Utility of extracorporeal membrane oxygenation for early graft failure following heart transplantation in infancy

Utility of extracorporeal membrane oxygenation for early graft failure following heart transplantation in infancy

Utility of Extracorporeal Membrane Oxygenation for Early Graft Failure Following Heart Transplantation in Infancy Max B. Mitchell, MD, David N. Campbe...

118KB Sizes 0 Downloads 46 Views

Utility of Extracorporeal Membrane Oxygenation for Early Graft Failure Following Heart Transplantation in Infancy Max B. Mitchell, MD, David N. Campbell, MD, Mark R. Bielefeld, MD, and Trischia Doremus, BA Background: Extracorporeal membrane oxygenation (ECMO) is widely used for postcardiotomy cardiogenic shock in children. However, the efficacy of ECMO for early post-heart transplant graft failure in infants has not been reported. Our aims were to determine: (1) the utility of ECMO in infants with severe donor-heart dysfunction, (2) predictors for requiring ECMO, and (3) the long-term outcome of surviving ECMO patients. Methods: All infants (age ⬍ 6 months at listing) undergoing heart transplantation were reviewed. Diagnostic categories were hypoplastic left heart syndrome (HLHS) and nonHLHS (complex congenital heart disease and cardiomyopathies). Continuous and categorical comparisons were by Wilcoxon’s rank sum test and Fisher’s exact test respectively. Results: 14 (12 HLHS, 2 non-HLHS) of 63 (46 HLHS, 17 non-HLHS) infants were placed on ECMO. Ten patients (71%) were successfully weaned from ECMO and 8 (57%) were discharged alive. All ECMO hospital survivors remain alive (mean followup 36.2 ⫾ 21.4 months, range 13.1–77.6 months). Mean duration of ECMO support was 68 hours in weaned patients vs 144 hours (p ⫽ 0.19) in nonweaned patients, and 64 hours in survivors vs 123 hours (p ⫽ 0.35) in nonsurvivors. ECMO deaths were due to sepsis (n ⫽ 3), intractable pulmonary hypertension (n ⫽ 2), and intracranial bleed (n ⫽ 1). Neurologic deficits occurred in 2 survivors. Median ICU and hospital stays for ECMO survivors were 29 and 33 days vs 7 (p ⫽ 0.0003) and 9 (p ⫽ 0.0004) days for non-ECMO patients. Age listed, age transplanted, wait time, body weight, donor/recipient weight ratio, total ischemia time, and diagnosis did not predict the need for ECMO. Conclusions: (1) ECMO is useful for post-heart transplant circulatory support in infants with early graft failure. (2) All survivors were weaned in fewer than 4 days. (3) Three-year survival of ECMO hospital survivors has been high, but neurologic complications are prevalent. J Heart Lung Transplant 2000;19:834–839.

H

eart transplantation is accepted treatment for end-stage heart disease in children.1,2,3 In our cen-

ter, short- and intermediate-term survival in infants undergoing heart transplantation has improved with

From the Division of Cardiothoracic Surgery, University of Colorado Health Sciences Center and The Children’s Hospital, Denver, Colorado Submitted March 27, 2000; accepted June 23, 2000. Reprint requests: Max B. Mitchell, MD, The Children’s Hospital, 1056 E. 19th Avenue, B200, Denver, Colorado 80218. Tele-

phone: 303-861-6624 (business), 303-393-6852 (home). Fax: 303-764-8022. E-mail: [email protected]. Copyright © 2000 by the International Society for Heart and Lung Transplantation. 1053-2498/00/$–see front matter PII S1053-2498(00)00171-6

834

The Journal of Heart and Lung Transplantation Volume 19, Number 9

increasing experience.4 This improvement has largely been due to lower operative mortality during our more-recent experience. In our prior report we noted that several infants with severe donor-heart failure were salvaged with mechanical circulatory support utilizing ECMO, and that this modality contributed to decreased operative mortality in these patients. In pediatric centers ECMO is widely used for mechanical circulatory support of children with postcardiotomy cardiogenic shock following operations for congenital heart disease.5–9 However, reports of ECMO for postcardiotomy cardiac failure primarily involve patients undergoing corrective procedures and very few patients with acute graft failure following heart transplantation. There are no single center reports of the use of ECMO for acute graft failure following heart transplantation in infancy. The purposes of this study were: (1) to review our experience with ECMO in infants with acute graft failure following heart transplantation, (2) to delineate risk factors for acute graft failure requiring ECMO in infants, and (3) to determine the long-term outcome of transplanted infants successfully salvaged with ECMO following acute graft failure.

METHODS Patient population The clinical records and computerized transplant database of all patients listed for heart transplantation at the age of 6 months or less who underwent cardiac transplantation in our center from the inception of our program through December 1999 were retrospectively reviewed. The ECMO group included all patients within this population who were placed on ECMO in the early postoperative period. The non-ECMO group included all patients not requiring ECMO who underwent transplantation after we extended our ECMO program to transplant patients (i.e., November 1993). All patients operated prior to this date were excluded. No patients in this age group were placed on ECMO for delayed hemodynamic deterioration due to rejection. All children surviving heart transplantation were followed regularly in our transplant clinic. No patients were lost to follow-up although several non-ECMO patients changed to other cooperating transplant centers that provided follow-up. Graft function was regularly assessed by transthoracic echocardiography. The heterogeneous group of congenital heart diseases in this series cannot allow meaningful risk analysis by individual lesion. There-

Mitchell et al.

835

fore, diagnoses were categorized as HLHS or nonHLHS. Patients in the HLHS spectrum with less severe hypoplasia of the left heart (i.e., without mitral or aortic atresia) but having a left ventricle judged inadequate for biventricular repair were categorized as HLHS. The non-HLHS group was comprised of infants with other complex uncorrectable congenital heart defects and cardiomyopathies.

Transplant protocol Preoperative, perioperative, and postoperative management of infant heart transplant patients was as previously described.4 Perioperative pulmonary artery pressures and mixed venous oxygen saturations were continuously measured in all patients using oximetric pulmonary artery catheters (Abbot Laboratories, Chicago, IL). Donor hearts were procured in standard fashion including extra-great vessel lengths as dictated by recipient anatomy. Roe’s solution was used for donor cardioplegia. Isoproterenol and dopamine were routinely used for inotropic support. Dobutamine and epinephrine were used when additional inotropic support was required. Inhaled nitric oxide and intravenous milrinone infusions were routinely used when elevated pulmonary vascular resistance was anticipated or encountered. Perioperative immunosuppression consisted of cyclosporin, azathioprine, and methylprednisolone. Induction therapy with antithymocyte serum (American Medical Resources, Nashville, TN) was used with discontinuation of methylprednisolone 48 hours post-transplant. In 1998 induction therapy was changed to thymoglobulin (SangStat Medical Corp., Fremont, CA). Induction therapy was routinely withheld until chest closure was achieved in all patients requiring delayed sternal closure.

Extracorporeal membrane oxygenation In our center ECMO was first used for an infant with early posttransplant graft failure in November 1993. The ECMO perfusion circuit consisted of a servo-regulated roller pump (Cobe Laboratories, Arvada, CO), polyvinyl tubing (Cobe Laboratories, Arvada, CO), a membrane oxygenator (Avecor Cardiovascular, Inc, Plymouth, MN), and heat exchanger. All patients requiring ECMO were cannulated for venoarterial ECMO through the open chest utilizing the right atrium and ascending aorta. Activated clotting times (ACT) were maintained between 180 and 220 seconds with intravenous heparin. After establishing hemodynamic stability inotropic support was minimized. Hemofiltration

836

Mitchell et al.

was used as required until renal function stabilized. Following recovery, patients were weaned and decannulated using inotropic support as needed. Chest closure was usually achieved within 48 hours of decannulation.

Definitions and statistical analysis Early death was defined as mortality occurring within 30 days of transplant. Late deaths were defined as deaths occurring after 30 days. Variables considered potentially predictive of the need for ECMO (age at listing, age at transplant, wait time, body weight at transplant, donor/recipient weight ratio, total graft ischemia time, and diagnostic category) were compared between ECMO patients and non-ECMO patients. Fisher’s exact test was used to compare categorical variables, and nonparametric Wilcoxon’s rank sum test was used for continuous variables. Intensive care and hospital stays were compared between ECMO survivors and nonECMO surviving patients using the Wilcoxon’s rank sum test. Survival curves were estimated using the Kaplan-Meier method. A p-value ⬍ 0.05 was considered significant. All statistical analyses were performed with S Plus version 4.5 (Mathsoft, Inc., Seattle, WA).

RESULTS From November 1993 through December 1999, 63 patients listed at age less than 6 months underwent heart transplantation in our center. Forty-six patients had HLHS. Seventeen patients had nonHLHS diagnoses (16 complex congenital heart disease, 1 cardiomyopathy). Fourteen patients were placed on ECMO during the early postoperative period. Eight patients placed on ECMO survived and were discharged alive (57%).

Survival of ECMO vs non-ECMO patients Age at listing, age at transplant, waiting time to transplant, body weight at transplant, donor/recipient weight ratio, and total graft ischemia time were compared between 14 ECMO and 49 non-ECMO patients. By univariate analysis no variable predicted the need for ECMO (Table I). Twelve of 46 patients with HLHS required ECMO (24%). Two of 17 non-HLHS patients required ECMO (12%). Therefore, 12 of the 14 (86%) patients requiring ECMO were patients with HLHS. However, the diagnosis of HLHS vs non-HLHS was not a significant risk factor for requiring ECMO (p ⫽ 0.32). Similarly, the waiting time to transplant analyzed only for those patients with HLHS was not a risk factor for ECMO

The Journal of Heart and Lung Transplantation September 2000

TABLE I Non-ECMO vs ECMO patients Variable Age at listing (days) Age at transplant (days) Wait time (days) Weight at transplant (kg) Donor/recipient weight ratio Ischemic time (minutes)

Non-ECMO (n ⴝ 49)

ECMO (n ⴝ 14)

p-value

37 ⫾ 50 15.4 ⫾ 13 110 ⫾ 69 80 ⫾ 53 74 ⫾ 53 65 ⫾ 48 4.7 ⫾ 0.9 4.3 ⫾ 1.0 2.1 ⫾ 0.6 1.9 ⫾ 0.9

0.45 0.24 0.70 0.08 0.91

239 ⫾ 86

0.31

258 ⫾ 72

Comparisons by Wilcoxon’s rank-sum test. Values are mean ⫾ standard deviation.

(p ⫽ 0.52). Not surprisingly, hospital and ICU stays of surviving patients requiring ECMO were longer than those of survivors not requiring ECMO. Median hospital stay was 33 days for the 8 ECMO survivors vs 9 days for the 47 non-ECMO early survivors (p ⫽ 0.0004). Median ICU stay was 29 days for ECMO survivors vs 7 days for non-ECMO survivors (p ⬍ 0.0003).

Timing and indications for ECMO The mean time from the cessation of cardiopulmonary bypass to the institution of ECMO was 9.4 hours (range: 0 – 48 hours). Indications for ECMO were failure to wean from cardiopulmonary bypass (n ⫽ 3), cardiac arrest in the intensive care unit following transplantation (n ⫽ 4), and progressive hemodynamic deterioration following transplantation (n ⫽ 7). Of the 3 patients who failed to wean from cardiopulmonary bypass and were placed directly on ECMO, 2 survived. Four of 7 patients with progressive hemodynamic deterioration survived, and 2 of 4 patients suffering cardiac arrest survived. There were no differences in survival based on the indication for ECMO. Ten of the 14 patients placed on ECMO were weaned from ECMO (71%). One child who was initially weaned from ECMO after 41.25 hours of support required reinstitution of ECMO 30 hours after initial weaning. This child subsequently died after an additional 186 hours of ECMO support. For purposes of comparison, this child was considered not weaned from ECMO, and the duration of ECMO support was considered to be the sum of both support times. The mean duration of ECMO support in weaned patients was 68 hours vs 144 hours in patients not weaned from ECMO (p ⫽ 0.19). Two patients were successfully weaned from ECMO but later died. The mean duration of ECMO

The Journal of Heart and Lung Transplantation Volume 19, Number 9

Mitchell et al.

837

TABLE II ECMO survivors

Patient 1 2 4 9 10 11 13 14

Diagnosis

Time to ECMO (hours)

ECMO duration (hours)

Chest closure (days)

Complications

Follow-up (months)

HLHS HLHS HLHS HLHS/TAPVD HLHS HLHS CM HLHS

6 0 0 8 12 11 4 6

43 83 70 68 77 50 55 65

1 3 1 1 1 1 5 2

paraplegia none none renal failure mediastinitis none chylothorax, CMV subdural hematoma

78 65 54 22 21 21 17 13

Patient numbers are chronological out of 14 ECMO patients. Complications do not include hemorrhage requiring reexploration. CM, cardiomyopthy; CMV, cytomegalovirus pneumonitis; HLHS, hypoplastic left heart syndrome; TAPVD, totally anomalous pulmonary venous drainage.

support in survivors was 64 hours vs 123 hours in nonsurvivors (p ⫽ 0.35).

Support was withdrawn 11 days later when the child was declared brain dead.

ECMO morbidity and mortality

Long-term outcome

Hemorrhage requiring reexploration was the most common complication of ECMO. Only 4 patients did not require reexploration for bleeding. All 4 of these patients were placed on ECMO for progressive hemodynamic deterioration that occurred after initial hemostasis had been achieved. In every case delayed chest closure was required. Complications and the need for additional procedures were common in surviving patients who required ECMO (see Table II). Clinical details of the 6 patients requiring ECMO who died are presented in Table III. Autopsies were obtained in all cases. The most common cause of death was sepsis (n ⫽ 3). Unrelenting pulmonary hypertension and right heart failure caused 2 deaths. One child weaned from ECMO after 40 hours of support suffered a massive intracranial hemorrhage prior to coming off ECMO.

At a mean follow-up of 36.2 months all 8 patients requiring ECMO who survived the early postoperative period remain alive. The long-term survival in this cohort was not different from patients who did not require ECMO and were discharged alive. Five of the 8 ECMO survivors have had essentially normal long-term clinical courses. The child (Table III, patient 1) who sustained an ischemic spinal-cord injury resulting in paraplegia has severe developmental delay. This child has subsequently been diagnosed with a hereditary neurologic disorder (Smith-Lemil-Opitz syndrome) that has contributed to her outcome. The child (Table III, patient 14) who suffered a subdural hematoma has a permanent right hemiparesis and 3rd cranial nerve palsy. Finally, the child (Table III, patient 13) who developed a chylothorax and cytomegalovirus (CMV)

TABLE III ECMO deaths

Patient 3 5 6 7 8 12

Diagnosis

Time to ECMO (hours)

ECMO duration (hours)

Weaned

Cause of death

HLHS Asplenia/Unbalanced AVSD HLHS HLHS HLHS HLHS/TAPVD

9 48 10 10 7 0

184 4 124 40 226 160

no no yes yes no no

sepsis sepsis pulmonary hypertension intracranial bleed sepsis pulmonary hypertension

Patient numbers are chronological out of 14 ECMO patients. AVSD, atrioventricular septal defect; HLHS, hypoplastic left heart syndrome; TAPVD, totally anomalous pulmonary venous drainage.

838

Mitchell et al.

pneumonitis has had significant right heart failure due to severe tricuspid regurgitation (biopsy-induced injury) and residual pulmonary hypertension. This child underwent a successful tricuspid valve repair 15 months after the initial transplant and is improved.

DISCUSSION Because of size constraints and technical considerations ECMO is the most prevalent form of circulatory support used in pediatric patients with postcardiotomy cardiogenic shock.9 Several large pediatric centers have published their experiences with ECMO in patients with postcardiotomy cardiogenic shock.5–9 However, in contrast to our series, these reports have all contained a heterogeneous mix of patient ages, diagnoses, and surgical procedures. Furthermore, very few patients with early graft failure following cardiac transplantation have been reported. The only prior study in the literature that specifically addressed the use of ECMO in pediatric heart-transplant patients is a report of combined data from the Extracorporeal Life Support Organization Registry; the Combined Registry of Mechanical Ventricular Assist Devices and Total Artificial Hearts; and published case reports.10 Of 10 patients with early graft failure supported by ECMO in this report, only 4 patients (2 survivors) were infants. In a series of 36 infants undergoing transplantation, Twedell et al mentioned using ECMO in 7 patients (all unpalliated HLHS) with 5 survivors.11 Our data taken together with that of Twedell et al indicate that ECMO is a useful adjunct for infant heart transplantation. Hospital survival in our patients requiring ECMO was comparable to that of other reports of ECMO for postcardiotomy support (range: 40 –70%).5–9 Other investigators have noted that the presence of residual cardiac defects in children placed on ECMO following corrective cardiac procedures is a strong predictor of mortality.7 Consequently, the expected lack of residual defects in transplant patients theoretically renders this group ideal for recovery when ECMO is required. However, our experience did not demonstrate superior survival. One reason is that immunosuppressed transplant patients are at increased risk of infection as indicated by the fact that sepsis was the leading cause of death and a significant cause of morbidity in our patients. A second factor is the difficulty of assessing the magnitude and reversibility of pulmonary hypertension in infants awaiting cardiac transplantation.11,12

The Journal of Heart and Lung Transplantation September 2000

All but 1 (Table III, patient 5) of our patients requiring ECMO exhibited severe right-heart dysfunction and pulmonary hypertension. Patients with unpalliated HLHS dominated our series, and these patients obligatorily have systemic pulmonary artery pressures while waiting for donor organs. Although the diagnosis of HLHS was not a risk factor for ECMO, the smaller number of non-HLHS patients may not have allowed an adequate comparison. Given the 24% incidence of ECMO in HLHS patients and that 12 of 14 patients placed on ECMO had HLHS, we consider patients with this diagnosis at higher risk of acute graft failure. In addition, the other non-HLHS patient (Table II, patient 13) who required ECMO also demonstrated clinical findings of pulmonary hypertension and donor right-heart failure that did not respond to maximal medical therapy. Lastly, 2 ECMO patients with HLHS who died had autopsy findings consistent with irreversible pulmonary hypertension. Previous reports indicate that the duration of ECMO support in postcardiotomy patients is related to the likelihood of successful weaning from ECMO. Black et al reported that no postcardiotomy patient requiring ECMO longer than 6 days survived.7 Similarly, Duncan and coworkers found that the lack of return of ventricular function within 72 hours of the initiation of support predicted failure to wean from ECMO.9 With the relatively small number of patients in our cohort, we could not detect a difference in the duration of support between weaned and nonweaned patients or between survivors and nonsurvivors. However, no patient supported longer than 83 hours survived. Also, the difference in ECMO durations between weaned and nonweaned patients may have been obscured by the patient with the shortest EMCO duration in our series. Of note, this child (Table III, patient 5) was placed on ECMO 48 hours after transplantation while all other patients had ECMO instituted within 12 hours of transplantation. He subsequently proved to have overwhelming sepsis, and support was terminated after 4 hours when adequate perfusion could not be restored. Retrospectively, this child clearly differed from all other patients in that hemodynamic deterioration was not due to acute graft failure from myocardial stunning or excessive pulmonary vascular resistance. Excluding this patient, the mean duration of support for weaned patients was 67.5 hours vs 190 hours for nonweaned patients (p ⫽ 0.007). However, there was only a suggestion of a difference (p ⫽ 0.09) for survivors (mean 63.4 hours) vs nonsurvivors (mean 147 hours). Thus, our

The Journal of Heart and Lung Transplantation Volume 19, Number 9

experience suggests that patients not weaned from ECMO within 4 days should either be relisted for transplantation or termination of support should be considered. Because the mean waiting time for a donor heart in this age group in our center is over 2 months, we have not relisted any of our patients. Secondly, contraindications to transplantation (e.g., sepsis) are frequently present after this duration of mechanical support. Long-term survival of our patients salvaged with ECMO has been 100%. Similarly, Ibrahim et al, have recently reported 96% survival at a median follow-up of 41 months.13 However, these authors also noted a 59% incidence of moderate to severe neurologic impairment in long-term ECMO survivors. Although our series consisted of fewer patients, the incidence of neurologic impairment in our survivors was 25%. Other series have described permanent neurologic impairment in 10 to 25% of cases.5,7 Despite the significant incidence of major complications experienced in our survivors, those not having neurologic complications have had posttransplant courses comparable to infant transplant survivors not requiring ECMO. In conclusion, our results indicate that ECMO is a useful modality for resuscitating infants with acute graft failure early after heart transplantation. We were unable to identify risk factors that would help predict the need for ECMO in infants undergoing heart transplantation. Despite the inherent risks of immunosuppressing infants with transthoracic cannulation, survival is comparable to other reports of patients with postcardiotomy cardiac failure supported with ECMO. The long-term outcome of patients salvaged with ECMO is good. However, neurologic impairment remains a significant problem. Efforts to identify patients who are at increased risk for early donor-heart failure deserve further attention. Finally, careful assessment of pulmonary vascular reactivity in infants awaiting transplantation and methods of protecting the pulmonary vascular bed from increased flow in these patients warrant further investigation.

Mitchell et al.

839

The authors are grateful to Todd Mackenzie, PhD for his assistance with the statistical analysis in this manuscript.

REFERENCES 1. Boucek MM, Faro A, Novick RJ, et al. The registry of the International Society of Heart and Lung Transplantation: third official pediatric report—1999. J Heart Lung Transplant 1999;18:1151–72. 2. Hosenpud JD, Bennett LE, Keck BM, et al. The registry of the International Society for Heart and Lung Transplantation: sixteenth official report—1999. J Heart Lung Transplant 1999;18:611–26. 3. Shaddy RE, Naftel DC, Kirklin JK, et al. Outcome of cardiac transplantation in children: survival in a contemporary multiinstitutional experience. Circulation 1996;94(suppl II):II-69 – II-73. 4. Mitchell MB, Campbell DN, Clarke DR, et al. Infant heart transplantation: improved intermediate results. J Thorac Cardiovasc Surg 1998;116:242–52. 5. Kanter KR, Pennington DG, Weber TR, et al. Extracorporeal membrane oxygenation for postoperative cardiac support in children. J Thorac Cardiovasc Surg 1987;93:27–35. 6. Rogers AJ, Trento A, Siewers RD, et al. Extracorporeal membrane oxygenation for postcardiotomy cardiogenic shock in children. Ann Thorac Surg 1989;47:903– 6. 7. Black MD, Coles JG, Williams WD, et al. Determinants of success in pediatric cardiac patients undergoing extracorporeal membrane oxygenation. Ann Thorac Surg 1995;60: 133– 8. 8. Walters HL, Hakimi M, Rice MD, et al. Pediatric cardiac surgical ECMO: multivariate analysis of risk factors for hospital death. Ann Thorac Surg 1995;60:329 –37. 9. Duncan BW, Hraska V, Jonas RA, et al. Mechanical circulatory support in children with cardiac disease. J Thorac Cardiovasc Surg 1999;117:529 – 42. 10. Galantowicz ME, Stolar CJH. Extracorporeal membrane oxygenation for perioperative support in pediatric heart transplantation. J Thorac Cardiovasc Surg 1991;102:148 –52. 11. Twedell JS, Canter CE, Bridges ND, Moorhead S, Huddleston CB, Spray TL. Predictors of operative mortality and morbidity after infant heart transplantation. Ann Thorac Surg 1994;58:972–7. 12. Chinnock RE, Larsen RL, Emery JR, Bailey LL. Pretransplant risk factors and causes of death or graft loss after heart transplantation during early infancy. Circulation 1995; 92(suppl II):II-206 –II-209. 13. Ibrahim AE, Duncan BW, Blume ED, Jonas RA. Long-term follow-up of pediatric cardiac patients requiring mechanical circulatory support. Ann Thorac Surg 2000;69:186 –92.