Understanding how aggressively to treat patients in their last years of life is challenging for a number of reasons. Beyond the difficulty of predicting the benefit that any intervention will provide to a patient who has already exceeded average life-expectancy, issues related to the proper allocation of scarce societal resources are also becoming increasingly relevant. In this commentary, Suzanne Arnold and David Cohen discuss when transcatheter aortic valve implantation (TAVI) may do more harm than good for a patient who is nearing the end of their life.
A recent analysis from Medicare showed that average healthcare spending per capita rises with age and peaks at age 96 at US$16,145 per year—more than twice the per capita spending at age 70.1 When one considers adding the cost of TAVI for such a patient (at a cost of US$60–US$70,000),2 the issue of the medicalisation of ageing becomes even more concerning. Although both clinicians and patients want to avoid the risks and expense of treatment when a procedure is futile, there has been little guidance to date about how and when not to offer TAVI.
To know how to “pull back” when dealing with patients at the end of life can be difficult, particularly when presented with an individual patient who just wants to feel better. While TAVI prolongs survival and improves quality of life for many patients,3–5 a number of patients still have poor outcomes despite a successful procedure.6 Ideally, we want to be able to predict and avoid these poor outcomes. However, this is not a simple task. First, simply defining a poor outcome after TAVI has challenges, particularly since survival is not the only relevant outcome to these patients.7 While it is possible to conceptualise a composite outcome that integrates survival and quality of life,8 a single definition of a poor outcome may not apply to all patients as some may value quality of life more or less than others.
Even if we accept this composite outcome as a meaningful endpoint, estimating the likelihood of a poor outcome prospectively is also challenging. Our research has shown that the key predictors of a poor outcome after TAVI are severe comorbid conditions (eg. renal disease, lung disease)6 and frailty and disability (including poor functional capacity),9 both of which limit the potential of patients to derive meaningful benefit with TAVI. Although we have successfully developed6 and externally validated10 models for estimating risk of a poor outcome after TAVI, these statistical models provide only moderate discrimination. As a result, they are particularly effective in estimating risk of a poor outcome but are limited in reliably identifying individuals who are certain not to benefit, which is typically what we want to know when determining futility.
This distinction is important, since there is no defined threshold of risk of a poor outcome that uniquely defines futility. In our work, we have used a risk of poor outcome of ≥50% at six months to define high-risk—a threshold that encompasses about 10% of the high- and extreme-risk TAVI population. Based on data from both the PARTNER and CoreValve pivotal trials, 41% of these high-risk patients were dead and an additional 16% had very poor quality of life or quality of life decline by six months after TAVI.11 While this may seem like an acceptable threshold of risk above which we should not to offer TAVI, it is likely that individual patients vary with respect to their tolerance for risk. From a societal perspective, if a patient’s life expectancy is less than two years or if quality of life remains poor after successful valve replacement, TAVI does not meet current benchmarks for cost-effectiveness.2 However from an individual patient perspective, the decision of whether or not to offer TAVI requires nuance and determination of the patient’s preferences, ideally supported by risk estimation that is based on the patient’s characteristics.
Finally, even if one has determined that there is a high likelihood of a poor outcome after TAVI, having the conversation about why a patient is unlikely to meaningfully benefit from TAVI can be difficult. In our experience, however, patients are often more accepting of these discussions than physicians are comfortable initiating them. Interestingly, cardiothoracic surgeons have substantial experience here, as turndown rates for aortic valve surgery are about 10% due to excessive risk.9 As the decision to offer TAVI is most often done as part of a multidisciplinary valve team, the role of the surgeon in these conversations offers a potential opportunity for cardiologists to learn how to approach these high-risk patients. Importantly, in these cases, the patient should not feel abandoned and also needs to be realistic about his or her limited life expectancy. Involving palliative care specialists in the treatment plan can be an effective means to facilitate these important discussions and to enable adequate symptom control.10 While risk models can be effective mechanisms to identify high risk patients who may not be good candidates for TAVI, these decisions are often still going to be difficult, and having an established treatment plan for patients who are not offered TAVI may ease this process for both patients and physicians.
- Neuman et al. Health Aff (Millwood) 2015; 34: 335–39.
- Reynolds et al. Circulation 2012; 125: 1102–09.
- Leon et al. N Engl J Med 2010; 363: 1597–07.
- Reynolds et al. Circulation 2011; 124: 1964–72.
- Osnabrugge et al. JACC Cardiovasc Interv 2015; 8(2): 315–23.
- Arnold et al. Circulation 2014; 129(25): 268–90.
- Lewis et al. J Heart Lung Transplant 2001; 20(9): 1016–24.
- Arnold et al. Circ Cardiovasc Qual Outcomes 2013; 6(5): 591–97.
- Bach. J Heart Valve Dis 2011; 20(3): 284–91.
- Lauck et al. Curr Opin Support Palliat Care 2016; 10(1): 18-23.
- Arnold et al. J Am Coll Cardiol. In press.
Suzanne Arnold and David Cohen are both at the Department of Cardiology, St Luke’s Mid America Heart Institute, Kansas City, USA