6.2 / IAGS 2019
Session 6: Structural Session 3
Routine vs Selective Use of Cerebral Protection for TAVR: A Debate
Problem Presenters: Mike Rinaldi, MD (Pro) / Gus Pichard, MD (Con)
Statement of the problem
Despite advances in technique and technology, stroke remains a persistent and prevalent complication of TAVR, with major stroke rates between 1%-3% and major plus minor stroke rates 5%-10%, particularly when neurologist adjudicated. MRI and transcranial Doppler studies show embolic events in virtually all patients. Cerebral embolic protection devices offer the potential to capture or deflect emboli and thus reduce stroke rates. Clinical trial evidence suggests significant reduction in emboli and stroke, yet the pivotal Sentinel randomized trial showed a reduction in stroke with the Sentinel cerebral embolic protection (EPD) system that failed to meet statistical significance. The entire body of evidence did lead to FDA approval and commercial availability in the US, but widespread adoption has been limited. Given the significant unreimbursed cost of EPD technology, some operators have adopted a selective strategy by applying EPD only in patients perceived to be at higher risk of stroke. No clinical or anatomic characteristics have been shown that define a population that does not benefit from protection, therefore, this strategy is difficult to justify.
Gaps in knowledge
Despite registry data, post hoc analysis, and meta-analysis of clinical trial data all showing 60%-80% reductions in stroke rates, the single randomized Sentinel trial showed no statistically significant differen ce in stroke or emboli. For this reason, there remain questions about the efficacy of cerebral EPD within the cardiology community which have limited widespread adoption. Larger randomized trials may provide more definitive evidence and are planned. Additionally, asymptomatic cerebral emboli, which are nearly ubiquitous in TAVR, may be associated with long-term cognitive dysfunction, and routine use of EPD may attenuate this. However, this hypothesis is unproven and requires further study. Finally, EPD technology is associated with additional unreimbursed cost, potentially offset by the cost savings of strokes averted, and so the true cost effectiveness of EPD remains unknown.
Solutions and Future Directions
A more adequately powered randomized trial of Sentinel and of other EPD devices may provide more compelling data, and if positive, would be expected to lead to widespread adoption. Such trials may define anatomic or clinical characteristics that identify populations with greatest benefit, or may confirm that all patients are at risk and thus derive benefit from EPD. Until such data are available, a selective approach to a “higher risk” population cannot be justified. Therefore, operators, hospital systems, and patients will need to judge the present body of evidence and decide if it is compelling enough to justify systemic application. If future randomized trials provide more definitive evidence for stroke reduction and protection of cognitive function, it is likely that adoption into routine use will be widespread. Device improvements that provide more complete protection of all cerebral territories and simplify use may encourage adoption. Further data on the cost of stroke related to TAVR, and the cost effectiveness of EPD in TAVR, will further support routine use. Finally, it must be acknowledged that unreimbursed cost is a significant impediment to use, and adequate reimbursement will mitigate this barrier. There have been no signals of harm with Sentinel, and registry data and clinical experience have shown that EPD can be incorporated into clinical practice without loss of procedural efficiency. Until additional efficacy studies are available, individual operators will have to decide if the present body of data suggesting 60%-80% reductions in procedural stroke are compelling enough to justify routine use, or if the absence of definitive evidence is enough to justify a more conservative strategy. Selective use for perceived high-risk population subsets is not based on any currently available evidence and cannot be justified.