Pyogenic Hepatic Abscess Drainage

Pyogenic hepatic abscesses result in considerable morbidity, and mortality occurs if abscesses are left untreated. Advances in cross-sectional imaging and interventional radiology have revolutionized the diagnosis and therapy of liver abscesses with a dramatic improvement in prognosis (see Chapter 66). Mortality has decreased from 65% (1952–1972) to 31% (1973–1993). This decrease occurred in the face of no significant change in patient characteristics, such as age and sex distribution, or incidence of associated conditions, such as diabetes mellitus, steroid use, chronic liver disease, chronic pancreatitis, and pyelonephritis (Huang et al, 1996). Another study from Denmark found the incidence rate of pyogenic abscesses increased from 6 to 18 per 1 million men and from 8 to 12 per 1 million women, and the 30-day mortality rate decreased from 40% for men and 50% for women to around 10% for both genders (Jepsen, 2005).

The liver is the most common site of abdominal visceral abscesses. Liver abscess is a result of either direct spread of infection from a contiguous focus, such as the biliary tree, or hematogenous seeding of the liver via the portal vein or hepatic artery. Biliary tract pathology is the most common cause of liver abscess. Of the nonbiliary diseases that lead to hepatic abscess, appendicitis with rupture and suppurative pylephlebitis, usually from infection in the female genital tract, are frequent causes (Zaleznik & Kasper, 1998).

The constellation of a liver abscess and portal vein thrombosis with cavernous transformation of the portal vein may mimic the findings seen in hepatocellular carcinoma (HCC; Fig. 28.1). Clues to the correct diagnosis are the clinical and laboratory parameters that suggest infection and the absence of cirrhosis and portal hypertension, both of which are commonly seen with HCC. With successful treatment, complete resolution of both the abscess (see Fig. 28.1B) and the cavernous transformation of the portal vein (see Fig. 28.1C) is seen in many cases.

FIGURE 28.1 Pylephlebitis. A, Computed tomography (CT) scan showing complex mass in the caudate lobe and portal vein thrombosis. The unusual pattern of parenchymal liver enhancement with central low density is characteristic in patients with portal vein occlusion. B, CT scan 1 month after treatment with antibiotics shows that the abscess in the caudate lobe has largely resolved. C, Findings of cavernous transformation of the portal vein on ultrasound.

Hematogenous seeding of the liver via the hepatic arterial system is unusual and should be suspected if nongastrointestinal flora—particularly a single organism, such as Staphylococcus aureus or streptococcal species—is isolated from a liver abscess. Cryptogenic abscesses may be seen in diabetic patients; however, a diligent search should always be made for a primary cause, such as diverticulitis.

The microbial flora often reflect the source of the infectious process (see Chapter 11). In a study of 111 patients with pyogenic liver abscesses, Klebsiella species and Escherichia coli were the most common offenders (Lok et al, 2008). Anaerobes generally are encountered with a previous history of biliary intervention (e.g., surgery or stenting). With nonbiliary gastrointestinal (GI) sources of infection, polymicrobial gram-negative aerobes and anaerobes predominate, whereas systemic bacteremia is more likely to give rise to a single isolate, often S. aureus. Other reported microbes reflect individual patient factors, such as Entamoeba histolytica with a history of residence in or travel to endemic areas, and Candida species in neutropenic patients.

Clinical symptoms and signs are nonspecific. Of 111 patients with pyogenic liver abscess (Lok et al, 2008), 91% were seen initially with fever, 79% with right upper quadrant pain, 18% with anorexia, and 18% with septic shock. Laboratory abnormalities also are nonspecific. In the same study, 93% had hypoalbuminemia, 75% leukocytosis, 72% elevated alkaline phosphatase, and 59% elevated alanine transaminase.

US and CT are the imaging modalities of choice (Zaleznik & Kasper, 1997). Radionuclide scans (gallium, indium, and technetium) and MRI can provide further diagnostic information, but rarely are necessary.

Decompression of the abscess cavity and systemic antibiotic therapy are the mainstays of treatment for liver abscess. Decompression evacuates the pus and endotoxins and relieves the intracavitary pressure; this ameliorates systemic toxic symptoms by decreasing hematogenous dissemination of toxins. It also allows better penetration of antibiotics into the wall of the abscess. In general, large abscess cavities should be drained, and aspiration alone should only be considered for smaller lesions.

Uncomplicated pyogenic liver abscesses may be single or multiple, often are unilocular, and may be septated (Fig. 28.2). It may be impossible to evacuate complex multiloculated abscesses completely at the time of initial catheter placement (see Fig. 28.2B). Typically, there is free communication between the loculi (see Fig. 28.2C); consequently, most resolve completely when treated with a single catheter (see Fig. 28.2D) and appropriate antibiotics. If in doubt about adequate drainage of multiple locules or multiple abscesses, more than one catheter is indicated. Communication with the biliary tree may be seen when the drainage catheters are injected with contrast material as a part of performing an abscessogram (see Fig. 28.2C). The catheter contrast injections should not be performed until the pus has been evacuated to ensure that septicemia is not provoked iatrogenically.

FIGURE 28.2 Liver abscess. A, Computed tomography (CT) scan showing multiloculated abscess in right lobe with dilated intrahepatic bile ducts. B, CT scan at the time of drainage catheter placement. C, Contrast study via the drainage catheter shows typical, irregular, loculated-type appearance and communication with biliary tree. D, Follow-up CT scan 3 months after treatment of hepatic abscess shows complete resolution.

Regarding the choice of percutaneous needle aspiration versus catheter drainage, opinions vary. Giorgio and colleagues (2006) described their experience with 118 patients treated with US-guided percutaneous needle aspiration and systemic antibiotic therapy. Cure was achieved in all 118 patients, and in 36 instances with only a single puncture. No recurrences were seen at 7 to 42 months of follow-up. The use of percutaneous needle aspiration and antibiotics is simple, less expensive, and more comfortable for the patient than catheter drainage. In practice, however, more than one aspiration is usually necessary for complete treatment with this strategy. Yu et al (2004) performed percutaneous needle aspiration in 64 patients with pyogenic liver abscess and concluded it was probably as effective as continuous percutaneous catheter drainage; however, their data were not statistically significant. Rajak and colleagues (1998) also compared the two methods of percutaneous treatment in 50 patients; they concluded that percutaneous catheter drainage was more effective than percutaneous needle aspiration. A failure of treatment was defined as a lack of response to a second attempt of percutaneous needle aspiration.

A more recent study (Zerem, 2007) also compared the two treatment methods in 60 patients, with the lack of response to percutaneous needle aspiration defined at three attempts. These authors agreed with Rajak and colleagues (1998) and concluded that percutaneous catheter drainage was more effective than percutaneous needle aspiration in the management of liver abscess. They opined that percutaneous catheter drainage is more efficient for multiloculated liver abscesses, and that percutaneous needle aspiration is a valid alternative for abscesses 5 cm or less in greatest diameter. Currently, treatment of pyogenic liver abscess with drainage catheter placement is the most commonly accepted method among interventional radiologists. Percutaneous aspiration without catheter drainage is not recommended except for occasional patients with one or more small liver abscesses that are too small for a catheter.

Surgical intervention for liver abscesses should be reserved for cases with coexisting surgically correctable causes of liver abscess (e.g., biliary tree abnormalities) or for cases of failed percutaneous therapy (see Chapter 66). A recent study demonstrated that partial liver resection is a safe and successful treatment for large, multiloculated complex pyogenic liver abscesses (Hope et al, 2008). Some associated biliary tree abnormalities may be amenable to endoscopic correction. In this subset of patients, endoscopic retrograde cholangiopancreatography (ERCP) might be preferred over surgery (Lam et al, 1999). Another study reported that previous endoscopic sphincterotomy was a good prognostic factor for early resolution of liver abscesses within 6 weeks (Wong et al, 2002).

Empiric antibiotic therapy is initiated on the suspected source of infection, with later adjustment based on the results of culture and sensitivity of the aspirated specimen. Several weeks of therapy may be required. Sending aspirates to the laboratory for aerobic and anaerobic cultures at the time of initial access is crucial for prescribing an appropriate antibiotic regimen.

Amebic Abscess

In a patient with a hepatic abscess and a history of travel to or residence in endemic areas, the possibility of infection with Entamoeba histolytica should be considered (see Chapter 67). Hepatic amebic abscess is the most common extraintestinal manifestation of amebiasis. Conditions that affect cell-mediated immunity—such as extremes of age, pregnancy, corticosteroid therapy, malignancy, HIV infection, and malnutrition—also may increase the chance that enteric E. histolytica infection results in extraluminal disease with liver involvement. The clinical findings are similar to the findings for pyogenic abscess; in addition, amebic abscesses may have superimposed bacterial infection, and they usually are solitary and most commonly located in the right lobe, adjacent to the liver capsule (Brindicci et al, 2006).

Serologic testing is the most widely used method to diagnose hepatic amebic abscess. Aspiration may be used to establish the diagnosis and to exclude secondary bacterial infection. When aspiration is used for diagnosis, the cavity should be aspirated as completely as possible. Amebic abscesses contain acellular, proteinaceous debris and an “anchovy paste,” a chocolate-colored fluid consisting predominantly of necrotic hepatocytes. Antigen tests and polymerase chain reaction on aspirated material may be helpful to confirm the diagnosis. Biopsy of the abscess wall may also establish the diagnosis if the other tests are inconclusive, which is the case in about 10% of patients (vanSonnenberg et al, 1985).

Amebic liver abscesses can generally be treated successfully with antibiotics alone (Chavez-Tapia et al, 2009). With clinical suspicion of amebic liver abscess, empiric treatment with metronidazole and a luminal agent (paromomycin or iodoquinol) may be started while awaiting confirmatory tests. Indications for catheter drainage of amebic abscess include failed medical therapy, secondary pyogenic infection, pregnancy, perforated abscess, false-negative serology, and left lobe abscess because of the proclivity to rupture or spread into the pericardium (vanSonnenberg et al, 1985; Ken et al, 1989; Hanna et al, 2000).

Echinococcal Disease

Echinococcal disease is an endemic zoonosis acquired from ingestion of Echinococcus granulosus eggs from infected animals (see Chapter 68). The most common location of hydatid cysts is the liver. Echinococcal cysts in the liver have a characteristic multiloculated or finely multiseptated appearance, and the right lobe is affected in 60% to 85% of cases.

Echinococcosis rarely manifests with an acute illness. Symptoms are unusual before the cyst has reached at least 10 cm in diameter. A large cyst can produce hepatomegaly with or without associated right upper quadrant pain, nausea, and vomiting. Echinococcus granulosus cysts can rupture into the biliary tree and produce biliary colic, obstructive jaundice, cholangitis, or pancreatitis. Pressure or mass effect on the bile ducts, portal and hepatic veins, or inferior vena cava can result in cholestasis, portal hypertension, venous obstruction, or Budd-Chiari syndrome. Liver cysts also can rupture into the peritoneum causing peritonitis, or they can rupture transdiaphragmatically into the bronchial tree, causing pulmonary hydatidosis or a bronchial fistula. Secondary bacterial infection of the cysts can result in liver abscesses. Serologic tests are often useful to establish the diagnosis, in conjunction with the typical clinical and radiologic presentation.

Traditionally, surgery has been the recommended treatment for hepatic cysts (World Health Organization, 1996), but the combination of percutaneous drainage with drug therapy is a more recent alternative. Although percutaneous aspiration previously was contraindicated because of the risk of anaphylaxis, this complication has been shown to be rare (Akhan et al, 1996). Bret and associates (1988) demonstrated the safety of aspiration or drainage of these cysts in 13 patients, and none of the patients who underwent drainage was premedicated. If a catheter was placed for drainage, a scolicidal agent was injected into the cyst cavity after cystography. No complications or recurrences were reported 6 months to 1 year after treatment. Khuroo and colleagues (1997) compared percutaneous drainage with surgery for hepatic hydatid cysts and reported that percutaneous drainage, combined with albendazole, is an effective alternative to cystectomy. They demonstrated similar efficacy in cyst regression and disappearance, with advantages of a shorter hospital stay and a lower complication rate. Zerem and Jusufovic (2006) treated 72 patients with univesicular and multivesicular hepatic hydatid cysts with US-guided percutaneous drainage combined with albendazole, and 81% of the cysts in the univesicular group and 63% in the multivesicular group disappeared.

Percutaneous evacuation of cyst content (PEVAC) has been reported as a safe and effective treatment of multivesicular echinococcal cysts with or without cystobiliary fistulas, which contain undrainable material (Schipper et al, 2002). Hydatid cysts commonly have a communication with the biliary tree, which is a contraindication to injection of sclerosing agents, because this may cause diffuse injury to the biliary epithelium (Belghiti et al, 1986).

Indications

The main indication for TIPS is the treatment of acute or recurrent variceal bleeding that has failed or is refractory to medical or endoscopic therapy, or both (see Chapter 75A, Chapter 75B, Chapter 75C, Chapter 76A, Chapter 76B, Chapter 76C, Chapter 76D, Chapter 76E ). This is especially true for patients who are bleeding from gastric or peristomal varices and those with severe portal hypertensive gastropathy. It is the preferred technique as a bridge to liver transplant, compared with surgical portosystemic shunts, because when performed correctly, TIPS does not alter extrahepatic vascular anatomy. It also is indicated for the control of refractory ascites and hepatic hydrothorax. TIPS may be used to treat Budd-Chiari syndrome in patients with moderate disease who fail to improve with anticoagulation (Boyer & Haskal, 2010).

Contraindications

There are several absolute and relative contraindications to TIPS placement. Absolute contraindications include congestive heart failure, sepsis, unrelieved biliary obstruction, severe pulmonary failure, and as the primary prevention of variceal bleeding (Boyer & Haskal, 2010). Relative contraindications to TIPS include the presence of HCC or hepatic metastases along the expected path of TIPS, hepatic vein obstruction, portal vein thrombosis, severe encephalopathy, polycystic disease of the liver, uncorrectable coagulopathy, and thrombocytopenia.

Technique (See Chapter 76E)

Various techniques and different systems are available for performing TIPS (Fig. 28.3). An imaging study using US, intravenous contrast-enhanced CT, or MRI to show a patent portal vein, hepatic veins, and hepatic venous anatomy is required. In an acutely bleeding patient, upper GI endoscopy confirms esophageal varices, excludes other sources of upper GI bleeding, and allows an attempt at endoscopic treatment. TIPS usually is performed under conscious sedation, but general anesthesia also may be used. Right internal jugular access is preferred, and a vascular sheath is placed. A selective catheter is advanced into one of the hepatic veins (see Fig. 28.3A).

FIGURE 28.3 Transjugular intrahepatic portosystemic shunting. A, Right hepatic venogram performed via a transjugular catheter. B, Portogram shows enlarged coronary vein with opacification of gastroesophageal varices. C, Balloon dilation of the transhepatic tract between the right hepatic and right portal veins. D, After shunt placement. The dilated coronary vein and varices are no longer opacified.

The right hepatic vein is most commonly accessed for TIPS creation. Right atrial and free and wedged hepatic venous pressures are obtained, followed by hepatic and wedged hepatic venography. A specially designed needle is passed via the transjugular route from the right hepatic vein across the liver parenchyma into the right portal vein 2 to 3 cm from the main portal vein confluence. A guidewire is advanced into the superior mesenteric vein or splenic vein and is used to place a catheter for pressure measurements and venography (see Fig. 28.3B). The tract is predilated with a balloon (see Fig. 28.3C); an appropriately sized (8 to 12 mm) self-expanding bare stent or a specially designed covered stent is deployed, and the balloon is dilated to an appropriate size. Pressures are obtained, with the goal being a portal–systemic gradient of less than or equal to 12 mm Hg. If the gradient is greater than 12 mm Hg, certain stents (Wallstents) may be dilated with a 12-mm balloon. If this dilation does not result in a satisfactory pressure gradient, and collaterals fill persistently, or when stents other than Wallstent have been used, parallel TIPS should be considered. If the result is satisfactory (see Fig. 28.3D), the procedure is concluded, and baseline US studies are performed within the following few days.