End-stage heart failure is a leading cause of death of modern society. In the US almost five million people suffer from heart failure, with an annual incidence of 500,000 new cases. About 10% of the population beyond 70 years of age are affected (www.nhlbi.nih.gov). Estimations for Europe amount to 10 million and heart failure has overtaken cancer as the leading cause of death (www.destatis.de). Medical therapy is standardised in the treatment guidelines of the respective national and international societies, and is highly effective in the early stages of the disease. For late stages, including New York Heart Association (NYHA) class IV and Stage D (American Heart Association, AHA), heart transplantation is the treatment of choice as it provides excellent long-term survival and a good quality of life.1 However, not all patients can undergo heart transplantation. The supply of donor organs can never meet the demand such that only four to eight patients per million inhabitants in Europe and the US can have a heart transplantation.2 In other continents and countries the situation is much worse, and heart transplantation is restricted to very few patients or even not performed at all. Furthermore, many traditional Muslims refuse organ donation as body desecration and also refuse to receive life-saving transplants.
The realisation of these facts led to the progressive use of mechanical ventricular assist devices, especially during the last decade. The main intention still remains to provide a bridge to transplantation, as the overall results for transplantation, i.e. survival and quality of life, are superior to any kind of long-tem mechanical support.3 Accordingly, every third to fourth heart transplant procedure is currently performed on a patient with prior implanted ventricular assist device. Despite focusing on transplantation, a lot of experience with longer-term mechanical support has been gained during the past two decades.4,5 Due to the scarcity of donor organs patients had to be supported for periods from many months to more than a year. During that time, many problems and complications become apparent and lead the patient as well as the cardiovascular team to realise what destination therapy means.6 Nevertheless, as many of the therapeutic challenges could be successfully met, finally, the way was carefully paved for chronic mechanical support with ventricular assist devices.7
The general indication for mechanical ventricular support is a low cardiac output defined as cardiac index <2.0 litres/minute/m2. This is especially true for patients in acute cardiogenic shock. In cases of chronic end-stage heart failure, where low cardiac outputs can be tolerated for quite some time, the clinical status is more important than the cardiac output or the central venous oxygen saturation.
Patients in NYHA class III and IV with worsening haemodynamics despite optimal (inotropic and/or vasodilator) medical therapy usually motivate surgeons to device placement.8 However, implantation of an assist device should not rely only on a general indication, particularly not when chronic support is intended; it is important to consider the risks and contraindications too. Database analyses have revealed several risk factors for left ventricular assist device (LVAD) implantation, as described below.
Emergent LVAD implantation in patients with severe shock and progressive multi-organ failure is difficult as the implantable rotary pumps deliver only a pump flow of about 3–4 litres/minute.9 This pump output is frequently not enough in a clinical condition with dominating systemic inflammatory response syndrome and vasoplegia. It has been recognised that outcome with rotary pumps is far better when LVADs are implanted on an elective/urgent basis.
Age has a strong influence on outcome. Several studies have shown that LVAD placement beyond 60 years of age is associated with worse results.10 The main reason for the poor results is co-morbidity, as many VAD patients experience complications that are not typically a consequence of the former surgery, such as renal and liver failure, respiratory insufficiency and cerebrovascular events.11
Right heart failure after LVAD implantation is a deleterious complication that is difficult to predict.12,13 Risk assessment is more often based on experience rather than on reliable scores. It is noteworthy that pulmonary hypertension per se is neither a risk factor nor a contraindication. On the contrary, it may be the leading diagnosis hindering transplantation and promoting LVAD placement.14,15
Although there are several risk factors, absolute contraindications for definitive LVAD therapy are only very few. The main obstacles are co-morbid conditions that do not allow long-term oral anticoagulation. Peripheral and cerebral arterial occlusive disease at a late stage may render a critically low blood supply to the respective area when a non-pulsatile system providing a constant arterial pressure of about 60mmHg is implanted. Thus, cerebral imaging diagnostics are mandatory if the patient has a history of transient ischaemic attack (TIA) or stroke. Finally, compliance, hygiene and a stable social environment are important prerequisites for long-term success. If patients are non-compliant and unreliable, the LVAD patient is jeopardised by infection, thromboembolism and pump failure.
Benefits and Expectations
The basic idea of LVAD implantation is the improvement of the patient’s haemodynamic situation, i.e. to provide them with a higher cardiac output. The native heart is partially unloaded and cardiac diameters and contractility can recover. The so-called reverse remodelling leads to regression of left ventricular hypertrophy and normalisation of β-adrenergic responsiveness and calcium cycling. As a consequence, impaired organ function can recuperate, as can neurohumeral and cytokine levels.16,17 Overall, the patient’s health status usually improves dramatically. Thus, an LVAD not only prolongs survival but also offers the patient a better quality of life despite being tethered to the drive line of the system. With the relief of dyspnoea and oedema, the patient can be reintegrated into normal life and sometimes can go back to work.
The majority of patients initially cope well with the support systems with the conviction of having survived a life-threatening situation. Later on, when life normalises, they increasingly realise the burden of the drive line. Patients stating early after VAD placement that living with a mechanical assist device is quite tolerable are later often desperate for a heart transplantation. Elderly people with limited activities far better accept their destiny.
Ventricular Assist Device Systems
During the past two decades, there has been a tremendous evolution in mechanical long-term support. In the early 1990s, the first generation of implantable LVADs became widely available. The Novacor and the HeartMate I were bulky, heavy and noisy displacement pumps, inserted into the posterior rectal sheath of the abdominal wall. Both systems were associated with unacceptably high complication rates and are largely abandoned nowadays. Similarly, the Lionheart (Arrow Int, Reading, PA), a totally implantable LVAD, failed to provide convincing results.18
With the first implantation of the MicroMed DeBakey system, a continuous flow pump, infectious complication rates significantly decreased and neurological complications decreased. More continuous flow pumps emerged, including the Jarvik 2000, the HeartMate II and the INCOR, of which the latter two gained widespread acceptance.
- The HeartMate II (Thoratec, Pleasanton, CA) is the most wildly used system worldwide designed to accommodate a broad patient population with a body surface area (BSA) starting at 1.2m2. It is a mid-size device (400g, 125cc) with a mechanical bearing that can create flows of 3–10 litres/minute (6,000–15,000rpm). Due to its size, a pocket in the anterior abdominal wall is required. The inflow cannula is made of titanium, while the outflow conduit consists of a flexible Dacron graft. Anticoagulation has been stepwise lowered, and is now standardised with coumadin and aspirin. The HeartMate II has received US Food and Drug Administration (FDA) approval for bridge-to-transplantation and destination therapy.
- The INCOR (BerlinHeart, Berlin, Germany) is the most technically advanced device. It is smaller (200g) and usually fits well into the pericardium without the need for a pump pocket. It can be implanted via median sternotomy or left lateral thoracotomy. It has magnetic bearings and thus has no friction and no wear. Another feature is the silicon cannulas and its snap connectors, which greatly alleviate device implantation and exchange if necessary. The rotational speed of the impeller can be varied between 5,000 and 10,000rpm providing a flow between 3 and 10 litres/minute. Several automatic programmes, such as antisuction control and periodical rpm reduction to clear the rotor blades from debris and allow regular aortic valve opening, are integrated into the software and can be separately activated. Anticoagulation is similar to the HeartMate II LVAD. The device has received the CE mark but is not FDA approved. Hence, it is mostly used in Europe and Asia.
- The Jarvik 2000 Flowmaker (Jarvik Heart Inc, New York) has a different set-up, as the pump chamber is placed into the left ventricular apex; hence, there is no inflow conduit. The outflow conduit consists of a Dacron prosthesis that is connected to the aorta. With a mid-line sternotomy approach the graft is anastomosed to the ascending aorta; in the case of a lateral thoracotomy approach it is attached to the descending aorta. The control unit allows a choice of flow rate between 8,000 and 12,000rpm, but is not automated. Instead, flow adjustments are made manually according to the patient’s perceived needs. There are no automated control algorithms included. As a unique option for lifetime therapy, a retroauricular connector for the drive line has been tested to further increase quality of life. The Jarvik 2000 is not yet fully FDA approved; only investigational use as bridge-to-transplant is allowed in the US – but the CE mark has been granted.
- The MicroMed DeBakey LVAD has undergone several changes and is now called the HeartAssist 5 (MicroMed Cardiovascular, Inc, Houston, TX). It is still the smallest pump with a length of 71mm and a weight of 92g. The inflow cannula is attached to the titanium housing in a fixed angle; the outflow conduit consists of a flexible Dacron tube. Pump flows up to 10l/minute can be generated with 7,500–12,500rpm. A flow probe measures the exact flow on the outflow graft, which is unique in this system. A CE mark has been obtained, but the system is still not FDA approved.
Meanwhile, further devices have emerged. Implantable centrifugal pumps are attributed to be the third LVAD generation. The lower pump speed and the non-mechanical levitation are presumed advantageous. Some LVADs such as the Coraide in 2005 were tested on investigational grounds and were found to be not yet matured. Other LVADs were much more successful, including the Duraheart in 2004, the Ventrassist (Ventracor) and the Heartware device. However, the Ventracor had to close down due to financial problems in 2009.
- The Duraheart (Terumo Heart, Inc, Ann Arbor, MI) is a centrifugal pump with a magnetically levitated impeller. The inflow conduit consists of a small titantium tube that is connected to the housing. The conduits are available in three sizes for optimal pump positioning. In any case, a pump pocket in the anterior chest wall has to be created to accommodate the relatively large pump chamber (540g). With a motor speed of 1,200–2,400rpm a flow of 2–8l/minute can be created. FDA approval has not been granted yet, but the CE mark has been obtained.
- The Heartware HVAD (Heartware Int, Inc, Framingham, MA) is a new system that convinced the cardiosurgical community due to its small size (145g). The pump has a displacement volume of only 50cc and is mounted with its integrated inflow cannula on the left ventricular apex, while the outflow graft prosthesis is connected to the patient’s ascending aorta. The device is capable of generating up to 10l/minute of blood flow with the impeller, which spins at rates between 2,000 and 3,000rpm. The impeller is suspended within the pump housing through a combination of passive magnets and a hydrodynamic thrust bearing. The system received a CE mark and FDA approval for bridge-to-transplant in 2010.
Paracorporeal pulsatile VADs have been used for chronic support only sporadically as they do not offer an acceptable quality of life.19 Likewise, total artificial hearts (TAH) did not fulfill expectations. The electrically driven Abiocor TAH (Abiomed, Inc, Danvers, MA) has been under trial as a destination therapy, but its use has been stopped. The Cardiowest TAH (SynCardia Systems, Inc, Tucson, AZ) is mainly used as bridge-to-transplant, and only in very few cases for destination therapy.
Results and Complications
Chronic mechanical circulatory support has been the ultimate goal of cardiac surgeons and patients for many years. Since the first implantations of mechanical support systems in the 1960s various devices have been tested and prolonged survival as well as a reduction in device-related complications has been achieved. However, the goal of replacing the heart with an implantable machine has yet to be reached.
The first remarkable results for chronic mechanical support as destination therapy were obtained with the HeartMate XVE LVAS (Thoratec, Pleasanton, CA) in the Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) trial. In this trial 129 patients with end-stage heart failure, who were ineligible for transplantation, were randomised to either LVAD therapy (n=68) or optimal medical management (n=61). The LVAD patients demonstrated 52.1% one-year survival compared with a 24.7% survival rate for patients under optimal medical treatment. At two years, the likelihood of survival dropped to 22.9% for the LVAD patients compared with 8.1% for the patients under medical therapy. Despite the obvious advantage for LVAD patients, two-year survival remained relatively poor. Moreover, prolonged life was obtained at the costs of tremendous complications. The probability of device infection was 28% at three months and of bleeding was 42% at six months, and the incidence of device failure was 35% at two years.4 Accordingly, the hype for destination therapy dramatically declined and it was noted that there is a learning curve, and that patient selection matters most. The cumulative US experience in the post-REMATCH era of 280 patients demonstrated similar survival to the REMATCH patients (56 and 27% at one and two years, respectively), but there were also large single-centre experiences with much more favourable results (77% at two years).20,21 Importantly, even if the adverse events could be reduced, the probability of device exchange or fatal failure was 72.9% at two years due to the poor durability of the system.
A better outcome was anticipated with the Novacor LVAD (World-Heart, Oakland, California, US) because of a different pump design. The Investigation of Nontransplant-Eligible Patients Who Are Inotrope Dependent (INTrEPID) trial, which was a prospective, non-randomised study comparing Novacor LVAD therapy (n=37) with medical treatment (n=18) revealed superior survival at six and 12 months (46 versus 22%, and 27 versus 11%, respectively) and no relevant device failure but a much higher incidence of cerebral thromboembolic complications (62 versus 11%).22
New hopes for long-term use of ventricular assist devices as an alternative to heart transplantation grew when axial flow pumps came into play. They were smaller and thus could be used in smaller patients. Technically distinct, there was less wear and due to continous flow design no volume chamber was needed, resulting in reduced pump size. In addition a considerably thinner drive line and simplified implantation enhanced usability. As a consequence, bleeding and infectious complications dramatically decreased. However, thromboembolic problems remained.
The HeartMate II trial with 200 patients is now the largest study in the US to investigate destination therapy with LVADs. The non-pulsatile axial flow pump HeartMate II demonstrated a superior two-year survival rate of 58% compared with 23% with the first-generation HeartMateXVE, as well as a significantly improved probability of freedom from stroke and device failure. Quality of life and functional capacity were significantly improved with both systems.23
In Europe, a multicentre study investigated the length of the inflow cannula in the INCOR system, in bridge-to-transplant and destination therapy patients. Patients with a long inflow cannula demonstrated a better survival rate than those with a short inflow cannula, as well as a significantly lower thromboembolic adverse event rate of only 3.8%.24
Overall, the leading causes of death according to the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) in patients treated with LVADs as destination therapy are cardiac failure (23%), cerebrovascular event (23%) and infection (15%). Compared with the whole cohort of LVAD patients, destination therapy seems to be especially prone to cerebrovasular complications, while multi-organ failure is less likely to occur. These facts reflect the selection of elderly patients to electively undergo LVAD placement. Predominant risk factors outlined by the INTERMACS registry are advanced age and cardiogenic shock. Further problems including surgical bleeding and device malfunction have declined over the years and now occur in less than 10% of patients.7 A new challenge will become non-cardiac surgery in patients treated with chronic mechanical support.25
Unsolved Problems and Outlook
Even if a support interval of more than seven years has already been achieved, there is still a long way to go until chronic mechanical support becomes daily routine. Apart from financial issues, patient co-morbidity is the greatest obstacle hindering liberal use like in pacemaker and implantable cardioverter–defibrillator (ICD) therapy. Commonly in patients with end-stage heart failure, significant end-organ dysfunction is present, such as renal and hepatic dysfunction and chronic obstructive lung disease, which is not always reversible after device placement. Post-operative renal failure requiring haemodialysis and prolonged ventilation via tracheostomy are not infrequent complications and cannot be adequately predicted or prevented. Extended stays in the intensive care unit with its typical complications are more rule than exception26 (see Figure 1). One of the most significant predictors of improved survival with acceptable complication rates is pre-operative patient condition, defined by the INTERMACS level. It has been demonstrated that patients in INTERMACS level 3–4 prior to LVAD implantation have significant superior survival in the post-operative course.1 Therefore, elective VAD implantation should be performed in patients with chronic heart failure before irrevocable deterioration of organ function has occurred. As the likelihood for jeopardising co-morbidity clearly increases with age, the vast majority of old patients are not eligible for device placement yet. Driveline infection, non-cardiac surgery and disabling neurologic complications are challenges still to be met.