Special techniques are required for circulatory support and protection of the myocardium and vital organs during detailed repair of the non-beating empty heart. During 35 years of open-heart surgery with cardiopulmonary bypass, methods have evolved that have reduced the mortality and morbidity of certain straightforward procedures almost to zero. Widespread application of modern cardiovascular technology now allows safer, more routine surgical correction of cardiac defects at the extremes of age, from the preterm neonate with complex congenital heart disease, to the late octogenarian with calcific aortic stenosis. Even hearts with extensive myocardial infarction and rupture can be repaired successfully without further damage using cold cardioplegic arrest. Nevertheless, exposure of blood to foreign surfaces still has widespread damaging effects (the whole body inflammatory response) and myocardial protection could be further improved. Short cardiopulmonary bypass and aortic cross-clamp times greatly reduce the potential for morbidity and mortality associated with cardiac surgery. There is scope for future research and development, and no room for complacency.

Cardiopulmonary bypass is a method of whole body perfusion, in which the pumping action of the heart and oxygenation of blood by the lungs are replaced by an extracorporeal circuit. Elimination of blood flow through the heart and lungs provides a controlled environment for surgery of the heart and great vessels, and less frequently for the major airways. The technique enables hundreds of thousands of cardiac operations to be performed annually with a mortality in some categories of patients of less than 1 per cent.

The main components of the bypass machine are an oxygenator and an arterial pump, which substitute for the lungs and heart respectively for the duration of the surgical procedure. Other components include provision for blood defoaming, filtration, temperature regulation, monitoring and safety devices. A simple circuit is shown in Fig. 1 1627.

Venous blood is diverted from the heart using a large bore cannula inserted in the right atrial appendage (Fig. 2) 1628. If the right side of the heart has to be opened separate cannulae are inserted into the superior and inferior vena cavae, usually introduced through the right atrium. Purse string sutures are snared around the incisions to produce a bloodtight and airtight seal. The venous cannula is connected to a reservoir sited below the operating table so that blood drains by siphon. Once established, this acts as a low pressure suction system controlled by the difference in height between the operating table and the venous reservoir.

There are two types of commercial disposable oxygenator, both of which contain an integral heat exchanger to regulate the blood outflow temperature.

These have been used since the early days of cardiopulmonary bypass. A rapid jet of oxygen is forced through a volume of blood to cause the bubbles, which produce a large blood–gas interface across which gas transfer occurs by partial pressure gradients. The smaller the bubbles the more efficient the oxygenation because of the relative surface area to volume ratio. However, smaller bubbles are more difficult to remove effectively before the blood is returned to the patient. Carbon dioxide also rapidly enters the gas phase and larger bubbles are more effective for its removal. In practice, the oxygen flow in bubble oxygenators creates sufficient bubbles in a range of sizes required to fulfil both functions. Clearance of bubbles before delivery of the arterialized blood to the circulation is achieved by exposure to surface-action silicone rubber, filtration, and settling (Fig. 4) 1630. High gas flows in the oxygenator increase both emission and dispersion time of microbubbles and high rates of blood flow decrease the time available for settling. Both are undesirable.

Although bubble oxygenators are relatively simple in construction, cheap, and easy to incorporate within the bypass circuit they have inherent disadvantages. Inefficient removal of bubbles may lead to widespread microembolization. A direct blood–gas interface damages the cellular elements of blood, particularly platelets, and may produce fibrinous microemboli. Blood gas control is imprecise. Blood levels of carbon dioxide can be maintained by varying the flow rate of oxygen, but this is at the expense of precise control of arterial oxygen tension. From a safety standpoint there is an inherent capacity for air or oxygen to be drawn through the oxygenator should the venous reservoir empty. This can result in massive arterial gas embolism. Equally, if the gas supply is interrupted, blood entering the bubble generator could induce catastrophic failure. Nevertheless, modern bubble oxygenators are reliable and are currently used in at least two-thirds of bypass operations worldwide.

Blood is passed over a membrane permeable to oxygen and carbon dioxide but which separates the two over a wide area. Venous blood drawn from a reservoir after filtration and de-airing is pumped through the oxygenator, where any remaining bubbles tend to pass through the inert microporous membrane into the gas phase. Gaseous nitrogen is highly insoluble and can be confined to the gas phase so an oxygen/air mixture is used to modulate the arterial oxygen tension. Carbon dioxide tension is regulated by varying gas flow or by altering carbon dioxide tension in the gas mixture. Blood gas and hydrogen ion concentrations can therefore be maintained with considerable accuracy. Far fewer microbubbles enter the circulation and massive gas embolism is almost impossible. Should venous obstruction lead to an empty reservoir air pumped into the oxygenator passes across the membrane into the gas phase, rendering passage to the patient unlikely.

The membrane itself may take the form of flat sheets or hollow fibres. Flat sheet membranes are usually arranged as a ‘fan fold’ with the blood on one side and the gas on the other providing an exchange surface area of 2 to 3 m². Baffles placed between the layers of the stack allow free flow of fluids, though the internal resistance of this type of oxygenator is inherently high. The oxygenator is therefore placed between the pump head and the patient. Hollow fibre oxygenators are compact devices containing membrane capillaries with an internal diameter of between 100 and 200 &mgr;m. Gas flows through the fibres, which are arranged in bundles. In both flat plate and hollow fibre oxygenators the blood flows in eddies, which bring more red cells to the gas exchange surface during transit. The intricacy of the blood path increases the time taken in setting up the apparatus because any air contained within the system must be purged with soluble carbon dioxide to prevent the formation of bubbles. Although the membrane surface presents a large surface of foreign material to the blood, the propensity for cellular activation or damage is less than that associated with a gas–liquid interface. The relative merits of bubble and membrane oxygenators are summarized in Table 1 475. Membrane oxygenators have distinct advantages and are now more competitively priced.

This is a simple mechanical double arm roller pump with speed control which allows the production of variable flow rates. For any given setting, the rollers rotate at a constant rate to provide an output that can equal or exceed the normal range for cardiac output. The rollers act on a smooth, flexible, and strong tube of silicone rubber which is less prone to spalliation (the splinting off of small particles) than is polyvinyl chloride, from which the remainder of the circuit tubing is made. The rollers are non-occlusive in order to avoid direct damage to the cellular elements of blood. The fundamental difference in mechanical pumping as opposed to physiological cardiac pumping is the absence of a pulsatile wave form, though pulsatile systems are now available.

Besides blood drained directly from the right atrium, the heart lung machine has two further pump heads to return blood from the pericardium or that vented from the left side of the heart in order to prevent cardiac distension or maintain a dry field. Because blood scavenged from these sites is mixed with air, fat globules, and debris from the operating site, it is returned to the cardiotomy reservoir for defoaming and filtration before re-entering the oxygenator. The cardiotomy reservoir may be integral with the oxygenator or separate. Oxygenated pumped blood is returned to the body on the arterial side of the circulation by a cannula situated in the ascending aorta or femoral artery (Fig. 2) 1628. Most surgeons employ a 40 &mgr;m pore-diameter filter in the arterial line to remove bubbles or particles that evade upstream filtration, particularly when a bubble oxygenator is used (Fig. 5) 1631. The need for an arterial line filter is less certain in membrane oxygenator circuits, since the filter itself may damage blood or generate gaseous or fibrinous emboli.

The extracorporeal circuit must be primed with fluid to exclude air before use. In the early days the machine was filled with whole blood; however, large transfusions of autologous blood generated problems and were associated with considerable haemolysis within the circuit. There was a high incidence of coagulopathy and organ damage related to capillary sludging and complement activation at low flow rates. This led to the clear fluid prime; most centres now use total haemodilution for adult surgery, unless the patient had a low preoperative haematocrit. Most adult circuits have a priming volume of between 2 and 2.5 litres. We have reduced this to 1.0–1.5 litres. At the onset of perfusion there is a rapid infusion of clear fluid, producing haemodilution and a haematocrit around 25 per cent. Babies and infants have a relatively small total blood volume in comparison to the circuit volume so that blood must be added proportionally. Introduction of the fluid prime has greatly reduced the morbidity and mortality of cardiopulmonary bypass.

Before bypass, the arterial and venous pipes are connected to the oxygenator and filled with a balanced electrolyte solution such as Ringer's lactate. The priming fluid is pumped around the blind loop circuit in order to dislodge and disperse adherent bubbles. Passage through the heat exchanger ensures that the prime is at a uniform temperature of between 35 and 37°C before bypass commences. A pre-bypass filter of 0.2 &mgr;m diameter may be used temporarily within the close circuit to remove bacteria and other particles. The tubes connecting the patient, pump and oxygenator are made of transparent polyvinyl chloride (PVC) and are usually supplied as customized sets assembled under clean conditions and sterilized by &ggr;-irradiation. The circuit is assembled using ‘no touch’ techniques.

For patients in advanced cardiac failure or with compromised renal or pulmonary function, maintenance of plasma colloid osmotic pressure is important. A haemofilter unit in the arterial line allows fluid to be removed during the bypass procedure. This reduces the final degree of haemodilution and eliminates extra volume in the form of cardioplegic solution. Ultrafiltrate is removed from the circulation by creating a pressure gradient across the semipermeable membrane of the haemofilter. Colloid osmotic pressure may be increased by the use of human plasma protein or cryroprecipitate-poor plasma. Artificial plasma expanders such as gelatine solution or hydroxyethyl starch can also be added to the prime, though the latter may be associated with more severe postoperative bleeding due to platelet dysfunction.

Anaesthesia is induced by neurolept analgesic technique, combining high dose analgesia with a neuroleptic or an anxiolytic agent. Anaesthetic maintenance during bypass usually relies on ‘top-up’ doses during rewarming. This, together with hypothermia, provides adequate but sometimes uneven anaesthesia. Continuous intravenous anaesthesia with volatile agents, or more recently with Propofol, has proved satisfactory despite cardiodepressant tendencies. Nitrous oxide should never be used during the peribypass period since it rapidly augments the size of intravascular bubbles. Deleterious effects have been reported following use of nitrous oxide even 20 min after the end of bypass.

Heparin (300–400 U/kg) is given 5 to 10 min before arterial cannulation. Activated clotting time should be at least 400 s. Blood gas analysis should be undertaken prior to commencing bypass to confirm that carbon dioxide tension lies within normal limits (4.5–6.0 kPa), since hypocarbia caused by hyperventilation interferes with autoregulation of cerebral blood flow.

The closed bypass circuit loop is then divided between clamps removing the pre-bypass filter. The surgeon inserts and connects the arterial cannula, taking care to exclude air bubbles. The perfusionist confirms arterial pressure in the line and then venous drainage is achieved by inserting a two-stage venous cannula into the right atrial appendage or by cannulating the superior and inferior venae cavae separately. The cannulae are connected to the venous line to complete the circuit.

The surgeon begins cardiopulmonary bypass by telling the perfusionist to ‘go on’; he opens the arterial line and starts the roller pump slowly. When the perfusionist is confident that an adequate volume of blood can be perfused without an excessive rise in resistance or pressure the venous line is opened, allowing blood to drain into the venous reservoir. This begins a siphon and as levels in the reservoir rise the speed of the arterial pump is increased, drawing fluid from the reservoir into the oxygenator and then the aorta. When the level in the reservoir is constant all the blood entering the machine is being returned to the patient. Full flow has been achieved and ventilation of the lungs is suspended. The arterial line pressure is checked, since excessive pressure may indicate dangerous malpositioning of the aortic cannula.

For most bypasses the temperature of the perfusate is reduced by the heat exchanger, which in turn causes systemic hypothermia in the patient. Moderate hypothermia (28–32°C) gives a measure of protection for the brain in case of episodes of poor perfusion or hypotension (Table 2) 476. Cerebral oxygen consumption is less than 50 per cent of normal at this temperature and circulatory arrest may be tolerated for 5 to 8 min (long enough for a skilled team to cope with equipment failure). Full unobstructed venous drainage empties the heart so that it does not eject. Oxygenated blood is returned from the heart lung machine at a rate of 2.4 l/m².min. Pressure is governed by the peripheral vascular resistance which is adjusted pharmacologically by the anaesthetist and perfusionist to obtain mean levels of between 50 and 70 mmHg (Fig. 6) 1632. Most oxygenators are run at 1:1 gas to blood ratio and there is seldom a problem achieving full saturation. The Po&sub2; often reaches 300 mmHg or more. While the desired temperature is reached the ascending aorta is clamped and cardioplegic arrest is induced by infusion of 1 litre of hyperkalaemic electrolyte solution at 4°C into the coronary arteries. If aortic regurgitation is present the aorta is opened and the solution delivered directly into the coronary ostia. Cardioplegia stops the heart in diastole for 30 min or more; this period can be increased by using topical hypothermia with iced slush in the pericardium. Further doses of cardioplegic solution are administered if electrical activity returns on the ECG. This provides ideal conditions for surgery to be performed on the cold flaccid heart which is emptied of blood, and easily manipulated.

Removal of the aortic cross-clamp reperfuses the myocardium and electrical activity resumes. Fifty per cent of patients show reversion to spontaneous rhythm; the remainder develop ventricular fibrillation requiring defibrillation. When the heart begins to beat, the aortic valve is open and blood is ejected against the resistance of the perfusion pressure. Great care must be taken to evacuate air from the cardiac chambers to prevent cerebral or coronary air embolism. This procedure is an integral and routine part of every open operation and air embolism by macrobubbles is rare.

Some 15 to 20 min before completion of the cardiac repair the surgeon instructs the perfusionist to rewarm the patient. This is achieved by passing warm water through the heat exchanger coils, though temperature gradients greater than 10°C between water and blood must be avoided. The duration of rewarming depends on the depth of cooling and the weight and body surface area of the patient. A heating blanket on the surface of the operating table assists in rewarming; the aim is to regain a temperature of 37°C by the time that the cardiac repair is complete.

After surgical correction the heart should be more efficient through relief of obstruction, elimination of intracardiac shunt, or increase in myocardial blood flow. This may be balanced by short-term injury from ischaemia, reperfusion, and surgical manipulation. Interstitial oedema reduces cardiac compliance and increased atrial filling pressures may be needed. Five minutes before weaning from cardiopulmonary bypass ventilation is restarted. A few forceful inflations of the lungs by hand helps to expel residual air from the left atrium or pulmonary veins. Cardiopulmonary bypass is discontinued by gradually reducing venous drainage, so that less blood is available for the heart-lung machine and more for the heart. As venous pressure rises the heart begins to eject; the venous line is then clamped and the atrial filling pressures gradually adjusted by controlled transfusions from the reservoir. The circulating blood volume is slowly augmented until arterial pressure is satisfactory. If systemic pressure is low coronary perfusion pressure may be inadequate. Catecholamine support is occasionally necessary in patients with myocardial dysfunction, though calcium is an excellent inotrope at this stage. Problems of rate or rhythm may be addressed by temporary epicardial pacing. Severe myocardial depression with left ventricular failure may require intra-aortic balloon counterpulsation.

Acute hypertension due to residual hypothermia and vasoconstriction is common, particularly after coronary surgery, and jeopardizes vascular anastomoses. Increased afterload also adversely affects the myocardium during the recovery period from aortic cross clamping and cardioplegia. This should be treated pharmacologically with intravenous glyceryl trinitrate or sodium nitroprusside.

Conventional perfusion techniques are associated with a reduction in the mean arterial blood pressure level to around 50 to 70 mmHg. This is attributable to a reduction in blood viscosity produced by haemodilution techniques. When considering optimal pressures and flow the brain is the critical organ. Early studies introduced the concept of an established formula for calculating adequate mean pump flow during perfusion. Flow rates of 2.0 to 2.4 l/m².min provide satisfactory maintenance of acid–base balance and oxygen consumption and have been shown to preserve acceptably normal patterns of cerebral function as indicated by the electroencephalogram. Flows greater than 2.4 l/m².min at normothermia produce no further increase in oxygen consumption, and increase blood trauma and the risk of gaseous emboli. Recommended rates of flow for certain groups are shown in Table 2 476.

It is usual to employ moderately hypothermic bypass (25–32°C), which reduces overall oxygen demand and permits the use of lower flows and pressures (see Table 2 476). With lower flow rates non-coronary collateral flow is reduced, thereby slowing cardiac rewarming during aortic cross-clamping. A flow rate of 1.2 l/m².min maintains global oxygenation adequately at 25°C. There is a close correlation between oxygen consumption and flow rates modified by hypothermia. Blood lactate concentration is an effective indicator of microcirculatory perfusion and should be less than 5 mmol/l. Mixed venous oxygen saturation can be used as an indicator of very poor perfusion by shunting.

A fall in mean pressure below 50 mmHg is associated with adverse EEG changes. Pressor agents such as noradrenaline are used to increase the cerebral blood flow, at the expense of mesenteric, renal, and extremity flow. Mean pressures above 100 mmHg should be avoided by pharmacological manipulation of peripheral vascular resistance. When reperfusing the ischaemic heart after cardioplegia or the fibrillating heart during intermittent aortic cross-clamping, a pressure of 70 to 80 mmHg is optimal, though greater pressures are necessary in patients with hypertrophied left ventricles. High systemic pressures should be avoided when discontinuing bypass since afterload reduction reduces cardiac work and oxygen consumption.

The standard roller pump produces non-pulsatile flow. Modifications of the roller pump systems have been used to impart pulsatility with synchronization that allows a systolic pressure pulse to be delivered during the patient's diastole. Early investigations suggested that increased peripheral resistance and lactic acidosis during cardiopulmonary bypass could be diminished with pulsatile flow. Measurements of change in vascular resistance, oxygen consumption, and lactate production suggested a generally beneficial effect of pulsatility. The increased energy imparted by pulsatile flow at a given mean arterial pressure is thought to improve lymph flow and provide more effective opening of capillary beds. In particular, urine output is improved by pulsatility. However, many perfusion-related problems that prompted the development of pulsatility have disappeared with recent improvements in oxygenators, priming solutions, and the general conduct of cardiopulmonary bypass. Its use is therefore not widespread.

Cerebral blood flow influences cerebral function following cardiopulmonary bypass, through adequacy of oxygen delivery for cerebral metabolism and by flow-dependent delivery of microemboli to the cerebral vasculature. The influence of carbon dioxide (Paco&sub2;) on cerebral autoregulation during cardiopulmonary bypass is important.

The majority of cardiac surgical procedures employ hypothermic cardiopulmonary bypass at average temperatures of 26 to 32°C, and hypothermia increases the solubility of carbon dioxide in blood. Until recently the influence of acid-base management techniques and changes in Paco&sub2; on cerebral blood flow regulation and flow/metabolism coupling have been poorly understood. At 37°C the intracellular ratio of H⫀/OH&supminus; concentrations is in equilibrium with that of the blood. This is approximately 1:16, maintained by protein buffering and modulated by variation of the Paco&sub2;. This establishes an intracellular to extracellular gradient for hydrogen ions of about 4:1. Disturbance of the blood hydrogen to hydroxyl ion ratio may be expected to cause intracellular biochemical inefficiency or damage. During hypothermic perfusion, it has been common practice to maintain blood gas values normally found at 37°C at all temperatures. This is known as ‘pH stat’ management. However, cerebral blood flow increases during cardiopulmonary bypass with pH stat, implying impairment of autoregulation. This is now thought to be due to artificial maintenance of blood pH at the physiological level of 7.42 during hypothermia by addition of CO&sub2; to the oxygenator. This changes the acid–base status towards a state of respiratory acidosis, which is known to have a vasodilating effect on cerebral blood vessels.

Rahn and Reeves advocated the so-called alpha-stat regulation concept, whereby preservation of a constant relative alkalinity of blood is favoured at different temperatures by allowing a decrease in Paco&sub2; and a simultaneous increase in pH during the development of hypothermia. In practice, this is achieved by maintaining Paco&sub2; constant by analysis of blood gases at 37°C without temperature correction throughout the procedure. Serial measurements of cerebral blood flow before, during, and after cardiopulmonary bypass with alpha-stat CO&sub2; regulation show that cerebral autoregulation remains intact when CO&sub2; is not added to the oxygenator. Cerebral hyperaemia and impaired autoregulation during hypothermic cardiopulmonary bypass can therefore be explained by the relative hypercarbic state resulting from addition of CO&sub2;. When a non-temperature-corrected Paco&sub2; of approximately 40 mmHg is maintained, cerebral blood flow is lower and cerebral autoregulation is better preserved than when the pH stat concept is used. With alpha-stat control cerebral blood flow is independent of pressure changes and dependent upon cerebral metabolism and oxygen consumption. This is important, since global increases in cerebral blood flow due to elevation of Paco&sub2; may occur at the expense of potentially ischaemic areas, and CO&sub2; induced cerebral dilation can critically reduce perfusion pressure in the circle of Willis. This may jeopardize areas of brain which are dependent on flow through stenosed vessels. In addition, cerebral hyperaemia may increase the flow-dependent delivery of microemboli.

The technique of deep hypothermia and circulatory arrest was developed in response to the need for safe correction of complex intracardiac congenital heart defects in small infants. This, together with cardioplegic arrest of the heart, provides an operative field that is completely bloodless and without cannulae to interfere with the intracardiac procedure. The technique is used for repair of all complex intra-atrial defects in infants weighing less than 7 kg.

Repair of other congenital defects in infants weighing less than 4 kg also uses circulatory arrest: cannulation of the inferior and superior vena cava may be technically difficult in small infants with tetralogy of Fallot or ventricular septal defects. Although intra-atrial surgery may not be required, the heart may be small and the technical aspects of operating during continuous cardiopulmonary bypass may be formidable. Circulatory arrest is also used in adult patients undergoing extensive surgery on the ascending, transverse, and descending thoracic or thoracoabdominal aorta.

Hypothermia causes a significant reduction in oxygen consumption, which is directly related to the magnitude of the decrease in temperature. The central nervous system is the least tolerant tissue to hypoxia at normothermia, followed by the kidneys, liver, and heart. The spectrum of ischaemic tolerance spreads from 5 to 6 min for brain cells to many hours for fat and skin. Hypothermia roughly doubles this tolerance time for each 5°C of cooling: the brain is protected from ischaemic damage for 6 to 9 min at 32°C, for 40 min at 20°C, and for 100 min at 12°C, though temperatures less than 18°C are associated with other problems. Hypothermia shifts the oxyhaemoglobin dissociation curve to the left and, as blood is cooled, there is a slight increase in the solubility of oxygen in the plasma. These effects counterbalance the reduction in systemic oxygen delivery. Acid–base status is generally maintained when cooling proceeds in a gradual fashion. Serum glucose is elevated and is associated with a decrease in plasma insulin levels. There is probably decreased systemic utilization of glucose, as well as inhibition of insulin release secondary to hypothermia. There are no major alterations in serum electrolytes. Circulating levels of catecholamines increase and free fatty acids are released, probably because of central nervous system activation of the sympathetic nervous system. The increase in glucose and free fatty acids during hypothermia may provide additional metabolic substrates for the heart and brain during the arrest interval.

Experimental evidence suggests that renal blood flow is substantially reduced during surface or core cooling. Reversible derangement of renal function occurs in patients undergoing hypothermia and circulatory arrest, though it is possible that decreases in cardiac output and increases in systemic vascular resistance during cooling may be chiefly responsible for renal dysfunction.

In 1971 Barrett-Boyes described the combination of surface cooling and core cooling with a pump oxygenator during surgical correction of congenital cardiac defects. Surface cooling to 32°C followed by core cooling on cardiopulmonary bypass to 18°C was used routinely for many years. There has been a recent trend away from prolonged surface cooling. Measurements of acid–base status, oxygen consumption, serum lactate concentration, and electrolytes, together with various serum enzymes including creatinine phosphokinase, transaminases, and dehydrogenases have shown no significant difference between patients cooled to 20°C by cardiopulmonary bypass alone and those undergoing combination of surface and core cooling. The surgical approach at most centres currently uses minimal surface cooling by exposure during anaesthesia in conjunction with core cooling on cardiopulmonary bypass.

During induction of anaesthesia and insertion of monitoring lines the infant is left undercovered, allowing the temperature to drift down. Rectal and nasopharyngeal temperatures are monitored with the knowledge that brain temperature always lags behind during cooling. The child is placed on the operating table on a temperature mattress which is adjusted for surface cooling. The head is packed in ice bags. When the temperature reaches 34 to 32°C median sternotomy is performed, the ascending aorta cannulated, and a single venous cannula inserted into the right atrial appendage. The bypass circuit is primed with packed red cells and a balanced electrolyte solution to create a haematocrit of 23 to 25 per cent during bypass. This degree of haemodilution improves peripheral perfusion during hypothermia while maintaining adequate oxygen delivery. During core cooling frusemide (furosemide) 1 mg/kg is administered if urine flow falls below 1 ml/kg.30 min. When cardiopulmonary bypass is commenced the perfusate temperature is gradually lowered, maintaining a 10°C temperature gradient between the water bath and the blood; this 10°C gradient should never be exceeded. Flow rates are adjusted to obtain 120 ml/kg.min in a newborn infant or 2.4 l/m².min in older infants. It is usual to add steroids to the pump prime prior to circulatory arrest since they have a protective effect against cellular oedema. When the rectal temperature reaches 18°C the aortic cross-clamp is applied, cardioplegic solution is infused into the aortic root and cardiopulmonary bypass is discontinued. The venous blood is drained from the right atrium into the oxygenator, after which the venous cannula can be removed, facilitating intracardiac manipulation in a completely bloodless field. When the surgical procedure is completed the venous cannula is reinserted and the pump slowly restarted. When venous return commences pump flow can be increased. Rewarming is carried out at approximately 2.4 l/m².min with a maximal 10°C gradient between rectal temperature, blood temperature, and the heat exchanger. The heart is de-aired and the aortic cross clamp released to reperfuse the myocardium. Spontaneous rhythm is rarely restored until the temperature reaches 30 to 32°C. At completion of rewarming the patient is weaned from cardiopulmonary bypass. Calcium chloride (100 mg/kg) is administered before termination of bypass for its inotropic effects and to increase the blood calcium level. After reversal of the action of heparin by protamine, platelet transfusions are administered to all infants since profound hypothermia results in thrombocytopenia.

The safe circulatory arrest time during profound hypothermia (18 to 20°C) is considered to be between 50 and 70 min. Safety is defined as the absence of any structural or functional damage as a result of the procedure. In practice, 30 min is regarded as completely safe, whereas 40 min has only a 90 per cent probability of safety. Sensitive tests of damage are difficult to apply in humans: experimental studies of cerebral morphology in animals have been used to extrapolate the limits of safety. In children who have recovered from a successful operation with an uncomplicated postoperative course, it is difficult to detect neurological deficit or specific abnormality of cerebral function when circulatory arrest has not exceeded 1 h at rectal temperature of 20°C. Psychological assessment and IQ measurements do not provide a sensitive index of injury sustained from this technique. Isolated incidents of cerebral damage usually occur in patients who have had a stormy postoperative course or who have suffered cardiorespiratory arrest at some time.

In adults undergoing total circulatory arrest, barbiturate infusion (thiopentone 30–40 mg/kg) is the only technique proven to reduce cerebral metabolic oxygen requirements. Profound hypothermia together with barbituates can reduce cerebral oxygen consumption to approximately 11 per cent of the normothermic level, provided that the critical ischaemic time remains at 30 to 45 min. After this, progressive cellular damage occurs, probably as a result of free radical generation. Superoxide is produced at reperfusion with subsequent uncontrolled intracellular release of calcium ions. It is hoped that administration of the calcium channel blocker nimodapine before ischaemia may improve outcome, but the search for clinically effective free radical scavengers continues.

Improved methods for protection of the ischaemic myocardium and advances in surgical technique have greatly reduced surgical mortality. Morbidity associated with many types of operation stems predominantly from the damaging effects of cardiopulmonary bypass. The importance of biocompatibility within this system consequently assumes great significance. Circulation of blood outside its natural endothelialized channels cannot be regarded as a biological process and damaging effects always occur. The artificial environment created by plastics, glass, and metal causes many alterations in the structure and function of blood, which flows through the circuit for periods of a few minutes to several hours (Fig. 7) 1633 (Table 3) 477. Collectively, the adverse clinical manifestations of extracorporeal circulation were known as the ‘post perfusion syndrome’ and more recently as the ‘whole body inflammatory response’. They may complicate postoperative convalescence and may contribute to death from multisystem organ failure, particularly in the very sick or those at extremes of age.

The interactions of the ‘contact system’ of plasma and of the activated humoral cascades with the cellular elements of blood and tissues of the body are complex (Figs. 8–11) 1634,1635,1636,1637. Many synthetic materials are exposed to blood within the various types of oxygenator. In bubble oxygenators the allegedly biocompatible materials include nylon, polyurethane, carbon-coated polyurethane, and polyethylene, all within a polycarbonate shell. The coils of the heat exchanger consist of carbon-coated aluminium. Membrane oxygenators may be of silicone rubber or microporous material, such as Teflon, polyethylene, polypropylene, or polytetrafluoroethylene (Gore-Tex). In bubble oxygenators the effects of direct gas–blood interaction include protein denaturation, complement activation, increased fragility of red cells, susceptibility to haemolysis, and denaturation of platelet membrane materials, causing platelet aggregation, clumping, and removal from circulation. These effects of gas interface oxygenators may be caused more by repeated renewal of the interface than by the nature of the interface itself. Membrane oxygenators have the advantage of eliminating blood–gas interface effects, though complement activation occurs by direct interaction of blood with foreign surfaces. Nevertheless changes in plasma proteins, lipoproteins, and red cells which occur with gas interface oxygenators are less evident with membrane oxygenators.

In clinical circuits with membrane oxygenators a gas–blood interface still occurs in the cardiotomy suction apparatus, which retrieves spilt blood from the operative field. This system consists of metal or plastic suckers and vents which aspirate and accelerate blood under negative pressure generated by a separate roller pump. An efficient and safe cardiotomy suction system should have a large storage capacity, minimal degree of suction to avoid blood trauma, minimal admixture of air, and reliable defoaming capacity. A wide variety of microemboli of gas, fat, fibrin, tissue debris, cellular clumps and foreign materials such as bone, wax, calcium from heart valves, glove powder, dust, and swab fragments is aspirated, and an efficient filter in the system is advantageous. Cardiotomy reservoirs have been designed to incorporate filters and defoaming chambers in an attempt to remove these microemboli.

The bubbles created by suction are qualitatively different from those generated in a bubble oxygenator, since they consist mainly of nitrogen, a relatively insoluble gas which may persist in microbubble form after passage through both the defoamer in the cardiotomy reservoir and the oxygenator. A large proportion of the red cell destruction which occurs during cardiopulmonary bypass results from mechanical trauma and turbulence induced by cardiotomy suction. Aspiration of blood from the pericardium and intracardiac chambers causes microaggregation of formed blood elements, and should be minimized.

An understanding of the nature of bioincompatibility in extracorporeal circuits requires an insight into the pathogenesis and effects of perfusion-related damage. When pronounced, this consists of a generalized increase in pulmonary capillary permeability (non-cardiogenic haemorrhagic pulmonary oedema), renal impairment, bleeding diathesis, neurological changes, and fever of non-infective origin. The ‘post perfusion syndrome’ causes generalized organ dysfunction which is transient and inconsequential after most straightforward adult operations, but may necessitate respiratory or renal support, blood transfusion, or re-entry for diffuse abnormal bleeding in small infants or after prolonged bypasses for complex repair.

‘Pump lung’ ranges from barely noticeable interstitial oedema to rare but fatal non-cardiogenic pulmonary oedema. Most patients have increased alveolar/capillary oxygen difference and increased fluid in the tracheobronchial tree. There is shunting due to both venous mixture and alveolar collapse. Total airflow resistance increases and there are measurable changes in pulmonary compliance. The changes in mechanical properties have been ascribed to non-cardiogenic increase in extravascular lung water (Fig. 12) 1638.

Both light and electron microscopy show consistent pathological changes in ultrastructure in lung biopsies taken from patients up to 4 h after bypass (Fig. 13) 1639. There is engorgement of the pulmonary vascular bed, microatelectasis and both interstitial and alveolar haemorrhage. There is damage and swelling of the capillary endothelial cells and, intracellularly, of the matrix compartment of the mitochondria and of the endoplasmic reticulum. The pericapillary space shows dispersion and there are accumulations of leucocytes and their degranulation products in apposition to the capillary membranes, with local endothelial damage. Oedema and rarefaction of the cytoplasm can be seen within Type I pneumocytes, and interstitial mast cells show loss in density and degranulation. These ultrastructural changes can be related to the duration of perfusion.

On total bypass the lung is hypoperfused and not ventilated, but is left partially inflated with 100 per cent oxygen. The lack of blood supply with subsequent damage to Type II pneumocytes produces a decrease or total lack of surfactant production, which may contribute to the development of alveolar collapse. The lung is also the primary filter for intravascular emboli during partial bypass with the lungs in circuit. Nevertheless, the likely mechanism for pulmonary dysfunction is that of the ‘whole body inflammatory response’, with complement activation as the stimulus for organ damage (Figs. 14, 15) 1640,1641. White cells activated by complement anaphylotoxins are known to aggregate in the pulmonary circulation; at the end of cardiopulmonary bypass and 50 per cent of the total circulating neutrophils are found transiently in the pulmonary capillaries (Fig. 16(a)–(c)) 1642,1643,1644. Release of protease enzymes (such as elastase) and oxygen free radicals can directly damage the pulmonary capillary membranes. Denaturation of &agr;-protease inhibitor by free radicals may well facilitate proteolytic damage.

In support of the whole body inflammatory response hypothesis, levels of the complement anaphylotoxin C3a have been shown to relate directly to postoperative organ system dysfunction (Fig. 17(a)–(c)) 1645. Complement activation occurs by the alternative pathway. Plasma concentrations of elastase rise by a factor of 10 to 30 times during the course of cardiopulmonary bypass. The end-point of the inflammatory response is a generalized increase in microvascular permeability. In normal lungs interstitial oedema does not create serious gas exchange problems, but after major surgery reduced compliance and increased work of breathing may lead to a need for prolonged intubation, increased risk of infection, and death. Such complications explain the need for routine postoperative ventilation of cardiac surgical patients in most units. However, ventilation is unnecessary for the great majority of adult cardiac operations with bypass times less than 1 h. In Oxford 85 per cent of adult patients are extubated in a recovery area within 2 h of leaving the operating room.

Treatment of a severe episode of non-cardiogenic pulmonary oedema may include infusion of adrenaline, which has anti-allergic, bronchial smooth muscle relaxing, and haemodynamic properties, and pharmacological doses of steroids (equivalent to methylprednisolone at 30 mg/kg) to improve pulmonary blood flow, stabilize the endothelial wall, and stop the leak of protease enzymes into the interstitial space. Positive end expiratory pressure is used to improve gas exchange by increasing functional residual capacity, decreasing intrapulmonary shunting, and increasing interstitial pressure. Aggressive tracheobronchial toilet is essential for clearance of pulmonary oedema fluid.

Severe acute renal failure requiring haemodialysis develops in about 2 per cent of patients after cardiopulmonary bypass, usually in association with multisystem organ failure. These patients are usually critically ill with poor left ventricular function and have a predicted mortality in excess of 70 per cent. The highest mortality rate is associated with the presence of respiratory failure, stroke, hypotension, and infection. Frank neurological deficit and major stroke occur in less than 2 per cent of patients undergoing cardiac surgery. Many of these problems are related to pre-existing carotid disease or macroscopic embolism from within a cardiac chamber. However, detailed psychometric testing discloses cerebral changes in the great majority of patients following cardiopulmonary bypass. Historically these have been related to microembolism, and recently microemboli have been demonstrated in the retinal vessels following cardiopulmonary bypass. The relationship between subjectively imperceptible and transient psychological changes to well-being remains unclear. It is not unusual to find patients reading a newspaper 24 h after cardiac surgery, though attention span is limited. Residual late psychometric impairment is usually mild, not apparent to the patient, and usually limited to mild cognitive impairment which has no significance in functional terms. Factors predisposing to major long-term intellectual dysfunction include cardiac failure before surgery and global impairment of left ventricular function. Patients and their relatives should be warned that some impairment of memory, concentration and reaction time may be noticed for a number of months after surgery, but that this usually recovers or improves to a level which will not interfere with normal activity. Patients in whom intraoperative control of Paco&sub2; is poor have a significantly higher incidence of postoperative neurological deficit on psychometric testing than those in whom Paco&sub2; levels are aggressively monitored and corrected. &agr;-Stat (as opposed to pH stat) regulation of blood gases preserves cerebral autoregulation and is the method of choice if cerebral effects are to be minimized.

Coagulation disorders seen after cardiopulmonary bypass are multiple and complex, and detailed description is beyond the scope of this chapter. Postoperative bleeding is minimized by careful surgical haemostasis, though abnormal bleeding may result from heparin–protamine imbalance, thrombocytopenia combined with abnormal platelet function, defects of the clotting cascade, and abnormal fibrinolysis. Bleeding tendencies are rare in patients undergoing short, moderately hypothermic bypass using membrane oxygenation and controlled cardiotomy suction to minimize platelet damage. When abnormal bleeding occurs in the presence of normal clotting factors, platelet dysfunction may be caused by binding of fibrinogen degradation products to surface membrane receptors. The protease inhibitor aprotinin protects against this, and can be used in selected patients at substantial risk from postoperative haemorrhage (some early reoperations and patients with bacterial endocarditis). We do not advocate routine use of this drug and we do not use it for patients undergoing coronary bypass grafts. We have not found it to be useful in patients undergoing profound hypothermia and circulatory arrest, in whom platelet dysfunction is particularly common.

Treatment of abnormal bleeding depends on the precise cause, but includes protamine administration according to activated clotting time, infusion of platelets, fresh frozen plasma, or cryoprecipitate, and, rarely, the administration of an antifibrinolytic agent such as epsilon aminocaproic acid. Cardiac surgeons know that warm fresh blood transfused directly from a donor is the best treatment for intractable diffuse abnormal bleeding, though haematologists are at odds with this practice. Topical administration of cryoprecipitate or aprotinin to bleeding surfaces within the pericardium has proved effective when the chest cannot be closed safely. Reinfusion of shed mediastinal blood by an autotransfusion system is used routinely in many centres.

For patients with profuse haemorrhage in the recovery area there should be a low threshold for re-entry, even if surgical bleeding is deemed unlikely. A surgical bleeding point is often identified; if not, packing of the open chest may be necessary. Multiple transfusions of donor blood and even continuous reinfusion of poor-quality shed mediastinal blood greatly increase overall morbidity from pulmonary and renal dysfunction.

Operations on the heart are facilitated by having the ascending aorta cross-clamped and the heart bloodless and quiet through cessation of coronary blood flow. Consequently the heart is subject to global myocardial ischaemia and, even under conditions of moderate or deep hypothermia, the anoxic safe period without irreversible and fatal damage does not exceed 30 min. As early as 1955 Melrose and colleagues at the Hammersmith Hospital (London) used chemical cardioplegia with potassium citrate to arrest the heart and reduce myocardial oxygen demand. Although this technique was initially successful, it was abandoned when reports from the United States of America described a high incidence of ventricular fibrillation postoperatively and myofibrillar necrosis histologically following 30 min of ischaemia. Cardiac standstill was then induced by anoxic aortic cross-clamping or by continuous electrical fibrillation. Myocardial metabolic requirements were reduced by systemic hypothermia plus topical cooling in the pericardium. Intermittent coronary perfusion was used during the procedure to supply oxygen.

Hypothermia reduces cardiac rate and myocardial oxygen consumption but at temperatures below 28°C, coronary resistance increases and subendocardial blood flow decreases. Topical hypothermia with iced saline slush in the pericardium can selectively decrease myocardial temperature below systemic temperature and increases myocardial protection. However, the heart fibrillates and continuous electrical fibrillation itself produces a 50 per cent rise in oxygen consumption and rapid fall in ATP, creatinine phosphate, and glycogen stores. During coronary artery perfusion in a fibrillating heart blood flow is diverted away from the subendocardial surface and any rise in left ventricular end diastolic pressure further decreases subendocardial perfusion.

Intermittent coronary perfusion with moderate hypothermia and cardiac standstill induced by ventricular fibrillation has been used successfully for myocardial protection for many years. Nevertheless, this technique provides inadequate myocardial oxygenation, metabolite substrate delivery, and washout of catabolic products. Intermittent perfusion for 5 min followed by ischaemia for 15 min decreases myocardial contractility, compliance, and increases myocardial oedema. If the cyclical periods of aortic clamping are prolonged beyond 90 min there is a progressive fall in myocardial adenine nucleotides. Cellular swelling and myocardial oedema follows depletion of energy stores and failure of active membrane transport of small molecules, particularly sodium, potassium and chloride. Severe myocardial oedema reduces ventricular diasystolic function, stroke volume, and cardiac output.

In spite of surgical disadvantages, cardiac operations have been performed on normothermic, perfused, empty, beating hearts, a method advocated as optimal as late as 1975. Valve surgery can be carried out on a perfused empty beating heart by the technique of individual coronary artery perfusion, using small individual cannulae placed in the ostia of the right and left coronary arteries and perfused with oxygenated blood by way of separate pumps and lines from the heart–lung machine. Continuous coronary perfusion has been combined with ventricular fibrillation, and similar techniques are still successfully applied by some to coronary artery surgery. Nevertheless, the need for prolonged periods of cardiac arrest and improved myocardial preservation renewed interest in chemical cardioplegia. The hypothesis underlying the use of cold cardioplegic myocardial preservation is that it reduces myocardial oxygen demand during the ischaemic period of aortic cross-clamping to such low levels that the myocardial energy stores are sufficient to maintain cell structure and the energy-dependent cell membrane pumps that preserve transcellular gradients of sodium, potassium, calcium, and magnesium. Thus myocardial cell viability and function are preserved. Myocardial histological studies have shown a higher incidence of endomyocardial damage with continuous coronary perfusion and intermittent aortic cross-clamping than with cold cardioplegic arrest.

Experimental studies undertaken in the 1970s illustrated that the colder the heart during the ischaemic period, the better the myocardial protection. They also showed that abrupt electromechanical dissociation of the heart at the onset of ischaemia combined with rapid myocardial cooling provided better myocardial protection than either method alone. When the heart is electromechanically quiescent the energy demands of the myocardium are determined primarily by temperature, and myocardial energy is derived primarily from anaerobic metabolism of glycogen and glucose. The low energy output of anaerobic metabolism is sufficient to maintain myocardial viability during prolonged ischaemia if the energy demands are substantially lowered by cooling.

Experimental research and clinical experience indicate that the cardioplegic solution should be at 4°C and the myocardial temperature should be lowered to between 12 and 15°C at the start of the procedure and then maintained at between 15 and 20°C throughout the ischaemic period.

The ideal cardioplegic solution must be cold when infused (4°C), should induce cardioplegia rapidly, should produce no direct cell membrane damage, and should minimize intracellular ionic changes. The chemical principles for inducing cardiac arrest are shown in Table 4 478. Potassium rapidly infused into the cross-clamped aorta depolarizes myocardial cells, producing sustained diastole. Asystole decreases the amount of ATP consumption and reduces oxygen demand. A potassium dose of 20 to 24 mEq in a crystalloid delivery solution produces a prompt arrest. Doses above 30 mEq/l increase coronary vascular resistance because of a calcium-activated rise in left ventricular wall tension. When blood is used in the cardioplegic solution to increase oxygen delivery the solution must contain 30 mEq/l KCl since some potassium is taken up by the red cells. Magnesium is used at a concentration of 50 mEq/l to depress both the inherent rhythmicity of pacemaker cells and myocardial contractility. Pacemaker cells are unstable due to a slow inward flux of sodium which eventually reduces action potentials to the firing levels. Magnesium blocks the inward flux of sodium and interferes with repolarization. Calcium and magnesium compete at cell membrane receptor sites to activate or slow neuromuscular transmission. Magnesium inhibits the release of calcium by the sarcoplasmic reticulum, thereby inhibiting calcium influx and stabilizing potassium channels. Magnesium also suppresses myocardial ATPase activity and activates enzymes which transfer phosphate from ATP to ADP.

Because of the excellent operating conditions and effectiveness of myocardial preservation, cold cardioplegia arrest has become the most commonly used method of operative myocardial preservation. Hypothermic solutions in clinical use vary considerably in their components. Table 5 479 shows the constituents of St Thomas's cardioplegic solution used at the Oxford Heart Centre. This uses a temperature of around 4°C. Hypothermia of the perfusate appears to be the most important factor in lowering of cellular energy requirement during ischaemia. In the 1960s, both cold blood and crystalloid solutions without any additional arresting agent were used experimentally and shown to provide satisfactory myocardial protection in the dog heart for ischaemic periods of 60 min. Arrest using potassium or procaine at normothermia proves only slightly better than unmodified normothermic ischaemic arrest. Nevertheless, the effects of an arresting agent plus hypothermia are additive and result in better post-ischaemic ventricular function than when either is used alone. Utilization of cellular energy stores is lowered by progressive levels of myocardial hypothermia and the lower the myocardial temperature of the arrested heart, the better the post-ischaemic myocardial performance.

The majority of cardioplegic solutions contain either potassium or procaine as the arresting agent. Potassium produces arrest by membrane depolarization, procaine by preventing the propagation of an electrical stimulus. Studies comparing the two agents have found no difference in post-ischaemic myocardial performance; both appear to decrease cellular energy requirements equally, in spite of their different mechanisms of action. Magnesium has also been used as an arresting agent and is present in high concentration in the Kirsh solution. Additives such as nifedipine and verapamil have been used effectively in animal experiments.

Cardioplegic solutions appear to work well both with and without calcium in the medium. Proponents of calcium-containing solutions have supported its use by citing the ‘calcium paradox’. This syndrome was described in isolated rat heart experiments following normothermic aerobic perfusion with calcium free solution followed by perfusion with a calcium-containing medium. This perfusion sequence resulted in an irreversible loss of electrical and mechanical activity of the heart. The syndrome appears to be modified by profound hypothermia which may account for the clinical results achieved with calcium-free solutions. Nevertheless, the presence of calcium has not been shown to be harmful and may well prove beneficial.

Cellular acidosis is markedly deleterious for anaerobic metabolism, and this supports the use of a solution with a pH of 7.4 or above. Animal experiments have shown poor post-ischaemic recovery following the use of acidic cardioplegic solutions. Solutions used in clinical practice therefore contain bicarbonate or tris(hydroxymethyl)aminomethane (THAM) as their buffer.

Glucose is used to provide substrate during anaerobic metabolism in some types of cardioplegic solution, though the benefit has not been conclusively proved.

Wide variations in osmolarity of the perfusate are clearly harmful and some of the deleterious effects of the original Melrose solution (450 mOsmol) have been attributed to hyperosmolarity. Post-ischaemic ventricular performance following use of hyperosmolar solutions is poor when compared to the effect of solutions with an osmolarity between 300 and 320 mOsmol. Hypo-osmolar solutions (272 mOsmol) cause myocardial oedema. It therefore appears the solution should be at least iso-osmolar, and if hyperosmolar should not exceed 380 mOsmol.

Cardioplegia undertaken with cold oxygenated blood plus an arresting agent has the theoretical advantage of intermittently replenishing cellular energy levels by aerobic metabolism during period of reinfusion. However, it has been suggested that the lower limit of blood temperature should be 22°C since at lower temperatures the rheology of blood may be altered, with sludging of red cells and deleterious effects on the microcirculation. Blood cardioplegia therefore has the potential disadvantage of allowing a higher rate of energy consumption during ischaemia because the myocardial cooling is not as efficient as that with crystalloid solutions. Whereas myocardial temperatures below 15°C can easily be achieved with crystalloid solutions at 4°C, this degree of cooling cannot be achieved with blood at 22°C. Nevertheless experimental findings and practical clinical experience with blood cardioplegia have shown little difference between the two methods.

Blood cardioplegia demands more elaborate systems for administration, and a higher concentration of potassium chloride (28–30 ml/l) is required to produce asystole. This need for a higher potassium chloride concentration is probably due to potassium influx into the red blood cell. Since the hyperkalaemic blood used for cardioplegia is usually returned to the cardiopulmonary bypass circuit, a significant serum potassium rise is more likely then with crystalloid cardioplegia.

Since the chemically arrested heart at 15°C requires 0.3 to 0.5 ml oxygen/min.100 g, the need to supply oxygen to the heart by blood cardioplegia is questionable. Oxygen demand of this magnitude could well be met be oxygenating asanguinous solutions, and may in any case be met by non-coronary collateral flow (estimated at about 2 ml/min.100 g). Clinical trials comparing blood with crystalloid cardioplegia have shown that crystalloid cardioplegia provides more sustained and consistent myocardial protection. With blood cardioplegia, multiple doses seem necessary for comparable protection.

Evidence has accumulated to suggest that normothermic potassium-containing oxygenated blood cardioplegia (the so-called ‘hot shot’) is beneficial at the end of hypothermic cardiac arrest. Replenishment of oxygen and substrate by this technique appear to reduce ischaemia-induced ventricular failure.

Cold crystalloid cardioplegia produces rapid induction and sustained maintenance of cardiac arrest with profound myocardial hypothermia. Operative techniques should avoid injury to the coronary arteries or myocardium at the time of infusion and must subsequently avoid damage from reperfusion with blood during release of the aortic cross-clamp. In practice variable rates and levels of myocardial cooling may occur, particularly in the presence of severe coronary artery disease and through rewarming by non-coronary collateral flow. Retrograde delivery of cardioplegia via the coronary sinus has been used to circumvent the problem of coronary occlusion. For this method a balloon catheter is inserted through a purse string suture in the right atrial wall and manipulated into the coronary sinus. The balloon is inflated and cardioplegia instilled under low pressure.

The efficacy of cold crystalloid cardioplegia is directly related to the level of myocardial cooling. Myocardial temperature monitoring can be performed using a needle thermister in the septum in order to sustain a myocardial temperature below 20°C and preferably as close to 10°C as possible. The thermister is inserted 1.5 cm to the right of the left anterior descending coronary artery and angled superiorly into the septum. When delivering cardioplegia, attention should be focused on those areas of the myocardium known from the coronary angiogram to have the poorest blood supply.

For coronary artery bypass, repair of most congenital conditions, and mitral valve surgery, infusion of cardioplegia occurs directly through a needle in the aortic root after the cross-clamp application. This route can also be used in the treatment of aortic stenosis following vent insertion to ensure against left ventricular distension. In patients with aortic regurgitation, the aorta is opened and cardioplegia delivered directly into the coronary ostia through hand-held cannulae. Reinfusion of cardioplegia is often necessary for maintenance of myocardial arrest and adequate hypothermia. Further infusions are usually given at intervals of 25 to 30 min whilst the cross-clamp is in place. Should electrical activity resume earlier, because of collateral flow, reinfusion is undertaken sooner. Cardioplegic reinfusion washes out metabolic end-products and, if glucose-containing solutions are used, may replenish glucose for anaerobic metabolism. When myocardial temperature is measured directly, cardioplegia is repeated when the septal temperature rises above 20°C. The rate of myocardial temperature rise can be slowed by systemic hypothermia and topical hypothermia with ice slush in the pericardial sac. This reduces the tendency to rewarm through non-coronary collaterals that connect the left ventricle to the bronchial circulation.

During reperfusion, after cross-clamp removal and recovery of the heart prior to weaning from cardiopulmonary bypass, the heart is particularly susceptible to high arterial pressure. At the time of cross-clamp release, perfusion pressure is therefore deliberately reduced to around 50 mmHg until the heart obtains tone and begins to beat or fibrillate. There is a high initial rate of spontaneous resumption of sinus rhythm, although ventricular fibrillation is common during the early reperfusion period. This probably results from electrical instability during resolving myocardial hypothermia and the transition of the cellular membrane from an arrested state to normal function. When the blood temperature is 35 to 37°C at aortic unclamping, rewarming of the hypothermic myocardium usually takes 5 to 10 min. Attempts to discontinue cardiopulmonary bypass before full recovery with stable rhythm may well prove unsuccessful.

When properly administered, cold potassium cardioplegia appears to maintain the entire myocardium in a protected and completely reversible state during periods of ischaemia lasting more than 2 h. Clinical reports have included patients with ischaemic periods of up to 3.5 h with normal postoperative myocardial function. Cardioplegia is superior to either intermittent coronary perfusion or prolonged ischaemia modified by profound topical hypothermia. Factors that decrease the effectiveness of coronary perfusion or topical hypothermia, including left ventricular hypertrophy, severe coronary disease, and prolonged aortic cross-clamping, do not appear to decrease the effectiveness of cold potassium cardioplegia. Further refinement in the composition of cardioplegic solutions and methods of administration will probably lead to even better clinical results.

Akins K. Non cardioplegic myocardial preservation for coronary revascularisation. J Thoracic Cardiovasc Surg, 1984; 88: 174–7.

Bretschneider HJ. Myocardial protection. Thoracic Cardiovasc Surg, 1908; 28: 295–8.

Brody WR, Reitz BA. Topical hypothermic protection of the myocardium. Ann Thoracic Surg, 1975; 20: 66–70.

Buckberg GD. A proposed ‘solution’ to the cardioplegic controversy. J Thoracic Cardiovasc Surg, 1979; 77: 803–6.

Fox LS, Blackstone EH, Kirklin JW, Stewart RW, Samuelson P. Relationship of whole body oxygen consumption to perfusion flow rate during hypothermic cardiopulmonary bypass. J Thoracic Cardiovasc Surg, 1982; 83: 239–43.

Fox LS, Blackstone EH, Kirklin JW, Bishop SP, Bradley EL. Relationship of brain blood flow and oxygen consumption to perfusion flow rate during hypothermic cardiopulmonary bypass. J Thoracic Cardiovasc Surg, 1984; 87: 658–64.

Hearse DJ. The protection of the ischaemic myocardium: surgical success vs clinical failure. Prog Cardiovasc Dis, 1988; 30: 381–402.

Kirklin JW, Conti VR, Blackstone EH. Prevention of myocardial damage during cardiac operations. N Engl J Med, 1979; 301: 315–21.

Kirklin JK, Westaby S, Blackstone EH, Kirklin JW, Crenoweth DE, Pacifico AD. Complement and the damaging effects of cardiopulmonary bypass. J Thoracic Cardiovasc Surg, 1983; 86: 845–51.

Mavroudis C. To pulse or not to pulse. Ann Thoracic Surg, 1978; 25: 259–61.

Murkin JM, Farrar JK, Tweed WA, McKenzie FN, Guiraudon G. Cerebral autoregulation and flow/metabolism coupling during cardiopulmonary bypass: The influence of PaCO&sub2;. Anesth Analg, 1987; 66: 825–32.

Rahn H, Reeves RB, Howell BJ. Hydrogen ion regulation, temperature and evolution. Ann Rev Resp Dis, 1975; 112: 165–9.

Sade RH, Bartles DM, Dearing JP, Campbell LJ, Loadholt CB. A prospective randomised study of membrane vs bubble oxygenators in children. Ann Thoracic Surg, 1980; 29: 502–7.

Swain JA. Cardiac surgery and the brain. N Engl J Med, 1993; 329: 1119–20.

Takamoto S, et al. Comparison of single dose and multiple dose crystalloid and blood potassium cardioplegic during prolonged hypothermic aortic occlusion. J Thoracic Cardiovasc Surg, 1980; 79: 19–24.

Utley JR, ed. Pathophysiology and Techniques of Cardiopulmonary Bypass. Vol. 1. Baltimore: Williams and Wilkins, 1982.

Utley, JR, ed. Pathophysiology and Techniques of Cardiopulmonary Bypass. Vol. 2. Baltimore: Williams and Wilkins, 1983.

Van Oeveren W, et al. Effects of aprotinin on haemostatic mechanisms during cardiopulmonary bypass. Ann Thoracic Surg, 1987; 44: 640–5.

Westaby S. Aspects of biocompatibility in cardiopulmonary bypass. Crit Rev Biocompatibility, 1987; 3: 193–234.

Westaby S. Organ dysfunction after cardiopulmonary bypass: a systemic inflammatory reaction initiated by the extracorporeal circuit. Intensive Care Med, 1987; 13: 89–94.

Myocardial protection during cardiac surgery with cardiopulmonary bypass. In Kirklin JW, Barratt Boyes BG, eds. Cardiac Surgery. New York: John Wiley and Sons, 1986; 83.