Predictors and Outcome of Extracorporeal Life Support After Pediatric Heart Transplantation Jacob Simmonds, MD, Troy Dominguez, MD, Joanna Longman, MBBS, Nitin Shastri, MS, Maura O’Callaghan, MS, Aparna Hoskote, MD, Matthew Fenton, MBBS, Michael Burch, MD, Victor Tsang, MD, and Kate Brown, MPH
CONGENITAL HEART
Cardiac Unit, Great Ormond Street Hospital, London, United Kingdom
Background. Extracorporeal life support (ECLS) has proven success after conventional cardiac surgery. Its use after pediatric heart transplantation is less well documented. We reviewed ECLS after pediatric heart transplantation, to understand better predisposing factors, morbidity, and mortality. Methods. The notes of all patients at Great Ormond Street Hospital undergoing orthotopic heart transplantation from 1999 to 2009 were reviewed (202 transplants; patients aged 0.06 to 17.91 years). Patients were grouped by diagnosis: restrictive cardiomyopathy (n [ 17), nonrestrictive cardiomyopathy (n [ 134), and anatomic heart disease (n [ 51). Results. Twenty-eight patients (13.9%) required ECLS after transplantation. Those requiring ECLS had longer ischemic times (4.2 versus 3.7 hours, p [ 0.02). More restrictive cardiomyopathy patients (35.3%) required ECLS—higher than dilated cardiomyopathy (10.4%) or anatomic heart disease (15.7%; c2 7.99; p [ 0.018).
Factors associated with posttransplant ECLS were restrictive cardiomyopathy, longer ischemic time, and extracorporeal membrane oxygenation before transplant. Graft survival was higher in the non-ECLS group, with 1-year survival of 98.2% versus 57.7%; however, medium-term survival was comparable, with 5-year survival for those surviving to hospital discharge being 84.7% versus 100%. Conclusions. The requirement for ECLS was higher than expected for conventional cardiac surgery. Although just over one half of patients requiring ECLS survived to discharge, they had excellent medium-term survival, with all still alive. Although ECLS is an expensive, invasive therapy, with significant morbidity and mortality, without it, those patients would have perished. Its judicious use, therefore, can be recommended.
M
factors, and outcomes of ECLS in this setting are not well documented, particularly in children. We reviewed ECLS as rescue therapy after pediatric heart transplantation at our institution, to uncover risk factors predisposing patients to ECLS.
odern improvements in cardiac surgery, intensive care, and transplant medicine have enhanced outcomes for pediatric heart transplantation over the last 2 decades [1]. In addition, conventional surgical techniques now palliate children that previously would not have survived infancy, a proportion of whom will require transplantation [2]. Unfortunately, in contrast to this growing pool of potential recipients, donation rates are relatively static [3]. This imbalance has led to transplant programs being burdened with a more heterogeneous—and technically demanding—group of recipients, waiting longer on the transplant list (and deteriorating further, with potential for increasing pulmonary vascular resistance [PVR]), and being forced to accept organs from marginal donors. One consequence of these changes is further strain on the critical immediate postoperative period, accentuating the possibility of early graft failure. In response, programs increasingly rely on additional rescue measures, such as nitric oxide and extracorporeal life support (ECLS) [4], while waiting for graft recovery. The indications, risk
Accepted for publication Feb 12, 2015. Address correspondence to Dr Brown, Cardiac Unit, Great Ormond Street Hospital, London WC1N 3JH, United Kingdom; e-mail: katherine.brown@ gosh.nhs.uk.
Ó 2015 by The Society of Thoracic Surgeons Published by Elsevier
(Ann Thorac Surg 2015;99:2166–72) Ó 2015 by The Society of Thoracic Surgeons
Patients and Methods This retrospective review of all pediatric orthotopic heart transplants at Great Ormond Street Hospital covers an 11-year period (January 1999 to December 2009 inclusive), using institutional transplant and ECLS databases. The start reflects the date when ECLS became standard treatment for early graft failure. In all, 202 orthotopic transplants were performed (103 female recipients), including 8 retransplants (patients aged 0.06 to 17.9 years, median 9.6); in addition, 2 heterotopic heart transplants were conducted—these were excluded from analysis. Patients were grouped by pretransplant diagnosis (Fig 1): of 151 transplants for cardiomyopathy, 17 were for restrictive cardiomyopathy (RCM); dilated cardiomyopathy (n ¼ 129) and hypertrophic cardiomyopathy (n ¼ 5) were considered together, as the nonrestrictive cardiomyopathy group (n ¼ 134). The third group consisted of transplants for anatomic disease (n ¼ 51), 48 congenital heart disease and 3 acquired surgical heart disease 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2015.02.047
Ann Thorac Surg 2015;99:2166–72
SIMMONDS ET AL ECLS AFTER PEDIATRIC HEART TRANSPLANTATION
2167
Fig 1. Study population by pretransplant diagnosis. Numbers indicate patients in each group. Red boxes indicate the final groupings used in analysis. (CM ¼ cardiomyopathy; DCM ¼ dilated cardiomyopathy; HCM ¼ hypertrophic cardiomyopathy; RCM ¼ restrictive cardiomyopathy.)
CONGENITAL HEART
(2 noncongenital valve pathology secondary to Marfan’s disease and 1 infective endocarditis). Pretransplant workup involves formal catheter-based PVR measurements for high-risk patients, including those with suspicion of pulmonary hypertension on echocardiography, and patients with RCM. Twenty-five patients had a ventricular assist device (VAD) before transplantation (all cardiomyopathies), and 38 had a period of pretransplant extracorporeal membranous oxygenation (ECMO); 9 patients overlap both groups, with a period of ECMO followed by VAD. In total, 54 patients had mechanical support before transplantation. Immediate postoperative care included inotropic support with milrinone, and low-dose epinephrine infusion, if required. Recipients with suspected elevated PVR were begun on inhaled nitric oxide from the operating theater. Renal failure was treated with continuous venoveno hemofiltration. Severe low cardiac output syndrome (with or without cardiac arrest), despite optimum conventional ventilatory and inotropic assistance, resulted in mechanical circulatory support, the majority of cases with ECMO through either transthoracic or peripheral cannulation, depending on patient factors and operator preference; anticoagulation therapy was monitored using activated clotting time, aiming for 215 s to 235 s, or if the patient was actively bleeding, 195 s to 215 s. Complete blood count, clotting times, and antithrombin 3 levels were monitored, and infusions given based on protocolized levels. During the study, our immunosuppression regime evolved from induction with antithymocyte globulin to basiliximab, and standard maintenance triple immunosuppression from cyclosporin to tacrolimus, and azathioprine to mycophenolate mofetil, in addition to prednisolone, which—in the absence of rejection—is weaned over 3 months. Patients on ECLS support after
transplant followed the standard contemporaneous protocol. Primary outcome measures were ECLS within 30 days of transplantation, and graft loss, comprising death or retransplantation. Secondary outcome measures were length of intubation and in-hospital stay, cerebrovascular accident, seizure (clinical or on electroencephalographic monitoring), laboratory-proven infection, and biopsyproven rejection (grade 2R or more). Primary graft failure was defined as impaired graft function within 24 hours (in the operating theater or intensive care unit) requiring high-dose inotropes or mechanical support. Statistical analyses were run using the entire data set, unless stated, using Stata version 12 (StataCorp, College Station, TX). Transplant patients were separated depending on requirement for ECLS immediately after transplant. Demographic and pretransplant variables were compared using two-sample Wilcoxon rank sum (Mann-Whitney U) tests for continuous data, and c2 tests for categoric data. A small number of variables was missing: ischemic time for 2 patients, donor weight (and donor:recipient weight ratio) for 4, and hours intubated for 3. Rejection and graft survival were analyzed with Kaplan-Meier curves and log rank (Mantel-Cox) tests. Factors associated with posttransplant ECLS were analyzed with a multivariate Cox regression analysis. A p value less than 0.05 was considered significant. This study was registered as a quality improvement audit, and ethics approval was waived.
Results Of 202 transplants, 28 (13.9%) resulted in posttransplant ECLS; 24 had venoarterial ECMO only, 2 had right ventricular assist devices only, and 2 had venoarterial ECMO
CONGENITAL HEART
2168
SIMMONDS ET AL ECLS AFTER PEDIATRIC HEART TRANSPLANTATION
followed by right ventricular assist device. Primary indication for ECLS was primarily left ventricular failure (n ¼ 2), right ventricular failure (n ¼ 10), biventricular failure (n ¼ 12), cardiac arrest (n ¼ 2), and arrhythmia (n ¼ 2). The ECMO cannulation site was neck (n ¼ 8), neck and femoral (n ¼ 2), or transthoracic (n ¼ 10). A left atrial vent was used in 9. There was no significant difference between the ECLS group and non-ECLS group with respect to age, infancy at transplantation, sex, or VAD before transplantation (Table 1). Retransplantation tended to be more common in the ECLS group (7.1% versus 3.4%), but did not reach statistical significance. Patients requiring ELCS after transplant were significantly more likely to have required ECMO before transplantation (35.7% versus 16.1%, p ¼ 0.014). Ten patients required ECMO before transplantation: 3 for congenital heart disease (failing Fontan 8 months after total cavopulmonary connection; low cardiac output after mitral valve repair; and inability to wean from bypass after pulmonary valve homograft replacement), and 7 for end stage cardiomyopathy. Donor and recipient weight and ratio, and ABO mismatch were similar between groups (Table 2). However, patients requiring ECLS after transplant had longer median ischemic time (4.15 versus 3.74 hours, p ¼ 0.02). Complement-dependent cytotoxicity testing on samples taken at transplant (and analyzed after transplant) was available for 185 patients. Positive cross matches were seen in 2 of 25 patients (8.0%) requiring posttransplant ECLS, compared with 11 of 160 non-ECLS patients (6.9%; c2 0.04, p ¼ 0.84). Ventricular assist device use before transplantation was associated with a trend toward positive crossmatch, being seen in 3 of 19 patients (13.6%) on VAD, versus 10 of 153 non-VAD patients (6.1%), although this did not reach statistical significance (c2 1.67, p ¼ 0.20). Of 47 patients receiving mechanical support (ECMO or VAD) before transplantation, 4 (8.5%) had positive cross matches, as opposed to 9 of 129 (6.5%) not on mechanical support (c2 0.21, p ¼ 0.65). The impact of worsening donor shortage was investigated by dividing the cohort into two periods, 1999 to 2003 and 2004 to 2009. The cutoff equates to the point at which mechanical assist devices were introduced as bridge to transplant. There was no difference in ischemic time (Table 3). However, median waiting time increased from Table 1. Pretransplant Variables Variable Age, years Infants Females Pretransplant ECMO Pretransplant VAD Retransplant
Non-ECLS Group (n ¼ 174) 9.6 22 97 28 21 6
(0.1–17.9) (12.6) (55.7) (16.1) (12.1) (3.4)
ECLS Group (n ¼ 28)
p Value
10.0 1 14 10 4 2
0.65 0.14 0.57 0.014 0.74 0.32
(0.7–17.0) (3.6) (50) (35.7) (14.3) (7.1)
Continuous variables are median (range); categoric variables are n (%). ECLS ¼ extracorporeal life support; ECMO ¼ extracorporeal membrane oxygenation; VAD ¼ ventricular assist device.
Ann Thorac Surg 2015;99:2166–72
Table 2. Transplant Variables Variable
Non-ECLS Group (n ¼ 174)
ABO mismatch Recipient weight, kg Donor weight, kg Donor age, years D:R weight ratio D:R weight ratio 3 Ischemic time, hours
17 23.1 50.0 16 1.69 15 3.7
(9.8) (3.1–95.5) (2.0–90.0) (0.3–56) (0.65–3.73) (8.6) (1.1–7)
ECLS Group (n ¼ 28) 2 30.9 55.0 21 1.65 6 4.2
(7.1) (6.0–74.7) (10.0–90.0) (3–53) (0.70–4.00) (21.4) (1.5-8)
p Value 0.69 0.27 0.22 0.08 0.95 0.04 0.02
ABO mismatch and donor:recipient (D:R) weight ratio 3 are n (%); other values are median (range). ECLS ¼ extracorporeal life support.
8.5 days (range, 0 to 404) from 1999 to 2003, to 27 days (range, 0 to 852) from 2004 to 2009 (p ¼ 0.002). Twenty-eight transplants resulted in ECLS: 7 patients died, 2 underwent retransplantation, and 19 were weaned, of whom 4 died before discharge. Causes of death were primarily graft failure (n ¼ 6), respiratory failure (n ¼ 1), neurologic (n ¼ 3), and presumed bacterial infection (n ¼ 1). No patients were reinitiated on ECMO, although 2 subsequently underwent VAD insertion. Two patients underwent retransplantation directly from ECLS: 1 died with increased pulmonary pressures 10 days after retransplant; the second remains alive, 10 years after transplant. Median ECMO duration for those weaned was 109 hours (range, 33 to 246), and for patients who died (or underwent retransplantation) on ECMO, it was 93 hours (range, 20 to 615). ECLS was initiated in the operating theater in 18 patients, of which 14 (77.8%) were subsequently decannulated successfully; for the remaining 10 patients, ECLS was initiated in the intensive care unit, of whom 5 (50%) were weaned successfully (Fisher’s exact test, p ¼ 0.139). The mean time for return of echo function (left ventricular fractional shortening greater than 30%, or subjectively good in presence of dyskinesis) in those weaned was 10 5 days. Patients receiving ECLS after transplantation had longer hospital stays, and more renal replacement therapy (82.1% versus 22.4%, p < 0.001), laboratory-proven infections (21.4% versus 2.9%, p ¼ 0.040), and strokes (14.3% versus 1.7%, p ¼ 0.007; Table 4). Cerebrovascular accidents were hemorrhagic in two thirds of the ECLS patients and in three quarters of the non-ECLS patients. Six of 17 (35.3%) transplants for RCM required postoperative ECLS (Table 5)—significantly higher than for dilated cardiomyopathy (14 of 134; 10.4%) or anatomic heart disease (8 of 51; 15.7%; Pearson’s c2 7.99; p ¼ 0.018). Table 3. Impact of Donor Shortage Impact
1999–2003 (n ¼ 80)
2004–2009 (n ¼ 122)
p Value
Waiting time, days Ischemic time, hours
8.5 (0–404) 3.7 (1.1–8)
27 (0–852) 3.9 (1.3–7.3)
0.002 0.11
Values are median (range).
Ann Thorac Surg 2015;99:2166–72
SIMMONDS ET AL ECLS AFTER PEDIATRIC HEART TRANSPLANTATION
Table 4. Posttransplant Morbidity
Variable
Table 6. Pulmonary Vascular Resistance Studies in Restrictive Cardiomyopathy Subgroup
Non-ECLS Group (n ¼ 174)
ECLS Group (n ¼ 28)
p Value
39 (22.4)
23 (82.1)
CONGENITAL HEART
Cardiac Unit, Great Ormond Street Hospital, London, United Kingdom
Background. Extracorporeal life support (ECLS) has proven success after conventional cardiac surgery. Its use after pediatric heart transplantation is less well documented. We reviewed ECLS after pediatric heart transplantation, to understand better predisposing factors, morbidity, and mortality. Methods. The notes of all patients at Great Ormond Street Hospital undergoing orthotopic heart transplantation from 1999 to 2009 were reviewed (202 transplants; patients aged 0.06 to 17.91 years). Patients were grouped by diagnosis: restrictive cardiomyopathy (n [ 17), nonrestrictive cardiomyopathy (n [ 134), and anatomic heart disease (n [ 51). Results. Twenty-eight patients (13.9%) required ECLS after transplantation. Those requiring ECLS had longer ischemic times (4.2 versus 3.7 hours, p [ 0.02). More restrictive cardiomyopathy patients (35.3%) required ECLS—higher than dilated cardiomyopathy (10.4%) or anatomic heart disease (15.7%; c2 7.99; p [ 0.018).
Factors associated with posttransplant ECLS were restrictive cardiomyopathy, longer ischemic time, and extracorporeal membrane oxygenation before transplant. Graft survival was higher in the non-ECLS group, with 1-year survival of 98.2% versus 57.7%; however, medium-term survival was comparable, with 5-year survival for those surviving to hospital discharge being 84.7% versus 100%. Conclusions. The requirement for ECLS was higher than expected for conventional cardiac surgery. Although just over one half of patients requiring ECLS survived to discharge, they had excellent medium-term survival, with all still alive. Although ECLS is an expensive, invasive therapy, with significant morbidity and mortality, without it, those patients would have perished. Its judicious use, therefore, can be recommended.
M
factors, and outcomes of ECLS in this setting are not well documented, particularly in children. We reviewed ECLS as rescue therapy after pediatric heart transplantation at our institution, to uncover risk factors predisposing patients to ECLS.
odern improvements in cardiac surgery, intensive care, and transplant medicine have enhanced outcomes for pediatric heart transplantation over the last 2 decades [1]. In addition, conventional surgical techniques now palliate children that previously would not have survived infancy, a proportion of whom will require transplantation [2]. Unfortunately, in contrast to this growing pool of potential recipients, donation rates are relatively static [3]. This imbalance has led to transplant programs being burdened with a more heterogeneous—and technically demanding—group of recipients, waiting longer on the transplant list (and deteriorating further, with potential for increasing pulmonary vascular resistance [PVR]), and being forced to accept organs from marginal donors. One consequence of these changes is further strain on the critical immediate postoperative period, accentuating the possibility of early graft failure. In response, programs increasingly rely on additional rescue measures, such as nitric oxide and extracorporeal life support (ECLS) [4], while waiting for graft recovery. The indications, risk
Accepted for publication Feb 12, 2015. Address correspondence to Dr Brown, Cardiac Unit, Great Ormond Street Hospital, London WC1N 3JH, United Kingdom; e-mail: katherine.brown@ gosh.nhs.uk.
Ó 2015 by The Society of Thoracic Surgeons Published by Elsevier
(Ann Thorac Surg 2015;99:2166–72) Ó 2015 by The Society of Thoracic Surgeons
Patients and Methods This retrospective review of all pediatric orthotopic heart transplants at Great Ormond Street Hospital covers an 11-year period (January 1999 to December 2009 inclusive), using institutional transplant and ECLS databases. The start reflects the date when ECLS became standard treatment for early graft failure. In all, 202 orthotopic transplants were performed (103 female recipients), including 8 retransplants (patients aged 0.06 to 17.9 years, median 9.6); in addition, 2 heterotopic heart transplants were conducted—these were excluded from analysis. Patients were grouped by pretransplant diagnosis (Fig 1): of 151 transplants for cardiomyopathy, 17 were for restrictive cardiomyopathy (RCM); dilated cardiomyopathy (n ¼ 129) and hypertrophic cardiomyopathy (n ¼ 5) were considered together, as the nonrestrictive cardiomyopathy group (n ¼ 134). The third group consisted of transplants for anatomic disease (n ¼ 51), 48 congenital heart disease and 3 acquired surgical heart disease 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2015.02.047
Ann Thorac Surg 2015;99:2166–72
SIMMONDS ET AL ECLS AFTER PEDIATRIC HEART TRANSPLANTATION
2167
Fig 1. Study population by pretransplant diagnosis. Numbers indicate patients in each group. Red boxes indicate the final groupings used in analysis. (CM ¼ cardiomyopathy; DCM ¼ dilated cardiomyopathy; HCM ¼ hypertrophic cardiomyopathy; RCM ¼ restrictive cardiomyopathy.)
CONGENITAL HEART
(2 noncongenital valve pathology secondary to Marfan’s disease and 1 infective endocarditis). Pretransplant workup involves formal catheter-based PVR measurements for high-risk patients, including those with suspicion of pulmonary hypertension on echocardiography, and patients with RCM. Twenty-five patients had a ventricular assist device (VAD) before transplantation (all cardiomyopathies), and 38 had a period of pretransplant extracorporeal membranous oxygenation (ECMO); 9 patients overlap both groups, with a period of ECMO followed by VAD. In total, 54 patients had mechanical support before transplantation. Immediate postoperative care included inotropic support with milrinone, and low-dose epinephrine infusion, if required. Recipients with suspected elevated PVR were begun on inhaled nitric oxide from the operating theater. Renal failure was treated with continuous venoveno hemofiltration. Severe low cardiac output syndrome (with or without cardiac arrest), despite optimum conventional ventilatory and inotropic assistance, resulted in mechanical circulatory support, the majority of cases with ECMO through either transthoracic or peripheral cannulation, depending on patient factors and operator preference; anticoagulation therapy was monitored using activated clotting time, aiming for 215 s to 235 s, or if the patient was actively bleeding, 195 s to 215 s. Complete blood count, clotting times, and antithrombin 3 levels were monitored, and infusions given based on protocolized levels. During the study, our immunosuppression regime evolved from induction with antithymocyte globulin to basiliximab, and standard maintenance triple immunosuppression from cyclosporin to tacrolimus, and azathioprine to mycophenolate mofetil, in addition to prednisolone, which—in the absence of rejection—is weaned over 3 months. Patients on ECLS support after
transplant followed the standard contemporaneous protocol. Primary outcome measures were ECLS within 30 days of transplantation, and graft loss, comprising death or retransplantation. Secondary outcome measures were length of intubation and in-hospital stay, cerebrovascular accident, seizure (clinical or on electroencephalographic monitoring), laboratory-proven infection, and biopsyproven rejection (grade 2R or more). Primary graft failure was defined as impaired graft function within 24 hours (in the operating theater or intensive care unit) requiring high-dose inotropes or mechanical support. Statistical analyses were run using the entire data set, unless stated, using Stata version 12 (StataCorp, College Station, TX). Transplant patients were separated depending on requirement for ECLS immediately after transplant. Demographic and pretransplant variables were compared using two-sample Wilcoxon rank sum (Mann-Whitney U) tests for continuous data, and c2 tests for categoric data. A small number of variables was missing: ischemic time for 2 patients, donor weight (and donor:recipient weight ratio) for 4, and hours intubated for 3. Rejection and graft survival were analyzed with Kaplan-Meier curves and log rank (Mantel-Cox) tests. Factors associated with posttransplant ECLS were analyzed with a multivariate Cox regression analysis. A p value less than 0.05 was considered significant. This study was registered as a quality improvement audit, and ethics approval was waived.
Results Of 202 transplants, 28 (13.9%) resulted in posttransplant ECLS; 24 had venoarterial ECMO only, 2 had right ventricular assist devices only, and 2 had venoarterial ECMO
CONGENITAL HEART
2168
SIMMONDS ET AL ECLS AFTER PEDIATRIC HEART TRANSPLANTATION
followed by right ventricular assist device. Primary indication for ECLS was primarily left ventricular failure (n ¼ 2), right ventricular failure (n ¼ 10), biventricular failure (n ¼ 12), cardiac arrest (n ¼ 2), and arrhythmia (n ¼ 2). The ECMO cannulation site was neck (n ¼ 8), neck and femoral (n ¼ 2), or transthoracic (n ¼ 10). A left atrial vent was used in 9. There was no significant difference between the ECLS group and non-ECLS group with respect to age, infancy at transplantation, sex, or VAD before transplantation (Table 1). Retransplantation tended to be more common in the ECLS group (7.1% versus 3.4%), but did not reach statistical significance. Patients requiring ELCS after transplant were significantly more likely to have required ECMO before transplantation (35.7% versus 16.1%, p ¼ 0.014). Ten patients required ECMO before transplantation: 3 for congenital heart disease (failing Fontan 8 months after total cavopulmonary connection; low cardiac output after mitral valve repair; and inability to wean from bypass after pulmonary valve homograft replacement), and 7 for end stage cardiomyopathy. Donor and recipient weight and ratio, and ABO mismatch were similar between groups (Table 2). However, patients requiring ECLS after transplant had longer median ischemic time (4.15 versus 3.74 hours, p ¼ 0.02). Complement-dependent cytotoxicity testing on samples taken at transplant (and analyzed after transplant) was available for 185 patients. Positive cross matches were seen in 2 of 25 patients (8.0%) requiring posttransplant ECLS, compared with 11 of 160 non-ECLS patients (6.9%; c2 0.04, p ¼ 0.84). Ventricular assist device use before transplantation was associated with a trend toward positive crossmatch, being seen in 3 of 19 patients (13.6%) on VAD, versus 10 of 153 non-VAD patients (6.1%), although this did not reach statistical significance (c2 1.67, p ¼ 0.20). Of 47 patients receiving mechanical support (ECMO or VAD) before transplantation, 4 (8.5%) had positive cross matches, as opposed to 9 of 129 (6.5%) not on mechanical support (c2 0.21, p ¼ 0.65). The impact of worsening donor shortage was investigated by dividing the cohort into two periods, 1999 to 2003 and 2004 to 2009. The cutoff equates to the point at which mechanical assist devices were introduced as bridge to transplant. There was no difference in ischemic time (Table 3). However, median waiting time increased from Table 1. Pretransplant Variables Variable Age, years Infants Females Pretransplant ECMO Pretransplant VAD Retransplant
Non-ECLS Group (n ¼ 174) 9.6 22 97 28 21 6
(0.1–17.9) (12.6) (55.7) (16.1) (12.1) (3.4)
ECLS Group (n ¼ 28)
p Value
10.0 1 14 10 4 2
0.65 0.14 0.57 0.014 0.74 0.32
(0.7–17.0) (3.6) (50) (35.7) (14.3) (7.1)
Continuous variables are median (range); categoric variables are n (%). ECLS ¼ extracorporeal life support; ECMO ¼ extracorporeal membrane oxygenation; VAD ¼ ventricular assist device.
Ann Thorac Surg 2015;99:2166–72
Table 2. Transplant Variables Variable
Non-ECLS Group (n ¼ 174)
ABO mismatch Recipient weight, kg Donor weight, kg Donor age, years D:R weight ratio D:R weight ratio 3 Ischemic time, hours
17 23.1 50.0 16 1.69 15 3.7
(9.8) (3.1–95.5) (2.0–90.0) (0.3–56) (0.65–3.73) (8.6) (1.1–7)
ECLS Group (n ¼ 28) 2 30.9 55.0 21 1.65 6 4.2
(7.1) (6.0–74.7) (10.0–90.0) (3–53) (0.70–4.00) (21.4) (1.5-8)
p Value 0.69 0.27 0.22 0.08 0.95 0.04 0.02
ABO mismatch and donor:recipient (D:R) weight ratio 3 are n (%); other values are median (range). ECLS ¼ extracorporeal life support.
8.5 days (range, 0 to 404) from 1999 to 2003, to 27 days (range, 0 to 852) from 2004 to 2009 (p ¼ 0.002). Twenty-eight transplants resulted in ECLS: 7 patients died, 2 underwent retransplantation, and 19 were weaned, of whom 4 died before discharge. Causes of death were primarily graft failure (n ¼ 6), respiratory failure (n ¼ 1), neurologic (n ¼ 3), and presumed bacterial infection (n ¼ 1). No patients were reinitiated on ECMO, although 2 subsequently underwent VAD insertion. Two patients underwent retransplantation directly from ECLS: 1 died with increased pulmonary pressures 10 days after retransplant; the second remains alive, 10 years after transplant. Median ECMO duration for those weaned was 109 hours (range, 33 to 246), and for patients who died (or underwent retransplantation) on ECMO, it was 93 hours (range, 20 to 615). ECLS was initiated in the operating theater in 18 patients, of which 14 (77.8%) were subsequently decannulated successfully; for the remaining 10 patients, ECLS was initiated in the intensive care unit, of whom 5 (50%) were weaned successfully (Fisher’s exact test, p ¼ 0.139). The mean time for return of echo function (left ventricular fractional shortening greater than 30%, or subjectively good in presence of dyskinesis) in those weaned was 10 5 days. Patients receiving ECLS after transplantation had longer hospital stays, and more renal replacement therapy (82.1% versus 22.4%, p < 0.001), laboratory-proven infections (21.4% versus 2.9%, p ¼ 0.040), and strokes (14.3% versus 1.7%, p ¼ 0.007; Table 4). Cerebrovascular accidents were hemorrhagic in two thirds of the ECLS patients and in three quarters of the non-ECLS patients. Six of 17 (35.3%) transplants for RCM required postoperative ECLS (Table 5)—significantly higher than for dilated cardiomyopathy (14 of 134; 10.4%) or anatomic heart disease (8 of 51; 15.7%; Pearson’s c2 7.99; p ¼ 0.018). Table 3. Impact of Donor Shortage Impact
1999–2003 (n ¼ 80)
2004–2009 (n ¼ 122)
p Value
Waiting time, days Ischemic time, hours
8.5 (0–404) 3.7 (1.1–8)
27 (0–852) 3.9 (1.3–7.3)
0.002 0.11
Values are median (range).
Ann Thorac Surg 2015;99:2166–72
SIMMONDS ET AL ECLS AFTER PEDIATRIC HEART TRANSPLANTATION
Table 4. Posttransplant Morbidity
Variable
Table 6. Pulmonary Vascular Resistance Studies in Restrictive Cardiomyopathy Subgroup
Non-ECLS Group (n ¼ 174)
ECLS Group (n ¼ 28)
p Value
39 (22.4)
23 (82.1)
