The use of embolic filters to reduce the stroke risk in TEVAR


Richard Gibbs and colleagues have previously demonstrated that 80% of patients undergoing thoracic endovascular aneurysm repair (TEVAR) have evidence of new post-procedural cerebral infarcts by comparing pre- and postoperative diffusion-weighted magnetic resonance imaging scans. Whilst the majority have no overt stroke, the new silent infarcts are associated with a fall in cognitive function that persists in the elderly, Gibbs writes. 

We report on a feasibility study of cerebral protection in patients undergoing TEVAR. Based on the transcatheter aortic valve implantation experience of reducing procedural cerebral infarction, we elected to us the Sentinel Cerebral Protection System (Sentinel CPS, Claret Medical). This filter system is deployed in the innominate and left common carotid artery via a right brachial approach, and any particulate emboli are retrieved at the end of the procedure.

After bench top road testing in a pulsatile aortic flow model to ensure physical compatibility between the device and four different stent grafts (made by Gore, Medtronic, Bolton, and Cook), we have utilised the protection system in 10 patients undergoing TEVAR.

All patients underwent intraprocedural transcranial Doppler with software that could discriminate between solid and gaseous emboli (high intensity transient signals; HITS). Pre- and postoperative diffusion-weighted MRI scans were undertaken to determine new cerebral infarction. Embolic matter retrieved from the filters was histologically analysed by a core lab. Inclusion criteria were that the innominate and left carotid met the device sizing criteria for the Sentinel CPS, and that the proximal landing zone was in zones 2, 3 or 4.

The Sentinel CPS was successfully deployed in 90% of the patients (one distal filter failed to properly deploy in an emergency case). No patient suffered a stroke, and seven minutes was added to the total procedure time. Additional contrast of 23ml was needed, and the procedure added 2.2% additional radiation exposure. Transcranial Doppler results demonstrated that the median number of total HITS for the full procedure was 453 (227–521). The majority of these were gaseous rather than solid. Deployment of the filter caused the greatest number of HITS, and these were 95% gaseous. The majority of solid HITS occurred during wire and stent manipulation and deployment. We noted a significant number of gaseous emboli HITS during TEVAR, and these were correlated with the number of small (<3mm) post-procedural infarcts. Diffusion-weighted MRI scans showed evidence of new acute brain infarcts in 80% of the patients. Compared to a matched group of our unprotected historical controls there was a reduction in both the number and size of lesions. In both the protected and unprotected groups, the majority of new infarcts were seen in the vascular territory of the left posterior circulation.

The median total number of particles captured and retrieved in the filters was 937 (146–1,687). Retrieved embolic debris proved to be thrombus and arterial wall on histological analysis. The study demonstrates that the Sentinel CPS can be safely and successfully deployed in patients undergoing TEVAR with a proximal landing zone in zones 2, 3 and 4. Use of the system reduces both the number and size of new cerebral infarcts associated with TEVAR. The left hindbrain is the most vulnerable area for new infarcts in both protected and unprotected patients, and this will mandate the development of particulate embolic protection for the left subclavian artery. Consideration will also be needed in how to reduce air embolisation during TEVAR.

Richard Gibbs is Clinical Director General and Vascular Surgery at Imperial College, London, UK