Thoracic Epidural Anesthesia and the Patient with Heart Disease

Abstract

In the Western world, cardiovascular morbidity is the most common primary cause of death [1,2]. Coronary artery disease is a widespread concomitant condition in patients undergoing surgery [1,3]. The patient with coronary artery disease is susceptible to myocardial infarction, dysrhythmias, ventricular failure, and cardiac death [1]. Patients with chronic cardiovascular disease have a high risk of adverse cardiac outcome after surgery, especially after acute perioperative ischemic events [4]. The optimal anesthetic and analgesic management for these patients remains undefined [5]. Numerous studies have shown that stimulation of the sympathetic nervous system plays a major role in the development of perioperative myocardial ischemia [6-9]. Activation of cardiac sympathetic efferents may reduce myocardial oxygen availability by inducing poststenotic coronary constriction [10], which can lead to redistribution of myocardial blood flow with a reduction of blood supply to the subendocardium. Myocardial oxygen consumption is increased by tachycardia and increased contractility. This mismatch between oxygen delivery and demand during sympathetic activation can lead to myocardial ischemia, decreased arrhythmia threshold, and ventricular failure. In several patient populations with a high prevalence of coronary artery disease (such as peripheral vascular disease), the use of regional anesthesia has been investigated to improve outcome [11-13]. Thoracic epidural anesthesia (TEA) aims at a more specific reversible blockade of cardiac sympathetic efferents and afferents and provides intra- and postoperative analgesia or-in the nonsurgical population-effective therapy of anginal pain [14]. This review focuses on TEA and does not examine studies of lumbar epidural anesthesia and cardiac outcome in detail. Physiology and Pathophysiology In a resting adult who weighs 70 kg, coronary blood flow (CBF) is approximately 225 mL/min, which is approximately 4%-5% of cardiac output. In the autoregulatory range, CBF is nearly constant over a wide range of perfusion pressures. At a constant myocardial oxygen consumption, a decrease in coronary perfusion pressure causes autoregulatory vasodilation, adjusting coronary vascular resistance to maintain constant myocardial perfusion [15] (Figure 1). The major variables that influence coronary flow are perfusion pressure, myocardial systolic compression, metabolic control, and neurohumoral factors [16]. Beyond the autoregulatory range, coronary perfusion is proportionally related to coronary perfusion pressure and inversely dependent on coronary vascular resistance. Under these circumstances, changes in mean arterial pressure (MAP) are usually followed by a proportional change in CBF. Coronary perfusion pressure is often defined as the difference between MAP and left ventricular end-diastolic pressure. Because nearly 70% of CBF occurs during diastole, and increases in heart rate are accompanied by a shortening of the duration of diastole, heart rate is also an important determinant of CBF.Figure 1: Normal coronary blood flow (CBF) autoregulation at a constant myocardial oxygen consumption in dogs. The zone in which autoregulation is intact represents the pressure range at which CBF remains nearly constant as coronary pressure changes. Coronary vascular reserve is the difference between the lower curve and the straight maximum vasodilation line [modified from [15]].Neural effects on cardiac performance and rhythm are mediated via alpha- and beta- adrenergic receptors. Sympathetic alpha-adrenergic stimulation leads to a vasoconstrictive influence on epicardial vessels, in which alpha-receptors dominate. Moreover, this vasoconstrictive activity limits vasodilation in the subepicardial vessels and prevents epicardial steal. In healthy subjects, this stimulation does not necessarily lead to an increase in coronary vascular resistance, because metabolic counter-regulation may mobilize substantial vasodilatory reserve, especially in the subendocardial tissue [17]. Intramyocardial and subendocardial coronary arteries are dominated by beta1-adrenergic receptors [18-20]. The effect of beta-adrenergic stimulation of coronary vessels is not clearly defined because multiple effects are mediated by beta-receptors in vivo. beta-stimulation causes positive inotropy and chronotropy, which leads to increased myocardial metabolism and oxygen consumption. Because perfusion and contraction are coupled, an increased myocardial metabolism is followed by an increase in CBF via metabolic regulation [21]. The exact mechanism of metabolic regulation remains controversial; adenosine triphosphate-sensitive K+-channel openers [22], adenosine [23], prostaglandins [24], neuropeptides [25], and nitric oxide [26] are involved. An overview of these mechanisms exceeds the scope of this review. Counteracting the stimulation of beta-receptors with beta-blockade in resting dogs, however, results in a decreased CBF, whereas the oxygen delivery/oxygen consumption ratio remains constant. Decreased CBF might be the result of reduced myocardial oxygen consumption [27]. The effects of sympathetic stimulation in patients with coronary artery disease differ from those in healthy subjects. Nabel et al. [10] have demonstrated that constriction of atherosclerotic arteries is induced by the cold-pressor test, whereas smooth segments are dilated. This effect of sympathetic activation agrees with the finding in patients with classic angina that the diameter of the atherosclerotic artery decreases during exercise [28]. Activation of sympathetic influences can also override local metabolic vasodilation, which has been shown for adenosine infusion [29] or stress, induced by the cold pressor test [30]. This might be caused by activation of alpha-receptors [31], which restrict the metabolically related flow increase by approximately 30% [32]. Besides sympathetic reflexes, the endothelium plays an important role in the tone of coronary arteries [33]. In animal studies, it has been demonstrated that after removal of the endothelium, the relaxing effects of beta-adrenergic agonists are reduced and the constrictive effects of alpha-adrenergic agonists are enhanced [34,35]. Therefore, dysfunction of the endothelium can have further deleterious effects in the mediation of sympathetic activity. High TEA has the potential for blocking cardiac afferent and efferent fibers, which originate from the first to fifth thoracic level (T1-5) [36]. The perception of cardiac pain and angina is mediated via sympathetic afferent nerves. Stimulation of sympathetic efferents leads to an increase of inotropy, cardiac output, and systemic vascular resistance (Figure 2). Epidurally applied local anesthetics targeted to the T1-5 segments produce sensory blockade, motor blockade (depending on concentration), and blockade of the cardiac sympathetic fibers. In patients with coronary artery disease, it has been reported that TEA leads to a reduction in heart rate, cardiac output, and systemic vascular resistance, it may therefore decrease myocardial oxygen demand [37,38]. However, others reported increased heart rates with reduced cardiac output [39] or no change in either value [40,41]. The effect of TEA on left ventricular contractility has been the subject of several animal and clinical studies, but it still remains controversial. Contractility has been reported to be unchanged [42,43], reduced [44-46], or improved [47,48]. The variability of these results might be due to the different types of anesthetics used, whether epinephrine was added, the differing number of segments blocked, and species differences.Figure 2: Schematic illustration of the cardiac sympathetic innervation. a, Sympathetic stimulation in the thoracic region (Th1 to Th5) leads to an increase in heart rate, inotropy, and metabolism. b, Lumbar sympathetic innervation targets intraperitoneal and retroperitoneal organs and sympathetic stimulation of this area results in a decrease of motility.Various variables have been used to measure left ventricular function, including isovolumetric and ejection phase indices such as dP/dtmax, stroke volume, systolic time intervals, or ejection fraction. All these variables are highly dependent on cardiac loading conditions [49]. In conclusion, positive effects of TEA cannot simply be deduced from these investigations, and differences in patient populations must be taken in account when evaluating studies of TEA. Ideally, in patients at risk of ischemia, TEA should dilate constricted coronary vessels, decrease heart rate and myocardial metabolism, and improve cardiac function by reducing pre- and afterload and optimizing oxygen availability. In general, hemodynamic stability and a reduced stress response should provide an improved outcome, although hypotension can occur after epidural blockade of T1-S5 segments by diminishing sympathetic counter-regulation in a substantial vascular reservoir [50], which may offset the positive hemodynamic effects of TEA. Epidural Anesthesia and Pathologic States of the Heart It has been well established that activation of the sympathetic nervous system plays an important role in the pathophysiology of myocardial infarction, angina pectoris, and fatal cardiac arrhythmias [6,8,9]. Conversely, inhibition of sympathetic stimulation can reduce cardiac morbidity [7,51]. Selective inhibition of the sympathetic nervous outflow to the heart can be achieved by TEA, which blocks the segment T1-5, and although oxygen supply to the ischemic myocardium is improved, total CBF is unaltered. In experimental settings, the ratio of endocardial to epicardial blood flow is increased [43,52], and blood flow to ischemic regions is improved [52]. Under the influence of TEA, the size of the infarcted area was smaller both in subepicardial and subendocardial regions after experimental coronary occlusion in dogs (Figure 3) [44]. However, this investigation was not conclusive, because TEA decreased the rate-pressure product (RPP), which is usually beneficial in myocardial ischemia. A study with equivalent hemodynamic conditions induced by beta-adrenergic blockers has not been performed. After myocardial stunning, recovery was significantly faster in dogs when the ischemic insult leading to the stunning was induced during TEA under equivalent hemodynamic conditions (Figure 4) [43].Figure 3: The percentage of the cardiac circumference occupied by infarction in seven dogs after left anterior descending artery occlusion. A significant decrease of the infarcted area was found in both epicardial and endocardial regions during thoracic epidural anesthesia (TEA) [modified from [44]].Figure 4: Recovery of wall thickening fraction (WTF) from myocardial stunning in conscious dogs with and without thoracic epidural anesthesia (TEA). During the 48-h reperfusion period after a 10-min occlusion of the left anterior descending artery, dogs recovered significantly faster with TEA [modified from [43]].In humans, improvement in the myocardial oxygen supply by TEA may be related to the dynamic nature of approximately 75% of coronary stenoses [53,54], which can be modulated by pharmacologic or hemodynamic interventions [55]. By inducing a high-level TEA (T1-T6) with bupivacaine, Blomberg et al. [47] were able to increase the luminal diameter of stenotic coronary arteries in 64% of patients, whereas no effects were seen on nonstenotic segments. Studies have suggested that beta-blockers cause a direct constriction of coronary arteries by unmasking the effect of postjunctional alpha-receptors [56,57]. Because as nearly all patients in the study of Blomberg et al. [47] received beta-blocker-therapy, a decrease in the stimulation of alpha-adrenergic receptors may have increased blood flow through arterioles with diseased endothelium. Another indicator that increased blood flow is the main determinant in improving the ratio of oxygen supply and demand during TEA is the observation that in patients with unstable angina pectoris, high-level TEA reduced ST depression at a comparable workload during an exercise stress test [14]. In patients with stable angina pectoris, TEA reduced ST depression during exercise [14,48], although the 15% higher ejection fraction with TEA could not be explained entirely by lower RPP [48]. In addition, TEA has been used to treat anginal pain after multiple attempts to stop nitrate infusion have failed [14] and for long-term treatment of anginal pain [58]. Further support that the beneficial effects of TEA are not confined to hemodynamic changes is provided by an experimental study by Tsuchida et al. [59]. Brief coronary occlusions (5 min) lead to ST segment elevation and a concomitant decrease in pH, which is attenuated under TEA. This beneficial effect was also present when TEA-induced hemodynamic changes were corrected by blood transfusion or pacing. Intramyocardial levels of adenosine triphosphate, creatine phosphate, and lactate remained unchanged. Intramyocardial energy use is affected by cardiac rhythm, and the autonomic nervous system plays a critical role in its modulation. Several findings predict that TEA should protect against ventricular tachydysrhythmias and reentry supraventricular tachycardias, but that it may cause atrioventricular block [60,61]. In halothane-anesthetized dogs, the dose of exogenous epinephrine required to induce arrhythmia was significantly larger with TEA using mepivacaine (Figure 5) [62]. As this effect was not observed with intravenous (IV) mepivacaine, the elevation of the arrhythmogenic threshold is not an effect of systemically absorbed local anesthetic and therefore represents an effect of sympathetic blockade by epidural anesthesia. This result agrees with an earlier study in rats, in which the incidence of ventricular arrhythmias after coronary ligation with and without TEA was compared [63]. The incidence of ventricular flutter in the first 30 min after ligation of the left coronary artery was 53% in the control group compared with 20% in the TEA group. Studies on awake animals with control of hemodynamic variables are needed before a direct antiarrhythmic effect can be definitively attributed to TEA and human documentation is lacking.Figure 5: Arrhythmogenic doses of epinephrine (adrenaline) in the presence of various kinds of mepivacaine administration during halothane anesthesia in dogs [modified from [62]]. TEA = thoracic epidural anesthesia, iv = intravenous, MEPI = mepivacaine.Hypercoagulability, manifested by increases in fibrinogen, platelet activity, and plasminogen activator inhibitor-1 (PAI-1), often occurs after major surgery. It has also been implicated in the genesis of unstable angina and myocardial infarction [64]. Beneficial effects of epidural anesthesia on coagulation status have been described [64,65]. Most studies addressing this issue used lumbar epidural anesthesia (LEA), not TEA. The Perioperative Ischemia Randomized Anesthesia Trial Study Group (PIRAT) [11] reported a significantly reduced rate of graft occlusions in patients undergoing lower extremity revascularization who received epidural versus general anesthesia. In patients with general anesthesia, PAI-1 activity was greater 24 h after the operation compared with patients who received epidural anesthesia [64]. In patients undergoing major vascular surgery, Tuman et al. [12] found only one thrombosis in a patient after regional anesthesia versus 11 in the general anesthesia group. In the epidural group, platelet aggregation was reduced postoperatively. Beneficial effects of TEA can be caused by the systemic absorption of the local anesthetic, which leads to plasma concentrations efficient enough to directly impair platelet aggregation [66,67]. On the other hand, reduction of stress response with epidural anesthesia and analgesia may indirectly affect platelet function [68] and attenuate the catabolism of proteins involved in coagulation [12]. In addition, blood inflow and venous emptying is improved [69]. Lumbar or Thoracic Approach? An important question is whether beneficial effects can be achieved by LEA instead of TEA. Sympathetic neural blockade leads to numerous effects on the vascular bed. Measurement of sympathetic impulses in postganglionic fibers to skin and muscle of the legs has shown a complete block of either spontaneous or induced sympathetic neural activity during epidural anesthesia [70]. Epidural blockade of lumbar segments resulted in an increased sympathetic activity in splanchnic nerves due to baroreceptor drive, but only TEA produces near total ablation of splanchnic activity [71,72]. Both techniques have an inherent risk of cardiovascular depression and hypotension. The most frequent cardiovascular complications, e.g. vasovagal syncope and arterial vasodilation, are usually treated successfully with atropine and IV volume replacement, as well as vasopressor drugs when indicated. Bradycardia due to the Bezold-Jarisch reflex is often refractory to atropine treatment and the administration of potent vasopressors. Patients experience hypotension and profound bradycardia associated with venous blood pooling and heightened cardiac contractility [73]. This results in reflex arterial dilation and vagally mediated bradycardia. Because of the extent of the sympathetic blockade and vasodilation, this reflex occurs more often with LEA than with TEA. In an experimental setting in swine, the injection of bupivacaine for LEA with blockade to the thoracic level induced a severe reduction of myocardial blood flow distal to a coronary artery stenosis [74]. The reduction in oxygen supply due to hypotension during LEA is not followed by a concomitant reduction in demand [38]. This might even be aggravated by an enhanced sympathetic reactivity in sympathetically intact areas [75-77], which has recently been demonstrated for LEA and TEA in cats [78]. During TEA, the decrease in CBF seems to be compensated by a decrease in myocardial oxygen demand and cardiac work [38]. Similar reductions in MAP are not followed by segmental wall motion disturbances during TEA [38], but they are during LEA [79]. Kock et al. [48] found improved cardiac performance during exercise with TEA compared with control exercise. Ischemia-induced left ventricular global and regional wall motion abnormalities were improved under TEA and were associated with less pronounced ST segment depression. In conclusion, TEA has more favorable effects than LEA in the patient with coronary artery disease. Epidural Anesthesia in Noncardiac Surgery TEA has several benefits in myocardial ischemia: anginal pain is improved, even when other therapies failed; the stress response to surgery is suppressed; and the incidences of myocardial infarctions and arrhythmias may be reduced. Unfortunately, there have been few studies that focus on the effect of TEA in noncardiac surgery. Thus, an evaluation should consider epidural anesthesia in general. Although the often-quoted study published by Yeager et al. in 1987 [80] can be criticized because of methodological drawbacks, its surprising results sparked further investigations (Table 1). Fifty patients undergoing thoracic, abdominal, or vascular surgery were randomized into two groups. The first group received combined epidural and general anesthesia, the second received general anesthesia alone. The number of severe cardiovascular complications was significantly higher in the group with balanced general anesthesia alone. There were four deaths in this group more than 2 wk postoperatively, whereas all patients in the combined anesthesia group survived. Three myocardial infarctions occurred in the general anesthesia group versus none in the combined anesthesia group.Table 1: Studies Comparing the Influence of General Versus Regional Anesthesia or Analgesia on OutcomeThe level of epidural anesthesia (lumbar or thoracic) was not controlled, and neither the techniques of general anesthesia nor the postoperative pain therapy was standardized or specified. The study population was heterogenous with respect to severity and invasiveness of surgery, and patients were not stratified according to concomitant diseases. The high incidence of severe complications and high mortality have cast doubt on the validity of this study. The results obtained by Yeager et al. were not confirmed in a better designed study by Baron et al. [81] of 173 patients undergoing aortic reconstruction. The patients were randomized in two groups receiving either a balanced general anesthesia or TEA in combination with light general anesthesia. Ejection fraction, associated coronary artery disease, hemodynamic variables, and cardiovascular treatment were equally distributed in both groups. All patients received a thoracic epidural catheter, which was not used intraoperatively in the general anesthesia group. There was no difference between groups concerning cardiovascular complications, infections, and mortality. Postoperative pain therapy, unfortunately, was not controlled or randomized; therefore, no definitive conclusions regarding the perioperative period can be based on this study. The effects of epidural anesthesia during the postoperative period was investigated in a study by Tuman and coworkers [12]. They randomized 80 patients suffering from atherosclerosis and undergoing major vascular surgery to compare the effects of general anesthesia with on-demand narcotic analgesia or general anesthesia combined with epidural anesthesia and analgesia. Patients in the latter group had high lumbar or low thoracic (L3-T10) epidural catheters. The concomitant diseases were comparable in both groups, except for a significantly higher rate of diabetes mellitus and previous myocardial infarction in the combined anesthesia group. The rates of cardiovascular, infectious, and overall postoperative complications, as well as the duration of the intensive care unit stay, were significantly reduced in the combined anesthesia group. A study of the combined use of epidural and general anesthesia versus general anesthesia alone was performed by the PIRAT group [11] in patients undergoing lower extremity surgery. LEA catheters were inserted at L2-3 or L3-4. No differences were discovered with regard to cardiac, renal, or pulmonary complications. However, the need for reoperation was reduced in the epidural group. A more recent study by Bode et al. [13] found no statistically significant difference in the effects of lumbar regional (epidural or spinal) versus general anesthesia on perioperative cardiac morbidity and overall mortality in patients undergoing peripheral vascular surgery. Although this study is a bit out of the context of this review, it is noted because it provided an interesting finding. The cardiac death rate associated with inadequate regional anesthesia requiring conversion to a general anesthetic was significantly higher compared with that of patients undergoing successful regional or general anesthetic. Recovery from epidural anesthesia is of clinical importance. Sprung et al. [82] reported two patients who developed cardiac arrest caused by coronary spasm during recovery from epidural anesthesia. They concluded that this might be caused by imbalances between the parasympathetic and sympathetic nervous systems. Although current data do not definitively establish the benefit of intraoperative LEA, additional investigation is clearly required to clarify the role of epidural analgesia postoperatively, especially when administered via the thoracic spinal segments. TEA in Cardiac Surgery Patients undergoing coronary artery bypass grafting (CABG) have an increased risk of perioperative cardiac complications [83], and, thus, strategies to reduce the perioperative risk have been the focus of several studies. Liem and coworkers [84-86] published a series on the influence of TEA in 30 patients undergoing CABG. TEA was administered with bupivacaine 0.375% and sufentanil 5 micro g/mL in combination with general anesthesia with midazolam/N2 O. Epidural block extended from T1 to T10 segments. Hemodynamic results, postoperative outcome, and adrenergic responses were compared with a control group receiving large-dose IV sufentanil and midazolam. Hemodynamic stability and lower heart rate resulted in the reduced use of nitroglycerin for the treatment of myocardial ischemia in the TEA group (Figure 6) [84]. Postoperatively, patients who received TEA during and after CABG awoke earlier and were extubated sooner. This effect of TEA was first described by Joachimsson et al. [87], who were able to tracheally extubate patients within 2 h after CABG. These patients were less sedated, and pain relief was better, than with IV analgesia [85]. Variables of pulmonary and cardiac function were improved. Investigating changes in stress hormones in 20 patients, Liem and coworkers [86] found less increase in norepinephrine and less variability in epinephrine plasma concentrations when general anesthesia was combined with TEA during CABG. Cortisol release was higher during the bypass period in the general anesthesia/TEA group than in the general anesthesia group. During the first and second postoperative days, lower epinephrine and cortisol plasma levels were found in the TEA group [86].Figure 6: a, Incidence of cardiac complications in the 24-h postoperative period after coronary artery bypass graft with general anesthesia and general anesthesia plus thoracic epidural anesthesia (TEA). b, Requirement of inotropic and vasodilating drugs. Tachycard = tachycardia, Dobut = dobutamine, NTG = nitroglycerine, SNP = sodium nitroprusside. Data taken from [85].Studies performed more recently confirmed the results of Liem and coworkers. Stenseth and colleagues [88] compared the effects of TEA as an adjunct to small- or large-dose fentanyl versus large-dose fentanyl in 30 patients undergoing CABG. TEA was induced by 10 mL bupivacaine 0.5% followed by 4 mL every hour. Although the authors were not able to test the extension of the block due to heavy sedation, they postulated "from experience with TEA" a block extending from T1 to T12. Hemodynamic stability in both TEA groups was again better, with a reduced need for propanolol or nitroglycerine in both TEA groups. In another study, Kirno and coauthors [89] described an attenuation in the hemodynamic responses to surgery by inducing epidural blockade extending from T1 to T5 segments with 3-3.5 mL of 2% mepivacaine as an adjunct to a standard fentanyl/nitrous oxide anesthetic. Better hemodynamic stability and earlier tracheal extubation in patients receiving small-dose fentanyl in conjunction with bupivacaine for TEA compared with large-dose fentanyl alone was again found by Stenseth et al. [90]. They also confirmed better postoperative pulmonary function, and patients in the TEA group were tracheally extubated earlier (P < 0.05). The dose of the local anesthetic was 2- to 2.5-fold higher in the studies of Liem et al. and Stenseth et al. compared with those of Kirno et al. and Blomberg et al. Achievement of a sufficient but not too extended block using smaller doses of local anesthetic seems to be preferable. Efforts to prevent endocrine or metabolic responses to surgery and to reduce perioperative morbidity have been the focus of a number of studies [80,91]. The neuroendocrine response to surgery, trauma, and myocardial infarction leads to an increase in the plasma concentration of free fatty acids, thereby increasing myocardial oxygen consumption [92,93], which may contribute to myocardial ischemia and infarction in patients with reduced coronary reserve. Hotvedt and coworkers [45] demonstrated a decrease in the plasma concentration of free fatty acids during TEA. The related decrease in myocardial oxygen consumption may have beneficial effects under ischemic conditions. Catecholamine response, reflected by epinephrine and norepinephrine release, is abolished or attenuated under TEA. In the study of Kirno et al. [89], norepinephrine release was reduced in the TEA group. Stenseth et al. [94] found less epinephrine and norepinephrine release in patients with TEA as an adjunct to large- or small-dose fentanyl for CABG. Similar results were obtained by Moore et al. [95], who showed an abolished catecholamine response in patients receiving TEA with bupivacaine in addition to IV sufentanil for cardiac surgery. Another study by Thorelius and colleagues [96] failed to show any benefits of TEA for cardiac surgery, but the investigators examined only two distinct measurements during the procedure, precluding the formation of any definitive conclusion from this study. It is generally accepted that the reduction of the catecholamine response has beneficial effects. However, it seems that a decrease of cortisol levels may be associated with detrimental effects on outcome. McLedingham and Watt [97] observed a higher mortality rate in trauma patients under sedation with etomidate. Later, it was shown that continuous infusion of etomidate inhibits cortisol synthesis [98,99]. Moreover, Jansen et al. [100] even reported improved hemodynamic stability after cardiopulmonary bypass and the postoperative course when additional corticoids (dexamethasone) were administered before bypass. The impact of TEA on cortisol synthesis remains to be investigated. In addition, no study of TEA has extended for longer than 72 h postoperatively. Defining the effect of prolonged postoperative use of TEA on perioperative outcome in patients at high risk of myocardial infarction on the third or fourth day postoperatively [101,102] also requires additional study. TEA and Lung Function Because as reduction of myocardial oxygen supply might be exacerbated by hypoxemia, the effect of TEA on lung function must be considered. Impaired ventilatory function, as reflected by reduced vital capacity, occurs after abdominal as well as after thoracic surgery [103-106]. Hypoxemia occurs postoperatively, especially during sleep [107-109], and is associated with myocardial ischemia [107]. Impairment of lung function starts with induction of general anesthesia and lasts for 1-2 wk postoperatively. An important approach in reducing postoperative pulmonary dysfunction is the improvement of analgesia by epidural application of local anesthetics or opioids. For opioids, the epidural route provides better recovery of pulmonary function compared with IV administration [110,111]. Periods of oxygen desaturation are reduced during epidural compared with IV anesthesia [109]. However, pain does not seem to be the only determinant of ventilatory deterioration postoperatively. The inhibition of diaphragmatic function despite profound analgesia plays an important role in ventilatory impairment [112-114]. The precise mechanism of diaphragmatic dysfunction is unknown; one hypothesis suggests that stimulation of afferent nerves in the chest and abdominal walls, viscera, and diaphragm leads to an inhibition of phrenic motor drive [103]. TEA improves postoperative diaphragmatic function in lambs [115] and humans [116], and improves respiratory function variables [117]. In contrast to these beneficial effects, TEA has potentially negative effects by paralysis of the thoracic muscles [115]. The denervation of the musculature of the rib cage theoretically alters lung volume. Measurements of functional residual capacity before and after induction of TEA in subjects without lung disease did not reveal any changes in this muscle tone-dependent lung volume [118-120]. In elderly patients, Sakura et al. [121] found a significant decrease of 13% in minute ventilation and a 14% decrease in tidal volume after induction of TEA, but no impairment in the response to hypercapnia or hypoxia. When LEA is extended to high thoracic levels (T1), functional residual capacity is significantly increased [122] and is associated with a reduction in intrathoracic blood volume after peripheral vasodilation. As with effects on cardiac function, the number of spinal nerve segments blocked seems to play an important role in the effects of epidural analgesia on lung function. The pulmonary vasculature is innervated by the autonomic nervous system, and hypoxic pulmonary vasoconstriction may be influenced by sympathetic neural blockade. Hypoxic pulmonary vasoconstriction, an important mechanism for maintenance of adequate oxygenation after induction of general anesthesia [123], is not affected by TEA in dogs [124]. In clinical studies, a positive effect of TEA during anesthesia is mainly present in patients at high risk of postoperative pulmonary complications, but such an effect may also be observed when adequate epidural analgesia is maintained into the postoperative period [5]. An important issue is the control of dynamic pain, i.e., adequate analgesia during movement in contrast to analgesia only at rest.1 This is important for pulmonary function, and in patients receiving epidural versus IV patient-controlled analgesia treatment, pulmonary function was significantly improved. Current data do suggest that TEA reduces postoperative pulmonary complications. Reducing hypoxemic episodes in the postoperative period may help to reduce the incidence of myocardial ischemia in high-risk patients. (1) Bell S. The correlation between pulmonary function and resting and dynamic pain scores in postaortic surgery patients [abstract]. Anesth Analg 1991;72:S18. Risks and Benefits from TEA Any invasive medical procedure with an inherent risk mandates a thorough assessment of the risk/benefit ratio. The most common complication of epidural anesthesia is accidental dural perforation. The incidence using the loss of resistance method is approximately 0.6% [126-128]. Reports on the incidence of neurological complications vary greatly. The incidence of paresthesias and neurologic injuries is approximately 0.01%-0.001% [126,129]. The most disastrous complication is paraplegia after development of a epidural hematoma [130,131]. The occurence of this complication is rare; no cases were reported in 100,000 patients receiving central neuraxial anesthesia or analgesia [132] or 4,185 patients receiving TEA [128]. However, 3 paraplegias in 18,000 epidurals were caused by spinal hematomas in patients with deranged hemostatic capacity [130]. Coagulation status is a problem especially applicable to cardiac patients. Patients at cardiac risk often receive anticoagulant therapy preoperatively or perioperatively. The effects of aspirin, a common anticoagulant, last for approximately 7-10 days. Assessment of bleeding time may provide some information about aspirin effects [133]; for elective surgery, consideration should be given to stopping aspirin therapy 1 wk before surgery [134]. Low molecular weight heparins are often used in the postoperative period. In this case, it is advisable that the anticoagulant be given 10-12 h preceding TEA (e.g., on the evening before surgery) or at least 1-2 h after the initiation of TEA [135]. Similarly, epidural or subarachnoidal catheters should probably be removed at least 10-12 h after the previous injection of low molecular weight heparins or immediately (1-2 h) before the next dose [135]. The issue of intraoperative anticoagulation requires discussion. A time interval of 1-2 h has been recommended for beginning systemic anticoagulation after a bloody tap [135]. Another method favored at our own institution, and in the studies by Liem et al. [84] and Stenseth et al. [88], is to increase the time interval to between 16 and 24 h by inserting the catheter on the day before surgery. There were no neurological problems in these studies. It is also recommended that catheter removal be performed when coagulation variables are within normal ranges. Vandermeulen et al. [135] concluded, "the correct approach to this dilemma cannot be dogmatic" and "continous awareness…should enable us to make anesthetic practice safer without withholding anesthetic techniques from patients who would most certainly benefit from them." Conclusions Earlier editorial comments about the optimal anesthetic regimen for peripheral vascular surgery stated that the question of the "best anesthetic management for this group of these seriously ill patients" remains unresolved [136], and more recently, it has been opined that further studies were unlikely to be useful [137]. However, these editorials were based on studies that had some limitations: the use of LEA and/or the absence of postoperative epidural analgesia. A number of studies, however, now support the use and initiation of further studies of TEA in patients with compromized cardiac function to obtain a definitive answer to this important question. TEA improves cardiac performance and may even have beneficial effects on the oxygen delivery/demand ratio. TEA has beneficial effects on several variables (catecholamines, nitrogen balance, coagulation) that correlate with improved patient outcome, although definite proof for better outcome has not been documented. Lung function may also benefit from TEA. Outcome benefit from TEA can be reasonably expected only when its intraoperative use is extended to the postoperative period. Considering the level and extent of epidural anesthesia, data indicate that TEA is superior to LEA in patients with compromised cardiac function. Questions still remain regarding the application of TEA in anticoagulated patients. Based on current knowledge, application of TEA in the face of anticoagulation should be undertaken with caution and with consideration of the caveats outlined above. The optimal duration and analgesic regimen of TEA in the postoperative period also remains undefined. Finally, the optimal combination of general anesthesia (inhaled/IV) with intraoperative regional anesthesia has yet to be established. The authors thank Dr. Sture Blomberg for his suggestions and constructive criticism during the preparation of this manuscript.