Sedation of the mechanically ventilated patient is challenging; this is especially true in the post-liver transplant patient. Mental status changes can be an early clue to graft dysfunction and efforts should be made to avoid excessive sedation. Benzodiazepines are not recommended as they have been shown to increase delirium in the ICU setting [40]. Propofol and dexmedetomidine have become popular sedatives due to their favorable pharmacokinetics. A recent meta-analysis suggested that use of dexmedetomidine or propofol infusions rather than a benzodiazepine infusion in critically ill adults reduced ICU length of stay and duration of mechanical ventilation [41]. There have been only a few recent case reports of safe use of dexmedetomidine infusions in post-liver transplant patients [42, 43]. The short acting agents have a favorable profile and allow for serial neurologic exams while still providing adequate sedation and anxiolysis.

Liver transplant is a major surgical procedure and may be accompanied by significant postoperative surgical pain. Pain control intra- and postoperatively is usually achieved with fentanyl, via infusion or intermittent bolus. Other opioids such as morphine and hydromorphone are avoided if possible due to their prolonged half-lives in liver failure. Fentanyl derivatives such as sufentanil, alfentanil, and remifentanil have superior pharmacokinetic properties but are not routinely used in the postoperative setting due to higher cost, insufficient staff experience, and lack of data showing improved efficacy. Some patients may require use of a patient controlled analgesia (PCA) pump along with longer acting agents, or transition to around-the-clock oral medications if pain persists.

Thoracic epidurals are beneficial for pain control following abdominal surgery [44], however they are not routine for liver transplant patients. The varied coagulation status of posttransplant patients raises concerns regarding the use of thoracic epidurals for postoperative analgesia. Hypotension from the epidural also raises concern that graft function may be compromised, particularly in posttransplant patients who have complex hemodynamic indices. For certain patients, other than transplant recipients (i.e.: hepatectomy patients) thoracic epidurals may be an acceptable option for postoperative analgesia.

Non-opioid adjuncts for pain control have received significant attention. While there have been few studies examining these agents in liver transplant patients, some generalizations can be made. Nonsteroidal anti-inflammatory drugs (NSAIDs) , while efficacious for pain, should probably be avoided in the setting of increased bleeding risk and potential renal insufficiency. Acetaminophen is usually given at lower doses (2 g/day) for liver failure patients and should be avoided in the immediate postoperative period. However, for patients with functioning grafts, it is reasonable to consider acetaminophen administration due to its synergy with opioids.

Unfortunately there does not exist a one-size-fits all approach to pain management in the liver transplant patients. Each patient’s individual risk for postoperative pain must be weighed against potential side effects. At this point, opioids such as fentanyl remain the mainstay of therapy until further studies are completed that can validate the safety of other interventions.

Patients with liver failure often suffer from hepatic encephalopathy . The underlying etiology of hepatic encephalopathy is not entirely understood but current theories suggest that increased ammonia in the systemic circulation crosses the blood brain barrier where it is converted into glutamine by astrocytes. The glutamine causes swelling of the astrocytes, which impairs neurotransmission regulation. Interestingly, the level of ammonia does not correlate with neurologic symptom severity, so trending ammonia levels may not be helpful. In the postoperative period, a patient with a newly functioning liver should have steady clearance of toxins and a continual improvement in mental status. If there is no improvement or mental status declines, then a workup for graft nonfunction and infection should be undertaken and electrolyte imbalances corrected. Given the extreme changes in coagulation status, there should also be a low threshold to obtain imaging if there is concern for intracranial hemorrhage.

Hyponatremia in the setting of liver failure will be discussed below. However, it is important to note a potential neurologic complication that is associated with rapid correction of hyponatremia: central pontine myelinolysis or osmotic demyelination syndrome. The exact etiology of osmotic demyelination syndrome is unknown. The symptoms are usually seen 1–6 days after the insult of rapid sodium correction [45]. The most common clinical manifestation is fluctuations in consciousness. Eventually, pseudobulbar palsy and quariparesis may develop. If a patient is known to be hyponatremic preoperatively, then clinicians must closely monitor electrolytes and choose intravenous fluids appropriately to avoid rapid over correction postoperatively.

Alterations in sodium levels are very common in pre- and posttransplant patients. In fact, there is clinical evidence to suggest that adding serum sodium to model for ESLD (MELD) scoring improves mortality prediction [46]. The first step in management is to determine the acuity of the situation. Patients with acute hyponatremia (development in under 48 h) are at risk for developing neurologic impairment and, consequently, require prompt correction of serum sodium levels. Administration of a hypertonic (3 %) saline infusion may be necessary for this situation. In patients with more chronic hyponatremia (development in over 48 h), rapid correction of hyponatremia is an independent risk factor for the development of posttransplant neurological complications [47]. Serum sodium correction should be performed in a controlled manner in this instance. The goal is usually 1–2 mmol/L per hour for the first 48 h. If the level rises too quickly, then hypotonic intravenous fluids should be started to restore the goal correction rate.

Hypernatremia is a less-frequent complication associated with liver transplant patients. The etiology is frequently related to excessive loss of free water in patients using an osmotic laxative (such as lactulose) to reduce hepatic encephalopathy. These patients are unable to adequately regulate their own free water balance due to impaired thirst mechanisms. This derangement may continue into the postoperative setting. As the mental status improves, the patient should begin to appropriately regulate water intake. For a hypernatremic patient who is unable to tolerate oral free water boluses, hypotonic maintenance fluids are recommended with close monitoring of electrolytes.

Hyperkalemia may be the most lethal electrolyte abnormality due to the rapid progression of arrhythmias and death. The causes of hyperkalemia in the posttransplant patient are often multifactorial. Many liver transplant patients either have pre-existing renal dysfunction [48] or will develop transient renal dysfunction in the perioperative period which can impair mechanisms of potassium homeostasis.

For patients that had significant blood loss and transfusion requirements during the operation, there may be a significant potassium burden in the form of lysed cells from aged units that are transfused. Many liver transplant centers have a high usage rate of blood products and will often be assigned aged units by the blood bank because they are unlikely to be wasted. While this is an excellent use of resources, these units contain less-functional cells and correspondingly represent a higher potassium load to the patient. Washing the cells before transfusion can partially attenuate the hyperkalemia, but frequent potassium monitoring remains necessary.

Hyperkalemia can be exacerbated acutely by reperfusion of the preserved graft and release of a significant potassium load from ischemic tissues. This is often managed with temporizing measures such as administration of calcium, sodium bicarbonate, and insulin with glucose, but the total body potassium may continue to be elevated in the postoperative setting. Dialysis may be needed if renal insufficiency and hyperkalemia persist in the ICU.

Hypocalcemia is frequently identified in liver transplant patients. However, it is important to remember that these patients often have low albumin levels and the total calcium is not necessarily reflective of free calcium levels [49]. Ionized calcium levels are more accurate in this situation. Low calcium levels can result from chelation with the anticoagulant citrate, found in blood products and renal replacement therapy infusions. Hypocalcemia should be suspected in a patient with hypotension despite adequate resuscitation. Calcium gluconate or calcium chloride can be used for replacement.

Glucose levels following liver transplantation have significant implications for both prognosis and complications. Hypoglycemia in the postoperative period may be a marker for sepsis or poor graft function [2]. Hyperglycemia, which is much more common in the postoperative setting, may be a reflection of underlying diabetes, stress response, or steroid administration. Severe hyperglycemia (glucose > 200 mg/dL) is associated with an increased risk of liver allograft rejection [50], surgical site infection [51], and increased mortality [52]. Hyperglycemia is known to aggravate ischemia reperfusion injury in several organ systems.

Although hyperglycemia has complications, tight glucose control (between 80 and 120 mg/dL) is not recommended due to poor outcomes in the ICU setting [53, 54]. The best approach is to achieve modest glucose control (150–180 mg/dL), which is consistent with current ICU guidelines. In the immediate postoperative setting, an insulin infusion with frequent blood glucose checks is often required, as fluctuations in the stress response make steady state dosing difficult.

Once renal failure begins to develop, nephrotoxic immunosuppressants, namely CNIs, should be withdrawn, and immunosuppression should be maintained with renal-sparing protocols. CNI-induced nephropathy results from afferent arteriolar vasoconstriction and subsequent decrease in renal perfusion [63]. Using a reduced dose of cyclosporine, or replacing cyclosporine with mycophenolate mofetil (MMF) and sirolimus reduces the incidence of CNI-induced renal injury [64]. While CNI-induced nephropathy was reduced with MMF and sirolimus, the incidence of biopsy-proven acute rejection in the liver increased. Fortunately, this was not associated with increased rates of graft loss. A recently conducted Cochrane review of the literature surrounding CNI toxicity did not reach a conclusion regarding the role of CNI minimization in preventing nephrotoxicity in liver transplant patients [65]. Many centers now delay the administration of these drugs following surgery. The dosages used today are also substantially lower than those prescribed in the past in order to reduce the subsequent risk of chronic kidney disease [56].

Opportunistic infections generally occur in the second through sixth months, when immunosuppression is most profound. Trimethoprim/sulfamethoxazole (TMP-SMX) prophylaxis for Pneumocystis jirovecii should be instituted for the first 6 months following transplant, and continued in patients requiring monoclonal OKT3 antibodies for rejection and in patients with graft dysfunction. An additional benefit of TMP-SMX administration is prophylaxis for Toxoplasma gondii, Listeria monocytogenes, and Nocardia asteroids [66].

CMV infection is notable for its association with increased opportunistic infections in liver transplant patients, including fungemia and bacteremia, and its association with transplant rejection [69]. Infection with CMV within the first year of transplant is associated with increased mortality. Effective prophylaxis can be provided with ganciclovir or valganciclovir for 3 months following transplant [70]. Herpes simplex virus (HSV) reactivation may occur posttransplantation, but antivirals used for CMV prophylaxis should also be effective in these patients. If the patient is not receiving CMV prophylaxis, acyclovir can be used for the prevention of HSV. Varicella vaccination should be administered prior to transplantation. Beyond 6 months, patients are no longer at risk for most opportunistic infections if the level of immunosuppression has been reduced.

Candida is the most common fungal pathogen following liver transplantation and accounts for nearly 80 % of postoperative fungal infections, followed by Aspergillus. Most fungal infections occur within the first 2 months following transplantation. Risk factors for opportunistic fungal infections are retransplantation, renal failure, and reoperation involving the thoracic or abdominal cavity [71]. The use of antifungal prophylaxis is highly variable between liver transplant centers, and can include nystatin suspension, fluconazole, amphotericin B, or no empiric prophylaxis [72].

Blood transfusions for bleeding are indicated to maintain adequate oxygen delivery. No firm transfusion threshold exists, but evidence in other patient populations suggests that a more restrictive strategy is appropriate. In the most recent clinical practice guidelines published, the taskforce, comprised of surgeons, anesthesiologists, and intensivists, felt there was good evidence to recommend a restrictive strategy of red blood cell (RBC) transfusion (hemoglobin < 7 g/dL) in critically ill patients with hemodynamically stable anemia [73]. Acute blood loss with hemodynamic instability should probably be addressed by more aggressive resuscitation with blood products. Further trials testing rigorous transfusion protocols are necessary, but the trend has been toward more restrictive transfusion practices.

Coagulopathy does not resolve immediately after transplantation and often persists into the postoperative ICU period. The etiology is multifactorial and can involve hyperfibrinolysis, disseminated intravascular coagulopathy, platelet activation, platelet sequestration within the graft, and the presence of heparin-like effect (HLE). Some patients are actually hypercoagulable posttransplant, which further complicates the evaluation of their coagulation status [81]. The cause of this hypercoagulability is not entirely clear but maybe due to impaired synthesis of antithrombin by the liver.

As the new graft improves in function, synthesis of coagulation factors should improve and laboratory values should return to baseline. While laboratory value correction may not correlate well with bleeding risk, it does correlate with improved graft function. Failure to see improvement in coagulopathy should prompt a work up for graft nonfunction and infection, two serious causes of impaired coagulation in the postoperative setting. Routine transfusion for laboratory abnormalities is not indicated unless there is evidence for ongoing bleeding and hemostatic problems [82]. Aggressive transfusion can worsen cardiac function and consequently graft perfusion, so it should be reserved as therapy for clinically significant bleeding.

In addition to hypofibrinogenemia from transfusions and blood loss, the new graft releases t-PA and tissue factor, which results in an accelerated fibrinolytic state that frequently causes significant consumption of fibrinogen in the post-reperfusion setting [83, 84]. Refractory bleeding should prompt an investigation for low fibrinogen and fibrinolysis. Administration of antifibrinolytic drugs has shown benefit in reduction of transfusion requirements, and with the small number of patients studied so far, there does not appear to be an increased risk in thrombotic events (Table 29.2). Due to the lack of definitive data, it is not routine practice to administer antifibrinolytics, but practice patterns may change with further results.

The prevalence of HLE in patients undergoing liver transplant is not uncommon, and can range from 25 to 95 % of cases [85]. Patients who have acute liver failure, primary nonfunction of the liver graft or require retransplant have a higher prevalence of HLE. The problem appears to be worse in patients with acute liver failure; however, the problem can persist in the posttransplant period regardless of the etiology of the liver failure [86].

The HLE can come from an exogenous source as well as an endogenous source. Residual heparin bound to the endothelium of the donor liver, which is perfused with heparin before clamping, is the exogenous source of heparin. The endogenous source comes from substances known as heparinoids. The increased release of heparinoids is thought to occur from activation of macrophages or hepatocytes following ischemic injury to the liver. There is currently no evidence for reversing the HLE and supportive care is the best treatment option. An infusion of protamine sulfate has been attempted, but did not result in reduced bleeding or transfusion requirements [87]. If impaired coagulation persists several days into the postoperative period, then a sepsis workup is indicated as infection can worsen the production of these heparin-like molecules.

Low platelet counts are a commonly seen abnormality in the posttransplant patient. The etiology for the thrombocytopenia is varied but is related to decreased circulation and decreased production. With severe cirrhosis, there is often significant sequestration of platelets in the spleen due to portal hypertension, and the new graft will also sequester platelets. There is decreased platelet production because of low thrombopoietin levels in liver failure patients [88]. In the postoperative period, massive blood transfusions can result in a dilutional thrombocytopenia. Finally, even if the platelet count is adequate, platelets in a patient with liver disease may have decreased function because of adenosine diphosphate-induced and collagen-induced aggregation [89]. Platelet function may be further impaired by uremia in the setting of coexistent renal dysfunction. Thromboelastography (TEG) may be beneficial in measuring platelet function [90], but definitive studies relating use of TEG in liver transplant patients are needed.

All coagulation factors except for factor VIII and von Willebrand factor are synthesized by the liver and are therefore decreased in the setting of severe hepatic impairment. Fresh frozen plasma (FFP) can replace these factors, but administration of plasma carries the risk of transfusion reactions and large volumes are often needed to reverse the laboratory coagulopathy [91].

For patients with refractory bleeding, many clinicians have used recombinant activated factor VII (rFVIIa) [92]. No randomized clinical trials have been conducted in postoperative liver transplant patients; however, case series have shown some benefit. There are risks associated with the off-label use of rFVIIa. Mayer and colleagues demonstrated increased risk of thrombosis with rFVIIa administration in patients presenting with intracerebral hemorrhages [77]. The exact role of rFVIIa in liver transplantation is unclear due to lack of data. Given the uncertainty, recommendations are that rFVIIa should be used only as “rescue therapy” in patients with severe life-threatening bleeding where other therapies have failed.

After liver transplantation, there are many types of graft rejection that may occur. Rejection of the allograft can be hyperacute, acute, chronic, or graft-versus-host (GVHD). Since chronic rejection is not usually an issue for the ICU patient, it will not be covered in this chapter.

Hyperacute rejection, mediated by antibodies, occurs within minutes to hours after the transplant procedure. Sixty percent of the cases of hyperacute rejection are due to ABO-incompatible allografts. In the presence of ABO-incompatible transplants, plasmapheresis, splenectomy, and the CD20 monoclonal antibody, rituximab, have been reported to prevent hyperacute rejection [104], but immediate retransplantation is often the only lasting option. Because antibody-mediated rejection with ABO-compatible allografts is so rare, due to the liver’s relative resistance to the humoral immune system, a positive crossmatch does not necessarily preclude liver transplantation. However, evidence does suggest that the presence of preformed donor-specific HLA-antibodies can increase the risk of acute cellular rejection and chronic rejection [26].

Unlike hyperacute rejection, which is B cell mediated, acute rejection is mediated by T cells. Acute rejection is usually seen within days or weeks of the transplant and occurs in 36–75 % of liver transplant patients. Acute rejection is characterized by mononuclear inflammation and active cell damage, and episodes refractory to antirejection medications (usually high-dose steroids) can progress to chronic rejection [105]. Risk factors for the development of steroid unresponsive acute rejection include pre-liver transplant steroid administration, ABO incompatibility, recurrent rejection, low serum cyclosporine levels, and high liver function tests. A rising or persistent elevation of alanine aminotransferase (ALT) levels should prompt a biopsy to exclude rejection. Treatment options for a positive biopsy depend on the severity and include: optimization of maintenance immunosuppression for mild rejection, steroid pulses for moderate or severe rejection, and T cell depletion therapies for severe rejection.

GVHD occurs in 1–2 % of liver transplant recipients and is associated with an 85 % mortality rate. In the case of solid organ transplant, donor lymphocytes remaining in the parenchyma become detectable in the recipient weeks after transplant. These immunocompetent cells react against the different cellular antigens found in the host. A humoral response leading to hemolysis can also occur due to organ ABO incompatibility [106]. GVHD is divided between acute (occurring within 100 days of transplant) and chronic (after 100 days) presentations. Risk factors associated with the development of GVHD include alcoholic liver disease, hepatocellular carcinoma, and diabetes mellitus. It has also been suggested that GVHD is more likely to occur in the setting of close HLA matching and autoimmune hepatitis. Symptoms usually develop 2–6 weeks posttransplant, and include fever, diarrhea, rash, and pancytopenia. Similar to treatment for acute rejection, treatment of GVHD includes administration of corticosteroids, increasing the current immunosuppressant regimen, or administration of medications for the antagonism of T cells. Mortality following the development of GVHD can be as high as 85 % [107]. Prevention includes limiting recipient exposure to donor lymphocytes, such as graft irradiation or treatment with monoclonal antibodies, and limitation of blood products to those that have been leukocyte reduced and irradiated.

Primary nonfunction occurs in 4–8 % of deceased-donor liver transplants. Although uncommon, it is the most serious and life threatening condition in the immediate postoperative period and can be the most challenging for the transplant service. It is caused by reperfusion injury of the new liver, and results in irreversible graft failure. The diagnosis of primary nonfunction can only be made in the absence of technical or immunologic causes for graft dysfunction [108]. The acute destruction of hepatocytes results in decreased bile production, coagulopathy, encephalopathy, hypoglycemia, lactic acidosis, and hemodynamic instability. Signs often are present intraoperatively, but correction of the metabolic disturbances will need to be aggressively continued in the intensive care unit. The risk factors for primary nonfunction are numerous, and include: prolonged cold ischemic time, increased donor age, donor hypernatremia, donor length of stay in the ICU, male recipients with female donors, reduced graft size, racial mismatch between donors, retransplantation, and hepatic steatosis [109–111].

The only treatment for primary nonfunction is early retransplantation, and primary nonfunction is the most common reason for early retransplantation [110]. Without retransplantation, mortality is high. In addition to the complications from liver failure, cardiovascular, renal, and respiratory failure can often result from the release of vasoactive mediators from the nonfunctioning liver. Often times, removal of the failing graft can lead to a dramatic improvement in the patient’s clinical status. In the absence of an immediately available liver, a rescue hepatectomy with portocaval anastomosis can be performed with subsequent liver transplantation occurring 24–48 h following hepatectomy [112].