Gastrointestinal and Hepatic Disorders

Introduction

Gastrointestinal (GI) issues and maladies and in particular GI bleeding, liver dysfunction, and pancreatitis can complicate the primary course of critically ill neurologic patients. In addition, patients with hepatic encephalopathy may be cared for in the neurocritical care unit (NCCU). The neurointensivist therefore must be prepared to recognize, coordinate specialty care, and provide local monitoring and support when these issues arise. This chapter reviews disorders of the GI system including the liver and pancreas that are of importance in the NCCU. A comprehensive review of GI and hepatic function is beyond the scope of this chapter. Nutrition is reviewed in Chapter 14 .

Gastrointestinal Bleeding

GI bleeding is relatively common in acutely ill hospitalized patients, and the risk may be augmented in the NCCU associated with severe sympathetic stress such as in acute subarachnoid hemorrhage (SAH) or neurotrauma patients, including both traumatic brain injury (TBI) and spinal cord injury (SCI) and in patients in whom steroids are administered. The bleeding source may be from the upper or lower GI tract. The ligament of Treitz is the landmark that distinguishes the upper from lower GI tract. Common causes for upper GI bleeding (UGIB) are peptic ulcer disease, including stress-induced ulcer bleeding, variceal hemorrhage, and esophagitis. Common causes for lower GI bleeding (LGIB) are diverticulitis, angiodysplasia, colitis that can have several causes, tumors, and local anorectal pathologies. UGIB presents with hematemesis, melena, hematochezia, and occasionally hypotension when more severe. In patients with a nasogastric tube in place, the bleeding may manifest in the drainage contents (e.g., coffee-ground gastric contents). LGIB presents with hematochezia and melena, and can be a reason for anemia.

When GI bleeding occurs the initial focus is to ensure hemodynamic stability with resuscitation using crystalloid or blood products as appropriate. Patient monitoring therefore is aimed at recognition and avoidance of hemorrhagic shock using reliable real-time continuous blood pressure and heart rate monitoring. In more severe GI bleeding, resuscitation can be facilitated using central venous pressure (CVP), mixed venous oxygen saturation (ScvO ₂), and newer noninvasive hemodynamic monitoring techniques. These noninvasive devices can provide data on cardiac performance, preload, and systemic vascular resistance, similar to that provided by Swan-Ganz catheterization (see Chapter 19 ).

Laboratory assessment includes hemoglobin and hematocrit, and to exclude an occult bleeding diathesis platelet count, coagulation indices, and if necessary disseminated intravascular coagulation (DIC). The brain contains high levels of tissue factor (TF), and in severe injury or associated with a prolonged or complex surgery (e.g., for a large arteriovenous malformation [AVM]), significant TF release may occur and induce the DIC cascade. The optimal platelet count after an injury to or a disease of the nervous system may depend on pathology and time after the initial insult. There is general agreement that a platelet count greater than 100,000 is a reasonable goal after TBI, after neurosurgery, or for anticipated intervention or instrumentation of the central nervous system (CNS). This includes for placement of an intracranial monitor or external ventricular drain. By contrast for other patients without CNS disorders or likely surgery, a platelet count greater than 50,000 is reasonable perhaps even when there is GI bleeding. For the neurointensivist this means that platelet support may differ depending on whether GI bleeding occurs in a patient with a disorder such as myasthenia gravis or severe TBI or is a postoperative neurosurgical patient. The various laboratory indices should be repeated at intervals until they are stabilized. This typically includes assessment of complete blood count with platelets and if necessary coagulation studies every 6 hours.

GI bleeding requires an assessment of the source of bleeding and should include an early consultation with a gastroenterology service. Endoscopy typically is the initial means to identify the source of UGIB, determine whether active bleeding still is present, and guide the frequency of monitoring, including repeat endoscopy or other diagnostic tests. In addition, the source of bleeding may be treated through endoscopic techniques. An intravenous bolus of proton pump inhibitors followed by an infusion is recommended to prepare the patient for endoscopy associated with nonvariceal bleeding. Somatostatin or octreotide can be considered for variceal bleeding. The Blatchford score, which is based on several factors including urea and hemoglobin levels, systolic blood pressure, heart rate, hepatic disease, cardiac failure, melena, and presentation with syncope, can be used to determine which patients require urgent endoscopy. Once endoscopy is complete the Rockall score, which includes the endoscopic diagnosis and evidence for hemorrhage among other variables, can be used to stratify patients into high or low risk for recurrent hemorrhage or mortality. Bedside colonoscopy is feasible when LGIB is identified, but whether urgent colonoscopy is associated with better outcome than a standard care algorithm based on angiography and expectant colonoscopy is unclear. Current guidelines for evaluation, monitoring, and management of UGIB, variceal bleeding, and LGIB are provided in Table 23.1 ; Table 23.2 ; and Figure 23.1 .

Table 23.1

Summary of Recommendations from 2010 Guidelines of the International Consensus Upper Gastrointestinal Bleeding Conference Group *

	Resuscitation, risk assessment, and pre-endoscopy management: 1. Urgent/emergent evaluation and initiation of resuscitation 2. Use appropriate prognostic scales for mortality and rebleeding to stratify patients as low or high risk. 3. Consider nasogastric tube placement. 4. PRBC transfusion for Hgb ≤7 g/dL 5. Correct coagulopathy, but do not delay endoscopy. 6. Promotility agents are not recommended as routine for pre-endoscopy. 7. Selected patients determined at low risk for rebleed after acute ulcer bleeding may be considered for discharge after endoscopy. 8. Pre-endoscopic PPI therapy may be useful to downstage the lesion and decrease need for endoscopic management, but its administration should not delay endoscopic evaluation.
Endoscopic management: 1. Institution-specific protocols for multidisciplinary UGIB are recommended; availability of an endoscopist trained in endoscopic hemostasis is recommended. 2. Urgent availability of endoscopic support staff 3. Endoscopy within 24 hours for most cases 4. Hemostatic treatment not indicated for low-risk lesions 5. Clots in the ulcer bed should be irrigated with the goal of revealing the underlying lesion for treatment. 6. Removing adherent clots is controversial; consider endoscopic therapy vs. intensive PPI therapy. 7. Hemostatic endoscopic treatment is indicated for high-risk lesions. 8. Epinephrine injection alone is suboptimal: best used as part of a combined therapy 9. Endoscopic coaptive therapy methods are equivalent. 10. Clips, thermocoagulation, or sclerosant injection should be used for high-risk lesions, alone or combined with epinephrine. 11. Second-look endoscopy as a routine is not recommended. 12. In case of rebleeding, a second endoscopic therapy attempt is generally recommended.	Pharmacologic management: 1. Histamine receptor antagonists are not recommended for treatment in acute UGIB. 2. Somatostatin and octreotide are not recommended for treatment in acute UGIB. 3. PPI infusion protocol (bolus plus continuous infusion) should be given post–successful endoscopic treatment to all high-risk lesion patients to reduce rebleeding and mortality. 4. Patients should be discharged on once-daily regimen of oral PPI for a duration appropriate to the treated etiology.
Nonendoscopic and nonpharmacologic in-hospital management: 1. Following endoscopy, patients at low risk may be fed within 24 hours. 2. Patients with high-risk lesions should remain hospitalized for 72 hours after endoscopy. 3. Surgical consultation is advised when endoscopic therapy has failed. 4. Percutaneous embolization may provide an alternative to surgery when endoscopic therapy has failed. 5. Patients with bleeding ulcers require testing for Helicobacter pylori and should receive eradication therapy if positive; eradication should be confirmed with a negative test result. 6. Negative acute testing for H. pylori should be repeated.	Postdischarge ASA and NSAIDs: 1. For patients with history of ulcer bleeding who require chronic treatment with NSAIDs, it must be recognized that whether treatment is with traditional NSAID plus PPI or with COX-2 inhibitor alone, the risk for recurrent ulcer bleeding remains clinically significant. 2. For patients with history of ulcer bleeding, the combination of a COX-2 antagonist plus PPI is recommended to reduce risk of rebleed compared with COX-2 agent alone. 3. For patients on low-dose ASA for cardiovascular risk who then develop UGIB, the ASA therapy should be resumed as soon as risk for cardiovascular complication is thought to outweigh risk for rebleed. 4. In patients who require cardiovascular prophylaxis after ulcer bleeding, it should be noted that clopidogrel alone has a higher risk for rebleeding than ASA plus PPI.

ASA, Acetylsalicylic acid; COX, cyclooxygenase; Hgb, hemoglobin; NSAIDs, nonsteroidal anti-inflammatory drugs; PRBC, packed red blood cell; PPI, proton pump inhibitor; UGIB, upper gastrointestinal bleeding.

Adapted from Barkun AN, Bardou M, Kuipers EJ, et al. International consensus recommendations on the management of patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med 2010;152:101–13; Barkun A, Bardou M, Marshall JK, et al. Consensus recommendations for managing patients with nonvariceal upper gastrointestinal bleeding. Ann Intern Med 2003;139:843–57.

* Nonvariceal UGIB.

Table 23.2

Summary of Recommendations for Acute Variceal Hemorrhage from the Practice Parameters Committee of the American College of Gastroenterology

HEMORRHAGE FROM ESOPHAGEAL VARICES

1.

Acute UGIB in patients with cirrhosis requires prompt attention with judicious intravascular volume support and blood transfusion aimed at maintenance of Hgb at ≈8 g/dL.
2.

Short, up to 1 week of antibiotic prophylaxis of SBP and other infections is recommended for all cirrhotic patients with gastrointestinal hemorrhage: oral norfloxacin 400 mg bid or intravenous (IV) ciprofloxacin or IV ceftriaxone 1 g/day in advanced cirrhotic patients or when quinolone resistance may be suspected.
3.

Pharmacologic therapy using somatostatin or its analogs octreotide, vapreotide and terlipressin (latter not available in United States) should be initiated immediately on suspicion of variceal bleeding and continued 3 to 5 days after diagnosis.
4.

Endoscopy should be performed within 12 hours to diagnose variceal hemorrhage and treat by endoscopic variceal ligation (EVL) or sclerotherapy.
5.

Transjugular intrahepatic portosystemic shunt (TIPS) is indicated for patients with uncontrolled bleeding, or who have failed combined pharmacologic and endoscopic treatment.
6.

Balloon tamponade should be used as a temporizing measure maximally for 24 hours in patients with uncontrollable bleeding while awaiting definitive endoscopic or TIPS treatment.

HEMORRHAGE FROM GASTRIC VARICES

1.

In patients bleeding from gastric fundal varices, endoscopic variceal obturation using tissue adhesives such as cyanoacrylate is preferred if available; EVL is an option.
2.

TIPS is indicated for patients with uncontrolled bleeding or who have failed combined pharmacologic and endoscopic treatment.

Hgb, Hemoglobin; SBP, spontaneous bacterial peritonitis; UGIB, upper gastrointestinal bleeding.

Adapted from Garcia-Tsao G, Sanyal AJ, Grace ND, et al. Prevention and management of gastroesophageal varices and variceal hemorrhage in cirrhosis. Am J Gastroenterol 2007;102:2086–102.

Hepatic Failure

Few patients with hepatic failure are cared for in the NCCU. However, patients with hepatic encephalopathy and cerebral edema frequently require the expertise of neurointensivists and neurosurgeons and management in the NCCU, for example, when therapeutic hypothermia is used for increased intracranial pressure (ICP) associated with hepatic encephalopathy. The pathology of hepatic encephalopathy is incompletely understood but appears to be associated with nitrogenous GI products and in particular ammonia that bypass liver metabolism in acute liver failure (ALF). These substances cross the blood-brain barrier, where they are thought to increase gama-aminobutyric acid (GABAergic) transmission and injure astrocytes that subsequently swell (i.e., cytotoxic edema). For this reason, corticosteroids are of no use. On the other hand, several lines of evidence suggest that proinflammatory mechanisms are involved in the pathogenesis of brain edema in ALF; this proposed mechanism provides a basis for using induced hypothermia to bridge patients with hepatic encephalopathy to transplant.

ALF, also know as fulminant liver failure, is the abrupt loss of liver function in a patient without previous liver disease. It typically is associated with coagulopathy (international normalized ratio [INR] >0.5) and encephalopathy. In the United States, acetaminophen accounts for approximately 50% of ALF cases: other causes include hepatitis, drug-induced liver injury, and viral or autoimmune hepatitis. Initial evaluation includes a liver function panel, prothrombin time [PT]/INR, complete blood count, fibrinogen, D-dimer, acetaminophen level, toxicology screen, electrolytes, and creatinine. In addition, a search for an etiology is necessary; this may include alpha-fetoprotein, ceruloplasmin, serum protein electrophoresis, virology (cytomegalovirus; Epstein-Barr virus; hepatitis A, B, or C virus antigens; or antibodies), or antinuclear antibody among others. An abdominal computed tomography (CT) scan to evaluate liver volume is useful, but if not feasible bedside ultrasound may be obtained. A head CT scan is necessary for patients with grade III or IV encephalopathy (see following text). During an ICU stay, patient follow-up evaluation may include arterial blood gas, arterial lactate, arterial blood oxygen (ABO) analysis (two separate tests), and every-6-hour PT/INR, transaminase level, total and direct bilirubin, and serum sodium. Frequent assessment of serum osmolarity or osmolality gap (when mannitol is used), and coagulation status also is required.

There are two important approaches to classification of ALF patients: (1) grading of hepatic encephalopathy and (2) determining whether a patient is eligible for transplant and whether transplant is needed to prevent death. There are several classification scales for hepatic encephalopathy, but the West Haven Simplified Criteria ( Table 23.3 ) is used most frequently. The severity of encephalopathy is the main barometer of disease severity and grades III and IV often mark the threshold for escalation of therapies. Determining who requires a liver transplant to prevent death is complex and requires consideration of ethical factors and an anticipation of future needs to ensure that definitive therapy is available in time to rescue the patient. The King’s College Criteria and the Model for End-Stage Liver Disease (MELD) are widely used scales to predict mortality in ALF and to help triage patients with severe liver disease for transplant ( Table 23.4 ). The Acute Physiology and Chronic Health Evaluation (APACHE) II score or Sequential Organ Failure Assessment (SOFA) score also can help predict outcome in acetaminophen-induced ALF. These scores (SOFA, MELD) and Simplified Acute Physiology Score (SAPS) II also can be used to predict outcome among patients with chronic liver insufficiency and cirrhosis who require ICU admission. Some studies suggest that ICU prognostic models such as SOFA may perform better than liver specific scores such as Child-Turcotte-Pugh (CTP) or MELD for outcome prediction among cirrhotic patients who are admitted to the ICU.

Table 23.3

West Haven Criteria for Semiquantitative Grading of Mental State in Acute Liver Failure

GRADE I