Low BMI Predicts Stroke, Death After TAVR

In high-risk patients who undergo transcatheter aortic valve replacement (TAVR), a low body mass index (BMI)—a marker of frailty—increases the chances of stroke and death over the next few years. Moreover, the risk imparted by an access-site or bleeding complication is exacerbated by the presence of low BMI as well as by older age, prior stroke, and atrial fibrillation (A-fib), according to a study published online November 20, 2012, ahead of print in Circulation: Cardiovascular Interventions.

In the absence of these comorbidities, TAVR patients have a life expectancy similar to that of age-matched counterparts without aortic stenosis.

For the prospective registry study, Stephan Windecker, MD, and colleagues at Bern University Hospital (Bern, Switzerland), looked at 389 patients with severe aortic stenosis and elevated surgical risk who underwent TAVR at their institution between July 2007 and 2011.

Risk Evolves with Events

Of the entire cohort, 42.5% of patients progressed directly to kidney injury (9.5%), access-site or bleeding complications (20.1%), or stroke or death (12.9%). In addition, among patients who developed periprocedural kidney injury, 40.5% died, as did 41.0% of those who experienced access-site or bleeding complications.

Over a median follow-up of 1.1 years, transapical access emerged as a predictor of kidney injury (HR 2.12; 95% CI 1.00-4.47; P = 0.049) and access-site and bleeding complications (HR 1.78; 95% CI 1.07-2.96; P = 0.026) but was not associated with increased risk of stroke or death. Body mass index (BMI) of 20 kg/m2 or less, however, independently increased the risk of stroke or death (HR 2.64; 95% CI 1.25-5.54; P = 0.011). In a multivariable model, age older than 80 years, BMI of 20 kg/m2 or less, prior stroke, and A-fib at baseline all became predictors of death or stroke after an intervening access-site or bleeding complication (table 1).

Table 1. Multivariable Predictors of Death or Stroke After an Access-Site or Bleeding Event



95% CI

P Value

Age > 80 Years




BMI ≤ 20 kg/m2




Prior Stroke




A-fib at Baseline




Identifying Vulnerability

For the study, Dr. Windecker and colleagues used multistate analysis to estimate mortality and stroke taking into account the impact of intervening events. In an interview with TCTMD, Philippe Généreux, MD, of Columbia University Medical Center (New York, NY), praised the authors for looking at outcome predictors in an innovative way.  “This model tells us that if you’re very sick at baseline, you have less chance of surviving another insult,” he said.

Basically, the findings provide guidance for predicting outcomes for patients with multiple comorbidities who experience multiple events following a procedure, Dr. Généreux said, adding, “We’re still searching for a way to identify those for whom [TAVR] is going to go well, because, as the authors point out, when it does, mortality is low.”

Currently, all TAVR candidates are high risk, Dr. Généreux noted. “And the sicker they are, the more important it is to minimize complications because they have less chance to survive the first hit,” he said. “But if you can minimize the intensity of the first hit, that may make a big [difference]. For example, if you have minor bleeding instead of major bleeding, you can give 1 transfusion rather than 5, and the [risk of a bad outcome] is proportional to the number of transfusions.”

Dr. Généreux noted that “we are entering an era of miniaturization of devices, and it will be interesting to see how frail patients respond” as new technology and growing operator skill diminish the severity of postprocedural complications. In-hospital care also plays an important role in preventing or minimizing complications, he added.

What Does Multistate Analysis Add?

But in a telephone interview with TCTMD, Peter C. Block, MD, of Emory University School of Medicine (Atlanta, GA), was less impressed with the authors’ analysis.

“Any [study] that gives us data about patients who are less likely to get good benefit from TAVR is important data,” he said. “Unfortunately, I’m not sure what to take away from this paper.”

“The fancy [multistate] algorithm just tells me that people who have lots of comorbidities are not likely to do well,” Dr. Block said. “But we already learned that from other databases.” Moreover, he said, focusing on predictors of mortality “begs the question, because the reason we replace valves is not to save lives—[these patients] die from their other comorbidities. The reason we replace valves is to make the rest of their lives more comfortable.

“On a positive note,” he continued, “this kind of data collection is important because it helps us to understand which patients should and shouldn’t have TAVR. This database is an attempt to winnow comorbidities out from valve replacement—or to see how the 2 interact—and that’s laudable. But at the end of the day, [the authors] simply say that people who are frail and have [certain other comorbidities] don’t do well.

“The endgame we’re all searching for is some kind of an index that is the equivalent of an STS score that will tell us what the real risk is for patients undergoing TAVR,” Dr. Block said. “We need a big enough database so that we can really understand how to score these folks with appropriate variables.”

Meanwhile, not all patients die from complications, he observed. “So do you take everyone that has the comorbid [profile outlined in the paper] and say, ‘Sorry, you don’t qualify for TAVR, you’re going to have to die of heart failure instead’? It’s a terribly difficult nut to crack, and I don’t think this paper cracks it,” Dr. Block said.

Study Details 

Patients received either the Edwards Sapien (Edwards Lifesciences, Irvine, CA) or CoreValve (Medtronic, Minneapolis, MN) bioprosthesis, delivered via transfemoral (79.2%, more than half CoreValve), transapical (19.5%, all Edwards), or subclavian (1.6%, all CoreValve) access.

The study population was roughly equally divided between men (42%) and women (58%), with a mean age of 82.4 years. Logistic EuroScore and STS score averaged 24.3 ± 14.2% and 6.8 ± 5.3%, respectively. 

All endpoints were defined according Valve Academic Research Consortium criteria.

Pilgrim T, Kalesan B, Wenaweser P, et al. Predictors of clinical outcomes in patients with severe aortic stenosis undergoing TAVI: A multistate analysis. Circ Cardiovasc Interv. 2012;Epub ahead of print.


  • Drs. Genereux and Block report no relevant conflicts of interest.

Related Stories:

Low BMI Predicts Stroke, Death After TAVR

In high risk patients who undergo transcatheter aortic valve replacement (TAVR), a low body mass index (BMI)—a marker of frailty—increases the chances of stroke and death over the next few years. Moreover, the risk imparted by an access site or
  • Dr. Windecker reports receiving research grants from Edwards Lifesciences and Medtronic.