Overview

Patients with cirrhosis or decompensated liver disease (see Chapter 2, Chapter 70B, Chapter 72, Chapter 73, Chapter 74, Chapter 75A, Chapter 75B, Chapter 75C ) are at particularly high risk for perioperative morbidity and mortality after major surgery, often necessitating admission to the intensive care unit (ICU). Compared with noncirrhotic patients, patients with cirrhosis who are admitted to the ICU have higher rates of complications and hospital mortality, extended lengths of stay, and greater resource use (Aggarwal et al, 2001; Cholongitas et al, 2006). Combined respiratory and renal failure requiring the use of mechanical ventilation and renal replacement therapy carries a mortality rate ranging from 61% to 89% (Arabi et al, 2004; Cholongitas et al, 2009). Meticulous perioperative assessment, planning, and care are essential to minimize complications and decrease mortality rates in these patients. Although this chapter addresses the special perioperative concerns of cirrhotic patients, unless otherwise specified, the data are applicable to all types of chronic liver disease, including chronic infections with hepatitis B or C virus, chronic autoimmune hepatitis, biliary cirrhosis, sclerosing cholangitis, and hemochromatosis.

Preoperative Assessment and Care

Global Assessment

The loss of hepatic reserve capacity and the associated systemic derangements that result from liver dysfunction cause patients with chronic liver disease to have an inappropriate response to surgical stress. Increased surgical risk has been reported for various types of surgery in cirrhotic patients, including cardiac operations, open cholecystectomy, hysterectomy, nephrectomy, and transurethral resection of the prostate (Shaheen et al, 2009; Hayashida et al, 2004; Thulstrup & Sorensen, 2001; Lund et al, 2003; Nielsen et al, 2001, 2002). Emergency surgery and unscheduled laparotomy for trauma carry a high mortality rate (Pronovost et al, 1999; Carbonell et al, 2005; Wahlstrom et al, 2000; Demetriades et al, 2004). In contrast, patients with cirrhosis who undergo laparoscopic cholecystectomy have generally favorable outcomes (Morino et al, 2000; Pavlidis et al, 2009).

Two risk-stratification scoring systems have been developed to evaluate the preoperative severity of liver dysfunction and to predict postoperative outcome in cirrhotic patients. The Child score and its modification, the Child-Turcotte-Pugh (CTP) score, has been most commonly used. In the CTP score, a composite score of the degree of hepatic synthetic dysfunction (based on prothrombin time and serum levels of albumin and bilirubin), degree of ascites, and presence of hepatic encephalopathy yields a total score that ranges from well-compensated cirrhosis (class A, 5 to 6 points) to mild (class B, 7 to 9 points) and severe (class C, 10 to 15 points) decompensation. Cirrhotic patients undergoing elective or emergency extrahepatic surgery have a mortality rate ranging from 0% to 7.1% for class A to 84% to 100% for class C (Franzetta et al, 2003). Other risk factors for perioperative complications and mortality among cirrhotic patients undergoing surgery include male gender, preoperative infection or upper gastrointestinal bleeding, elevated creatinine concentration, and thrombocytopenia (Ziser et al, 1999).

In 2000, the Model for End-Stage Liver Disease (MELD) score was developed and proved to be superior to the CTP score in predicting the 3-month survival of patients undergoing transjugular intrahepatic portosystemic shunt (TIPS) (Malinchoc et al, 2000; Salerno et al, 2002). The MELD score is based on the prothrombin time–international normalized ratio (PT-INR) and serum bilirubin and creatinine levels, and it is calculated as follows:

The MELD score has been shown to correlate well with the CTP score in predicting postoperative mortality in cirrhotic patients undergoing elective and emergent surgery (Farnsworth et al, 2004). Other studies have shown that high MELD scores predict a poor outcome after hepatic resection for hepatocellular carcinoma (HCC) and following abdominal nonhepatic surgery in cirrhotic patients (Befeler et al, 2005; Teh et al, 2005). Finally, functional assays, such as the monoethylglycinexylidide (MEGX) test and indocyanine green clearance, may be used to assess hepatic functional reserve to predict postoperative morbidity and mortality prior to liver resection (Ercolani et al, 2000; Oellerich & Armstrong, 2001; Ravaioli et al, 2003; Lorf et al, 2008).

Risk Assessment (See Chapter 2)

Although surgical and anesthetic risks are largely related to the severity and not the cause of cirrhosis, determination of the etiology of the liver disease may be helpful in guiding the perioperative management and the timing of surgery (Table 23.1). A history of alcohol use should be obtained for all cirrhotic patients. Surgery should be deferred, if possible, in patients who are actively drinking alcohol because alcohol increases tolerance to narcotic and anesthetic agents, increases the risk of general anesthesia, and predisposes patients to alcohol withdrawal in the postoperative period (May et al, 2001; Spies et al, 2001). Prolonged alcohol abstinence (≥6 months) may lead to a significant reduction in the severity of portal hypertension and improvement in synthetic function in patients with alcoholic hepatitis. However, postponing surgery to allow alcohol abstinence is generally not practical in case of surgery for trauma or liver cancer.

Table 23.1 Risk Factors for Surgery and Preventive and Therapeutic Interventions

Risk Factors	Detrimental Effects	Interventions
Active alcohol use	Acute alcohol hepatitis	Cancel elective surgery
	Increased tolerance to sedatives and anesthetics	Prolonged alcohol abstinence
	Alcohol withdrawal and delirium	Use sedative and defer elective surgery
Iron overload	Cardiomyopathy and arrhythmias	Therapeutic phlebotomies
Hepatitis B infection	Flare of hepatitis	Antiviral agents
Autoimmune hepatitis	Flare of hepatitis	Stress-dose steroids

Similar to active alcohol use, chronic viral and autoimmune hepatitis increase the risk of perioperative complications. Among patients undergoing hepatectomy for hepatocellular carcinoma, the presence of high-grade inflammation of the remaining healthy liver is associated with postoperative hepatic dysfunction (Nanashima et al, 2003). Patients with cirrhosis as a result of chronic hepatitis B may benefit from antiviral and immunologic agents such as lamivudine, adefovir, entecavir, telbivudine, tenofovir, interferon alfa-2b, and peginterferon alfa-2a before surgery (Keeffe et al, 2008; Lau et al, 2005; Hadziyannis et al, 2005, 2006; Schalm et al, 2000; Chan et al, 2005). Patients with chronic autoimmune hepatitis who have been on long-term steroid therapy should receive stress-dose steroids in the perioperative period, and adjunctive immunomodulatory agents should be resumed as soon as possible (Czaja, 2009).

Portal Hypertension, Varices, and Ascites (See Chapter 70A, Chapter 70B, Chapter 74, Chapter 75A, Chapter 75B, Chapter 75C )

The severity of portal hypertension, as manifested by ascites and varices, is a major concern in the preoperative evaluation of cirrhotic patients. Preoperative endoscopy and measurement of hepatic venous pressures identify patients with severe portal hypertension who are at risk for variceal bleeding (Dittrich et al, 2003). Elevated hepatic venous pressures may portend the development of postoperative complications and higher hospital mortality in cirrhotic patients undergoing various surgical procedures including cholecystectomy, coronary artery bypass grafting, abdominal aortic aneurysm repair, and colorectal surgery (Csikesz et al, 2009; Nguyen et al, 2009). A platelet count of less than 100,000/µL is another useful surrogate marker of severe portal hypertension and of eventual hepatic decompensation (Zaman et al, 2001).

In patients with significant risk factors for variceal hemorrhage, nonselective β-blockers, such as propranolol or nadolol, are recommended to decrease the resting heart rate by 25% of pretreatment level or to approximately 50 to 60 beats/min. These agents, used either alone or in combination with nitrates, may reduce the variceal bleeding rate by half (Bureau et al, 2002; Patch et al, 2002). Alternatively, esophageal variceal band ligation or TIPS placement (see Chapter 76E) could be considered (Escorsell et al, 2002; Lo et al, 2002).

Ascites predisposes to bacterial peritonitis, respiratory compromise as a result of reduced lung expansion, and delayed wound healing. Adequate control and evaluation of ascites is important in cirrhotic patients scheduled to undergo abdominal surgery. In general, control of ascites may require several months and can be achieved with a low-sodium diet, diuretics, and/or large volume paracentesis (Table 23.2). In approximately two thirds of patients with ascites that is refractory or intolerant to diuretics, TIPS may be beneficial. Doppler ultrasound is advocated prior to TIPS procedure, as findings of abnormal blood flow in the portal venous system correlate well with increased mortality (Harrod-Kim & Waldman, 2005). Patients who undergo TIPS are also at increased risk for hepatic encephalopathy and should be monitored closely (Sahagun et al, 1997; Maleux et al, 2004). Patients with high MELD scores (>15) are at high risk for 30-day mortality and should not undergo TIPS for refractory ascites (Ferral et al, 2004; Schepke et al, 2003). In general, cirrhotic patients whose ascites is refractory to diuretics or TIPS should avoid abdominal surgery because of the high rates of postoperative morbidity and death. Finally, a high MELD score and use of a proton pump inhibitor are also associated with increased risks of spontaneous bacterial peritonitis (Bajaj et al, 2009; Obstein et al, 2007).

Table 23.2 Preoperative Risk Factors and Recommendations

Risk Factors	Evaluation and Management
Esophageal varices	Esophagogastroduodenoscopy
	Nonselective β-blockers
	Band ligation
Ascites	Abdominal ultrasound
	Paracentesis to rule out spontaneous bacteral peritonitis
	Sodium restriction
	Diuretics
Refractory ascites with low MELD score	TIPS
Hepatopulmonary syndrome	Arterial blood gas and bubble electrocardiogram
	Quantitative tagged microaggregated lung perfusion scan
	Echocardiogram
	Right heart catheterization for confirmation

MELD, Model for End-Stage Liver Disease; TIPS, transjugular intrahepatic portosystemic shunt

Avoiding preoperative intravascular volume depletion is important, as hypovolemia can lead to intraoperative and postoperative hypotension and shock, particularly in cases of surgery for emergent or life-threatening indications. Paracentesis and diuretics should be used judiciously, because they can trigger hypovolemic shock. Paracentesis may be performed with concomitant intravenous administration of selective colloid solutions. Albumin solutions are the preferred colloids used for resuscitation, as isotonic saline or dextran solutions may lead to ongoing accumulation of ascites (Garcia-Compean et al, 2002; Planas et al, 1990).

Nutrition (See Chapter 24)

Severe malnutrition affects up to 40% of cirrhotic patients and is associated with a high incidence of morbidity and mortality in patients with cirrhosis who are undergoing surgery (Kalaitzakis et al, 2007; Merli et al, 2002; Sam & Nguyen, 2009). In patients with advanced cirrhosis, several of the common parameters of nutritional status—such as body weight, body mass index, and albumin level—are not valid and reliable tools for assessment.

Perioperative interventions to enhance the nutritional status of cirrhotic patients may be associated with improved outcomes. Although special supplements with branched-chain amino acids are available and may reduce the risk of hepatic encephalopathy (Marchesini et al, 2003), in practice, their benefit is at most marginal. Nutrition via the enteral route is advocated, and nasogastric tubes can be employed safely in patients with esophageal varices. Total parenteral nutrition may be difficult to administer in cirrhotic patients because of concerns with fluid overload and infection.

Anemia

Bleeding, hemolysis, splenic sequestration, malnutrition, and underlying liver disease are among the multifactorial causes of anemia in cirrhotic patients. Longitudinal studies of patients with compensated cirrhosis without esophageal varices demonstrate that anemia may develop an average of 39 months after the onset of cirrhosis (Qamar et al, 2009). In 20% to 40% of patients with advanced cirrhosis of nonbiliary origin, a spur-cell hemolytic anemia associated with iron overload may become evident. Iron should be administered only in the presence of documented deficiency. In patients scheduled for partial hepatectomy for HCC, autologous blood transfusion and erythropoietin therapy have been used successfully to prevent the need for red blood cell transfusions (MacLaren et al, 2004; Shinozuka et al, 2000; Kajikawa et al, 1994).

Coagulopathy

Coagulation abnormalities are frequently found in cirrhotic patients undergoing surgery. Etiologies include decreased hepatic synthetic function, abnormal synthesis of coagulation factors, malnutrition, vitamin K deficiency, thrombocytopenia, and dysfibrinogenemia (Amitrano et al, 2002; Qamar et al, 2009; Richard et al, 2007). The degree of coagulopathy is proportional to the severity of cirrhosis. For patients with prolonged PT-INR and suspected vitamin K deficiency, a 3-day trial of vitamin K is recommended to determine whether the coagulopathy is correctable. Intravenous administration of vitamin K is the favored method in the ICU: subcutaneous administration has an inconsistent rate of absorption, and intramuscular injection carries the risk of hematoma formation. Oral menadiol, the water-soluble formulation of vitamin K, is effective for patients with cholestasis. If the PT-INR fails to correct with vitamin K, mixing studies may help identify possible inhibitors and guide therapy with fresh frozen plasma (FFP) or recombinant factors. Of note, the number of units of FFP commonly used in clinical practice infrequently corrects the coagulopathy (Youssef et al, 2003). In general, administration of cryoprecipitate is recommended when serum fibrinogen levels are less than 100 mg/dL, particularly in the presence of bleeding.

Excessive plasma transfusions can lead to volume overload and may exacerbate ascites. Human recombinant activated factor VIIa (rFVIIa) has been shown to be beneficial as an alternative to plasma. Rapid correction of the PT-INR has been documented within 15 minutes of a small, single intravenous dose of rFVIIa (Surudo et al, 2003), and rFVIIa has been used successfully in cirrhotic patients with coagulopathy to achieve hemostasis for colonic polypectomy, liver biopsy, dental extraction, and bleeding esophageal varices refractory to standard endoscopic therapy (Anantharaju et al, 2003; Berthier et al, 2002; Carvalho et al, 2002; Ejlersen et al, 2001; Romero-Castro et al, 2004). Transfusion of platelets is recommended before invasive procedures. A platelet count of less than 50,000/µL is the usual trigger for transfusion (Rebulla, 2001). Finally, desmopressin may be considered for refractory hemorrhage, particularly when azotemia is present.

Cardiorespiratory Evaluation

Patients with advanced cirrhosis typically have a hyperdynamic hemodynamic profile similar to the systemic inflammatory response syndrome: high cardiac output and vasodilatation of the pulmonary, splanchnic, and peripheral beds (Cazzaniga et al, 2009). Patients with alcoholic liver disease and iron overload are predisposed to cardiomyopathy and cardiac arrhythmias (Moller & Henriksen, 2008). Patients with cirrhosis also are at risk for coronary artery disease because of a high prevalence of cigarette smoking and diabetes mellitus (Hickman & Macdonald, 2007).

Hypocapnia is the most frequent gas-exchange abnormality in patients with cirrhosis (Aller et al, 1999), and pulmonary function can be compromised by tense ascites. Large-volume paracentesis may lead to improvement in the forced expiratory volume in 1 second (FEV₁), functional residual capacity (FRC), and arterial oxygenation (Gupta et al, 2000).

Patients with cirrhosis are also at risk for the development of hepatopulmonary syndrome (HPS) (Gupta et al, 2001). Subclinical or overt HPS is often associated with portopulmonary hypertension and causes significant pulmonary abnormalities as a result of alterations in the pulmonary vasculature by various toxins that accumulate and bypass the cirrhotic liver. Patients with HPS are seen with hypoxemia, platypnea, orthodeoxia, clubbing, and spider angiomata. The predictive value of the alveolar-arterial oxygen gradient in identifying patients with HPS remains a matter of controversy (Schenk et al, 2002; Lima et al, 2004); however, transthoracic echocardiography can be performed to screen for HPS and portopulmonary hypertension; bubble contrast echocardiography identifies the presence of intrapulmonary shunts indicative of HPS (see Table 23.2).

The presence of HPS worsens the overall prognosis of patients with liver cirrhosis (Schenk et al, 2003). The severity of HPS can be quantified with a tagged, macroaggregated albumin lung perfusion scan and arterial blood gas analysis. Postoperative mortality from cardiorespiratory complications can be estimated by a preoperative Pao₂ of 50 mm Hg or a macroaggregated albumin scan quantification shunt fraction of 20% or higher (Arguedas et al, 2003). General anesthesia for hepatic and extrahepatic procedures in patients with HPS is associated with increased perioperative risk (Mazzeo et al, 2004).

Portopulmonary hypertension, an integral component of HPS, is often asymptomatic and difficult to diagnose; however, this syndrome predisposes to intraoperative cardiac arrhythmias and cardiac arrest. Mortality of portopulmonary hypertension is high, with 1- and 5-year survival rates of 54% and 14%, respectively (Swanson et al, 2008). Although transthoracic echocardiogram may be the initial study performed, right heart catheterization is the best diagnostic tool for portopulmonary hypertension, and it helps discriminate this process from other causes of pulmonary hypertension (Krowka et al, 2006). Elective surgery under general anesthesia in patients with severe HPS-associated pulmonary hypertension should be deferred until the pulmonary hypertension is controlled with vasodilator agents.

Electrolytes

Hyponatremia is quite common in patients with cirrhosis because of their impaired ability to excrete free water, and it is associated with increased mortality (Borroni et al, 2000). Hyponatremia often occurs in the setting of severe ascites, hepatic encephalopathy, spontaneous bacterial peritonitis, and hepatorenal syndrome (Angeli et al, 2006). Hyponatremia and other electrolyte abnormalities, such as hypokalemia and hypomagnesemia, can result from the use of loop diuretics and as a consequence of the chronic respiratory alkalosis associated with cirrhosis. Patients with active alcohol use are also at increased risk of hypophosphatemia. Electrolyte imbalances should be evaluated and corrected before surgery to limit cardiac arrhythmias, peritonitis, and hepatic encephalopathy.