Post-TAVI Stroke More Often Noted at Specialized Centers

Post-TAVI stroke rates are higher at hospitals classified as comprehensive stroke centers (CSCs) than at those without that designation, data from Michigan show, highlighting concerns about using the complication as a quality metric for the procedure.

The likelihood of in-hospital stroke after TAVI was higher at CSCs even after accounting for patient and clinical characteristics (adjusted OR 2.21; 95% credible interval 1.03-4.62), according to researchers led by P. Michael Grossman, MD (University of Michigan, Ann Arbor).

But there were no differences between center types for other important clinical outcomes, suggesting that the CSCs had higher stroke rates not because they’re delivering lower-quality TAVI care, but because they’re better at recognizing the events.

“I think all of us performing these procedures need to be vigilant for this complication, and when and if it occurs, ensure that the patients are getting appropriate management and follow-up,” Grossman told TCTMD.

These findings, published in the January 23, 2023, issue of JACC: Cardiovascular Interventions, also have implications for discussions around quality and public reporting of outcomes in the TAVI realm. “Using stroke post-TAVR as a quality metric that’s publicly reported and then could potentially be used to compare performance of individual TAVR centers is problematic,” said Grossman.

No Differences in Other Major Outcomes

Stroke remains a feared complication of the procedure, and there has been talk of including 30-day stroke rates as a metric of TAVI quality that could be publicly reported. Stroke has been reported at lower rates in real-world data sets than what was seen in the pivotal trials, possibly due to advancements in TAVI technology and techniques or to less-rigorous neurological assessment out in the community. But there is also substantial variability across centers, hinting that the difference could be related to how individual hospitals are equipped to recognize stroke.

To delve into this question, Grossman and colleagues turned to the Michigan TAVR Collaborative (now called the Michigan Structural Heart Consortium), a quality-improvement effort that collects and analyzes data sent by centers in the state to the TVT Registry. The current analysis included 6,231 patients who underwent TAVI between January 2016 and June 2019 at 22 Michigan hospitals; nine of those centers were classified as CSCs.

At both CSC and non-CSC sites, the mean patient age was about 79 years, and slightly more men than women were treated. The average STS predicted risk of mortality score was about 6% regardless of center type.

All of us performing these procedures need to be vigilant for this complication, and when and if it occurs, ensure that the patients are getting appropriate management and follow-up. P. Michael Grossman

Before accounting for differences in patient and clinical characteristics, the rate of post-TAVI stroke occurring before discharge was more than double in the CSC centers (2.65% vs 1.15%), a disparity that remained after adjustment and across various sensitivity analyses.

In contrast, CSC status was not associated with other key outcomes, including in-hospital acute kidney injury or transfusion, 30-day mortality, and a favorable outcome at 1 year (survival without a substantial decrease in quality of life assessed using the Kansas City Cardiomyopathy Questionnaire).

The investigators say more research is needed to understand why there is a difference in stroke rates, but they discount the idea that it could be a true difference in the quality of TAVI since there were no differences in these other outcomes. It could be related to the treatment of more-complex patients at CSCs, they acknowledge, because even though the analyses accounted for differences in patient characteristics, they might not have captured all relevant variables. The third possible explanation is that CSCs are more likely find strokes both because they have systems in place to do so and because they were more likely to have participated in the pivotal TAVI trials.

‘Perverse Incentive’ to Underreport Strokes

If CSCs are just better at recognizing strokes, it could create a “perverse incentive” for hospitals to underreport the events so they’ll look better to patients, Grossman suggested.

“Using stroke as part of a composite measure that’s publicly reported for performance in hospitals that are doing TAVR may be fraught and may potentially put centers that are appropriately identifying strokes at a disadvantage,” he said.

Neurologist Steven Messé, MD (University of Pennsylvania, Philadelphia), who served on steering committee of the PROTECTED TAVR trial, said he’s not surprised by the current study because “it suggests, as we’ve suspected, that stroke is being underascertained and underreported in the voluntary reporting quality-improvement databases.”

It suggests, as we’ve suspected, that stroke is being underascertained and underreported in the voluntary reporting quality-improvement databases. Steven Messé

Differences in recognition of strokes and in case mix between CSCs and non-CSCs both likely explain at least part of the observed disparity in stroke rates, he said.

As for the implications of the findings, Messé said, “Voluntary reporting of adverse events leads to underestimation of the true incidence of this important outcome. Publicly making those available and rating a hospital based on these results is likely to make this even more problematic.” He said mandating release of stroke rates as a quality metric could pressure hospitals to not treat high-risk patients who would otherwise be helped with TAVI and also incentivize them to underreport the complication.

Alexandra Lansky, MD, and Yousif Ahmad, MD, PhD (both from Yale School of Medicine, New Haven, CT), raise this same possibility in an accompanying editorial, noting that public reporting of some PCI outcomes has led to risk-avoidance behavior among physicians and some undertreatment of high-risk patients.

“With the expansion of TAVR in the United States and moves to increase public reporting of TAVR outcome data, it is necessary to thoughtfully select metrics that will lead to the intended improvements in quality and patient outcomes rather than metrics that may affect physician or institutional behavior to the detriment of best practices and patient care,” they write.

Taking a broader view, Grossman said efforts are still needed to address post-TAVI strokes even though rates have been falling over time.

“Identification of strokes, particularly big strokes, early on actually can make a big impact for these patients so we still think it’s quite important,” he said. “We need to do more work as a community with embolic protection, number one, and then working together so that we’re all empowering our staff and our teams to screen for strokes, educating them about how to do that, and then activating our colleagues in stroke neurology when and if there are signs of an acute stroke post-TAVR, because something can often be done if we catch it early enough.”