Skip to main content
Skip main navigationClose Drawer MenuOpen Drawer Menu

Public Reporting of Stroke After Transcatheter Aortic Valve Replacement: A Cautionary Tale Free Access

Editorial Comment

J Am Coll Cardiol Intv, 16 (2) 177–178


Stroke is a catastrophic complication of any cardiac intervention or surgery, and patients consider stroke a worse outcome than death.1 Patients undergoing transcatheter aortic valve replacement (TAVR) specifically value maintaining independence above survival.2 Therefore, therapeutic strategies to reduce the occurrence and clinical impact of stroke are highly relevant, and patients may consider institutional stroke rates when evaluating treatment options. The study from Grossman et al3 in this issue of JACC: Cardiovascular Interventions touches on both of these important issues (ie, stroke rates after TAVR and the implications of public reporting of these data).

The incidence of postprocedure stroke in the early TAVR clinical trials was ∼5% to 6%; these rates declined in the more contemporary trials that included lower-risk patients. These declining rates are also likely a reflection of some combination of advances in TAVR technology, improvement in operator technique with experience, and the lower-risk phenotype of patients now being treated with TAVR. In registries, stroke rates have remained ∼2.0% to 2.5%,4,5 generally below the rates reported in randomized clinical trials (RCTs).

The lower observed stroke rates reported in real-world practice compared with RCTs are somewhat of a paradox because event rates in trials are generally lower than clinical practice because of a combination of factors, including narrow inclusion criteria for RCTs excluding patients with multiple severe comorbidities, enhanced medical attention received through participation in RCTs with regularly scheduled follow-up and contact with clinical research teams, and trial participants being a self-selected group with better health motivation than patients typically encountered in routine clinical practice.6,7 The discrepancy between reported stroke rates in RCTs and real-world registries is presumed to reflect differences in ascertainment methods, but this has never been proven.

The current study advances some persuasive evidence to explain this phenomenon and some cautionary arguments against public reporting. Grossman and colleagues hypothesized that increased stroke rates after TAVR would be associated with hospitals with a comprehensive stroke center (CSC) certification that have the greatest available expertise and resources for the diagnosis and management of acute stroke. Using the Michigan TAVR collaborative data, the 22 participating sites were compared on the basis of CSC certification and, indeed, demonstrated that in-hospital stroke rates occurred more frequently at CSC sites (2.65% vs 1.15%; P < 0.001). This difference in stroke rate persisted in sensitivity analyses adjusting for patient and clinical characteristics as well as when restricting the population to those with transfemoral access.

The authors postulate that CSC sites are more likely to identify patients with stroke because of greater expertise in rapid diagnosis and treatment of stroke. This is analogous to stroke rates in RCTs being higher than in registries because of mandated systematic neurologic assessments as part of trial protocols. The absence of differences in other TAVR quality metrics between CSC and non-CSC sites including mortality, acute kidney injury, and blood transfusions in their study further supports the impact of the stroke ascertainment method on reported rates.

A word of caution in the current study pertains to its reliance on site-reported outcomes collected into a national registry. The observed differences in strokes between CSC and non-CSC sites could also be credibly explained by inherent differences in the treated patient populations. For example, alternative access TAVR is known to be associated with an increased risk of stroke and was more common in patients treated at CSC sites. Although the transfemoral-only subgroup analysis was consistent with the primary analysis, there are other important differences in baseline characteristics that could reasonably be expected to impact stroke rates such as prior stroke, carotid disease, prior cardiac surgery, and a history of peripheral arterial disease. Despite these baseline differences, the Society of Thoracic Surgeons Predicted Risk of Mortality score was similar in the 2 groups, which may be a demonstration that this risk model does not necessarily perform well in patients undergoing TAVR.8 It also remains plausible that CSC sites draw higher-risk patients and that the differences in stroke rates are predominantly explained by residual confounding of baseline imbalances in the 2 groups that are not adequately accounted for in the analysis models.

A broader issue highlighted by this study is the prospect of stroke rates being used as a public reporting metric for TAVR centers. The authors raise appropriate concern with this approach because sites treating higher-risk patients or with greater expertise in assessing and diagnosing stroke would be considered worse performers when the opposite may be true. We are all familiar with the scenario of patients attending clinic to discuss potential procedural therapies armed with printouts from publicly reported metrics or requesting specific adjunctive interventions (such as cerebral embolic protection devices). This study raises the possibility that public reporting of stroke rates post-TAVR could lead to the potentially disastrous consequence of incentivizing sites to systematically under-report strokes, or worse yet, avoid treatment of clinically indicated high-risk patients. This scenario has played out in the realm of percutaneous coronary intervention in which public reporting has been introduced in some states in an effort to improve quality of care but may have had the unintended consequence of leading to risk aversion among physicians and potential patient harm from undertreatment of those at highest risk.9

The current study speaks to some of the challenges of using postprocedural stroke as a basis for ranking or comparing TAVR programs. With the expansion of TAVR in the United States and moves to increase public reporting of TAVR outcome data,10 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. In this case, centers reporting higher stroke rates after TAVR, because they may be caring for higher-risk patients or have more integrated neurologic assessment and treatment approaches, should not be incentivized to lower their standards.

Funding Support and Author Disclosures

The authors have reported that they have no relationships relevant to the contents of this paper to disclose


  • 1. Ahmad Y., Nijjer S., Cook C.M., et al. "A new method of applying randomised control study data to the individual patient: A novel quantitative patient-centred approach to interpreting composite end points". Int J Cardiol 2015;195:216-224.

    CrossrefGoogle Scholar
  • 2. Coylewright M., Palmer R., O’Neill E.S., Robb J.F., Fried T.R. "Patient-defined goals for the treatment of severe aortic stenosis: a qualitative analysis". Health Expect 2016;19:1036-1043.

    CrossrefMedlineGoogle Scholar
  • 3. Grossman P.M., Sukul D., Lall S.C., et al. "The relationship between hospital stroke center designation and TVT reported stroke: the Michigan TAVR Experience". J Am Coll Cardiol Intv 2023;16:2: 168-176.

    View ArticleGoogle Scholar
  • 4. Huded C.P., Tuzcu E.M., Krishnaswamy A., et al. "association between transcatheter aortic valve replacement and early postprocedural stroke". JAMA 2019;321:2306-2315.

    CrossrefMedlineGoogle Scholar
  • 5. Auffret V., Lefevre T., Van Belle E., et al. "Temporal trends in transcatheter aortic valve replacement in France: FRANCE 2 to FRANCE TAVI". J Am Coll Cardiol 2017;70:42-55.

    View ArticleGoogle Scholar
  • 6. Sherman R.E., Anderson S.A., Dal Pan G.J., et al. "Real-world evidence — what is it and what can it tell us?"N Engl J Med 2016;375:2293-2297.

    CrossrefMedlineGoogle Scholar
  • 7. Kennedy-Martin T., Curtis S., Faries D., Robinson S., Johnston J. "A literature review on the representativeness of randomized controlled trial samples and implications for the external validity of trial results". Trials 2015;16:495.

    CrossrefMedlineGoogle Scholar
  • 8. Martin G.P., Sperrin M., Ludman P.F., et al. "Inadequacy of existing clinical prediction models for predicting mortality after transcatheter aortic valve implantation". Am Heart J 2017;184:97-105.

    CrossrefGoogle Scholar
  • 9. Waldo S.W., McCabe J.M., O’Brien C., Kennedy K.F., Joynt K.E., Yeh R.W. "Association between public reporting of outcomes with procedural management and mortality for patients with acute myocardial infarction". J Am Coll Cardiol 2015;65:1119-1126.

    View ArticleGoogle Scholar
  • 10. Desai N.D., O’Brien S.M., Cohen D.J., et al. "Composite metric for benchmarking site performance in transcatheter aortic valve replacement: results from the STS/ACC TVT Registry". Circulation 2021;144:186-194.

    CrossrefMedlineGoogle Scholar


The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the Author Center.