Calculating Risk for Poor Outcomes After Transcatheter Aortic Valve Replacement

From Saint Luke’s Mid America Heart Institute/University of Missouri–Kansas City, Kansas City, MO.

Abstract

• Objective: To outline the tools available to help understand the risk of transcatheter aortic valve replacement (TAVR) and the gaps in knowledge regarding TAVR risk estimation.
• Methods: Review of the literature.
• Results: Two models developed and validated by the American College of Cardiology can be used to estimate the risk of short-term mortality, a 6-variable in-hospital model designed for clinical use and a 41-variable 30-day model designed primarily for site comparisons and quality improvement. Importantly, neither model should be used to inform the choice of TAVR versus surgical aortic valve replacement. Regarding long-term outcomes, a risk model to estimate risk of dying or having a persistently poor quality of life at 1 year after TAVR has been developed and validated. Factors that most significantly increase a patient’s risk for poor outcomes are very poor functional status prior to TAVR, requiring home oxygen, chronic renal insufficiency, atrial fibrillation, dependencies in activities of daily living, and dementia. If a patient has ≥ 2 or 3 major risk factors for a poor outcome, this risk and the uncertainty about the degree of recovery expected after TAVR should be discussed with the patient (and family).
• Conclusion: It is important to understand the patient factors that most strongly drive risk of poor outcomes after TAVR and use this information to set appropriate expectations for recovery.

Keywords: aortic valve stenosis; risk factors; postoperative complications; TAVR.

Among patients with severe aortic stenosis, trans­catheter aortic valve replacement (TAVR) has emerged as a less invasive option for aortic valve replacement. This procedure offers substantial reductions in mortality and improvement in quality of life compared with medical therapy^1,2 and at least similar long-term outcomes compared to surgical aortic valve replacement (SAVR).^3-9

As with any emerging technology, selecting the appropriate patients for TAVR—a procedure with high initial costs¹⁰—has been an area of active investigation. As TAVR was first introduced in patients who were considered inoperable, initial efforts focused on trying to identify the patients who did not improve functionally or live longer following TAVR. Termed Cohort C patients, these patients were thought to have too many comorbidities, be too sick, and have too little reserve to recover from TAVR, and in the early trials, represented a substantial minority of the patients. For example, in pivotal clinical trials of patients at high or extreme surgical risk, approximately 1 in 4 patients who were treated with TAVR were dead at 1 year.^1,3,11 Furthermore, a number of patients who received TAVR were alive at 1 year but continued to have significant heart failure symptoms and functional limitations.^2,4 Practitioners,^12,13 regulators,¹⁴ and third-party payers¹⁵ have recommended that TAVR should not be offered to patients in whom valve replacement would not be expected to positively impact either their survival or quality of life, but how best to identify these patients has been less clear.

More recently, as the use of TAVR has moved down the risk spectrum, patient selection for TAVR has shifted to understanding which patients should be preferentially treated with TAVR versus SAVR. While patients often prefer a less invasive treatment option with faster recovery—which is what TAVR offers—there are lingering questions about valve longevity, need for a pacemaker (and the associated long-term implications), and the ability to treat other cardiovascular conditions (eg, Maze, mitral valve repair) that potentially make a patient a more appropriate candidate for valve surgery. This review outlines the tools currently available to help understand the risk of TAVR and the gaps in knowledge.

Short-Term Outcomes

When TAVR was initially introduced, the 30-day mortality rate was 5% to 8%.^1,11,16 This high mortality rate was a function of treating very ill patients and more invasive procedures with larger sheath sizes and routine use of general anesthesia, transesophageal echocardiography, pulmonary artery catheterization, and so on. Over time, however, this rate has gone down substantially, with the 30-day mortality rate in intermediate- and low-risk patients now ranging from 0.5% to 1%.^8,17-19 Although this low mortality rate indicates that the vast majority of patients will survive to discharge from the hospital, 2 models can be used to estimate the risk of short-term mortality: an in-hospital²⁰ and a 30-day model,²¹ both developed and validated by the American College of Cardiology. The in-hospital model was developed for clinical use, as it includes only 6 variables (age, renal function, severe lung disease, non-femoral access, New York Heart Association class IV, and acuity of the procedure [elective versus urgent versus shock versus emergent])²⁰ and has an online calculator (https://tools.acc.org/tavrrisk/). The 30-day model was developed for risk adjustment (primarily for site comparisons and quality improvement) and includes 41 variables (including pre-TAVR patient health status and gait speed).²¹

While 30 days is a better time frame for assessment because outcome is less impacted by differences in local post-acute care facilities, we explicitly did not create a parsimonious 30-day mortality model for clinical use due to concern that having such a model would allow for indirect comparisons with estimated risk of SAVR using the Society of Thoracic Surgeons risk model (https://riskcalc.sts.org/stswebriskcalc). It would be tempting to estimate a patient’s risk of mortality with the TAVR calculator and the SAVR calculator and use those risk estimates to inform the choice of treatment; however, these risk estimates should not be directly compared to make treatment selections, as they were built on entirely different patient populations. In real-world practice, there is minimal overlap in the characteristics of patients who are treated with TAVR and SAVR. For example, in an analysis that merged surgical and transcatheter databases, less than 25% of patients treated with TAVR could be matched to a clinically similar patient treated with SAVR.²² As such, these TAVR models should be used to estimate a patient’s risk for short-term mortality, but should not be used to contribute to the decision on TAVR versus SAVR.