Simple Risk-Scoring System Predicts Mortality After TAVR

The use of a simple risk-scoring system, one that includes baseline assessments of comorbidity, frailty, and disability, predicted the short- and long-term risks of mortality in extreme- and high-risk patients undergoing transcatheter aortic valve replacement, a new study shows.

Importantly, say investigators, the four positive predictors of death at 30 days are not part of the traditional STS Predicted Risk of Mortality (STS-PROM) risk-assessment algorithm and three of the four metrics, excluding low serum albumin levels, can be assessed without invasive testing.

“What was very interesting was in looking at the 30-day mortality, the independent predictors are very simple. Are they on home oxygen? Are they in assisted living? Are they older than 85 years? Is the albumin less than 3.3 [g/dL]?” said lead investigator James Hermiller, MD (St. Vincent’s Heart Center of Indiana, Indianapolis). “These are really simple factors, something that clinicians can use as part of patient-centered decision-making. In the high-risk patient, we can delve more deeply into whether or not they are an appropriate candidate for TAVR. Would the procedure be futile? Looking at the [variables] very quickly would allow us to delve into it a little bit more than we have.”

The new study, published in the July 26, 2016, issue of the Journal of the American College of Cardiology, is an analysis of 3,687 patients from the CoreValve US Pivotal Trial and continued access registry. The researchers analyzed multiple variables, including extensive data on patient comorbidity, frailty, and disability, to identify predictors of death at 30 days and 1 year. These predictors were then used to devise a weighted risk-scoring system to classify patients as having low, moderate, or high risk of death following TAVR. 

Speaking with TCTMD, Hermiller said traditional risk-scoring systems, such as STS-PROM and EuroSCORE, are known to be poor tools for patient risk assessment in TAVR. Michael Mack, MD, and Elizabeth Holper, MD (Heart Hospital Baylor Plano, TX), make a similar assessment in an accompanying editorial. In the past, an initial patient assessment involved the “eyeball test” or the “end-of-the-bed-o-gram” to predict whether a patient would survive a planned procedure. While risk algorithms such as STS-PROM and EuroSCORE have been developed to standardize this subjective assessment, these algorithms were not developed or validated for assessing patient outcomes after TAVR.

More importantly, write Mack and Holper, functional limitations, such as frailty, are increasingly recognized as an important part of clinical decision-making, yet neither STS-PROM nor EuroSCORE include frailty as part of the risk-assessment algorithm. 

Risk Predictors Dominated by Frailty and Disability

In the CoreValve extreme- and high-risk patients, home oxygen use, residence in an assisted living facility, serum albumin level less than 3.3 g/dL, and age greater than 85 years were significant predictors of death. On the basis of the weighted risk score, the researchers observed a three-fold higher risk of death at 30 days among patients classified as high risk compared with those considered low risk. Overall, 10.9% of the 1,205 patients in the validation cohort classified as high risk died at 30 days compared with 3.6% of patients categorized as low risk (P = 0.002). Results were similar among the 2,482 patients in the derivation cohort.

At 1 year, predictors of mortality differed slightly, with more traditional risk-scoring systems, such as the STS-PROM score greater than 7% and a high Charlson comorbidity score, associated with an increased risk of death. Home oxygen use, low serum albumin levels, and falls within the past 6 months were also associated with an increased risk of death at 1 year. Again, those classified as high risk based on the devised risk score had a three-fold higher risk of death compared with low-risk patients. At 1 year in the validation cohort, the mortality rate was 36.6% in the high-risk group compared with 12.3% in the low-risk patients (P < 0.001).  

Regarding the predictors of mortality, Hermiller said they are dominated by markers of disability, frailty, and comorbidity. Given that many of these variables are easy to ascertain and can be assessed quickly, it’s the heart team’s responsibility to “delve deeper” into a patient if these predictors are identified in order to determine if a reasonable outcome can be expected, he said.

“There is an expanding realization that frailty and disability are potent drivers of how well patients do after the procedure,” said Hermiller. “What we’ve tried to do over time is to move from the eyeball test—they look frail, or they don’t look frail—to use more quantitative measures to truly refine whether they are in fact frail or have disabilities.” In the CoreValve trial, for example, frailty was evaluated using gait speed, grip strength, body mass index, unplanned weight loss, anemia requiring transfusion, hypoalbuminemia, and mobility. Disability was assessed using the Katz Index of Independence in Activities in Daily Living questionnaire and an examination for dementia. 

In their editorial, Mack and Holper state it is “reassuring” that significant predictors of 30-day mortality included variables that can be checked on an initial evaluation as this allows for greater clinical utility. The editorialists do point out that variables involved in the assessment of mortality risk at 1 year are more diverse and time-consuming to obtain. “It is clear that a single unified scale or scoring system cannot adequately predict both operative risk and long-term risk in an older patient with a complex clinical situation and aortic valve stenosis,” write Mack and Holper.

The current study suggests the need to incorporate frailty into existing risk scores, as well as the need to create a “unified, efficient means of evaluation,” add the editorialists.

Related Stories:

• Frailty Index Best at Predicting Functional Decline After TAVR
• Poor Mobility Strongly Predicts Mortality in TAVR Patients
• Risk Model Improves Practical Prediction of In-Hospital Mortality After TAVR