In 1969, Denton Cooley implanted a pneumatically-driven right and left ventricle prosthesis in a 47-year-old man with a left ventricular aneurysm who could not be weaned from cardiopulmonary bypass. The prosthesis kept the patient alive for 64 hours, at which time a donor heart was transplanted. The transplanted heart functioned for 32 hours, but the patient died of overwhelming bacterial pneumonia. This historic implantation was the first of its kind in humans, and it represented the culmination of years of intensive research at many centres in the effort to assist the failing heart.

A number of innovative advances prior to 1969 provide the basis for present day mechanical circulatory assistance. In 1934, DeBakey designed the roller pump to support the circulation with continuous blood flow. Utilizing the roller pump, Gibbon performed the first successful clinical total cardiopulmonary bypass to correct an intracardiac lesion. In 1958 Akutsu and Kolff performed the first orthotopic replacement of a dog heart with a totally artificial heart made of polyvinylchloride and powered by a compressed air source placed externally. In 1961, Dennis demonstrated a complete left heart bypass by inserting an inflow cannula into the left atrium through the atrial septum and returning blood through the femoral artery. In 1967, the intra-aortic balloon counterpulsation device was used by Kantrowitz to assist the failing left ventricle. Collectively, these innovations have served to provide a diverse armamentarium to assist the growing number of patients who require circulatory support.

Modern circulatory support strategies now include a continuum of modalities to support the failing heart. Initial efforts to resuscitate a patient with myocardial dysfunction rely on optimization of the volume status and infusion of vasoactive agents to improve myocardial function and to reduce ventricular afterload. Balloon counterpulsation can next be employed alone or in combination with these agents to augment myocardial function. When more extensive short-term assistance is required, the centrifugal pump can be used to assist both the right and left ventricle. Patients who require assistance for longer periods of time, such as transplant candidates or those with biventricular failure after bypass, can benefit from support by pulsatile left ventricular assist devices or the total artificial heart. Each of these strategies, while offering circulatory support for variable periods of time, allows the myocardium to recover from acute injury or ultimately requires biological replacement either by skeletal muscle flap augmentation, cardiac allografting, or xenografting.

In a retrospective review of 14 168 patients undergoing cardiopulmonary bypass between 1975 and 1979, 326 patients (2.3 per cent) demonstrated inadequate response to conventional pharmacological therapy and required intra-aortic balloon support during weaning. An additional 94 patients demonstrated an inadequate response to balloon support and 21 patients (0.1 per cent) eventually underwent profound mechanical circulatory support with a left ventricular assist device. While the need for profound circulatory support was relatively rare, a substantial number of critically ill patients who require operations involving cardiopulmonary bypass need subsequent support.

As the technical advances in mechanical circulatory assistance grow, the potential application of these assist devices increases. In the United States, approximately 200 000 patients die from ischaemic heart disease each year and another two million people are afflicted with congestive heart failure. In Europe, end-stage ischaemic disease kills nearly two million people annually. In Central and South America, Chagas' disease afflicts two million people with an extensive myocarditis leading to congestive heart failure. While cardiac transplantation is an effective therapy for many patients with heart failure, only 15 000 are suitable candidates for heart transplantation due to current selection criteria and age limitations, and the maximum supply of suitable donor hearts has been estimated to be 2000 per year in the United States.

Cardiac failure amenable to circulatory assistance can be grouped into three broad pathophysiological categories: acute myocardial ischaemic and/or infarction, myocardial failure following open heart surgery, and primary myocardial diseases leading to chronic congestive failure. Acute myocardial infarction occurs when perfusion is reduced below the critical needs of the myocardium. Within 24 h of cellular necrosis, there is an inflammatory response which persists over several days. The infarcted area is remodelled by phagacytosis of necrotic material and deposition of collagen after 1 to 2 weeks. A dense collagen scar forms by 4 to 6 weeks. By comparison, transient ischaemia and subsequent reperfusion leads to myocardial dysfunction; however, recovery does occur. This has been referred to as the post-ischaemic ventricular dysfunction or ‘stunned myocardium’, and it resolves within several days.

Myocardial dysfunction after cardiopulmonary bypass has been reduced significantly by shorter operative times on bypass and improved myocardial protection. There are, however, a small number of patients who require assistance beyond conventional inotropic therapy and volume replacement. Pathologically, these patients may sustain myocardial ischaemia with cellular necrosis, similar to that seen in acute myocardial infarction, but the changes develop more slowly. Recovery occurs with progressive reduction of myocardial oedema and replenishment of the high energy phosphate content of myocardial cells.

Primary myocardial diseases or ‘cardiomyopathies’ cause pathological changes within the myocardium itself. Idiopathic dilated cardiomyopathies are associated with congestive heart failure due to non-specific aetiologies. Patients with ischaemic cardiomyopathies develop cardiac failure secondary to coronary atherosclerosis and resultant myocardial damage. Myocarditis is an inflammatory process within the heart which may be associated with viral or protozoal infections, or toxic drugs. Myocardial inflammation leads to destruction of myocardial cells and replacement of contractile elements with fibrous tissue. This can remain asymptomatic for years; chronic congestive failure may develop after years of inflammation.

Each of these pathophysiological conditions can lead to congestive heart failure with an extremely poor prognosis. More than 50 per cent of these patients die unless attempts are made to augment or improve myocardial function with some form of circulatory assistance. These potential therapeutic options include vasoactive agents, balloon counterpulsation, ventricular assist devices, or the total artificial heart.

The adequacy of myocardial function is dependent upon heart rate, ventricular preload, ventricular afterload, and myocardial contractility. When myocardial failure results in a low cardiac output state, preliminary manipulations are aimed at these key determinants. Optimization of ventricular filling pressure by the administration of volume increases both the stroke volume and ventricular contractility according to the Frank–Starling principles. Maintenance of normal sinus rhythm permits the appropriate increases in heart rate to maximize cardiac output with co-ordinated atrial contraction. When the appropriate manipulations of ventricular preload and heart rate are not adequate to improve cardiac output, attention is focused on the inherent contractile state of the myocardium and the resistance to ventricular emptying.

Traditionally, cardiac glycosides have been used to augment the contractile state of the myocardium. Digitalis functions at the molecular level to inhibit the sodium transport enzyme (Na⫀ATPase), slowing the heart rate and indirectly leading to a positive inotropic effect by changing the intracellular calcium concentration. It is effective in short-term treatment of dilated cardiac failure secondary to ischaemia, mitral valve disease, and hypertension. Unfortunately, it is of little value in cardiogenic shock or acute myocardial infarction because it may increase myocardial oxygen demand, increase vascular resistance, and increase the risk of arrhythmias. It is usually unable to increase the inotropic state in acute postoperative myocardial dysfunction.

Sympathomimetic amines and their derivatives are commonly employed to improve myocardial contractile function: their effect is dependent upon the interaction between the drug and its receptor site. The naturally occurring catecholamines noradrenaline, adrenaline, and dopamine act at five receptor sites. &bgr;&sub1;-receptors are primarily located in the myocardium whereas &bgr;&sub2;-receptors are found in the lungs and peripheral vascular bed. &agr;&sub1;-Postsynaptic receptors cause peripheral arterial and venous constriction, and &agr;&sub2;-presynaptic receptor stimulation decreases neurotransmitter release (Table 1) 480. Vasodilatation is stimulated by dopamine receptor activity in the renal, mesenteric, cerebral, and coronary circulations. On a cellular level, &bgr;-receptor–drug interaction leads to activation of the enzyme adenylate cyclase on the cell membrane which increases the synthesis of cAMP in the effector cell. Drug interaction with an &agr;-receptor increases the transmembrane influx of ionized calcium. The efficacy of catecholamines in circulatory support is dependent upon the physiological state of the patient, the number of receptors available to interact with the drugs, and their binding affinity. Hyperthyroidism increases the number and affinity of adrenoreceptors in the heart, while hypothyroidism leads to the opposite effect. Acidosis impairs myocardial function and therefore decreases the efficacy of catecholamine receptor stimulation.

Catecholamines are primarily used to increase the inotropic state of the myocardium in the treatment of cardiogenic shock following myocardial infarction, myocardial failure following cardiac surgery, and in circulatory failure complicating sepsis. Increasing the inotropic state of the myocardium is not without cost. Stimulation of &bgr;&sub1;-receptors also has a positive inotropic effect which increases the heart rate and myocardial oxygen consumption. This increase in heart rate also decreases coronary perfusion by decreasing diastolic filling time. When coronary blood flow is limited, as in atherosclerosis, this can lead to myocardial ischaemia and further depress myocardial function.

Specific agents used most commonly in cardiothoracic surgical practice include noradrenaline, adrenaline, isoproterenol, dopamine, dobutamine, and phosphodiesterase inhibitors. The vasodilators nitroprusside, nitroglycerine, and prostaglandin E&sub1; will also be discussed (Table 2) 481.

Noradrenaline is a potent catecholamine, interacting predominantly with &agr;-receptors and causing vasoconstriction, although it also has &bgr;&sub1;-receptor activity. Because of this &agr;-mediated arterial and venoconstriction, the arterial blood pressure is markedly elevated. This increased afterload may decrease cardiac output by increasing resistance to left ventricular ejection. With prolonged infusion, renal, hepatic, and muscle perfusion are reduced, and this may lead to acidosis. Noradrenaline may be useful to improve coronary perfusion by raising the central aortic pressure in patients with decreased systemic vascular resistance and preserved cardiac output. This can be seen in septic patients and those with severe neurological disturbance. Occasionally, noradrenaline has been used in combination with phentolamine (a potent vasodilator) to counteract peripheral vasoconstriction in patients who are unable to be weaned from cardiopulmonary bypass. The usual dosing range is 0.01 to 0.03 &mgr;g/kg.min.

Noradrenaline has also been administered to patients who develop acute pulmonary hypertension, right ventricular failure, and resultant systemic hypotension. This situation has most commonly been reported in patients with pulmonary hypertension after mitral valve replacement, massive pulmonary embolism, or in cardiac transplant recipients. The initial problem in these patients is pulmonary hypertension which leads to right ventricular failure; systemic hypotension develops as left atrial filling pressures fall. Fonger has supported the selective infusion of noradrenaline through the left atrium or aorta to provide systemic vasoconstriction in combination with a pulmonary vasodilator, such as nitroglycerine or prostaglandin E&sub1;, administered in the venous circulation to relieve the pulmonary hypertension. We have occasionally used selective infusion of noradrenaline through the central lumen of the intraaortic balloon pump, in combination with central venous infusion of vasodilators, for right ventricular failure with associated pulmonary hypertension.

Adrenaline is another naturally occurring catecholamine which has both &agr;- and &bgr;-receptor activity. At lower doses, &bgr;&sub1;-receptor effects predominate, leading to increased myocardial contractility, increased heart rate, and increased coronary blood flow. It causes less vasoconstriction and hypertension at lower doses. At higher doses, &agr;-receptor mediated vasoconstriction predominates, leading to decreased renal blood flow and glomerular filtration. Adverse effects include tachycardia and arrhythmias which compromise coronary perfusion and oxygen supply–demand relationships. Adrenaline is often the primary agent used in combination with vasodilators, in patients who need inotropic support during weaning from cardiopulmonary bypass. The usual dose range is 0.01 to 0.1 &mgr;g/kg.min.

Isoproterenol is a synthetic catecholamine with singular &bgr;-receptor activity. It causes strong inotropic and chronotropic effects as well as increased atrioventricular conduction and automaticity. The &bgr;&sub2;-receptor activity leads to vasodilatation in the systemic and pulmonary vascular beds, with a resultant marked increase in right and left ventricular cardiac output. In spite of the increased cardiac output, reflex tachycardia may occur, as may ventricular arrhythmias at higher doses. It is useful in the emergency treatment of bradycardia. Patients with chronic pulmonary hypertension and acute right ventricular failure may also benefit from its use, provided that they do not have coexistent coronary artery disease. Isoproterenol has also been used to improve myocardial contractility and to increase heart rate after cardiac transplantation, exerting a direct effect on &bgr;-adrenergic receptors and decreasing afterload in both the systemic and pulmonary circulation. The usual dose is 0.01–0.1 &mgr;g/kg.min. Isoproterenol may not be appropriate for patients with ischaemic myocardium or uncorrected coronary artery stenosis because it increases myocardial oxygen consumption and decreases coronary perfusion pressures.

Dopamine is an endogenous catecholamine which stimulates myocardial contractility by directly acting on &bgr;-adrenergic receptors. Dopaminergic receptor stimulation in coronary, renal, and mesenteric vascular beds also causes vasodilatation. These effects are mediated via the stimulation of adenylate cyclase which leads to an elevation of intracellular cAMP. Dopaminergic effects are dose-dependent. At low doses (2–5 &mgr;g/kg.min), &bgr;&sub1; stimulation predominates with enhanced myocardial contractility, increased cardiac output, and mesenteric vasodilatation. At higher doses (5–10 &mgr;g/kg.min), there is a greater effect on blood pressure as peripheral vasoconstriction predominates. As with other catecholamines, increases in heart rate, myocardial contractility, and arterial pressure all lead to increased myocardial oxygen consumption. Moderate doses of dopamine are commonly used in refractory congestive failure to increase cardiac output and produce sodium diuresis through its effect on the renal vascular system. In cardiogenic shock, when higher doses are necessary for more pronounced inotropic and vasoconstrictor effects, it may be combined with nitroprusside or nitroglycerine to counteract vasoconstriction. Dopamine is also used for treatment of postcardiopulmonary bypass myocardial dysfunction, alone or in combination with intra-aortic balloon counterpulsation.

Dobutamine is a synthetic catecholamine which stimulates &bgr;&sub1;-&bgr;&sub2;, and &agr;-receptor sites. &bgr;&sub1;-Receptor activity predominates, leading to augmentation of myocardial contractility, with less tachycardia than noradrenaline or dopamine: Leier reported that dobutamine caused less tachycardia and fewer premature ventricular contractions compared to dopamine in patients treated for congestive heart failure. Over a 24-h period, dopamine and dobutamine produced similar dose-related haemodynamic improvements in cardiac function, but dobutamine was found to be less arrhythmogenic. Its &agr;-receptor effects are much weaker than those of noradrenaline, resulting in improved cardiac output, increased arterial pressure, reduced left ventricular end diastolic pressures, and enhanced coronary perfusion. Like other catecholamines, dobutamine can increase myocardial oxygen demand and arrhythmias. It can be used effectively in patients with myocardial dysfunction after cardiopulmonary bypass in patients with acute myocardial infarction to augment cardiac function. In chronic congestive failure, it is effective in stimulating myocardial contractility, although this effect is limited because of receptor ‘down regulation’ over time. The usual dose of dobutamine is 2.5 to 10 &mgr;g/kg.min.

The phosphodiesterase inhibitors amrinone, milrinone, and enoximine have been developed recently as inotropic agents effective in the treatment of acute and chronic cardiac failure. At the cellular level, they inhibit the membrane-bound phosphodiesterase enzyme which is responsible for the breakdown of AMP, therefore increasing its intracellular concentration. This action enhances myocardial contractility, by facilitating calcium availability at the level of the sarcoplasmic reticulum. Amrinone, the only agent currently approved for clinical use, also enhances myocardial relaxation. It has a direct dilating effect on peripheral arteries and veins, which leads to increased cardiac output and decreased right and left sided pressures. This combination of positive inotropic effect and vasodilatation prevents any net increase in myocardial oxygen consumption. Amrinone is often useful in patients with reversible myocardial depression after cardiopulmonary bypass or acute myocardial infarction. Patients who are not adequately volume loaded will develop hypotension following amrinone infusion. While its effect is very similar to dobutamine, patients do not develop tolerance over time. Intravenous administration of amrinone requires an initial bolus of 0.75 mg/kg followed by continuous infusion at a dose of 5 to 10 &mgr;g/kg.min.

In patients with myocardial failure and low output states, systemic vascular resistance is often elevated to maintain arterial pressure. As the left ventricle becomes dilated and left ventricular volume increases, wall stress within the ventricle becomes elevated. These phenomena lead to increased resistance to ventricular ejection, exacerbating the depression of myocardial function. These basic physiological observations suggest the rationale for vasodilator therapy to reduce ventricular afterload. Elevations in venous tone also play a part in the patient with congestive failure, leading to intrathoracic venous overload and pulmonary congestion. A number of arterial and venous dilating agents have been employed in patients with myocardial failure either in the acute or chronic setting (Table 3) 482. The most commonly used agents in the surgical setting include sodium nitroprusside and nitroglycerine.

Sodium nitroprusside is one of the most commonly used vasodilators in acute congestive heart failure. It has a direct relaxing effect on peripheral vascular smooth muscle, thereby decreasing arterial pressure and venous return, and increasing cardiac output by decreasing afterload. The initial infusion rate is 1 to 10 &mgr;g/kg.min, and it is titrated according to systemic pressure effect. It has a rapid onset of action and short half-life. Combination therapy using nitroprusside with dopamine has been reported to be effective in congestive heart failure. It is also useful in patients who are hypertensive during the immediate postoperative period to prevent excessive bleeding. Degradation of nitroprusside can liberate cyanide radicals with toxic side-effects. When infusions continue for longer than a few days, serum thiocyanate levels should be measured and not allowed to exceed 6 mg/100 ml.

Nitroglycerine is a vasodilator that primarily affects vascular smooth muscle tone in venous capacitance vessels. At doses 1 &mgr;g/kg.min, venodilatation predominates; however, at higher doses some arterial vasodilatation occurs. Nitroglycerine also exerts a beneficial effect on oxygen supply and demand by decreasing preload, wall tension, and myocardial oxygen consumption. It may also dilate coronary arteries, thereby increasing blood flow to ischaemic myocardium. Usual intravenous infusion rates are 1 to 5 &mgr;g/kg.min.

Right ventricular failure has become more frequently recognized in patients who suffer from left ventricular failure, as well as in those with primary pulmonary parenchymal disease and pulmonary hypertension. It may also complicate mitral valve replacement, coronary artery bypass grafting, and heart transplantation. The common pathophysiological variable is markedly increased pulmonary vascular resistance. This elevation in right ventricular afterload in combination with hypoxia and acidosis (which are common in chronic pulmonary disease) lead to impaired right ventricular contractility. Prostaglandin E&sub1;, a potent pulmonary vasodilator, relaxes pulmonary arteriolar smooth muscle. It also has systemic vasodilator properties at higher doses; however, there is a significant first pass metabolism in the lungs, which helps to minimize its systemic effect. When systemic hypotension persists secondary to vasodilation, left atrial infusion of noradrenaline may be used to provide pressure support. As mentioned previously, agents acting on &agr;-receptors may also be given directly into the thoracic aorta via the central lumen of the intra-aortic balloon pump, if this device has been inserted. As with all left sided infusions, extreme caution must be used to avoid the introduction of air. The usual infusion dose is 30 to 150 ng/kg.min. It has also been used in patients with right ventricular failure post bypass and heart transplant recipients with pulmonary hypertension.

The intra-aortic balloon counterpulsation device was first used successfully by Kantrowitz in 1967. Since that time, it has become the most commonly used mechanical assist device for patients with acute left ventricular failure in postmyocardial infarction syndromes, in cardiogenic shock, in patients with refractory acute mitral insufficiency, and in patients unable to be weaned from cardiopulmonary bypass (Fig. 1) 1646,1647. Development of a technique to enable the balloon to be placed percutaneously has enhanced its use in clinical practice.

The balloon is an inflatable plastic device that is positioned in the descending thoracic aorta just distal to the left subclavian artery. This can be accomplished either by femoral arterial cut-down or a percutaneous approach using fluoroscopy to assist in positioning. The balloon is connected to a console which synchronizes inflation and deflation with the electrocardiogram or the arterial pressure tracing. Optimal timing provides for the balloon to be inflated during diastole and deflated in systole. Inflation increases the diastolic pressure, therefore augmenting coronary perfusion pressure and enhancing coronary blood flow. Deflation just before systole provides systolic unloading by decreasing the relative afterload to left ventricular ejection.

These actions improve myocardial oxygen supply by increasing coronary perfusion and decrease myocardial oxygen demand by decreasing the left ventricular workload. Intra-aortic balloon counterpulsation has been used in patients with acute myocardial infarction complicated by persistent angina pectoris refractory to conventional therapy, patients in cardiogenic shock, patients who fail to wean from cardiopulmonary bypass, and those who develop ventricular septal rupture and acute mitral regurgitation after myocardial infarction. Gold reported a significant reduction in left ventricular ischaemia in patients suffering cardiogenic shock secondary to myocardial infarction using intra-aortic balloon counterpulsation. While the majority of patients in cardiogenic shock responded to balloon counterpulsation initially, they required coronary revascularization to bypass severely obstructed vessels leading to ischaemia.

The complications of intra-aortic balloon counterpulsation have steadily decreased since their introduction. There is, however, a significant incidence of major complications that occur most commonly with percutaneous insertion. Kantrowitz reported a complication rate of 8.8 per cent, the majority of which were related to vascular trauma, thromboembolism, or severe reduction in femoral arterial flow. Two-thirds of these complications were resolved either by removal of the balloon or, occasionally, with thrombectomy. Patients with a history of diabetes mellitus, peripheral vascular disease, and hypertension, and female patients have a higher incidence of vascular complications. Relative contraindications to intra-aortic balloon counterpulsation include severe aortic valve insufficiency, aortic dissection, and previous prosthetic grafts in the thoracic aorta.

Cardiac assist devices have traditionally been used to support the left ventricle. However, right ventricular failure is an important, though less frequent, indication for cardiac assist devices. In particular, right ventricular failure has been diagnosed in patients being weaned from cardiopulmonary bypass, those with acute inferior wall myocardial infarction, and those with refractory pulmonary hypertension. In the rare instance when standard medical manoeuvres such as volume loading, hyperventilation, and infusion of vasoactive drugs such as nitroglycerine, isoproterenol, and prostaglandin E&sub1;, are unable to overcome right ventricular failure balloon counterpulsation has obvious theoretical potential. Early clinical experience uses a standard intra-aortic balloon positioned in a graft sewn to the pulmonary artery as a reservoir for blood and the balloon (Fig. 2) 1648. Sporadic case reports in this setting have demonstrated definite haemodynamic improvements in patients with biventricular failure following bypass. However, the technical complexities of such a procedure have prevented this technique from enjoying widespread clinical application. A pulmonary artery balloon that can be positioned in the artery via a percutaneous approach is under development and it holds significant promise for clinical investigation in the future. This percutaneous pulmonary artery balloon has produced significant systolic unloading of the right ventricle in experimental animals.

Higher levels of circulatory assistance are necessary in a small number of patients who do not respond to volume loading, infusion of inotropic agents, and balloon counterpulsation. Liotta reported the first attempt to support a patient with prolonged partial left ventricular bypass by means of an intrathoracic pump. The patient was a 42-year-old male with severe congestive heart failure who sustained a cardiac arrest after aortic valve replacement. He was supported for 4 days. In 1971, DeBakey developed the prototype pneumatic ventricular assist device and reported long-term patient survival. In the 1970s and 1980s, the centrifugal pump was used to support left ventricular function in a number of patients, resulting in good functional recovery and survival. Cooley and Norman later reported the first successful left ventricular assist device used as a ‘bridge’ to cardiac transplantation. This report was followed by several accounts of long-term survival after transplantation using a preoperative or ‘bridging’ ventricular assist device.

In general, ventricular assist devices function in parallel with the native ventricle with the potential to support the systemic and/or pulmonary circulation. They unload the failing ventricle, decrease myocardial oxygen demand and consumption and allow the ‘stunned myocardium’ to recover. Optimal recovery can be obtained by maintaining normal sinus rhythm and normal atrial and ventricular pressures to reduce ventricular wall stress. Clinical experience suggests that myocardial function often improves within 3 to 7 days.

The left atrium, left ventricle, or pulmonary veins may be cannulated to provide adequate inflow for the assist device. The ascending or descending aorta provides the outflow conduit for the device. Decompression of the left side of the heart directly reduces left atrial and left ventricular pressure and subsequently reduces left ventricular work and myocardial oxygen demand. Left ventricular bypass does, however, increase right ventricular work and this can lead to right ventricular failure and subsequent reductions in left-sided filling pressures if the right ventricle is not supported.

The indications for left ventricular circulatory support are postoperative cardiac failure in patients unable to wean from cardiopulmonary bypass in spite of pharmacological and/or intra-aortic balloon counterpulsation support; cardiac failure due to cardiomyopathy; cardiogenic shock complicating acute myocardial infarction; acute viral myocarditis; failed donor heart after transplantation; and as a bridge to transplantation. Right ventricular assist devices are indicated in cases of isolated right ventricular failure secondary to acute right ventricular infarction, or in association with inferior wall myocardial infarction. A third indication is biventricular failure after cardiopulmonary bypass. Haemodynamic indications include patients with left atrial pressures above 25 mmHg, arterial systolic pressure below 90 mmHg, and a cardiac index below 1.8 l/min.m². The only absolute contraindication to implantation of a ventricular assist device is documented preoperative organ dysfunction such as renal failure, or infections such as pneumonia or bacterial endocarditis. Advanced age (>70 years) is a relative contraindication, since recovery is less likely and biological replacement is usually not considered.

Ventricular assist devices have developed rapidly over the past 25 years. Four basic forms of assist device warrant description—the roller pump, the centrifugal pump, the pneumatic sac-type pump, and the electromechanical pump.

Since it was developed by DeBakey in 1934, the roller pump has been the cornerstone of cardiopulmonary bypass and early ventricular assist strategies. Systemic anticoagulation is required. Blood inflow is usually provided by a left atrial cannula and outflow runs to the ascending aorta via a trans-sternotomy cannula. Constant observation is necessary to maintain adequate left atrial volume and to prevent air aspiration around the cannulation site. The roller pump system has been used as a left and right heart assist device in postcardiotomy bypass failure and postmyocardial infarction, because of its relative effectiveness, inexpense, and familiar technology. In one report 41 per cent of patients who had a roller pump left heart assist device inserted for profound heart failure were weaned from the device and 29.2 per cent were discharged from hospital. Fatal complications of this system included severe coagulopathy, irreversible extensive myocardial infarction, refractory arrhythmias, severe ‘shock’ lung, and multisystem organ failure. Percutaneous cardiopulmonary bypass systems have been developed to provide successful short-term circulatory support in acute myocardial infarction and cardiac arrest patients; however, further experience is necessary to assess their efficacy and potential to allow long-term survival.

Absence of pulsatile flow in these systems is an important consideration. A number of investigators have studied this question in experimental and clinical models. The majority of investigators support pulsatile blood flow as a more physiological and effective method during cardiopulmonary bypass since it provides more efficient metabolism, better renal and cerebral protection, and more rapid warming and cooling.

Centrifugal pumps spin blood as a solid body vortex inside a rigid chamber (Fig. 3) 1649. The kinetic energy developed by this spinning is transmitted into pressure as the blood leaves the chamber via the exit cannula. Inflow fluid enters the chamber through the central inlet, and systemic anticoagulation is recommended, since clot can form in the chamber. Cannulation sites are similar to those used for cardiopulmonary bypass. The pumps produce non-pulsatile blood flow, but they are often used in concert with intra-aortic balloon counterpulsation devices. These pumps have been used primarily for postcardiotomy myocardial failure. Approximately 50 per cent of patients have been successfully weaned from the device, and 25 per cent have survived to be discharged from the hospital. Postoperative bleeding, haemolysis, and thromboembolic events limit long-term use of these pumps. Centrifugal pump circulatory support systems have been used successfully as a bridge to heart transplantation. In one series, seven of nine patients underwent successful transplantation after an average of 1.6 days of circulatory support (range 0.5 to 3 days).

The pneumatic pulsatile ventricular assist devices are much more complex and rely on an external console to pump compressed CO&sub2; from the power unit to displace a flexible diaphragm which empties the blood sac. The Thoratec/Pierce-Donachy Ventricular Assist Device ( Figs. 4 1650,1651,5) developed by Pierce and Donachy at Pennsylvania State University, is the most commonly used pump of this type. The device is composed of a flexible, smooth, segmented polyurethane blood sac and diaphragm which lie within a rigid case. The stroke volume of the device is 65 ml and the maximum blood flow it can generate is 6.5 l/min; it uses Bjork–Shirley tilting disc inlet and outlet valves to promote undirectional flow. The left atrial appendage or left ventricular apex is cannulated for blood inflow, and the outlet cannula is anastomosed to the ascending aorta. The cannulae exit the skin below the left costal margin to enter the device positioned on the anterior abdominal wall. The pump ejection is initiated when the blood sac is full and sensed by a switch located between the diaphragm and outer casing. Patients are anticoagulated with low dose heparin in the early phase followed by full anticoagulation when postoperative bleeding stops.

The Thoratec ventricular assist device has been implanted in patients throughout Europe, Asia, and North America during the past 8 years. Results vary with the indication for implantation: patients who required assistance for profound postcardiotomy myocardial failure during the early experience had little improvement in cardiac function and a poor prognosis. Subsequent reports have shown better overall cardiac recovery and patient survival and this may be due to more experience and better patient selection. Pennington reported that 16 of 30 patients (53 per cent) with profound cardiogenic shock refractory to conventional therapy who were supported by the Pierce-Donachy ventricular assist device had improved cardiac function. Fifty per cent were weaned from the device and eventually 11 patients (37 per cent) were discharged. Survival was adversely affected by the extent of myocardial infarction and renal failure. Common complications were bleeding (73 per cent) and biventricular failure (83 per cent).

The Thoratec ventricular assist device has also been used as a ‘bridge’ to transplantation. In a recent review, Farrar reported on 72 heart transplant candidates who were supported by the devices. Fifty-eight of the 72 patients (81 per cent) received biventricular devices and 14 received left ventricular support only. Fifty-four patients (75 per cent) recovered sufficiently to undergo heart transplantation after a median of 4.4 days (range 8 h to 81 days). Forty-five patients were discharged from the hospital (63 per cent survival from implant to discharge). Of note, 22 cases initially required only left ventricular support; however, eight patients later required right ventricular assist device insertion. The most common complication was bleeding, which required reoperation in 16 patients. Other complications included infection (23 per cent) and renal failure (12 per cent). There were also six patients who sustained neurological events of probable embolic origin in spite of systemic anticoagulation.

In 1984, Oyer and associates performed the first successful ‘bridge’ to transplantation using the Novacor implantable electrical left ventricular assist device in a patient with ischaemic end-stage heart disease (Fig. 6) 1652. This electromechanical device has been designed as a totally implantable left ventricular assist system for long-term support.

The Novacor left ventricular assist device consists of an implantable pump drive unit which is positioned in the anterior abdominal wall of the left upper quadrant. It is connected to the control console by power and control leads contained within a percutaneous vent tube. Dacron inflow and outflow conduits traverse the diaphragm to connect the pump to the left ventricular apex and the ascending aorta. The blood pump consists of a seamless polyurethane sac bonded to dual, symmetrically opposed pusher plates in a lightweight housing that incorporates valve fittings. The pump has a smooth blood–interface surface and pericardial tissue valves with custom silicone flanges. The dual pusher plate blood pump is integrally coupled to a spring-decoupled pulsed-solenoid energy converter. The control console provides electrical power and automatic control to the unit with continuous physiological monitoring of the patient and the device. The maximum stroke volume is 70 ml and the pump can generate flows up to 10 l/min. Patients receive low-dose heparin anticoagulation and/or platelet-modifying drugs.

Since 1984, this device has continued to be implanted as a bridge to cardiac transplantation. Portner reported a multicentre experience of 20 patients who required profound circulatory support. Three patients required implantation of a centrifugal right ventricular assist device in addition to the Novacor device. Orthotopic cardiac transplantation was performed on 10 patients (50 per cent) after implants ranging from two to 90 days (mean 30.3 days). Eight patients (40 per cent) who received transplants were discharged from the hospital. Serious complications occurred in 16 of 20 patients, precluding transplantation in 10 patients. The most common complication was bleeding, which required reoperation in four patients. Infections occurred in 6 per cent severe right heart failure occurred in 25 per cent, and renal failure occurred in 20 per cent of patients. Two patients sustained perioperative fatal cerebrovascular accidents, one from mural thrombus embolization from the native heart and the other from an air embolus during a de-airing procedure.

The Novacor system has been developed as a device for long, term support. Recently, a long-term survivor (>6 months) was discharged to a convalescent home to wait for a donor heart with the device providing circulatory support. Preclinical research is being conducted into an implanted electronic controller with rechargeable batteries to replace the control console. The power is transmitted from an externally worn battery pack across the intact skin by inductive coupling. The continued development of this system will allow patients to be fully ambulatory and rehabilitated prior to transplantation.

The Thermedics left ventricular assist device is a pusher plate blood pump that is pneumatically powered by an external control system (Fig. 7) 1653. The inflow and outflow conduits, which consist of Dacron grafts, are attached to the left ventricular apex and the ascending aorta, and each conduit contains a glutaraldehyde-preserved porcine xenograft. The pump bladder and metal components are covered with prosthetic interface polyester fibrils which facilitate the formation of a biological lining. Because of this unique lining, patients require much less aggressive antithrombotic therapy which consists of heparin for 5 days, followed by antiplatelet therapy. The device has a maximum stroke volume of 85 ml and provides synchronized counterpulsation or fixed rate pumping with a maximum cardiac output of 10 l/min.

At the Texas Heart Institute, 12 patients were supported with the Thermedics device as a bridge to transplantation; four of these were supported for more than 30 days (range 35–132 days), and three underwent successful cardiac transplantation. Post-implantation complications included minor postoperative bleeding and one device-related infection. A battery powered device has been approved for long-term support of the failing left ventricle at the Texas Heart Institute. This will allow patients to be fully ambulatory until the time of transplantation.

The ultimate goal of mechanical circulatory assist devices is safe and effective replacement of the human heart. Efforts in this direction have been stimulated by the realization that while orthotopic heart transplantation is the best means of replacing the heart, donor organ supply significantly limits this potential, until other alternatives, such as xenografting, become available. As the number of patients needing circulatory assistance for congestive heart failure grows, research and development of the total artificial heart continues to be of considerable interest.

As noted previously, the first clinical implantation of a total artificial heart was performed by Cooley in 1969. Since this historic implantation, a number of general problems have been identified and addressed. Of utmost importance has been the development of a power source adequate to fulfil the power requirements of the artificial heart. Electrical and pneumatically powered artificial hearts have been shown to be efficient and effective in long-term animal studies, and they serve as the power source for most clinically implanted artificial hearts today. These power sources are, however, often large and cumbersome, tethering the patient to a power console, limiting patient mobility, and preventing resumption of normal daily activities.

The second major problem area has been the materials used for the surfaces of the heart in contact with the blood. Segmented polyurethane has become the most commonly used material for blood sacs because of its flexibility, durability, and resistance to formation of thrombus. Unfortunately, these materials are not ideal, and all of the artificial hearts used in clinical practice have required anticoagulation to inhibit thrombus formation. Textured membranes on the pumping chamber surface have recently been developed to address this problem, and preliminary reports have been encouraging. Finally, a major drawback of many early artificial heart designs was their prohibitive size and resultant difficulty of implantation in the chest. Further refinements in the size of the hearts and development of prosthetic ventricles which could be attached to fabric cuffs sewn to the atria by a rapid connection system constitute a major advancement.

In 1982, DeVries electively implanted a pneumatically-driven ‘Jarvik–7’ artificial heart, developed at the University of Utah, into a 61-year-old man with congestive heart failure secondary to idiopathic cardiomyopathy (Fig. 8) 1654. The Jarvik–7 heart consists of two separate prosthetic ventricles constructed of a flexible diaphragm and a smooth blood surface fabricated of segmented polyurethane. Pulses of compressed air displace the diaphragm to empty the blood sac. Each ventricle has a stroke volume of 100 ml, although the smaller Jarvik–70 has a maximum stroke volume of 70 ml. The cardiac valves are monostrut tilting disc protheses. Connections to the natural atria are achieved by polymer coated Dacron felt cuffs. These are connected to the total artificial heart by a system of ‘quick connect’ polycarbonate segments. Connections to the great vessels are made by Dacron vascular prosthetic grafts. The drive lines of the ventricular air chambers consist of polyurethane tubing brought out percutaneously in the left upper quadrant and covered with velour skin buttons to retard infection and promote tissue ingrowth. The lines are connected to the control which regulates each ventricle independently. The pumping parameters are set to provide for near-complete filling of the ventricle depending on atrial pressure, and complete ventricular emptying. As the atrial or inflow pressures increase, the stroke volume and cardiac output increase proportionally.

This initial patient's postoperative course was complicated by multiple problems, including pulmonary insufficiency, renal failure, and seizures. However, he survived for 112 days, demonstrating that the total artificial heart could support patients for extended periods of time without mechanical failure. This landmark procedure provided a number of insights which led to significant improvements in total artificial heart function from a technical and clinical standpoint. Unfortunately, total artificial heart systems have been limited in their clinical use because of concerns about thromboembolic events, infection, and multisystem organ failure in subsequent implantations for permanent circulatory support.

In 1985, Copeland and associates performed the first successful total artificial heart bridge to transplantation in a 27-year-old man with a viral cardiomyopathy. Since 1985, a combined registry for clinical use of mechanical ventricular assist pumps and the total artificial heart in conjunction with heart transplantation has compiled data on 400 patients who required circulatory support. Of these 400 patients, 178 received the total artificial heart and 128 patients (71.9 per cent) subsequently underwent heart transplantation. The hospital discharge rate was significantly lower for patients who receive the total artificial heart (47 per cent) compared to those who received other forms of assistance (65–80 per cent). Major complications from the largest individual series included postoperative bleeding, mediastinal infection, renal failure and multisystem organ failure. Analysis of all total artificial heart recipients reported to this registry since 1985, however, suggests that there was a significantly higher rate of multiorgan failure (26 per cent) precluding transplantation than was seen in patients who received univentricular or biventricular assistance. The total artificial heart recipients also had a significantly higher rate of kidney failure (40 per cent). The infectious complication rate after bridge to transplantation with the total artifical heart is significantly higher, and in the face of post-transplant immunosuppression, it continues to be a major problem in bridging procedures.

There are a number of other artificial hearts under development in centres throughout the world, including Cleveland, Tokyo, Osaka, Berlin, Brno, and the University of Utah. The only other total artificial heart which has been implanted in a human is the Penn State heart (Fig. 9) 1655. This artificial heart is another electronically driven device consisting of two separate ventricles constructed of polyurethane. The left and right blood pumps are alternately actuated by an assembly of two pusher plates situated between the pumps. Rotation of a brushless direct current motor is transformed into motion of the pusher plates. The stroke volume of the ventricle is 100 ml and it can generate a maximum output of 12 l/min. The haemodynamic function of the heart is controlled by a built-in electronic automatic controlled system. For the left ventricular prosthesis, the system maintains a desired aortic pressure. On the right, the left atrial pressure controls pump rate. This heart has been implanted in three patients. The first patient was subsequently transplanted 10 days after implantation and died 17 days later of sepsis. The second patient received the heart after acutely rejecting a donor heart. He died 13 months later with the artificial heart in place: a second donor heart was not located due to the presence of preformed antibodies. The third patient underwent heart and kidney transplantation after 12 days of support, but eventually died of multisystem organ failure.

Further advances in circulatory support will be dependent upon continued progress in biological replacement, using allograft transplantation, xenografting, or skeletal muscle augmentation since these approaches offer a greatly superior quality of life. However, as the potential for biological replacement increases, new mechanical circulatory support systems will continue to grow as well. Clinical experience is accruing with a miniature axial flow pump which is inserted into the left ventricular cavity across the aortic valve to provide up to 4 l/min of mechanical circulatory support (Fig. 10) 1656,1657. The National Cardiovascular Center in Osaka, Japan is developing a paediatric ventricular assist device with a stroke volume of 20 ml to treat children with postoperative heart failure.

Appreciating the rapid development of present day total artificial hearts, it is likely that future total artificial hearts will be totally implantable, with improved geometry and materials able to eliminate thromboembolic complications and size constraints. Electrical energy provided by way of percutaneous inductive coupling will eliminate the necessity for percutaneous leads and the resulting infectious complications. With these technical improvements, mechanical replacement may become a competitive alternative to biological replacement, comparable to present day mechanical heart valves in relation to tissue valve prostheses.

Bolman R, et al. Circulatory support with a centrifugal pump as a bridge to cardiac transplantation. Ann Thoracic Surg 1989; 47: 108–12.

Braunwald E. Vasodilator therapy. A physiologic approach to the treatment of heart failure. N Engl J Med 1977; 297: 331–2.

Braunwald E., Kloner RA. The stunned myocardium—prolonged post-ischemic ventricular dysfunction. Circulation 1982; 66: 1146–9.

Cooley DA, et al. First human implantation of a cardiac prosthesis for staged total replacement of the heart. Trans Am Soc Artif Intern Organs 1969; 15: 252.

D'Ambra M, LaRaia P, Philbin M, Walkins W, Helgenberg A, Buckley M. Prostaglandin E&sub1;—a new therapy for refractory right heart failure and pulmonary hypertension after mitral valve replacement. J Thoracic Cardiovasc Surg 1985; 89: 567–72.

Dasse K, Chipman S, Sherman C, Levine A, Frazier O. Clinical experience with textured blood contacting surfaces in ventricular assist devices. ASAIO J; 10: 418–25.

Davies M. The cardiomyopathies: a review of terminology, pathology and pathogenesis. Histopathology 1984; 8: 363–93.

DeVries W, et al. Clinical use of the total artificial heart. N Engl J Med 1984; 310: 273–8.

Evans R, Manninen D, Gersh B, Hart L, Rodin J. The need for and supply of donor hearts for transplantation. J Heart Transpl 1984; 4: 57–60.

Farrar D, Rawson J, Letivak P, Cederwall G. Thoratec VAD system as a bridge to heart transplantation. J Heart Transplant 1990; 9: 415–23.

Fishbein M, Maclean D, Maroko P. The histopathological evolution of myocardial infarction. Chest 1978; 73: 843.

Fonger J, et al. Acute right ventricular failure following heart transplantation: improvement with prostaglandin E1 and right ventricular assist. J Heart Transpl 1986; 5: 317–24.

Fowler MB, et al. Dobutamine and dopamine after cardiac surgery: greater augmentation of myocardial blood flow with dobutamine. Circulation 1984 70 (Suppl 1): 1–103.

Frazier D, Cooley D. Use of cardiac assist devices as bridges to cardiac transplantation review of current status and report of the Texas Heart Institute's experience. In: Unger F, ed. Assisted Circulation. Berlin: Springer-Verlag 1989; 247–59.

Frazier O, et al. Treatment of cardiac allograft failure by use of an intraaortic axial flow pump. J Heart Transpl 1990; 9: 408–14.

Goenen M, Pedemonte O, Baele P, Col J. Amrinone in the management of low cardiac output after open heart surgery. Am J Cardiol 1985; 56: 33B–38B.

Gold H, Leinbach R, Mundth E, Sanders C, Buckley M. Reversal of myocardial ischemia complicating acute infarction by intraaortic balloon pumping. Circulation 1972; 45–46 (Suppl. 11): 22.

Gottlieb S, et al. Identification of patients at high risk for vascular complications of intraaortic balloon counterpulsation. Am J Cardiol 1984; 53:1135.

Griffith B, Kormos R, Hardesty R, Armitage J, Dummer J. The artificial heart: infection-related morbidity and its effect on transplantation. Ann Thorac Surg 1988; 45: 409–14.

Kantrowitz A, Tjonneland S, Freed P, Phillips S, Butner A, Sherman J. Initial clinical experience with intraaortic balloon pumping in cardiogenic shock. JAMA 1968; 203: 113–8.

Kantrowitz A, Wasfie T, Freid P, Rubenfire M, Wajszezuk W, Schork M. Intraaortic balloon pumping 1967 through 1982: analysis of complications in 733 patients. Am J Cardiol 1986; 57: 976–83.

Leier CV, Heban P, Huss P, Bush C, Lewis R. Comparative systemic and regional hemodynamic effects of dopamine and dobutamine in patients with cardiomyopathic heart failure. Circulation 1978; 58: 466–75.

Letsou G, Franco K, Detmer W, Wolvek S, Baldwin J. Pulmonary artery balloon counterpulsation: 1990; safe after peripheral placement. Ann Thoracic Surg 1992; 55: 741–6.

Liotta D, Hall C, Henly W, Cooley D, Crawford S, Debakey M. Prolonged assisted circulation during and after cardiac or aortic surgery. Am J Cardiol 1963; 399–405.

McGee M, et al. Retrospective analyses of the need for mechanical circulatory support (intraaortic balloon pump/abdominal left ventricular assist device or partial artificial heart) after cardiopulmonary bypass. Am J Cardiol 1980; 46: 135–42.

Miller R, et al. Combined dopamine and nitroprusside therapy in congestive heart failure. Circulation 1977; 55: 881–4.

Miller DC, Moreno-Cabral R, Stinson E, Shinn J, Shumway N. Pulmonary artery balloon counterpulsation for acute right ventricular failure. J Thoracic Cardiovasc Surg 1980; 80: 760–3.

Miller C, Pae W, Pierce W. Combined registry for the clinical use of mechanical ventricular assist pumps and the total artificial heart in conjunction with heart transplantation. Fourth official report—1989. J Heart Transpl 1990; 9: 453–8.

Mueller H, et al. The effects of intra-aortic counterpulsation on cardiac performance and metabolism in shock associated with acute myocardial infarct. J Clin Invest 1971; 50: 1885.

Mundth E. et al. Intra-aortic balloon pump assistance and early surgery in cardiogenic shock. Adv Cardiol 1975; 15: 159.

Norman J, et al. Total support of the circulation of a patient with post-cardiotomy stone heart syndrome by a partial artificial heart (ALVAD) for five days followed by heart and kidney transplantation. Lancet 1978; i: 1125–7.

Oaks T, et al. Results of mechanical circulatory assistance before heart transplantation. J Heart Transpl 1989; 8: 113–5.

Pennington D, et al. The importance of biventricular failure in patients with postoperative cardiogenic shock. Ann Thoracic Surg 1985; 39: 16–26.

Pennington DG, et al. Use of the Pierce-Donachy ventricular assist device in patients with cardiogenic shock after cardiac operations. Ann Thoracic Surg 1989; 47: 130–5.

Pennock J, Pierce W, Wisman C, Bull A, Waldenhausen J. Survival and complications following ventricular assist pumping for cardiogenic shock. Ann Surg 1983; 198: 469–78.

Portner P, et al. Implantable electrical left ventricular assist system: bridge to transplantation and the future. Ann Thoracic Surg 19889; 47: 142–50.

Schoen F, Bernhard W. Pathological considerations in temporary cardiac assistance. In Unger F, ed. Assisted Circulation. Berlin: Spring-Verlag 1989; 103–14.

Stinson EB, et al. Hemodynamic observations in the early period after human heart transplantation. J Thoracic Cardiovasc Surg 1975; 69: 264–70.

Sturm J, et al. Combined use of dopamine and nitroprusside therapy in conjunction with intra-aortic balloon pumping for the treatment of post cardiotomy low output syndrome. J Thoracic Cardiovasc Surg 1981; 82: 13–17.

Taylor S, et al. Intravenous amrinone in left ventricular failure complicated by myocardial infarction. Am J Cardiol 1985; 56: 29B.

General reading

Gibson DG. Heart failure: ventricular disease. In: Weatherall D, Ledingham J. Warrell D, eds. Oxford Textbook of Medicine, 2nd edn. Oxford: Oxford University Press, 1988; 13.84–13.98.

Smith T, Brunwald E, Kelly R. The management of heart failure. In: Brunwald E, ed. Heart Disease: A textbook of Cardiovascular Medicine. Philadelphia: WB Saunders, 1988; 485–543.

Unger F, ed. Assisted Circulation. 3rd edn. Berlin: Springer Verlag, 1989.

Weiner N. Norepinephrine, epinephrine and the sympathomimetic amines. In: Goodman L. Gilman A, Rall T, Murad F, eds. The Pharmacologic Basis of Therapeutics. London: Macmillan Publishing, 1985; 478–523.