| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
Department of Anesthesiology, Mayo Clinic, Rochester, Minnesota
Address correspondence and reprint requests to David J. Cook, MD, Department of Anesthesiology, Mayo Clinic, 200 First St. SW, Rochester, MN 55905. Address e-mail to cook.david{at}mayo.edu
 |
Introduction |
|---|
 Top Introduction Historical Background:The Physiology of Normothermic...Clinical Considerations for Warm...Warm CPB and the...Summary and DiscussionReferences |
|---|
In this review, I briefly present the historical background of normothermic CPB and the initial reports that triggered the change in CPB practice. I also discuss how the physiology of warm CPB differs from hypothermic bypass and examine the impact of systemic normothermia in the postoperative period. The studies on the effects of warm bypass on neurologic outcome and the rationale for "tepid" bypass are also reviewed. Finally, some practical recommendations for the intraoperative management of these patients are provided.
|
Historical Background: |
|---|
|
TopIntroductionHistorical Background:The Physiology of Normothermic...Clinical Considerations for Warm...Warm CPB and the...Summary and DiscussionReferences |
|---|
The Rationale for and Practice of Systemic Hypothermia for Cardiac Surgery
For theoretical and practical reasons, the introduction of hypothermia and hemodilution mostly occurred in tandem. Considerable morbidity was demonstrated with early CPB, and the effect of hypothermia in reducing O2 demand and in increasing ischemic tolerance was established (6,8). In practice, it was argued that the reduction in metabolic rate associated with hypothermia would also allow for the use of an asanguineous prime and/or reduced CPB flows (9). In fact, hypothermia required hemodilution because hypothermia with a whole blood prime resulted in hypertension during CPB. The introduction of hypothermia was also of practical importance because cardiac surgery was placing a tremendous strain on blood banking resources. It was typical for these cases to use whole fresh blood (<24 h old) in volumes of 1500–2500 mL as a circuit prime (3–5,10). This blood use limited cardiac surgical capacity (10).
Hypothermia had become an established practice for adult cardiac surgery by the late 1960s and, at most institutions, constituted the largest part of the surgical practice until the reintroduction of warm CPB in 1991 and 1992.
The Arguments for Warm Cardioplegia
The introduction of systemic normothermia for CPB in the early 1990s was indirect. Studies that eventually resulted in systemic normothermia during CPB included investigations of alternate cardioplegia techniques. Specifically, the research group at the University of Toronto, extending some of the earlier observations in canines by Partington et al. (11) and Buckberg et al. (12), advocated the myocardial advantages of a warm continuous cardioplegic technique (13–15). Clinical literature indicated advantages of a warm induction of cardioplegia (16), as well as a terminal "hot shot," that is, perfusion of the heart with warm aerobic cardioplegia before removing the aortic cross-clamp (17). The Toronto group reasoned that warm continuous cardioplegia would extend these advantages (13–15). First, hypothermia, by its effect on enzyme kinetics, was known to alter myocardial substrate use and to impair the generation of adenosine triphosphate (14,18,19). Second, hypothermia was also known to shift the oxyhemoglobin dissociation curve leftward, potentially limiting O2 offloading in the ischemic myocardium (20). Third, the greatest reduction in myocardial O2 consumption is achieved by diastolic arrest (12,21,22). The additional reduction in maximal venous oxygen consumption (MVO2) provided by hypothermia was relatively small (12,21) and difficult to justify based on its potentially harmful effects on myocardial contractility (23–25). It was hypothesized that the MVO2 of the heart arrested in diastole could be met by warm blood cardioplegia provided continuously (15,18,22,26). Fourth, the addition of continuous cardioplegia had the potential to eliminate or attenuate the reperfusion injury thought to be associated with intermittent cardioplegia techniques (26,27).
The Toronto Results of Warm Cardioplegia and CPB
Clinical reports from Toronto with this new technique showed very favorable results (24,28). In those studies, warm continuous blood cardioplegia was delivered from antegrade or antegrade and retrograde cannula. Because blood cardioplegia is composed of a high potassium solution mixed in a greater volume with blood shunted from the oxygenator, the cardioplegia and systemic temperatures in the initial reports were similar, although the major rationale for systemic normothermia was to allow for prompt separation from CPB. Although systemic normothermia represented a major shift in practice, the practical physiologic impact of systemic normothermia was not addressed until a subsequent clinical report (29). With the warm continuous cardioplegia technique and systemic normothermia, the Toronto group reported significant improvements in myocardial outcome in 121 consecutive coronary artery bypass graft patients undergoing relative to a historical cohort of 133 patients managed with the cold intermittent cardioplegia technique (24). In the postbypass period, cardiac output was significantly improved in the warm group. The incidence of myocardial infarction was 1.7% in the warm group versus 6.8% in the cold group. Low cardiac output syndrome occurred in 14% of hypothermic technique patients versus 3% of the warm technique patients, and the need for intraaortic balloon counterpulsation was 1% in the warm group versus 9% in the cold group. Additionally, 99% of patients undergoing the warm technique converted to sinus rhythm spontaneously. A subsequent report using this technique in patients with ventricular dysfunction secondary to acute myocardial infarction showed similar advantages compared with a historical control group (28).
At nearly the same time, Flack et al. (30) compared the effects of combined antegrade and retrograde normothermic with hypothermic cardioplegia on postoperative conduction abnormalities and CK-MB release. The outcomes of 120 patients undergoing surgery with normothermic (37°C) cardioplegia and systemic normothermia (37°C) were compared with a historical cohort of 98 consecutive patients undergoing CPB with hypothermic cardioplegia (6–8°C) and a body temperature of 28°C. Although the aortic cross-clamp times were longer in the normothermic group, CK-MB release was less. Although it did not reach statistical significance, the hypothermic group also showed a threefold higher requirement for postoperative intraaortic balloon pump support. Finally, the normothermic group had a much lower incidence of conduction abnormalities postoperatively and at long-term follow-up (30). Because of the physiologic foundation for the practice and the reported myocardial advantages, the techniques advocated by the Toronto group were rapidly adopted.
Because these results were obtained with warm cardioplegia and warm systemic temperature, it was not clear whether the hemodynamic advantages were completely secondary to the cardioplegia technique. The relative roles of myocardial and systemic temperature are still not certain. However, elements of the Toronto reports, as well as previous and subsequent physiologic studies, suggested additional benefits related to higher body temperature.
To some extent, experimental studies on the warm cardioplegia technique followed its clinical introduction (14,18,19,22,31). Brown et al. (31) at Emory compared warm and cold cardioplegia techniques in a canine model of acute ischemia and largely confirmed the physiologic foundation for the clinical practice change. Using pressure-volume, electrocardiographic, and histologic studies in animals, they compared normothermic (37°C) blood cardioplegia with systemic normothermia (37°C) with cold (4°C) blood or cold crystalloid cardioplegia with systemic hypothermia (28°C). Although the histologic myocardial injury score in their acute ischemia model did not differ between groups, the warm technique resulted in significantly better overall ventricular performance, improved diastolic relaxation, and an improved electrophysiologic state (31). Magnetic resonance studies in canine hearts also indicated that warm, continuous blood cardioplegia better preserved contractile and metabolic function (18).
Since the initial reports from Toronto, cardioplegic temperature management has continued to evolve. It is more difficult to maintain electromechanical silence with the warm antegrade technique alone; intermittent warm ischemia results from interruption of delivery, and retrograde delivery alone may result in incomplete protection of the right ventricle and posterior septum (11,26,32,33). Higher hemoglobin concentrations and flow rates of the cardioplegia are also recommended with a warm cardioplegic technique (19,33). Finally, animal and clinical studies of cardioplegia temperature and delivery suggest that a combination of antegrade and retrograde techniques (34,35) at tepid (29°C) temperatures may be optimal (26,36,37).
|
The Physiology of Normothermic CPB: Systemic and Regional O2 Delivery |
|---|
|
TopIntroductionHistorical Background:The Physiology of Normothermic...Clinical Considerations for Warm...Warm CPB and the...Summary and DiscussionReferences |
|---|
However, normothermic bypass as it reemerged in the 1990s was practiced largely like hypothermic CPB, with the exception of higher temperatures. This differed from the early experience with warm bypass, because now patients were hemodiluted and warm, with pump flows approximating those seen in the intact circulation with a normal hematocrit (Hct). Calculation of DO2 under these conditions demonstrates that it is reduced by approximately 25%–35% (depending on the degree of hemodilution) relative to the nonhemodiluted, non-CPB state (38).
Under nonbypass conditions, hemodilution is associated with significant increases in cardiac output, such that whole body O2 delivery will be supported over a range of Hcts (39,40). Although the heart of the cardiac surgical patient may not adequately compensate for hemodilution (41), the normal physiologic response to a reduction in Hct is an increase in cardiac output. Therefore, the situation during normothermic CPB with asanguinous prime and a cardiac index of 2.2–2.5 L · min-1 · m2 differs significantly from the normothermic CPB without hemodilution, as well as from the physiology of the intact organism.
Although reductions in DO2 are generally well tolerated and whole body O2 consumption is maintained by increases in oxygen extraction, the balance between systemic O2 supply and demand is narrowed by warm CPB. Reductions in mixed venous O2 saturation (SvO2) are demonstrated, but uncommonly decrease to the range suggesting O2 extraction is approaching maximal. However, organs such as the brain, myocardium, intestinal mucosa, kidneys, and muscle differ in their ability to regulate regional perfusion and have differing metabolic rates and capacities to increase O2 extraction (42,43). SvO2 monitoring represents the body as a homogeneous compartment, and does not provide an adequate measure of oxygen delivery at the level of individual organs. This is probably more important during warm CPB, in which the margin between systemic O2 supply and demand is narrowed unless CPB flows are increased (38,44) (Table 1).
|
Although not reported by all investigators, MAP during warm bypass is typically reduced in the absence of interventions to maintain it. Because much of the neurohumoral response to CPB is related to aortic cross-clamping, normothermic patients may demonstrate the same catecholamine response to CPB as hypothermic patients (45,46); however, MAP is depressed secondary to the low systemic vascular resistance (SVR) associated with uncompensated hemodilution. Although the Hagen-Poiseuille equation describes isolated, nonbiological systems, it provides a useful framework for understanding MAP during warm CPB.
The Hagen-Poiseuille equation is: 
|
|
P is the pressure gradient, r is vessel radius, L is vessel length, and 
is viscosity (47). The Hagen-Poiseuille equation indicates that, for the body as a whole, if blood viscosity decreases and there is no increase in cardiac output (pump flow), then MAP will decrease in proportion to the reduction in blood viscosity. The effect of moderate hemodilution on blood viscosity is well established. Over the range of Hcts that patients experience during CPB, there is an approximately linear relationship between Hct and blood viscosity (48). The hemodynamic effect of this viscosity change is well presented by Gordon et al. (49), who described the relationship among patient red cell mass, priming volume of the circuit, and the SVR. Those authors developed nomograms that allow the prediction of the MAP change during bypass using these variables (Figure 1). The authors described a 50% reduction in MAP with a 40% reduction in blood viscosity in the early phase of CPB.
|
During hypothermic bypass, the sympathetic response to cold and increases in blood viscosity with induced hypothermia (50,51) will offset the reduction in SVR seen with isolated hemodilution. Thus, during hypothermic CPB, hemodilution-related hypotension is typically only present during the early and late phases of CPB when the patient approximates normothermia. During warm bypass, the effect of hemodilution on blood viscosity and SVR may be demonstrated throughout the entire period of extracorporeal circulation.
The effect of warm CPB on SVR is evident in clinical practice. After the introduction of normothermic bypass in 1991 and 1992, reports promptly appeared on the differing hemodynamics of CPB using this technique. Although the experience has not been universal (52–54), most reports indicated that, relative to hypothermic patients, warm bypass patients require larger doses of vasoconstrictors, greater fluid administration, or higher pump flows to maintain MAP during CPB (29,44,55–58). To maintain a MAP of 50 mm Hg, Christakis et al. (29) demonstrated a threefold increase in total phenylephrine requirements in 101 patients at 35°C relative to those in 103 patients at 28–30°C (Figure 2).
|
-agonists may offset their vasoconstrictive effects, such that tissue blood flow is unchanged or increased. This is suggested by the report of Redl-Wenzl et al. (62), who determined creatinine clearance, cardiac output, and whole body O2 consumption in septic patients placed on norepinephrine infusions, but it has not been determined whether those results are transferrable to the setting of CPB. Ultimately, the relative balance between vasoconstriction and increased perfusion pressure should be considered on a case-specific basis when MAP is reduced during CPB.
Warm CPB and Splanchnic Perfusion
Simple calculation of systemic O2 balance indicates that, during CPB with hemodilution and flow rates of 2.2–2.5 L · min-1 · m2, whole body O2 delivery is reduced relative to the non-CPB state. Therefore, if oxygen delivery of critical organ beds such as brain is to be preserved, flow to other tissues must be reduced. Animal studies indicate that, during hypothermic CPB, splanchnic perfusion is reduced or altered in distribution (42,59,63,64). The same is true during normothermic CPB, in which liver (63), renal (59,65,66), and intestinal mucosal (67–69) blood flow are decreased. This is of further physiologic significance because arterial oxygen content is also reduced. This, together with the lower MAP often associated with warm CPB, may increase the likelihood of visceral ischemia.
Mackay et al. (70) examined the effect of dopamine and perfusion pressure on visceral blood flow during normothermic (37°C) CPB in pigs and found that kidney, stomach, intestine, and pancreas blood flow was approximately doubled as the MAP was increased from 45 to 90 mm Hg by increases in pump flow. Conversely, dopamine 5 µg · kg-1 · min-1 did not increase visceral blood flow under either MAP condition. A related result, also in a swine model, was obtained in a second study by the same group. In that investigation, the effect on regional perfusion of using either pump flow or phenylephrine to increase MAP during normothermic CPB was determined (43). With a MAP of 65 mm Hg, visceral blood flow was significantly greater when MAP was supported by pump flow rather than with an -agonist. Although the authors do not provide a non-CPB measurement, data obtained with CPB flows of 70 and 100 mL · kg-1 · min-1 suggest that visceral organ perfusion may be compromised with conventional flows and the perfusion pressures that are common during warm CPB.
Despite these physiologic observations, an increase in renal insufficiency or splanchnic morbidity has not been reported with warm CPB or the use of vasoconstrictors during these operations in small studies of relatively low-risk patients (52,58,71).
|
Clinical Considerations for Warm CPB |
|---|
|
TopIntroductionHistorical Background:The Physiology of Normothermic...Clinical Considerations for Warm...Warm CPB and the...Summary and DiscussionReferences |
|---|
The effect of CPB temperature on anesthetic adjuncts is clearer. In particular, the use of neuromuscular blocking drugs can be expected to differ during warm CPB relative to the hypothermic practice. Clinically, under non-CPB conditions, even 2°C of hypothermia may double both the duration of action of vecuronium and the time to spontaneous recovery (85). Across studies, hypothermia reduces the amount of atracurium, rocuronium, pancuronium, and vecuronium infusion required to maintain a given level of neuromuscular blockade by approximately 30%–75% (86–89), so that differences in dosing of these drugs (relative to hypothermic practice) should be anticipated with warm CPB.
Although the reports from Toronto did not indicate that the normothermic technique required larger doses of heparin to maintain a target activated clotting time (29), this result might be expected given the effect of hypothermia on drug metabolism (90,91) and organ blood flow (42,59,63). A retrospective study of 354 patients undergoing coronary artery bypass grafting examined this question and reported very little correlation (r = 0.13) between minimal CPB temperature (24–37°C) and heparin requirements (92). However, when the heparin requirement was adjusted for bypass duration, the relationship was more revealing. Although CPB was shorter (P < 0.0001) at higher bypass temperatures (>32°C), the total heparin requirement and, therefore, heparin required/minute of CPB, was greater than when patients underwent hypothermic (<32°C) CPB (P = 0.018). The authors suggest that greater attention to coagulation monitoring is required in patients undergoing warm CPB.
The observation that normothermic CPB is associated with shorter bypass times has been noted with some consistency (53,93,94). Although this experience is not universal (95–97), a shorter CPB time may be anticipated if the rewarming phase is eliminated. Because there is an association between CPB duration and morbidity (98) and because the rewarming phase may be a significant neurologic stressor (99–102), the elimination of the rewarming period has theoretical advantages.
Although not an issue of anesthetic drug pharmacology, the effect of CPB temperature on glucose homeostasis deserves comment. Hyperglycemia is common during hypothermic CPB. This is related to catecholamine secretion, inhibition of insulin secretion, and a decrease in peripheral glucose use (103–105). Although the inhibition of insulin secretion is attenuated by normothermia, a partial inhibition remains, and peripheral glucose use is still impaired. Therefore, during normothermic CPB, there remains a relative glucose intolerance such that a glucose load may lead to a persistent hyperglycemia (104). Therefore, the increase in insulin requirements demonstrated during CPB is largely independent of temperature.
Normothermic Cardioplegia and Cardiac Rhythm
Most reports on the warm technique document clear advantages with regard to cardiac rhythm after removal of the aortic cross-clamp (24,29,57,95). The Warm Heart Investigators (106) noted a decreased requirement for electrical defibrillation in their warm cardioplegia group. Of these patients, 99% defibrillated spontaneously, versus 11% of the cold group. This effect on rhythm was further emphasized in the subsequent report from Christakis et al. (29) and by investigators in the United Kingdom (57) and Israel (95). Figure 3 is from the work of Gozol et al. (95), who examined recovery of rhythm in 36 patients randomized to receive normothermic (34°C) or hypothermic (10°C) cardioplegia during CPB with systemic temperatures of >32°C or 24°C, respectively. In addition to the spontaneous defibrillation previously reported with the warm technique, Gozol et al. documented spontaneous sinus recovery in 95% of patients versus 25% in the group undergoing the cold technique. The warm cardioplegia technique also profoundly reduced the incidence of complete atrioventricular block. Perhaps even more important, in the study by Flack et al. (30), the normothermic cardioplegic technique (37°C cardioplegia and systemic body temperature) reduced conduction abnormalities immediately postoperatively, at discharge, and at late (26–420 days) follow-up in patients receiving 37°C cardioplegia and CPB versus a historical cohort of patients receiving cardioplegia at 6–8°C and CPB at 28°C (Figure 4).
|
|
Normothermia and the CPB Inflammatory Response
CPB triggers an inflammatory response involving a variety of pathways, including complement, contact activation, and polymorphonuclear neutrophil and monocyte activation, resulting in cytokine release. Whether normothermic CPB accentuates or attenuates this response is not clear.
Between 26 and 37°C, clinical CPB has been reported to increase the expression of C3a, counts of circulating neutrophils, lymphocytes, and granulocyte elastase (97,107–109). Cytokines, interleukin (IL)-6, a nonspecific acute-phase reactant, and IL-8, a neutrophil-activating and chemotactic factor, are also demonstrated after CPB in patients undergoing CPB at 28–37°C (107,110,111). During CPB at 28–37°C, other inflammatory markers, such as IL-1ß and tumor necrosis factor may (111,112) or may not be (107,110,113) increased.
In vitro, hypothermia decreases complement activation (109,112) and cytokine production (112). Monocyte cultures also generate less interleukin-1 under hypothermic conditions. Although generation of complement fragments may also be decreased by hypothermia in vitro (112), the study by Quiroga et al. (108) did not demonstrate a difference in C3a between hypothermic (26–28°C) and normothermic (37°C) CPB patients. Similarly, IL-1ß production (109,111) and tumor necrosis factor (111) are inhibited by hypothermia in vitro, but these markers may (111,112) or may not be (107,113,114) demonstrated in patients undergoing CPB.
At least three studies have demonstrated an effect of CPB on levels of cytokines IL-6 and IL-8 (107,110,113). However, Frering et al. (110) only studied patients undergoing normothermic (>36.5°C) CPB, and Kawamura et al. (113) provided no information on bypass temperature management. Menasche et al. (111) and Ohata et al. (107) examined IL-6 levels in both hypothermic (28°C and 28–30°C, respectively) and normothermic (34° and 37°C, respectively) patients. The former reported greater increases in IL-6 levels in normothermic patients 2 h after bypass but no difference in subsequent measurements (111). Ohata et al. (107) could not document any effect of bypass temperature per se on IL-6 levels, although this inflammatory marker was increased in both groups. In that study, IL-8 and granulocyte elastase were reduced in the warm group relative to the hypothermic group at 12 h, although the groups did not differ at 24 h. In contrast, Menasche et al. (115) reported that elastase was increased in a warm CPB group at 30 min and 4 h post-CPB. Finally, Le Deist et al. (97) reported that expression of neutrophil adhesion molecules was increased by both hypothermic (27°C) and tepid (33°C) CPB, but meaningful temperature-dependent differences were not shown.
From these reports, no clear picture emerges for the effect of CPB temperature on the inflammatory response. In vitro studies on the effect of hypothermia on inflammatory mechanisms have not been shown to translate into what is demonstrated clinically. Clinical studies have typically been small and nonrandomized and often report different inflammatory indices at differing post-CPB intervals. Finally, these inflammatory processes are not unique to CPB. They are also associated with other forms of major surgery (114). CPB increases levels of circulating neutrophils and results in complement activation. Cytokines IL-6 and IL-8 are also generated in response to CPB. Hypothermia seems to delay some of these responses (Figure 5), but it is not clear whether this delay has a significant effect on the magnitude of the response or a meaningful effect on clinical outcomes. This remains a subject of intense investigation.
|
Temperature after-drop after cardiac surgery is well described (116–119). Patients may be weaned from CPB with a nasopharyngeal temperature of 37–38°C, but in the intensive care unit (ICU), core temperature may be <35°C (117,120). In addition to losing heat to the environment in the post-CPB period, there is evidence that patients undergoing hypothermic CPB are not adequately rewarmed (116,118,119). Measurement of muscle temperature at the completion of rewarming indicates that a large proportion of body mass remains hypothermic. This leads to temperature redistribution and core cooling in the postoperative period (116,118,119). Postoperative hypothermia increases catecholamine release, SVR, and heart rate (117,121,122). During the initial 2- to 6-h postoperative period when body temperature is increasing, whole body oxygen demand, ventilatory requirements, and myocardial work may increase dramatically, particularly if shivering is triggered (120,122). Although this has not been studied in cardiac surgical patients, these hypothermia-associated variables are associated with an increased incidence of cardiac arrest, myocardial ischemia, and infarction after noncardiac surgery (123,124).
Although mild hypothermia on arrival in the ICU is also seen after normothermic CPB, there is a clinical impression that postoperative hypothermia is less common and less severe. A body of research directly comparing postoperative temperature in patients undergoing warm and cold CPB has not been generated, but some literature suggests that there is a difference. Baker et al. (125) described core temperature changes in 56 patients after CPB at 37–38°C. Over a 12-h ICU observation period, mean temperature remained >37°C 4, 8, and 12 h postoperatively. Only on arrival in the ICU was core temperature 36.2°C (125). Tonz et al. (126) also described higher temperatures for the first postoperative day in their randomized study of CPB temperature. In another small study, Lehot et al. (44) demonstrated that core temperature was higher in the normothermic (37°C) CPB group throughout the first three postoperative hours, but they did not extend their observations beyond this period. Because patient temperature in the ICU is determined by many factors in addition to bypass temperature, such as body mass, blood loss, closure time, and room temperature, it remains to be seen whether consistent differences between hypothermic and warm CPB will be seen. If these differences are demonstrated, they will probably be of greatest importance in the sickest patients who have less capacity to reestablish thermal homeostasis.
The Role of Warm CPB in Early Extubation.
Because most protocols for postcardiac surgical extubation include a requirement for systemic normothermia, patients undergoing warm CPB might be tracheally extubated earlier than their hypothermic CPB counterparts. Additionally, respiratory acidosis and increased respiratory demand in the early ICU period (44,120) might also delay the tracheal extubation of hypothermic patients relative to patients managed with warm CPB. Finally, if the hemodynamic advantage of reduced SVR, stable rhythm, and increased cardiac output are carried into the postoperative period, this might facilitate earlier extubation of patients undergoing warm CPB. Despite these arguments, only a few studies have specifically reported the effect of CPB temperature on extubation time. These investigations have shown either a shorter time to extubation (53) or no difference (57,96) with the warm technique. Physiologic studies also provide conflicting information. Birdi et al. (127) could not identify an effect of CPB temperature on the postbypass arterial-alveolar O2 partial pressure (A-a)O2 difference in patients without pulmonary disease. Conversely, Rannucci et al. (54) evaluated transpulmonary shunt fraction, as well as (A-a)O2 and CO2 gradients, in 50 smokers and patients with chronic obstructive pulmonary disease randomized to CPB at either 36°C or 28°C. Those investigators found that the transpulmonary shunt was reduced in patients after warm CPB. The same was true for (A-a)O2 and (A-a)CO2 differences. However, these differences were not demonstrated 3 h after arrival in the ICU.
Regardless of these observations, time to extubation is a problematic end point in evaluating the advantages or disadvantages of an anesthetic or surgical management technique. Although the decision to extubate is based on objective criteria, determinants such as time of day, position on the surgical list, and motivation of ICU personnel can introduce a great deal of variability into the process. It remains to be determined whether warm CPB alters extubation time or the costs associated with the early post-CPB period. In the absence of a compelling group difference or rigorous extubation criteria, time to extubation must be viewed as a soft end point. In some ways, the same can be said of using the volume of blood transfused as a clinical end-point.
Normothermic CPB and Postoperative Bleeding.
Hypothermia has well established effects on coagulation. In experimental models, hypothermia to the range commonly used during CPB (25–32°C) induced reversible platelet dysfunction (128) and inhibited activated clotting factors (129,130). In noncardiac surgery, there is an association of hypothermia with excessive bleeding (131). Functional tests of coagulation suggest that warm CPB may have advantages, but the clinical outcome data are not compelling.
Boldt et al. (132) examined platelet aggregation in four groups of patients undergoing CPB at >34°C or <28°C, with and without aprotinin. They determined that platelet aggregation was most inhibited in the hypothermic patients, and that these patients showed the slowest recovery of platelet function. Postoperative blood loss was also highest in the hypothermic group. Finally, aprotinin blunted the effects of hypothermia on platelet function in the hypothermic group but was without beneficial effect on platelet aggregation in patients undergoing warm CPB.
A later study by the same group examined the effect of CPB temperature on circulating thrombomodulin, protein S, and protein C; platelet aggregation, and the thrombin/antithrombin complex (55). Relative to the warm group (>35°C), hypothermic (28°C) patients demonstrated increased levels of the thrombin/antithrombin III complex (Figure 6), greater reduction in protein S and C levels, greater inhibition and slower recovery of platelet function, and greater blood loss, and they received more homologous transfusion (55).
|
Clinical studies not specifically designed to examine blood loss have also reported this variable. These studies either report no difference in blood loss between temperature groups (71,96) or reduced blood loss in the warm group (57,126) (Figure 7), but none reported greater blood loss in the warm CPB group.
|
|
Warm CPB and the Brain |
|---|
|
TopIntroductionHistorical Background:The Physiology of Normothermic...Clinical Considerations for Warm...Warm CPB and the...Summary and DiscussionReferences |
|---|
Cerebral O2 Balance and Warm CPB
Although clinical studies had previously reported cerebral physiologic variables during warm conditions, these studies were often performed during the brief period after CPB rewarming (134,135), rather than during continuous normothermic CPB. After the introduction of normothermic CPB at the Mayo Clinic, Cook et al. (100) examined jugular bulb O2 saturation in patients randomized to CPB at either 27 or 37°C and found that cerebral venous desaturation occurred in more than half of normothermic CPB patients within the first 20 min of CPB but resolved by 60 min of CPB. This pattern differed from the desaturation pattern demonstrated in hypothermic patients who only showed low venous O2 saturation during rewarming.
To explain these results, Cook et al. (136) conducted a second randomized study to determine the global cerebral O2 balance during normothermic CPB. Interestingly, this demonstrated that during CPB at 37°C, cerebral metabolic rate and oxygen delivery were unchanged from the prebypass state despite a 30% reduction in hemoglobin (Hgb) concentration (prebypass Hgb 11.7 g/dL, CPB Hgb 8.1 g/dL) (Figure 8). Essentially, if MAP was maintained, increases in cerebral blood flow compensated for the reduction in arterial O2 content so that cerebral oxygen balance was unchanged. These findings were initially difficult to reconcile with the previous report of cerebral venous desaturation during early CPB (100). However, examination of the data indicated that approximately 2°C of warming was required to achieve strict normothermia during the early phase of CPB (136). Increasing brain temperature was probably instrumental in the early cerebral venous desaturation reported previously. These clinical data and subsequent animal experiments (137) suggest that, from the standpoint of hemispheric O2 balance, normothermic CPB with conventional hemodilution could be well tolerated in a patient population without diabetes or a history of cerebral vascular disease.
|
Warm Bypass and Neurologic Outcome
Although the warm CPB technique seemed to offer advantages, concern over neurologic outcomes came to the forefront in 1994 when large randomized reports were published. The Warm Heart Investigators from Toronto published an outcome study of 1732 patients (860 warm, 872 cold) (106). Patients underwent CPB at either 33–37°C or 25–30°C, and the incidence of stroke at discharge was equivalent (Table 2). However, in the same year, Martin et al. (142) from Emory reported a randomized trial of 1001 patients undergoing CPB at either 
35°C (n = 493) or 
28°C (n = 508). The Emory study was designed to examine the myocardial effects of warm cardioplegia with systemic normothermia but was interrupted because of a threefold greater incidence of stroke or encephalopathy in the warm group.
|
Further analysis of differences between the Toronto and Emory studies was provided by Guyton et al. (144), who pointed out that the Toronto population included fewer women, a smaller proportion of patients older than 70 yr (16% vs 38%), and fewer patients undergoing reoperations. Additionally, 94% of the Toronto patients received only anterograde cardioplegia, whereas the Emory patients also received retrograde cardioplegia. Reduced manipulation of the aorta by avoidance of a partial occluding clamp may also have contributed to better outcome in the Toronto group (33). Both retrograde cardioplegia and the application of a partial occluding clamp could place patients at greater risk of cerebral embolism (33,145–147). Finally, the warm group in the Toronto study was cooler than the warm group in the Emory study (144).
Another difference in the two patient study populations was the higher incidence of diabetes in the Emory study. Diabetes is an independent risk factor for stroke in the noncardiac surgical population (148) and seems to be an independent risk factor in cardiac surgical patients as well (149,150). The incidence of diabetes in the Emory study was approximately 26%, whereas in the Toronto population, it was approximately 6%. Although the higher incidence of diabetes at Emory might be expected to increase the stroke risk at both warm and cold CPB temperatures, diabetic cerebral vascular disease could be of greater relevance during normothermic CPB.
More limited neurocognitive outcomes of the Emory and Toronto study populations have also been provided. Mora et al. (151) provided neurocognitive data on a subset of the original 1001 patients enrolled at Emory. Neurocognitive testing was conducted preoperatively, during the postoperative hospitalization, and again at 1 mo for 109 patients. Study drop-out allowed complete reporting for approximately 80 subjects. Although CPB resulted in neurocognitive deficits in both groups, there was no difference between the warm and cold populations (151). Wong et al. (152) and McLean et al. (153) also reported neurocognitive outcomes on a subset of the Toronto study patients. McLean et al. completed neurocognitive testing preoperatively, during the postoperative hospitalization, and again at 3 mo for 153 patients. Similar to the Emory authors, these investigators found that CPB temperature was not a determinant of neurocognitive injury.
A variety of other reports have commented on stroke incidence or neurocognitive outcomes with warm CPB but are limited by lack of randomization or small study size. Plourde et al. (93) randomized 62 patients to CPB at either 34–35°C or 28°C and could not identify neuropsychologic differences between patient groups on the seventh postoperative day. Hvass et al. (52) reported 100 patients who underwent surgery with a warm body (37°C)-cold heart technique and documented a 1% incidence of stroke, whereas Birdi et al. (57) documented a 0%–1% incidence of stroke in 300 patients randomized to CPB at either 37, 32, or 28°C. Singh et al. (71) compared 2585 consecutive patients managed with a warm body (37°C)-cold heart technique with a historical cohort of 1605 patients who underwent surgery with systemic hypothermia (25–30°C). In that study, the stroke incidence was 1% and 1.3% in the warm and cold groups, respectively. Regragui et al. (58) conducted neurocognitive testing on 96 patients randomized to CPB at either 37, 32, or 28°C and found that the overall neurocognitive score was worse in the 37°C group (n = 31), better in the 32°C group (n = 36), and equivalent between the 28°C (n = 29) and 32°C groups. From these results, the authors make an argument for mild hypothermia. However, their article was followed by a commentary that faulted the study's statistical power, neurocognitive testing methods, and data analysis.
The weight of the evidence indicates that warm CPB (nasopharyngeal temperature 33–35°C), as opposed to strict normothermic CPB, in the absence of hyperglycemia, in a relatively healthy patient population, does not seem to significantly increase neurologic risk. It is unclear whether the same can be said of this technique in the high-risk population, and it is unlikely that studies of this magnitude will be conducted in a population at higher risk.
Temperature Monitoring During CPB
In discussions of temperature management, it must be emphasized that institutions differ regarding the site of temperature measurement, and that many studies fail to report this site. In fact, this is true of both the Emory (142) and Toronto (106) studies. The site of temperature measurement is important, because there may be large temperature differentials among different regions (118,154–158). This was well described by Stone et al. (156), who measured body temperature at nine sites in neurosurgical patients with profound hypothermia and circulatory arrest undergoing CPB. In that report, nasopharyngeal or tympanic membrane temperatures, typically used as an indicator of brain temperature, both greatly over- and underestimated brain temperature. Similarly, with rewarming from 27°C, nasopharyngeal temperature systematically underestimates the temperature of cerebral venous effluent measured at the jugular bulb (159). Thus, brain temperatures may be >39.5°C during rewarming. This temperature difference between brain and other sites may be even greater if body temperature is measured at the bladder or rectum (155). Because the warmest blood enters the aortic root just proximal to the carotid origins, and because the brain has a high blood flow on a per-gram basis, a disproportionate amount of kCal are delivered to the head (159,160). Peripheral sites such as bladder and rectum have lower blood flows and so change temperature more slowly. Furthermore, perfusate temperature has been dissipated proximal to these sites, so their absolute temperature may be lower than more proximal tissues. Therefore, maintaining strictly normal bladder temperature during CPB could result in a prolonged period of cerebral hyperthermia.
The experimental neurology literature suggests that cerebral hyperthermia may be the most simply avoided aggravator of ischemic neurologic injury, whereas mild hypothermia may be the single most important attenuator. If a primary goal is the regulation of brain temperature, then nasopharyngeal or tympanic membrane temperatures sites probably provide the greatest yield, although acknowledging the assumption that these sites may systematically underestimate actual brain temperature.
Mild Hypothermia and Cerebral Protection
One of the primary differences in CPB management between the studies from Emory (142) and Toronto (106) is related to the definition of normothermia. Although a host of other factors probably contributed to the outcome difference between the institutions, the 2–4°C temperature difference has been emphasized. There is convincing evidence in the experimental stroke literature that small degrees of hypothermia have important effects on neuropathologic and neurophysiologic outcomes in ischemia models (161–164).
In a rat model, Busto et al. (165) examined the effect of small differences in brain temperature (30, 33, 36, and 39°C) on the extent of neuronal ischemic injury. Although high-energy phosphate depletion did not differ among temperature groups, mild hypothermia had a profound effect on histopathologic outcome. At 30–31°C, ischemic cell damage was infrequently observed; at 33°C, ischemic injury was evident in additional areas; and at 36°C, severe histopathologic injury was demonstrated diffusely. At 39°C, histopathologic injury was most generalized and severe. Minamisawa et al. (161) reported that the protective effect of mild hypothermia is relevant even for brief periods of ischemia.
Figure 9 shows the influence of temperature on hippocampal neuronal necrosis in a rat model. These investigations also clearly demonstrated the adverse effects of small degrees of hyperthermia in ischemic models (161). The latter observation is particularly relevant to temperature management during rewarming, when cerebral temperature can exceed 39°C (159).
|
|
Summary and Discussion |
|---|
|
TopIntroductionHistorical Background:The Physiology of Normothermic...Clinical Considerations for Warm...Warm CPB and the...Summary and DiscussionReferences |
|---|
Practical Recommendations for the Management of CPB at Warmer Temperatures
Although experience with the warm CPB technique varies among institutions and definitive studies are not available for all management issues, several general observations can be made about the management of the patient undergoing warm CPB. First, the warm cardioplegia technique works well, but with it, cardioplegia must be given frequently and in large volumes. Additionally, adequate myocardial protection may depend on coronary sinus cannulation and retrograde cardioplegia. However, even with combined antegrade and retrograde cardioplegia, complete electromechanical silence can sometimes be difficult to achieve. This is of greatest issue in patients with left ventricular hypertrophy or with previous coronary grafts (166). The optimal temperature, flows, and route of delivery for warm cardioplegia continue to evolve.
Second, patients undergoing warm CPB usually have a lower SVR than patients undergoing surgery at lower temperatures. In the absence of support with higher pump flows or a vasoconstrictor, this will result in significantly lower MAP during CPB and in the post-CPB period. These patients may also have greater perioperative fluid requirements.
Third, during CPB, the ratio between systemic O2 supply and demand is narrowed as a function of hemodilution in the absence of a hypothermia-induced decrease in O2 demand. This may be reflected in a higher O2 extraction. Therefore, if CPB flows cannot be increased, then a higher hemoglobin concentration may sometimes be required.
Fourth, the pharmacokinetics of warm CPB may differ from that of hypothermic CPB. A relatively greater anesthetic requirement can be anticipated, as can a greater requirement for neuromuscular blockers. Additionally, the relative requirement for antifibrinolytic therapy may be reduced in this patient population.
Fifth, in the immediate postoperative period, less bleeding and greater stability of patient temperature and ventilatory variables may be seen, and this might allow for earlier extubation, although the data are currently not compelling.
Finally, the temperature-monitoring site during CPB is essential. Nasopharyngeal or tempanic membrane temperature sites are probably the most practical sites to use, although it is assumed that these sites may systematically underestimate actual brain temperature.
Unanswered Questions and Further Study
The ideal temperature for CPB is probably an indeterminate value that varies with the physiologic goals. Even accepting that proposition, the answer to at least four practical questions should improve patient care. First, although the neuroscience literature strongly support CPB at 32–34°C, no large randomized clinical study has been designed to directly compare outcomes of CPB performed at this temperature range with those at lower or higher temperatures. Thus, the practice of tepid CPB is somewhat of an educated guess. Second, because it is difficult to achieve whole body normothermia (in a reasonable time frame) without cerebral hyperthermia, terminating CPB at 34–35°C is being contemplated by some clinician investigators. If this option is pursued, it must be determined whether this practice has adverse effects on postoperative bleeding, O2 consumption, or hemodynamics that will offset any potential neurologic advantage. Third, the data on the influence of temperature on bleeding and inflammatory modulation must be considered as preliminary rather than definitive. With the increasing proportion of second operations and patients with multisystem disease, additional investigation is warranted. Fourth, and perhaps most important, it remains to be defined whether risk stratification can be applied to choose what temperature may be best for particular groups of patients. The choice of CPB temperature will be a compromise between competing goals. Risk stratification may allow us to better weigh surgical, hemodynamic, hemostatic, and neurologic costs and benefits and to choose an appropriate temperature management paradigm.
|
Acknowledgments |
|---|
|
References |
|---|
|
TopIntroductionHistorical Background:The Physiology of Normothermic...Clinical Considerations for Warm...Warm CPB and the...Summary and DiscussionReferences |
|---|
—
—
P < 0.05 versus 28°C. 
P < 0.002 versus 28°C, P < 0.04 versus 32°C. Modified from Birdi I, Regragui I, Izzat MB, et al. Influence of normothermic systemic perfusion during coronary artery bypass operations: a randomized prospective study. J Thorac Cardiovasc Surg 1997;114:475–81.
