Author + information
- Jeffrey J. Teuteberg, MD∗∗ (, )
- Garrick C. Stewart, MD†,
- Mariell Jessup, MD‡,
- Robert L. Kormos, MD∗,
- Benjamin Sun, MD§,
- O.H. Frazier, MD‖,
- David C. Naftel, PhD¶ and
- Lynne W. Stevenson, MD†
- ∗Heart and Vascular Institute, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania
- †Division of Cardiovascular Medicine, Brigham and Women's Hospital, Boston, Massachusetts
- ‡Division of Cardiovascular Medicine, Hospital of The University of Pennsylvania, Philadelphia, Pennsylvania
- §Abbott Northwestern Hospital, Minneapolis, Minnesota
- ‖Department of Surgery, Texas Heart Institute, Houston, Texas
- ¶Department of Surgery, University of Alabama at Birmingham, Birmingham, Alabama
- ↵∗Reprint requests and correspondence:
Dr. Jeffrey J. Teuteberg, Heart and Vascular Institute, University of Pittsburgh, Scaife Hall, Suite 556, 200 Lothrop Street, Pittsburgh, Pennsylvania 15213.
Objectives This study investigated how the initial intended strategy at left ventricular assist device (LVAD) implantation influenced patient outcomes.
Background Left ventricular assist device implantation strategy impacts candidate selection, reimbursement, and clinical trial design; however, concepts of device strategy are continuing to evolve.
Methods For patients entered in the INTERMACS (Interagency Registry for Mechanically Assisted Circulatory Support) receiving a primary continuous flow LVAD between March 2006 and March 2011, initial strategies were bridge to transplant (BTT), bridge to candidacy (BTC) for transplant, and destination therapy (DT). Primary analyses compared BTT, BTC, and DT outcomes at 6, 12, and 24 months.
Results Among 2,816 primary LVAD recipients, implant strategy was 1,060 (38%) BTT, 1,162 (42%) BTC (likely to be listed 796, moderately likely 282, unlikely 84), and 553 (20%) DT. Compared with BTC/DT, those listed at implant (BTT) had similar degrees of ventricular dysfunction and hemodynamic derangement but generally less comorbidity. Survival (alive with LVAD or transplanted) was superior at 24 months for BTT versus BTC versus DT (77.7% vs.70.1% vs. 60.7%, respectively, p < 0.0001). Strategic intent changed over time, at 2 years 43.5% of BTT patients were no longer listed for transplant, but 29.3% of BTC patients were listed for transplant.
Conclusions The currently accepted indications only account for 58% of LVAD implants. Across indications, patients differ by the number and types of comorbidities rather than the need for hemodynamic support. Regardless of initial implant strategy, patients often have long durations of support, and strategies often change over time, challenging the regulatory categorization of LVAD recipients as either BTT or DT.
Although most patients currently receive mechanical circulatory support (MCS) as a bridge to transplantation (BTT), the evolution of the technology is toward long-term support (1) Although many BTT patients are listed at the time of implant and might be excellent candidates for transplantation, limited donor organ availability often results in wait times of over 1 year (2). Relative contraindications to transplantation such as pulmonary hypertension, sensitization, and uncertain social support might prevent a patient from being listed for transplant for some time after MCS, further prolonging the duration of support. Such patients are often referred to colloquially as a “bridge to candidacy” (BTC) (3). Although many BTC patients are eventually listed for transplant, some patients implanted as BTC might never become transplant candidates or might develop complications while on MCS that make them permanently ineligible for transplant, effectively changing their strategy to that of long-term mechanical support. After approval of destination therapy (DT) as a distinct implant strategy in 2002 (4), the adoption of DT remained low, but with the approval of a continuous flow device as DT there has been substantial growth in the DT population (5,6).
From a regulatory standpoint, all patients who receive MCS must have their intended strategy, either as BTT or DT, declared at the time of implant. However this dichotomization of patients as either BTT or DT at implant is problematic and does not reflect the clinical reality that many patients whose intent is for transplant are not listed at the time of implant due to some relative contraindication that will hopefully resolve after a period of support. The likelihood that BTC candidates will be listed for transplant varies with the number and types of relative contraindications present at the time of implant. The initial implant strategy might also change over time and patients who were intended as BTT may no longer be transplant candidates; conversely those intended as DT might eventually be considered for transplant. This regulatory framework thus artificially constrains trial design with each successive MCS technology requiring separate trials for each indication, typically first with a BTT trial, followed by a separate DT trial (7,8). The maintenance of these strategies as wholly separate entities has also been endorsed by the United States Centers for Medicare and Medicaid Services, which subsequently influences candidate selection and defines reimbursement for this resource-intensive therapy (9).
Despite these regulatory and reimbursement structures, little is known about the patient characteristics, likelihood of transplantation, or the change in intent over time in patients who are implanted as BTT compared with those who are BTC or DT. We hypothesized that the patient characteristics, duration of support, and outcomes would be reflective of the initial assessment of transplant eligibility and that the original implant intent would be dynamic over time.
The Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) database, funded by the National Heart Lung and Blood Institute, is a registry for patients who receive durable, Food and Drug Administration–approved MCS for the treatment of advanced heart failure (10). This analysis included all adult (≥19 years of age) patients prospectively enrolled in INTERMACS who received a primary continuous flow left ventricular assist device (LVAD) as either BTT or DT between June 2006 and March 2011 (Fig. 1). Patients receiving biventricular support or a total artificial heart were excluded, because the purpose of the analysis was to study continuous flow devices that could potentially be implanted both as BTT and DT. Patients with an LVAD who were implanted as bridge to recovery, as rescue therapy, or for reasons other than BTT or DT were also excluded. The date of last follow-up for all patients was March 31, 2012.
Comprehensive clinical and demographic data were collected at the time of initial implantation as well as serially for each participant in the INTERMACS database, as previously described (11–13).
The INTERMACS registry used specific, predefined classifications for device implant strategy that was prospectively assigned before implant. Implant strategy was determined by each center at the time of implantation. Patients were classified according to their initial implant strategy for the primary analysis.
The BTT patients were defined as those already listed for cardiac transplantation at the time of device implant or listed within 24 h after implant. Patients receiving an LVAD who were not listed for cardiac transplantation within 24 h were defined as BTC. The BTC strategy was further subdivided into 3 levels of the likelihood of eventual transplantation, as determined by the implanting center: those likely to be eligible (BTC: likely), moderately likely to be eligible (BTC: moderate), and those unlikely to be eligible (BTC: unlikely). After implant, if a patient was no longer considered to be a transplant candidate by the center, their strategy could be changed to DT in INTERMACS. Device intent was reassessed at 6, 12, and 24 months if the patient remained on support.
Baseline demographic data and clinical features were stratified by device strategy (BTT, BTC, or DT) and compared with analysis of variance test for continuous variables and chi-square for categorical variables. Primary analysis evaluated competing outcomes according to initial implant strategy through 24 months, comparing BTT patients with BTC and DT patients. Competing outcomes methodology was used to estimate the time related simultaneous probability of a patient moving into 3 mutually exclusive terminal events. The events were death (before transplant), transplantation, and explant of device due to recovery. This method is an adaptation of Kaplan-Meier estimation where the product limit calculation is used to increment the time related proportion of patients who “flow” into each terminal state (14).
Pre-implant clinical features
Between June 2006 and March 2011, a total of 2,775 primary LVAD recipients met the inclusion criteria with 37.6% (n = 1,060) as BTT, 41.8% (n = 1,162) as BTC (of which: 68.5% [n = 796] BTC: likely; 24.3% [n = 282] BTC: moderate; 7.2% [n = 84] BTC: unlikely), and 19.6% (n = 553) as DT. Baseline clinical features according to initial implant strategy are summarized in Table 1.
Overall 80% of the entire cohort of patients were on inotropes, with only 14% of patients being INTERMACS profile 1, whereas 68% were profile 2 or 3. The DT patients were less likely to be profile 1 and more likely to be profiles 3 and 4. Comorbidities that might adversely affect transplant candidacy—such as obesity, diabetes, carotid disease, peripheral vascular disease, cancer, and chronic obstructive pulmonary disease—and social factors such as drug and alcohol abuse became more common as patients were judged less likely to be transplanted. Laboratory data indicative of more advanced heart failure, such as a higher creatinine and lower prealbumin, also correlated with implant intent. However, serum sodium, liver function tests and international normalized ratio, hemoglobin, and B-type natriuretic peptide, which also reflect disease severity, were not significantly different across the groups. Pre-implant hemodynamics were similar across the groups with the exception of a trend toward a higher right atrial pressure in the BTC and DT groups. There was not a consistent worsening of echocardiographic features, such as left and right ventricular function or degree of valvular regurgitation for those deemed less likely to be transplanted. The BTC group had the highest use of intra-aortic balloon pump and mechanical ventilation and the lowest use of beta-blockade, likely reflective of a sicker patient population.
Survival after implant
Two-year survival (alive on support, transplant, or recovered) after primary LVAD implant was 77.7% for BTT, 70.1% for BTC, and 60.7% for DT; p < 0.0001 (Fig. 2A). The 2-year survival for the BTC patients were analyzed separately: 73.7% for BTC: likely; 62.8% for BTC: moderate; and 62.9%% for BTC: unlikely; p < 0.0001 (Fig. 2B).
Competing outcomes after MCS
Several mutually exclusive outcomes are possible after primary LVAD implant. These include cardiac transplantation, death before transplant, device explant after ventricular recovery, and continued mechanical support. At 6 months there was already a dramatic difference in the percentage of patients transplanted across groups when categorized by their original implant strategy, with over 25% of the BTT group and approximately 1% of the DT group receiving a transplant (Fig. 3A).
By 12 months transplant was still the most common and death the least common in the BTT group; there was a progressive decline in transplant rates and an increase in mortality rates across the other categories (Fig. 3B). However, 68% of the BTC: unlikely group were still on device support at 1 year, comparable to 70% for those whose initial intent was DT.
At 24 months the relative rates of transplant and death across the initial implant strategies were similar to those seen at 6 and 12 months (Fig. 3C). At 2 years nearly 30% of BTT patients and between one-third and one-half of BTC patients remained on support. However, only 6% of the DT population had received a transplant.
Change in device strategy
Device intent was not only determined at baseline but reassessed at 3, 6, 12, and 24 months if the patient remained on support. The change in device strategy over time for those still on support by the original implant strategy is seen in Figures 4A to 4E. By 12 months, 25.8% of patients implanted as BTT and still on support were no longer considered BTT; the vast majority of these patients were reclassified as BTC. By 2 years the percentage of BTT patients no longer listed had increased to 43.5%. For those implanted as DT and still on support at 12 months, 14.6% had been listed for transplant or deemed potentially eligible for transplantation. Of the patients in the BTC: likely group, 45.6% and 32.7% at 1 and 2 years, respectively, were listed for transplant, whereas at the same time points 11.1% and 23.4% had their strategy changed to DT. For the BTC: moderate and BTC: unlikely groups who were still on support at 12 months, 24.1% and 39.1%, respectively, had been changed to a strategy of DT.
Trends in strategic intent over time
Before the approval of continuous flow devices as DT, almost all of the patients had received implantation as BTT or BTC (Fig. 5). However, within a year of approval (January 2010) of a continuous flow device as DT, approximately one-third of patients who received an LVAD were implanted as DT. Interestingly, the availability of a continuous flow device as DT corresponded to a decrease in patients receiving an LVAD as BTT from over 50% to <25%, whereas the percentage of those who had an LVAD as BTC remained essentially unchanged.
In this study of primary continuous flow LVAD recipients from the INTERMACS registry, many of the hemodynamic and echocardiographic characteristics of the patients were similar across BTT and BTC strategies, but comorbidities and social risks were more common as the perceived likelihood of transplant decreased. Survival after LVAD was significantly better for BTT subjects compared with BTC and DT. The outcomes of the BTT patients were nearly equivalent to the BTC: likely group, and the BTC: moderate and BTC: unlikely groups had similar 2-year outcomes to the DT group. Competing outcomes after LVAD implant also demonstrate an increasing rate of transplant and a decreasing rate of survival across the pre-implant likelihood of transplantation. Lastly, and most importantly, strategic intent changed over time. Those who were still supported at 12 and 24 months and who had an initial implant strategy of BTT or BTC: likely had the highest proportion of patients who were still BTT or considered likely to receive a transplant; whereas 24% of the BTC: moderate and 39% of the BTC: unlikely groups still on support at 12 months were reassigned to a strategy of DT. In contrast, approximately 14.6% of patients initially assigned to DT who were on support at 12 months were being considered for transplantation.
The categorization of strategic intent grew from the evolution of device technology. Originally, devices were designed as short- to medium-term support to bridge patients for several months to perhaps 1 year while awaiting transplant and thus received approval as a BTT. When MCS was applied to the transplant ineligible it was considered a new application in a newly defined population; hence DT was considered a separate indication by the Food and Drug Administration and subsequently by Centers for Medicare and Medicaid Services and other payers. The current generation of continuous flow devices offers improved survival and lower rates of adverse events with superior mechanical longevity. These factors, in turn, have led to an evolution in the clinical application of MCS, with earlier implantation in a broader range of patients.
Our data highlight the artificial dichotomy of the currently accepted implant strategies of BTT and DT, which are increasingly less representative of the clinical circumstances in which MCS is used. Over 40% of patients are not clearly transplant candidates (BTT) or DT but rather receive implantation as a BTC. Additionally, we have shown that the initial implant intent is dynamic, with some patients becoming more likely to be transplanted and others becoming less likely to be transplanted or changed to a strategy of DT.
In contrast to the regulatory dichotomization of BTT as short-term support and DT as long-term support, 45% of the BTT patients were still on support at 1 year, and 28% were still on support at 2 years. Thus, BTT is often long-term support. Although rates of transplantation and death might be influenced by device type, a substantial proportion of patients in bridge trials are still supported beyond 1 year. In the HeartMate II BTT trial, patients had to be listed 1A or 1B to be included, but 20.6% of patients were still on support at 18 months (7). The long duration of support for many patients in the BTT and BTC: likely group is not necessarily attributable to transplant ineligibility due to development of contraindications after MCS or the failure of pre-existing contraindications to resolve with MCS. When device strategy over time was examined, 74.2% of BTT patients were still BTT at 1 year, and only 6.3% had become DT. For those still on support at 1 year in the BTC: likely group, 67% of patients were still BTC: likely or were actively listed for transplant, whereas the rate of DT in this group was 11.1%. In contrast, by 1 year 24.1% of those still supported in the BTC: moderate group had already changed to a strategy of DT.
Our data strongly suggest that the population of patients who receive MCS have a continuum of risk as reflected by their pre-implant strategies. Overall, the BTC patients represent the transition in risk and outcomes between the BTT and DT populations. Furthermore the initial implant strategy is dynamic in clinical practice and often changes over time. Although many patients who have received implantation as BTT are supported for a short time and subsequently receive transplantation, many other patients are supported for long periods of time and might even change strategies to DT. The BTC population might have lower rates of transplant, but they are more likely to change implant strategies over time compared with BTT or DT. These data challenge the currently accepted regulatory and reimbursement paradigm of device indication as either BTT or DT and argues strongly for a change in policy with regard to device indication and approval. Ideally the indication for MCS should be on the basis not of transplant candidacy but rather of end-stage heart failure and the need for mechanical support, recognizing that a substantial proportion would receive transplant in the short- to medium-term, whereas the remainder would receive long-term support regardless of their eventual transplant eligibility. If the ultimate goal of MCS is to provide long-term patient survival with good quality of life free from significant adverse events, this might be best assessed by a trial of patients who require mechanical support as opposed to separate BTT and DT trials. Such a unified trial design would also avoid the lag in the introduction of new technologies inherent in the need to complete separate BTT and DT trials; however, it would require not only an openness on the part of regulatory agencies to consider such a trial design but also a willingness on the part of payers to reimburse if devices were approved in such a trial.
The INTERMACS registry only includes patients from centers that participate in the registry and who consent to their data being entered; thus patients who were implanted at nonparticipating centers or those who were implanted emergently and for whom consent could not be obtained are not part of the database. However, because centers are mandated to be part of the INTERMACS database to be reimbursed for DT implants, the INTERMACS registry does capture the vast majority of all MCS implanted in the United States. Determination of strategic intent was made at the discretion of the implanting center rather than upon standardized criteria, and not all patients had a strategy assigned at the time of follow-up. The patient-specific reasons for a given strategic intent, such as the level of sensitization, were not available, nor were the reasons for the change in intent over time or data on patient preferences. Continuous flow devices were commercially available for both BTT and DT only after January 2010, which might have affected implant strategy. Change in strategy over time might also be influenced by center-specific criteria for transplantation and differential access to transplantation.
Although most LVADs are intended as a bridge to transplant, nearly one-half of such recipients receive implantation as a BTC. However, patients who receive an LVAD and are listed at the time of implant have superior survival to those who are BTC or DT. Factors that negatively impact transplant candidacy were more common in those who were deemed less likely to receive transplantation at the time of implant. Despite being listed at the time of implant, nearly 30% of patients receiving implantation as BTT were still on support at 2 years. Furthermore the strategic intent remains fluid over time; at 2 years 43.5% of BTT who were still on support were no longer listed for transplant. These results challenge the current paradigm of implant strategy as either BTT or DT.
The authors would like to acknowledge all of the centers, investigators, coordinators, and patients who are part of the Interagency Registry for Mechanically Assisted Circulatory Support.
The Interagency Registry for Mechanically Assisted Circulatory Support is supported by U.S. Department of Health and Human Services/National Institutes of Health/National Heart, Lung, and Blood Institute Grant HHSN268201100025C. Dr. Teuteberg has served as consultant to Sunshine Heart; and has a relationship with HeartWare. Dr. Jessup has received research support from Thoratec and HeartWare. Dr. Sun has served as consultant to Thoratec and Sunshine Heart. Dr. Naftel has received statistical support from Thoratec and Berlin Heart. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- bridge to candidacy
- bridge to transplant
- destination therapy
- Interagency Registry for Mechanically Assisted Circulatory Support
- left ventricular assist device
- mechanical circulatory support
- Received May 13, 2013.
- Accepted May 21, 2013.
- American College of Cardiology Foundation
- Kirklin J.K.,
- Naftel D.C.
- Lietz K.,
- Miller L.W.
- Mancini D.,
- Lietz K.
- Lietz K.,
- Long J.W.,
- Kfoury A.G.,
- et al.
- Felker G.M.,
- Rogers J.G.
- Deng M.C.,
- Loebe M.,
- El-Banayosy A.,
- et al.
- ↵Interagency Registry for Mechanically Assisted Circulatory Support. Available at: http://www.intermacs.org. Accessed on May 1, 2013