Author + information
- Chetan B. Patel, MD∗ ( and )
- Kishan S. Parikh, MD
- ↵∗Address for correspondence:
Dr. Chetan B. Patel, Division of Cardiology, Department of Medicine, Duke University Medical Center, DUMC 3034, Durham, North Carolina 27710.
Left ventricular assist device (VAD) implantation occurs at a time when the patient is sickest during the heart failure journey and most greatly affected by active comorbid conditions. It is at this critical time that the body is asked to withstand high-risk cardiothoracic surgery. Surgical recovery and adjustment to new routines and mechanical equipment continue for several more weeks in both inpatient and ambulatory settings. As the patient’s cardiac insufficiency is relieved, the systemic effects of chronic heart failure such as muscle wasting and malnutrition begin showing improvement. Therefore the body, aided by the newly implanted VAD, has an enhanced capacity for improvement and a greater reserve to tolerate comorbid conditions. Although the experience with left VAD therapy has grown exponentially, much remains unknown regarding best practices to maximize post-surgical quality and quantity of life.
In this issue of JACC: Heart Failure, Bachmann et al. (1) investigate the use of cardiac rehabilitation (CR) among 1,164 Medicare beneficiaries who received VAD therapy in 2014. These investigators report on patients’ characteristics, CR participation rates, and associated risks for subsequent hospitalizations and mortality. Major findings are as follows: One-third of VAD-treated patients participated in CR, although on average they completed just two-thirds of the prescribed regimen. Of the examined variables, only Medicare census region was predictive of increased CR use. CR participants had a 23% lower adjusted risk for being hospitalized, and the only other variable more strongly associated with hospitalization risk was the presence of chronic pulmonary disease (42% increased adjusted risk). CR use was also associated with a 47% lower adjusted risk for death at 1 year. Conversely, markers associated with increased mortality rates were discharge to inpatient rehabilitation or skilled nursing facility, peripheral vascular disease, and weight loss—pre-implant characteristics that represent physical and logistic barriers to early and effective CR.
These findings further inform strategies for better delivery of post-VAD care, which remains an area in need of more rigorous study. Recent publications have identified limitations of the current infrastructure to manage this complex group of patients. This is true even among highly experienced and high-volume VAD centers, emphasizing the impact of non–heart failure comorbidities on outcomes after VAD therapy (2,3). Accordingly, Bachmann et al. (1) appropriately recognize healthy cohort bias as an inherent shortcoming of the current analysis and use several methods to address this limitation. The Elixhauser comorbidity index, which approximates risk for adverse outcomes by the presence of weighted comorbid conditions, was calculated using codes billed during the hospitalization for implantation and the preceding year. Bachmann et al. (1) also implemented 2 sensitivity analyses. First, they used a marginal structural model that addressed time-dependent confounders before CR initiation and found similar results. Next, a sensitivity analysis was performed to evaluate for the potential effect of frailty, a confounder recognized to be unmeasurable in the present analysis. Drawing on prior estimates of the prevalence of frailty in patients with heart failure and VAD, these investigators found that CR participation would no longer be associated with a survival advantage if frailty was present in <37% of their cohort.
The overall message of this analysis is that CR is underused and may be associated with significant morbidity and mortality benefits by treating the frailty associated with end stage heart failure. The magnitude of this treatment effect is consistent with findings in patients post-coronary artery bypass grafting (4), but more pronounced than findings in the patients with chronic heart failure who were included in the HF-ACTION study (Heart Failure: A Controlled Trial Investigating Outcomes of Exercise Training) (5). Additional characterization of the post-VAD course can also be gleaned from this study. Interestingly, increasing age was not associated with increased risk for hospitalization at 1 year, but comorbidities, including chronic pulmonary disease, depression, and renal failure, were significant risks and reinforced the concept that understanding frailty in patients with heart failure is critical even after VAD implantation. This is also reflected in the markedly worse prognosis for hospitalizations and survival among patients discharged to inpatient rehabilitation or skilled nursing facilities.
The study by Bachmann et al. (1) has several strengths, including the large number of patients and inclusion of data from all VAD centers in the United States that allowed for a representative sample. Use of Medicare claims files for description of patients’ characteristics, resource use, and clinical outcomes has been well validated and successful in many prior investigations. The multiple sensitivity analyses assessed the potential impact of comorbidities and did not identify confounding including frailty as an explanation for the results.
The study also has important limitations. Healthy cohort bias remains a significant concern despite the statistical methods used, and only randomized data, which may not be feasible, would be able to show the effects of CR definitively in these patients. The CR participation rates should be taken in context of the relative newness of the heart failure indication for CR. Although Bachmann et al. (1) used perhaps the most recently available Medicare claims files for the analysis (2014 to 2015 year), a repeated analysis using data from a subsequent year would be needed to provide a more accurate understanding of CR uptake. Furthermore, given the lack of granularity into individual site practices, it is unclear what, if anything, is the importance of geographic location despite its relatively strong statistical association with CR participation. It is remarkable, however, that VAD-treated patients in the Midwest had nearly a 60% greater chance for CR initiation than did their South census counterparts.
This analysis also does not provide insight into the reasons that VAD recipients did not participate in CR. It is possible that consistency of systematic referrals may vary by implanting institution or that some of the nonparticipants were recommended to undergo CR but declined. Further, it is possible that providers did not recommend CR to patients because of ongoing medical reasons not captured by the claims data, such as worsening right-sided heart failure, inadequate social support at home, or the need for frequent ambulatory encounters interfering with an exercise program.
Finally, several factors may account for improved adjusted 1-year outcomes with CR in the VAD-treated patients. The Kaplan-Meier plot for mean cumulative hospitalizations shows an early separation of curves once CR is initiated. Although skeletal muscle function and peak oxygen uptake have been shown to improve with exercise training in VAD recipients, it is still unknown whether an additional benefit is obtained that is unique to VAD physiology, in addition to affecting the deconditioned state of a patient with heart failure. For example, could exercise training in these patients improve right-sided heart function or augment improvements in lung mechanics and gas exchange? Would patients with chronic heart failure derive similar benefit from exercise training if they had already undergone CR pre-VAD implantation under the heart failure indication? It is also possible the timing of CR initiation post-cardiac surgery is important. In both this retrospective analysis and the Rehab-VAD randomized controlled trial (Cardiac Rehabilitation in Patients With Continuous Flow Left Ventricular Assist Devices: Rehab-VAD Trial) (6), exercise programs were initiated between 70 and 100 days post-implantation. Whether an earlier CR start leads to better outcomes is unknown and would be worth exploring prospectively.
In summary, the study by Bachmann et al. (1) suggests that patients post-VAD implantation substantially benefit from CR participation. The hope is that future mechanistic and randomized studies of exercise training in VAD-treated patients will provide important insights into the global physiological impact of heart failure and ways to optimize the journey to recovery.
↵∗ Editorials published in JACC: Heart Failure reflect the views of the authors and do not necessarily represent the views of JACC: Heart Failure or the American College of Cardiology.
Dr. Patel has served as a consultant to Abbott and Medtronic. Dr. Parikh has reported that he has no relationships relevant to the contents of this paper to disclose.
- 2018 American College of Cardiology Foundation
- Bachmann J.M.,
- Duncan M.S.,
- Shah A.S.,
- et al.
- Cowger J.A.,
- Stulak J.M.,
- Shah P.,
- et al.
- Pack Q.R.,
- Goel K.,
- Lahr B.D.,
- et al.
- Kerrigan D.J.,
- Williams C.T.,
- Ehrman J.K.,
- et al.