Secondary Logo

Journal Logo

Adult Circulatory Support

Survival and Functional Status After Bridge-to-Transplant with a Left Ventricular Assist Device

Suarez-Pierre, Alejandro; Zhou, Xun; Fraser, Charles D. III; Grimm, Joshua C.; Crawford, Todd C.; Lui, Cecillia; Valero, Vicente III; Choi, Chun W.; Higgins, Robert S.; Kilic, Ahmet

Author Information
doi: 10.1097/MAT.0000000000000874


The strategy of left ventricular assist device (LVAD) implantation as bridge-to-transplant (BTT) has become more common for patients with end-stage heart failure who are not immediately able to undergo heart transplantation.1 Initial reports, that focused on earlier generation LVADs, raised concerns about the impact that a device may have on survival following heart transplantation.2 The basis for this notion was that the duration of mechanical support increased rate of humoral sensitization3,4 as well as post-transplantation mortality.5 Technological advancements in combination with a better understanding and management of heart failure pathophysiology has allowed for better contemporary outcomes in patients receiving LVAD therapy. The duration of mechanical support does not have a direct effect of post-transplantation survival6 and there is no difference between two of the most commonly used devices.7 Continuous-flow pumps have been broadly used for the past decade with over 17,000 implants worldwide.8 The aim of this study was to examine long-term survival and functional status of patients who were BTT with an LVAD in comparison to patients undergoing de novo heart transplantation.

Materials and Methods

Patient Population and Data Sources

Adult patients (age ≥18 years) who underwent isolated heart transplantation between January 1, 2007, and September 30, 2017 were included in this study. We focused on patients bridged-to-transplant with an LVAD (Heartmate II [HMII] or HeartWare Ventricular Assist System [HVAD] only). This group was compared against patients undergoing de novo heart transplantation—defined as a patient without a prior sternotomy or LVAD implantation. Patients supported with extracorporeal or percutaneous LVADs, right ventricular assist devices, biventricular assist devices, and total artificial hearts were excluded. The Organ Procurement and Transplantation Network (OPTN) provided the publicly available Standard Transplant Analysis and Research files with follow-up available until December 1, 2017. These files comprise a prospectively collected dataset of all thoracic transplantation in the United States. This study was approved by the Johns Hopkins Medicine Institutional Review Board (reference IRB00159748).

Study End-Points and Variable Definitions

The primary outcome of interest was overall survival 1, 2, and 5 years after transplantation. Conditional survival estimates were examined to determine whether mortality hazards changed over time for either cohort. Survival was conditioned on 90 day survival (for 1 year survival estimates) and 1 year survival (for 2 and 5 year survival estimates). Secondary outcomes were functional status, return to work, hospital readmission, and graft rejection requiring treatment. Secondary outcomes were examined at the same intervals (1, 2, and 5 years). Functional status was characterized using the Karnofsky Performance Scale (KPS), which assigns a score ranging between 10% and 100%.9 Patients with a KPS score ≤40% are deemed to require complete assistance to manage daily tasks, those between 50% and 70% require partial assistance, and between 80% and 100% require no assistance. The KPS has been previously validated in heart transplantation.10 Postoperative outcomes were identified by crosslinking the index hospitalization to postoperative follow-up encounters for each subject. Data collection and variable definitions for the OPTN data set have been described previously.11,12

Data Presentation and Analysis

Continuous variables are presented as median [25th, 75th percentile] or mean ± standard deviation and categorical variables as count of patient (percentage). Descriptive comparisons were conducted with Kruskal-Wallis rank sum test or one-way analysis of variance for continuous variables and Pearson’s χ2 for categorical variables, as appropriate. Statistical significance was determined at a p value <0.05 (two sided). Analyses were performed using R version 3.4.313 with the following packages: rms,14survival,15 and survminer.16

Recipient-, donor-, and transplant-specific characteristics were compared between BTT and de novo cohorts. These characteristics were systematically collected at the time of transplant. Glomerular filtration rate was estimated using the Chronic Kidney Disease Epidemiology Collaboration equation17 with the pretransplant creatinine level and pertinent patient characteristics. The Model for End-Stage Liver Disease Excluding INR (MELD-XI) score was calculated as follows:

According to convention, serum creatinine was imputed to 4 mg/dl for patients with chronic renal failure requiring renal replacement therapy. The lower limit of serum creatinine and bilirubin was set to 1 mg/dl. The prognostic utility of the MELD-XI has been previously validated in heart transplantation.18

To examine differences on mortality, actuarial survival estimates were generated using the Kaplan-Meier method. Pairwise comparisons were made using the log-rank test and p values were adjusted with the Benjamini-Hochberg procedure.19 The impact of BTT on risk-adjusted mortality was examined using Cox proportional hazards models. Covariates for each model were only considered if they were collected by OPTN during the study period and selected through bootstrap aggregation.20 In brief, 250 samples were generated through random sampling with replacement. Each bootstrap sample containing 17,879 observations was used to construct a parsimonious model through automated stepwise regression and backward elimination. The retention criterion for covariates was set at a p value <0.05. The results of each bootstrap model were stored and counted. Covariates that entered >50% of the bootstrap models were chosen for the final model. Once covariates were selected, group identity (i.e., BTT or de novo) was forced into the model to evaluate independent association with risk-adjusted mortality. All covariates considered as candidates for model development are listed in Table 1 (Supplemental Digital Content, Separate models were developed for each individual outcome with this method. Estimates are presented as coefficients and hazard ratios (HR) with 95% CI. Strength of association between covariates and risk-adjusted mortality was determined using Wald’s χ2 statistic.


Study Population

Over the study period, 5,584 patients were successfully BTT using isolated left ventricular support and 12,295 received a de novo heart transplant. Within the BTT cohort, 75% (4,201/5,584) were bridged using a HMII device and 25% (1,383/5,584) with a HVAD device. Additionally, 56% (3,139/5,584) underwent LVAD implantation before registration for heart transplantation and 44% (2,445/5,584) underwent implantation while listed. Follow-up had a mean duration of 3.68 years ± 2.91 and a maximal duration of 10 years and 8 months.

Recipient, Donor, and Transplant Characteristics

There were notable differences between both recipient populations (see Table 2 Supplemental Digital Content, Bridge-to-transplant patients were more commonly male (82% vs. 70%; p < 0.001), African American (24% vs. 20%; p < 0.001), blood group O (46% vs. 34%; p < 0.001), had a higher body mass index (BMI; 28.4 vs. 26.3 kg/m2; p < 0.001), had a higher prevalence of tobacco use (52% vs. 44%; p < 0.001), and diabetes mellitus (32% vs. 27%; p < 0.001), and were less frequently hospitalized at the time of transplant (82% vs. 42%; p < 0.001). De novo patients had a higher MELD-XI score (12.5 vs. 12.1 points; p < 0.001), a lower estimated glomerular filtration rate (65 vs. 68 ml/min/1.73 m2; p < 0.001), a larger proportion required an intraaortic balloon pump at the time of transplant (10% vs. 1%; p < 0.001), and required care at an intensive care unit more frequently (41% vs. 7%; p < 0.001). The total waitlist time (220 vs. 56 days; p < 0.001) and status 1A time (31 vs. 2 days; p < 0.001) was longer for BTT patients.

A higher proportion of donors accepted for BTT patients were male (77% vs. 66%; p < 0.001), blood group O (55% vs. 48%; p < 0.001), and had slightly higher BMI (26.8 vs. 25.9 kg/m2; p < 0.001; see Table 3 Supplemental Digital Content, The median recipient–donor BMI ratio, utilized as a surrogate for size mismatch, was slightly higher for BTT patients (1.05 vs. 1.00; p < 0.001). A higher proportion of de novo patients received an ABO compatible organ (17% vs. 12%; p < 0.001) versus receiving a heart from an ABO identical donor. The median distance between donor hospital and transplant center was longer for the de novo group (150 vs. 139 km; p < 0.001). Yet, the median organ ischemic time was slightly longer for the BTT group (3.2 vs. 3.1 hours; p = 0.019).

Impact of Support on Survival

Unconditional survival estimates were lower for BTT patients at 1 year (90% vs. 91%; p < 0.001), 2 years (86% vs. 88%; p < 0.001), and 5 years (77% vs. 79%; p = 0.003) (Figure 1 and Table 1). One year survival, conditional on 90 day survival, was similar between groups (96% vs. 96%; p = 0.054). Two and 5 year survival, conditional on 1 year survival, were also similar between groups (Table 2). Bridge-to-transplant was independently associated with unconditional, risk-adjusted mortality at 1 (HR: 1.33, 95% CI: 1.19–1.49; Table 3 and see Table 4, Supplemental Digital Content, and 5 years (HR: 1.72, 95% CI: 1.57–1.87; Table 3 and see Table 5, Supplemental Digital Content, This association was also evident for conditional, risk-adjusted mortality at 1 (HR: 1.26, 95% CI: 1.04–1.51; see Table 6, Supplemental Digital Content, and 5 years (HR: 1.39, 95% CI: 1.22–1.59; see Table 7, Supplemental Digital Content,

Table 1.
Table 1.:
Unconditional Survival Estimates Stratified by Groups
Table 2.
Table 2.:
Conditional Survival Estimates Stratified by Groups
Table 3.
Table 3.:
Risk-Adjusted Mortality as a Function of Group
Figure 1.
Figure 1.:
Actuarial survival curves for bridge-to-transplant patients (blue) and de novo transplant patients (red). A: Unconditional survival at 5 years (log-rank p < 0.01). B: 5 year survival, conditional on 90 day survival (log-rank p = 0.15). C: 5 year survival, conditional on 1 year survival (log-rank p = 0.68).

The impact of LVAD implantation era and transplantation era on post-transplantation survival was examined by clustering patients by years of implantation (2005–2009, 2010–2013, and 2014–2017) or by year of transplantation (2007–2010, 2011–2013, and 2014–2017; Figure 2). Left ventricular assist device implantation in the 2005–2009 was associated with lower survival estimates at 1 (87.0%, 95% CI: 84.9%–89.2%) and 2 years (83.4%, 95% CI: 81.0%–85.7%) after transplantation than the 2010–2013 era (p = 0.014). Survival estimates were similar across all eras of transplantation (all p values >0.05; Figure 2B).

Figure 2.
Figure 2.:
Actuarial survival curves for bridge-to-transplant patients only. A: Patients are stratified based on year of left ventricular assist device implantation. Left ventricular assist device implantation between 2005 and 2009 had significantly lower 2 year survival (83.4%) than implantation between 2010 and 2013 (87.1%, log-rank p = 0.014). B: Patients are stratified by year of transplantation. Pairwise comparisons between the three eras demonstrated similar survival after heart transplantation. p values were adjusted with the Benjamini-Hochberg procedure.19

Secondary End-Points

Functional status was examined by stratifying performance according to the KPS categories described previously. At the time of listing, close to 40% required complete assistance in both groups and only 15% required no assistance. These proportions shifted by the time of transplantation with a larger proportion of BTT patients requiring no assistance (23% vs. 12%; p < 0.001) and a smaller proportion requiring complete assistance (33% vs. 58%; p < 0.001; Figure 3). From there on, functional status of both groups was similar. The proportion of patients requiring no assistance to manage daily tasks for BTT and de novo groups was 86% vs. 86%, 89% vs. 90%, and 87% vs. 90% at 1, 2, and 5 years, respectively.

Figure 3.
Figure 3.:
Functional status assessed through the Karnofsky Performance Scale. Bars represent the proportion of patients who belong to each functionality level cohort in relation to their need for assistance to manage daily tasks. *p < 0.001.

Employment rate was low at the time of listing, but relatively higher for de novo patients (13% vs. 10%; p < 0.001), and remained low at the time of transplantation for both groups (~8%; Figure 4). The proportion of patients returning to work 1 year after transplantation doubled and was relatively higher in the de novo group (23% vs. 19%; p = 0.004). Rates of employment continued to improve over time and were similar between groups at 2 (21%) and 5 years (26%).

Figure 4.
Figure 4.:
Employment status after heart transplantation. Bars represent the proportion of patient who worked for income at each time point. A higher proportion of de novo patients (23% vs. 19%) were employed 1 year after transplant. This difference was not observed again at 2 or 5 years. *p < 0.05.

Rates of hospital readmission were similar between groups at every time point. Readmission estimates at 1, 2, and 5 years were 40%, 55%, and 73%, respectively. These estimates only account for the interval between transplantation and the first subsequent hospital admission. Iterative events were not examined (Table 4). Rates of graft rejection requiring treatment were also similar between groups at every time point. Rejection estimates at 1, 2, and 5 years were 25%, 33%, and 41%, respectively. Iterative events were not examined.

Table 4.
Table 4.:
Time-to-Event Estimates for Hospital Readmission and Graft Rejection


Key Findings

Bridge-to-transplant using isolated left ventricular support with newer generation, continuous-flow pumps provides comparable long-term post-transplantation survival to de novo transplantation. Bridge-to-transplant incurs an up-front mortality that is 2% higher in the first 3 months from the surgical risk of redo sternotomy with device explantation; however, long-term survival conditional on 90 day survival is similar between groups. Hence, BTT patients who survive the early postoperative phase experience similar hazards of mortality over time to those of de novo patients.

Overall, patients had adequate post-transplantation functional performance with the majority of patients (>80%) requiring no assistance to manage daily tasks (Figure 3). Both groups had similar function at the time of listing but the de novo group suffered a decrease in functionality between registration and transplant, in contrast with the BTT group, which experienced an increase in functionality. The proportion of patients who were employed (Figure 4) decreased for de novo patients between the time of waitlist registration (13%) and transplantation (7%), while the proportion of BTT patients who remained employed did not suffer a similar decrease. These findings serve as indirect evidence for superior rehabilitation and conditioning in patients receiving left ventricular support while waiting for transplantation. It is congruent with the small fraction of BTT patients who required care within a hospital or intensive care unit, preferable preoperative hepato-renal function (examined with estimated glomerular filtration rate and MELD-XI score), and being able to survive longer periods waiting for a heart (see Table 2, Supplemental Digital Content,

Left ventricular assist device implantation during the early era (2005–2009), during which clinical trials for these devices were being conducted, was associated to a 3.6% decrease in 2 year survival (Figure 2A). Perhaps within this era, transplantation was more frequently done in the setting of pump complications and at the outset of widespread use of continuous-flow LVAD’s the outcomes of BTT improved from optimized management and device familiarity.

Prolonged mechanical support with LVAD has been associated with an increased rate of humoral sensitization in observational studies from previous decades.3,4 However, in neither of these has BTT with an LVAD been associated with an increased rate of graft rejection requiring treatment. Arnaoutakis et al.4 described an association with a higher rate of primary graft dysfunction with no negative effects on survival. In the current study, there was no difference in the rates of graft rejection requiring treatment nor hospital readmission at 1, 2, or 5 years between de novo and BTT patients (Table 4). Future studies should examine the extent of humoral sensitization with current generation devices.

Risks and Benefits of Mechanical Support

The use of LVAD as BTT represents a cost-effective alternative to “non-bridge” heart transplantation for intermediate- and high-risk patients,21 particularly for those who are young, have prolonged expected waiting times, or renal dysfunction.22 A comparative, cost-effectiveness analysis estimated that BTT-LVADs offer 3.8 additional life-years to patients who are expected to wait for longer than 6 months.23

Continuous-flow devices provide increased durability and lower rates of infections than those of earlier generations. However, offering BTT with LVAD poses significant challenges to patients who are constantly enduring longer periods of mechanical support. Stroke continues to be the major cause of death for patients on LVAD, closely followed by infection, which is the most common adverse event.8 The hemodynamics of continuous-flow devices have been associated to acquired von Willebrand syndrome24 and aortic insufficiency.25 Additionally, prolonged support is associated to increased risk for gastrointestinal bleeding, pump thrombosis,26 and adverse neurologic events.27 An observational study that compared bridging strategies (LVAD versus inotropic agents) concluded that patients bridged with inotropes who decompensated and required subsequent LVAD implantation acquired a twofold increase in mortality.28 Previous observational studies compared both devices and found no significant differences in pretransplantation end-organ function or post-transplantation patient survival.7,29 Selecting an appropriate bridging strategy for patients who are unlikely to undergo expeditious transplantation requires thoughtful consideration.


The Food and Drug Administration approved HMII and HVAD devices as BTT therapies in 2008 and 2012, respectively. The study period incorporates patients enrolled in regulatory clinical trials and patients who underwent commercial implantation; we could not identify nor adjust for this factor. The volume of LVADs implanted at each center on a yearly basis, a fitting surrogate for center experience, was not available within in the OPTN dataset. National databases may experience inherent variability in data collection. The fields used in these analyses had a low proportion of missing values (all <10%); furthermore, those used for regression modeling had <1% of missing values. This study focused solely on patients who were successfully bridged-to-transplant and these findings should not be extrapolated to BTT patients waiting for a heart. The physiology and clinical management of de novo and BTT patients are categorically different; hence, certain factors may impart different effects to patient outcomes. Future studies should examine the interactions between these groups and predictors for post-transplantation survival.


Bridge-to-transplant with LVAD provides excellent survival and similar quality of life to that of patients undergoing de novo heart transplantation. Bridge-to-transplant patients experience a slightly higher mortality rate within 90 days of transplantation. This difference in survival is offset by the live-years gained during mechanical support. BTT with LVADs may provide a means for pretransplantation rehabilitation and conditioning for otherwise moribund patients.


1. Uriel N, Jorde UP, Woo Pak S, et al. Impact of long term left ventricular assist device therapy on donor allocation in cardiac transplantation. J Heart Lung Transplant 2013.32: 188–195.
2. Patlolla V, Patten RD, Denofrio D, Konstam MA, Krishnamani R. The effect of ventricular assist devices on post-transplant mortality an analysis of the United network for organ sharing thoracic registry. J Am Coll Cardiol 2009.53: 264–271.
3. Joyce DL, Southard RE, Torre-Amione G, Noon GP, Land GA, Loebe M. Impact of left venticular assist device (LVAD)-mediated humoral sensitization on post-transplant outcomes J Hear Lung Transpl 2005.24: 2054–2059.
4. Arnaoutakis GJ, George TJ, Kilic A, et al. Effect of sensitization in US heart transplant recipients bridged with a ventricular assist device: Update in a modern cohort J Thorac Cardiovasc Surg 2011.142: 1236–1245.e1.
5. Gammie JS, Edwards LB, Griffith BP, Pierson RN 3rd, Tsao L. Optimal timing of cardiac transplantation after ventricular assist device implantation. J Thorac Cardiovasc Surg 2004.127: 1789–1799.
6. Grimm JC, Magruder JT, Crawford TC, et al. Duration of left ventricular assist device support does not impact survival after US heart transplantation. Ann Thorac Surg 2016.102: 1206–1212.
7. Magruder JT, Grimm JC, Crawford TC, et al. Survival after orthotopic heart transplantation in patients undergoing bridge to transplantation with the HeartWare HVAD versus the Heartmate II. Ann Thorac Surg 2017.103: 1505–1511.
8. Kirklin JK, Pagani FD, Kormos RL, et al. Eighth annual INTERMACS report: Special focus on framing the impact of adverse events. J Heart Lung Transplant 2017.36: 1080–1086.
9. Schag CC, Heinrich RL, Ganz PA. Karnofsky performance status revisited: Reliability, validity, and guidelines. J Clin Oncol 1984.2: 187–193.
10. Kilic A, Conte JV, Baumgartner WA, Russell SD, Merlo CA, Shah AS. Does recipient age impact functional outcomes of orthotopic heart transplantation? Ann Thorac Surg 2014.97: 1636–1642.
11. UNOS: UNOS Transplant Pro - Data Collection Forms, 2017. Available at: Accessed January 10, 2018.
12. Leppke S, Leighton T, Zaun D, et al. Scientific Registry of Transplant Recipients: Collecting, analyzing, and reporting data on transplantation in the United States. Transplant Rev (Orlando) 2013.27: 50–56.
13. R Core Team: R: A language and environment for statistical computing, 2017.
14. Harrell F. Regression Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis. 2017.2nd ed. New York, NY, Springer.
15. Therneau T. A Package for Survival Analyis in S, 2015.
16. Kassambra A, Kosnski M. survminer: Drawing Survival Curves using “ggplot2,”2017.
17. Levey AS, Stevens LA, Schmid CH, et al.; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration): A new equation to estimate glomerular filtration rate. Ann Intern Med 2009.150: 604–612.
18. Grimm JC, Shah AS, Magruder JT, et al. MELD-XI score predicts early mortality in patients after heart transplantation. Ann Thorac Surg 2015.100: 1737–1743.
19. Benjamini Y, Hochberg Y. Controlling the false discovery rate: A practical and powerful approach to multiple testing. J R Stat Soc 1995.57: 289–300.
20. Breiman L. Bagging predictors Mach Learn 1996.24: 123–140.
21. Alba AC, Alba LF, Delgado DH, Rao V, Ross HJ, Goeree R. Cost-effectiveness of ventricular assist device therapy as a bridge to transplantation compared with nonbridged cardiac recipients. Circulation 2013.127: 2424–2435.
22. Seco M, Zhao DF, Byrom MJ, et al. Long-term prognosis and cost-effectiveness of left ventricular assist device as bridge to transplantation: A systematic review. Int J Cardiol 2017.235: 22–32.
23. Long EF, Swain GW, Mangi AA. Comparative survival and cost-effectiveness of advanced therapies for end-stage heart failure. Circ Heart Fail 2014.7: 470–478.
24. Uriel N, Pak SW, Jorde UP, et al. Acquired von Willebrand syndrome after continuous-flow mechanical device support contributes to a high prevalence of bleeding during long-term support and at the time of transplantation. J Am Coll Cardiol 2010.56: 1207–1213.
25. Pak SW, Uriel N, Takayama H, et al. Prevalence of de novo aortic insufficiency during long-term support with left ventricular assist devices. J Heart Lung Transplant 2010.29: 1172–1176.
26. Maltais S, Anwer LA, Haglund NA, et al. Temporal differences in outcomes during long-term mechanical circulatory support. J Card Fail 2017.23: 852–858.
27. Coffin ST, Haglund NA, Davis ME, et al.; Mechanical Circulatory Support Research Network: Adverse neurologic events in patients bridged with long-term mechanical circulatory support: A device-specific comparative analysis. J Heart Lung Transplant 2015.34: 1578–1585.
28. Wozniak CJ, Stehlik J, Baird BC, et al. Ventricular assist devices or inotropic agents in status 1A patients? Survival analysis of the United Network of Organ Sharing database. Ann Thorac Surg 2014.97: 1364–1371; discussion 1371.
29. Topkara VK, O’Neill JK, Carlisle A, Novak E, Silvestry SC, Ewald GA. HeartWare and HeartMate II left ventricular assist devices as bridge to transplantation: A comparative analysis. Ann Thorac Surg 2014.97: 506–512.

ventricular assist device; bridge-to-transplant; heart failure; heart transplantation

Supplemental Digital Content

Copyright © 2018 by the ASAIO