Cardiovascular System

Hyperdynamic circulation, cardiomyopathy, and CAD are common clinical problems that impact postoperative outcomes in orthotopic liver transplantation (OLT) candidates (Fouad et al, 2009).

Pulmonary System

As many as 70% of patients with chronic liver disease (CLD) suffer from respiratory problems, that manifest as impaired respiratory mechanics and gas exchange (Kim et al, 2009). Most pulmonary diseases are independent from the liver disease but are in some instances related to liver disease (e.g., α₁-antitrypsin deficiency and cystic fibrosis) (Yeshua et al, 2009). Pulmonary evaluation includes history and physical examination, radiologic studies, arterial blood gas, and pulmonary function tests. Right-heart catheterization is indicated when clinical or ECG evidence suggests pulmonary hypertension.

Hemostasis in End-Stage Liver Disease

The liver synthesizes most of the factors involved in clot formation and, equally importantly, factors opposing clot formation or mediating its lysis. The traditional concept that patients with advanced liver disease are hypocoagulable has been challenged. The current view emphasizes complex disturbances of platelet function, procoagulant, anticoagulant, and fibrinolytic pathways that result in rebalancing of the hemostatic system (Senzolo et al, 2006). The net effect may be a hypocoagulable or hypercoagulable state with a tendency for bleeding or thromboembolism (Warnaar et al, 2008b). Understanding the hemostatic system in ESLD is essential to perioperative coagulation management and optimization of blood product transfusion practices, which are directly related to morbidity and mortality in liver transplantation (de Boer et al, 2008; Massicotte et al, 2005).

Deficiency in primary hemostasis is due to quantitative and qualitative platelet changes (Pereboom et al, 2008). Mild to moderate thrombocytopenia is common. Causes are splenomegalic platelet sequestration in portal hypertension, impaired megakaryocytopoiesis as a result of reduced production of thrombopoietin by the cirrhotic liver, folic acid deficiency in alcoholic cirrhosis or acute hepatitis C infection, and reduced platelet half-life related to autoantibodies (Lisman & Leebeek, 2007; Senzolo et al, 2006). Evidence for impaired platelet aggregation and adhesion in cirrhosis is controversial. The proposed mechanisms are complex and multifactorial, including the important role of two potent endothelium-derived platelet inhibitors, nitric oxide and prostacyclin. Contrary to this, more recent work suggests that the functional capacity of platelets in cirrhotic patients may be normal when platelet number and hematocrit are normalized (Lisman & Leebeek, 2007; Tripodi et al, 2007), suggesting that the main defect in primary hemostasis is thrombocytopenia. An important compensatory mechanism for the overall deficient platelet function in cirrhosis is an increase in endothelium-synthesized von Willebrand factor (vWf). The markedly upregulated vWf promotes adhesion of platelets to injured vasculature and seems to be more effective than plasma from healthy volunteers (Lisman & Leebeek, 2007).

Hepatocytes synthesize all procoagulant factors except vWF, which is synthesized by the endothelium. Factor VIII is synthesized by hepatocytes but also by nonhepatic sinusoidal endothelial cells. Reduced levels of coagulation factors are observed in acute and chronic liver disease. Several factors—II, VII, IX, and X—require vitamin K as an essential cofactor and exist in plasma as inactive precursors when vitamin K is deficient as a result of decreased absorption (cholestasis) or antagonism (warfarin). In acute liver failure (ALF), plasma concentrations of coagulation factors with the shortest half-life (factor VII, 4 to 6 hours; factor V, 12 hours) decrease first, followed by those with a longer half-life. An exception are factors VIII and vWf, which may be elevated in chronic and acute liver disease because of inflammatory cytokine-mediated upregulated extrahepatic production and because of relative sparing as a result of inhibition of thrombin by antithrombin III. Fibrinogen is usually well preserved, and decreased fibrinogen is a mark of advanced cirrhosis or ALF; however, even when levels are normal, fibrinogen is often functionally aberrant because of impaired fibrin polymerization and results in an abnormal thrombin time (TT) despite normal international normalized ratio (INR) and partial thromboplastin time (PTT).

The deficiency in procoagulant factors is, at least in part, offset by decreased production of anticoagulants, including proteins C and S, antithrombin III, and heparin cofactor II. Proteins C and S are vitamin K dependent and may be deficient in cholestatic disease, and although genetic deficiency of these two proteins is rare, such deficiency is found in both Budd-Chiari syndrome and portal vein thrombosis. Antithrombin III is synthesized by hepatocytes and endothelium, deficiencies are usually mild, and thromboembolic complications rare.

All proteins involved in fibrinolysis are synthesized by the liver except tissue plasminogen activator (tPA) and plasminogen activator inhibitor type 1 (PAI-1). Consequently, reduced levels of liver-synthesized factors—plasminogen, plasmin inhibitor, thrombin activatable fibrinolysis inhibitor (TAFI), and factor XIII—are found in both acute and chronic liver failure. Cirrhosis is classically associated with hyperfibrinolysis, whereas patients in ALF may have hypofibrinolysis as a result of increased production of PAI-1. The recent expert viewpoint is contrary to the traditionally held view and emphasizes that the existing in vitro evidence of hyperfibrinolysis in cirrhosis is weak (Lisman & Leebeek, 2007). Increased levels of fibrinogen degradation indicators—d-dimers, plasmin-antiplasmin and thrombin-antithrombin complexes, thrombin fragment F1+2—could be explained by their delayed clearance by the diseased liver.

Coagulation tests, such as the prothrombin time (PT) and activated partial thromboplastin time (aPTT), are used to assess the severity of synthetic dysfunction, and PT-INR is a part of the Child-Turcotte-Pugh (CTP) and Model for End-Stage Liver Disease (MELD) prognostic indexes (Ng, 2009). As tests of in vivo coagulation function, PT-INR and PTT are conceptually deficient because they measure only the procoagulant pathway and ignore the anticoagulant function. This inadequacy is best illustrated by the observation that when the (prolonged) PT is modified by including thrombomodulin to allow full activation of protein C, the thrombin generation in cirrhotic patients is close to normal (Tripodi et al, 2005). This may in part explain why PT-INR and PTT are poor predictors of bleeding in cirrhosis and in liver transplantation (Massicotte et al, 2008). Thromboelastogram (TEG; Haemoscope Corp., Niles, IL) and rotation thromboelastometry (ROTEM; Pentapharm, Munich, Germany) have the advantage of assaying whole-blood clot formation and lysis, including the platelet contribution, but these assays do not incorporate the endothelial component (see Intraoperative Monitoring, below). Importantly, these assays allow easy recognition of hypercoagulable states (Fig. 97B.1A).

FIGURE 97B.1 Intraoperative transesophageal echocardiography (TEE) in a patient with hepatopulmonary syndrome and intrapulmonary shunting demonstrates a positive bubble study. A, Midesophageal bicaval view. B, Bolus of contrast injectate (agitated saline-blood mixture) is visible entering the right atrium via the superior vena cava (SVC) following an injection through the central venous port. C, Several heartbeats later, the injectate is present in the left atrium (LA), indicating transpulmonary passage. RA, right atrium; IS, interatrial septum; IVC, inferior vena cava.

Renal System

Renal failure confers increased risk of death to cirrhotic patients, both while on the transplant waiting list and after liver transplantation. Implementation of the MELD score in 2002 (Moreau et al, 2003) has prioritized patients with renal failure waiting for a liver transplant, resulting in reduced overall mortality on the transplant list (Ojo, 2009).

Hepatorenal syndrome (HRS) is a specific type of renal failure in patients with cirrhosis, ALF, or alcoholic hepatitis. It is marked by functional renal vasoconstriction and a severe reduction in glomerular filtration rate (GFR <30 mL/min) but only minimal renal histologic abnormalities (Gines & Schrier, 2008). The primary pathophysiologic process in HRS is a circulatory disturbance: portal hypertension facilitates translocation of bacteria from the intestinal lumen into the mesenteric lymph nodes and induces production of inflammatory cytokines and local vasodilators, such as nitric oxide, carbon monoxide, and endogenous cannabinoids. Resulting splanchnic vasodilatation decreases the overall systemic vascular resistance and activates a number of compensatory mechanisms. The initial response is hyperdynamic and comprises tachycardia, increased cardiac output, low SVR, and relative hypotension.

With progression of cirrhosis, sympathetic and rennin-angiotensin vasoconstrictor systems are activated that temporarily maintain arterial blood pressure, but they also cause renal vasoconstriction and hypoperfusion. Activation of nonosmotic hypersecretion of arginine vasopressin, aimed at preserving circulatory volume, results in retention of solute-free water with attendant hyponatremia, edema, and ascites. HRS is almost invariably associated with ascites and rarely develops in its absence (see Chapter 74). HRS is classified into one of two types: HRS 1 is characterized by a rapid decline in renal function reflected in doubling of serum creatinine (SCr) to above 2.5 mg/dL over a 2-week period; this type of HRS has a particularly ominous clinical course, leading to multiorgan dysfunction that affects the heart, brain, and other organs. HRS 2 has a more benign, slowly progressive course. It is manifested by a milder elevation of SCr (1.5 to 2.5 mg/dL) and refractory ascites. Onset of HRS is often triggered by a discrete event, such as spontaneous bacterial peritonitis, use of nonsteroidal antiinflammatory drugs (NSAIDs), or hypovolemia as a result of diuretic use, gastrointestinal (GI) bleeding, or fluid losses.

It is important to recognize that HRS may be superimposed on a chronic disease, such as glomerulonephritis as a result of hepatitis B virus (HBV) or hepatitis C virus (HCV) infection, interstitial nephritis as a side effect of certain drugs, or nephropathy resulting from diabetes or CLD. Diagnosis of HRS is of great prognostic significance: median survival for HRS 1 is 1 month; for HRS 2, it is 6 months.

Measurement of SCr remains the primary test of renal function despite its well-known limitations. By consensus, SCr above 1.5 mg/dL (133 µmol/L) defines renal insufficiency in cirrhosis (Gines & Schrier, 2008; Salerno et al, 2007). GFR and creatinine clearance are more accurate but less practical tests when repeated assessments are needed. Other tests measure serum and urinary electrolytes, fractional sodium excretion (FENa), albumin, osmolarity, sediment, and biologic markers. Specific diagnostic tests for HRS do not exist, rather diagnosis is based on the presence of liver disease, a precipitating factor, and FENa less than 1%, signifying preserved tubular reabsorption; however, the latter test is invalidated by the use of diuretics.

General treatment measures are aimed at 1) preventing infection, in particular subacute bacterial peritonitis; 2) maintaining adequate intravascular volume; and 3) avoiding nephrotoxic agents. Circulatory volume losses, often the result of large-volume paracentesis or a GI bleed, are best treated with albumin or, if indicated, by blood products and not with crystalloids, to avoid retention of solute-free water.

The main pharmacologic approach utilizes vasoconstrictors, including a selective vasopressin V1 receptor agonist, such as terlipressin, or α-adrenergic agonists, such as norepinephrine and midodrine. Vasoconstrictors in conjunction with albumin are only modestly beneficial but are the best medical therapy currently available. Two recent studies showed a marked improvement in renal function in 40% of patients treated with terlipressin and albumin compared with controls treated with albumin alone (Martin-Liahi et al, 2008; Sanyal et al, 2008). Although the overall survival in the treatment group was not better than in the control group, the probability of 3-month survival in patients who responded with improved renal function was significantly higher (58% vs. 15%) than in nonresponders (Martin-Liahi et al, 2008); however, organ ischemia is a concerning side effect of vasoconstrictor administration.

Placement of a TIPS improves renal function and GFR and reduces sympathetic and renin-angiotensin-aldosterone axis activation in about 60% of patients (Cardenas & Gines, 2009; see Chapter 76E). TIPS is less applicable in HRS 1 as a result of severe liver disease, but it may delay progression of HRS 2 to HRS 1. No advantage in survival has been demonstrated in patients with refractory ascites treated by TIPS compared with repeated paracentesis and intravascular volume replacement with albumin.

Renal replacement therapy (RRT) offers a bridge to transplant, but the optimal RRT method in HRS and the benefit to patient outcomes are not known (Gines & Schrier, 2009). Continuous venovenous hemodialysis (CVVHD) seems to be hemodynamically the most favorable form of RRT (Davenport, 2009). CVVHD can be used to effectively control intravascular volume, pH levels, and solutes (Na, K, SCr, urea, ammonia); in addition, it corrects sodium in a time-controlled fashion, which is particularly relevant to minimize the risk of pontine demyelination in patients with hyponatremia.

Liver transplantation is the treatment of choice for patients with cirrhosis and HRS (Cardenas & Gines, 2009; see Chapter 97A). About 90% of patients with HRS recover kidney function after successful liver transplantation. The remaining 10% do not recover their kidney function and require prolonged RRT and eventual kidney transplantation. When the cause of renal failure in cirrhosis is HRS, the best predictor of need for a combined liver-kidney transplant is pretransplantation RRT for longer than 8 weeks.

Central Nervous System

Hepatic encephalopathy (HE) is a progressive but potentially reversible metabolic disorder of the central nervous system (CNS), and it is associated with variable degrees of brain edema. HE is an overt manifestation of cirrhosis in up to 45% of patients, and it is a defining sign and main cause of death in ALF (Eroglu et al, 2009). The World Congress of Gastroenterology recognizes three classifications of HE based on the etiology (Ferenci et al, 2002): type A is associated with ALF, type B with TIPS, and type C with cirrhosis. Based on the severity of neurologic impairment, HE ranges from normal (grade 0) to mildly impaired cognitive and neuromuscular function (grade 1), progressive lethargy (grade 2), confusion with muscle rigidity or clonus (grade 3), to coma and brain death (grade 4). Although presence and severity of HE has prognostic and therapeutic significance in cirrhosis and ALF, it is not a part of the MELD score.

The pathophysiology of HE is complex and incompletely understood. Multiple putative mediators have been implicated that include ammonia, endogenous benzodiazepine-like or γ-aminobutyric acidlike agonists, reactive oxygen species, inflammatory cytokines, hyponatremia, and manganese. The exact pathogenesis of cerebral edema remains controversial, but good evidence supports the role of ammonia, which promotes astrocyte and neuron swelling by increasing intracellular glutamine content. Transported to the mitochondria, glutamine contributes to production of reactive oxygen species and activation of mitochondrial transition permeability pore and various kinases, resulting in cellular swelling. Increased ammonia levels have been shown to alter cerebral blood flow and brain glucose utilization (Bjerring et al, 2009; Sundaram et al, 2009). High arterial and brain levels of ammonia in ALF correlate well with cerebral edema and increased ICP, with consequent high risk of mortality from brain herniation and hypoxia (Wendon et al, 2008).

HE in cirrhosis features slow onset, a progressive course, and only partial responsiveness to therapy. Significant cerebral edema typically does not develop, and a precipitating event—such as infection, excessive dietary protein, dehydration, or GI bleeding—can be often identified. TIPS is a known risk factor and is associated with 30% of new or worsening encephalopathy (see Chapter 76E). In addition to correction of any precipitating factors, treatment includes nonabsorbable disaccharides, which reduce the intestinal production and absorption of ammonia; antibiotics that target urease-producing bacteria; ornithine and acarbose; benzodiazepine receptor antagonists; and probiotics.

The hallmarks of HE in ALF are cerebral edema and increased intracranial pressure (ICP). Both are life-threatening complications of grade 3 and 4 encephalopathy. Treatment includes a number of general and specific interventions (Table 97B.1) aimed at promoting cerebral perfusion pressure (CPP) above 60 mm Hg; this is achieved by maintaining mean arterial pressure (MAP) and reducing ICP to below 20 to 25 mm Hg by reducing brain edema. Grade 3 or 4 HE requires endotracheal intubation for airway protection. Mild hyperventilation temporarily relieves brain hyperemia, but prolonged hyperventilation is not beneficial and may cause CNS ischemia. Mannitol is effective in reducing ICP in the short term, but its prophylactic use is not recommended. Recent evidence demonstrates that N-acetylcysteine treatment may also prolong transplant-free survival in ALF unrelated to paracetamol/acetaminophen toxicity (Lee et al, 2007).

Table 97B.1 Management of Encephalopathy and Elevated Intracranial Pressure in Acute Liver Failure

CVVHD, continuous venovenous hemodialysis; ICP, intracranial pressure

The role of ICP monitoring is controversial, as ICP monitor placement is associated with high risk of bleeding in this patient population, and no data exist to support its role in improving patient outcomes. The Acute Liver Failure Study Group recommends this intervention only for liver transplant candidates but recognizes regional practice differences (Stravitz et al, 2007). An attractive, noninvasive alternative to ICP monitors is transcranial Doppler (TCD) sonography. By measuring blood flow velocities in middle and anterior cerebral arteries and calculating resistance and pulsatility indexes, TCD allows serial estimates of cerebral blood flow in the setting of impaired autoregulation and developing brain edema (Bhatia & Gupta, 2007). In a small series of ALF patients, confirmation of adequate cerebral blood flow by TCD was successfully used as a basis to proceed with OLT (Bindi et al, 2008). Intraoperative use of TCD in OLT has not been reported.

Mild to moderate hypothermia (32° C to 35° C) is an emerging therapeutic modality. Hypothermia reduces brain ammonia and cytokine levels and in part restores cerebrovascular autoregulation and brain glucose metabolism. Although experimental evidence provides a strong rationale, only limited clinical data on efficacy and safety of hypothermia in humans in ALF are available, and future studies are warranted (Stravitz et al, 2009).

Hemodynamic Management

Hemodynamic management is a cornerstone in the perioperative care of liver transplant patients, but no consensus has been reached on the standards of hemodynamic monitoring (Claus-Georg & De Wolf, 2008). Continuous invasive arterial pressure monitoring may be the only ubiquitously used hemodynamic monitor in liver transplantation. When measured at the most common site, the radial artery, peripheral arterial pressure may underestimate the true aortic pressure, particularly during periods of hypotension. A femoral arterial line reflects central aortic pressure with higher fidelity and consistently records higher systolic pressures than the radial artery. The difference between the femoral and the arterial pressure waveform is particularly pronounced after graft reperfusion or when high doses of vasopressors are used (De Wolf, 2006; Shin et al, 2007). Because of this discrepancy, and because of the need for frequent arterial blood sampling, some centers routinely use both femoral and radial arterial lines.

The pulmonary artery catheter (PAC, see Chapter 22) has long been the mainstay of hemodynamic monitoring in liver transplantation (De Wolf et al, 1993); however, the clinical utility of PACs has recently been questioned because of criticisms that its invasiveness is not justified by improved outcomes (Sandham et al, 2003; Wheeler et al, 2006) and that measured pressures are not good estimates of ventricular preload (Kumar et al, 2004). Although a number of less invasive technologies have emerged, resulting in a decrease in their use, PACs are still advocated by many for OLT (Della Rocca et al, 2008; De Wolf, 2008; Pinsky, 2006), and the ability of PACs to continuously display pulmonary artery pressures is particularly advantageous when managing patients with portopulmonary hypertension. Transesophageal echocardiography (TEE) provides different but complementary information to that provided by PACs (see below), and concurrent use of both devices is considered superior (De Wolf & Aggarwal, 2008). A notable shortcoming of PACs in OLT is the inaccuracy of thermodilution cardiac output in the period following liver reperfusion because of postreperfusion hypothermia and subsequent instability of central temperature.

Newer, less invasive hemodynamic monitors for OLT have been introduced. The PiCCO system (Pulsion Medical Systems, Munich, Germany) combines transpulmonary thermodilution and arterial pressure contour analysis, and it provides continuous cardiac output in alignment with PAC thermodilution and a number of indices of preload that include global end-diastolic volume (GEDV) and intrathoracic blood volume (ITBV), which agree better with cardiac index than filling pressures (Della Rocca et al, 2002a, 2002b). PiCCO requires central venous and femoral or brachial arterial access, which arguably is not much less invasive than the PAC. The LiDCO device (LiDCO Group, London, UK) estimates cardiac output based on the peripheral arterial pressure waveform pulse power analysis after a lithium dilution calibration, and it has been used successfully in patients recovering from liver transplantation in the ICU (Costa et al, 2008). A device that estimates cardiac output based on the arterial pressure waveform without need for calibration, Flo-Trac/Vigileo (Edwards Lifesciences, Irvine, CA), did not perform well intraoperatively during OLT (Biais et al, 2008; Biancofiore et al, 2009).

The accuracy of all arterial pressure–based cardiac output monitors depends critically on accurate determination of arterial compliance, which changes rapidly in liver transplantation, particularly at reperfusion or clamping of major vessels; and because it cannot be readily measured, the assumed value of vascular compliance used in the algorithms of the devices to calculate cardiac output may lead to inaccurate calculations. The inaccuracy may be mitigated by repeated calibrations with the LiDCO or PiCCO.

Cardiac output determined by a novel suprasternal ultrasound cardiac output monitor (USCOM Pty, Coffs Harbour, Australia) correlates well with the PAC cardiac output (Wong et al, 2008; Su et al, 2008). Other USCOM measurements, such as Doppler estimation of peripheral artery pressures and left ventricular contractility, remain to be validated.

TEE has been gaining more popularity in liver transplantation (Burtenshaw & Isaac, 2006). Data from 2002 estimated use of TEE in only 21% of high-volume centers in the United States (Schuman, 2003), but a more recent survey showed that 86% of anesthesiologists in 40 large-volume U.S. centers use TEE in some or all liver transplantations (Wax et al, 2008).

The main advantages of TEE are that it provides continuous, real-time visualization of cardiac structures and dynamic function. TEE is a superior modality for estimating ventricular volume status, global LV contractility and ejection fraction, right ventricular function, segmental wall motion abnormalities, and septal motion. Doppler modalities allow for interrogation of transvalvular flows, dynamic left ventricular outflow obstruction, intracardiac shunts, and diastolic function. In addition, TEE is the only intraoperative monitor that allows an instantaneous diagnosis of pulmonary or intracardiac thrombus (Ellenberger et al, 2006; Planinsic et al, 2004) and pericardial or pleural effusion. TEE detection of a transpulmonary shunt as a positive bubble test in a patient with HPS is presented in Figure 97B.1.

Dynamic monitoring of the right ventricular function and detection of right ventricular failure are critically important in managing patients with portopulmonary hypertension, particularly during reperfusion. Assessment of right ventricular volume and contractility, severity of tricuspid regurgitation, and measurement of lows in the inferior vena cava (IVC) may help guide volume, inotrope, and pulmonary vasodilator therapy.

In addition to requiring significant training and resources, TEE has several important disadvantages. An obvious concern in the liver transplant patient population is the risk of rupture of gastroesophageal varices. Although this complication is very rare, it does emphasize the need for careful risk/benefit assessment of this monitoring modality in each individual recipient. It is prudent to withdraw the TEE probe from the stomach during surgical gastric retraction, making transgastric views inaccessible during most of the procedure. Another disadvantage is that TEE is not adept at providing and trending traditional numerical data (e.g., cardiac output), which are readily provided by the PAC. Visual assessments used in TEE are also highly operator dependent. Finally, unlike the PAC, the TEE probe is typically removed at the end of surgery; therefore the continuity of monitoring is interrupted, and the transfer of specific parameters of hemodynamic management to the postoperative ICU is difficult. This problem may be mitigated by the introduction of less expensive, disposable TEE probes.

Monitoring of Hemostasis

Conventional tests of hemostasis, PT and aPTT, have several important limitations (Dalmau et al, 2009; Ng et al, 2009). First, these assays utilize plasma samples devoid of platelets. The activated platelet surface, the naturally occurring medium of the in vivo coagulation, is replaced by addition of phospholipids. The contribution of platelets is thus not considered by these tests. Second, PT and PTT measure only the activity of procoagulant factors and do not take into account the anticoagulant factors. When PT is modified to include anticoagulant factors, thrombin generation in cirrhotic patients is not different from normal controls (Tripodi et al, 2005). The reported higher intralaboratory variability of INR than of PT may be of more relevance for MELD score determination than for acute intraoperative management (Lisman & Leebeek, 2007).

An important shortcoming of the conventional PT and PTT is a long turnaround time, typically about 45 minutes—suboptimal for an ongoing transplant surgery. Point-of-care (POC) assays require only a minimal delay between blood sampling and attaining the results of the test. POC assays of PT and PPT are quick (5 to 15 minutes), cartridge-type, whole-blood tests offered by several manufacturers. An acceptable agreement of such assays with corresponding plasma-based tests has been reported in OLT (Herbstreit et al, 2010), suggesting that the POC tests can effectively replace the slower, conventional testing.

Activated clotting time (ACT) was developed for POC monitoring of high-dose unfractionated heparin therapy, and it should be a part of the POC battery; with reperfusion of the graft, a massive dose of unfractionated heparin is flushed into the circulation. In conjunction with other POC tests, ACT may help guide the use and dosing of protamine.

TEG and ROTEM are POC assays that measure viscoelastic properties of whole-blood clot formation and lysis. Both technologies have been used to drive transfusion algorithms (Kang et al, 1985; Coakley et al, 2006). TEG measures time to onset of clot, rate of clot formation, clot tensile strength, and rate of clot lysis (Fig. 97B.2). Treatment of samples with an intrinsic pathway activator, kaolin, shortens clot formation, whereas addition of heparinase dissects the effects of exogenous heparin from intrinsic heparinoids. TEG results are available more rapidly than the conventional laboratory tests. In our institution, progress of TEG clot formation and lysis is displayed on a dedicated operating room monitor and is available to the anesthesiologist in real time.

FIGURE 97B.2 Intraoperative thromboelastogram (TEG) recordings from three orthotopic liver transplantation patients with different hemostatic derangements. In each panel, a “normal” TEG trace, generated by the instrument, is presented with an interrupted line. A, Hypercoagulable state in a patient with primary biliary cirrhosis. B, Postreperfusion TEG shows profoundly suppressed clot formation and marked fibrinolysis. C, Mild hypocoagulability and fibrinolysis in a patient with hepatitis C.

ROTEM generates a similar trace to that of TEG, although a ROTEM assay may be modified by a heparinase reagent (hep-TEM), an intrinsic pathway activator (in-TEM), and a tissue-factor activator (ex-TEM). In addition, ROTEM includes a qualitative test of fibrinogen activation and fibrin polymerization (fib-TEM), which is based on adding a blocker of platelet function. ROTEM is very useful in determination of poor tensile clot strength as a result of combined platelet and fibrinogen deficiency, and it is the most reliable method for detecting hypofibrinogenemia or dysfibrinogenemia, both common in liver disease (Kalina et al, 2008).

Both TEG and ROTEM are quick and superior monitors of hypercoagulability, fibrinolysis, and therapeutic effects of antifibrinolytics. The main conceptual deficiency of the two monitors is the inability to evaluate endothelial function. It is worth emphasizing that although both instruments plot clot formation in a similar fashion, TEG and ROTEM do not generate the same results when analyzing blood without exogenous activation, therefore they are not interchangeable (Coakley et al, 2006; Nielsen, 2007).

Monitoring and management of hemostasis remain great challenges of OLT, and it has been well established that blood loss and blood product transfusion are an important cause of morbidity and mortality (de Boer et al, 2008; Massicotte et al, 2005; Ramos et al, 2003). Adherence to an algorithm-driven transfusion practice may reduce unnecessary perioperative blood and blood-component transfusion (Avidan et al, 2004; Kang, 1997; Kang et al, 1985), but no consensus has been reached on what constitutes standard intraoperative monitoring of hemostasis or a standard hemostatic intervention.

Difficulties stem from the incomplete understanding of complex coagulation abnormalities in liver disease and OLT and from the limitations of currently available tests to emulate the in vivo conditions. For example, preoperative coagulation tests do not predict intraoperative bleeding (Massicotte et al, 2004; Steib et al, 2001), and transfusion requirements are actually lower when abnormal coagulation test results do not lead to routine correction by infusion of fresh frozen plasma (FFP) and other blood products (de Boer et al, 2005). Finally, bleeding in OLT is often related to factors other than coagulopathy. These findings emphasize the need for careful interpretation of coagulation tests within the clinical context rather than relying on interventions formulated solely on the basis of test results.

Metabolic Monitoring

Temperature Monitoring

Monitoring of temperature is usually achieved via a PAC or an esophageal probe. Surprisingly little has been reported on temperature management in OLT, considering the fact that hypothermia is a key sign of liver failure and a significant problem during surgery. Well-known detrimental effects of hypothermia include impaired wound healing and impaired coagulation. An acute decrease in central temperature results from the cold perfusate returning to the heart following graft reperfusion. This precipitous hypothermia contributes to postreperfusion hemodynamic instability and bradyarrhythmias that may progress to cardiac arrest (Fig. 97B.3). The hypothermic insult can be partially mitigated by achieving upper-limit normothermia in the prereperfusion period, thus the value of continuous temperature monitoring and measures to assure normothermia cannot be overemphasized. A special circumstance is deliberate mild hypothermia, which acts by decreasing brain metabolism and mediators of oxidative stress; thus mild hypothermia is an accepted treatment of increased intracranial pressure in ALF (Stravitz & Larsen, 2009).

FIGURE 97B.3 A two-lead intraoperative electrocardiogram (ECG) recorded on the operating room monitor shortly after reperfusion, shows profound bradycardia and tall T waves. Such changes are not uncommon in the period immediately following graft reperfusion.

Ventilatory Management

Mechanical ventilation should optimize respiratory gas exchange and maintain lung volumes and capacities to facilitate early postoperative extubation. Preexisting respiratory comorbidities, such as hydrothorax, have to be considered in setting the ventilatory parameters. Although not much is reported about the ventilator setting in OLT, it seems reasonable to adopt a protective ventilatory strategy and to avoid high airway peak pressures (below 30 to 35 cm H₂O). Tidal volume of 6 to 8 mL/kg are a reasonable initial setting (Slinger, 2008; Shultz, 2008), but tidal volumes of least 8 mL/kg are necessary for reliable pulse pressure variation (PPV) or stroke volume variation (SVV) monitoring (Michard, 2005).

Multiple factors can cause increased pulmonary capillary permeability and predispose the lungs to interstitial edema. Positive end-expiratory pressure (PEEP) is a rational tool to prevent lung edema, and a PEEP of up to 10 mm Hg has been proven safe in maintaining hemodynamics and graft perfusion after cadaveric and living-donor liver transplant (LDLT) (Saner et al, 2006a, 2006b; 2008). Fio₂ can be increased to modulate pulmonary resistance in case of pulmonary hypertension. It is not known whether high intraoperative values of inspiratory fraction of oxygen may have a role in modulating ischemia-reperfusion injury after reperfusion (Kabon & Kurz, 2006).

Intraoperative Hemodynamic, Fluid, and Transfusion Management

Hemodynamic management of fluids and blood-product transfusions are critically interconnected and should be considered together. Excessive fluid administration may result in increased pressure in the IVC, portal vein, and collateral veins, possibly contributing to blood loss during liver dissection, hemodilution of coagulation factors and platelets, and tissue edema (see Chapter 22; Westercamp et al, 2009). Conversely, inadequate intravascular volume repletion may result in hemodynamic instability and organ hypoperfusion.

Better understanding of pathophysiology in ESLD, improvement in surgical and anesthetic techniques, and availability of POC laboratories have resulted in marked reductions in blood-product administration, including those for transplantations without transfusion. This is of great significance, as reduced blood-product transfusion is associated with better long-term outcomes (Bennett-Guerrero et al, 2001; Massicotte et al, 2007; Pereboom et al, 2009).

A strategy of maintaining low central venous pressure (CVP ≤5 mm Hg) through restrictive fluid administration and phlebotomy has been advocated by Massicotte and colleagues (2006) as an effective and safe strategy for transfusion-free OLT (see Chapter 22). In contrast, Schroeder and colleagues (2004) reported increased mortality and renal failure associated with low-CVP management. Whereas the low-CVP strategy may be possible in select patients with normal kidney function and good cardiovascular reserve, targeting one specific low CVP (i.e., 5 mm Hg) may not be physiologically advantageous in patients with poor ventricular compliance and preexisting renal dysfunction.

A low-CVP regimen usually requires vasopressor administration to maintain arterial blood pressures (Ponnudurai et al, 2005; Schroeder et al, 2004) but with the possible side effect of end-organ hypoperfusion. As CVP is not a reliable index of preload, volumetric assessment with TEE or fluid-responsiveness indexes (dPP or SVV) may be a more physiologic approach to guide fluid management (Biais et al, 2008; Della Rocca, 2009). Fluid management strategy and interpretation of CVP values may vary based on the surgical technique; for example, during a complete IVC clamp, blood is sequestered below the clamp, and low CVP does not accurately reflect volume status. After unclamping of the IVC and portal vein, the CVP increases rapidly and may abolish the pressure gradient between the portal vein and the right atrium and thereby compromise liver perfusion. Hemodynamic changes during a piggyback technique rarely reach such magnitude.

Hypotension is common during all phases of OLT, but little is reported on the impact of intraoperative hypotension on long-term patient outcomes. Reich and colleagues (2003) reported the worst surgical outcomes early after liver transplantation in patients who had at least one episode of arterial hypotension with an MAP below 40 mm Hg.

Little evidence is available to provide guidance regarding the choice of intraoperative vasopressors. Agents are chosen based on their pharmacologic properties and side effects (Zhang et al, 2005). Commonly used agents are neosynephrine, norepinephrine, vasopressin, and dopamine. Of note, vasopressin 1 receptors and α₁-adrenergic agonists are useful in preserving renal function in HRS (Gines & Schrier, 2009); vasopressin reduces splanchnic blood flow, which may reduce bleeding from portal collaterals during the dissection phase but may also impede portal vein flow during the neohepatic phase (Wagener et al, 2008).

Most centers use a combination of crystalloids and colloids for fluid replacement. A recent study showed that use of new-generation hetastarch (HES 130/0.4) may be a safe alternative to conventionally used 5% albumin (Mukhtar et al, 2009), but more studies are needed.

Severe pulmonary hypertension (PVR >240 dynes/sec/cm⁻⁵) is associated with high perioperative mortality and is a contraindication for proceeding with liver transplant. In moderate PPHTN with preserved right ventricular function, the goal of intraoperative management is to lower pulmonary vascular resistance and promote myocardial contractility. General measures include increasing Fio₂, avoidance of hypercapnia and acidosis, and administration of nitrous oxide. Selective pulmonary vasodilators, inhaled nitric oxide, and prostacyclin are superior to intravenous vasodilators, which may cause systemic hypotension and compromise right-heart perfusion and preload. Milrinone and dobutamine are used as inotropes and pulmonary vasodilators but with caution, as they also cause systemic hypotension. Pulmonary pressure management is particularly challenging in the postreperfusion period, and pulmonary vasodilators are weaned gradually to avoid rebound pulmonary hypertension. Extubation should be delayed if inhalational vasodilators agents are needed (Mandell, 2004; Teo & Greenhalg, 2010).

With implementation of the MELD scoring system, more patients with renal insufficiency are being treated with OLT. Renal failure is associated with increased short- and long-term mortality after OLT, and perioperative renal failure often requires postoperative RRT. Intraoperative renal protection strategy is based on the maintenance of adequate intravascular volume and renal perfusion. No specific intraoperative renoprotective agents have been proven effective, but it is reasonable to maintain perfusion pressure with vasopressin and adrenergic agonists, which are effective in management of HRS. Surgical techniques that preserve vena caval flow, such as a piggyback technique, favor better renal outcomes. Intraoperative RRT is feasible and has been proven safe and successful in maintaining electrolyte and fluid balance (Ramsay & Garcia-Valdecasas, 2009; Townsend et al, 2009), although many transplant units try to avoid the need for it during the course of an OLT if at all possible. New markers of renal injury, such as neutrophil gelatinase–associated lipocalin, may be more useful than SCr in assessing acute kidney injury (AKI) (Niemann et al, 2009). The prognostic value of this rapid assay may prove useful in future studies to confirm the validity of intraoperative renoprotective strategies.

The administration of RBCs, FFP, and platelets has been associated with increased mortality (Bennett-Guerrero et al, 2001; de Boer et al, 2008; Massicotte et al, 2005; Pereboom et al, 2008; Ramos et al, 2003). Average blood-product requirements have decreased dramatically over the past decade, and now OLT can often be performed without transfusion (De Boer et al, 2005; Ozier & Tsou, 2008). Although it would be useful to prospectively identify patients at high risk of major bleeding, factors predictive of higher transfusion requirements are still debatable. Low pretransplant hemoglobin and a MELD score higher than 30 are better established predictors than previous abdominal surgery, prolonged INR, and low platelet counts (Massicotte et al, 2009; Ozier & Tsou, 2008; Xia et al, 2006).

No thresholds have been universally accepted for transfusing RBCs, FFP, and platelets, and practice differs among various centers and clinicians. POC testing, including TEG and ROTEM, have been used to define transfusion algorithms (see Monitoring of Hemostasis above).

The selective use of antifibrinolytic agents in the presence of persistent fibrinolysis is indicated, but routine prophylactic use of these agents is controversial. Aprotinin and tranexamic acid have been shown to reduce transfusion requirements in high-risk patients (Dalmau et al, 2000; Molenaar et al, 2007). Based on several case reports focused on the association of antifibrinolytics with vascular thrombosis (Dalmau et al, 2000; O’Connor et al, 2000) of either the arterial or venous circulation, and considering the current low transfusion rate, we question the validity of an indiscriminate prophylactic treatment in the absence of clinical and laboratory signs of hyperfibrinolysis. A recent meta-analysis (Molenaar et al, 2007) and a retrospective study of aprotinin by the same group (Warnaar et al, 2009) did not provide evidence for an increased risk of thromboembolic events associated with antifibrinolytic drugs in liver transplantation. It should be emphasized that in both of these studies, trends were observed that suggested a possible association between antifibrinolytics and thrombosis that did not reach statistical significance, perhaps because of the limited number of subjected enrolled and other confounding factors.

No consensus exists regarding the choice of any specific antifibrinolytic agent, dose, or time of administration; therefore monitoring of fibrinolysis and early detection of hypercoagulability are of critical importance. The prevailing view is that TEG/ROTEM may best provide such information, and TEE may be useful for detection of intracardiac clots adhering to the central lines. Use of activated factor VII as a prophylactic treatment has not been proven advantageous in a liver transplant setting, but it has been used successfully as a rescue treatment in a few individual cases (da Silva Viana, 2006). Cryoprecipitate, a source of factors I and VIIIc and of vWF, and prothrombin complex concentrates that contain vitamin K–dependent factors II, VII, IX, and X are increasingly used as an alternative to FFP. The advantage of these products over FFP is that a smaller volume is required, and risk of infectious complications is reduced (Ofosu et al, 2008; Samama et al, 2008).

Pulmonary Embolism

Pulmonary air and thromboembolism are rare but potentially fatal complications of OLT. With air embolism, a simultaneous laceration of a large vein, such as the IVC or hepatic veins, and a low CVP allow active entrainment of air into the central circulation; in addition, partial disconnection of the venovenous bypass circuit can also lead to air entrapment in the central circulation.

Intravascular clotting may manifest with pulmonary embolism, and intracardiac thrombi may occur with or without venous thrombosis and are frequently adherent to central venous lines. The mechanism of intravascular clotting in OLT patients is not fully understood. Some ESLD diagnoses, such as primary biliary cirrhosis or primary sclerosing cholangitis, are associated with hypercoagulability. In other cases, intravascular clotting has been related to administration of antifibrinolytics or transfusion of blood products (Warnaar et al, 2008a). Use of aprotinin in patients undergoing OLT may be associated with an increased incidence of venous thrombosis (Warnaar et al, 2009). Still, in a single-center case series, four episodes of PE were reported in the absence of antifibrinolytic use (Lerner et al, 2005).

Paradoxical embolus is a possible complication when an anatomic or functional atrial septum defect is present. Such complication is more likely to happen in pediatric recipients and in the early postreperfusion period, when right artery pressure is often higher than left artery pressure. During the pretransplant evaluation of pediatric patients, it is worthwhile to rule out such a defect or to plan for an atrial septal defect percutaneous closure when the defect size is significant and unlikely to heal before the transplant date.

In a recent single-center series, reported incidence of pulmonary embolism was 3.3% (El-Baghdadi & Sakai, 2010); however, very high overall and intraoperative mortality were reported in a recent review focused on pulmonary embolism in an OLT setting (Warnaar et al, 2008a): hospital mortality was 68%, and intraoperative mortality rate 82%.

The most common initial indication of intraoperative pulmonary embolism is usually systemic hypotension with a simultaneous increase in pulmonary pressure (Ellenberger et al, 2006; Warnaar et al, 2008a). End-tidal CO₂ may also decrease. A reasonable preventive strategy is based on prompt identification of hypercoagulable conditions with TEG/ROTEM and avoidance of unnecessary prophylactic administration of antifibrinolytics (Planinsic et al, 2006). TEE can be a diagnostic tool, but its role as a preventive tool is less clear. Treatment is based on hemodynamic support, and successful intraoperative administration of recombinant tissue plasminogen activator has been reported (Jackson et al, 2006).

Acute Liver Failure

Acute liver failure is defined as a severe liver dysfunction causing encephalopathy and coagulopathy (INR >1.5) within 24 weeks of the onset of symptoms in patients without previous liver disease (see Chapters 72 and 97C); it is considered fulminant ALF if it develops in less than 2 weeks. Overall transplant-free survival in ALF is dependent on etiology and is just above 40%; it is highest for acetaminophen/paracetamol overdose, which requires transplantation in less than 10% of patients (Trotter, 2009). Liver transplantation is the only proven therapeutic intervention in advanced ALF, accounting for just over 5% of all liver transplantations, and carries a 1-year survival of 60% to 80% (De Gasperi et al, 2008; Trotter, 2009; see Chapter 97C).

Increased ICP and severe coagulopathy are the hallmarks of ALF and are also the main perioperative challenges. As cerebral edema is the most common cause of death in ALF, control of intracranial hypertension and maintenance of adequate cerebral perfusion pressure (CPP) are critical intraoperative tasks. Widely accepted minimal goals are ICP of less than 25 mm Hg and CPP of 50 to 60 mm Hg or higher. Continuous invasive monitoring of ICP is ideal for guiding management, but it is associated with a high risk of bleeding.

ICP typically arises during the dissection phase and peaks after reperfusion. A simultaneous and precipitous drop in MAP makes postreperfusion the most critical and challenging period for management of CPP. A protocol that included intraparenchymal ICP monitoring, maintenance of CPP greater than 60 mm Hg by administration of vasopressors, hyperventilation (Pco₂ at 30 to 35 mm Hg), mannitol and 3% saline, mild hypothermia (33° to 34° C), and a high dose of pentobarbital titrated to ICP effect resulted in favorable neurologic outcomes in a small series of ICU and liver transplant patients (Raschke et al, 2008).

No randomized trials or consensus statements are available that stipulate the safety and efficacy of different types of ICP monitors or specific management protocols during liver transplantation, but it is increasingly recognized that clinically meaningful inferences may require multimodal approaches, including brain tissue oxygen monitoring and microdialysis (Stevens & Merritt, 2008). Interventions aimed at management of CPP are outlined in Table 97B.1.

Despite profound coagulation abnormalities in ALF (Box 97B.1), serious bleeding is rare during liver transplantation, because ALF patients have minimal or no portal hypertension (Muñoz et al, 2009). No evidence-based protocols have been outlined for correcting coagulopathy in ALF, but it is recognized that prophylactic administration of FFP and other products simply to correct the prothrombin time is not justified. Correction of coagulopathy is important before placement of an ICP transducer. Whole-blood assays (TEG, ROTEM) and clinical assessment are more useful guides for intraoperative treatment of coagulopathy. A major limitation of FFP is the potential for volume overload and transfusion-related acute lung injury (TRALI). Cryoprecipitate is typically indicated when the fibrinogen level falls below 100 mg/dL; activated recombinant factor VII (rFVIIa) and thrombin concentrates are beneficial, but optimal use of these last two remain to be determined. Antifibrinolytics should be considered if clinically significant fibrinolysis is proven.

Both high-dose steroids given during the anhepatic period and a functioning graft may cause hyperglycemia in the neohepatic phase. As hyperglycemia has an adverse effect on ischemia-reperfusion brain injury, plasma glucose should be kept in the normal range, and continuous infusion of insulin is often required. Restoration of liver function by the functioning graft gradually eliminates encephalotoxins and improves metabolic homeostasis, which eventually leads to reduction of brain edema and normalization of intracranial and cerebral perfusion pressures.

Combined Liver-Kidney Transplantation

With the introduction of the MELD scoring system, the percentage of simultaneous liver-kidney transplantation doubled, to about 4.4% of all liver transplantations performed in the United States (Gines & Schrier, 2009). Although it is obvious that combined transplantation should be reserved only for patients with irreversible renal failure, reliable predictive factors for reversibility are sometimes difficult to fully define. Current criteria recommended by the International Liver Transplant Society expert panel (Gonwa & Davis, 2009) include patients who, in addition to having liver disease, are on dialysis because of end-stage renal disease; patients with AKI that requires dialysis for at least 6 weeks; and those not on dialysis but who have a low glomerular filtration rate (<30 mL/min) and proteinuria (3 g/day, 24-hour urine protein/creatinine ratio >3).

A matched case-control analysis of patients dialyzed for more than 3 months demonstrated a favorable impact of combined liver-kidney transplantation compared with liver transplantation alone. Patient survival was improved (87.2% vs. 74.5%), as was liver-graft survival (84.5% vs. 70.8%), 1 year after transplantation (Locke et al, 2008). In contrast, 1-year survival of the kidney graft was lower than after kidney transplantation alone (77.2% vs. 89.3%). The inferior renal results and the priority given to the kidney on the long waiting list point to a need for a standardized approach to combined liver-kidney allocation and for a better understanding of reversibility of renal dysfunction (Davis, 2008).

Patients presenting for a combined liver-kidney transplantation present a particular anesthetic challenge. Electrolyte abnormalities and metabolic acidosis may mandate preoperative hemodialysis, whereas hyperkalemia greater than 5.5 mmol/L and hyponatremia of less than 120 mmol/L are typically considered prohibitive to transplantation. Intraoperative monitoring of preload and strategies for judicious volume administration and correction of electrolyte abnormalities are essential. Hyperkalemia is common and should be managed aggressively by maintaining a normal pH and administering insulin or glucose, inhaled β-adrenergic agonists, and furosemide, if urine output is preserved. Banked blood is a major source of potassium; ideally, it should be washed, and units with the longest remaining shelf life should be preferentially used.

Intraoperative continuous RRT is possible, and its use is routine in some centers. In a series of 41 OLTs, this modality was well tolerated hemodynamically, and it allowed achievement of neutral fluid balance without a need for systemic anticoagulation (Townsend et al, 2009). Fluid strategies for the two parts of the combined procedure may seem conflicting; although the kidney transplantation part may prompt a somewhat more liberal fluid regimen, in our opinion, goals for both OLT and kidney transplantation should be maintenance of a normal range preload to allow optimal organ perfusion.

Retransplantation

Most common reasons for retransplantation are primary nonfunction, recurrence of viral hepatitis, acute or chronic rejection, and thrombosis of the hepatic artery (Ma et al, 2008; Pfitzmann et al, 2007). The time interval between the primary transplantation and retransplantation—that is, early versus delayed retransplantation—depends on the etiology of the graft failure.

Almost 11,000 retransplantations have been performed since the inception of the UNOS registry—about 11% of all the OLTs ever performed. A declining trend toward the need for retransplantation is apparent; for example, in 1988, the retransplantation rate was 24.5%, whereas in 2009, the rate was only 7.3%.

The survival rate for retransplanted patients is lower than for primary transplant recipients: at 1 year, the survival is 67.2% versus 83.4%; at 5 years, it drops to 46% versus 67.4% (UNOS, 2010). With appropriate recipient selection, better outcomes are possible, and 1- and 5-year survival rates have been reported as high as 78% and 67%, respectively, with retransplantation (Ma et al, 2008; Pfitzmann et al, 2007). General agreement has been reached on two points: 1) that whole graft should be used in adults (Pfitzmann et al, 2007) and 2) that retransplantation should not be undertaken with extended-criteria donors (Zimmerman & Ghobrial, 2005).

The current modest results emphasize the need for strategies to improve retransplantation survival. In particular, a strong interest has been shown in the development of valid predictive criteria; for example, according to Zimmerman and Ghobrial (2005), patients who require mechanical ventilation, those who have advanced renal failure, and patients of advanced age have worse outcomes and should not be considered for retransplantation. MELD score has prognostic value in retransplantation: patients with of a score below 20 have better outcomes than those with higher scores (Ma et al, 2008); in another study, worse results were reported when MELD was more than 25 (Zimmerman & Ghobrial, 2005).

A few studies have also looked at survival as a function of the time interval between the primary transplantation and retransplantation; survival was better when retransplantation occurred less than 8 or more than 30 days after the initial transplantation (Ma et al, 2008).

Intraoperative anesthetic challenges depend on the overall patient condition on the day of surgery; factors that may impact transplantation include difficult vascular access, a need to support liver and coagulation function, increased blood losses, and a higher requirement for blood-product transfusions. The main causes of death were bacterial infection, multisystem organ failure, bleeding, and recurrence of disease (Pfitzmann et al, 2007).

Pediatric Transplantation

Indications for liver transplantation in the pediatric population include 1) primary liver disease, such as extrahepatic biliary atresia, and intrahepatic cholestasis, which includes sclerosing cholangitis, Alagille syndrome, and other similar syndromes; 2) metabolic diseases, such as Wilson disease, α₁-antitrypsin deficiency, Crigler-Najjar syndrome, disorders of the urea cycle, oxaluria type 1, and disorders of carbohydrate metabolism; 3) ALF; and 4) liver disease brought about through other causes, including primary liver tumor and cystic fibrosis (Spada et al, 2009; Yudkowitz & Chietero, 2005; see Chapter 71).

According to UNOS data, more than 12,000 OLTs have been performed since 1988 in patients younger than 18 years, with the majority carried out in patients younger than 5 years.

Living-Related Donor Liver Transplantation

Since 2006, fewer than 300 OLTs with a graft from a living relative donor have been performed annually in the United States—about 3% of adult and 10% of pediatric OLTs (see Chapter 98B). This is a marked decline with respect to the year 2001, when a record 524 OLTs came from living donors—about 9% of adult and 18% of pediatric OLTs. Although survival of the recipients of a living graft is superior to that of cadaveric graft recipients (UNOS, 2010), a justified concern for donor morbidity and mortality drives this trend. Challenges for the anesthesiologist depend more on the recipient’s characteristic and comorbidities than on the donor category (Table 97B.2).

Table 97B.2 Early Complications After Liver Transplantation

ARDS, acute respiratory distress syndrome; ATN, acute tubular necrosis; CHF, congestive heart failure; CNI, calcineurin inhibitor; IVC, inferior vena cava; TRALI, transfusion-related acute lung injury

Early Extubation

Safe early or even immediate postoperative extubation of a majority of liver transplant patients has been reported by some centers; however, the exact criteria, the benefits of early extubation to liver recipients, and any cost benefits remain a matter of debate. A multicenter cohort study by a proponent group (Mandell et al, 2007) reported a lower rate (7.7%) of postoperative adverse events in the immediate extubation group than that commonly reported for ventilated liver transplant patients. This study showed great variability in rates of extubation (5% to 67%) and rates of complications between centers, suggesting a significant impact of the regional patterns of practice, intraoperative events, and physician experience. Multiple studies agree that encephalopathy, whether the result of acute or chronic liver failure, is a strong predictor of the need for prolonged ventilation. Other factors, such as volume of blood transfused and RRT, are less consistent predictors among studies and may be only indirect markers of poor graft function, severity of illness, and a lack of standardized intraoperative protocols (Mandell et al, 2009).

No randomized studies are currently available that demonstrate that early extubation improves patient outcomes or reduces cost (Ozier & Klinck, 2008). Although experienced clinicians can select appropriate patients for early extubation, no evidence-based criteria exist in support of this practice; more studies are needed to establish whether this intervention is beneficial and whether it should be a therapeutic goal in liver transplant recipients.

Medical Complications

Pulmonary complications that include acute respiratory distress syndrome (ARDS), TRALI, pneumonia, and pleural effusions are common in the immediate postoperative period. These complications are reported to occur in over 50% of patients after liver transplantation and lead to hypoxemia that often requires prolonged mechanical ventilation (Biais et al, 2009; Bozbas et al, 2008; Li et al, 2010). ARDS and TRALI may be related to blood-product transfusion and fluid overload. Pneumonia is the most frequent postoperative infection in patients following OLT, and it carries a high mortality. Pleural effusion is another common pulmonary complication; it promotes atelectasis and contributes to impaired respiratory exchange. The addition of PEEP monitoring may be effective in improving oxygenation in postoperative patients. PEEP of up to 10 cm of H₂O was found not to affect the perfusion of the grafted liver, as assessed by Doppler (Saner et al, 2006b).

Cirrhotic patients continue to exhibit hyperdynamic circulation during the first hours or days after OLT. Arterial hypotension and the need for vasopressors are common. With the graft functioning and no evidence of surgical complications, the hemodynamics tend to normalize, which allows downgrading of invasive monitoring and of hemodynamic support. Maintenance of kidney perfusion pressure is paramount because AKI is common (17% to 95%) (Pham et al, 2009) and is associated with increased ICU length of stay and mortality (Smith et al, 2009). The causes of post-OLT acute renal failure are multifactorial and include preexisting renal dysfunction, HRS, perioperative acute tubular necrosis, and CNI nephrotoxicity. RRT is necessary in 5% to 35% among patients who develop postoperative AKI.

Endocrine and metabolic complications are common in the early posttransplant period. Surgical stress, corticosteroid administration, and CNIs contribute to early posttransplant hyperglycemia, which frequently evolves to diabetes mellitus, and hypertension. Diabetes occurs in up to 30% of posttransplant patients, and hypertension is seen in over 50%. Hyperglycemia requires an intravenous insulin regimen and improves as steroids are tapered; hypertension should be treated aggressively to prevent cardiovascular complications (Kallwitz & Cotler, 2008).

Neurologic complications may occur in up to 50% of liver transplant recipients (Saner et al, 2009). Such complications contribute to longer ICU and hospital stays and may result in long-term poor quality of life. The most common neurologic complications are encephalopathy and seizures, followed by psychotic disorders, stroke, and poorly understood posterior leukoencephalopathy. The etiology is multifactorial and includes poor graft function, use of steroids and possibly CNIs, infection and systemic inflammatory response, electrolyte imbalance, and embolic events.

Meticulous perioperative neurologic assessment and quick recognition of neurologic impairment are essential. Specific therapeutic modalities are limited (i.e., tapering steroids, exchanging CNIs), but if applied in a timely manner, they may improve outcome. In addition, a key factor of neurologic improvement is the recovery of graft function.

Infections are the primary cause of death after liver transplantation. Many infections that occur in the first days after transplantation are related to the surgery itself, such as wound infections, seromas, and perigraft fluid collections. Postoperative and opportunistic infections are common in the immunocompromised and debilitated transplant patient. Such infections may lead to sepsis and multiorgan failure; therefore they require a high level of suspicion and routine postoperative prophylactic protocols that include broad-spectrum antibacterial, antiviral, and antifungal therapy.

Posttransplant rejection ranges from 10% to 30% (Mueller et al, 2004). Cellular rejection is suspected with increasing transaminases and normal hepatic vascular Doppler studies, and diagnosis is confirmed with liver biopsy; however, early episodes respond well to boluses of corticosteroids and rarely lead to graft failure.

Surgical Complications

Postoperative hemorrhage is primarily a clinical diagnosis that can be confirmed by abdominal imaging techniques, and it occurs in 10% to 15% of OLT patients. A priority is to distinguish whether hemorrhage is due to persisting coagulopathy or to surgical causes. Hemodynamic instability and transfusion of more than 4 to 6 U packed blood cells over a 24-hour period are usual indications for surgical reexploration (Mueller et al, 2004).

Early vascular complications include hepatic artery and portal vein stenosis or thrombosis and venocaval obstruction or kinking. Poor graft function and protracted elevation of liver enzymes point to a vascular complication; however, the diagnosis is made based on a bedside vascular Doppler and can be confirmed with conventional or MRI angiography. Hepatic artery thrombosis occurs in 2.5% to 10% of adult patients and in 5% to 20% of children; portal vein thrombosis is much less common (0.3% to 2.2%), and flow obstruction at the level of the infrahepatic and suprahepatic vena cava occurs in 1% to 2%.

All vascular complications require urgent surgical intervention. If vascular repair is delayed, the ensuing graft failure may necessitate retransplantation. Many vascular complications have been successfully treated using various endovascular approaches, including percutaneous angioplasty, stent placement, or thrombolysis (Amesur & Zajko, 2006).

Biliary complications occur in about 10% to 15% of patients following liver transplantation. Stricture of the bilary tree accounts for most of the complications, but biliary leak is a more common early complication. The most severe manifestations of biliary leak are peritonitis and sepsis, but asymptomatic bilomas are also a common presentation. The diagnosis of the leak is based on percutaneous transhepatic cholangiography (PTC) or endoscopic retrograde cholangiography (ERCP). Ultrasound and CT scan are useful in diagnosis of biliary dilations due to obstruction, bile collections, or abscesses. Treatment of biliary obstruction includes percutaneous balloon dilation and drainage or surgical revision. Treatment of biliary leaks involves percutaneous transhepatic drainage (PTD), a technically difficult alternative to surgical repair. Nonanastomotic leaks and leaks secondary to hepatic artery thrombosis are particularly ominous, because they are less amenable to percutaneous repair and often require retransplantation (Amesur & Zajko, 2006; Mueller et al, 2004).

Recipients of grafts from donors after cardiac death (DCD) have worse graft and patient survival, a higher rate of postoperative complications, and higher resource utilization than the recipients of dead-brain donation (DBD) grafts (Reich et al, 2009). Primary graft nonfunction, nonanastomotic biliary strictures, and hepatic artery stenosis are the most significant postoperative complications of DCD liver transplantation. Higher immediate perioperative mortality in DCD transplants has been reported (Pine et al, 2009), but most complications lead to increased mortality or to the need for retransplantation in the weeks or months following the transplantation (Grewal et al, 2009). Strategies to improve DCD outcomes currently under development include a super-rapid recovery technique by experienced procurement teams, premortem cannulation techniques, ex vivo organ perfusion, and minimization of cold- and warm-ischemia time (Reich et al, 2009). Although DCD grafts are associated with inferior outcomes, DCD donation remains a valuable resource, and with selective and judicious use, these can be used to achieve acceptable results (Pine et al, 2009).