Abstract
Background:
Continuous-flow mechanical circulatory support (CF-MCS) is associated with impaired vascular function and increased risk of vasoplegia. One contributing factor to early graft failure (EGF) is severe vasoplegia. We tested the hypothesis that CF-MCS is associated with increased risk of EGF.
Methods:
Adult primary heart transplant recipients in the ISHLT Registry from 2005–2013 were stratified into 3 groups based on pre-transplant MCS: No MCS (n=11,748), pulsatile (P)-MCS (n=718), and CF-MCS (n=3,818). EGF was defined as death/retransplantation due to graft failure within 30 days after HT. Comparisons were made using descriptive statistics and associations. EGF was assessed with multivariable Cox proportional hazard regression.
Results:
The incidence of EGF within 30 days was similar between groups (No MCS 2.2%, P-MCS 3.3%, CF-MCS 2.1%, p=0.10). Following multivariable adjustment the risk of EGF was not statistically different for those with CF-MCS compared with P-MCS (HR 0.75, 95% CI 0.46–1.21, p=0.24). The risk of EGF was numerically, but not statistically significantly higher for CF-MCS compared with No MCS (HR 1.24, 95% CI 0.92–1.67, p=0.16).
Conclusion:
CF-MCS use was not associated with a statistically significant increased risk of EGF resulting in death or retransplantation in the first 30 days after transplant.
Keywords: Mechanical circulatory support, heart transplantation, early graft failure
Introduction:
Mechanical circulatory support (MCS) has become a major treatment for advanced heart failure, improving patient quality of life and survival, and is now used as a bridge in more than 50% of heart transplant recipients.[1] Early generation pulsatile MCS devices (P-MCS) preserved arterial pulsatility; however their limited durability and adverse event profile limited use. Continuous-flow MCS devices (CF-MCS) have improved pump design, resulting in improved device durability and patient survival. However, their conversion of circulation from pulsatile to continuous flow reduces endothelial cyclic strain and shear stress. On the cellular level, the loss of pulsatility results in altered transcriptional regulation of genes & translation of proteins related to anti-oxidant responses,[2] impaired endothelial nitric oxide production,[3, 4] increased matrix degradation,[5] and decreased smooth muscle content & activity.[6] Histologically, elastic fiber fragmentation, medial fibrosis, and aortic wall thinning are observed.[6, 7] Physiologically, axial CF-MCS leads to increased sympathetic nerve activity that declines with pulsatility[8, 9] and impairs the ability to vasoconstrict and raise blood pressure in response to norepinephrine and phenylephrine.[5, 10] Furthermore, compared to age-matched normal and to New York Heart Association (NYHA) class II-IV heart failure patients, those with CF-MCS have decreased peripheral vascular function,[11] while those with PF-MCS have improved vascular function.[12]
Primary graft dysfunction (PGD) is the leading cause of early death following transplantation, accounting for over 45% of death or retransplantation within the first 30 days.[13] The reported prevalence varies from 2.3% to 28.2%, with the variability in part a result of a lack of a standardized definition prior to a 2014 International Society for Heart and Lung Transplantation (ISHLT) consensus conference (Figure 1).[14] A number of risk factors for PGD have been identified;[13–16] one notable factor believed to contribute is a recipient inflammatory response leading to vasoplegia. As a result of limitations of ISHLT registry outcomes data, the ISHLT created the term early graft failure (EGF), the composite end-point of death or retransplantation associated with graft failure within the first 30 days after transplant, as a surrogate for PGD.[13] This study tested the hypothesis that bridge to transplant CF-MCS compared to P-MCS and no MCS is associated with increased risk of EGF.
Figure 1.
ISHLT 2014 primary graft dysfunction criteria. Adapted from Kobashigawa J, Zuckermann A, Macdonald P, et al. Report from a consensus conference on primary graft dysfunction after cardiac transplantation. The Journal of Heart and Lung Transplantation. 2014;33:327–340.
Patients and Methods:
We conducted a retrospective cohort study using data collected in the ISHLT Transplant Registry. All adult (age ≥18 years) primary heart transplant recipients from 2005–2013 were included, with follow-up through July 15th, 2016. Patients undergoing multi-organ transplantation, with an unknown MCS status, with temporary MCS support, or with right ventricular assist device or biventricular assist device support were excluded. Patients were stratified into 3 groups based on pre-transplant durable MCS: no MCS, P-MCS, and CF-MCS. The outcome was EGF. Studies using registry data have previously been deemed exempt from review by the Institutional Review Board of Columbia University Medical Center.
Statistical Analysis:
Demographic and clinical variables were summarized with standard descriptive statistics and expressed as median (with interquartile range). Group comparisons were made using the Chi-squared or Fisher’s exact test for categorical variables and the Kruskal-Wallis test (all group comparison with Bonferroni adjustments for multiple comparisons) or the Wilcoxon test (pairwise comparison) where appropriate for continuous variables. A Cox proportional hazard regression model was used to assess the relationship between MCS status and the outcome EGF within 30 days post-transplant. Continuous risk factors were included in the models using a restricted cubic spline, to allow for the most flexible fit of the functional form. In all models, missing values were imputed using multiple imputation. All factors in Supplemental Table 1 were considered for inclusion in the model. MCS status was retained in the final model irrespective of significance. For other variables in the model, a backward selection method with p≤0.10 was used to determine which to retain in the model. We tested the interaction between MCS group (no MCS, CF-MCS, and P-MCS) and select baseline recipient, donor, and transplant procedure characteristics identified as significant in the multivariable model.
Statistical analyses were performed using SAS Enterprise Guide 5.1 (SAS Institute, Inc., Cary, NC.) and R v0.99.486 (RStudio Team (2015). RStudio: Integrated Development for R. RStudio, Inc., Boston, MA URL http://www.rstudio.com). A 2-sided p-value of <0.05 was considered significant.
Results:
The study cohort was composed of 16,287 patients stratified into three groups based on durable MCS: no MCS (n=11,748, 72.1%), P-MCS (n=718, 4.4%), and CF-MCS (n=3,818, 23.4%). Between groups there was significant variation in pre-transplant recipient, donor, and transplant characteristics (Table 1 [selected covariates], Supplemental Table 1 [all covariates]). A greater proportion of patients with MCS were male (87.9% [P-MCS] vs. 81.1% [CF-MCS] vs. 72.7% [No MCS], p<0.0001), had an ischemic & non-ischemic cardiomyopathy etiology of their heart failure (96.3% [PF-MCS] vs. 95.9% [CF-MCS] vs. 84.7% [No MCS], p<0.0001), as opposed to other etiologies (e.g. congenital, restrictive, hypertrophic, valvular), had diabetes (32.2% [P-MCS] vs. 30.4% [CF-MCS] vs. 24.0% [No MCS], p<0.0001), had infections requiring IV antibiotics (24.9% [P-MCS] vs. 14.4% [CF-MCS] vs. 6.8% [No MCS], p<0.0001), and used amiodarone (37.4% [P-MCS] vs. 36.9% [CF-MCS] vs. 29.9% [No MCS], p<0.0001). More patients with CF-MCS did not require assistance with daily activities (44.7% [CF-MCS] vs. 34.0% [PF-MCS] vs. 31.0% [No MCS], p<0.0001), fewer patients were on a ventilator (0.9% [CF-MCS] vs. 2.1% [P-MCS] vs. 2.3% [No MCS], P<0.0001), and fewer patients were in the ICU (9.1% [CF-MCS] vs. 19.9% [P-MCS] vs. 34.4% [No MCS], <0.0001). Those with CF-MCS also had a longer duration of device support prior to transplant (228.5 days (45.0 – 839.0) vs. 149.5 days (22.5 – 513.0), p<0.0003) compared with P-MCS. Lastly, fewer patients with MCS were inotrope dependent (7.7% [CF-MCS] vs. 15.6% [P-MCS] vs. 52.4% [No MCS], p<0.0001) and more of those with MCS had prior cardiac surgery (excluding LVAD implantation, 72.0% [CF-MCS] vs. 65.6% [P-MCS] vs. 35.7% [No MCS], p<0.0001). Donors for patients with P-MCS were less likely to have an infection, were more likely to have required inotropes, and more commonly had head trauma as a cause of death.
Table 1.
Baseline Characteristics
| Recipient Gender | <0.0001 | |||
| Female | 3,208 (27.3%) | 722 (18.9%) | 87 (12.1%) | |
| Male | 8,543 (72.7%) | 3,096 (81.1%) | 631 (87.9%) | |
| Recipient Age (years) | 55.0 (26.0 – 68.0) | 56.0 (27.0 – 68.0) | 53.0 (29.0 – 66.0) | <0.0001 |
| Recipient Height (cm) | 172.7 (155.0 – 188.0) | 175.3 (157.5 – 190.5) | 177.8 (162.6 – 190.5) | <0.0001 |
| Recipient Weight (kg) | 78.3 (52.6 – 108.0) | 85.8 (59.0 – 115.8) | 88.9 (62.1 – 120.0) | <0.0001 |
| Recipient Cardiac Output (L/min) | 4.1 (2.3 – 6.9) | 4.6 (2.6 – 7.4) | 4.7 (2.4 – 7.5) | <0.0001 |
| Recipient Creatinine (mg/dl) | 1.2 (0.7 – 2.1) | 1.2 (0.7 – 2.0) | 1.2 (0.7 – 2.0) | 0.0004 |
| Recipient Bilirubin (mg/dl) | 0.8 (0.3 – 3.0) | 0.7 (0.3 – 2.2) | 0.7 (0.3 – 2.8) | <0.0001 |
| Recipient PASP (mmHg) | 42.0 (22.0 – 69.0) | 38.0 (20.0 – 65.0) | 40.0 (22.0 – 68.0) | <0.0001 |
| Recipient Diagnosis Grouping | <0.0001 | |||
| Ischemic CM | 4,648 (39.6%) | 1,647 (43.1%) | 327 (45.5%) | |
| Non-Ischemic CM | 5,293 (45.1%) | 2,016 (52.8%) | 365 (50.8%) | |
| Other | 1,797 (15.3%) | 155 (4.1%) | 26 (3.6%) | |
| Recipient History of Diabetes | 2,796 (24.0%) | 1,156 (30.4%) | 229 (32.2%) | <0.0001 |
| Recipient on IV Inotropes | 6,156 (52.4%) | 294 (7.7%) | 112 (15.6%) | <0.0001 |
| Recipient Amiodarone Use | 3,003 (29.9%) | 1,121 (36.9%) | 248 (37.4%) | <0.0001 |
| Recipient on Ventilator | 275 (2.3%) | 34 (0.9%) | 15 (2.1%) | <0.0001 |
| Recipient Functional Status | <0.0001 | |||
| No Assistance | 3,384 (31.0%) | 1,658 (44.7%) | 232 (34.0%) | |
| Some Assistance | 2,536 (23.2%) | 963 (26.0%) | 171 (25.1%) | |
| Total Assistance | 5,008 (45.8%) | 1,088 (29.3%) | 279 (40.9%) | |
| Infection Requiring IV Antibiotics | 758 (6.8%) | 535 (14.4%) | 174 (24.9%) | <0.0001 |
| Recipient Medical Condition | <0.0001 | |||
| In ICU | 3,842 (33.4%) | 348 (9.1%) | 143 (19.9%) | |
| Hospitalized, Not in ICU | 1,672 (14.5%) | 466 (12.2%) | 180 (25.1%) | |
| Not Hospitalized | 5,982 (52.0%) | 3,004 (78.7%) | 395 (55.0%) | |
| Recipient Time on Device (days) | N/A | 228.5 (45.0 – 839.0) | 149.5 (22.5 – 513.0) | <0.0001 |
| Recipient Prior Non-LVAD Cardiac Surgery | 4,153 (35.7%) | 2,736 (72.0%) | 470 (65.6%) | <0.0001 |
| Donor Age (years) | 30.0 (16.0 – 54.0) | 29.0 (17.0 – 51.0) | 28.0 (17.0 – 52.0) | 0.0007 |
| Donor Weight (kg) | 77.5 (55.0 – 109.0) | 82.0 (59.3 – 114.0) | 84.8 (63.0 – 113.0) | <0.0001 |
| Donor Inotrope Use | 5,764 (53.1%) | 1,843 (49.3%) | 395 (56.2%) | <0.0001 |
| Donor Clinical Infection | 5,506 (50.0%) | 2,273 (61.0%) | 253 (36.4%) | <0.0001 |
| Ischemic Time (hours) | 3.2 (1.6 – 5.0) | 3.3 (1.6 – 5.0) | 3.4 (1.7 – 5.2) | <0.0001 |
| Donor Cause of Death | <0.0001 | |||
| Anoxia | 1,785 (15.3%) | 723 (19.0%) | 80 (11.2%) | |
| Cerebrovascular/Stroke | 2,854 (24.5%) | 762 (20.0%) | 143 (20.0%) | |
| Head Trauma | 6,398 (54.8%) | 2,208 (57.9%) | 469 (65.5%) | |
| Other | 629 (5.4%) | 120 (3.1%) | 24 (3.4%) |
EGF occurred within 30 days in 2.2% of the entire cohort. Unadjusted analysis found that there was no statistically significant difference in the incidence of EGF between either CF-MCS or P-MCS (2.1% vs. 3.3%, p=0.14) or CF-MCS and No MCS (2.2% vs. 2.1%, p=1.0, Figure 2). Similarly, EGF did not significantly differ between P-MCS and No MCS (3.3% vs. 2.2%, p=0.11). Multivariable analysis including all covariates with a p-value of 0.10 or less failed to demonstrate an association between CF-MCS and EGF. When compared to P-MCS the risk of EGF was decreased, but not statistically significantly different for CF-MCS (HR 0.75, 95% CI 0.46–1.21, p=0.24, Supplemental Table 2). Similarly there was a 24% increased risk in EGF for those with CF-MCS compared to patients with No MCS, but this also failed to meet statistical significance (HR 1.24, 95% CI 0.92–1.67, p=0.16). There was an associated 66% increased risk of EGF for patients with P-MCS compared to those with No MCS (95% CI 1.04–2.63, p=0.03).
Figure 2.
Kaplan-Meier estimates for EGF rates within 30 days stratified by MCS group
Recognizing that management strategies of EGF may have evolved during the study period (e.g. increased use of percutaneous MCS and temporary surgical MCS), the impact of year on EGF was analyzed. Univariable analysis demonstrated that year was not associated with EGF (Supplemental Table 3). Further, when analyzed by era (2005–2009 vs. 2010–2013), there again was no association with EGF (Table 2).
Table 2.
Freedom From EGF by Era
| 2005–2009 | 2010–2013 | |||||
|---|---|---|---|---|---|---|
| Number at Risk | Freedom from EGF | 95% CI | Number at Risk | Freedom from EGF | 95% CI | |
| No MCS | 6561 | 97.7% | 97.3–98.0% | 4520 | 98.0% | 97.6–98.4% |
| P-MCS | 641 | 96.8% | 95.1–97.9% | 30 | 94.3% | 79.0–98.5% |
| CF-MCS | 875 | 97.0% | 95.6–97.9% | 2750 | 98.2% | 97.6–98.6% |
No-MCS=No mechanical circulatory support; P-MCS=Pulsatile mechanical circulatory support; CF-MCS=Continuous flow mechanical circulatory support
Rates were estimated using the Kaplan-Meier method and compared using log-rank test statistic and adjusted for multiple comparisons using the Scheffe’s method
Further exploratory sub-group analysis was performed. For the risk of EGF, there were no significant interactions between patients with a CF-LVAD and inotrope use, functional status, prior cardiac surgery, or amiodarone use. There was an interaction between patients located in the ICU and each group of patients. Further analysis was performed analyzing the differential risk of EGF based on a patient’s pre-transplant location, however again there was a numerical but non-statistically significant association between EGF and CF-MCS use was present (Supplemental Table 4).
Discussion:
Early graft failure represents the most severe form of primary graft dysfunction following heart transplantation. Through the comparison of patients with CF-MCS to patients with P-MCS or No MCS, the impact of no or limited pulsatility was compared to artificial pulsatility and the natural pulsatility of advanced heart failure. The principal finding of this study was that in unadjusted and adjusted analyses CF-MCS was not associated with a statistically significant increased risk of EGF compared to those with P-MCS or no MCS. However, in adjusted analysis, the risk was numerically greater.
There are several potential reasons that the hypothesis of this study was not confirmed by the results. First the hypothesis may be flawed and simply incorrect. PGD is a complex process involving early (within 24 hours) impairment of cardiac function and hemodynamics, with causes and mechanisms still unknown. Refractory vasodilation or vasoplegia is a component or contributing factor to PGD, however it does not uniformly occur in all patients supported by MCS nor does it occur in all cases of PGD. Thus it may be a necessary component in some cases but by itself is not sufficient. Additionally it may be that the hypothesis was flawed because the vascular and neurologic changes associated with continuous-flow reverse after heart transplantation. Previous work has demonstrated that the return of pulsatility following P-MCS implant improves vascular function (at four months post implant)[12] and also that the vascular dysfunction of advanced heart failure improves after heart transplantation.[17] Further, excessive sympathetic activation has been shown to recover (within minutes) with the return of pulsatility.[8]
The possibility exists that the study hypothesis may be true, but the effect was not captured by this study. EGF, as defined by ISHLT, only captures the most severe cases of PGD resulting in mortality or retransplant at 30 days and thus is a poor surrogate for PGD. Patients with moderate or severe PGF may recover. At the 2013 ISHLT PGD consensus conference 47 transplant centers reported the rate of PGD to be 7.4%[14], compared to a 2.2% prevalence of EGF in this study Further, the lack of difference between the groups in risk EGF may be related to improved post-transplant care, specifically with respect to the use of temporary MCS. While strategies vary by center[14], the predominant devices used in PGD are an intraaortic balloon pump for mild PGD and veno-arterial extracorporeal membrane oxygenation (ECMO) or temporary ventricular assist device (right ventricular only or biventricular) for more severe cases. Severe PGD requiring MCS has been reported to occur in 2.9%−16.3%[18–23] of all transplants and among these patients survival to hospital discharge ranged from 44.9%−81.5%% for ECMO[18–22] and 56%−58.8%% for biventricular MCS.[18, 20, 23] This raises the possibility that temporary MCS may help mitigate the risks associated with profound vasodilation and confound the endpoint of EGF. Finally, despite the large sample size from the worldwide ISHLT Registry, the number of events was small and the HR of 1.24 for CF-MCS vs. no MCS may have failed to reach statistical significance due to insufficient statistical power.
Limitations of this study include the retrospective nature of the study, limiting conclusions to association and not causation. The ISHLT registry that was used is of high-quality, however it is limited to the data collected and the precision of data collection and entry. As such, a number of covariates including post-transplant MCS use, inotrope use, and hemodynamic data that would have allowed us to study PGD were not available. Most notably, data regarding pre-transplant treatment with oral pulmonary vasodilators and neurohormonal antagonists were not available. Similarly there were missing data, though the use of multiple imputation in this scenario has a relative efficiency (how well the true population parameters are estimated) of over 99%.[24] Further, including patients with data missing not at random, improves the generalizability and external validity of the findings. In this study the baseline characteristics varied by group, however the fact that the findings were maintained in the multivariable model suggest that those differences did not significantly bias the conclusions.[25] Lastly, interaction testing requires greater statistical power than typical significance testing and it is possible that the ISHLT registry was underpowered for the interaction testing.
Conclusion:
In a contemporary cohort, early graft failure leading to death or retransplantation within 30 days of transplant is relatively rare, taking place in 2.2% of patients. Among these patients, CF-MCS use was not associated with a statistically significant increase in the risk of EGF.
Supplementary Material
Acknowledgements/Funding:
KJC is supported by National Institutes of Health Grant T32 HL007854. LHL is funded by the Swedish Research Council, The Swedish Heart and Lung Foundation, and Stockholm County Council.
Footnotes
Conflict of Interest:LHL has received research funding from Abbott.
References:
- 1.Lund LH, Edwards LB, Dipchand AI, et al. , The Registry of the International Society for Heart and Lung Transplantation: Thirty-third Adult Heart Transplantation Report-2016; Focus Theme: Primary Diagnostic Indications for Transplant. J Heart Lung Transplant, 2016. 35(10): p. 1158–1169. [DOI] [PubMed] [Google Scholar]
- 2.Patibandla PK, Rajasekaran NS, Shelar SB, et al. , Evaluation of the effect of diminished pulsatility as seen in continuous flow ventricular assist devices on arterial endothelial cell phenotype and function. J Heart Lung Transplant, 2016. 35(7): p. 930–2. [DOI] [PubMed] [Google Scholar]
- 3.Nakano T, Tominaga R, Morita S, et al. , Impacts of pulsatile systemic circulation on endothelium-derived nitric oxide release in anesthetized dogs. Ann Thorac Surg, 2001. 72(1): p. 156–62. [DOI] [PubMed] [Google Scholar]
- 4.Thacher T, Gambillara V, da Silva RF, et al. , Reduced cyclic stretch, endothelial dysfunction, and oxidative stress: an ex vivo model. Cardiovasc Pathol, 2010. 19(4): p. e91–8. [DOI] [PubMed] [Google Scholar]
- 5.Gambillara V, Thacher T, Silacci P, et al. , Effects of reduced cyclic stretch on vascular smooth muscle cell function of pig carotids perfused ex vivo. Am J Hypertens, 2008. 21(4): p. 425–31. [DOI] [PubMed] [Google Scholar]
- 6.Nishimura T, Tatsumi E, Takaichi S, et al. , Morphologic changes of the aortic wall due to reduced systemic pulse pressure in prolonged non pulsatile left heart bypass. Asaio j, 1997. 43(5): p. M691–5. [PubMed] [Google Scholar]
- 7.Segura AM, Gregoric I, Radovancevic R, et al. , Morphologic changes in the aortic wall media after support with a continuous-flow left ventricular assist device. The Journal of Heart and Lung Transplantation, 2013. 32(11): p. 1096–1100. [DOI] [PubMed] [Google Scholar]
- 8.Cornwell WK, Tarumi T, Stickford A, et al. , Restoration of Pulsatile Flow Reduces Sympathetic Nerve Activity Among Individuals With Continuous-Flow Left Ventricular Assist Devices. Circulation, 2015. 132(24): p. 2316–2322. [DOI] [PubMed] [Google Scholar]
- 9.Markham DW, Fu Q, Palmer MD, et al. , Sympathetic Neural and Hemodynamic Responses to Upright Tilt in Patients With Pulsatile and Nonpulsatile Left Ventricular Assist Devices. Circulation: Heart Failure, 2013. 6(2): p. 293–299. [DOI] [PubMed] [Google Scholar]
- 10.Nishinaka T, Tatsumi E, Nishimura T, et al. , Change in vasoconstrictive function during prolonged nonpulsatile left heart bypass. Artif Organs, 2001. 25(5): p. 371–5. [DOI] [PubMed] [Google Scholar]
- 11.Witman MAH, Garten RS, Gifford JR, et al. , Further Peripheral Vascular Dysfunction in Heart Failure Patients With a Continuous-Flow Left Ventricular Assist Device: The Role of Pulsatility. JACC: Heart Failure, 2015. 3(9): p. 703–711. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Amir O, Radovancevic B, Delgado Iii RM, et al. , Peripheral Vascular Reactivity in Patients With Pulsatile vs Axial Flow Left Ventricular Assist Device Support. The Journal of Heart and Lung Transplantation, 2006. 25(4): p. 391–394. [DOI] [PubMed] [Google Scholar]
- 13.Lund LH, Edwards LB, Kucheryavaya AY, et al. , The Registry of the International Society for Heart and Lung Transplantation: Thirty-second Official Adult Heart Transplantation Report-2015; Focus Theme: Early Graft Failure. The Journal of Heart and Lung Transplantation, 2015. 34(10): p. 1244–1254. [DOI] [PubMed] [Google Scholar]
- 14.Kobashigawa J, Zuckermann A, Macdonald P, et al. , Report from a consensus conference on primary graft dysfunction after cardiac transplantation. The Journal of Heart and Lung Transplantation, 2014. 33(4): p. 327–340. [DOI] [PubMed] [Google Scholar]
- 15.Segovia J, Cosío MDG, Barceló JM, et al. , RADIAL: A novel primary graft failure risk score in heart transplantation. The Journal of Heart and Lung Transplantation, 2011. 30(6): p. 644–651. [DOI] [PubMed] [Google Scholar]
- 16.Cooper LB, Mentz RJ, Edwards LB, et al. , Amiodarone use in patients listed for heart transplant is associated with increased 1-year post-transplant mortality. J Heart Lung Transplant, 2017. 36(2): p. 202–210. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Sansone R, Stanske B, Keymel S, et al. , Macrovascular and microvascular function after implantation of left ventricular assist devices in end-stage heart failure: Role of microparticles. The Journal of Heart and Lung Transplantation, 2015. 34(7): p. 921–932. [DOI] [PubMed] [Google Scholar]
- 18.Takeda K, Li B, Garan AR, et al. , Improved outcomes from extracorporeal membrane oxygenation versus ventricular assist device temporary support of primary graft dysfunction in heart transplant. J Heart Lung Transplant, 2016. [DOI] [PubMed] [Google Scholar]
- 19.Marasco SF, Vale M, Pellegrino V, et al. , Extracorporeal Membrane Oxygenation in Primary Graft Failure After Heart Transplantation. The Annals of Thoracic Surgery, 2010. 90(5): p. 1541–1546. [DOI] [PubMed] [Google Scholar]
- 20.Mihaljevic T, Jarrett CM, Gonzalez-Stawinski G, et al. , Mechanical circulatory support after heart transplantation(). European Journal of Cardio-Thoracic Surgery : Official Journal of the European Association for Cardio-thoracic Surgery, 2012. 41(1): p. 200–206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.D’Alessandro C, Aubert S, Golmard JL, et al. , Extra-corporeal membrane oxygenation temporary support for early graft failure after cardiac transplantation. European Journal of Cardio-Thoracic Surgery, 2010. 37(2): p. 343–349. [DOI] [PubMed] [Google Scholar]
- 22.Loforte A, Murana G, Cefarelli M, et al. , Role of Intra-Aortic Balloon Pump and Extracorporeal Membrane Oxygenation in Early Graft Failure After Cardiac Transplantation. Artificial Organs, 2016. 40(8): p. E136–E145. [DOI] [PubMed] [Google Scholar]
- 23.Thomas HL, Dronavalli VB, Parameshwar J, et al. , Incidence and outcome of Levitronix CentriMag support as rescue therapy for early cardiac allograft failure: a United Kingdom national study. European Journal of Cardio-Thoracic Surgery, 2011. 40(6): p. 1348–1354. [DOI] [PubMed] [Google Scholar]
- 24.Graham JW, Olchowski AE, and Gilreath TD, How Many Imputations are Really Needed? Some Practical Clarifications of Multiple Imputation Theory. Prevention Science, 2007. 8(3): p. 206–213. [DOI] [PubMed] [Google Scholar]
- 25.Elze MC, Gregson J, Baber U, et al. , Comparison of Propensity Score Methods and Covariate Adjustment. Evaluation in 4 Cardiovascular Studies, 2017. 69(3): p. 345–357. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.