Abstract
Objectives
Primary graft failure (PGF) after heart transplantation is a detrimental complication, and carries high morbidity and mortality. The aim of this study was to analyze the results of our multidisciplinary approach in supporting patients affected with PGF after heart transplantation.
Methods
Out of 114 consecutive patients receiving orthotopic heart transplantation between January 2006 and July 2013, 18 (15.7%) developed PGF requiring veno-arterial extracorporeal membrane oxygenator (VA-ECMO) support. Fourteen patients were male and the mean age was 49 ± 11 years. General principles in treating the patients were based on a low dose of adrenaline (0.05 mic/kg per min) infusion; femoral intra-aortic balloon pump (13 of the 18 patients); low dose of vasoconstrictors; careful fluid balance; daily echocardiographic transesophageal monitoring.
Results
Mean graft recipient pulmonary vascular resistance was 3.6 ± 3.2 WU. Five patients had absolute contraindication to IABP placement. The mean left ventricle ejection fraction pre-VA-ECMO was 18.4% ± 10.2%. The mean VA-ECMO and IABP support times were 6.7 ± 3.2 and 9.2 ± 7.6 days, respectively. Mean VA-ECMO flow was 4164 ± 679 l/min. The mean left ventricle ejection fraction increased to 43.4% ± 17.7% at the end of support. Weaning and discharge rates in patients treated with VA-ECMO+IABP were 84% and 53%, respectively. Causes of death were primarily end-stage organ failure.
Conclusions
A multidisciplinary evaluation of ECMO patients done by intensivists, cardiologists, and surgeons may influence weaning and survival rate. Our approach seems to be a safe and reproducible strategy for avoiding left ventricle distension and fluid overload, and for detecting complications that negatively affect outcomes.
Introduction
The use of a veno-arterial extracorporeal membrane oxygenator (VA-ECMO) in severe primary graft failure (PGF) after heart transplantation (HTx) is widely accepted (1–7). ECMO allows right-chamber unloading and sufficient circulatory support, though left ventricle (LV) distension can occur, resulting in LV stasis, thrombosis and failure of the aortic valve to open (8–9). In PGF, the loss of LV compliance, but maintenance of mitral valve competence, leads to an increase in endoventricular pressure, a rise in wall tension, a decrease in coronary flow, and reduction in the rate of cardiac recovery (10).
Many surgical or percutaneous solutions have been described to relieve LV overload during ECMO support (mainly in non-PGF patients), though most are anecdotic, and still imply further risk of bleeding from puncture site, heart trauma or tamponade (11–12). A multidisciplinary assessment with the goal of preventing rather than treating ECMO-support-related complications can increase the recovery rate.
Since the start of our heart transplant program, in 2006, 114 transplants have been performed at our institute. Nineteen patients (16.6%) suffered from PGF, requiring a ventricle assist support device: one patient was assisted by a biventricular centrifugal pump, while the other 18 received VA-ECMO. These 18 patients were our study population.
The aim of our study was to analyze the role of a multidisciplinary approach to the prevention and detection of complications, cardiac recovery, and readiness for ECMO weaning in severe PGF patients.
Materials and Methods
Approval of the Institutional Review Board was obtained. We retrospectively analyzed all patients who had undergone isolated HTx at our institute, and needed extracorporeal circulatory support for severe PGF. Severe PGF was defined, according to the International Society of Heart and Lung Transplantation consensus (13), as dependence on left or biventricular mechanical support including ECMO, LVAD, BiVAD, or percutaneous LVAD, despite the maximum medical management (included counterpulsation). The male/female ratio was 14/4, and the mean age was 49 ± 11 years. Body surface area and body mass index were 1.8 ± 0.3 m2 and 26 ± 4, respectively. Donor and recipient characteristics are listed in Table I.
Patient Characteristics
IV = intra venous; IABP = intra aortic balloon pump; VAD = ventricle assist device; CI = cardiac index; PAPm = mean arterial pulmonary pressure; PCWP = pulmonary capillary wedge pressure; PVR = pulmonary vascular resistance.
Donor with high inotropic support is intended: more than 0.06 mic·min-1·kg-1 of norepinephrine and/or more than 10 mic·kg-1·min-1 of dopamine/dobutamine until the time of the retrieval.
Surgical Technique
Procurement and Transplantation
The CelsiorTM cardioplegic solution (Genzyme, Cambridge, MA, USA) was used for the procurement of the donor hearts, with the infusion of two liters of solution to achieve plegic arrest, and one liter in the bag for the transport. To protect the heart during the implantation, the heart was covered by an iced gauze (to avoid direct contact between the ice and the epicardium) and intermittent cold bloody cardioplegia was administered every 20 to 30 minutes.
The transplants were done using both the Lower/Shumway and bicaval technique, depending on the surgeon. Four surgeons are authorized to do heart transplantation at our institute.
ECMO
Cannulation was performed peripherally or centrally through median sternotomy. Peripheral cannulation was used at the beginning of the activity, between 2006 and 2008, for the first patients. In those years we observed an elevated rate of peripheral complications (despite the perfusion cannula) as well as worse venous drainage and few cases of arterial flow competition with severe upper body hypoxiemia. For these reasons, we changed our approach and we swapped to central cannulation in all patients undergoing postcardiotomy and primary graft failure. The peripheral approach was reserved to acute heart failure patients.
Considering the high risk of bleeding, especially during the first few hours, and the high risk of tamponade with consequent ECMO low flow and severe low output syndrome, all patients were left electively with the sternum open and the skin closed with a transparent plastic patch. This strategy allows a quick check for mediastinal clotting formation and initial tamponade; furthermore, it allows direct access to the mediastinum if an urgent bedside re-exploration is necessary.
Cannulation Techniques
In cases of peripheral cannulation, the femoral vessels were surgically isolated and cannulated under direct visualization. A reperfusion cannula was connected to the arterial line and placed through the superficial femoral artery (routinely done since the beginning of 2007). A Doppler check of the pedal pulse was routinely performed twice a day.
Regarding the central cannulation, the cannula used for the aortic perfusion and for the inferior vena cava drainage during the cardiopulmonary bypass (CPB) were re-utilized for the ECMO. The aortic cannula was left in situ to avoid new aortic puncturing, while the inferior vena cava cannula was placed in the right atrium through the right appendage. During the repositioning of the venous cannula, CPB was made possible via the superior vena cava cannula. Eventually, the CPB was suspended to switch the circuits and then the ECMO assistance was started.
System and cannulas
The ECMO system used for all patients was the Quadrox PLS Set (Maquet, Rastatt, Germany). The cannulas used for venous and arterial peripheral accesses were the femoral vein with DurafloTM Coating and Fem-Flex II (Edwards LifescienceTM, Irvine, CA, USA), respectively. DLP flexible arch arterial and venous single stage cannulas (Medtronic, Minneapolis, MN, USA) were used for central ECMO.
Anticoagulation
Anticoagulation was achieved with intravenous unfractionated heparin (PTT range of 1.5-2 times the normal value), and usually started when chest drainage loss dropped to less than 50 ml/h for four consecutive hours. PTT was checked every six hours.
ECMO Weaning Protocol
If the Ejection Fraction (EF) on full support reaches 40%, the patient is considered for a weaning trial under transesophageal echocardiographic (TEE) view. The institutional protocol consists in an initial reduction of the support to 50% of the theoretical flow for about 10 minutes. If the EF does not worsen and no mitral regurgitation or LV distention occurs, the ECMO flow is further reduced to 25% for about five minutes and the IABP held. If the echocardiograms do not show any cardiac distress, the weaning can start.
The protocol for the weaning consists in a progressive, daily reduction of the support and at least a daily TEE check:
First reduction to 75% support for 24 h
If hemodynamically stable, with low lactate and diuresis preserved, a TEE is performed to confirm the EF >40% and no worsening of mitral regurgitation or LV distension
Further reduction to 50% support for 24 h.
If hemodynamically stable, with low lactate and diuresis preserved, a TEE is performed to confirm the EF >40% and no worsening of mitral regurgitation or LV distension.
The patient is sent to the operating room. With the chest open, the ECMO support is kept at 25% for about one hour and if the TEE confirms good functioning, the ECMO is stopped the cannulas are pulled out.
During partial support, we usually increase the inotropic support (epinephrine 0.1 mics·kg-1·min-1).
Echocardiography
A comprehensive transthoracic (TTE) or TEE examination was done with commercially available cardiac ultrasound machines (Vivid System Seven and Vivid I, GE/Vingmed, Milwaukee, WI, USA). Pulsed, continuous, and color-flow Doppler techniques were used to investigate transvalvular flows and deceleration times (14–15). All images were digitally recorded for off-line analysis.
The timing of the echocardiogram corresponded with the time of the ECMO placement, and on each of the following 3 days. For cases in which complications were detected, the exam was repeated as needed.
Intensive Care Unit Management
Vasoactive drugs were administered in all patients, principally adrenaline to keep some ventricular contractility, and norepinephrine to counteract the frequent vasoplegia and reduce fluid intake. In two cases, vasopressin was necessary. The dosage of epinephrine was kept low during the full assistance (0.05 mic·kg-1·min-1, mainly the first three days of support), and increased during the weaning trials. Fluid management is shown in Table II. Blood products were infused primarily on the first two postoperative days, when the coagulopathy is generally more evident and the hemodynamics are more unstable. Generally, crystalloids were not used for fluid resuscitation, while albumin was infused only when a severely low serum albumin level was detected. The mean total fluid balance was positive during the first two days, given the amount of blood transfused because of perioperative bleeding. On the third postoperative day, the balance became negative in most of the patients, likely due to stabilization of the hemodynamics, and resolution of the coagulopathy.
Intensive Care Management and Laboratory Tests
IV = intravenous; PRBC = packed red blood cells; CRRT = continuous renal replacement therapy; SGOT = serum glutamic oxaloacetic transaminase; LVEF = left ventricle ejection fraction.
Day 1 vs. end of support.
Time of implant vs. end of support.
Intensive Care Multidisciplinary Management
The main principles of our multidisciplinary management in the intensive care unit tended to minimize fluid overload, allow the aortic valve to open, and optimize the residual cardiac function to keep some blood flow through the heart chambers.
The general principles were:
Low dose of adrenaline (0.05 mic/kg per min) infusion to optimize residual cardiac function;
Femoral IABP placement to allow aortic valve opening, decompressing the LV, and increasing coronary flow;
Low dose of vasoconstrictor (noradrenaline or vasopressin) and careful fluid balance to avoid fluid overload;
TEE monitoring of aortic valve opening, mitral regurgitation (MR), and LV distension.
Data collection
Several data sets were collected and analyzed: pre- and post-assistance, including demographic and laboratory data (creatinine, lactate, and liver function test); hourly perioperative invasive hemodynamic parameters (cardiac output, cardiac index, pulmonary pressure, wedge pressure, right atrial pressure, systemic arterial pressure); daily fluid balance; blood product transfusions; inotropic infusions; use of IABP and days of support; days of ECMO support, and mean ECMO flow. The number and type of complications and survival were registered. Categorized variables are expressed as percentage, and continuous variables as mean +/- standard deviation.
Results
PGF was characterized mainly by right and left ventricle dysfunction and a mean LV ejection fraction of 15 ± 9% before initiation of ECMO. All ECMO supports were set up in the operating room after three failed attempts of weaning from CPB. The IABP could be placed in the operating room in 13 patients (72.2%), while 5 had anatomical contraindications (mainly small vessels or peripheral disease). The mean IABP support time was 9.2 ± 7.6 days. All patients received TEE monitoring, even those with no complications.
For ECMO support, 14 patients were supported using central cannulation, while 3 were peripherally cannulated, and one patient received central arterial cannulation and peripheral venous cannulation (for severe femoral artery calcification). Details of the extracorporeal support are shown in Table III. Mean ECMO flow was 4164 ± 679 l/min, and in four cases the flow spontaneously dropped to less than 2 l/min due to bleeding or cardiac tamponade, with the patient requiring urgent re-exploration.
ECMO Characteristics and Details
ECMO = extracorporeal membrane oxygenator.
The support achieved good end-organ perfusion, with reduction of lactate, creatinine, and SGOT (Tab. II).
Outcomes
Thirteen patients (72.2%) were able to be weaned from the support, and eight of them (44%) were eventually discharged. A better trend was noted in patients with IABP: the weaning rate was 84% (11 of the 13 patients) and the survival was 53% (7 patients). In the group of patients without IABP (5 patients) the weaning rate was 40% (2 patients), and the survival 20% (1 patient) (Tab. IV). Cause of death of the patients weaned from the support was multi-organ failure, with sepsis in four of them and acute mycotic rupture of pulmonary artery in the last one.
Ventricle Assist Device, Use of Iabp and Outcome
IABP = intra aortic balloon pump; VA-ECMO = venoarterial ECMO.
The complication rate is shown in Table V. The most frequent complications were renal and respiratory failure, and bleeding requiring surgical re-exploration. Peripheral vascular complications were recorded in four patients, three of whom after peripheral ECMO placement: before the preventive use of the reperfusion cannula in the femoral artery in two patients, and after IABP placement in the other. One patient, without IABP and no aortic valve opening, experienced a ventricular thrombosis requiring thrombectomy and subsequent placement of IABP. There was no LV distension or need to place a left sided additional inflow cannula.
Postoperative ECMO Complications
ECMO = extracorporeal membrane oxygenator; LV = left ventricle.
All the patients required tracheostomy after a failed respiratory weaning.
Echocardiography
All patients received at least one TEE per day and repeated as needed. Daily echocardiogram confirmed trace or mild MR in 15 patients, and moderate MR in 4 patients. No severe mitral regurgitation was detected. Mean LV function increased during circulatory support in all but two patients after three days (mean EF 18.4% +/- 10.2% vs. 43.4% +/-16.7%, p<0.001). Figure 1 shows the echographic parameters.
Daily echocardiographic evaluation of mitral regurgitation, tricuspid regurgitation, aortic valve opening, and ejection fraction. AV = aortic valve; ECMO = extracorporeal membrane oxygenator.
The aortic valve was systematically analyzed only for the last two years of the study period. A total of 13 patients were analyzed. In three patients we needed to reduce the ECMO RPM to allow a correct opening of the aortic valve (at least every two to three beats). In two cases of severely compromised LV, despite a reduction in the afterload in both, and IABP placement in one, the AV valve failed to open, and after 24 h, a ventricular thrombosis was suspected. These patients were sent to the operating room. The patient with the IABP had no thrombi in the chambers, while the other one needed a thrombectomy and subsequent IABP placement to allow opening of the AV. Cardiac tamponade, despite acceptable ECMO flows, was diagnosed in four cases. No cannula malpositions were detected.
The decision for weaning the patient was made when the LV EF reached 40% and no worsening of MR or LV distension occurred after a trial of ECMO flow reduction.
Discussion
Primary graft failure is one of the strongest risk factors for 30-day mortality after heart transplantation, with a slightly increasing incidence over the recent years, probably because of the extensive use of marginal donors (2–3). The rate of PGF after heart transplantation is reported as between 3% and 26%, with mortality between 20% and 50%, which is in line with our data (4–7, 16).
Both centrifugal pumps and ECMO have been extensively described in the treatment of PGF, with different approaches mainly determined by the expertise of the center. At present, ECMO is widely accepted for postcardiotomy and PGF, given the easier setup, the versatility, and the possibility of biventricular assistance and lung assistance with a single pump. Despite the obvious advantages in terms of end- organ perfusion, ECMO still carries a high risk of severe complications and mortality. Consequently, every effort should be made to guarantee appropriate flows, prevent the most life-threatening complications, such as intraventricular thrombosis and cardiac tamponade, and, overall, facilitate proper cardiac recovery.
Indeed, avoiding left ventricular distension and stasis during ECMO, ensuring that the aortic valve opens and some blood flows through the chambers, should be one of the principal goals, particularly when left ventricular function is severely compromised (8–10). This can be achieved with a continuous multidisciplinary approach involving intensivists, cardiologists, and surgeons. With severely compromised hemodynamics, it is not easy to strictly follow a general protocol, especially the goal of negative fluid balance. With a longitudinal analysis, we observed a kind of learning curve with more strict observation of some general principles, and identification of the roles of physicians in the management of the patients.
The responsibility of the intensivist is to achieve strict fluid balance to avoid fluid overload, the appropriate use of inodilators, and prompt communication with the ECMO team. TEE has become part of the management of monitoring ECMO, from the initiation of support to the decision to wean in order to detect complications before there is clinical evidence. Several echocardiography parameters, as well as an ejection fraction >35% to 40%, an LV outflow tract velocity-time integral >10 cm, absence of LV dilatation, the aortic valve opening rate, and, eventually, mitral regurgitation have been suggested as predictors of successful weaning (14, 17, 18). To the best of our knowledge, most of the echographic evaluations are reported for ECMO in decompensated heart failure (14–15), while PGF is a different condition, with specific patterns (19). So, given the high probability of recovery, making sure that the LV is well perfused but not overloaded, following the opening of the aortic valve and the LV dimension, and ruling out the most common complications can help improve the outcome.
Finally, the surgeon is integrated with the other physicians, and is responsible for the right timing of surgical revisions to reduce blood product transfusions, and for evacuation of pericardial collections before a drop in the flows and perfusion, which would necessitate an excessive use of vasoconstrictors.
With this approach, we have been able to ensure good cardiovascular support in all patients, with a significant reduction in lactate in the first 24 h, and a progressive normalization of the end-organ function test in most of the patients. The support has allowed almost 70% of the patients with PGF to recover heart function, and be weaned from support.
The role of IABP placement in aiding recovery is still a matter of debate. The beneficial effects of adding IABP support to VA ECMO on coronary perfusion and on reduction of the afterload have been well demonstrated in animal models (20–21), and single-center experiences have shown the efficacy in terms of outcome compared with no IABP use (22–23). Nevertheless, some have abandoned the use of IABP with ECMO because of the increased risk of vascular complications (16). Analyzing our data, in the group supported with IABP, the weaning rate was 84%, with a survival of 53%, while in the group without IABP the weaning rate was 40%, and the survival 20%. Moreover, analyzing the recovery of LV function patient by patient, the only two patients with no improvement in the ejection fraction were in the non-IABP group. Of course these data have no statistical weight, but still need to be considered. We found no other differences between the two groups, probably because the reduction in the afterload, and the increase in the coronary and pulsatile flows, offered by the IABP have a protective effect. In addition, we observed the effect of the absence and then of the presence of the IABP in the patient with the LV thrombosis. Initially the balloon was not placed because of some concerns about the size of the femoral vessels, then, after the trombectomy, the IABP was placed, and the aortic valve started to open, with no further occurrence of thrombi. It is likely that allowing even a small flow through the LV, reducing the afterload, and giving a pulsatile flow, can help the heart to recover.
Our study has several limitations. First, it is retrospective and non-controlled. Ideally, a prospective randomized trial would be indicated to better understand the impact of IABP placement on the outcome in PGF patients, though the nature of the pathology, with its inherently extreme hemodynamic instability would not allow for randomization and strict adherence to the two different protocols for IABP and non-IABP patients. Second, ventricular failure was not the same in all the patients: in some cases the right ventricle was worse than the left, so the role of IABP can become more questionable. Nonetheless, a multidisciplinary approach is still valid, and even when the right ventricle is worse than the left, the balloon helps increase the coronary perfusion of the right ventricle. Third, the rate of PGF requiring mechanical support in our series is higher than that reported in the literature, perhaps because of our extensive use of donors with high dosages of α-agonists and inotrops, as we reported in a previous, published study (24). Despite the fact that risks inherent in PGF are well known, the shortage of donors has increasingly worsened, and, like many centers, we have extended our criteria to accept marginal donors, particularly donors with high α-agonist or inotropic dosages.
In conclusion, a multidisciplinary approach, with routine use of an IABP, together with low doses of inotropic infusions and careful fluid management seems to help the recovery of the left ventricle after PGF. We may need a multicenter study, with a larger population, for more definitive results.
TEE monitoring of the right ECMO, from the initiation of support to the decision to wean, can help detect complications before there is clinical evidence. In PGF, recovery is usually more frequent than in other cases of postcardiotomy syndrome, due to the more probable reversibility of the damage, which is why weaning should be the first goal upon initiation of ECMO support.