Most patients with an acute abdomen are diagnosed outside the ICU. These patients may require treatment in an ICU for one of four reasons:
In this chapter we will consider only patients in the fourth group, since the other groups are discussed in other areas of the text. We will primarily consider the ICU management of patients who have undergone treatment of a septic abdominal disease.
A classification of the sources of IAS is given in Table 89-2. Sepsis resulting from disease originating in the pancreas and infections arising in the urinary tract are discussed in other chapters.
The principles of treatment of IAS are well established and are as follows:
The importance of prompt diagnosis and treatment cannot be overemphasized. This is one of the few prognostic variables that physicians can control, and prompt treatment has been shown repeatedly to result in decreased mortality.5–7
Primary peritonitis is a group of diseases characterized by infection in the peritoneal cavity without an obvious source such as a gastrointestinal (GI) tract perforation.8,9 This most frequently occurs in patients who have ascites secondary to cirrhosis of the liver, congestive heart failure, and renal dialysis, among other disorders. Patients suffering from primary peritonitis rarely require intensive care. However, primary peritonitis may well be present in patients requiring intensive care for other reasons. For example, a cirrhotic patient with portal hypertension and ascites may develop primary peritonitis that precipitates hepatic decompensation, leading to variceal bleeding and hypovolemic shock necessitating ICU admission.
The clinical presentation is usually one of fever and physical signs of peritoneal irritation. However, approximately one third of patients with primary peritonitis have no signs or symptoms of sepsis referable to the abdomen. Diagnosis is based on clinical suspicion, the patient's presentation, and the Gram stain and culture results obtained from ascitic fluid aspiration. Culture of infected ascitic fluid usually yields aerobic enteric organisms; however, approximately 35% of patients with these diseases will have negative ascitic fluid cultures.9 Blood cultures are also frequently positive in these patients. Primary bacterial peritonitis may be assumed to be present when the ascitic fluid neutrophil count is >250/μL. The diagnosis may be confirmed in culture-negative patients by the response to appropriate antibiotic treatment, which should be evident within 48 hours and is characterized by both clinical improvement and a decrease in the number of white blood cells present in the ascitic fluid.
It is most important to distinguish primary from secondary bacterial peritonitis. Secondary bacterial peritonitis is caused by contamination from the gut lumen, and multiple organisms thus are usually found in the ascitic fluid Gram stain or culture. Patients with secondary bacterial peritonitis are unlikely to respond to antibiotic administration alone, and usually require surgical treatment.
Antibiotic treatment should be initiated on clinical suspicion of primary peritonitis and before final culture and sensitivity results are available. Broad-spectrum treatment with either ampicillin and an aminoglycoside, a third-generation cephalosporin, or ciprofloxacin is usually sufficient. The prognosis of these patients depends almost entirely on the underlying liver disease responsible for the ascites.
Patients who develop peritonitis secondary to an infected peritoneal dialysis catheter may or may not respond to antibiotic treatment (usually instilled into the dialysis fluid). Nonresponse necessitates removal of the catheter and temporary hemodialysis.
Both sepsis and hyperbilirubinemia are common findings in critically ill patients. When these conditions coexist, the question of biliary tract sepsis arises. However, most jaundiced ICU patients do not have pathology in their biliary tract.10 Before we discuss biliary tract sepsis, we will briefly outline the approach to the jaundiced patient.
Infection in the biliary tree results in one or more of three rather different clinical entities. Most commonly, infection is only present in the gallbladder and is due to obstruction of the cystic duct by a gallstone. This entity, acute calculous cholecystitis, is usually treated surgically by removal of the gallbladder, and only results in ICU admission if the patient has other major medical problems. Of more relevance to the intensivist are the syndromes of acute cholangitis and acute acalculous cholecystitis.
The Jaundiced Intensive Care Unit Patient
Abnormalities on liver function tests and even clinically evident jaundice are quite common in the ICU patient population. LaMont and Isselbacher, in a classic review article on postoperative jaundice,10 provide a simple classification, distinguishing jaundice caused by increased pigment load, by impaired hepatocellular function, and by extrahepatic obstruction. The vast majority of ICU patients with abnormal liver function test results represent the first two categories. Exclusion of extrahepatic biliary obstruction is best accomplished11 by history, physical examination, and routine laboratory tests (i.e., the clinical context) and ultrasonography to look specifically for evidence of bile duct dilation. In certain unusual circumstances,11 obstructed bile ducts may not be dilated, and clinical suspicion will necessitate direct visualization of the biliary tree with either endoscopic retrograde cholangiopancreatography (ERCP) or percutaneous transhepatic cholangiography (PTC).
Cholangitis is defined as a bacterial infection within bile ducts. It occurs when bacteria are introduced into a partially or totally obstructed duct system. In this situation the bacteria multiply rapidly in the bile and induce an inflammatory reaction around the small biliary radicles in the liver. When the hydrostatic pressure in the ducts exceeds a certain point, bacteria are literally forced out of the bile canaliculi and into the hepatic sinusoids, resulting in systemic bacteremia. The microbiology of this disease is similar to that of peritonitis caused by GI tract perforation. The portal of entry of bacteria into the biliary tree remains a topic of debate.
The clinical presentation is usually obvious, with right upper quadrant abdominal pain, jaundice, and fever (Charcot triad). Patients may be only mildly ill with bactobilia, or critically ill with frank pus under pressure in the biliary tree. Diagnosis may be confirmed with ultrasound, ERCP, or PTC. Treatment consists of biliary decompression and the administration of appropriate broad-spectrum antibiotics to cover aerobic gram-negative rods, anaerobic gram-negative bacilli, and enterococci. Bile and blood for culture are obtained before antibiotics are started (if possible) to be sure the antibiotic coverage is optimal.
These patients may require intensive care if they develop septic shock, either from the disease itself, or from bacteremia induced by an invasive procedure used to diagnose or treat the disease. From the intensivist's point of view, the management is straightforward—support the cardiorespiratory system until the ducts have been decompressed and the patient is hemodynamically stable and ensure that appropriate antibiotics are administered. It must be emphasized that the key to success is adequate biliary decompression. This may be accomplished either surgically (exploration of the common bile duct and insertion of a T tube), endoscopically (insertion of a nasobiliary drain, sphincterotomy and stone extraction, or placement of an internal stent), or percutaneously (placement of a transhepatic external drain). The adequacy of drainage is confirmed by rapid clinical and biochemical improvement (obvious within 24 to 48 hours). A patient with cholangitis who does not improve clinically and may require ongoing cardiopulmonary support must be assumed to have inadequate drainage. In this case, it is essential to image the biliary tree to determine the adequacy of the drainage and to plan for further decompression. It may also be necessary to obtain a CT scan to look for hepatic abscess(es) induced by the cholangitis.
This disease is a child of modern intensive care. It was well summarized in a 1978 article by Long and coworkers:
Acute acalculous cholecystitis is a treacherous and potentially lethal disease. It may occur in patients without known biliary tract disease who are severely compromised by trauma or gastrointestinal dysfunction and require prolonged intensive care. Its onset is insidious, its presenting symptoms are inconstant, and its neglect can lead to necrosis of the gallbladder with sepsis and death.12
This disease occurs in approximately 0.5% to 1.5% of long-term (>1 week) ICU patients.13,14 The etiology of this disease is unknown, and hence there are no known ways to prevent it.15 It appears that in the occasional critically ill patient the gallbladder wall becomes inflamed and infected with enteric organisms. The cystic duct becomes edematous and occluded, but it is not known whether this event is the cause or a result of the disease. This condition may occur at any time during the patient's ICU stay. In about one-third of patients, the inflammation has induced partial or complete necrosis of the gallbladder by the time of diagnosis.
The real problem with this disease is the difficulty in achieving a definitive diagnosis without resorting to laparotomy or laparoscopy. Attention should be focused on the gallbladder when sepsis or organ failure occurs without a known cause. The patient may have right upper quadrant abdominal tenderness, but an ICU patient can have this sign without cholecystitis. Distention of the liver capsule due to increased venous pressure or sepsis can produce identical physical findings. Liver function tests are not helpful. A biliary radionuclide scan is helpful, in that filling of the gallbladder (perhaps with the aid of morphine-induced spasm of the sphincter of Oddi16) implies cystic duct patency, which effectively rules out cholecystitis. The most valuable tests are ultrasonography and CT. Findings of pericholecystic fluid (without ascites), intramural gas, or a sloughed mucosal membrane are virtually diagnostic (Fig. 89-1). Unfortunately, not all patients have these findings.14 A thick-walled gallbladder is suggestive, unless the patient has generalized edema. Percutaneous bile aspiration for culture has high false-positive and false-negative rates and is therefore not helpful.
CT scan diagnostic of acalculous cholecystitis. The patient has an enlarged thick-walled gallbladder with intraluminal gas. The liver around the gallbladder fossa is markedly edematous.
It is our practice to operate on these patients if our clinical index of suspicion is sufficiently high and the patient is deteriorating without obvious cause. About half the patients we operate on for this reason have the disease. These patients are treated with cholecystectomy. In the remaining half, we place a cholecystostomy tube to prevent the disease in the future. To date, direct visualization of the gallbladder in the operating room is the only completely accurate method of diagnosis. This can be carried out under local anesthesia, but this approach seems somewhat pointless in the ventilated patient. In the rare patient who is felt to be “too sick to undergo laparotomy,” percutaneous transhepatic drainage of the gallbladder may result in significant clinical improvement.17 This procedure can be carried out at the bedside under ultrasound guidance. It is our experience, however, that these patients should later undergo cholecystectomy before the drainage tube is removed.
Secondary Bacterial Peritonitis
In this section we will discuss the ICU management of patients with secondary bacterial peritonitis, defined as the presence of pus or gastrointestinal contents in the peritoneal cavity.18 This condition may be either localized (an abscess) or diffuse (generalized peritonitis). Patients with physical signs of peritoneal irritation due to localized gastrointestinal tract infections are discussed elsewhere in this text.
Patients with secondary bacterial peritonitis requiring intensive care constitute a significant fraction both of all surgical ICU admissions and of all patients with peritonitis. For example, at our hospital, 300 patients with generalized peritonitis or abdominal abscess were treated in a recent 5-year period. Of these patients, 107 (mean Acute Physiology and Chronic Health Evaluation [APACHE] II score, 19.3) required posttreatment ventilatory support for an average of 9.6 days (78 treated surgically, 29 treated percutaneously), and 193 (mean APACHE II score, 10.4) did not require ventilation (137 treated surgically, 56 treated percutaneously). Patients requiring postoperative ventilatory support were severely ill and had a 64% mortality rate, compared to patients not requiring such treatment, who had only an 11% mortality rate. Therefore the need for mechanical ventilation may be a marker for general severity of illness, which results in a poor prognosis.
Pathophysiology and Treatment
Generalized peritonitis is usually caused by either bowel infarction, or more commonly, bowel perforation. Occasionally, generalized bacterial peritonitis results from perforation of an infected gallbladder, infected pancreatic pseudocyst, or other rare disease. We shall restrict our attention in this chapter to the vast majority of cases which are due to gastrointestinal tract perforation.
Patients with generalized peritonitis may become very ill because of the anatomy of the peritoneal cavity. Its large surface area permits massive loss of fluid into the abdomen and rapid absorption of bacteria, endotoxin, and inflammatory mediators into the systemic circulation. The hemodynamic effects of generalized peritonitis have been compared with those of a burn covering 50% of the body surface.
Treatment of this disease is well established.19 Following rapid fluid resuscitation and the initiation of antibiotic therapy, patients undergo laparotomy to close the perforation (or resect, exteriorize, etc) and to remove as much of the contaminating material and inflammatory exudate as possible. Broad-spectrum antibiotics are used to eliminate residual bacterial contamination after laparotomy and because septicemia is common in these patients. The gold standard in antibiotic therapy is the combination of an aminoglycoside and an antianaerobe agent such as metronidazole or clindamycin. To avoid nephrotoxicity, aminoglycosides should not be given to patients with a significantly decreased glomerular filtration rate (e.g., patients with elevated serum creatinine level, advanced age, or hypotension). In the last 15 years, a plethora of new and very effective antibiotics have appeared.20,21 Aminoglycosides may be replaced with a third-generation cephalosporin or with ciprofloxacin. Single-agent therapy with a second-generation cephalosporin (cefotetan) is effective for mild to moderate infections. Imipenem should be reserved for severe infections. Since the administration of broad-spectrum antibiotics may lead to systemic candidiasis, pseudomembranous colitis, and the selection of multiply resistant organisms, the drugs should be discontinued as soon as the acute episode of abdominal infection has subsided. Unfortunately, there is no straightforward guide to antibiotic discontinuation. The conventional criteria of clinical improvement, no fever, and no leukocytosis are associated with an exceedingly small rate of relapse of peritoneal infection, but may never be achieved in many ICU patients.22 In this group, we believe that it is reasonable to stop antibiotics when the patient no longer has physical signs of peritonitis or ileus. If the patient still has fever or leukocytosis, a CT scan should be obtained to rule out intra-abdominal abscess. If no abscess is found, antibiotics can be stopped and a search begun for an extraperitoneal source of infection.
The mortality rate of generalized peritonitis is about 30%. Known prognostic factors include age, pre-existing disease, severity of physiologic derangement at the time of diagnosis, steroid dependency, and peritonitis occurring in the postoperative period.23 Interestingly, the site of perforation is not a determinant of mortality (with the exception of perforated appendicitis). Perforated peptic ulcers are just as deadly as perforated colons. The cause of death is usually uncontrolled sepsis. Of our 107 ICU patients with peritonitis, 68 died. IAS was considered the main cause in 37 patients (55%) and was a contributing cause in 24 patients (35%). Only 7 patients (10%) suffered a nonseptic death. Essentially all the septic deaths were due to multiple organ dysfunction syndrome (MODS).
It is widely assumed that the cause of MODS in these patients is bacterial infection. (It is also possible that the uncontrolled infections are just another result of MODS caused by an unknown disorder of immune regulation). When considering treatment of these patients, it is helpful to classify the possible anatomic sites of bacterial infection (Table 89-3). At present, we do not have specific treatments for all intra-abdominal sites of infection. The prevention and treatment of infection in fibrin deposits, tissues, or the gut lumen is a very active research field at this time. For example, many postoperative abscesses likely start as small collections of fibrin-encased bacteria. Translocation of bacteria from the gut lumen may be prevented with appropriate enteral nutrition. Interestingly, patients who developed nosocomial pneumonia complicating intra-abdominal sepsis were more likely to die than patients who developed recurrent intra-abdominal infection—perhaps because pneumonia is more difficult to treat than intra-abdominal abscess.24
Table 89–3. Anatomic Sites of Bacterial Infection in Postoperative Patients |Favorite Table|Download (.pdf)
Table 89–3. Anatomic Sites of Bacterial Infection in Postoperative Patients
- Peritoneal fluid
- Peritoneal fibrin
- Extraperitoneal tissues (e.g., hepatic macrophages)
- Visceral abscess
- Within the gastrointestinal tract lumen (bacterial translocation, Clostridium difficile colitis)
- Infected prosthetic vascular graft
- Acalculous cholecystitis
- Soft tissue infection
- Intravascular catheter infection
- Disseminated candidiasis
It is the intensivist's role to support these patients postoperatively and be aware of the possible need for further surgical intervention as well as the potential for a large variety of possible complications. Close collaboration between the intensivist and the surgeon is essential for the optimal management of the patient. Supportive treatment is usually straightforward and includes hemodynamic support (recovering from shock), respiratory support, nutritional support, administration of antibiotics and activated protein C (beginning 12 hours postoperatively in selected patients), and if necessary, renal support with dialysis or hemofiltration. However, complications are many and varied and include complications of the supportive care (e.g., nosocomial infection), complications that may occur in any surgical patient (e.g., pulmonary embolus), and complications that are specific to patients treated surgically for peritonitis. We only consider the latter group of problems (Table 89-4).
Table 89–4. Postoperative Complications Specifically Related to the Surgical Treatment of Peritonitis |Favorite Table|Download (.pdf)
Table 89–4. Postoperative Complications Specifically Related to the Surgical Treatment of Peritonitis
- Wound complications
- Wound infection
- Necrotizing soft tissue infection
- Fascial dehiscence/evisceration
- Gastrointestinal tract complications
- Mechanical obstruction
- Enterocutaneous fistula
- Gastrointestinal bleeding
- Anastomotic disruption or perforation
- Ischemic bowel
- Antibiotic-associated colitis
- Complications arising in the peritoneal cavity
- Abscess formation
- Intra-abdominal bleeding
- Compartment syndrome
- Postoperative pancreatitis
- Acalculous cholecystitis
These patients must be examined daily by the intensivist as well as the surgeon, with this list of complications in mind. First, all dressings covering the abdomen should be removed and the wound examined. The skin incision is usually packed open at operation to minimize the incidence of wound infection.25 Fascial dehiscence may be observed. This complication is not rare and should prompt an immediate surgical consultation. Dehiscence occurred in 17 of our 107 ventilated peritonitis patients. Dehiscence is most common on postoperative days 4 to 8 but may occur at any time. It is heralded by the drainage of serosanguineous fluid through the incision. The diagnosis is confirmed by palpation of the wound (if the skin incision was closed) or the observation of intra-abdominal contents in the wound (if the wound was packed open). The presence of loops of bowel in the wound poses the acute risk of evisceration. In some patients with particularly severe peritonitis, it is not technically possible to reapproximate the fascial edges at the end of the surgical procedure. The surgeon then sutures an artificial mesh or other pliable material to the fascia or skin to prevent postoperative evisceration, and the wound is packed with saline-soaked gauze.26 These patients are particularly at risk for the formation of enterocutaneous fistulae at the surface of their open wounds, and this complication is easily diagnosed by inspection.27 Finally, all tubes or drains should be inspected to make sure that they have not been dislodged and are functioning as intended. For example, sump drains should be checked to make sure that the air inlet ports are not occluded.
Second, the intensivist needs to determine whether the gastrointestinal tract is functioning well enough to allow enteral feedings. This is sometimes very difficult to determine in the sedated, ventilated patient, and it is frequently necessary to challenge the patient by starting tube feedings and simply checking the gastric residual volume every 4 hours. As discussed in Chap. 11, enteral feeding is preferred over parenteral feeding in this patient population, if technically feasible. Jejunal feeds are almost always tolerated, even in patients with severe peritonitis. Early enteral feeding may improve outcome.28
Third, the intensivist must determine if the patient is septic and if the septic focus is intra-abdominal. The most common IAS complication is abscess formation, which occurred in 21 of the 107 peritonitis patients who required postoperative ventilation in our series. It is often difficult to determine when the patient's original septic response is abating and when a new septic response is being mounted.29 In general, if the patient is not improving steadily following surgery or if the patient begins to deteriorate in any way, a CT scan of the abdomen should be obtained to identify and localize a possible abscess (Fig. 89-2). However, it is usually not fruitful to scan the patient sooner than 5 to 7 days after laparotomy. Patients in this early postoperative period frequently have multiple intra-abdominal fluid collections, most of which are sterile. The patient who is deteriorating in the first week following surgery for peritonitis and who is thought to have persistent or recurrent peritoneal infection usually requires repeat laparotomy. Beyond this early phase, when abscesses are better formed and sterile collections have been resorbed, percutaneous drainage offers a safe and effective method for the diagnosis and control of abscesses.30 The tendency to broaden the antibiotic coverage when a patient continues to have fever or develops recurrent fever should be resisted.31
CT scan demonstrates massive left subphrenic abscess. The patient developed septic shock 11 days after total gastrectomy for adenocarcinoma of the stomach. Following resuscitation in the ICU, this scan was obtained and the abscess was drained percutaneously. The anastomotic leak was later repaired surgically.
In certain circumstances patients with peritonitis require “open abdomen” treatment. The skin and fascia are not closed and evisceration is prevented by suturing an artificial mesh to the fascia or skin. These patients fall into two categories—patients whose fascia could not be closed for technical reasons but who are otherwise stable, and patients whose peritonitis is so severe that in the surgeon's opinion the abdomen should be left open to facilitate repeated laparotomies for peritoneal toilet. The latter group presents a major problem to the intensivist and the ICU nursing staff. These patients undergo laparotomy (through the mesh) every 1 to 3 days until the surgeon feels that the peritoneal cavity is sufficiently clean. Weaning from ventilatory support is almost always impossible until after the last scheduled relaparotomy.32 Furthermore, during this period of repeated laparotomies, large quantities of proteinaceous fluids are lost through the open abdominal wound, and the patients may therefore require support with plasma as well as aggressive nutritional support. Fortunately, such drastic treatment for peritonitis is rarely required and has not been proved to be more efficacious than conventional management.33
Pyogenic liver abscess is a rare condition that can occur in an ICU patient. Almost any bacterium may be cultured. A wide variety of conditions may give rise to a hepatic abscess (Table 89-5), but it should be noted that at least 20% of cases are cryptogenic.34
Table 89–5. Etiology of Hepatic Abscess |Favorite Table|Download (.pdf)
Table 89–5. Etiology of Hepatic Abscess
- Perihepatic sepsis
- Systemic bacteremia
- Portal bacteremia
Patients with hepatic abscess usually present with sepsis associated with right upper quadrant pain and occasionally with peritoneal findings such as an enlarged liver. Liver function test results are frequently abnormal. The diagnosis is confirmed by CT or ultrasound examination (Fig. 89-3).
CT scan demonstrates large pyogenic liver abscess. The patient presented with septic shock and abnormal liver function test results. Following resuscitation in the ICU, an ultrasound examination failed to support the clinical diagnosis of acute cholangitis. This CT scan was therefore obtained, and the abscess was drained percutaneously. No cause for the abscess was found.
The preferred treatment of hepatic abscess is percutaneous drainage for large abscesses. It should be noted that the presence of more than one large abscess does not necessitate open surgical treatment. Antibiotics should be administered as described for patients with peritonitis until culture results are available. This may be the only treatment required for patients with multiple small abscesses, usually secondary to cholangitis, after bile duct drainage has been established.
Splenic abscess formation is rare. It may be due to trauma, direct extension of a septic process such as pancreatic abscess, infection of a splenic infarct, or bacteremia. These patients present with left upper quadrant abdominal pain, left pleural effusion, or sepsis of unknown etiology, and the diagnosis is established by CT or ultrasound examination of the abdomen. The optimal treatment is splenectomy, but percutaneous drainage may be used as a temporizing maneuver.35