Chapter 22 Intraoperative and immediate postoperative management

Overview

Improvements in patient selection criteria and advances in hepatic surgical techniques and perioperative care have expanded the pool of patients who are candidates for hepatobiliary surgical procedures. Patients with known underlying hepatic parenchymal disease are at greater operative risk, particularly for liver-related and biliary tract–related procedures (see Chapters 2, 70B, and 90A). More recent studies have documented the progressive safety of liver resection in noncirrhotic patients (Belghiti et al, 2000; Jarnagin et al, 2002; Mullen et al, 2007). The anesthetic problems in connection with liver transplantation (Chapter 97B) and cirrhosis (Chapter 90A) are discussed elsewhere in this book. This chapter addresses the unique anesthetic considerations of noncirrhotic patients undergoing hepatobiliary operations, with an emphasis on problems related to joint anesthetic and surgical management.

Preoperative Evaluation

With the advent of modern techniques, the anesthetic risk attributed to any surgical procedure depends heavily on the preoperative status of the patient. In hepatobiliary patients, the presence of preoperative pulmonary, cardiac, or hepatic disease significantly augments the incidence of postoperative morbidity and mortality. Although in the past, the presence of comorbidities favored elective intensive care unit (ICU) admission, currently they more appropriately designate a select group of patients who may potentially fail fast tracking. These patients need to be recognized at the preoperative visit and clinically optimized before surgery. Postoperative morbidity adversely affects long-term outcome after hepatic resection, therefore efforts aimed at reducing perioperative morbidity will likely further enhance the therapeutic benefit of resection (Ito et al, 2008).

Cardiac Evaluation

Patients undergoing liver surgery are at significant risk of cardiovascular morbidity and mortality. Although the perioperative event rate has declined as a consequence of better anesthetic and surgical techniques, perioperative cardiac complications remain a significant problem. The first step in preoperative care is to identify patients at risk for perioperative cardiac events. The risk of perioperative cardiac complications is the summation of an individual patient’s risk and the cardiac stress related to the surgery. Clinical history, physical examination, and review of the electrocardiogram usually provide enough data to estimate cardiac risk. Active cardiac conditions may preclude all but emergent surgery.

For most stable patients, cardiac risk can be derived from the Revised Cardiac Risk Index (Lee et al, 1999b). This simple index identifies six independent risk factors: 1) high-risk type of surgery, 2) history of ischemic heart disease, 3) history of congestive heart failure, 4) history of cerebrovascular disease, 5) diabetes mellitus, and 6) renal insufficiency. The number of clinical risk factors, the risk level of the surgery, and the functional capacity of the patient will determine the need for further testing prior to surgery or pharmacologic intervention perioperatively (Poldermans et al, 2008). For those patients with clinical risk factors, the 2007 American College of Cardiology/American Heart Association guidelines on perioperative cardiovascular evaluation and care for noncardiac surgery are an excellent framework for evaluating cardiac risk in the perioperative period for patients undergoing liver surgery (Fleischer et al, 2007).

Over the years, a shift in perioperative management has been seen, away from treating coronary obstruction with interventions such as surgery and stents and toward medical therapy aimed at stabilizing coronary plaque and preventing myocardial oxygen supply and demand mismatch. Today, preoperative cardiac testing, cardiac stenting, and coronary revascularization are performed for the same indications as those used in the nonoperative setting (Poldermans et al, 2008), but β-blockers, statins, and aspirin are now widely used in the perioperative setting. Compelling evidence supports perioperative β-blocker therapy, unless contraindicated, in patients with coronary artery disease. β-Blocker therapy has been shown to reduce the incidence of perioperative ischemic events and myocardial infarction (Goldman, 2001; Lee et al, 1999a). Recently the risk benefit of this therapy in patients without coronary artery disease has been challenged. The perioperative ischemic evaluation (POISE) study showed an increased incidence of ischemic stroke in combination with intraoperative bradycardia and hypotension (Devereaux et al, 2006). In a retrospective review of patients who underwent noncardiac surgery, perioperative β-blockers showed no benefit—and possible harm—in low-risk patients, but in high-risk patients, a benefit was seen (Lindenauer et al, 2005).

Communication among the patient’s cardiologist, surgeon, and anesthesiologist is essential for the management of hepatobiliary surgical patients who have coronary artery stents. Clinical judgment is of the utmost importance to balance the risk–benefit ratio of antiplatelet therapy interruption versus continuation. Aspirin should never be interrupted, unless the risk of bleeding far outweighs the risk of stent thrombosis (Newsome et al, 2008).

In hepatobiliary patients with a history of alcohol abuse, cardiac assessment should include the evaluation of myocontractile function. Cirrhotic cardiomyopathy classically occurs in men 30 to 55 years old who have been alcohol abusers for more than 10 years (Braunwald, 1997). Putative mechanisms of damage are direct toxic effects of 1) ethanol and its metabolite acetaldehyde; 2) nutritional deficiencies, particularly of thiamine (beriberi); and 3) additives in commercial products, such as cobalt in beer. Alcohol and acetaldehyde have been shown to decrease calcium binding and transport, myocardial lipid metabolism, myocardial synthesis, and myocardial adenosine triphosphatase (ATPase) activity. The two basic patterns of cardiac dysfunction are 1) left ventricular dilation with impaired systolic function and 2) left ventricular hypertrophy with diminished compliance and normal or increased contractile performance. Presenting manifestations range from insidious onset to catastrophic left-sided cardiac failure. Anginal chest pain does not occur, unless the patient has concomitant coronary artery disease or aortic stenosis.

Pulmonary Evaluation

Despite steady advances in care, patients with respiratory disease remain at increased risk for postoperative pulmonary complications (PPCs). Indeed postoperative pulmonary complications may rival cardiovascular complications in frequency and severity (Lawrence et al, 1995). Studies that examine risk factors for PPCs have limitations, but some consistent patterns are seen. The three most important risk factors for PPCs are 1) pulmonary disease, 2) cigarette smoking, and 3) the site of surgery, with abdominal surgery being one of the highest risk sites (Warner, 2000). Of note, pulmonary function testing per se is not useful in predicting risk (McAlister et al, 2003); thus routine pulmonary function testing is not justified unless it is part of an effort to optimize preoperative pulmonary status. The definition of “optimization” depends on the type of respiratory disease and the individual patient. Consultation with the surgeon and pulmonologist will help determine whether a course of systemic corticosteroids or antibiotics is warranted perioperatively. Despite the increased risk of PPCs in patients with preexisting pulmonary disease, no prohibitive level of pulmonary function has been established for which surgery is contraindicated (Nunn et al, 1988; Kroenke et al, 1993). A prospective study did not find an elevated Paco₂ level to be a risk factor among surgical candidates, hence clinicians should not use arterial blood gas analyses to identify patients for whom the risk of surgery is prohibitive (Kearney et al, 1994).

Obese patients are at risk of suffering from a number of respiratory derangements, including obstructive sleep apnea (OSA), obesity hypoventilation syndrome, and restrictive impairment. The increase in body mass also results in increased oxygen consumption and carbon dioxide production. With these isssues in mind, it is not surprising that acute PPCs are twice as likely in obese and OSA patients (Rose et al, 1994; Dindo et al, 2003). Many patients with OSA are undiagnosed, but a strong relationship exists between obesity and OSA (Strohl et al, 1996). The American Society of Anesthesiologists addressed this issue with practice guidelines to assess patients for possible OSA before surgery and to provide careful postoperative monitoring for those suspected to be at risk (Gross et al, 2006).

An association was also found between obesity and OSA and several medical conditions, such as increased risk of venous stasis, pulmonary embolism, hypertension, cerebral vascular acccidents, cardiomyopathy, arrythmias, and ischemic heart disease (Dominguez et al, 1993; Strohl et al, 1996; Gupta et al, 2001). With these issues in mind, the anesthesiologist should have an informed discussion with the patient about the increased risk of morbidity and mortality and should also work with other members of the patient care team to determine whether any interventions need to be initiated before surgery in an effort to minimize complication risk.

Polysomnography is the gold standard for diagnosis of OSA, but it is an expensive and limited resource. The most reasonable approach is to check room air pulse oximetry. If the patient has an oxygen saturation level less than 96%, further evaluation is warranted; a 2-week period of continuous positive airway pressure therapy has been shown to be effective in correcting abnormal ventilatory drive and improving cardiac function (Lin, 1994; Loadsman et al, 2001). In addition, deep venous thrombosis prophylaxis should be discussed with the surgeon.

Hepatic Evaluation

Risk factors and symptoms of liver disease are not as well defined as those in other organ systems. The diagnosis of liver disease requires a high degree of suspicion and careful probing of the clinical history to identify specific risk factors for liver disease, such as previous blood transfusions, jaundice, travel, tatoos, high-risk sexual behavior, illicit drug use, excessive alcohol intake, and chemotherapy (Suman & Carey, 2006).

In North America and Europe, most hepatic resections are performed for metastatic disease. The operative outcome for hepatobiliary surgery is linked to severity of hepatic parenchymal disease and the extent of postoperative functional reduction (see Chapter 2). In patients with mild or well-compensated chronic hepatic disease, operative outcome is likely to be indistinguishable from the outcome in the general population. Patients with hepatocellular carcinoma (HCC) have poorer outcome than those with metastatic disease given the propensity of HCC to develop in the cirrhotic liver (Fattowich, 2003).

The goal of preoperative screening is to determine the presence of preexisting liver disease without the need for extensive or invasive monitoring. Liver function tests can measure different aspects of hepatic function, but as a group of tests, they lack specificity and are often affected by nonhepatic function. These biochemical markers cannot quantify the hepatic disease. Another issue is finding abnormal hepatic function test results in asymptomatic patients. In general, for asymptomatic patients with mildly elevated alanine and aspartate aminotransferase levels and a normal bilirubin concentration, cancellation of surgery is rarely indicated. In patients with significant abnormalities, detailed investigation is warranted to evaluate for underlying cirrhosis, given the high perioperative risk observed in patients with cirrhosis. The need for further investigation in these patients is highlighted by studies that report an incidence of undiagnosed cirrhosis of 34% in asymptomatic patients with abnormal liver function results (Hay et al, 1989; Hultcrantz et al, 1986). It is important to identify patients with underlying hepatic impairment, because they may be at increased risk for further injury from alterations in splanchnic blood flow that occur during anesthesia and intraabdominal surgery.

Measurement of metabolic liver function would provide the necessary predictive information to assess this risk. Metabolic liver function tests (see Chapter 2), such as the aminopyrine breath test and indocyanine green clearance, may have predictive value in this regard but are not routinely performed. To date, risk assessment has been primarily performed in patients with known liver disease (Child-Turcotte-Pugh score), but the quantification of risk in asymptomatic patients should not be extrapolated from these studies. The risk of postoperative hepatic insufficiency is closely related to the volume and function of the remnant liver, and several methods for liver volume determination are available. Most use computerized tomography (CT) combined with three-dimensional CT volumetry (Denys et al, 2002). When the functional liver remnant volume is below a certain threshold, generally 25% to 30% for normal liver and 40% for diseased liver, the risk of postoperative hepatic failure is high, and portal vein embolization may be indicated prior to hepatic resection (Vauthey et al, 2000).

Bloodless Surgery

In the early years of hepatic resection, major intraoperative blood loss was common, transfusion of blood products was routine, and morbidity and mortality were unacceptably high (Foster & Berman, 1977). General improvements in hepatic resection techniques have reduced intraoperative blood loss and transfusion rates; nevertheless, perioperative transfusion remains a potent predictor of increased perioperative morbidity and mortality (Jarnagin et al, 2002). In recent years, liver resection without the need for blood transfusion has become increasingly possible (Belghiti et al, 2000; Cunningham et al, 1994; Fan et al, 1999; Jarnagin et al, 2002). Patient blood management is based on three pillars: 1) detecting and treating preoperative anemia, 2) reducing the loss of red blood cells perioperatively, and 3) optimizing the treatment of anemia (Spahn et al, 2008).

Thorough preoperative planning is essential to avoid perioperative allogeneic transfusion. Any history of bleeding disorders and management of anticoagulation must be evaluated, including discontinuation of drugs that adversely affect clotting (e.g., acetylsalicylic acid, nonsteroidal antiinflammatory drugs, and anticoagulants). In patients with anemia, recombinant human erythropoietin or iron therapy may optimize the starting operative hemoglobin (Monk, 1996). Preoperative autologous donation also has been used to reduce the need for allogeneic red blood products (Brecker & Goodnough, 2001). Preoperative autologous donation may not avoid allogeneic blood, however, because almost half of the patients who donate blood before surgery are anemic on the day of operation, and preoperative strategies to augment the red blood cell mass require more time than is generally reasonable for optimal efficacy. Patients with low hemoglobin levels at the start of surgery are at an increased risk of receiving allogeneic blood (Armas-Loughran et al, 2003). In addition, preoperative autologous donation is costly; it can be associated with clerical errors; and for every 2 units donated, usually only 1 unit gets transfused (Goldman et al, 2002). If the patient is optimized and surgery is bloodless, the autologous units are discarded. Other blood conservation strategies, such as intraoperative blood salvage (cell saver) and acute normovolemic hemodilution (ANH) have been used successfully for hepatobiliary surgery in Jehovah’s Witness patients (Jabbour et al, 2005). ANH has been shown in two prospective studies to reduce the amount of red blood cells transfused per patient in major liver resections (Matot et al, 2002a; Jarnagin et al, 2008).

Intraoperative Management

Hepatic blood flow in surgical patients can be altered by a variety of factors, including arterial blood pressure, posture changes, carbon dioxide levels, intravascular volume shifts, positive pressure ventilation, positive end-expiratory pressure, regional anesthesia, and volatile anesthetics (Gelman, 1992). Surgical stimulation and manipulation of the liver markedly increases hepatic oxygen extraction ratio and splanchnic vasoconstriction (see Chapter 4) (Kainuma et al, 1991; Whittle & Moncada, 1998). A 16% decrease in hepatic blood flow is associated with anesthesia and mechanical ventilation (Gelman, 1976), and a further 40% decrease is seen with splanchnic surgery. This decrease may be the direct effect of sympathetic control of the hepatic venous bed mediated through the hepatic innervation (Greenway et al, 1986).

Positive end-expiratory pressure decreases hepatic blood flow in a stepwise manner. The response cannot be explained by a decrease in cardiac output alone; investigators postulate a vasoconstrictive response in the preportal vasculature. Intraperitoneal insufflation and head-up tilt result in impairment of hepatic blood flow secondary to decreases in cardiac output (Berendes et al, 1996b; Eleftheriadis et al, 1996). The hepatic blood flow decrease caused by hypotensive spinal anesthesia is attenuated by ephedrine administration (Nakayama et al, 1993). Two human studies using the indocyanine green plasma disappearance rate to evaluate the effects of high lumbar epidural anesthesia (EDA) on hepatic blood flow have found a decrease of 25% to 35% despite a constant cardiac output (Kennedy et al, 1971; Tanaka et al, 1997). Patients treated with norepinephrine to compensate for the EDA-induced decrease in arterial blood pressure had an additional marked decrease in hepatic blood flow (Meierhenrich et al, 2009).

It is controversial whether catecholamines increase hepatic blood flow and oxygen supply, but dobutamine does not cause a significant increase in hepatic arterial blood flow (Kainuma et al, 1992). Total hepatic blood flow and portal venous blood flow are increased, however, resulting in an augmentation of hepatic oxygen delivery. The increase in hepatic oxygen delivery is countered by an increase in hepatic oxygen uptake with no overall improvement in hepatic oxygen supply–uptake ratio. Dopamine and norepinephrine increase hepatic venous oxygen saturation, suggesting that the vasoactive treatment of patients in septic shock may not compromise splanchnic oxygenation (Ruokonen et al, 1991, 1993). Hepatic portal flow does not decrease despite a 20% to 60% reduction in blood pressure as long as cardiac index is maintained during sodium nitroprusside hypotension (Chauvin et al, 1985). A well-planned anesthetic maximizes the relationship between oxygen transport and oxygen use with the premise that reductions in systemic pressure would reduce hepatic blood flow. Anything resulting in significant reductions of systemic pressure and cardiac output–induced hypotension, hypovolemia, and anesthetic overdoses should be avoided.

Volatile Anesthetics

Anesthetic Hepatotoxicity

Fulminant hepatic necrosis and jaundice after halothane, so-called halothane hepatitis, is rare—one in 6000 to 35,000 anesthetics—but often fatal. Halothane hepatitis is an immunologic phenomenon initiated by halothane metabolism and the binding of its metabolite to liver proteins to form trifluoroacetylated proteins, which stimulate antibodies in susceptible individuals. Upon subsequent halothane reexposure, these antibodies mediate massive hepatic necrosis. Because the extent of metabolism of enflurane, isoflurane, and desflurane is so much less than that of halothane, fulminant hepatitis from these anesthetics is far less common than with halothane (Elliot & Strunin et al, 1993). In 1987, the U.S. Food and Drug Administration determined that there was no conclusive association between isoflurane exposure and postoperative hepatitis (Shingu et al, 1983; Stoelting et al, 1987). Sevoflurane metabolism is different from that of the other volatile anesthetics, however, because it does not result in trifluoroacetyled liver proteins, and immune-based hepatitis after sevoflurane has not been reported. With the disappearance of halothane and enflurane from clinical practice in developed countries, and lack of hepatotoxicity from either isoflurane, desflurane or sevoflurane, anesthetic volatile hepatotoxicity is not a significant concern (Elliot & Strunin, 1993).

Hemodynamics

Volatile anesthetics reduce hepatic blood flow in a dose-dependent fashion by affecting cardiac output and systemic pressure. Surgical manipulation causes an additional decrease in estimated hepatic blood flow. Isoflurane has been considered the agent of choice in cases in which preservation of splanchnic blood flow is required, and liver blood flow and the hepatic artery buffer response are maintained better in the presence of isoflurane than with any other volatile anesthetic agent (Berendes et al, 1996a). In addition, isoflurane is shown to attenuate the increases in hepatic oxygen consumption associated with surgery and liver manipulation.

Desflurane is shown to have no deleterious effects on liver function and hepatocyte integrity, and it is associated with significantly greater gut blood flow than equipotent isoflurane. This difference cannot be explained by systemic hemodynamics alone. There is no difference in total hepatic flow between isoflurane and desflurane groups, however, which suggests that an intact hepatic arterial supply buffers response with desflurane (O’Riordan et al, 1997). Sevoflurane is similar to isoflurane and desflurane with a few exceptions (Ebert et al, 1995). Indocyanine green clearance is better preserved during sevoflurane anesthesia. In Beagles subjected to sevoflurane anesthesia and ligation of the hepatic artery, hepatic oxygen supply–uptake ratio, hemoglobin oxygen saturation, and oxygen partial pressure in hepatic venous blood were significantly lower than with halothane or isoflurane.

Nitrous oxide is used extensively in patients with hepatic disease (Khalil et al, 1994; Lampe et al, 1990). Although it has not been shown to contribute to hepatic disease exacerbation, the sympathomimetic effects of nitrous oxide may increase hepatic metabolic requirements.