Cerebral Embolic Protection During TAVI

Prevent the Unpreventable?

Mohamed Abdel-Wahab; Holger Thiele


Eur Heart J. 2019;40(17):1340-1341. 

Peri-procedural stroke is a feared complication of aortic valve replacement, and can occur after both surgical and transcatheter procedures. It is commonly caused by ischaemic brain injury secondary to procedure-related cerebral embolism, and is associated with significant morbidity and mortality. The occurrence of a peri-procedural stroke also consumes a significant proportion of healthcare resources, which extends well beyond the peri-procedural period.

Since the majority of cerebral events in patients undergoing transcatheter aortic valve implantation (TAVI) have an embolic origin that includes thrombi, calcific and atherosclerotic material, leaflet tissue, and calcific deposits from the native valve,[1,2] embolic cerebral protection devices (EPDs) have been developed to prevent these materials from reaching the cerebral vasculature, which may consequently minimize the extent of neurological injury. Of these EPDs, the Claret Sentinel dual filter (Boston Scientific, MA, USA) is the device with the most available scientific evidence, and is currently the only approved device for clinical use in both Europe and the USA. The device consists of a proximal and a distal filter delivered to the brachiocephalic and left common carotid arteries, respectively. Thereby, the device provides important—though incomplete—protection of the cerebral vasculature. Published small-sized randomized clinical trials have demonstrated device safety and captured embolic debris of variable size and nature in nearly all extracted filters.[3–5] However, a significant reduction in the ischaemic and total volume of cerebral lesions detected by cerebral magnetic resonance imaging could only be found in one single-centre trial,[4] and all these studies were not powered for the detection of a significant clinical advantage of the EPD in preventing peri-procedural stroke.

It is on this background that Seeger et al. have conducted a propensity-matched analysis of one non-randomized and two randomized clinical studies of the dual filter EPD, published in this issue of the European Heart Journal.[6] In this study, patient-level data from the multicentre randomized SENTINEL US IDE trial[5] were pooled with those of the single-centre randomized CLEAN TAVI trial[4] and the single-centre non-randomized SENTINEL ULM study,[7] leading to an overall study population of 1306 patients. A propensity score matching was possible without significant reduction in the number of included patients, resulting in a comparative analysis of 533 patients receiving TAVI with the dual filter EPD vs. 533 patients receiving TAVI without. The matching procedure resulted in two largely comparable groups, with some minor differences in baseline clinical and echocardiographic characteristics. In this study, TAVI with the use of the dual filter EPD was associated with a significantly lower rate of stroke [1.88% vs. 5.44%, odds ratio (OR) 0.35, 95% confidence interval (CI) 0.17–0.72, P = 0.0028] as well as death and stroke (2.06% vs. 6.00%, OR 0.34, 95% CI 0.17–0.68, P = 0.0013) compared with unprotected TAVI, with the endpoints being assessed during the peri-procedural period (72 h). Both relative (65.0%) and absolute (3.5%) risk reduction of stroke were clinically meaningful.

The authors should be congratulated on this important analysis, which may represent a further step towards establishing the clinical efficacy of EPDs. Important strengths of their study include the relatively large number of included patients, the successful matching of the study population, and the inclusion of all classes of currently available TAVI devices. The major strength of this study lies in its main endpoint of peri-procedural stroke. First, this is the clinical efficacy measure of an EPD that matters most to both patients and clinicians. Secondly, from a mechanistic point of view, EPDs would only be able to prevent procedure-related embolic events. Previous studies on EPDs used heterogenous time points for assessing peri-procedural stroke, ranging from immediate to 30 days post-TAVI. Defining the peri-procedural period as 72 h post-procedure is both plausible from a pathophysiological stand-point and in line with the Valve Academic Research Consortium definition of the peri-procedural period.[8]

On the other hand, the limitations of this individual patient-based propensity-matching study by Seeger et al. are also obvious, and the authors correctly point out the majority of these. By design, the included patient populations are heterogenous, and this is aggravated by the inclusion of the SENTINEL ULM study, which is a non-randomized single-centre registry including patients treated during two different sequential time periods. Even with the most sophisticated matching techniques, the bias created by various unmeasured confounders in the SENTINEL ULM study cannot be corrected for. With CLEAN TAVI failing to detect any significant difference in stroke, and the SENTINEL IDE trial showing—if any—a trend towards stroke reduction with the dual filter device, the stroke reduction in this pooled analysis seems to be mainly driven by the SENTINEL ULM registry, which also contributed to almost two-thirds of the included patients. The heterogeneity of currently available studies on EPD is similarly evident in two recently published meta-analyses, both reaching different conclusions on the clinical efficacy of EPDs.[9,10]

A further important limitation of the study by Seeger et al. lies in the definition of stroke in itself. The three included studies used various definitions of stroke, and neurological assessment was not mandated in all three studies. No information is provided in accordance with the recently published NeuroARC initiative, which sets out clear recommendations for the assessment of various types of cerebral injury, particularly those related to cardiovascular procedures.[11]

In this context, and despite the encouraging findings of this study, the jury is still out for the case of cerebral embolic protection during TAVI. The concept remains appealing, the dual filter device appears to be safe, and efficacy is pointing in the right direction (Figure 1). However, lessons learned from various apparently intuitive cardiovascular technologies underscore the necessity of adequately powered randomized controlled trials to prove the clinical benefit of these devices in reducing peri-procedural stroke. This applies to the dual filter EPD as well as to all other EPDs, which may be even easier to deploy and provide a more complete cerebral protection.[12] In addition, without more evidence it remains unclear if a routine or a selected approach in high-risk for stroke patients is adequate. Without these trials, it will be difficult to obtain clear guideline recommendations, advocate widespread clinical use, and justify complete device-specific reimbursement. We are now approaching a crossroad, and it should be clear if we are satisfied with the current level of evidence, which is more or less like beating around the bush, or if we have the courage to start beating into it.

Figure 1.

Summary of currently available evidence on cerebral embolic protection during transcatheter aortic valve implantation. MRI, magnetic resonance imaging.