Cardiopulmonary Bypass: Principles and Practice

Previous Chapter | Next Chapter >




Quick Links to Sections in this Chapter










T. Gourlay and K. M. Taylor:Department of Cardiac Surgery, National Heart and Lung Institute, Imperial College, Hammersmith Hospital, London W12 ONN, United Kingdom.

Interest in the pulsatile nature of blood flow and the physiologic importance of this flow modality is not new. This area of research predates the clinical application of cardiopulmonary bypass (CPB) by a considerable time. Indeed, the very earliest physicians expressed an interest in the pulse and the pulsatile motion of the blood vessels. Aristotle (384–322 B.C.) noted, "The blood in animals throbs within their veins," and with particular reference to the heart, "The veins pulsate as a whole synchronously and successively inasmuch as they depend on the heart. It keeps moving and so do they." These apparently simple and logical conclusions, made without the aid of modern instruments, demonstrate that the pulsatile nature of the circulation did not go unnoticed by these early observers. However, it was not until much later that its significance could be thoroughly investigated.

Many early investigations of the significance of the pulsatile nature of blood flow were carried out by using isolated organ preparations. Hamel (1) demonstrated that pulsatile flow was of considerable importance to renal function. These findings were confirmed by Gesell (2), who suggested that this maintenance of function was the consequence of improved gas exchange at the capillary level, together with freer flow of lymph. Using a depulsed isolated organ preparation, Kohlstaedt and Page (3) further confirmed the importance of pulse pressure to kidney function, and in particular demonstrated that renin secretion is affected by the flow modality employed. Renal function and blood flow were the focus of many of the early flow studies. However, other factors were also studied in an effort to confirm the importance of the pulsatile nature of blood flow. McMaster and Parsons (4) demonstrated that lymph flow is greatly reduced during periods of nonpulsatile blood flow. The pulsatile nature of blood flow was studied intensely during the period leading up to the development of clinical CPB, and the results, although not universally accepted, offer considerable evidence supporting the importance of pulsatile blood flow, rather than simply bulk blood flow, in the maintenance of normal physiologic organ function.

These studies were of greater academic interest than practical importance until the development of clinical CPB. With the development and application of clinical CPB, it became necessary to support the total circulation to enable heart surgery to be performed. The early heart–lung machine of Gibbon (5) utilized a nonpulsatile pumping mechanism of the type designed by DeBakey (6) for this purpose. Despite the fact that this type of mechanism appeared to work well during CPB, interest still existed in examining the application of pulsatile blood flow during CPB. Mindful of the early results of the study of pulsatile blood flow in isolated organ preparations, the pioneers of CPB and cardiac surgery investigated the possibility of utilizing a pulsatile blood flow regime during CPB. Several attempts were made to produce pulsatile blood pumps for clinical CPB, but clinical acceptance was hampered by clinicians' fears of the complexity of such devices and of the potential for damage to formed blood elements. These fears persisted until a reliable commercial pulsatile blood pump (Fig. 10.1) was made available in the late 1970s, more than 20 years after the first clinical application of CPB.

FIG 10.1. Stockert pulsatile roller pump system (Stockert, Munich, Germany). This system is seen in a modern console-based configuration, but the pump modules and control system do not differ significantly from those in the first pump of this type produced.

The availability of this relatively simple modified roller pump system spurred a number of investigators to readdress the issue of pulsatile blood flow in the context of CPB. This area of research persists to this day and has led to some degree of understanding of the importance of pulsatile blood flow in the maintenance of normal physiologic response patterns during CPB. Nonetheless, the issue of pulsatile blood flow during CPB remains controversial (7).

During the years since the introduction of clinical CPB, research into the significance of pulsatile blood flow has focused mainly on two areas—metabolic effects and hemodynamic effects. In addition to these areas of research, the development of technology to generate pulsatile blood flow that is compatible with other components of the CPB circuit continues.


Back to Quick Links

Before considering the benefits attributed to pulsatile blood flow during CPB, one should consider the principles underlying these benefits. The two fundamental differences between pulsatile and nonpulsatile flow relate to flow–pressure architecture (shape of the flow–pressure complex) and energy delivery.

Blood flow–pressure architecture

In pulsatile blood flow or pressure, a specific shape or architecture characterizes the pulse. The architecture is determined by the mechanism generating the blood flow and its interaction with the environment in which it operates. Some characteristics of pressure wave shape that can be used to describe the architecture of a pulse waveform (Fig. 10.2) include frequency, amplitude, rise time, decay time, and mean pressure or flow.

FIG 10.2. Some of the factors that can be employed to describe the architecture of a pulse pressure waveform.

All these parameters have been employed to describe pulsatile blood flow during clinical and experimental CPB (8–10), and they represent the descriptive parameters most accessible to the clinician. In the absence of more advanced measurement and monitoring techniques, the arterial blood pressure waveform is commonly used to indicate "how pulsatile" a perfusion is during clinical practice. Although this constitutes a reasonably useful indicator capable of discriminating between pulsatile and nonpulsatile perfusion, it is not sufficient to describe fully the pulsatile properties of a particular flow modality.

Since the introduction of the pulsatile roller pump, most clinical investigations of pulsatile blood flow have employed the pulse pressure simply as a marker to confirm whether some form of pulse is present. Our group has employed a standard pulsatile pump control during more than two decades of clinical and experimental CPB in animals (11). The settings used have been based on safety considerations and represent the midrange of control options available to the clinician without claiming to offer optimum or physiologic pulsatility. The arterial pulse waveform generated with this configuration is not physiologic insofar as it does not resemble that of the native heart (Fig. 10.3).

FIG 10.3. Graph showing typical roller pump (A) and heart-generated (B) pulsatile blood pressure taken from the radial artery of a patient undergoing open heart surgery.

Roller pump-generated pulsatile flow has been described as "ripple flow" (12). Nonetheless, this suboptimal flow modality has been associated in many investigations with significant clinical advantages, ranging from maintenance of relatively normal hemodynamics to reduced patient morbidity and mortality (13–16).

It is conceivable that the benefits attributed to roller pump-generated pulsatile flow would be further enhanced if the pulse output were made more physiologic. This would seem logical, and some in vitro and clinical experiments have assessed the degree to which the roller pump mechanism generates physiologic pulse pressure–flow architecture. Our experiments, in which an in vitro model of the human systemic circulatory system was used (17,18), suggest that although the roller pump offers a range of control for almost every aspect of pressure architecture, it is not capable of generating truly physiologic pulsatility. In our experience, when the maximum pulsatile control configurations are employed clinically, the pressure architecture differs little from the conventional configuration (18), and both configurations produce unphysiologic waveforms. As it exists today, the roller pump mechanism may not therefore be the best one for delivering physiologic pulsatile blood flow architecture. This apparent limitation makes it difficult to understand the clinical benefits that have been demonstrated with this system. Optimizing the output profile within the performance envelope of this mechanism does not appear to enhance the situation significantly.

Delivery of energy

The perfusion pump, whether the heart itself or some mechanical system, does no more than deliver mechanical energy to the circulation. This energy accrues in the form of blood flow and intravascular pressure. The rate at which the energy is delivered is defined as hydraulic power (19). The overall delivery of hydraulic power consists of both pulsatile and nonpulsatile, or mean, components, and these can be measured and used to characterize the flow delivery of the system. Separation of hydraulic power into its mean and pulsatile components is a complex matter involving phase shifting and Fourier analysis of the pressure and flow complexes. Total hydraulic power is the sum of mean power (product of mean pressure and flow) and pulsatile hydraulic power (derived from the sum of a number of pressure and flow harmonics). Computing total hydraulic power is not a simple matter, as it ideally requires that both blood flow and pressure be measured invasively at a common site. Although some investigators have managed to perform these measurements clinically (18,20), such an approach is not always possible or advisable in the operating theater, and the assessment of hydraulic power is generally reserved for research environments. However, extensive studies of this parameter have been carried out, and it has been established in both in vitro and animal models as a useful tool to assess the pulsatile blood flow-generating capabilities of several pumping systems. Studies by Gourlay (18) and Wright (19) demonstrate the limited pulsatile hydraulic power-generating capability of the standard pulsatile roller pump in comparison with that of a pulsatile ventricular system (Fig. 10.4).

FIG 10.4. Graph showing the pulsatile and nonpulsatile hydraulic power components of pulsatile blood flow generated by a roller and a ventricular pump. The ventricular pump groups are numbers 13 through 17 and are clearly associated with significantly more hydraulic power in the pulsatile domain. The mean pulsatile power remains similar between the two pump types. It can be seen that the roller pump is associated with very little hydraulic power in the pulsatile domain. These results are taken from an in vitro study of two pump systems.

For a number of reasons, particularly expense, pulsatile pumps of the ventricular type are not generally available for routine clinical use; therefore, the true clinical benefits of using this type of pumping mechanism during clinical CPB have not, as yet, been fully determined. However, the importance of energy transfer associated with pulsatile blood flow has been the focus of considerable research effort (21–23), and it is generally believed that it is of considerable importance. Shepard et al. (24) concluded that at the same mean blood pressure and flow rate, a pulsatile perfusion regime delivered up to 3.4 times as much energy to the circulation, and that this improved energy delivery might be responsible for maintaining normal peripheral blood flow distribution under pulsatile blood flow conditions. This factor may be responsible for many of the observed and perceived advantages associated with pulsatile CPB. Pulsatile flow is a complex issue that remains the focus of much study, and clearly the roller pump pulsatile flow system is not the ideal one from a fluid dynamic or hydraulic standpoint. Ventricular systems appear to offer more control of the pulse architecture and enhanced energy delivery, but use of such systems is confounded by expense and technologic complexity. In the meantime, the roller pump system is the only clinically available option, and its use, despite suboptimal output characteristics, has been associated with reduced patient mortality and morbidity 7,25,26. The reasons for this are clearly open to some degree of speculation, but many investigations indicate that energy transfer and pulse shape are involved.


Back to Quick Links

In a recent review of pulsatile perfusion, Hornick and Taylor (7) observed that progressive systemic arterial vasoconstriction is the inevitable consequence of nonpulsatile CPB. This vasoconstriction may ultimately lead to reduced visceral perfusion. At separation from CPB, afterload might be increased at a time when the left ventricle is already functionally compromised from the insult of the operative procedure, which predisposes to a low cardiac output syndrome and potential visceral organ damage. This sequence of events, familiar to most clinicians, had been described by Taylor et al. (27) as a vicious cycle that is generally controlled by a "cocktail" of vasodilator and inotropic drugs in the immediate postperfusion period. Taylor and colleagues also showed that the use of pulsatile perfusion during CPB appears to be associated with a reduction in the vasoconstriction commonly present at the termination of bypass. When combined with the improved myocardial contractility resulting from modern myocardial preservation techniques, this improved post-CPB hemodynamic state should hasten the return to normal myocardial performance. Common approaches to increased systemic vascular resistance in the period after bypass (e.g., sodium nitroprusside) do improve cardiac performance (28,29). Pulsatile flow during the CPB period can potentially offer a similar benefit by avoiding the development of increased systemic vascular resistance in the first place.

The mechanisms underlying the hemodynamic effects of pulsatile and nonpulsatile blood flow during CPB are complex and as yet not fully understood; however, a number of mechanisms have been proposed. Mechanisms purportedly involved in the development of postperfusion vasoconstriction consequent to nonpulsatile blood flow include activation of the renin–angiotensin system and the release of catecholamines, vasopressin, and local tissue vasoconstrictors.

Renin secretion increases under nonpulsatile blood flow conditions (30,31). This may ultimately increase plasma concentrations of angiotensin II, one of the most potent endogenous vasoconstrictors. Taylor et al. (32) associated the use of an angiotensin I- and angiotensin II-specific converting-enzyme inhibitor during CPB with significant and rapid reduction in systemic vascular resistance during CPB and an increase in cardiac index immediately following CPB. Plasma vasopressin levels are also elevated during nonpulsatile CPB (33); however, the importance of this finding to the development of postoperative vasoconstriction is unresolved. Similarly, many reports demonstrate that catecholamines are secreted during CPB and that this secretion may be attenuated by the use of pulsatile flow (34,35).

The hemodynamic response to nonpulsatile CPB and the mechanisms involved are varied and complex. However, a substantial body of evidence suggests that many of the undesirable postoperative sequelae of nonpulsatile flow can be prevented by using pulsatile blood flow during CPB. This may reduce the need for pharmacologic interventions in the critical post-CPB period. Clinical studies have demonstrated that pulsatile CPB reduces the requirement for inotropic agents and intraaortic balloon pump support after CPB 25,26,36. In addition, these studies demonstrate a reduction in hemodynamically related mortality in patients receiving pulsatile perfusion in comparison with those exposed to nonpulsatile CPB.


Back to Quick Links

The metabolic effects of pulsatile and nonpulsatile CPB can be characterized at the cellular level and at the vital organ level. The effects of pulsatile and nonpulsatile blood flow on most vital organs have been studied. The cellular metabolic effects of pulsatile flow continue to be studied, particularly as new metabolic assays become available.

Pulsatile blood flow and cell metabolism

Nonpulsatile CPB has been associated with the development of metabolic acidosis and reduced tissue oxygen consumption (37–40), whereas pulsatile flow during CPB has been associated with a higher rate of oxygen consumption and a reduction in the level of metabolic acidosis (41).

The mechanisms underlying the maintenance of relatively normal cell metabolism during pulsatile bypass have not been clearly ascertained. However, it has been postulated that the enhanced energy associated with a pulsatile regimen may be responsible for maintaining the patency of the microcirculation, thereby improving the delivery of nutrients (42,43).

Pulsatile blood flow and organ function

The effect of pulsatile CPB on organ function has been studied in a wide variety of experimental models and in a number of target organs. Many of the earliest animal studies are now being repeated as the technology for evaluating the effects of flow modality on organ function evolve into tools that can be utilized clinically, although many of the effects of pulsatile and nonpulsatile blood flow on major organ function during CPB have been well described.

Pulsatile blood flow and the kidney

As early as 1889, Hamel (1) determined that the pulse has an important influence on kidney function. His conclusions were confirmed by Gesell (2), who postulated that improved kidney function is the result of better gas exchange at the capillary level together with the maintenance of normal lymph flow. Kohlstaedt and Page (3), by performing a series of experiments on isolated kidneys, confirmed the importance of pulse pressure to kidney function and in particular to the secretion of renin. During these experiments, blood flow to the kidney was depulsed. They found that renin secretion is much higher in the group exposed to nonpulsatile blood flow. In all these experiments, the mean pressure was similar in both the pulsatile and nonpulsatile groups.

Mavroudis (36), in his excellent 1978 review, pointed out that some investigators contested these early findings. Ritter (44), Goodyer and Glenn (45), Oelert and Eufe (46), and Selkurt (47) all found that renal function is not affected by the presence or absence of a pulse in an isolated preparation, provided that the mean blood pressure is maintained. Mavroudis observed, however, that the blood flow architecture varied widely in the pulsatile flow groups involved in these experiments, an issue that is of considerable concern to this day. The pulse pressure profiles employed by the latter investigators were quite different from those used by Hooker in 1910 (48) in experiments in which normal pressure architecture was maintained in isolated tissue preparations, which leads to the conclusion that pulsatile blood flow enhances kidney function.

In common with most other investigators, Many et al. (30) confirmed the importance of pulsatility to kidney function in a series of experiments in which they found that animals undergoing nonpulsatile perfusion experience increased renin levels, along with attendant fluid and electrolyte imbalances. Many investigators considered improved distribution of blood flow to be the prime reason for the enhancement of renal function associated with pulsatile blood flow. Boucher et al. (49), using radioactively labeled microspheres, found renal blood flow to be preserved under pulsatile flow conditions. Nakayama et al. (50) had already reported that renal venous return is preserved under pulsatile blood flow conditions.

Earlier, Fintersbusch et al. (51) demonstrated a loss in normal renal artery configuration associated with nonpulsatile blood flow in a perfused dog model. Barger and Herd (52), using the same model, associated this finding with a shift in intrarenal blood flow that results in decreased sodium excretion. Mori et al. (53) discovered that after a period of hypothermic circulatory arrest, renal blood flow is substantially higher and the kidneys recover function more fully and rapidly in dogs exposed to pulsatile blood flow during the reperfusion period. Studies of renal function in open heart surgical patients by German et al. (54) confirmed that nonpulsatile blood flow is associated with a more rapid onset of renal hypoxia and acidosis than is pulsatile flow, despite an adequate systemic rate of blood flow and oxygen extraction. Paquet (55) had already described a similar phenomenon in an isolated porcine kidney model. Mukherjee (56) et al. demonstrated decreased tissue oxygen pressure in the renal medulla, together with increased local lactate levels and decreased oxygen uptake, when blood flow is nonpulsatile. Landymore et al. (15), working with essentially the same model, found that urine output is enhanced with pulsatile flow and once again that plasma renin levels are higher with nonpulsatile flow.

Clinical studies have tended to confirm the results of animal investigations. Williams et al. (57), in a study of infants undergoing open heart surgery during profound hypothermia, found urine output, used as an indicator of renal function, to be 100% greater in the pulsatile flow group than in the nonpulsatile group. Many clinical investigators have had difficulty in discerning any particular advantage directly attributable to the pulsatile nature of blood flow. This may be a consequence of many factors, including variations in clinical management during cardiac surgical procedures. Indeed, Louagie et al. (58) concluded that pulsatile blood flow is associated with reduced urine output in patients undergoing open heart surgery. Although these findings are at odds with those in most of the scientific literature, they may be an indication of how critical the clinical model is in establishing the beneficial effects of flow modality.

The effects of pulsatile flow on renal function may be most apparent and significant under the most challenging conditions. A number of clinical studies have focused on the importance of pulsatile flow in patients undergoing cardiac surgery who have renal insufficiency preoperatively. On the basis of a substantial clinical trial, Matsuda et al. (59) recommended that pulsatile blood flow be employed in this group of patients, having established that renal function is best preserved in these patients under pulsatile flow conditions.

The effect of pulsatility on kidneys being preserved for transplantation has also been studied. Belzer et al. (60) demonstrated that isolated kidneys subjected to nonpulsatile perfusion show a gradual but clearly defined rise in vascular resistance that is not present with pulsatile perfusion. This appears to confirm the development of rising vascular resistance with nonpulsatile flow, even in an isolated organ preparation. They further demonstrated that once transplanted, kidneys perfused in a pulsatile manner return to normal function more quickly than those perfused with nonpulsatile flow. This confirms the enhanced preservation associated, they surmised, with improved tissue perfusion. This is perhaps one area in which the frequently reported association between pulsatile blood flow and improved tissue perfusion may still be of considerable practical benefit.

Pulsatile blood flow and the brain

The brain, although protected by autoregulation, is still susceptible to injury during cardiac surgery. Taylor et al. (61) showed that the autoregulatory mechanisms can be modified by several factors, including temperature, pattern of blood flow, viscosity, oxygen and carbon dioxide tension, and various pharmacologic agents. Simpson (62) determined that certain brain tissues require substantially higher blood flow rates than others. He further suggested that these may be compromised by the breakdown of autoregulatory mechanisms and flow disruption associated with CPB. As early as 1969, Hill et al. (63) described significant neuropathologic manifestations associated with cardiac surgery. These findings included histologic evidence of focal brain lesions resulting from extracorporeal circulation. Sanderson et al. (64), employing a canine model, demonstrated that the diffuse brain cell damage associated with nonpulsatile perfusion can be prevented by pulsatile flow. Taylor et al. (61), using a similar canine perfusion model, demonstrated that the level of creatine phosphokinase BB isoenzyme in the cerebrospinal fluid, a sensitive marker of cerebral injury, is significantly higher in animals exposed to nonpulsatile blood flow than in those receiving pulsatile blood flow. These findings may be related to those of DePaepe et al. (65), who described significantly smaller cerebral capillary diameters with nonpulsatile than with pulsatile flow, which suggests the possibility of reduced cerebral blood flow under nonpulsatile conditions. Taylor and colleagues (66,67) found markedly different hypothalamic–pituitary axis responses to surgical stress with pulsatile and nonpulsatile blood flow conditions. They found that the anterior pituitary gland fails to respond to thyrotropin-releasing hormone under nonpulsatile conditions, in contrast to the normal profile exhibited during major surgery not utilizing CPB. The secretion of cortisol was also found to be significantly reduced under nonpulsatile flow conditions. In all cases, they noted a return of normal responses within 1 hour of cessation of nonpulsatile bypass. Taylor et al. (68) further demonstrated that the pituitary–adrenal axis responds normally under pulsatile perfusion conditions. Philbin et al. (69) demonstrated a similar response pattern of vasopressin secretion during nonpulsatile CPB.

All these findings demonstrate differences in the cerebral metabolic response to different CPB perfusion modalities. Several investigations have compared the adequacy of brain blood flow with pulsatile and nonpulsatile flow. Briceno and Runge (70) showed that pulsatile flow prevents the cerebral acidosis often observed during the early phase of nonpulsatile CPB. This may be a consequence of better preservation of regional cerebral blood flow. In comparison with nonpulsatile perfusion in patients, Kono et al. (71) found that pulsatile flow reduces cerebral vascular resistance by as much as 25%. This highly significant difference and the proposed improvement in regional blood flow distribution may be responsible for the reduced cerebral lactate production demonstrated by Mori et al. (72). They suggested that regional blood flow is maintained and anaerobic metabolism is suppressed with the application of pulsatile blood flow, particularly during the critical cooling and rewarming phases of the operative procedure. In a canine stroke model, Tranmer et al. (73) used computerized mapping to demonstrate that pulsatile perfusion better maintains cerebral blood flow in ischemic regions. In another canine CPB study, Onoe et al. (74) found that pulsatile blood flow preserves cerebral circulation, even during profound hypothermia, which suggests a cerebral protective effect of pulsatile blood flow.

A number of studies focusing on the brain contest the evidence supporting pulsatile blood flow. Invariably, these studies do not offer nonpulsatile blood flow as a superior mode of perfusion, but rather indicate that pulsatile blood flow provides little or no advantage over nonpulsatile flow. In particular, two studies showed no differences in neurologic or neuropsychological outcomes between pulsatile and nonpulsatile flow (75,76), but Murkin et al. (77) did find a reduction in cardiovascular morbidity in their pulsatile flow group. Hindman et al. (78), in a 1995 study focusing on cerebral blood flow and cerebral oxygen consumption in a rabbit model, found no difference between pulsatile and nonpulsatile flow regimes. In comparing roller pump-generated pulsatile and nonpulsatile blood flow, Chow et al. (79) found no differences in a number of cerebral metabolic markers. Despite this conflicting evidence, considerable evidence favors pulsatile blood flow for the maintenance of normal cerebral function, metabolism, and blood flow distribution during CPB.

Pulsatile blood flow and the liver and pancreas

Interest in the effects of CPB on pancreatic function was stimulated by sporadic findings pointing to increased plasma levels of amylase after nonpulsatile bypass. Feiner (80) reported a 16% incidence of ischemic pancreatitis in patients who had undergone CPB. Baca et al. (81), using a canine model, found that post-CPB pancreatic function was significantly better immediately and 48 hours after bypass in dogs exposed to pulsatile CPB than in those that underwent nonpulsatile CPB. Saggau et al. (82), monitoring insulin, glucose, glucagon, and growth hormone levels in human and animal studies, concluded that pulsatile CPB preserves pancreatic function better than nonpulsatile CPB. They found "normal function" of the pancreatic -cells in the pulsatile blood flow group and reduced function in the nonpulsatile group. Using a clinical model, Murray et al. (83) were able to demonstrate improved pancreatic function associated with a reduced incidence of elevated amylase levels in patients undergoing CPB with pulsatile flow. Mori et al. (72) concluded that pancreatic function is preserved in dogs perfused under both hypothermia and normothermia in the presence of pulsatile blood flow. In contrast, they found pancreatic function to be reduced in dogs exposed to a nonpulsatile regimen. Further evidence supporting the role of pulsatile flow in the maintenance of normal hepatic function was provided by Pappas et al. (84), who, employing serum glutamic oxaloacetic transaminase (SGOT) as a marker of hepatic injury, concluded that pulsatile blood flow preserves hepatic tissues and function. Mathie and colleagues (85) found similar results in a canine CPB model. This was echoed in the results of a series of clinical studies carried out by Chiu et al. (86), who found that hepatic function is preserved with pulsatile blood flow during CPB, as reflected in postoperative SGOT levels. They demonstrated that hepatic blood flow shows a typical vasoconstrictive response to nonpulsatile CPB, coupled with a reduction in hepatic oxygen consumption.

Pulsatile blood flow and the gut

Abdominal complications associated with CPB have become increasingly recognized as a significant component of operative morbidity and mortality. In one study, Gauss et al. (87) reported that 1.8% of 500 patients who underwent cardiac surgery had some type of abdominal complication. The mortality rate was 44% in this group. From his own retrospective study of 5,924 patients and two other reports, Baue (88) noted that postoperative gastrointestinal morbidity occurs in 0.29% to 2% of patients undergoing CPB, with an associated mortality of 23.5% to 44%. He suggested that the primary cause of these complications is mesenteric hypoperfusion leading to ischemia and subsequent gut-related morbidity or mortality. Bowles et al. (89) reported that CPB is associated with endotoxemia in some patients and that mesenteric ischemia may be an important contributor to this problem. Endotoxemia associated with CPB in children has been the focus of study for some years. Anderson and Baek (90) postulated that the high levels of endotoxin found in children after CPB may derive from ischemia-induced increases in gut permeability. The incidence of this undesirable post-CPB endotoxemia remains undetermined.

However, there have been reports of substantially elevated endotoxin levels in patients with no apparent preoperative infection (91). Two studies (92,93) highlighted the period immediately following removal of the aortic cross-clamp as a critical time in the development of endotoxemia. In a canine CPB model, Ohri et al. (94) demonstrated a disparity between mesenteric oxygen consumption and oxygen delivery during the rewarming phase of CPB. Further studies using the same model confirmed an increase in gut permeability during this period. Tao et al. (95) demonstrated in pigs that gut mucosa becomes ischemic during CPB, apparently from blood flow redistribution and shifting tissue oxygen demand.

Riddington et al. (96) confirmed these findings in the clinical model. They found that patients undergoing CPB exhibit increased gut mucosal ischemia and gut permeability, and endotoxin was detectable in the plasma of 42% of these patients. They further found that an elevated intestinal pH did not return to normal until the nonpulsatile flow regimen was terminated and the heart took over the circulation. Quigley et al. (97) established that perfusion pressure is one important factor in preventing endotoxemia. When the perfusion pressure was maintained in excess of 60 mm Hg throughout the perfusion period, there was no measurable endotoxemia in either the pulsatile or nonpulsatile groups. A number of preventive measures have been proposed to modify and reduce the impact of this potentially injurious course of events. Fiddian-Green (98) suggested that pulsatile blood flow results in improved blood flow to the gut, reducing mucosal ischemia and increasing oxygen delivery. He further suggested the application of preoperative gut lavage and parenteral antibiotics. Reilly and Bulkey (99) proposed that the vasoactive gut response to circulatory shock is mediated by activation of the renin–angiotensin system, which increases gut permeability. Pulsatile blood flow, therefore, may offer a simple but significant method for avoiding the potentially injurious results of impaired gut perfusion. There is little to suggest any harmful effects of pulsatile blood flow on the gut, although a 1998 case report (100) attributes a case of bowel necrosis to pulsatile blood flow dislodging plaque during CPB. The authors claim that this is a risk with pulsatile blood flow, but they provide no substantiation other than this perhaps coincidental complication.


Back to Quick Links

At the beginning of this chapter, we described the current state of clinical pulsatile blood flow as suboptimal in terms of pulse architecture. This derives from a combination of factors, including the interaction between the pump and other circuit components and the ever-changing hemodynamic status of patients undergoing cardiac surgery. The hemodynamics change considerably during CPB as a result of hemodilution, pharmacologic interventions, and changes in perfusion flow and temperature. All these factors may importantly influence the quality of the perfusion delivered by a pulsatile pump system. However, the quality of the system employed and its inherent pulse-generating capability are critical to the architectural integrity of pulsatile flow delivery. Many systems claiming to have some degree of pulsatile blood flow capability have evolved, some of which have been tested and characterized. However, the roller pump remains the most commonly used system for this purpose.


Back to Quick Links

A number of pumping systems are available to clinicians involved in CPB practice. The degree to which each of these systems is capable of delivering pulsatile blood flow varies greatly and has been the focus of much study.

Roller blood pump

The roller pump mechanism is simple and reliable, features that are extremely important for a clinical support mechanism. These pumps were used for circulatory support for some time before extracorporeal circulation and oxygenation were added. DeBakey (101) and Wesolowski (102) both reported the early use of a roller pump mechanism in total circulatory support in animals. The roller pump mechanism, which works on the principle of two or more diametrically opposed rollers "milking" a constrained piece of tubing (Fig. 10.5), offers the advantages of simplicity and predictability.

FIG 10.5. Typical roller pump head arrangement. The broken arrows show the direction of blood flow, and the solid arrows show the direction of rotation of the head components. In this case, two rollers are shown on a common axle. Each roller is capable of independent rotation in either a clockwise or counterclockwise direction. By rotating the roller mechanism over a constrained piece of tubing, a positive displacement of the perfusate contained within the tube is achieved by a "milking" process; each roller operates as a one-way valve when in contact with the tubing.

The occlusive nature of the mechanism induces negative pressures at the inlet side, rendering this type of mechanism appropriate for both arterial blood pumping systems and suction pumping (103). This inherent flexibility led to the mechanism being adopted as the pump of choice for heart–lung machines from a very early point in the evolution of CPB. Initially, commercially available roller pumps did not offer a pulsatile flow option. However, interest in the investigation of pulsatile flow as a modality for clinical CPB led to the development of modified systems. Ogata et al. (104) reported the modification of a roller pump to generate pulsatile blood flow. This study was followed by those of Nonoyama (105) and Nakayama et al. (106), who each employed essentially the same roller pump system with limited mechanical success but with reported clinical benefit. These systems were associated with poor mechanical reliability, possibly resulting from the effects of inertia of the heavy pump heads.

In the 1970s, a pulsatile roller pump for clinical use was made available for the first time by the Stockert company (Stockert Instruments, Munich, Germany). This innovation was made possible by the development of a pump head with low inertia made of aluminum and stainless steel and the incorporation of a "stepping" motor mechanism. The principal design requirement for this pumping system was that it be capable of generating some degree of pulsatile blood flow during CPB. The combination of a light pump head and the stepping motor permitted the pump head to be controlled precisely during the rapid acceleration and deceleration phases of the pulse cycle. The manufacturers also recognized the need for control of flow architecture and offered, in a limited way, some degree of user determination of output profile. While using the pump control module, it is possible to adjust the output frequency, pulse duration, and baseline flow rate in addition to total flow rate. A common fear among potential users was that the rapid acceleration and deceleration of the pump head during pulsatile flow might lead to an increase in hemolysis as the result of an increase in shear rate occurring under pulsatile blood flow conditions (107). However in vitro (108) and ex vivo studies (109,110) showed this not to be the case.

A number of clinical and laboratory experiments in which the Stockert pulsatile flow system was used have been carried out with considerable success during almost two decades. Demand for these systems has been such that several other manufacturers of blood pumps now offer a pulsatile blood flow option on roller-type blood pumps. However, Wright (19) has indicated that the roller pump may not be the most efficient mechanism for generating pulsatile flow. There is some question regarding whether the output generated by a roller pump in the pulsatile mode can truly be described as pulsatile in the physiologic sense. At best, the roller pump may be capable of generating only a "ripple" flow pattern. Subsequent studies (18) using a model of the systemic circulation and circuitry that mimics the typical perfusion add weight to this assertion. It has been established that the roller pump is not capable of matching the hydraulic power output of the human heart (20), but there is clear evidence that, even though the output of such systems is less than optimal, clinical benefits are derived from its use.

Ventricular blood pumps

These devices are probably the most physiologic method for generating pulsatile blood flow in that they operate in a similar manner to the ventricle of the heart. In simple terms, ventricular systems consist of compressible sac and two one-way valves permitting blood to flow into and out of the ventricle in only one direction. There are many configurations of ventricular mechanisms, one of which is shown in Fig. 10.6.

FIG 10.6. Ventricular pumping mechanism, mode of action. The ventricle is typically driven by either water or air. The driving medium fills a sac within the blood chamber, thereby displacing a volume of blood (A). The blood flow is directed by one-way inlet and outlet valves. During the fill cycle, the driving medium is drawn out of the sac and is replaced by blood in the blood chamber (B). The driving force for the displacement of blood is the displacement of the driving medium.

Typically, these systems are driven by either hydraulic or pneumatic means. Hydraulic systems generally employ a noncompressible fluid, such as distilled water, as the drive medium. The total blood flow generated by ventricular systems depends on frequency and stroke volume. Any alteration in stroke volume or frequency (e.g., when the system is synchronized with a patient's electrocardiogram) will affect the total output of the system. This problem can be solved by ensuring that total blood flow is maintained independently of all other control demands by some form of compensating software. This requires a substantial reserve capacity within the system to enable adjustments of output.

Ventricular systems have been associated with architecturally physiologic pulse pressure–flow profiles during experimental procedures. Pumps of this type have been employed in clinical practice with some success (111). However, widespread use of such systems has been hampered by their considerable cost, in terms of both mechanical and disposable components. It is possible that as the interest in physiologic pulsatile blood flow increases (112), interest in ventricular pumps will also increase and more systems will find their way into clinical practice.

Compression plate pumps

The principle of the compression plate pumping system is simple and not dissimilar to that of the ventricular models. Like ventricular pumps, compression plate pumps can produce only pulsatile flow. Simply put, a length of tubing of known diameter is placed on a rigid back plate and compressed by a moving plate that descends for a preselected stroke length, thereby ejecting a volume of perfusate from the tube (Fig. 10.7). The direction of blood flow is ensured by valves positioned at the inlet and outlet of the ventricle or sac.

FIG 10.7. Typical compression plate mechanism. This diagram shows both the fill and ejection cycles. The valves that ensure flow direction can be either pressure-driven or cam-driven.

The pulse rise time can be controlled by the rate of compression of the tube, and the flow rate by altering either the frequency or the length of travel of the compression plate. These systems offer significant control of pulse wave architecture. The filling aspect of the pumping cycle, like that of the ventricular mechanisms, can be either passive or active. Passive filling systems depend on a head of pressure at the inlet side of the device to fill the ventricle after the ejection cycle has been completed. Such systems employ ventricles with little or no elastic memory; therefore, the filling is entirely passive, limiting the application of such pumps. Active filling systems do not entirely depend on a head of pressure at the inlet side to effect a filling cycle. In these systems, the material employed in constructing the ventricle has an elastic memory that effects filling by generating a negative pressure. This condition augments the positive pressure at the inlet side of the device. A similar outcome can be achieved by connecting the ventricle to both the constraining and compression plates, whereby the return of the compression plate to the neutral position will effect the filling of the ventricle by generating a negative pressure within the ventricle. Active filling systems are the only compression plate systems that can be considered for routine CPB applications as stand-alone systems. Passive filling systems commonly require an additional or "priming" pump if the system is to operate without being affected by position within the perfusion circuitry (Fig. 10.8). The passive filling system has attractive attributes; notably, negative pressures are not generated within the system, and it can pump only the volume of blood supplied by the inlet conditions, matching output with inlet flow. Active filling ventricles powered by the elastic memory of the ventricular material have been the basis of several systems in the past (18,113) and have proved adequate for CPB, although their flow capabilities are somewhat limiting. The valves employed in ventricular pumping systems have produced significant problems. Internal valves have proved prohibitively expensive through the years, which has led to many promising systems being shelved on economic grounds before clinical use. One possible solution to this issue is to position the valves on the outside of the tubing or ventricle. This arrangement, used in the University of Texas preload-responsive pump (114), has proved to be both economical and efficient. The use of mechanical, externally positioned valves may require another compromise. The tubing employed in the pump head must be soft enough to allow the passively operated valves to compress it to the point of closure. This removes to a great extent the strong elastic memory required to power an active fill cycle, and the system reverts to being a passive filling system.

FIG 10.8. Circuit diagram of the feeder reservoir system for employing a passive-filling ventricle pump system. Venous blood flows into the venous reservoir in the normal manner and is pumped from there by a roller pump (a) through the membrane oxygenator. Rather than passing directly to the patient, the blood is held in a second reservoir (arterial reservoir). The arterial reservoir is the most elevated aspect of the perfusion circuit and fills the pump head (b) by gravity flow. Such a system permits passive-filling ventricles to be used for routine cardiopulmonary bypass.

Ventricular pumps are pulsatile and capable of generating pulsatile flow of physiologic proportions. As research into the beneficial effects of pulsatile blood flow continues, it is conceivable that interest in ventricular pumps will expand with the realization that this pump is uniquely capable of generating nearly physiologic pulse architecture.

Centrifugal blood pumps

Centrifugal blood pumps are a relatively recent innovation in pumping technology for routine CPB applications. Leschinski et al. (115) described them as pumps in which the working elements rotate a drive shaft; they can be axial, nutational, or rotary in nature. The drive for these pumps is invariably provided by an electric motor coupled to the drive shaft. The coupling of the motor to the drive shaft offers a complex design problem. The drive components must be sealed so that sterility of the blood path can be ensured. The accepted solution to this problem is to couple the drive motor to the drive shaft magnetically (Fig. 10.9). Pennington (116) determined that centrifugal pumps are afterload-dependent, as the output is related to pump speed and pressure gradient. It is, therefore, necessary to measure the pump output with a flow probe during use.

FIG 10.9. Schematic diagram of a centrifugal pump head (Bio- Medicus pump, Medtronic, Minneapolis, Minnesota) showing the rotating cone arrangement, inlet and outlet orientation, and drive magnets. The rotating components are held in place by a bearing assembly.

There are as many pump head designs as there are pump types, but in general, designs fall into one of three categories—vaned pump mechanisms, concentric disk pump mechanisms, or combinations of both.

Centrifugal pumps have been used for a considerable time with success. These pumps may be susceptible to the generation of stagnant zones and high vortex areas, which can result in hemolysis and the formation of thrombi on the working parts of the pump mechanism. Centrifugal pumps have been extensively studied in the long-term assist arena, where they have been found to offer significant advantages over roller pumps (117). Increasingly, however, centrifugal pumps are being employed for short-term, routine CPB procedures. Interest in the use of centrifugal pumps for routine clinical practice has been fired by the safety features associated with these systems, particularly a reduced risk for air embolism. Centrifugal pumps, particularly those of the disk type, have been associated with reduced hemolysis in clinical use. Berki et al. (118) noted a reduction in operative hemolysis with the disk centrifugal pump, and also an improvement in postoperative hemostasis and preservation of the platelet count, in comparison with the roller pump. They further noted a reduced microembolic load with the centrifugal pump, which they thought was probably a consequence of the elimination of tubing spallation traditionally associated with the roller pump. These findings confirmed those of Mandl (119) and Noon et al. (120), who noted that centrifugal pumps, particularly constrained vortex pumps, are associated with a lower particulate and gaseous embolic load than are roller pumps. Improved blood handling with the centrifugal pump during routine clinical use was confirmed by Matsukura (121), who noted improved platelet preservation in comparison with the roller mechanism. Maas et al. (122) confirmed this, observing reduced hemolysis and improved postoperative platelet counts.

Because fluid displacement in such systems depends on centrifugal impulsion, which requires a very high rotational speed to produce blood velocity, a pulsatile output is difficult to achieve. To do so, the impeller would have to accelerate very rapidly to an extremely high rotational speed. Attempts to achieve this have been made with varying degrees of success; however, pulsatile blood flow with centrifugal systems has been slow to develop. Recent studies by Nishida et al. (123) with the Terumo Capiox (Terumo Corporation, Kanagawa, Japan) centrifugal pump in the pulsatile mode confirmed that physiologic arterial wave profiles are difficult to achieve clinically. Use of a rapidly accelerating rotor head achieved a radial artery pressure of only 10 mm Hg, which is no greater than the ripple pressure amplitude seen with roller pump mechanisms in a nonpulsatile mode. Gobel et al. (124) described a new centrifugal pump that they determined to be capable of generating pulsatile flow of physiologic proportions. In a comparative study, they further noted that pumps without vanes are incapable of generating pulsatile flow at all because energy is transferred to the perfusate by friction only in such systems. Vaned pumps could produce varying degrees of pulsatility during testing but tended to "decouple" under the increased strain, potentially resulting in unacceptably low blood flows. Komada et al. (125) found that pulsatile blood flow generated by a centrifugal blood pump is not associated with the normally observed reduced peripheral vascular resistance and that the pressure profiles observed are damped. They found, however, that the centrifugal system does offer other advantages normally associated with roller pump-generated pulsatile blood flow, such as reduced angiotensin levels and a reduction in the need for postoperative inotropic agents. Ninomiya and colleagues (126) recognized the limitations of the centrifugal pump in generating pulsatile blood flow and employed a centrifugal pump in conjunction with a pulsatile assist device to generate pulsatile flow. They concluded that under these complex conditions, the centrifugal pump can produce sufficient pulsatile blood flow. In comparison with a nonpulsatile centrifugal pump, Dreissen et al. (127) observed reduced complement activation with a pulsatile centrifugal pump. They further found a higher incidence of postoperative respiratory tract infection in the nonpulsatile group, and the classic whole body inflammatory response (i.e., complement activation) appeared to be reduced when a pulse was added to centrifugal blood flow.

Centrifugal pump-generated pulsatile flow has not been well accepted, probably because generating physiologic pulsatility is not possible with currently available mechanisms. This belief appears to have been borne out in clinical practice. However, this area of research continues to attract much attention.

Pulsatile assist device

The pulsatile assist device is an intermittent occlusive device that employs an intraaortic balloon pump apparatus to produce pulsatile blood flow in the arterial line of the CPB circuit (Fig. 10.10). The pulse is generated by occluding the arterial line of the circuit under flowing conditions, thereby creating a large pressure and volume delay within the arterial side of the circuit. On deflation of the balloon in the arterial line, the pressure and blood volume are released into the aorta of the patient as a "pulse." Bregman and colleagues (128) described using this system during CPB to produce pulsatile blood flow at the beginning of as well as before and after CPB to produce arterial counterpulsation. The control module of the balloon pump permits control of the balloon inflation and deflation times, together with the rate of inflation and volume of gas infused into the balloon. The level of control offered by such a system is significant, and clinical studies involving this technology have been promising. Maddoux et al. (129) and Philbin et al. (130) were both impressed by the versatility of the system. However, there have been some concerns regarding its use, particularly regarding the possibility of balloon rupture (131) and generation of hemolysis (132). The pulsatile assist device chamber is positioned in the arterial line of the circuit and, in the event of a balloon rupture, the contents of the balloon can be discharged into the arterial line and from there into the aorta, with potentially disastrous consequences. However, the most convincing argument against this technology from the clinical standpoint is that it adds complexity to a traditional roller pump mechanism. Despite excellent clinical results with these devices, the increase in circuit complexity together with the fear of balloon rupture have led to its relative disuse.

FIG 10.0. Diagram of mode of operation of pulsatile assist device (PAD). Blood from the roller pump passes through the device when the balloon is deflated (A). When the balloon is inflated, the blood flow is prevented from passing through the device (B). This process is repeated at the desired pulse frequency, and a considerable degree of pulsatility can be achieved. Timing of the mechanism is critical, and extremely high circuit arterial "line" pressures are normal. RA, right atrium.

Compatibility of pulsatile blood flow with perfusion circuit components

There are a number of ways in which perfusion circuit components other than the blood pump can influence the quality of pulsatile blood flow delivery to patients undergoing CPB. Of particular interest, because of their apparent involvement in the beneficial effects of pulsatile blood flow, are the energy adsorption characteristics of devices positioned in the arterial line of the perfusion circuit. The three devices commonly positioned between the arterial blood pump and the patient are the membrane oxygenator, arterial filter, and aortic cannula. In addition to these devices, the tubing connecting the patient to the heart–lung machine has energy-adsorbing properties that can affect energy delivery. Wright (133) noted that the tubing should be as short and rigid as possible with the minimum number of connectors if energy delivery is to be maximized. When tubing is long and unrestrained between pump and patient, it can often be seen swinging under pulsatile flow conditions, with the energy utilized in producing this pendular motion originating at the arterial pump. Highly flexible tubing can be seen to expand and contract under pulsatile flow conditions. Some of the pulsatile energy will be lost as a result of these processes, which then will impair the ability of the system to deliver physiologic pulsatility to the patient. Most recently, Inzoli et al. (134) found that the arterial tubing has a much greater overall damping effect on pulsatile structure and energy delivery than does the membrane oxygenator. Studies of the effects of tubing type and configuration on pulsatile flow delivery are few; however, many more studies have concentrated on the effects of membrane oxygenators on this important factor. Two studies (17,133) showed that the inclusion of a membrane oxygenator between the pump and the patient reduces hydraulic power delivery and alters pressure architecture. Specifically, Gourlay (18) showed that the membrane reduces hydraulic power only in the pulsatile domain when measured in a model of the human systemic circulation (Fig. 10.11). However, because the proportional contribution to total hydraulic power arising from the pulsatile domain is small with a roller pump system, the resultant reduction in hydraulic power may not be clinically significant (135). In the same series of experiments, the effect of an arterial line filter on energy delivery was evaluated, and no significant effect was found.

FIG 10.1. The effects of including a membrane oxygenator on total hydraulic power. Group S has no oxygenator in line, group ST has an oxygenator positioned in the venous line, and group SA has the membrane in the arterial line. Clearly, energy absorption is greatest when the oxygenator is positioned in the arterial line. It is interesting to note that power in the mean domain is unaffected.

The third circuit component between the pump and the patient, the aortic cannula, is critical to the transmission of pulsatile blood flow. Runge et al. (136) stated that it may be impossible for truly physiologic pulsatile flow to be achieved with current aortic cannulas. They further suggested that the excellent results achieved with pulsatile flow in animal studies cannot be repeated in the clinical environment without radically altering cannulation techniques. Clearly, the optimally sized aortic cannula for pulsatile blood flow transmission has an internal diameter that is the same as that of the arterial infusion tubing. Not surprisingly, it might be difficult to find a cardiac surgeon willing to insert such a large "pipe" into a patient's aorta.

Special attention must be paid to sheer stress and velocity associated with catheter size and tip geometry under pulsatile blood flow conditions. Kayser (137) stated that the high velocity of blood passing through aortic cannulas under pulsatile blood flow conditions may increase hemolysis. Wright (107) determined that the degree to which blood undergoes hemolysis in an arterial cannula depends on several factors—shear rate, velocity profile, and dimensions of the cannula. Most importantly, however, this depends on whether the flow within the cannula becomes turbulent or not. Wright determined by experimentation that hemolysis can be expected to occur when a shear rate of 200 Nm–2 is exceeded (107). Taylor et al. (67) measured the shear stress under pulsatile blood flow conditions at varying flow rates with two commonly used aortic cannulas and showed that not all cannulas are compatible with roller pump-generated pulsatile blood flow. One of these cannula types, although both were largely similar in design, was associated with high levels of hemolysis under the test conditions.

Aortic cannulas are generally considered benign in relation to perfusion. This may well be the case, but only if attention is paid to velocity profile, shear rate, and pressure drop during use. Runge et al. (136) summed up the importance of aortic cannulas to pulsatile blood flow by saying simply that cannula size should be maximized to optimize transmission of pulsatility. In effect, the only approach to aortic cannulation that can deliver physiologic pulsatile architecture is one that preserves the diameter of the aortic return line all the way into the patient's aorta. Clearly, without a radical change in cannulation technique and approach, this will not be achievable.


Back to Quick Links

Generating pulsatile flow with roller pump technology is relatively simple. However, there are several safety considerations that must be carried out. In a previous section, we described interactions that take place between the pump and other circuit components that adsorb hydraulic energy. Other interactions between these devices can, under some circumstances, induce considerably more hazardous consequences. In particular, the interaction between the pulsatile flow modality and the membrane oxygenator can generate gaseous microemboli (17,138).

Interaction between pulsatile flow and membrane oxygenators

Membrane oxygenators are generally microporous in nature and rely on a positive pressure gradient between the blood and gas phases to maintain integrity. It is possible for oxygenating gas to enter the blood phase in the form of microbubbles if the pressure gradient is, even momentarily, positive in favor of the gas phase (138). Pearson (139) described this phenomenon and the presence of microbubbles in the arterial tubing of the perfusion circuit in association with large transient negative pressure "spikes" generated by pulsatile roller pumps. In our own series of experiments (18), we found a relationship between the flow configuration and the number of microbubbles present in the arterial line of the circuit (Fig. 10.12.).

FIG 10.2. Graph showing free circulating gaseous microembolic activity (GMA) in an in vitro circuit running at the clinical range of blood flow rates with normal clinical pulsatile control configuration (Stockert roller pump mechanism, Stockert, Munich, Germany). In this case, the pump was used with and without an Affinity membrane oxygenator (Avecor Cardiovascular Inc., Minneapolis, MN) in the arterial line. It can be seen that inclusion of the membrane is associated with an increase in activity across the flow range, but a particularly high response can be seen at the higher levels of blood flow.

The gaseous microembolic activity associated with the combination of pulsatile blood flow and a membrane oxygenator is significant, but it can be dealt with quite adequately by including a screen filter downstream from the membrane oxygenator. The generation of microbubbles associated with this combination of technologies must be considered when pulsatile flow is used in the clinical setting; in particular, the fact that the number of microbubbles present increases with increasing pulsatility and blood flow (18) must be borne in mind during any decision regarding which pulsatile flow configuration to use.

Although increased microbubble production and energy adsorption are disadvantages to combining pulsatile flow technology with membrane oxygenators, there may be advantages. Pulsatile blood flow can enhance gas exchange within the membrane by generating secondary flows at the membrane–blood interface, and by breaking down boundary layers (see Chap. 4). This may not be immediately apparent in routine clinical practice, during which the membrane oxygenator's reserve capacity is not challenged. However, if the device is stressed by an increase in oxygen demand, pulsatile flow within the membrane compartment can assist in meeting this challenge (140).

The effect of pulsatile blood flow on the other main function of the modern blood oxygenator, heat exchange, has also been the focus of study. Sheperd (141) determined that pulsatile blood flow enhances the performance of the heat exchanger of one commercially available membrane oxygenator. He reasoned that under pulsatile blood flow conditions, the boundary layer effect within the heat exchanger is broken down, leading to improved heat transmission. This factor is potentially important because the reduced systemic vascular resistance encountered during pulsatile CPB may place greater demand on the heat exchanger, particularly during the rewarming phase of the procedure.

Overall, the combination of pulsatile blood flow and membrane oxygenation has advantages and disadvantages. It is particularly important that all these be understood fully before the pulsatile flow route is undertaken. The combination is safe and effective provided that adequate safety measures are taken to ensure that the optimum combination of devices and control configuration are employed.


Back to Quick Links

The clinical use of pulsatile flow during CPB has a considerable history; however, evolving technologies require that the issue be continually readdressed. Pulsatile CPB remains a controversial issue (7). Studies of pulsatile blood flow during CPB have focused on a broad spectrum of factors, ranging from cellular metabolism to organ function and including both hemodynamic and metabolic responses. However, most of this work, particularly the clinical studies, has been carried out without a truly physiologic pulsatile blood flow having been achieved. This results from the lack of a device capable of delivering such a flow pattern during routine CPB. Despite evidence in its favor, pulsatile CPB still remains controversial; however, it is conceivable that production of a pulsatile blood pump capable of generating physiologic pulsatile flow during routine CPB will resolve the controversy. This leaves considerable opportunity for future investigation. Despite the safety issues raised in the latter half of this chapter, it seems logical that a flow modality closely resembling the natural state will deliver the best result. It remains to be seen whether this can be safely achieved within the size constraints imposed by arterial cannulas. With advancing technologic development, it may not be long before this controversy is resolved.


Back to Quick Links

  • It has long been recognized that pulsatile flow is of physiologic importance and that reproducing pulsatility during cardiopulmonary bypass may be beneficial, but the safe delivery of pulsatile blood flow during cardiopulmonary bypass presents some formidable technologic obstacles.

  • Simply generating a normal pulse pressure may not be sufficient to sustain normal circulatory physiology during cardiopulmonary bypass; rather, reproducing a normal pulsatile "architecture" should be the goal.

  • Nonpulsatile cardiopulmonary bypass induces a progressive increase in systemic vascular resistance, potentially compromising organ perfusion during cardiopulmonary bypass and setting the stage for increased myocardial oxygen demand after bypass. Mechanisms involved in the increase of systemic vascular resistance include an increase in circulating catecholamines and vasopressin and activation of the renin–angiotensin system.

  • Investigations in animals have generally found that urine output and renal blood flow are better preserved with pulsatile than with nonpulsatile bypass, yet clinical studies most often have not noted improved renal outcomes with pulsatile bypass, perhaps because the pulsatile flow has not been sufficiently physiologic.

  • Pulsatile cardiopulmonary bypass has been shown experimentally to reduce both cerebral acidosis and markers of brain injury and neurohumoral dysfunction, but again, clinical studies have not demonstrated improved outcomes.

  • Investigations suggest that hepatic, pancreatic, and gut function are better preserved during pulsatile cardiopulmonary bypass, quite possibly because of a reduction in the development of mucosal ischemia, which can induce endotoxemia. The clinical implications of these findings remain unclear.

  • Several different types of systems can be used to deliver pulsatile cardiopulmonary bypass. These are briefly highlighted as follows:

    • Rhythmically varying the speed of a traditional roller pump head is the most widely available mechanism, but the architectural form and hydraulic power generated by these pumps have been suboptimal. Nevertheless, studies have shown some improvement in process variables such as systemic vascular resistance and stress hormone release, although improved clinical outcomes remain elusive.

    • Ventricular pumps use a compressible sac and one-way valves to generate pulsatility that is more architecturally physiologic than that obtained with other devices. Clinical trials show promise, but high cost remains a limiting factor.

    • Compression plate pumps and pulsatile assist devices compress the extracorporeal arterial flow between the oxygenator and the arterial inflow cannula to produce pulsatility. For various reasons, these devices have not become widely used.

    • Pulsatility can be produced with centrifugal pumps under ideal conditions, but their dependence on afterload renders them unreliable for this purpose.

  • Obstacles to transmitting pulsatile flow into the patient include distensibility of the arterial tubing and resistance and damping imposed by membrane oxygenators, arterial filters, and arterial infusion cannulas that intervene between the source of pulsatility and the patient's arterial circulation. The small size of the aortic cannula in relation to the arterial tubing creates the extracorporeal equivalent of severe aortic stenosis, resulting in loss of pulsatility and possible hemolysis when sufficient energy is transmitted across the cannula to induce architecturally physiologic intraarterial pulsatility.


Back to Quick Links

    1. Hamel G. Dei Bedeutung des pulses fur den blutstrom. Z Biol NSF 1889;474–497.

    2. Gesell RA. On relation of pulse pressure to renal function. Am J Physiol 1913;32:70.

    3. Kohlstaedt LA, Page IH. The liberation of renin by perfusion of kidneys following reduction of pulse pressure. J Exp Med 1940;72:201–211.

    4. McMaster PD, Parsons RJ. The effect of the pulse on the spread of substances through tissues. J Exp Med 1938;68:377–400.

    5. Gibbon JH Jr. Application of a mechanical heart and lung apparatus to cardiac surgery. Minn Med 1954;37:171–185.

    6. DeBakey ME. A simple continuous flow blood instrument. New Orleans Med Surg J 1934;87:387–389.

    7. Hornick P, Taylor KM. Pulsatile and nonpulsatile perfusion: the continuing controversy. J Cardiothorac Vasc Anesth 1997;11:310–315.

    8. Gourlay T, Gibbons M, Taylor KM. Pulsatile flow compatibility of a group of membrane oxygenators. Perfusion 1987;2:115–126.

    9. Hutcheson IR, Griffith TM. Release of endothelium-derived relaxing factor is modulated both by frequency and amplitude of pulsatile flow. Am J Physiol 1991;261:H257–H262.

    10. Runge TM, Cohen DJ, Hantler CB, et al. Achievement of physiologic pulsatile flow on cardiopulmonary bypass with a 24 French cannula. J ASAIO 1992;38:M726–M729.

    11. Taylor KM, Bain WH, Maxted KJ, et al. Comparative studies of pulsatile and nonpulsatile flow during cardiopulmonary bypass. I. Pulsatile system employed and its hematologic effects. J Thorac Cardiovasc Surg 1978;75:569–573.

    12. Wright G, Furness A. What is pulsatile flow? Ann Thorac Surg 1985;39:401–402.

    13. Gourlay T, Taylor KM, Russell M, et al. Comparative retrospective study of pulsatile and non-pulsatile flow in 380 consecutive cardiac patients. Proceedings of the First World Congress on Extracorporeal Circulation , Brighton, England, 1983.

    14. Taylor KM, Bain WH, Davidson KG, et al. Comparative clinical study of pulsatile and non-pulsatile perfusion in 350 consecutive patients. Thorax 1982;37:324–330.

    15. Landymore RW, Murphy DA, Kinley CE, et al. Does pulsatile flow influence the incidence of postoperative hypertension? Ann Thorac Surg 1979;28:261–268.

    16. Kono M, Orita H, Shimanuki T, et al. A clinical study of cerebral perfusion during pulsatile and non-pulsatile cardiopulmonary bypass. Nippon Geka Gakkai Zasshi 1990;91:1016–1022.

    17. Gourlay T, Taylor KM. Pulsatile flow and membrane oxygentors. Perfusion 1994;9:189–196.

    18. Gourlay T. Controlled pulsatile architecture in cardiopulmonary bypass: in vitro and clinical studies [Thesis]. Strathclyde University, 1997.

    19. Wright G. The hydraulic power outputs of pulsatile and nonpulsatile cardiopulmonary bypass pumps. Perfusion 1988;3:251–262.

    20. Wright G, Ping J, Campbell C, et al. Computation of haemodynamic power and input impedance in the ascending aorta of patients undergoing open heart surgery. Cardiovasc Res 1988;22:179–184.

    21. Wilcox BR, Coulter NA, Peters RM, et al. Power dissipation in the systemic and pulmonary vascualture of dogs. Surgery 1967;62:25.

    22. Westerhof N, Elzinga G, Sipkema P. An artificial arterial system for pumping hearts. J Appl Physiol 1971;31:776–781.

    23. Wright G. Hemodynamic analysis could resolve the pulsatile blood flow controversy. Ann Thorac Surg 1994;58:1199–1204.

    24. Shepard RB, Simpson DS, Sharp JF. Energy equivalent pressure. Arch Surg 1966;93:730–740.

    25. Gourlay T, Taylor KM, Russell M, et al. Comparative retrospective study of pulsatile and non-pulsatile flow in 380 consecutive cardiac patients. Proceedings of the First World Congress on Extracorporeal Circulation , Brighton, England, 1983.

    26. Taylor KM, Bain WH, Davidson KG, et al. Comparative clinical study of pulsatile and non-pulsatile perfusion in 350 consecutive patients. Thorax 1982;37:324–330.

    27. Taylor KM, Bain WH, Morton JJ. The role of angiotensin II in the development of peripheral vasoconstriction during open heart surgery. Am Heart J 1980;100:935–937.

    28. Stinson EB, Holloway EL, Derby GC, et al. Control of myocardial performance early after open heart operations by vasodilator treatment. J Thorac Cardiovasc Surg 1977;73:523–530.

    29. Taylor KM, Bain WH, Russell M, et al. Peripheral vascular resistance and angiotensin II levels during pulsatile and non-pulsatile cardiopulmonary bypass. Thorax 1979;34:594–598.

    30. Many M, Soroff HS, Birtwell WC, et al. The physiologic role of pulsatile and nonpulsatile blood flow. II. Effects on renal function. Arch Surg 1968;95:762–767.

    31. Kohlstaedt LA, Page IH. The liberation of renin by perfusion of kidneys following reduction of pulse pressure. J Exp Med 1940;72:201–211.

    32. Taylor KM, Casals J, Morton JJ. The haemodynamic effect of angiotensin blockage after cardiopulmonary bypass. Br Heart J 1979;41:380–385.

    33. Levine FH, Philbin DM, Kono K, et al. Plasma vasopressin levels and urinary sodium excretion during cardiopulmonary bypass with and without pulsatile flow. Ann Thorac Surg 1981;32:63–67.

    34. Minami K, Korner M, Vyska K. Effects of pulsatile perfusion on plasma catecholamine levels and haemodynamics during and after cardiac operations with cardiopulmonary bypass. J Thorac Cardiovasc Surg 1990;99:82–91.

    35. Taylor KM. Pulsatile perfusion. In: Taylor KM, ed. Cardiopulmonary bypass—principles and management . London: Chapman and Hall, 1986.

    36. Mavroudis C. To pulse or not to pulse. Ann Thorac Surg 1978:25:259–271.

    37. Jacobs LA, Klopp EH, Seamone W, et al. Improved organ function during cardiac bypass with a roller pump modified to deliver pulsatile flow. J Thorac Cardiovasc Surg 1969;58:703–712.

    38. Dunn J, Kirsh MM, Harness J, et al. Hemodynamic, metabolic and hematologic effects of pulsatile cardiopulmonary bypass. J Thorac Cardiovasc Surg 1974;68:138–147.

    39. Steed D, Follette D, Foglia R. Effects of pulsatile and nonpulsatile flow on subendocardial perfusion during cardiopulmonary bypass. Ann Thorac Surg 1978;26:133–141.

    40. Shepard RB, Kirklin JW. Relation of pulsatile flow to oxygen consumption and other variables during cardiopulmonary bypass. J Thorac Cardiovasc Surg 1969;58:694–702.

    41. Ogata T, Ida Y, Takeda J, et al. A comparative study of the effectiveness of pulsatile and nonpulsatile blood flow in extracorporeal circulation. Arch Jpn Chir 1960;29:59–65.

    42. Prior FGR, Moorcroft V, Gourlay T, et al. Further testing of pulse reverse osmosis. A new theory of the maintenance and control of blood pressure. Int J Artif Organs 1995;18:469–473.

    43. Prior FGR, Moorecroft V, Gourlay T, et al. The therapeutic significance of pulse reverse osmosis. Int J Artif Organs 1996;19:487–492.

    44. Ritter ER. Pressure-flow relations in kidney: alleged effects of pulse pressure. Am J Physiol 1952;168:480–489.

    45. Goodyer AVN, Glenn WL. Relation of arterial pulse pressure to renal function. Am J Physiol 1951;167:689–697.

    46. Oelert H, Eufe R. Dog kidney function during total left heart bypass with pulsatile and non-pulsatile flow. J Cardiovasc Surg (Torino) 1974;15:674–678.

    47. Selkurt EE. Effects of pulse pressure and mean arterial pressure modification on renal haemodynamics and electrolyte and water excretion. Circulation 1951;4:541–546.

    48. Hooker DR. A study of the isolated kidney: the influence of pulse pressure upon renal function. Am J Physiol 1910;27:24–44.

    49. Boucher JK, Rudy LW, Edmunds LH. Organ blood flow during pulsatile cardiopulmonary bypass. J Appl Physiol 1974;36:86–90.

    50. Nakayama K, Tamiya T, Yamamoto K, et al. High-amplitude pulsatile pump in extracorporeal circulation with particular reference to hemodynamics. Surgery 1963;54:798–805.

    51. Fintersbusch W, Long DM, Sellers RD. Renal arteriography during extracorporeal circulation in dogs with preliminary report upon effects of low-molecular-weight dextran. J Thorac Cardiovasc Surg 1961;41:252–260.

    52. Barger AC, Herd JA. Study of renal circulation in the unanaesthetized dog with inert gases (Proceedings of the Third International Congress of Nephrology). Nephrology 1966;1:174.

    53. Mori A, Watanabe K, Onoe M, et al. Regional blood flow in the liver, pancreas and kidney during pulsatile and nonpulsatile perfusion under profound hypothermia. Jpn Circ J 1988:52:219–227.

    54. German JC, Chalmers GS, Hirai J, et al. Comparison of nonpulsatile and pulsatile extracorporeal circulation on renal tissue perfusion. Chest 1972;61:65–69.

    55. Paquet KJ. Hemodynamic studies on normothermic perfusion of the isolated pig kidney with pulsatile and nonpulsatile flows. J Cardiovasc Surg (Torino) 1969;10:45–53.

    56. Mukherjee ND, Beran AV, Hirai J. In vivo determination of renal tissue oxygenation during pulsatile and non-pulsatile left heart bypass. Ann Thorac Surg 1973;15:334–363.

    57. Williams GD, Seifen AB, Lawson NW, et al. Pulsatile perfusion versus conventional high-flow non-pulsatile perfusion for rapid core cooling and rewarming of infants for circulatory arrest in cardiac operation. J Thorac Cardiovasc Surg 1979;78:667–677.

    58. Louagie YA, Gonzalez M, Collard E, et al. Does flow character of cardiopulmonary bypass make a difference? J Thorac Cardiovasc Surg 1992;104:1628–1638.

    59. Matsuda H, Hirose H, Nakano S, et al. Results of open heart surgery in patients with impaired renal function as creatinine clearance below 30 mL/min. The effects of pulsatile perfusion. J Cardiovasc Surg (Torino) 1986;27:595–599.

    60. Belzer FO, Ashby BS, Huang JS. Etiology of rising perfusion pressure in isolated organ perfusion. Ann Surg 1968;168:382–391.

    61. Taylor KM, Devlin BJ, Mittra S, et al. Assessment of cerebral damage during open heart surgery. A new experimental model. Scand J Thorac Cardiovasc Surg 1980;14:197–203.

    62. Simpson JC. Cerebral perfusion during cardiac surgery using cardiac bypass,. In: Longmore E, ed. Towards safer cardiac surgery. MTP, Lancaster, 1981.

    63. Hill JD, Aguilar KJ, Baranco A, et al. Neuropathological manifestations of cardiac surgery. Ann Thorac Surg 1969;7:409–419.

    64. Sanderson JM, Wright G, Sims FW. Brain damage in dogs immediately following pulsatile and non-pulsatile blood flows in extracorporeal circulation. Thorax 1972;27:275–286.

    65. DePaepe J, Pomerantzeff PMA, Nakiri K, et al. Observations of the microcirculation of the cerebral cortex of dogs subjected to pulsatile and non-pulsatile flow during extracorporeal circulation. In: A Propos du Débit Pulse. Belgium: Cobe Laboratories1979.

    66. Taylor KM, Wright GS, Reid JS, et al. Comparative studies of pulsatile and non-pulsatile flow during cardiopulmonary bypass. II. The effects on adrenal secretion of cortisol. J Thorac Cardiovasc Surg 1978;75:574–578.

    67. Taylor KM. Pulsatile perfusion. In: Taylor KM, ed. Cardiopulmonary bypass—principles and management. London: Chapman and Hall, 1986.

    68. Taylor KM, Wright GS, Bain WH, et al. Comparative studies of pulsatile and non-pulsatile flow during cardiopulmonary bypass. III. Anterior pituitary response to thyrotrophin-releasing hormone. J Thorac Cardiovasc Surg 1979;75:579–584.

    69. Philbin DM, Coggins CH, Emerson CW, et al. Plasma vasopressin levels and urinary sodium excretion during cardiopulmonary bypass: a comparison of halothane and morphine anesthesia. J Thorac Cardiovasc Surg 1979;77:582–585.

    70. Briceno JC, Runge TM. Monitoring of blood gases during prolonged cardiopulmonary bypass and their relationship to brain pH, PO2 and PCO2. J ASAIO 1994;40:M344–M350.

    71. Kono M, Orita H, Shimanuki T, et al. A clinical study of cerebral perfusion during pulsatile and non-pulsatile cardiopulmonary bypass. Nippon Geka Gakkai Zasshi 1990;91:1016–1022.

    72. Mori A, Watanabe K, Onoe M, et al. Regional blood flow in the liver, pancreas and kidney during pulsatile and nonpulsatile perfusion under profound hypothermia. Jpn Circ J 1988:52:219–227.

    73. Tranmer BI, Gross CE, Kindt GW, et al. Pulsatile versus nonpulsatile blood flow in the treatment of acute cerebral ischemia. Neurosurgery 1986;19:724–731.

    74. Onoe M, Mori A, Watarid S, et al. The effect of pulsatile perfusion on cerebral blood flow during profound hypothermia with total circulatory arrest. J Thorac Cardiovasc Surg 1994;108:119–125.

    75. Henze T, Stephan H, Sonntag H. Cerebral dysfunction following extracorporeal circulation for aortocoronary bypass surgery: no differences in neuropsychological outcome after pulsatile versus nonpulsatile flow. Thorac Cardiovasc Surg 1990;38:65–68.

    76. Murkin JM, Martzke JS, Buchan AM, et al. A randomized study of the influence of perfusion technique and pH management strategy in 316 patients undergoing coronary artery bypass surgery: II. Neurological and cognitive outcomes. J Thorac Cardiovasc Surg 1995;110:349–362.

    77. Murkin JM, Martzke JS, Buchan AM, et al. A randomized study of the influence of perfusion technique and pH management strategy in 316 patients undergoing coronary artery bypass surgery: I. Mortality and cardiovascular morbidity. J Thorac Cardiovasc Surg 1995;110:340–348.

    78. Hindman BJ, Dexter F, Smith T, et al. Pulsatile versus nonpulsatile flow. No difference in cerebral blood flow or metabolism during normothermic cardiopulmonary bypass in rabbits. Anesthesiology 1995;82:241–250.

    79. Chow G, Roberts IG, Harris D, et al. Stockert roller pump generated pulsatile flow: cerebral metabolic changes in adult cardiopulmonary bypass. Perfusion 1997;12:113–119.

    80. Feiner H. Pancreatitis after cardiac surgery. Am J Surg 1976;131:684–688.

    81. Baca I, Beiger W, Mittmann U, et al. Comparative studies of pulsatile and continuous flow during extracorporeal circulation. Effects on liver function and endocrine pancreas secretion. Chir Forum Exp Klin Fortschr 1979;49–53.

    82. Saggau W, Baca I, Ros E, et al. Clinical and experimental studies on pulsatile and continuous flow during extracorporeal circulation. Herz 1980;5:42–50.

    83. Murray WR, Mittra S, Mittra D, et al. The amylase creatinine clearance ratio following cardiopulmonary bypass. J Thorac Cardiovasc Surg 1982;82:248–253.

    84. Pappas G, Winter SD, Kopriva CJ, et al. Improvement of myocardial and other vital organ functions and metabolism with a simple method of pulsatile flow (IAPB) during clinical cardiopulmonary bypass. Surgery 1975;77:34–44.

    85. Mathie R, Desai J, Taylor KM. Hepatic blood flow and metabolism during pulsatile and non-pulsatile cardiopulmonary bypass. Life Support Systems 1984;2:303–305.

    86. Chiu IS, Chu SH, Hung CR. Pulsatile flow during routine cardiopulmonary bypass. J Cardiovasc Surg (Torino) 1984;25:530–536.

    87. Gauss A, Druck A, Hemmer W, et al. Abdominal complications following heart surgery. Anaesthesiol Intensivmed Notfallmed Schmerzther 1994;29:23–29.

    88. Baue AE. The role of the gut in the development of multiple organ dysfunction in cardiothoracic patients. Ann Thorac Surg 1993;55:822–829.

    89. Bowles CT, Ohri SK, Nongchana K, et al. Endotoxaemia detected during cardiopulmonary bypass with a modified Limulus amoebocyte lysate asay. Perfusion 1995;10:219–228.

    90. Andersen LW, Baek L. Transient endotoxaemia during cardiac surgery. Perfusion 1992;7:53–58.

    91. Andersen LW, Baek L, Degen H, et al. Presence of circulating endotoxins during cardiac operations. J Thorac Cardiovasc Surg 1987;93:115–119.

    92. Rocke DA, Gaffin SL, Wells MT, et al. Endotoxaemia associated with cardiopulmonary bypass. J Thorac Cardiovasc Surg 1987;93:832–837.

    93. Jansen NJ, van Oeveren W, Gu YJ, et al. Endotoxin release and tumor necrosis factor formation during cardiopulmonary bypass. Ann Thorac Surg 1992;54:744–747.

    94. Ohri SK, Bowles CT, Siddiqui A, et al. The effect of cardiopulmonary bypass on gastric and colonic mucosal perfusion. A tonometric assessment. Perfusion 1994;9:101–108.

    95. Tao W, Zwischenberger JB, Nguyen TT, et al. Gut mucosal ischemia during normothermic cardiopulmonary bypass results from blood flow redistribution and increased oxygen demand. J Thorac Cardiovasc Surg 1995;110:819–828.

    96. Riddington DW, Venkatesh B, Boivin CM, et al. Intestinal permeability, gastric intramucosal pH, and systemic endotoxemia in patients undergoing cardiopulmonary bypass. JAMA 1996;275:1007–1012.

    97. Quigley RL, Caplain MS, Perkins JA, et al. Cardiopulmonary bypass with adequate flow and perfusion pressure prevents endotoxaemia and pathologic cytokine production. Perfusion 1995;10:27–33.

    98. Fiddian-Green RG. Gut mucosal ischemia during cardiac surgery. Semin Thorac Cardiovasc Surg 1990;4:389–399.

    99. Reilly PM, Bulkey GB. Vasoactive mediators and splanchnic perfusion. Crit Care Med 1993;21:S55–S68.

    100. Ogino H, Miki S, Ueda Y, et al. A case of bowel necrosis due to acute mesenteric ischemia following pulsatile cardiopulmonary bypass. Ann Thorac Cardiovasc Surg 1998;4:34–36.

    101. DeBakey ME. A simple continuous flow blood instrument. New Orleans Med Surg J 1934;87:387–389.

    102. Wesolowski SA, Miller HH, Halkett AE. Experimental replacement of the heart by a mechanical extracorporeal pump. Bull N Engl Med Center 1950;12:41–50.

    103. Gourlay T, Taylor KM. Perfusion pumps. In: Jones RA, Elliott MJ, eds. Cardiopulmonary bypass in neonates and young children. London: Butterworth–Heineman, 1994.

    104. Ogata T, Ida Y, Takeda J. Experimental studies on the extracorporeal circulation by use of our pulsatile arterial pump. Lung 1959;6:381–383.

    105. Nonoyama A. Haemodynamic studies on extracorporeal circulation with pulsatile and non-pulsatile blood flows. Arch Jpn Chir 1960;29:381–406.

    106. Nakayama K, Tamiya T, Yamamoto K, et al. High-amplitude pulsatile pump in extracorporeal circulation with particular reference to hemodynamics. Surgery 1963;54:798–803.

    107. Wright G. Blood cell trauma. In: Taylor KM, ed. Cardiopulmonary bypass. London: Chapman and Hall, 1986:249–276.

    108. Gourlay T, Taylor KM. Pulsatile flow and membrane oxygenators. Perfusion 1994;9:189–196.

    109. Taylor KM, Bain WH, Maxted KJ, et al. Comparative studies of pulsatile and nonpulsatile flow during cardiopulmonary bypass. I. Pulsatile system employed and its hematologic effects. J Thorac Cardiovasc Surg 1978;75:569–573.

    110. Adams S, Fleming J, Gourlay T, et al. Clinical experience with the Sarns pulsatile pump during open heart surgery. Perfusion 1986;1:53–56.

    111. Rottenberg D, Sondak E, Rahats S, et al. Early experience with a true pulsatile pump for heart surgery. Perfusion 1995;10:171–175.

    112. Mutch WA, Lefevre GR, Theissen DB, et al. Computer-controlled cardiopulmonary bypass increases jugular venous oxygen saturation during rewarming. Ann Thorac Surg 1998;65:59–65.

    113. Sanderson JM, Morton PG, Tolloczko TS, et al. The Morton-Keele pump—a hydraulically activated pulsatile pump for use in extracorporeal circulation. Med Biol Eng 1973;Mar,11(2):182–190.

    114. Runge TM, Grover FL, Cohen DJ, et al. Preload-responsive, pulsatile flow, externally valved pump: cardiovascular bypass. J Invest Surg 1989;2:269–279.

    115. Leschinski BM, Itkin GP, Zimin NK. Centrifugal blood pumps—a brief analysis: development of new changes. Perfusion 1991;6:115–123.

    116. Pennington DG, Merjavy JP, Swartz MT, et al. Clinical experience with centrifugal pump ventricular assist devices. Trans Am Soc Artif Intern Organs 1982;28:93–99.

    117. Nose Y. Nonpulsatile mode of blood flow required for cardiopulmonary bypass and total body perfusion. Artif Organs 1993;17:92–102.

    118. Berki T, Gurbuz A, Isik O, et al. Cardiopulmonary bypass using a centrifugal pump. Vasc Surg 1992;26:123–134.

    119. Mandl JP. Comparison of emboli production between a constrained vortex pump and a roller pump. Proceedings AmSECT , 1977.

    120. Noon GP, Sekela ME, Glueck , et al. . Comparison of in vitro blood trauma and spallation produced by the Bio-medicus and roller pump during hypothermia. International Workshop on Rotary Blood Pumps, Baylor College, 1988.

    121. Matsukura H. An experimental study of the preservation of platelet during extracorporeal circulation. Jpn J Thorac Surg 1983;36:11.

    122. Maas C, Kok R, Segers P. Comparison between a centrifugal and a roller pump. Abstract of presentation at Pathophysiology and Techniques of Cardiopulmonary Bypass, San Diego, California, February, 1993.

    123. Nishida H, Uesugi H, Nishinaka T, et al. Clinical evaluation of pulsatile flow mode of Terumo Capiox centrifugal pump. Artif Organs 1997;21:816–821.

    124. Gobel C, Eilers R, Reul H, et al. A new blood pump for cardiopulmonary bypass: the HiFlow centrifugal pump. Artif Organs 1997;21:841–845.

    125. Komada T, Maet H, Imawaki S, et al. Haematologic and endocrinologic effects of pulsatile cardiopulmonary bypass using a centrifugal pump. Nippon Kyobu Geka Zasshi 1992;40:901–911.

    126. Ninomiya J, Shoji T, Tanaka IS, et al. Clinical evaluation of a new type of centrifugal pump. Artif Organs 1994;18:702–705.

    127. Dreissen JJ, Dhaese H, Fransen G, et al. Pulsatile compared with nonpulsatile perfusion using a centrifugal pump for cardiopulmonary bypass during coronary artery bypass grafting. Effects on systemic haemodynamics, oxygenation and inflammatory response parameters. Perfusion 1995;10:3–12.

    128. Bregman D, Bowman FO Jr, Parodi EN, et al. An improved method of myocardial protection with pulsation during cardiopulmonary bypass. Circulation 1977;56[Suppl 2]:157–160.

    129. Maddoux G, Pappas G, Jenkins M, et al. Effect of pulsatile and nonpulsatile flow during cardiopulmonary bypass on left ventricular ejection fraction early after aortocoronary bypass surgery. Am J Cardiol 1976;37:1000–1006.

    130. Philbin DM, Levine FH, Emerson CW, et al. Plasma vasopressin levels and urinary flow during cardiopulmonary bypass in patients with valvular heart disease: effects of pulsatile flow. J Thorac Cardiovasc Surg 1979;78:779–783.

    131. Tomatis L, Nemiroff M, Raihi M, et al. Massive arterial air embolism due to rupture of pulsatile assist device. Successful treatment in the hyperbaric chamber. Ann Thorac Surg 1981;32:604–608.

    132. Zumbro GL, Shearer G, Fishback ME, et al. A prospective evaluation of the pulsatile assist device. Ann Thorac Surg 1979;28:269–273.

    133. Wright G. Factors affecting the pulsatile hydraulic power output of the Stockert roller pump. Perfusion 1989;4:187–195.

    134. Inzoli F, Pennati G, Mastrantonio F, et al. Influence of membrane oxygentors on the pulsatile flow in extracorporeal circuits: an experimental analysis. Int J Artif Organs 1997;20:455–462.

    135. Gourlay T, Taylor KM. Pulsatile flow and membrane oxygenators. Perfusion 1994;9:189–196.

    136. Runge TM, Cohen DJ, Hantler CB, et al. Achievement of physiologic pulsatile flow on cardiopulmonary bypass with a 24 French cannula. J ASAIO 1992;38:M726–M729.

    137. Kayser KL. Pulsatile perfusion problems. Ann Thorac Surg 1979;27:284–285.

    138. Karichev Z, Muler A, Vishnevsky M. Unusual oxygen transfer in microporous blood oxygenator. Int J Artif Organs 1997;21:522.

    139. Pearson DT. Microemboli. In: Taylor KM, ed. Cardiopulmonary bypass—principles and management. London: Chapman and Hall, 1986.

    140. Bellhouse BJ. Design of the Oxford membrane oxygenator and its performance during prolonged support of newborn lambs. In: Zapol WM, ed. Membrane lungs for acute respiratory failure. New York: Academic Press, 1976.

    141. Shepherd S, Pierce JMT. Pulsatile flow during cardiopulmonary bypass speeds thermal energy transfer: a possible explanation for the reduced afterdrop. Perfusion 1995;10:111–114.