Alternatives have emerged for patients ineligible for cardiac transplantation under standard criteria. The purpose of our study was to compare outcomes in patients ineligible for cardiac transplantation under standard criteria, treated either with extended criteria cardiac transplantation (ECCT) or a continuous flow destination therapy left ventricular assist device (CF DT-LVAD). From 2005 to 2012, patients treated with either ECCT or CF DT-LVAD at our institution were retrospectively analyzed. In the overall unmatched cohort, we examined mortality and other outcomes, including index hospitalization length of stay, renal function, stroke, and readmission rates. After propensity score (PS) matching, outcomes were compared between ECCT and CF DT-LVAD recipients. Overall, 62 patients underwent ECCT, and 146 patients were treated with CF DT-LVAD. The 2-year mortality estimate for ECCT recipients was 27.3% (95% confidence interval 15.5% to 39.1%) and for CF DT-LVAD recipients was 11.2% (95% confidence interval 4.8% to 17.6%). After PS matching of 39 patients from each treatment group, there was no significant difference in overall survival after 2 years (p = 0.346). In both unmatched and PS-matched analyses, CF DT-LVAD patients compared with ECCT had a significantly higher estimated glomerular filtration rate at 1 year but also had significantly higher hospital readmission rates. Stroke also more commonly occurred after CF DT-LVAD compared with ECCT (17 vs 5, unmatched; and 2 vs 1, PS matched). However, there was no significant difference between PS-matched groups in 2-year stroke-free survival (p = 0.371). In conclusion, ECCT and CF DT-LVAD in select patients are comparable therapies with respect to 2-year survival.
For patients with heart failure (HF) with symptoms refractory to maximal medical therapy (advanced or end-stage HF), cardiac transplantation under standard criteria (SCCT) is currently considered the most reliable therapy. However, SCCT is limited both by suitable donor availability and selective recipient characteristics. Alternative therapies have emerged for patients ineligible for SCCT. Extended criteria cardiac transplantation (ECCT) matches marginal donor organs that would otherwise go unused with recipients who have co-morbidities that would typically exclude them from SCCT. Another therapeutic option is mechanical circulatory support with a left ventricular assist device (LVAD) as destination therapy (DT). Since the advent of continuous flow (CF) devices, LVAD therapy has shown increasing reductions in morbidity and mortality for the patient with end-stage HF. Comparison of outcomes after ECCT and DT-LVAD has not been performed though since approval of CF devices for DT. Thus, there is limited information to provide patients who could be considered eligible for either therapy. Given our present experience with ECCT and the improved outcomes with CF devices, we hypothesized that the survival of patients ineligible for SCCT are similar with either ECCT or CF DT-LVAD. Additionally, we examined nonsurvival outcomes because of the unique differences in postoperative complications between ECCT and CF DT-LVAD.
Methods
We retrospectively evaluated patients with end-stage HF deemed ineligible for SCCT who underwent either ECCT or CF DT-LVAD from January 1, 2005, to December 31, 2012, at Duke University Medical Center. Determination of which therapy a patient received was dependent on patient and physician preference. No prospective protocol dictated their subsequent treatment. An extended-criteria protocol was developed and initiated at our center in January 2000. Patients were considered for ECCT if they were aged ≥65 years or they had significant co-morbidities including significant renal insufficiency, peripheral arterial disease (including carotid artery disease), or poorly controlled diabetes mellitus. Cytomegalovirus seropositivity is not considered a risk factor for consideration of ECCT. Patients who underwent ECCT signed informed consent for the transplant team to use organs that may be deemed unsuitable for use in SCCT recipients. Unfavorable characteristics of these hearts included older donors, single vessel coronary artery disease, LV hypertrophy, LV dysfunction, high-dose inotropic support, or donors considered high risk by the United Network for Organ Sharing. All the LVADs implanted were second-generation continuous, axial-flow devices (HeartMate II, Thoratec Corp., Pleasanton, California). Patients who had an existing first-generation, pulsatile LVAD were excluded from the analysis to isolate patients receiving de novo CF DT-LVAD or ECCT therapy. In addition, the number of patients listed for initial ECCT or SCCT who died while waiting for a suitable organ was retrospectively determined; these patients were also not included for the comparative analysis.
Baseline patient data were extracted from the medical record. In addition, we obtained details concerning perioperative use of an intra-aortic balloon pump (IABP), postoperative length of stay, hospital readmissions to Duke University Medical Center, renal function at 1 year, stroke events, and mortality. Survival analysis was supplemented using the Social Security National Death Index. Estimated glomerular filtration rate (eGFR) was calculated using the Modification of Diet in Renal Disease equation. Strokes were defined as a neurologic deficit accompanied by a brain imaging study showing a new infarction and further defined as either ischemic or hemorrhagic. United Network for Organ Sharing listing status, Interagency Registry for Mechanical Assisted Circulatory Support (INTERMACS) profile, and HeartMate II risk score were assigned to provide insight into clinical status and were determined on the basis of standard definitions or calculated as previously described. The Duke University Institutional Review Board approved the study design and protocol and waived the need for patient consent.
Outcomes of interest were 2-year overall survival, eGFR at 1 year, index hospitalization length of stay (days) and mortality, incidence and rate of stroke, 2-year stroke-free survival, and readmission rate per year of support. Overall survival was defined from time of ECCT or CF DT-LVAD to death or the last follow-up; stroke-free survival was defined from time of ECCT or CF DT-LVAD to death, stroke, or the last follow-up.
In the unmatched analysis, baseline characteristics and event rates are reported using medians with interquartile range for continuous variables and frequencies with proportions for categorical variables. We tested for differences in baseline variables between groups using chi-square tests for categorical variables and Wilcoxon rank sum for continuous variables. Kaplan–Meier estimates of survival, overall and stroke free, were computed for each group, with Wald 95% confidence intervals (CIs) for mortality calculated using Greenwood formula for standard error. The median length of stay and eGFR at 1 year were compared between ECCT and CF DT-LVAD groups using nonparametric Wilcoxon rank sum test. The proportion of stroke cases was estimated by group and the rate of stroke per 100 patient-years was estimated using exact Poisson CIs. Hospital readmission rate per year of support was compared between treatment groups using the generalized estimating equation method in a Poisson regression with an offset term for at risk (time to death or the last follow-up for readmission and time to death, stroke, or the last follow-up for stroke event). To account for bias in selection of patients to receive ECCT, a propensity score (PS) model was constructed using a multivariable logistic regression with treatment as the outcome and the following covariates: age, African-American race, creatinine, blood urea nitrogen, total bilirubin, albumin, and international normalized ratio. Covariates with a large rate of missing or those that were already balanced across treatment were excluded from the PS model. PSs were obtained from the predicted probabilities of performing CF DT-LVAD or ECCT, and using the resulting scores, patients receiving ECCT were matched in a 1:1 fashion with corresponding patients receiving CF DT-LVAD. Cramer phi and R 2 measures of association were used to assess covariate balance across treatment for categorical and continuous variables, respectively. Baseline characteristics, clinical status, and outcomes between patients receiving CF DT-LVAD and ECCT were reestimated for the matched cohort. We again tested for differences in baseline variables using chi-square tests for categorical variables and Wilcoxon rank sum for continuous variables. Lengths of stay and eGFR at 1 year were compared between PS-matched treatment groups using Wilcoxon signed-rank tests. The rates of hospital readmissions per year of support were also compared using the generalized estimating equation method in Poisson regression as detailed, accounting for matched pairs. Kaplan–Meier survival estimates of overall and stroke-free survival were compared using a paired analysis.
A 2-tailed p value <0.05 was considered statistically significant, and we report 95% CIs. All statistical analyses were performed using SAS version 9.3 (SAS Institute, Cary, North Carolina).
Results
During the study period, 62 patients underwent ECCT and 146 patients underwent CF DT-LVAD placement. Baseline and preoperative characteristics of the groups in the overall unmatched cohort are detailed in Table 1 . There were no significant differences in age or the numbers of male, black, or ischemic cardiomyopathy patients. Echocardiographic, hemodynamic, and laboratory data showed evidence of advanced HF and cardiogenic shock in both groups, with a suggestion of a greater severity of illness in the CF DT-LVAD group. Only 7 patients listed for ECCT (not included for comparative analysis) died while awaiting transplantation; overall 26 patients aged >18 years died on our center’s cardiac transplant wait list during our study period.
| Variable | Extended criteria cardiac transplantation (n = 62) | Continuous flow destination therapy left ventricular assist device (n = 146) | P value |
|---|---|---|---|
| Age (years) | 68 [66, 70] | 67 [59, 73] | 0.487 |
| Male | 48 (77%) | 108 (74%) | 0.600 |
| Black | 13 (21%) | 41 (28%) | 0.284 |
| Ischemic cardiomyopathy | 37 (60%) | 91 (62%) | 0.719 |
| Left ventricular ejection fraction (%; n = 199) | 20 [15, 25] | 17 [10, 20] | < 0.001 |
| Left ventricular end-diastolic dimension (cm, n = 183) | 6.3 [5.6, 7.0] | 6.7 [6.1, 7.3] | 0.009 |
| Serum creatinine (mg/dL) | 1.3 [1.1, 1.6] | 1.5 [1.1, 2.0] | 0.033 |
| Estimated glomerular filtration rate (mL/min/1.73 m 2 ) | 54 [43, 74] | 48 [34, 70] | 0.108 |
| Blood urea nitrogen (mg/dL) | 26 [20, 37] | 30 [20, 45] | 0.151 |
| Total bilirubin (g/dL) | 0.9 [0.6, 1.3] | 1.3 [0.9, 2.0] | < 0.001 |
| Albumin (g/dL; n = 200) | 3.5 [3.1, 3.9] | 3.1 [2.6, 3.5] | < 0.001 |
| International normalized ratio | 1.4 [1.1, 1.7] | 1.2 [1.1, 1.3] | < 0.001 |
| Pre-procedure inotrope (n = 174) | 19 (36%) | 53 (44%) | 0.327 |
| Cardiac index (L/min/m 2 ; n = 179) | 2.1 [1.8, 2.5] | 1.8 [1.5, 2.2] | < 0.001 |
| Pulmonary vascular resistance (Wood units; n = 174) | 2.8 [1.8, 3.9] | 3.3 [2.1, 4.2] | 0.052 |
| Mixed venous oxygen saturation (%; n = 177) | 55 [48, 61] | 51 [44, 57] | 0.007 |
Patients in each group were subclassified by the use of preoperative interventions (IABP) and by advanced HF severity classification schemes ( Table 2 ). In the CF DT-LVAD group, more patients required placement of a preoperative IABP for hemodynamic or end-organ function stabilization and were specifically classified into INTERMACS level 1 or 2, denoting failure of medical therapy or temporary mechanical assistance for cardiogenic shock. In contrast, nearly 1/3 of ECCT recipients (n = 11, 28%) were outpatients not requiring long-term inotropic support at the time a donor organ was identified.
| Variable | Extended criteria cardiac transplantation (n = 62) | Continuous flow destination therapy left ventricular assist device (n = 146) |
|---|---|---|
| Intra-aortic balloon pump | 12 (19%) | 46 (32%) |
| United Network for Organ Sharing listing status | ||
| 1A | 21 (34%) | – |
| 1B | 21 (34%) | – |
| 2 | 20 (32%) | – |
| Interagency Registry for Mechanical Assisted Circulatory Support profile | ||
| Not determined | – | 1 (0.1%) |
| 1 | – | 13 (9%) |
| 2 | – | 67 (46%) |
| 3 | – | 39 (27%) |
| 4 | – | 18 (12%) |
| 5 | – | 8 (6%) |
| Heartmate II risk score | ||
| Low or medium risk | – | 90 (62%) |
| High risk | – | 56 (38%) |
In the overall unmatched cohort, the 2-year mortality estimate for ECCT recipients was 27.3% (95% CI 15.5% to 39.1%) and for CF DT-LVAD recipients was 11.2% (95% CI 4.8% to 17.6%; Figure 1 ). In the CF DT-LVAD group, 9 patients died during index hospitalization for device implant, compared to no index hospitalization deaths in patients in the ECCT group. The inhospital mortality rate was 6.2% (95% CI 2.3% to 10.1%) for patients who underwent CF DT-LVAD and 0.0% (95% CI 0.0% to 5.8%) in the ECCT group.
After PS matching, comparative statistics were performed on 39 pairs of patients. Baseline characteristics of the groups are detailed in Table 3 and clinical status in Table 4 . The only significant difference in baseline characteristics between PS-matched groups was a lower LV ejection fraction in those patients receiving a CF DT-LVAD. Survival analysis in these groups showed a nearly identical postoperative survival at 2 years between groups ( Figure 2 , p = 0.346).
| Variable | Extended criteria cardiac transplantation (n = 39) | Continuous flow destination therapy left ventricular assist device (n = 39) | P value |
|---|---|---|---|
| Age (years) | 68 [67, 70] | 68 [60, 73] | 0.857 |
| Male | 29 (74%) | 28 (72%) | 0.799 |
| Black | 10 (26%) | 11 (28%) | 0.799 |
| Ischemic cardiomyopathy | 25 (64%) | 25 (64%) | 1.000 |
| Left ventricular ejection fraction (%; n = 72) | 20 [20, 25] | 15 [10, 20] | 0.001 |
| Left ventricular end-diastolic dimension (cm, n = 69) | 6.3 [5.6, 6.7] | 6.5 [6.2, 7.1] | 0.089 |
| Serum creatinine (mg/dL) | 1.4 [1.2, 1.7] | 1.3 [1.0, 1.7] | 0.748 |
| Estimated glomerular filtration rate (mL/min/1.73 m 2 ) | 50 [41, 60] | 51 [37, 75] | 0.589 |
| Blood urea nitrogen (mg/dL) | 29 [20, 42] | 28 [18, 38] | 0.988 |
| Total bilirubin (g/dL) | 0.9 [0.6, 1.4] | 1.0 [0.7, 1.4] | 0.372 |
| Albumin (g/dL) | 3.3 [3.0, 3.6] | 3.4 [2.9, 3.6] | 0.869 |
| International normalized ratio | 1.3 [1.1, 1.4] | 1.2 [1.1, 1.4] | 0.769 |
| Pre-procedure inotrope (n = 67) | 13 (38%) | 18 (55%) | 0.181 |
| Cardiac index (L/min/m 2 ; n = 68) | 2.0 [1.7, 2.5] | 2.0 [1.5, 2.2] | 0.122 |
| Pulmonary vascular resistance (Wood units; n = 66) | 2.9 [1.6, 4.2] | 3.2 [2.1, 3.9] | 0.316 |
| Mixed venous oxygen saturation (%; n = 67) | 55 [48, 61] | 52 [45, 57] | 0.337 |
| Variable | Extended criteria cardiac transplantation (n = 39) | Continuous flow destination therapy left ventricular assist device (n = 39) |
|---|---|---|
| Intra-aortic balloon pump | 12 (31%) | 11 (28%) |
| United Network for Organ Sharing listing status | ||
| 1A | 17 (44%) | – |
| 1B | 11 (28%) | – |
| 2 | 11 (28%) | – |
| Interagency Registry for Mechanical Assisted Circulatory Support profile | ||
| 1 | – | 3 (8%) |
| 2 | – | 18 (46%) |
| 3 | – | 13 (33%) |
| 4 | – | 2 (5%) |
| 5 | – | 3 (8%) |
| Heartmate II risk score | ||
| Low or medium risk | – | 31 (80%) |
| High risk | – | 8 (21%) |
