Biventricular Berlin Heart EXCOR Pediatric Use Across the United States Farhan Zafar, MD,* John L. Jefferies, MD, MPH,* Christine J. Tjossem, BS, Roosevelt Bryant, III, MD, Robert D. B. Jaquiss, MD, Peter D. Wearden, MD, PhD, David N. Rosenthal, MD, Antonio G. Cabrera, MD, Joseph W. Rossano, MD, Tilman Humpl, MD, PhD, and David L. S. Morales, MD Cincinnati Children’s Hospital, University of Cincinnati College of Medicine, Cincinnati, Ohio; Berlin Heart, Inc, The Woodlands, Texas; Duke Children’s Hospital, Duke University School of Medicine, Durham, North Carolina; Children’s Hospital of Pittsburgh, University of Pittsburgh School of Medicine, Pittsburgh, PA; Lucile Packard Children’s Hospital, Stanford University School of Medicine, Palo Alto, California; Texas Children’s Hospital, Baylor College of Medicine, Houston, Texas; The Children’s Hospital of Philadelphia, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania; and Hospital for Sick Children, University of Toronto, Toronto, Ontario, Canada
Background. Biventricular assist device (BiVAD) support was a strong predictor of early mortality in the Berlin Heart EXCOR Pediatric investigational device exemption (IDE) study (Assess Safety and Probable Benefit of the EXCOR Pediatric Ventricular Assist Device (VAD)). In adults, it has been identified that 5% to 10% of the VAD population is benefited by BiVAD support over left ventricular assist device (LVAD) support. An analysis of the Berlin Heart study cohort was performed to characterize patients supported with BiVAD, examine risk factors of mortality in this group, and identify subsets of patients in whom BiVAD is associated with survival. Methods. All EXCOR Pediatric devices (Berlin Heart, Inc, The Woodlands, TX) placed in North America between May 2007 and December 2010 comprised the study cohort of 204 patients (128 [63%] LVADs and 76 [37%] BiVADs). The following patient cohorts were analyzed to determine the effect of BiVAD use on survival: Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) profile 1 patients, patients with abnormal bilirubin levels, patients who received previous extracorporeal membrane oxygenation (ECMO), and patients with a 10-mL pump size.
Results. There were more patients with BiVADs in INTERMACS profile 1 (63% compared with 46%; p [ 0.018). The incidence of major bleeding, neurologic and renal dysfunction, and infection was similar between BiVAD and LVAD groups. White race, abnormal glomerular filtration rate (GFR), sites with experience of less than 5 implantations, and use of 10-mL pumps were predictors of mortality in patients who received BiVADs. BiVADs were not associated with improved survival in any patient cohort; however, they were associated with increased mortality in patients who had undergone ECMO before receiving a VAD. Conclusions. BiVAD support was not associated with improved survival in any identified subset of patients. Although not randomized, these results (which were corrected for multiple possible risk factors) suggest that some children supported with BiVADs might have done better with LVADs alone. Further prospective studies will be needed to identify patient cohorts that will be better served with BIVAD support.
C
consisted of 37% of all EXCOR Pediatric (Berlin Heart, Inc, The Woodlands, TX) implantations [1]. Before the IDE study, the initial experience with the EXCOR Pediatric device between 2000 and 2007 demonstrated significantly higher mortality associated with BiVAD (35%) support compared with LVAD (14%) support. Moreover, the use of BiVADs was the strongest predictor of mortality of all risk factors in this series [2]. Similar results have been reported in adult mechanical circulatory support (MCS) experiences, with 25% better survival with
hoice of device strategy between left ventricular assist devices (LVADs) and biventricular assist devices (BiVADs) is a controversial subject. BiVAD use in the United States during the Berlin Heart EXCOR Pediatric multiinstitutional investigational device exemption (IDE) study (Assess Safety and Probable Benefit of the EXCOR Pediatric Ventricular Assist Device) (2007–2010) Accepted for publication Sept 23, 2014. *Drs Zafar and Jefferies are first coauthors and have contributed equally to this study. Address correspondence to Dr Morales, Cincinnati Children’s Hospital Medical Center, The University of Cincinnati College of Medicine, 3333 Burnet Ave–MLC 2004, Cincinnati, OH 45229; e-mail: david.morales@ cchmc.org.
Ó 2015 by The Society of Thoracic Surgeons Published by Elsevier
(Ann Thorac Surg 2015;-:-–-) Ó 2015 by The Society of Thoracic Surgeons
Drs Tjossem, Rosenthal, and Morales disclose financial relationships with Berlin Heart.
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2014.09.078
2
ZAFAR ET AL BIVENTRICULAR SUPPORT IN PEDIATRICS
Abbreviations and Acronyms BiVAD CHD ECMO
= biventricular assist device = congenital heart disease = extracorporeal membrane oxygenation GFR = glomerular filtration rate IDE = investigational device exemption INTERMACS = Interagency Registry for Mechanically Assisted Circulatory Support LVAD = left ventricular assist device MCS = mechanical circulatory support PAS = postapproval study RV = right ventricular RVAD = right ventricular assist device VAD = ventricular assist device
LVAD support compared with BiVAD support [3]. As the field of adult ventricular assist device (VAD) therapy has progressed, the use of BiVADs in adults has decreased from 21% in 2007 to less than 5% in 2010. This is largely attributed to improved decision making, earlier intervention, and the maturation of programs [3]. Similarly, with improvements in decision making, the goal of pediatric programs may be to reduce the use of BiVADs to about 5% to 10% of all VADs. Most clinicians, including the authors, believe that there are cohorts of pediatric patients in whom BiVAD support is better than LVAD support alone, but who they are remains unclear. The aims of our study were (1) to characterize patients supported with BiVADs in the EXCOR Pediatric device multiinstitutional study, (2) to examine risk factors associated with mortality in this group, and (3) to identify cohorts in whom BiVAD support is associated with survival.
Patients and Methods Study Cohort The Berlin Heart EXCOR Pediatric device regulatory database was queried to identify all US children who received the Berlin Heart EXCOR Pediatric device between May 9, 2007 and December 31, 2010. The trial design, inclusion and exclusion criteria, and end points have been reported previously [2]. The IDE study was conducted at 17 centers across the United States and Canada. Additionally there were 27 non-IDE study sites that had access to compassionate use of the EXCOR Pediatric device during the study period. Every patient from all 44 sites was included in the study cohort. Data were collected by study coordinators at each site using a data platform adapted from the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) database [4]. All patients were followed from the time of EXCOR Pediatric device implantation until transplantation, death, or recovery. Patients were censored at
Ann Thorac Surg 2015;-:-–-
the time of transplantation or recovery. The Institutional Review Board at Cincinnati Children’s Hospital approved this study.
Primary End Point and Definitions The primary end point for this analysis was all-cause mortality in patients receiving EXCOR Pediatric device support. Renal dysfunction in this study was estimated by the glomerular filtration rate (GFR) adjusted for age with the formula of Schwartz and colleagues [5] and was categorized as moderately impaired (GFR, 30%–89% of predicted) or severely impaired (GFR < 30% of predicted). Hepatic dysfunction in this study was defined as elevation in the serum bilirubin level and was categorized as moderate (0.7–1.1 mg/dL) and severe ( 1.2 mg/dL). All clinical and demographic variables were defined at the time of EXCOR Pediatric device implantation unless otherwise specified. All data were analyzed as reported by the implanting center, except the serious adverse events, which were adjudicated by a central committee for the patients who underwent implantation at the IDE institutions.
Statistical Analysis Summaries of continuous variables are presented as median and range, whereas characteristics with discrete or binary variables are presented as number (percent). All characteristics were compared by use of the c2 test or Fisher’s exact test (if appropriate) for categorical variables and the Kruskall-Wallis test (nonparametric tests used given nonnormally distributed data) for continuous variables. Time to death while on EXCOR Pediatric device support and time to transplantation or recovery were analyzed as competing events. Logistic and Cox regression models used a stepwise backward elimination with a p value of 0.05 to retain variables. Multiple regression models were created using variables identified as predictive of death on univariate analysis. Additionally, the authors of this article and the Berlin Heart Investigators group identified 9 patient cohorts that clinically have been referenced as populations that might benefit from BiVAD support over LVAD support. These subgroups included weight less than 5 kg (33 patients), a severely impaired GFR (11 patients), congenital heart disease (CHD) (59 patients), INTERMACS profile 1 (107 patients), bilirubin greater than or equal to 1.2 mg/dL (86 patients), previous ECMO (83 patients), 10-mL pump size (76 patients), age less than 1 year (77 patients), and EXCOR Pediatric device support longer than 60 days (72 patients). Subgroup regression analysis was performed to identify if BiVAD support was protective for mortality in any of these groups. Subgroups of weight less than 5 kg, severely impaired GFR, CHD, age less than 1 year, and EXCOR Pediatric device support longer than 60 days were not reported because of invalidity of the regression model resulting from an inadequate sample size in each of these groups.
Ann Thorac Surg 2015;-:-–-
Table 1. Preimplantation and Implantation Characteristics Variable
a
LVAD (n ¼ 128)
18.6 33 97 124 107 59 18 86 54 49 49
(0.2–191.8) (16.2%) (47.6%) (62.6%) (52.5%) (28.9%) (8.8%) (44.3%) (26.5%) (24.0%) (24.0%)
6.1 22 63 68 59 34 16 44 37 28 36
(0.2–173.2) (17.2%) (49.2%) (54.8%) (46.1%) (26.6%) (12.5%) (37.0%) (29.0%) (21.9%) (28.1%)
157 83 187 7
(77.0%) (40.7%) (91.7%) (3.4%)
96 47 120 7
(75.0%) (36.7%) (93.8%) (5.5%)
51 (25.0%) 61 (29.9%) 92 (45.1%)
BiVAD (n ¼ 76) 23.9 11 34 56 48 25 2 42 17 21 13 61 36 67 0
31 (24.2%) 43 (33.6%) 54 (42.2%)
p Valuea (LVAD Versus BiVAD)
(0.5–191.8) (14.5%) (44.7%) (75.7%) (63.2%) (32.9%) (2.6%) (56.0%) (22.4%) (27.6%) (17.1%)
0.091 0.611 0.535 0.003b 0.018b 0.335 0.016b 0.015b 0.562 0.352 0.075
(80.3%) (47.4%) (88.2%) (0.0%)
0.388 0.134 0.162 0.047b 0.317
20 (26.3%) 18 (23.7%) 38 (50.0%) 0.478
76 103 25 159.0 38.5
(37.3%) (50.5%) (12.2%) (71–599) (0–435)
48 67 13 146.0 47.5
(37.5%) (52.3%) (10.2%) (73–316) (1–435)
p values from c2 test or Fisher’s exact test for binary variables and the Kruskal-Wallis test for continuous variables.
28 36 12 198.0 35.0 b
(36.8%) (47.4%) (15.8%) (71–599) (0–233)
10 Pump size 10 mL 25 or 30 mL 50 or 60 mL Cardiopulmonary bypass time, median (range) EXCOR Pediatric device support, d, median (range)
Overall (N ¼ 204)
3
4
ZAFAR ET AL BIVENTRICULAR SUPPORT IN PEDIATRICS
Ann Thorac Surg 2015;-:-–-
Results
Fig 1. Kaplan-Meier curve comparing survival after left ventricular assist device (LVAD) implantation versus biventricular assist device (BiVAD) implantation.
Fig 2. Competing outcomes analysis for (A) groups overall, (B) left ventricular assist device (LVAD) group, and (C) biventricular assist device (BiVAD) group.
There were 204 patients during the study period who were supported with the EXCOR Pediatric device. Seventy-six, (37%) of these patients had BiVADs, whereas 128 (63%) patients had LVADs. Both groups were similar in preimplantation and implantation characteristics, except that patients with BiVADs had a higher proportion of INTERMACS profile 1 at implantation. Patients with BiVADs were also more likely to be white and to have an abnormal bilirubin value (> 1.2 mg/dL) (Table 1). Overall survival for the 2 groups was significantly different, with a 100-day survival of 80% for LVADs versus 60% for BiVADs (log-rank p ¼ 0.03) (Fig 1). Patients on BiVAD support had higher mortality while on the device (32%) compared with patients with LVADs (23%) (Fig 2). The incidence of serious adverse events was similar in both groups (Table 2). The total days on support
Ann Thorac Surg 2015;-:-–-
ZAFAR ET AL BIVENTRICULAR SUPPORT IN PEDIATRICS
5
Table 2. Adverse Events LVAD (n ¼ 128) No. of Events (%)b
Event Major bleeding Major infection Localized, not device related Percutaneous site or pocket Internal pump component Sepsis Neurologic dysfunctionc TIA Ischemic CVA Hemorrhagic CVA Renal dysfunction Pump replacement for thrombus
55 61 51 4 5 17 41 8 27 4 12 56
BiVAD (n ¼ 76) No. of Events (%)c
(43) (48) (40) (3) (4) (13) (32) (6) (21) (3) (9) (43.8)
35 31 27 1 2 6 18 4 14 1 11 25
p Valuea (LVAD Versus BiVAD)
(46) (41) (36) (1.3) (2.6) (8) (24) (5) (18) (1.3) (14) (32.9)
0.668 0.341 0.539 0.653 1.000 0.240 0.204 1.000 0.645 0.653 0.266 0.122
b p values from c2 or Fisher’s exact test when appropriate. percent (%) refers to events per patient— a patient with multiple identical events is c counted as 1 event. neurologic events besides TIA, ischemic CVA, and hemorrhagic CVA are not shown here individually.
a
BiVAD ¼ biventricular assist device;
CVA ¼ cerebrovascular accident;
was not different for BiVADs (median, 35.0; range, 0–233 days) versus LVADs (median, 47.5; range, 1–435 days) (p ¼ 0.094). Univariate and multivariate predictors of mortality for each group and overall are summarized in Table 3. White patients, patients who had an abnormal GFR for age, implantation at sites with experience of less than 5 implantations, and a pump size of 10 mL were independent predictors of mortality in patients on BiVAD support. Subgroup analysis of patients at risk demonstrated that BiVAD support was not associated with better survival in any of the subgroups. However, BiVAD use was associated with increased mortality in patients who had received previous VAD ECMO (Table 4).
LVAD ¼ left ventricular assist device;
TIA ¼ transient ischemic attack.
Comment With the evolution of the field of pediatric MCS, more effort is now being directed toward early decision making, appropriate patient identification, and device selection strategies. Our study, which analyzed the largest cohort of pediatric patients with VADs available, aimed to identify risk factors associated with mortality in patients supported with BiVADs and if BiVAD support is associated with survival in any subgroup of patients. Our data show that groups with characteristics that are generally thought to indicate the use of BiVAD support did not have any survival benefit with BiVAD use, and an abnormal GFR, small pump size, and centers with less experience are independently associated with mortality in
Table 3. Univariate and Multivariate Predictors of Mortality After Each Device Strategy and Overalla Univariate Predictors
Multivariate Predictors
Factors
Overall
LVAD
BiVAD
Overall
LVAD
BiVAD
Weight 1.2 mg/dL ECMO before implantation Site experience 5 implantations Pump size 10 mL Age CPB time Days on EXCOR Pediatric device support
Background. Biventricular assist device (BiVAD) support was a strong predictor of early mortality in the Berlin Heart EXCOR Pediatric investigational device exemption (IDE) study (Assess Safety and Probable Benefit of the EXCOR Pediatric Ventricular Assist Device (VAD)). In adults, it has been identified that 5% to 10% of the VAD population is benefited by BiVAD support over left ventricular assist device (LVAD) support. An analysis of the Berlin Heart study cohort was performed to characterize patients supported with BiVAD, examine risk factors of mortality in this group, and identify subsets of patients in whom BiVAD is associated with survival. Methods. All EXCOR Pediatric devices (Berlin Heart, Inc, The Woodlands, TX) placed in North America between May 2007 and December 2010 comprised the study cohort of 204 patients (128 [63%] LVADs and 76 [37%] BiVADs). The following patient cohorts were analyzed to determine the effect of BiVAD use on survival: Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) profile 1 patients, patients with abnormal bilirubin levels, patients who received previous extracorporeal membrane oxygenation (ECMO), and patients with a 10-mL pump size.
Results. There were more patients with BiVADs in INTERMACS profile 1 (63% compared with 46%; p [ 0.018). The incidence of major bleeding, neurologic and renal dysfunction, and infection was similar between BiVAD and LVAD groups. White race, abnormal glomerular filtration rate (GFR), sites with experience of less than 5 implantations, and use of 10-mL pumps were predictors of mortality in patients who received BiVADs. BiVADs were not associated with improved survival in any patient cohort; however, they were associated with increased mortality in patients who had undergone ECMO before receiving a VAD. Conclusions. BiVAD support was not associated with improved survival in any identified subset of patients. Although not randomized, these results (which were corrected for multiple possible risk factors) suggest that some children supported with BiVADs might have done better with LVADs alone. Further prospective studies will be needed to identify patient cohorts that will be better served with BIVAD support.
C
consisted of 37% of all EXCOR Pediatric (Berlin Heart, Inc, The Woodlands, TX) implantations [1]. Before the IDE study, the initial experience with the EXCOR Pediatric device between 2000 and 2007 demonstrated significantly higher mortality associated with BiVAD (35%) support compared with LVAD (14%) support. Moreover, the use of BiVADs was the strongest predictor of mortality of all risk factors in this series [2]. Similar results have been reported in adult mechanical circulatory support (MCS) experiences, with 25% better survival with
hoice of device strategy between left ventricular assist devices (LVADs) and biventricular assist devices (BiVADs) is a controversial subject. BiVAD use in the United States during the Berlin Heart EXCOR Pediatric multiinstitutional investigational device exemption (IDE) study (Assess Safety and Probable Benefit of the EXCOR Pediatric Ventricular Assist Device) (2007–2010) Accepted for publication Sept 23, 2014. *Drs Zafar and Jefferies are first coauthors and have contributed equally to this study. Address correspondence to Dr Morales, Cincinnati Children’s Hospital Medical Center, The University of Cincinnati College of Medicine, 3333 Burnet Ave–MLC 2004, Cincinnati, OH 45229; e-mail: david.morales@ cchmc.org.
Ó 2015 by The Society of Thoracic Surgeons Published by Elsevier
(Ann Thorac Surg 2015;-:-–-) Ó 2015 by The Society of Thoracic Surgeons
Drs Tjossem, Rosenthal, and Morales disclose financial relationships with Berlin Heart.
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2014.09.078
2
ZAFAR ET AL BIVENTRICULAR SUPPORT IN PEDIATRICS
Abbreviations and Acronyms BiVAD CHD ECMO
= biventricular assist device = congenital heart disease = extracorporeal membrane oxygenation GFR = glomerular filtration rate IDE = investigational device exemption INTERMACS = Interagency Registry for Mechanically Assisted Circulatory Support LVAD = left ventricular assist device MCS = mechanical circulatory support PAS = postapproval study RV = right ventricular RVAD = right ventricular assist device VAD = ventricular assist device
LVAD support compared with BiVAD support [3]. As the field of adult ventricular assist device (VAD) therapy has progressed, the use of BiVADs in adults has decreased from 21% in 2007 to less than 5% in 2010. This is largely attributed to improved decision making, earlier intervention, and the maturation of programs [3]. Similarly, with improvements in decision making, the goal of pediatric programs may be to reduce the use of BiVADs to about 5% to 10% of all VADs. Most clinicians, including the authors, believe that there are cohorts of pediatric patients in whom BiVAD support is better than LVAD support alone, but who they are remains unclear. The aims of our study were (1) to characterize patients supported with BiVADs in the EXCOR Pediatric device multiinstitutional study, (2) to examine risk factors associated with mortality in this group, and (3) to identify cohorts in whom BiVAD support is associated with survival.
Patients and Methods Study Cohort The Berlin Heart EXCOR Pediatric device regulatory database was queried to identify all US children who received the Berlin Heart EXCOR Pediatric device between May 9, 2007 and December 31, 2010. The trial design, inclusion and exclusion criteria, and end points have been reported previously [2]. The IDE study was conducted at 17 centers across the United States and Canada. Additionally there were 27 non-IDE study sites that had access to compassionate use of the EXCOR Pediatric device during the study period. Every patient from all 44 sites was included in the study cohort. Data were collected by study coordinators at each site using a data platform adapted from the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) database [4]. All patients were followed from the time of EXCOR Pediatric device implantation until transplantation, death, or recovery. Patients were censored at
Ann Thorac Surg 2015;-:-–-
the time of transplantation or recovery. The Institutional Review Board at Cincinnati Children’s Hospital approved this study.
Primary End Point and Definitions The primary end point for this analysis was all-cause mortality in patients receiving EXCOR Pediatric device support. Renal dysfunction in this study was estimated by the glomerular filtration rate (GFR) adjusted for age with the formula of Schwartz and colleagues [5] and was categorized as moderately impaired (GFR, 30%–89% of predicted) or severely impaired (GFR < 30% of predicted). Hepatic dysfunction in this study was defined as elevation in the serum bilirubin level and was categorized as moderate (0.7–1.1 mg/dL) and severe ( 1.2 mg/dL). All clinical and demographic variables were defined at the time of EXCOR Pediatric device implantation unless otherwise specified. All data were analyzed as reported by the implanting center, except the serious adverse events, which were adjudicated by a central committee for the patients who underwent implantation at the IDE institutions.
Statistical Analysis Summaries of continuous variables are presented as median and range, whereas characteristics with discrete or binary variables are presented as number (percent). All characteristics were compared by use of the c2 test or Fisher’s exact test (if appropriate) for categorical variables and the Kruskall-Wallis test (nonparametric tests used given nonnormally distributed data) for continuous variables. Time to death while on EXCOR Pediatric device support and time to transplantation or recovery were analyzed as competing events. Logistic and Cox regression models used a stepwise backward elimination with a p value of 0.05 to retain variables. Multiple regression models were created using variables identified as predictive of death on univariate analysis. Additionally, the authors of this article and the Berlin Heart Investigators group identified 9 patient cohorts that clinically have been referenced as populations that might benefit from BiVAD support over LVAD support. These subgroups included weight less than 5 kg (33 patients), a severely impaired GFR (11 patients), congenital heart disease (CHD) (59 patients), INTERMACS profile 1 (107 patients), bilirubin greater than or equal to 1.2 mg/dL (86 patients), previous ECMO (83 patients), 10-mL pump size (76 patients), age less than 1 year (77 patients), and EXCOR Pediatric device support longer than 60 days (72 patients). Subgroup regression analysis was performed to identify if BiVAD support was protective for mortality in any of these groups. Subgroups of weight less than 5 kg, severely impaired GFR, CHD, age less than 1 year, and EXCOR Pediatric device support longer than 60 days were not reported because of invalidity of the regression model resulting from an inadequate sample size in each of these groups.
Ann Thorac Surg 2015;-:-–-
Table 1. Preimplantation and Implantation Characteristics Variable
a
LVAD (n ¼ 128)
18.6 33 97 124 107 59 18 86 54 49 49
(0.2–191.8) (16.2%) (47.6%) (62.6%) (52.5%) (28.9%) (8.8%) (44.3%) (26.5%) (24.0%) (24.0%)
6.1 22 63 68 59 34 16 44 37 28 36
(0.2–173.2) (17.2%) (49.2%) (54.8%) (46.1%) (26.6%) (12.5%) (37.0%) (29.0%) (21.9%) (28.1%)
157 83 187 7
(77.0%) (40.7%) (91.7%) (3.4%)
96 47 120 7
(75.0%) (36.7%) (93.8%) (5.5%)
51 (25.0%) 61 (29.9%) 92 (45.1%)
BiVAD (n ¼ 76) 23.9 11 34 56 48 25 2 42 17 21 13 61 36 67 0
31 (24.2%) 43 (33.6%) 54 (42.2%)
p Valuea (LVAD Versus BiVAD)
(0.5–191.8) (14.5%) (44.7%) (75.7%) (63.2%) (32.9%) (2.6%) (56.0%) (22.4%) (27.6%) (17.1%)
0.091 0.611 0.535 0.003b 0.018b 0.335 0.016b 0.015b 0.562 0.352 0.075
(80.3%) (47.4%) (88.2%) (0.0%)
0.388 0.134 0.162 0.047b 0.317
20 (26.3%) 18 (23.7%) 38 (50.0%) 0.478
76 103 25 159.0 38.5
(37.3%) (50.5%) (12.2%) (71–599) (0–435)
48 67 13 146.0 47.5
(37.5%) (52.3%) (10.2%) (73–316) (1–435)
p values from c2 test or Fisher’s exact test for binary variables and the Kruskal-Wallis test for continuous variables.
28 36 12 198.0 35.0 b
(36.8%) (47.4%) (15.8%) (71–599) (0–233)
10 Pump size 10 mL 25 or 30 mL 50 or 60 mL Cardiopulmonary bypass time, median (range) EXCOR Pediatric device support, d, median (range)
Overall (N ¼ 204)
3
4
ZAFAR ET AL BIVENTRICULAR SUPPORT IN PEDIATRICS
Ann Thorac Surg 2015;-:-–-
Results
Fig 1. Kaplan-Meier curve comparing survival after left ventricular assist device (LVAD) implantation versus biventricular assist device (BiVAD) implantation.
Fig 2. Competing outcomes analysis for (A) groups overall, (B) left ventricular assist device (LVAD) group, and (C) biventricular assist device (BiVAD) group.
There were 204 patients during the study period who were supported with the EXCOR Pediatric device. Seventy-six, (37%) of these patients had BiVADs, whereas 128 (63%) patients had LVADs. Both groups were similar in preimplantation and implantation characteristics, except that patients with BiVADs had a higher proportion of INTERMACS profile 1 at implantation. Patients with BiVADs were also more likely to be white and to have an abnormal bilirubin value (> 1.2 mg/dL) (Table 1). Overall survival for the 2 groups was significantly different, with a 100-day survival of 80% for LVADs versus 60% for BiVADs (log-rank p ¼ 0.03) (Fig 1). Patients on BiVAD support had higher mortality while on the device (32%) compared with patients with LVADs (23%) (Fig 2). The incidence of serious adverse events was similar in both groups (Table 2). The total days on support
Ann Thorac Surg 2015;-:-–-
ZAFAR ET AL BIVENTRICULAR SUPPORT IN PEDIATRICS
5
Table 2. Adverse Events LVAD (n ¼ 128) No. of Events (%)b
Event Major bleeding Major infection Localized, not device related Percutaneous site or pocket Internal pump component Sepsis Neurologic dysfunctionc TIA Ischemic CVA Hemorrhagic CVA Renal dysfunction Pump replacement for thrombus
55 61 51 4 5 17 41 8 27 4 12 56
BiVAD (n ¼ 76) No. of Events (%)c
(43) (48) (40) (3) (4) (13) (32) (6) (21) (3) (9) (43.8)
35 31 27 1 2 6 18 4 14 1 11 25
p Valuea (LVAD Versus BiVAD)
(46) (41) (36) (1.3) (2.6) (8) (24) (5) (18) (1.3) (14) (32.9)
0.668 0.341 0.539 0.653 1.000 0.240 0.204 1.000 0.645 0.653 0.266 0.122
b p values from c2 or Fisher’s exact test when appropriate. percent (%) refers to events per patient— a patient with multiple identical events is c counted as 1 event. neurologic events besides TIA, ischemic CVA, and hemorrhagic CVA are not shown here individually.
a
BiVAD ¼ biventricular assist device;
CVA ¼ cerebrovascular accident;
was not different for BiVADs (median, 35.0; range, 0–233 days) versus LVADs (median, 47.5; range, 1–435 days) (p ¼ 0.094). Univariate and multivariate predictors of mortality for each group and overall are summarized in Table 3. White patients, patients who had an abnormal GFR for age, implantation at sites with experience of less than 5 implantations, and a pump size of 10 mL were independent predictors of mortality in patients on BiVAD support. Subgroup analysis of patients at risk demonstrated that BiVAD support was not associated with better survival in any of the subgroups. However, BiVAD use was associated with increased mortality in patients who had received previous VAD ECMO (Table 4).
LVAD ¼ left ventricular assist device;
TIA ¼ transient ischemic attack.
Comment With the evolution of the field of pediatric MCS, more effort is now being directed toward early decision making, appropriate patient identification, and device selection strategies. Our study, which analyzed the largest cohort of pediatric patients with VADs available, aimed to identify risk factors associated with mortality in patients supported with BiVADs and if BiVAD support is associated with survival in any subgroup of patients. Our data show that groups with characteristics that are generally thought to indicate the use of BiVAD support did not have any survival benefit with BiVAD use, and an abnormal GFR, small pump size, and centers with less experience are independently associated with mortality in
Table 3. Univariate and Multivariate Predictors of Mortality After Each Device Strategy and Overalla Univariate Predictors
Multivariate Predictors
Factors
Overall
LVAD
BiVAD
Overall
LVAD
BiVAD
Weight 1.2 mg/dL ECMO before implantation Site experience 5 implantations Pump size 10 mL Age CPB time Days on EXCOR Pediatric device support
