Clinical Stroke Rate After Open AVR Higher than Expected

 Download this article's Factoid (PDF & PPT for Gold Subscribers)

When patients undergoing surgical aortic valve replacement (AVR) are evaluated postoperatively by neurologists, clinical stroke is more than twice as common as reported in the Society for Thoracic Surgery (STS) database, according to a paper published online April 1, 2014,ahead of print in Circulation.

Investigators led by Steven R. Messé, MD, of the Hospital of the University of Pennsylvania (Philadelphia, PA), conducted a prospective observational cohort study of clinical stroke and radiographic cerebral infarction in 196 patients aged 65 years or older undergoing AVR for calcific aortic stenosis. Patients were evaluated preoperatively and postoperatively on days 1, 3 and 7 by neurologists for clinical stroke and TIA.

Clinical stroke was defined as new focal neurologic symptoms with vascular origins lasting longer than 24 hours or symptoms lasting less than 24 hours with radiographic evidence of infarction in the appropriate territory.

Most Strokes Mild, Early

Neurologists identified clinical stroke in 34 patients (17%), with 6% of those cases showing intracerebral hemorrhage on imaging. TIA occurred in 2% of patients, and the in-hospital mortality rate was 5%.

Most strokes were identified on day 1 after AVR (58%) and classified as mild, with a median National Institutes of Health Stroke Scale (NIHSS) score of 3 (IQR 1-9). Overall, clinical stroke did not have an effect on in-hospital mortality, though substantially greater risk was seen among the 8 patients (4%) with NIHSS scores of 10 or higher (table 1). In addition, clinical stroke as a whole was associated with longer hospital stay (12 days vs. 10 days; P = .02)

Table 1. In-Hospital Mortality After AVR


Clinical Stroke
(n = 34)

No Clinical Stroke
(n = 162)

P Value

All NIHSS Scores




NIHSS Scores > 10




Multivariate analysis identified several independent predictors of increased stroke risk:

  • Older age (OR 1.07 per year;95% CI 1.01-1.15)
  • Increased duration of cardiopulmonary bypass (OR 1.13 per 10 minutes;95% CI 1.04-1.22)
  • Higher mean arterial pressure nadir (OR 1.07 per mm Hg;95% CI 1.01-1.13)

Only 6.6% of patients in the same cohort were reported as having had a clinical stroke in the STS database. The median NIHSS score for these patients was 5, as compared to a median score of 1 for patients whose clinical strokes were not reported. Two of the patients with unreported stroke died in the hospital, and 9 had their symptoms resolve by post-operative day 7.

Postoperative MRI was performed in 129 patients, including 109 identified by neurologists has being stroke-free. Of these ‘stroke-free’ patients, silent infarct was identified on postoperative MRI in 54%. However, silent infarct did not increase the risks of in-hospital mortality or longer hospital stay.

Better Assessment Could Explain Higher Rate

In an email with TCTMD, Thomas F. Floyd, MD, of the Hospital of the University of Pennsylvania (Philadelphia, PA), said that the incidence of clinical stroke in this study may be higher than in others due to the fact that neurologists employed a uniform approach to the surveillance and diagnosis of stroke.

In fact, according to Dr. Messé, more clinical events were identified here than elsewhere because neurologic evaluations were performed early after the procedure.

“The vast majority of neurologic injury likely occurs intraoperatively and symptoms from brain ischemia are generally maximal at onset and then improve over time,” Dr. Messé said in an email with TCTMD. “So it is understandable that you can miss an event if you do not evaluate patients early.”

Also in an email with TCTMD, Peter Block, MD, of Emory Healthcare (Atlanta, GA), said that most physicians caring for post-AVR patients are guilty of assuming that maybe the patient is just “having a bad day”, rather than wondering if a neurological event has occurred. However, since only 196 out of 714 possible candidates were enrolled in the study, he also questioned whether a selection bias toward a higher risk population might have occurred.

Difficult to Compare with TAVR

Based on these results, there is reason to be concerned that the actual stroke rate after TAVR might also be higher than expected, according to Dr. Messé.

The PARTNER randomized study (N Engl J Med. 2011;364:2187-2198) found nearly double the risk of stroke when TAVR was compared to open surgical AVR; however, the more recent CoreValve trial showed similar stroke rates between open surgical AVR and TAVR.

“Importantly, the 2 studies actually showed similar rates of stroke within 30 days for TAVR— 4.7% in the PARTNER study and 4.9% in the CoreValve study,” Dr. Messé said. “I suspect that a meaningful number of minor events were missed for the reasons stated above and I think it is reasonable to expect that the stroke rate in clinical practice with new devices is likely to be higher than what was reported in the clinical trials.”

However, Dr. Block called the attempt to relate the current results to TAVR outcomes is a “meaningless exercise,” as is any discussion of whether the PARTNER trials versus the CoreValve trials show disparate stroke rates.

“Importantly, as far as TAVR is concerned, most of the data are again self-reported, though in the CoreValve trial a neurological exam was performed post-TAVR. But the extent of testing in all of the TAVR trials is far less comprehensive than in the present report, and I would expect that stoke rates post-TAVR are likely higher than we think,” Dr. Block said. “Clearly a careful prospective trial of equally careful examinations of post-TAVR patients might help clarify this controversy.”

Steps Can Be Taken to Minimize Stroke and Its Fallout

In an editorial accompanying the study, Ron Waksman, MD, and Sa’ar Minha, MD, of Medstar Washington Hospital Center (Washington, DC), write that the study by Messé et al highlights the vast amount of information that is still unknown.

“AVR, either via [surgical] or TAVR, is a lifesaving procedure that remains the most viable modality for the treatment of aortic stenosis,” Drs. Waksman and Minha note. “Therefore, it is imperative that rigorous, prospective, well-designed clinical trials are conducted to address the unknowns of stroke-related issues post AVR and to better understand how we can further minimize this unwarranted event.”

Until those trials can be conducted, Dr. Block said, treatment decisions should be directed at the need for AVR rather than stroke rates, given that aortic stenosis carries a high risk of mortality. “Nonetheless, we must continue to be diligent is trying to minimize stroke—even if post-AVR stroke rates were indeed only 1% to 2%, for the patient with a post-AVR [cerebrovascular accident,] the result is disastrous,” he added.

Dr. Floyd agreed, pointing out that patients who undergo coronary artery revascularization are monitored closely for the occurrence of new cardiac ischemia, and that the same effort could be directed at the brain.

For AVR patients who do experience stroke, Dr. Messé said, there are many ways that physicians could likely improve outcomes.

“Specifically, for major strokes identified early, patients may safely undergo intra-arterial thrombectomy, which can be performed with devices that do not require use of a thrombolytic agent,” he suggested. “In addition, supportive measures that probably improve outcome from acute stroke include raising mean arterial pressure goals and avoiding drops in blood pressure, and aggressively treating fevers and hyperglycemia.”  

1. Messé SR, Acker MA, Kasner SE, et al. Stroke after aortic valve surgery: results from a prospective cohort. Circulation. 2014;Epub ahead of print.

2. Waksman R, Minha S. Stroke after aortic valve replacement: the known and unknown [editorial]. Circulation. 2014;Epub ahead of print.



Related Stories:

  • Drs. Block, Floyd, Minha, and Waksman report no relevant conflicts of interest.
  • Dr. Messé reports serving as a consultant to GlaxoSmithKline and receiving salary support as co-principal investigator of a study on neuroprotection in high risk thoracic aortic repair sponsored by GlaxoSmithKline.