Critically ill patients are susceptible to a variety of causes and manifestations of abdominal dysfunction. However, the diagnosis and treatment of these conditions can be challenging secondary to nonspecific clinical findings, concurrent complex disease processes, and altered mental status. The purpose of this chapter is to discuss select causes of abdominal dysfunction in the critically ill patient, including evaluation for acute abdominal pathology in the critically ill patient, AAC, severe acute pancreatitis, feeding intolerance, paralytic ileus and diarrhea, ACS, and care of the long-term open abdomen.
Evaluation for Acute Abdominal Pathology in the Intensive Care Unit
Critically ill patients are susceptible to acute abdominal pathology, including bowel perforation, biliary tract disease, pancreatitis, ischemia, and hemorrhage. Acute abdominal pathology may be the patient's initial insult, or the patient may develop abdominal dysfunction as a complication of critical illness. Patients with recent surgery may manifest intra-abdominal complications such as anastomotic leak or abscess, or may develop iatrogenic abdominal complications such as bowel perforation from paracentesis, or pancreatitis after endoscopic retrograde cholangiopancreatography (ERCP). Critically ill patients may pose a diagnostic dilemma, as some patients may be challenging to evaluate due to concomitant critical illness; even patients with evaluable mental status may have unreliable clinical examinations. Steroid use and immunosuppression may blunt a patient's clinical examination even in the presence of an intra-abdominal catastrophe. Evaluation of the patient with suspicion for acute abdominal pathology should occur expeditiously; failure to consider the abdomen as a potential source of sepsis or hemorrhage can lead to missed diagnoses and poor outcomes.
Although physical examination findings in this population can be nonspecific, patients with unexplained sepsis or abdominal pain (if they can communicate same) should undergo a thorough physical examination to evaluate for abdominal distension, tenderness, and inspection of all wounds and incisions. Laboratory findings will be nonspecific as well, but may provide important clues to the diagnosis and should be obtained and trended; white blood cell count, liver enzymes, amylase, lipase, lactate, and arterial blood gases may be of value.
Evaluation of the abdomen in the critically ill patient should include diagnostic radiologic imaging if clinically tolerable (in terms of positioning, or the need to transport the patient to the radiology suite) and appropriate for the suspected underlying process.1 In general, computed tomography (CT) is the test of choice for patients with a suspected intra-abdominal source of sepsis. Pancreatitis, diverticulitis, and other inflammatory bowel processes; intra-abdominal abscesses; and bowel obstructions are visualized easily with contrast-enhanced CT. Patients with recent abdominal surgery and postoperative intra-abdominal infection typically manifest signs and symptoms of dehiscence, leak, or abscess after 5 to 10 days and CT scan can be instrumental for the diagnosis; CT scans performed earlier than 5 to 7 days in postoperative patients are of limited value due to identification of nonspecific abdominal fluid, inflammation, and even residual pneumoperitoneum introduced during celiotomy. CT scans are sensitive for detecting the presence of air, which can be pathologically located outside the bowel in the peritoneal cavity, within the bowel wall (pneumatosis intestinalis), or in the portal venous system. In order to obtain the most useful images, both oral and intravenous contrast should be utilized. Oral contrast allows differentiation of the intraluminal bowel fluid from extraluminal fluid collections and will help to identify the transition point in cases of bowel obstruction. Intravenous contrast optimizes visualization of infectious processes by highlighting areas of inflammation; additionally, intravenous contrast aids in the assessment of solid organs and areas of potential ischemia or hemorrhage. Intravenous contrast is especially useful in cases of acute pancreatitis to delineate areas of devitalized or necrotic pancreatic tissue or retroperitoneal fat. Allergy or the patient's renal function may limit the ability to utilize intravenous contrast.
Specific conditions may require targeted diagnostic evaluation. Patients with a suspected biliary infection may benefit more from ultrasound (US) evaluation of the abdomen, rather than CT scan. Calculous or acalculous cholecystitis or dilation of bile ducts can be evaluated easily with US. Patients with gastrointestinal bleeding or suspicion of ischemic colitis or Clostridium difficile infection would benefit from endoscopy. Endoscopy for bleeding can be both diagnostic and therapeutic and can be performed at the bedside if necessary. Angiography with interventional radiology techniques can also be diagnostic and therapeutic for arterial bleeding.
Patients who are too unstable hemodynamically or clinically for imaging pose a greater diagnostic predicament. Patients requiring multiple or increasing doses of vasopressors, patients with ongoing hemorrhage, or patients with severely compromised respiratory status may have excess or unacceptable risk of morbidity from transport to radiology for testing.2 These patients should be given goal-directed resuscitation and stabilized with fluids and antibiotics if possible and imaged when feasible. For patients with a high suspicion for intra-abdominal pathology, an emergent bedside laparotomy may be indicated; however, this is a high-risk procedure, requires substantial resources to be mobilized, and has a risk of mortality. Other described diagnostic options for these patients include diagnostic peritoneal lavage (DPL), which places a catheter through the abdominal wall into the peritoneal cavity to allow evaluation of peritoneal fluid; some centers also advocate bedside diagnostic laparoscopy. These techniques have been used in the ICU to evaluate the abdomen for otherwise unevaluable patients with some success, but are not yet standard of care.1,3,4
Acute Acalculous Cholecystitis
AAC is the development of acute inflammation of the gallbladder in the absence of gallstones. AAC is generally considered to be a complication of serious medical and surgical illnesses, especially in the setting of trauma, burns, sepsis, prolonged fasting, or total parenteral nutrition. Among medical patients, a variety of systemic diseases have been associated with the development of AAC such as diabetes mellitus, abdominal vasculitis, congestive heart failure, and cholesterol embolization of the cystic artery. Resuscitation from hemorrhagic shock or cardiac arrest has been associated with AAC.5,6 Acalculous cholecystitis may also develop as a secondary infection of the gallbladder during systemic sepsis for a wide range of infections.
AAC poses major diagnostic challenges. Most afflicted patients are critically ill and unable to communicate their symptoms. Cholecystitis is but one of many potential causes in the differential diagnosis of systemic inflammatory response syndrome (SIRS) or sepsis in such patients. Rapid and accurate diagnosis is essential, as gallbladder ischemia can progress rapidly to gangrene and perforation. Acalculous cholecystitis is sufficiently common that the diagnosis should be considered in every critically ill or injured patient with a clinical picture of sepsis or jaundice and no other obvious source.
US and CT are generally useful for the diagnosis of AAC. Ultrasound of the gallbladder is generally the first-line modality for the diagnosis of AAC in the critically ill patient as it is rapid, low risk, and portable. In calculous cholecystitis, US is useful for detecting gallstones and measuring bile duct diameter; in AAC, these measurements are not valuable. Thickening of the gallbladder wall is the single most reliable criterion, with reported specificity of 90% at 3 mm and 98.5% at 3.5-mm wall thickness, and sensitivity of 100% at 3 mm and 80% at 3.5 mm.7 Accordingly, gallbladder wall thickness greater than or equal to 3.5 mm is generally accepted to be diagnostic of AAC. Other helpful US findings for AAC include pericholecystic fluid or the presence of intramural gas or a sonolucent intramural layer, or “halo,” that represents intramural edema. Distension of the gallbladder of more than 5 cm in transverse diameter has also been described.6,7 False-positive US examinations may occur in particular when conditions including sludge, nonshadowing stones, cholesterolosis, hypoalbuminemia, or ascites mimic a thickened gallbladder wall. CT appears to be as accurate as US in the diagnosis of AAC. Diagnostic criteria for AAC by CT are similar to those described for US. Low cost and the ability to perform US rapidly at the bedside make it the preferred diagnostic modality in possible AAC in the ICU setting. Preference may be given to CT if other thoracic or abdominal diagnoses are under consideration.
Historically, the treatment for AAC was cholecystectomy, owing to the ostensible need to inspect the gallbladder and perform a resection if gangrene or perforation was discovered. In the modern era, percutaneous cholecystostomy can be a lifesaving, minimally invasive approach, as it controls the AAC in 70% to 90% of patients.8,9 For this procedure, the gallbladder is intubated under US (occasionally laparoscopic) guidance via an anterior or transhepatic approach (through the right hepatic lobe) in order to minimize leakage of bile. Rapid improvement should be expected when percutaneous cholecystostomy is successful. Percutaneous treatment is an especially useful modality in the elderly patient with sepsis or patients who are unstable for a surgical procedure. Cholecystostomy will not decompress the common bile duct if cystic duct obstruction is present, therefore the common duct must be decompressed in addition by some manner (eg, ERCP) with sphincterotomy, or laparoscopic or open common bile duct exploration if cholangitis is suspected.
If percutaneous cholecystostomy does not lead to rapid improvement, the tube may be malpositioned, not draining properly, or the patient may have gangrenous cholecystitis. Other reported causes of failure include catheter dislodgement, bile leakage with peritonitis, or an erroneous diagnosis. Perforated ulcer, pancreatic abscess, pneumonia, and pericarditis have been discovered in the aftermath of percutaneous cholecystostomy when patients failed to improve. Reported major complications occur after 8% to 10% of procedures, including dislodgment of the catheter, acute respiratory distress syndrome (ARDS), bile peritonitis, hemorrhage, cardiac arrhythmia, and hypotension due to procedure-related bacteremia.8,9 A cholecystostomy or cholecystectomy may be required if other sources of sepsis are ruled out and the patient continues to deteriorate. If an operation is warranted, open cholecystostomy may be accomplished under local anesthesia through a short right subcostal incision, but the ability to visualize elsewhere in the abdomen is limited. A laparotomy or laparoscopy would be required to drain distant fluid collections or identify other pathology that may mimic acute cholecystitis in the case of a misdiagnosis (eg, perforated ulcer, cholangitis, pancreatitis). In stable patients with AAC who require surgery, laparoscopic cholecystectomy has been described.
Antibiotic therapy does not substitute for drainage of AAC, but is an important adjunct. The most common bacteria isolated from bile in acute cholecystitis are Escherichia coli, Klebsiella spp., and Enterococcus faecalis, although prior antibiotic administration may allow for other opportunistic pathogens to be present.5 However, critical illness and prior antibiotic therapy alter host flora, and resistant or opportunistic pathogens may be encountered. Anaerobes are particularly likely to be isolated from bile of patients with diabetes mellitus, in those older than 70 years, and from patients whose biliary tracts have been instrumented previously.
Patency of the cystic duct can be determined immediately after the cholecystostomy is performed by tube cholangiography. This should be performed again after the patient has recovered to determine the presence of gallstones that may not have been detected initially. If gallstones are present, an elective cholecystectomy is usually recommended, with the drainage tube remaining in place during the inter-procedure interval. For patients without gallstones, interval cholecystectomy is usually not indicated, and the cholecystostomy tube can be removed 4 to 6 weeks postcholecystostomy after tube cholangiography confirms that gallstones are absent. Recurrent episodes warrant cholecystectomy.
Acute pancreatitis can vary in presentation from mild to severe. Patients with severe acute pancreatitis often require a prolonged ICU stay to provide supportive therapy as patients may develop sepsis and progressive organ failure. The pancreas can develop focal or diffuse areas of nonviability, known as necrotizing pancreatitis; when bacteria from the gut infiltrate the nonviable tissue, it becomes infected necrosis. Whereas mild acute pancreatitis has a low mortality rate and usually resolves after a short period of bowel rest, mortality from severe acute pancreatitis with sterile necrosis is estimated to be 10% and can be as high as 70% in the presence of infected pancreatic necrosis.10,11
Acute pancreatitis typically presents with epigastric pain that radiates to the back or shoulder, concomitant with nausea, vomiting, fever, and leukocytosis. Laboratory evaluation should include a complete blood count, a complete metabolic panel with liver enzymes, amylase, lipase, and lactate dehydrogenase. For diagnosis of pancreatitis, amylase has a higher sensitivity and lipase has higher specificity; levels greater than 3 times the upper limit of normal support the diagnosis. However, the magnitude of elevation of these laboratory values does not correlate with severity of illness; similarly, normalization of values does not signify resolution of disease. Concurrently elevated liver enzymes, especially alanine aminotransferase, can suggest a biliary etiology.
The most important first step in evaluation of patients with pancreatitis is to identify risk of progression to severe pancreatitis so that aggressive treatment can be instituted expeditiously. Multiple scoring systems exit that are designed to assess the disease severity, although no one system is universally accepted as superior. Common scoring systems include the Ranson criteria, Acute Physiology and Chronic Health Evaluation (APACHE)-II, and the CT severity index (also known as the Balthazar score) (Table 35–1). Although the Ranson criteria were first described in 1974, the score remains clinically useful. Early CT scans for prognostic reasons are not indicated in patients with pancreatitis, as early imaging lacks sensitivity to detect necrosis and infected necrosis is not typically present early in the course of the disease. Patients who develop persistent SIRS or clinical deterioration after 72 hours may benefit from further characterization of disease by CT.
Table 35–1Severe acute pancreatitis scoring systems. |Favorite Table|Download (.pdf) Table 35–1 Severe acute pancreatitis scoring systems.
|System ||Criteria ||Scoring and Interpretation |
|Ranson criteria || |
Age > 55
WBC > 16 × 109/L
LDH > 350/L
AST > 250/L
Glucose > 200 mg/dL
During initial 48 h
Hgb falls below 10 mg/dL
BUN rises by > 5 mg/dL
Calcium < 8 mg/dL
PaO2< 60 mm Hg
Base deficit > 4
Fluid sequestration > 6 L
1 point for each factor listed
Score > 3 indicates severe acute pancreatitis
Requires 48 h to calculate entire score
|Acute physiology and chronic health evaluation II score (APACHE II) || |
Calculated from 12 measurements
Mean arterial pressure
White blood cell count
Glasgow Coma Scale
Score > 8 predicts increased risk for complications and mortality
Score calculated at 24 h
|Computer tomography scoring index (CTSI, or Balthazar index) || |
Based on CT scan
a. Normal pancreas (0 point)
b. Enlarged pancreas (1 point)
c. Pancreatic/peripancreatic inflammation (2 points)
d. One peripancreatic collection (3 points)
e. 2 or more peripancreatic collections or retroperitoneal air (4 points)
Percentage of necrosis
a. None (0 point)
b. < 30% necrosis (2 points)
c. 30%-50% necrosis (4 points)
d. > 50% necrosis (6 points)
Grade points are added to necrosis points to determine the total score
Cannot be performed prior to 72-96 h after the onset of symptoms because initially pancreatic necrosis is indistinguishable from edema
The initial treatment of severe acute pancreatitis is supportive and includes aggressive fluid resuscitation, pain control, and supplemental oxygen administration. A substantial inflammatory response leads to increased vascular permeability and fluid sequestration within the intra-abdominal space. Resuscitation should target adequate end-organ perfusion to maintain acceptable physiologic targets such as blood pressure, heart rate, and urine output. Patients with severe acute pancreatitis are at risk for development of multiple organ dysfunction syndrome (MODS); acute respiratory failure, circulatory shock, and acute kidney injury are observed in severe cases. In extremely ill patients, central venous pressure and mixed venous oxygen saturation may be of additional value. Sequential bladder pressure measurements are advisable to monitor for ACS.
Historically, a mainstay of treatment for severe acute pancreatitis was bowel and pancreatic rest, as pancreatic stimulation from enteral nutrition was believed to exacerbate pancreatic inflammation. Accordingly, patients with severe acute pancreatitis were given parenteral nutrition until the pancreas recovered. However, over the last decade, several randomized controlled trials comparing enteral to parenteral nutrition have demonstrated that enteral nutrition is associated with decreased infectious morbidity, shorter length of stay, fewer overall complications, and faster resolution of the disease process.12 Accordingly, enteral nutrition is preferred over parenteral nutrition. Enteral nutrition should be administered early; while the exact timing for initiating enteral nutrition is unknown, it is currently believed that feeding should be commenced soon after admission. Delays in initiation of enteral nutrition can lead to prolonged ileus and decrease tolerance to feeding; additionally, small studies suggest that the benefits of enteral nutrition may be reduced if the enteral nutrition is delayed. The preferred route of enteral administration is also unknown; several trials have failed to detect a difference between nasogastric and nasojejunal feeding.11 It has been suggested that probiotics could protect patients from overwhelming sepsis from opportunistic pathogens. Ultimately, probiotics have not proved to be effective and may actually increase mortality; routine administration of probiotics is not recommended.13
The most common cause of death from severe acute pancreatitis is MODS that develops as a result of infected pancreatic necrosis. This led to the hypothesis that antibiotic prophylaxis for patients with pancreatic necrosis might prevent the later sequelae and morbidity of infection. Early studies in the 1990s showed lower rates of infected pancreatic necrosis with lower mortality; subsequently, administration of antibiotics for prophylaxis became common.14,15 More recently, 3 double-blind randomized trials were published that could not confirm any beneficial effects of antibiotic prophylaxis; these well-done studies showed that prophylaxis is not associated with a statistically significant reduction in mortality, in the incidence of infected pancreatic necrosis, in the incidence of nonpancreatic infections, or in need for surgical interventions.16,17,18 Therefore, antibiotic prophylaxis for severe acute pancreatitis is not recommended.
It can be challenging clinically to differentiate if a severe SIRS response from active infection of pancreatic necrosis as patients will appear ill in both circumstances. Patients suspected of having infected necrosis should have the tissue sampled for culture via percutaneous fine-needle aspiration. Patients with infected necrosis should have antibiotic treatment and management of the infected tissue.
Traditionally, open debridement with pancreatic necrosectomy was performed to remove infected tissue but this procedure was associated with high morbidity and mortality, as well as high rates of long-term pancreatic insufficiency for survivors.10,11 Overall, there has been a trend toward less invasive approaches to management of necrosis and infected necrosis with or without delayed surgical intervention, known as the “step-up” approach. Patients receive initial medical management with fluid resuscitation and antibiotic administration, followed by percutaneous catheters for drainage of infected fluid. The percutaneous catheters can be irrigated using large volumes of sterile saline and patients are monitored for improvement. Patients who fail to improve may be candidates for operative debridement. In general, postponement of operative debridement for as long as possible is encouraged, as it is associated with decreased morbidity and mortality; however, prolonged antibiotic treatment leads to increased infections with resistant organisms and fungi.19 Skilled endoscopists have also started to explore transgastric and transduodenal necrosectomy, although these methods are not yet standard of care.
It is well established that enteral nutrition, provided as early as possible, is beneficial to patients in the ICU. Malnutrition is associated with increased infections, reduced wound healing, increased mortality, prolonged hospital stay, and increased cost. Many critically ill patients present to the hospital with preexisting protein-calorie malnutrition, which is exacerbated by their acute illness. ICU admission leads to periods of starvation (nil per os for a variety of reasons), with simultaneous increased metabolic demand from the acute illness. Early enteral nutrition is encouraged and most studies support the dictum “if the gut works, use it.” The indications and methods for enteral feeding are discussed elsewhere in this text.
Studies have demonstrated that a substantial percentage of patients in the ICU develop intolerance to enteral feeding, which can be manifested by high gastric residuals, emesis, abdominal distension, or abdominal pain.20 Most studies of critically ill patients report the ability to deliver only 40% to 60% of goal nutrients secondary to gastrointestinal dysmotility or other barriers to early nutrition.21 Once mechanical reasons for obstruction are ruled out, feeding intolerance in the critically ill patient stems from the presence of gastric dysmotility or ileus. Rapid resolution of dysmotility is crucial to allow progression of enteral feeding.
Gastric residual volumes are used as a surrogate marker for feeding intolerance, and thus “high gastric residuals” are frequently the rationale for interruption of feeding. Unfortunately, the relationship between gastric residual volumes and delayed gastric emptying is unclear, and the residual volumes that should trigger the cessation of enteral feeding are controversial.22,23 Some consensus groups agree that a gastric residual volume greater than 200 mL is abnormally high, but still recommend that this threshold should not trigger automatic cessation of enteral nutrition. Because there is no accepted definition for high gastric residuals, thresholds for holding enteral nutrition can vary among institutions, and caregivers at the same institution. Creation of an enteral tube feeding protocol that is not physician-dependent may also improve delivery of nutrients by minimizing feeding interruptions and standardizing the thresholds of residual volumes within an ICU.24
Feeding intolerance in the critically ill patient can be attributable to the patient's critical illness, medications, intra-abdominal pathology, or underlying disease. Feeding intolerance may also be due to ileus or bowel obstruction; mechanical obstruction must be ruled out prior to administering enteral nutrition aggressively. Gastric emptying is believed to be delayed in patients with traumatic brain injuries (TBIs) and elevated intracranial pressure (ICP); hyperglycemia may also contribute to delayed gastric empyting.25 Endogenous or exogenous catecholamines are likely to decrease gastric emptying, whereas dopamine reduces antral contractions. Anticholinergics, calcium channel blockers, sedatives, and opiates, all of which are used commonly in the ICU, also slow gastric motility.
Therapeutic options for the treatment of delayed gastric emptying are imperfect. Prokinetic options are most widely used, but no regimen is perfect. Common prokinetic agents include metoclopramide or erythromycin as monotherapy, or in combination. Metoclopramide antagonizes the effect of dopamine on the gastric antrum and improves gastric motility when given as a 10 mg q6h, although the effect is blunted in TBI patients. With repeated administration, tachyphylaxis develops quickly and reliably such that after 7 days of routine metoclopramide administration, only 20% to 25% of patients continue to have successful nasogastric feeding.25,26 Another commonly utilized promotility agent, erythromycin, acts as a motilin agonist and can be given intravenously in doses of 1 to 3 mg/kg/d and likely has better prokinetic activity than metoclopramide. Unfortunately, patients also become resistant to the effects of erythromycin over time, such that after 7 days, only 30% to 45% of patients can be successfully fed via a nasogastric tube. Additionally, there are concerns that erythromycin can exacerbate antibiotic resistance or lead to cardiotoxicity. Cardiac toxicity can be minimized by using smaller doses; reportedly, an intravenous dose of 70 mg of erythromycin is as effective as a 200-mg dose in accelerating gastric emptying. Once feedings are tolerated, the prokinetic agent can be discontinued.
Combination therapy of metoclopramide plus erythromycin may be superior to the use of either drug alone and decreases the incidence of tachyphylaxis as first-line treatment of feeding intolerance or after the failure of monotherapy.25,26,27 Newer therapies with potential to improve gastric emptying include opioid antagonists such as methylnaltrexone or alvimopan. Methylnaltrexone and alvimopan theoretically do not cross the blood-brain barrier so that analgesia is maintained while the effects of opioids on the gastrointestinal tract are blunted. Both drugs have been tested in in postoperative elective bowel resection patients and help accelerate recovery of bowel function or decrease postoperative ileus without increased pain. Unfortunately, these opioid antagonist agents have not been proved in the critically ill patient.28 When drug treatment fails, alternatives can include parenteral nutrition or postpyloric feeding access. In fact, some clinicians advocate enteral feeding into the small bowel as first-line enteral access. Controversy exists on the optimal anatomic location for feeding access, and this is discussed elsewhere in this text.
Other Enteral Feeding Issues: Hemodynamic Instability, Ileus, and Diarrhea
Patients in the ICU may undergo prolonged periods of starvation for reasons other than feeding intolerance, including imminent or recent surgical procedures, hemodynamic instability, diarrhea, and lack of functional enteral access. Some clinicians are reluctant to initiate early enteral feeding until the “acute” phase of the injury response has subsided, owing to the fear that immediate feeding will result in higher complications from delayed gastric emptying, such as aspiration of gastric contents. Despite this, large studies suggest that the benefits of early feeding outweigh the risks and the practice of early nutrition is becoming more common.
Giving enteral feeds on vasopressors continues to be contentious. Hemodynamic instability requiring vasopressors is often considered to be a relative contraindication to enteral feeding; however, the data supporting this are inconclusive. In healthy adults, enteral nutrition is associated with an increase in blood flow to the intestines. The concerns for feeding patients who are on vasopressors are 2-fold: Patients can develop intestinal ischemia from the increased demand when the circulatory system is not able to increase blood flow, or patients may experience a “steal” phenomenon, where blood is diverted to the splanchnic circulation leading to ischemia of other end organs. There have been a few case reports of very early enteral feeding being associated with intestinal ischemia in underresuscitated trauma patients but it is unclear whether the feeding or the inadequate resuscitation was the culprit leading to intestinal ischemia.29 Whereas it is probably prudent to not feed patients in hemorrhagic shock or with rapidly escalating vasopressor requirements, some data suggest that enteral feeding is safe in patients receiving vasopressors.30
Another common reason for delay of enteral nutrition is adynamic or paralytic ileus of the small bowel. Ileus can be caused by a number of intra-abdominal and retroperitoneal processes, including intestinal ischemia, ruptured viscus, hemorrhage, pancreatitis, peritonitis, medications (opioids, anticholinergics), and electrolyte abnormalities. It may also occur after abdominal surgery. Ileus is heralded by abdominal pain, vomiting, and abdominal distension with obstipation and must be differentiated from a mechanical small bowel obstruction. Radiographic studies may help differentiate obstruction from ileus. Laboratory studies should be obtained, including a complete blood count, chemistries, and perhaps an arterial blood gas and lactate if there is concern for intestinal ischemia, although no laboratory test is definitive. Patients with worsening signs on serial abdominal examinations, or who develop a fluid requirement and acidosis, should have a surgical evaluation for obstruction or ischemia. Plain abdominal films performed at the bedside have limited value. On plain radiographs, ileus will show diffuse small bowel dilation (> 3 cm) with air-fluid levels and no distinct cutoff point. Air and stool present in the colon on a plain radiograph can be a reassuring sign that there is no mechanical obstruction present but is not particularly specific, as a patient with acute obstruction or a closed loop obstruction may not have had time to decompress the distal bowel. Patients who are suspected of having a mechanical obstruction should undergo a CT scan with oral and intravenous contrast. Nasogastric tube decompression can be used to ease symptoms of nausea and may help resolve the ileus.
Colonic pseudo-obstruction (Ogilvie syndrome) occurs typically in bedridden or elderly patients, but may also happen as a sequela of spinal cord injury, prolonged opioid use, or postoperatively from abdominal or nonabdominal surgery. This manifests characteristically with abdominal distension, abdominal pain, obstipation, and possibly tenderness; vomiting may or not be present. Plain radiographs demonstrate a dilated colon, usually most pronounced at the cecum, and should be performed as an initial diagnostic test to rule out a mechanical cause of obstruction such as volvulus. Perforation risk is approximately 3% but risk increases if the diameter of the cecum exceeds 12 cm or if the distension has been present for more than 6 days. Therapy should include correction of electrolyte abnormalities, cessation of narcotics, and nasogastric decompression. Neostigmine, 2 mg IV over a period of 5 minutes, can be utilized to resolve the ileus. Administration of neostigmine should take place in the ICU, as some patients develop symptomatic bradycardia and require administration of atropine. A repeat dose of neostigmine may be administered after 3 hours.31 If neostigmine does not resolve the pseudo-obstruction (~ 15% of cases), colonoscopic decompression may be considered as success rates are high (~ 70%); complications include perforation, inability to decompress the colon, and recurrence. Surgical intervention may be necessary for patients who have perforated, patients who have had unsuccessful attempts at decompression, or patients with recurrent pseudo-obstruction after decompression.
Critically ill patients often develop diarrhea, defined as an increase in the fluidity, frequency, or quantity of bowel movements. The actual volume of stool that constitutes diarrhea is debated but typically is described as greater than 300 mL/d or greater than 4 loose stools per day. Diarrhea can lead to electrolyte imbalance, wound contamination, and dehydration, and may prompt providers to hold enteral nutrition, which can exacerbate malnutrition. Diarrhea may be attributable to one or multiple simultaneous causes, including enteral feeding formulas or regimens, malabsorption, underlying disease, medications, or infections.32 Medications that may cause diarrhea should be discontinued, such as metoclopramide, erythromycin, oral magnesium or phosphorus, or antibiotics. Diarrhea is the most commonly reported complication of enteral tube feeding; research on various feeding formula issues have been studied, including temperature, osmolality, fiber content and type, density, delivery rate, and formula content, but no factor in the feeding formula is consistently associated with development of diarrhea.32 Infectious diarrhea in the ICU can be secondary to C difficile, Klebsiella oxytoca, Clostridium perfringens, Salmonella spp., Staphylococcus aureus, or gastrointestinal viruses.
For patients with profuse diarrhea, all possible offending medications should be discontinued and infectious sources should be treated. Several nutritional interventions have been studied, with limited success. Fiber-enriched formulas are supported by the Society of Critical Care Medicine and the American Society for Parenteral and Enteral Nutrition in patients who have developed diarrhea,33 but not all clinical trials demonstrate an effect of fiber on diarrhea. Probiotics have been suggested as a way to recolonize the intestine and normalize the intestinal flora. However, the administration of probiotics is controversial in critically ill patients with abnormal immune responses; one study of patients with severe acute pancreatitis even showed increased mortality with probiotic prophylaxis.13 For patients with diarrhea that is deemed noninfectious, antidiarrheal agents may be considered but should be used with great caution. Opioid analogs may be considered; the most common form is dephenoxylate with atropine (1 tablet 3-4 times/d) or loperamide (up to 4 mg, 4 times/d).
Abdominal Compartment Syndrome
In the critically ill patient, pressure within the abdominal cavity can be increased pathologically above the patient's baseline. Prolonged elevations of intra-abdominal pressure can result in organ dysfunction and failure, known as ACS. In the healthy patient, physiologic intra-abdominal pressure is close to 0 mm Hg, although this can be increased chronically to 10 to 15 mm Hg in obesity or pregnancy. In the critically ill patient, pressures are generally increased to levels of 5 to 7 mm Hg.34 Acidosis, multiple blood transfusions, sepsis, major trauma or burns, pancreatitis, ileus, liver dysfunction, and aggressive ventilator settings with high positive end expiratory pressures are factors that can further increase intra-abdominal pressure. IAH is defined as intra-abdominal pressure greater than 12 mm Hg. ACS is defined as intra-abdominal pressure greater than 20 mm Hg with evidence of end-organ dysfunction or failure.34 Clinically, this may manifest with a tense, distended abdomen, progressive hypotension, oliguria, and increased airway pressures. Primary abdominal compartment syndrome is characterized by acute or subacute development of intra-abdominal hypertension in a relatively brief period of time secondary to pathology within the abdominopelvic cavity, usually due to abdominal trauma, ruptured abdominal aneurysm, hemoperitoneum, acute pancreatitis, acute peritonitis, or retroperitoneal hemorrhage. Secondary ACS is characterized by IAH that develops over a subacute or chronic time period as a result of an extra-abdominal cause such as sepsis, capillary leak syndrome, or aggressive fluid resuscitation following a major burn. Patients may develop recurrent ACS after abdominal decompression, even with an open abdomen, if resuscitation is ongoing or dressings are too tight, or if abdominal closure was performed and intra-abdominal pressure becomes elevated again.
Intra-abdominal pressure monitoring is essential to the diagnosis of ACS. A variety of methods exist to measure pressure within the abdominal cavity, which include direct (via needle puncture into the abdominal cavity) or indirect (transduction of intra-abdominal pressure through a surrogate, such as the bladder or colon). Bladder pressure measurement has the most widespread adoption owing to the simplicity and minimal cost; additionally most critically ill patients already have a bladder catheter in place, so the measurement of pressure is relatively noninvasive. However, there is increased risk of urinary tract infection. Regardless of the technique utilized, several key principles need to be followed as outlined by the International Conference of Experts on Intra-abdominal Hypertension and Abdominal Compartment Syndrome in 2004.34 Intra-abdominal pressure should be expressed in mm Hg and measured at the end of end expiration, with the transducer zeroed at the midaxillary line and ensuring that there are no abdominal muscle contractions. There reference standard is instillation of 25 mL sterile saline into the bladder; higher volumes of instilled fluid can overfill the bladder such that the transduced pressure becomes the pressure of the bladder wall instead of the intra-abdominal pressure. Changes in body position, and the presence of abdominal or bladder wall contractions have been demonstrated to impact the accuracy of intra-abdominal pressure measurements.35
Treatment and management of ACS consists of 4 general principles: (1) serial monitoring of intra-abdominal pressures, (2) optimization of systemic perfusion and organ function in the presence of intra-abdominal hypertension, (3) institution of medical procedures to decrease intra-abdominal pressure and reduce end-organ dysfunction, and (4) prompt surgical decompression for refractory intra-abdominal hypertension or ACS. Generally, the “critical” intra-abdominal pressure at which end-organ dysfunction occurs differs among patients, but some trials have evaluated abdominal perfusion pressure (similar in concept to ICP monitoring and cerebral perfusion pressure), where the abdominal perfusion pressure is equal to the mean arterial pressure minus the intra-abdominal pressure. Several retrospective trials have suggested that abdominal perfusion pressure above 50 to 60 mm Hg is associated with improved survival, but this has yet to be validated prospectively.35
Several medical therapies have been hypothesized to temporize and treat mild IAH; however, if a patient is showing signs of end-organ damage, medical therapies are unlikely to have beneficial or expeditious enough effects to prevent the need for decompressive laparotomy. Pain, agitation, ventilator dyssynchrony, and use of accessory muscles during work of breathing all may contribute to increased intra-abdominal pressure. Patients at risk for ACS must be afforded adequate sedation and pain control. It has been suggested that neuromuscular blockade may allow for muscle relaxation to decrease the compliance of the abdominal wall, thus decreasing intra-abdominal pressure. However, the risks of pharmacologic paralysis must be weighed carefully against the uncertain benefits. Body positioning may also contribute to intra-abdominal pressure; elevation of the head of the bed and supine positioning may increase intra-abdominal pressure, but again the benefits of these positions in specific patient populations may outweigh the risk of supine and flat positioning. Other interventions that may theoretically decrease intra-abdominal pressure such as nasogastric and colonic decompression, diuretics, and renal replacement therapy have not been adequately studied in this patient population.
Surgical decompressive laparotomy is the standard treatment for patients who develop ACS. Once a patient's IAH has become refractory to medical therapies, laparotomy is a lifesaving intervention. Surgical decompression results in an “open abdomen,” which is discussed further in the next section. Decompressive laparotomy is also a reasonable therapeutic option in a patient with intra-abdominal hypertension for whom the risk of ACS is high but the patient has not yet manifested end-organ damage.35
The Long-Term Open Abdomen
Use of damage control surgery with temporary abdominal closure has gained popularity since the late 1980s as a way to salvage critically ill trauma patients with physiologic compromise due to massive hemorrhage in the abdomen. This approach has been adapted for use in other nontrauma surgical patients with abdominal catastrophes, such as ACS or pancreatic necrosectomy with expected serial debridement procedures. There are generally 3 stages to damage control surgery: abbreviated surgery, resuscitation, and delayed definitive closure. The initial surgery is abbreviated to allow for rapid control of hemorrhage or abdominal contamination, and may require packing for hemostasis. At this point, the abdominal contents are covered with a temporary dressing. This is followed by resuscitation, warming, and correction of any existing coagulopathy in order to allow the patient to have normalization of physiology. Implicit in this approach is the planned surgical reexploration, which typically occurs 12 to 72 hours after the index operation. Aggressive resuscitation between the first and second operation may render the abdomen unable to be closed and serial dressing changes and staged operations may be necessary to obtain closure or coverage of the abdominal contents. Ideally, patients should undergo definitive fascial closure whenever possible, even if closure must be performed in stages; in some cases, closure is impossible. Surgical options for these patients are limited and unappealing but include temporary mesh closure, skin-only closure, or split-thickness skin grafts with planned ventral hernia. Many of these patients develop severe loss of domain and loss of function of the abdominal wall, eventually necessitating extensive reconstructive surgery to regain some abdominal wall functionality.
All patients who undergo damage control surgery with an open abdomen are at high risk for infectious complications. Surgical site infections have been reported to occur in as many as 83% of cases, and postoperative fascial dehiscence is reported in up to 25% of patients who have had an open abdomen.36 However, no data support antibiotic prophylaxis of the open abdomen. Particularly concerning in patients who are unable to undergo fascial closure is the development of bowel fistulae. The incidence of fistulae in patients with an open abdomen is approximately 5% to 19%, and varies according to the initial indication for damage control surgery.37 Of unique concern in this patient population is the “enteroatmospheric” fistula, which occurs because there is no tissue that overlies the exposed bowel and spontaneous healing becomes impossible. Patients with enteroatmospheric fistulae should have radiologic evaluation with enteral contrast to clarify the location of the fistula (proximal or distal bowel) to create strategies for enteral nutrition, if possible, and minimize fluid and electrolyte losses. Continuous leakage of enteric contents into the wound contributes to elevated catabolism, protein loss, infection/sepsis, and increased mortality. These patients likely require serial abdominal operations to control enteric leakage and reestablish bowel continuity.