvular disease (mainly mitral valve disease with severe MAC) and were burdened with multiple comorbidities, with a mean STS score of 6.0%. Six patients had preexisting atrial fibrillation, 3 patients had a prior stroke, and 2 patients had prior cardiac surgery. The preprocedural echocardiography revealed severe, circumferential mitral annular calcification in all but 2 patients (severe aortic stenosis due to degenerated bioprosthetic valve; endocarditis on mitral prosthesis). The mean ejection fraction was 59%, and the mean LVEDD was 42 mm. Detailed baseline echocardiographic parameters are listed in Supplementary Table (2) One patient had a porcelain aorta.

Table 1 Basic demographic data.
Full size table

All patients successfully underwent the designated procedures. Few intraprocedural complications were observed in ViMAC patients. In one individual, transesophageal echocardiography after Sapien 3 Ultra valve deployment revealed a severe paravalvular leak (PVL) extending posteriorly (P2 adjacent to the large immobile calcium bar, forming a crevice) which required re-cross-clamp. The defect was closed with a felt patch that was fashioned and sewn to the left atrial wall and the valve frame. Another patient suffered a lateral wall perforation, likely from the Sapien 3 Ultra delivery system nosecone, which was realized after coming off CPB (and after Sentinel filter removal) with concomitant severe hypotension. The patient was immediately re-cannulated, placed back on CPB, and the defect was repaired using a horizontal mattress repair with felt and bovine pericardium. After this, all bioprostheses remained well positioned and no valvular regurgitation or greater than mild PVL was reported throughout the in-hospital postoperative course (Table 2).

Table 2 Intra-operative data.
Full size table

The Sentinel device was successfully deployed and recaptured in all cases. No Sentinel-related vascular or other complications were observed in any of the cases. Clearly visible, large deposits of calcium debris were captured by 3 CEP devices (see Fig. 3 for a sample image). Formal histopathologic studies will be initiated in the future.

Fig. 3
figure 3

Large calcium debris captured by the sentinel cerebral embolic protection device.

Full size image

The postprocedural status was continuously monitored in each patient with enhanced neurological surveillance employed. No significant neurological deficits (above mild neurological dysfunction; NIHSS > 5) were found. A prolonged delirium was reported in a single individual, in whom a head CT on postoperative day (POD) 12 did not yield significant findings. However, a subacute-chronic infarct of the right corpus callosum was seen in MRI performed on POD 24. This patient had suffered an LV perforation that required surgical repair, as mentioned above. In one patient, despite a negative head CT scan and transient arm weakness, an MRI showed small, acute to subacute multifocal ischemic infarcts in the right middle cerebral artery (MCA) and posterior cerebral artery (PCA) territories suggesting covert central nervous system injury.

Nine out of ten patients suffered postprocedural complications that ranged from new-onset left bundle branch block to cardiogenic shock. One individual gradually deteriorated with her stay being complicated with acute kidney injury, wound infection and pneumonia that led to respiratory failure. Eventually, on postoperative day 64, the patient expired. Importantly, her neurological status remained intact throughout the course of the hospitalization. Despite various complications reported (Table 3) all other patients survived and were successfully discharged home in good clinical standing.

Table 3 Post-operative data.
Full size table

Discussion

In the current case series, the Sentinel Cerebral Embolic Protection system was utilized in open, high-risk cardiac surgery capturing embolic debris and potentially reducing the risk of stroke or other brain injury. Sentinel deployment and recapture was feasible in all subjects. No CEP device-related complications were reported in any of the patients. Despite the high risk of surgery and the presence of extensive calcifications in each patient, no moderate or severe neurological dysfunction (NIHSS > 5) or disabling stroke were reported. These preliminary results represent an important step forward in the effort to reduce intraoperative, embolic stroke in high-risk cardiac surgery.

Periprocedural stroke constitutes a common and potentially devastating complication. Most of the cardiac surgery related strokes are of the ischemic origin with thromboembolism and hypoperfusion constituting the prevailing mechanisms. Thromboembolism accounts for 70–80% of intraoperative strokes and is considered to likely result from surgical manipulation of valves and the subsequent detachment of atheromatous debris. Watershed strokes, which are attributable to prolonged global cerebral hypoperfusion (secondary to low perfusion pressure during CPB or to extensive distal cerebrovascular disease, or after multiple microemboli), as well as mixed causes are less common and represent around 20–30% of cases13,14. Not only is the perioperative stroke responsible for a 5- to 10-fold greater likelihood of in-hospital mortality, but also increases cost and length of stay, as well as risk of cognitive decline at 1 year after surgery13. The cardiac surgery associated stroke risk varies greatly depending on the procedure type and patient cohort ranging from 1.1% up to 17%1,2. This is especially evident in the populations at risk, such as those with significant MAC or those who require complex re-operative valvular surgery. The surgical implantation of transcatheter valve-in-mitral annular calcification has emerged as an alternative to both surgical MVR and TMVR in patients deemed at prohibitive surgical risk and with risk of LVOT obstruction15. Although the technique offers high technical success rates and reduced mortality, the rate of complications remains high (periprocedural stroke rate of 5% in mitral stenosis MAC patients)4,15.

In our patients, the risk of stroke was estimated using the STS calculator as 3.5%, which may underestimate the embolic stroke risk given the massive calcific burden in these patients. The Sentinel device grossly appeared to capture significant amounts of embolic debris and may prevent large, disabling stroke. Although non-disabling stroke was evident by imaging in two patients. One patient had no neurologic finding and the other had prolonged delirium. Head CT in affected individuals did not present any significant acute findings, while MRI revealed subacute infarct areas. Nevertheless, the unknown onset of these ischemic changes combined with the preserved neurological status pointed to limited clinical significance and an unlikely relation to largescale, major embolic phenomenon.

Cerebral embolic protection devices are approved for use during transcatheter aortic valve replacement where they proved to be a useful adjunct to high-risk TAVR procedures. Nevertheless, the body of conducted studies have failed to show an undisputed benefit of using CEP during TAVR.

The PROTECTED TAVR was the first randomized, open-label, multicenter, all-comer trial powered for clinical endpoints. The study failed to prove the significant reduction of periprocedural stroke in patients undergoing transfemoral TAVR with CEP (2.3% vs. 2.9%; difference, − 0.6% points; 95% confidence interval, − 1.7 to 0.5; P = 0.30). Still, it did not rule out the potential benefit of CEP and showed that the incidence of disabling stroke was significantly reduced with the use of it (0.5% of the patients in the CEP group and 1.3% of those in the control group; difference, − 0.8% points; 95% CI − 1.5 to − 0.1)5. These findings were corroborated in a recently published meta-analysis of 6 RCTs that suggested that Sentinel CEP might reduce disabling stroke in patients undergoing TAVR and that patients with high and intermediate surgical risks were most likely to derive benefits16.

The data concerning cerebral embolic protection devices in open heart surgeries are limited. To date, there have been only several studies and reports investigating this topic7,8,16. The most studied devices are the Embol-X intra-aortic filtration device and the CardioGard device, both integrated into the aortic cannula. A RCT by Mack et al. including 383 patients undergoing surgical aortic valve replacement, showed that there was no significant difference in freedom from stroke between suction-based extraction and a standard aortic cannula (32.0% vs. 33.3%, respectively) or between intra-aortic filtration and a standard aortic cannula (25.6% vs. 32.4%, respectively)15. However, only 9% of patients exhibited clinical findings, and the majority of strokes were only detectable by postoperative diffusion-weighted MRI. Additionally, large volume infarcts were more numerous in patients in the control group, and infarct volume may be associated with a higher stroke risk.

Several small case series and case reports have been published demonstrating the use of the Sentinel CEP system in open heart surgeries17,18. Lau et al. presented a series of 6 patients with a high risk of perioperative stroke with CEP use18. The device proved its feasibility and safety and was able to capture debris in all cases. There was no device related complication reported or development of ischemic stroke postoperatively. Nevertheless, due to liberal criteria, the study included patients requiring therapies ranging from left ventricular vent placement, through biventricular assist device placement to MVR (3 procedures being surgical valve replacements).

The use of CEP during high-risk cardiac surgery is an appealing perspective that needs to be investigated in further studies on larger populations. Especially in patients with greatly increased risk of calcium debris detachment, such as MAC patients, the use of neuroprotection might prove justified. Further patients’ stratification utilizing tools like CT-based MAC score can help tailor the intervention and increase the CEP use efficacy. For patients with intermediate to severe MAC score the use of CEP seems reasonable as the high calcium burden and frequent need for extensive manipulation and aggressive debridement during procedures such as ViMAC surgery results in high risk for embolic liberation of calcium. The design of future trials should account for the complexity of patients that can prove the establishing of direct causality between the intraprocedural factor and the neurological consequence difficult. In addition, the use of general anesthesia impedes early clinical assessment of stroke. A more stringent radiologic evaluation with the routine use of MRI imaging could be useful. Finally, more focus on neurological complications other than stroke, such as delirium and cognitive decline would be an important adjunct clinical consideration.

Limitations

The present study is a single center, retrospective case series and is subject to all the inherent limitations of such a design (selection bias, single type CEP device, no control group). Furthermore, neuroimaging was performed only in patients in whom stroke was suspected. Thus, covert (asymptomatic) brain infarction could not be identified. Moreover, the detailed analysis of debris components was not conducted.

Conclusions

The current study extends the early experience demonstrating the feasibility and safety of cerebral embolic protection with the Sentinel CEP device in high-risk open surgery. Particularly in the highest risk patient sub-sets, CEP devices may offer advantages, reducing the risk of periprocedural episodes and improve clinical outcomes. Further clinical research is underway to study its use in a larger population.