World J. Surg.
15, 162-169, 1991
Hepatic Abscess
A r t h u r J. D o n o v a n , M . D . , Albert E. Yellin, M . D . , and Philip W. Ralls, M.D.
Departments of Surgery and Radiology, Los Angeles County University of Southern California Medical Center and the University of
Southern California School of Medicine, Los Angeles, California, U.S.A.
Hepatic abscess--amebic or pyogenic---can be diagnosed with great
accuracy by either ultrasonography or computed tomographic (CT)
scanning. Ultrasound is the modality of choice and will detect almost
100% of abscesses. Confirmation of a diagnosis of amebic liver abscess is
made by the indirect hemagglutination test that should be positive in
almost 100% of cases. Cultures of pus from the abscess and from the
blood must be obtained in cases of pyogenic liver abscess. A positive
culture of pus from the abscess has been achieved in 90% of cases.
Ultrasound or CT guidance is utilized in aspiration of a hepatic abscess.
In the treatment of an amebic liver abscess, metronidazole is the
amebicide of choice. Open drainage is contraindicated. For cases that fail
to respond to therapy with amebieides, closed drainage guided by CT or
ultrasound is performed. Secondary bacterial infection of an amebic liver
abscess is an extremely rare event.
The identification and determination of the antibiotic sensitivity of
organisms responsible for pyogenic liver abscess is a crucially important
step. Unless a celiotomy is necessary to correct an intraabd0minal process
or the abscess is extremely large, the initial treatment of pyogenic liver
abscess is a 2 week course of appropriate antibiotics followed by a 1 month
course of oral antibiotics. The majority of pyogenic liver abscesses will
respond to such treatment. If drainage of a pyogenic abscess is required,
the preferable technique is with a percutaneous CT- or ultrasound-
directed catheter. Open surgical drainage should be reserved for those
cases in which a celiotomy is required for other purposes or for the patient
who has failed a course of appropriate antibiotic therapy and closed
percutaneous drainage is not feasible,
The mortality for treatment of amebic liver abscess should be approx-
imately zero and for pyogenic liver abscess should be less than 10%.
Abscess of the liver has been recognized since the time of
Hippocrates. Such an abscess may be amebic due to infestation
with Entamoeba histolytica or pyogenic due to bacterial infec-
tion. This article will review and reflect on the experience of the
University of Southern California and the Los Angeles County
and University of Southern California Medical Center with both
forms of liver abscess and will summarize the content of a
number of publications from this institution [1-13]. These
reports have included recommendations with respect to Utiliza-
tion of diagnostic procedures and the treatment of both pyo-
genic and amebic liver abscess.
Reprint requests: Arthur J. Donovan, M.D., Department of Surgery,
1200 North State Street, Room 9900, Los Angeles, California 90033,
U.S.A.
World Journal
of Surgery
9 1991 by the Soci6t6
Internationale de Chirurgie
Pathogenesis
Amebic liver abscess is due to infestation with Entamoeba
histolytica. This protozoan parasite is present worldwide, but is
most frequently encountered in tropical climates and in associ-
ation with unsanitary living conditions. Cysts containing the
parasite are passed in the stools of the human and can survive
for prolonged periods. They are ingested with contaminated
water or food. Mobile protazoites are released from the cysts in
the intestinal tract and may or may not invade the wall of the
large bowel. The protazoites may survive in the gastrointestinal
tract for long periods of time without mural invasion. The
amebae invade the portal system and pass to the liver where the
trophozoites cause cellular necrosis. There is coalescence of
multiple areas of focal necrosis with the development of a
hepatic abscess. The lesion which forms in the liver is usually
solitary and most frequently involves the right lobe. Multiple
abscesses may occur in advanced cases. The abscess is sur-
rounded by a thin wall of granulation tissue. The amebae are
usually present within the hepatic substance at the edge of the
abscess but not within the abscess cavity itself. The content of
the amebic abscess is generally homogeneous, does not have a
foul odor, and is variable in appearance. Although traditionally
described to be like anchovy sauce, the content of the amebic
abscess may be creamy white. The more usual reddish-brown
color reflects red blood cells or necrotic hepatic tissue, or both.
Pyogenic liver abscess is almost always the result of bacterial
infection although, on rare occasions, and in particular in an
immunologically compromised individual, fungal infection may
occur. As in the case of amebic abscess, the right lobe of the
liver is most often the site. The source of the bacterial infection
is variable--through the portal vein with or without septic
pylephlebitis, through the hepatic artery, from the biliary tract,
by direct extension from an inflammatory process in adjacent
structures such as the gallbladder, or consequent to blunt or
penetrating trauma. Formerly, the portal vein was the source in
about one-half of cases, most often as a consequence of the
advanced pathology that occurred in cases of acute appendici-
tis. Diverticulitis or any other intraperitoneal inflammatory
focus can be a source for portal venous seeding. Bacteria that
enter the portal system are usually engulfed by the Kupffer cells
in the liver. A hepatic abscess may form if the quantity of
A.J. Donovan et al.: Liver Abscess 163

organisms exceeds the capacity of the Kupffer cells, there is TRACHEO-BRONCHIAL

\ PULMONARY
underling hepatic disease, or the host is immunocompromised.
With the advent of effective antibiotics, the incidence of
~,ARDIAL
pylethrombosis and septic pylephlebitis as a complication of
intraabdominal sepsis has dramatically decreased. Sporadic
PLEUP
cases of pylephlebitis may still be identified when ultrasound
and CT scan are used in a search for a source of sepsis. A recent
report in this regard from this institution included a case of
pylethrombosis and septic pylephlebitis with hepatic abscess
that followed acute appendicitis [1].
The hepatic artery is a source of contamination of the liver in
patients with systemic bacteremia. Subacute bacterial en-
docarditis would be an example. Choledochal obstruction,
bacterial contamination of bile, and cholangitis can result in
ED ABSCESS
multiple hepatic abscesses that may be overt or microscopic. In
the latter instance, the abscess may not be apparent on gross
examination of liver. Biliary tract sepsis is now the most
frequently documented cause of pyogenic liver abscess. Orien-
tal cholangiohepatitis with its associated biliary strictures can
lead to a large pyogenic abscess, in particular involving the
lateral segment of the left lobe of the liver [2]. The segment may
be essentially destroyed and replaced by the abscess. The ENTERIC ,PERITONEAL
pathogenesis of hepatic abscess due to invasion from an adja-
cent septic focus such as empyema of the gallbladder or in
conjunction with residual devitalized tissue following blunt or
penetrating trauma is apparent. .~-
A variable number of cases of pyogenic hepatic abscess have
always been described as cryptogenic, that is, without apparent Fig. 1. The routes of rupture of an amebic liver abscess are depicted.
cause. The incidence of cryptogenic abscess remains approxi- Reproduced with permission of publisher [12].
mately one-third of cases at our institution [3].
Formerly, organisms were not cultured from as many as
one-half of reported cases of pyogenic liver abscess. This may of patients with proven amebic liver abscess, diarrhea is a
have reflected inadequate bacteriologic techniques and, in complaint in an almost equal number of cases of pyogenic liver
particular, inadequate screening for anaerobic bacteria. Failure abscess. Symptoms may have existed in excess of 2 weeks in
to culture bacteria from the pus Dom either the abscess or from patients with amebic or pyogenic liver abscess, Both pyogenic
the blood of patients with pyogenic liver abscess should be rare and amebic abscesses most characteristically result in an acute
with the use of proper techniques. Both aerobic and anaerobic right upper quadrant syndrome consisting of fever, abdominal
culture is essential. pain, nausea, and mild to moderate right upper quadrant
An important characteristic of amebic liver disease, in con- tenderness on palpation. With amebic liver abscess, there may
trast to pyogenic liver abscess, is its propensity for rupture. be either an acute or a chronic thoracic syndrome [4]. Chronic
This may reflect the thin band of granulation tissue surrounding cough can reflect transpleural and pulmonic extension of the
the amebic abscess in comparison to the more fibrous wall process. Pain over the lower rib cage on the right, a sympathetic
surrounding a pyogenic liver abscess. An amebic abscess may pleural effusion, and a right basilar pulmonary infiltrate can be
rupture freely into the peritoneal cavity, form a localized seen alone or in combination. Rupture through the lung into the
intraperitoneal abscess adjacent to the liver, penetrate into the bronchial tree with evacuation of pus may result in clinical
retroperitoneum, rupture into the gastrointestinal tract, or resolution of the hepatic abscess.
rupture through the diaphragm. In the latter instance, the Derangements of hepatic function occur in both forms of
perforation may be into the free pleural space or into the lung hepatic abscess and include an elevation of the alkaline phos-
that is adherent to the pleura. A hepatopulmonary bronchial phatase in the majority of cases as well as hypoalbuminemia.
fistula can develop and the patient may evacuate the abscess Elevations of bilirubin above normal or of the alanine ami-
through the bronchial tree. Although extremely rare, a left lobe notransferase and aspartate aminotransferase to greater than
abscess can rupture into the pericardium and provoke cardiac 100 units/liter occur in the minority of cases.
tamponade. These various forms of rupture are depicted in Selected clinical and laboratory findings in amebic as con-
Figure 1. trasted to pyogenic liver abscess among 96 patients with amebic
and 48 patients with pyogenic liver abscess are depicted in
Clinical Features Table 1. There is clearly such overlap in these clinical manifes-
tations as to preclude a distinction between amebic and pyo-
For unexplained reasons, amebic liver abscess is much more genic abscess on these bases alone. Nevertheless, a young male
frequent in males between the ages of 20 and 40 years. Although from an area with a high incidence of amebiasis who has a
a history of diarrhea may be present in approximately one-third clinical picture consistent with a liver abscess, and the abscess
164 World J. Surg. Vol. 15, No. 2, Mar./Apr. 1991

Table 1. Clinical features in patients with amebic and pyogenic liver or oval with well defined margins and with the lack of prominent
abscess, a peripheral echoes. The lesions are primarily hypoechoic but
Characteristicb Amebic Pyogenic they may have areas of clumped increased echogenicity similar
to pyogenic abscess. Amebic abscesses tend to "fill in" with
Male :female ratio 19:1 1.4:1 relatively homogeneous low echogenicity at higher gain set-
Median age (yr) 28 44
Ethnicity 98% No ethnic tings. Almost all amebic abscesses show some distal sonic
Hispanic predisposition enhancement, a finding that is less consistent in pyogenic
Birthplace 95% from No predominant abscess.
Latin birthplace Statistically, only a round, oval shape of hypoechogenic
America
Right upper quadrant pain 59% 27% appearance with fine homogeneous low-density echoes at high
Symptoms >14 days 14% 37% gain is more frequent in amebic liver abscess [6]. Figure 2A
Right upper quadrant 67% 42% shows an ultrasound examination of an amebic liver abscess.
tenderness On CT scan, an amebic abscess may manifest low density with
% neutrophils >89 6% 35% smooth margins and a contrast-enhancing peripheral rim. Irreg-
Albumin <3.0 g/dl 16% 50%
Direct bilirubin > 1.0 g/dl 2% 15% ularity and internal inhomogeneity may occur [7]. Figure 2B is
Lactic dehydrogenase 7% 41% a CT scan of an amebic abscess.
>500 U/L The pyogenic liver abscess by ultrasound usually has an
Aspartate aminotransferase 9% 39% irregular margin and is bypoecboic as depicted in Figure 3A.
>100 U/L
Positive amebic serology 94% 6% (all with prior There are irregular areas of increased echogenicity within the
amebiasis) lesion. If microbubbles are present from gas-forming organ-
isms, a diffusely hyperechoic appearance may be seen and may
aThis table summarizes clinical characteristics of amebic and
pyogenic liver abscess at our medical center and is reproduced from a lead to confusion with a solid tumor. On CT scan, pyogenic
report by Barnes and associates [3]. liver abscess may have well-defined round or oval cavities or be
bAll features listed are significantly different in the 2 groups (p < lobulated with poorly-marginated edges. There is usually low
0.05). internal density. The abscess is better identified with contrast-
enhanced CT scan. Pyogenic abscesses usually appear more
complex on ultrasound than on CT scan. Marginal lobulation
documented by one of the imaging techniques to be discussed, and irregularity are shown to better advantage with CT scan-
will be presumed to have an amebic abscess until proven ning. Internal gas may be present depending on the organisms
otherwise. Conversely, a pyogenic liver abscess will be as- present. Figure 3B is a CT scan of a pyogenic liver abscess.
sumed in an individual with a hepatic abscess that has devel- In summary, there is such overlap in the features of the
oped in the setting of an infection such as acute appendicitis amebic and pyogenic liver abscess by ultrasound that an
with abscess, perforated diverticulitis, or bacterial endocardi- assured diagnosis of one or the other cannot be made based
tis. purely on the sonographic findings [6]. Either ultrasound or CT
scan may document pylethrombosis in association with liver
abscess.
Imaging
Multiple hepatic abscesses may be imaged. A particularly
Newer radiologic techniques such as ultrasound and CT scan- characteristic pattern of multiple hepatic abscess is one in
ning have greatly enhanced our ability to establish the diagnosis which the liver appears like Swiss cheese. This pattern has been
of hepatic abscess and have increased our understanding of the recently recognized and may be seen in immunocompromised
natural history of this process. Ultrasonography is the prefera- individuals such as those with HIV infection. An example is
ble initial diagnostic study for the diagnosis of liver abscess [5]. depicted in Figure 4.
Indeed, among 96 consecutive amebic abscesses treated at this Ultrasound has proven a very useful technique for documen-
institution, ultrasound was performed in 94 cases and estab- tation of the course of the amebic liver abscess [8]. The image
lished the diagnosis in each case [3]. It is cost effective as may be stable, increase, or decrease in size during the initial
compared to a CT scan. When an ultrasound examination is period of successful therapy. The appearance of the abscess
performed in a patient with an acute right upper quadrant does not correlate with clinical cure as evidenced by lack of
syndrome, it is essential to image both the liver as well as the fever and the continued absence of signs and symptoms follow-
region of the gallbladder and pancreas. Failure to do so may ing cessation of therapy. There is not a change that would
result in failure to detect an abscess within the liver and, thus, herald an adverse outcome. The majority of amebic liver
lead to additional nondiagnostic procedures. abscesses resolve to a normal sonographic liver pattern over a
Hepatic abscess may occasionally be detected by CT scan- period of up to 2 years with a median of 7 months. A hepatic
ning in patients with nonlocalized findings in whom the CT scan defect with marked abnormalities may persist for years.
is performed in search for a source of intraabdominal sepsis.
The detection of the disease in the liver may, in this instance, be Diagnosis
serendipitous. Hepatic scans utilizing radioisotopes are obso-
lete for definition of hepatic abscess. Magnetic resonance Serology is the key to the diagnosis of amebic liver abscess.
imaging does not provide information of greater usefulness than There are both indirect hemagglutination, complement fixation,
does ultrasound or CT scanning. and gel diffusion tests for detection of antibodies to Entamoeba
Amebic liver abscess by ultrasonography tends to be round histolytica. The preferable test is the indirect hemagglutination
A.J. Donovan et al.: Liver Abscess 165

Fig. 2A. Oblique liver sonogram showing a large amebic abscess. This high gain scan shows homogeneous low level echoes throughout the
abscess. B. Computed tomography scan of an amebic liver abscess with a surrounding enhanced rim.

Fig. 3A. Longitudinal sonogram of the liver, Pyogenic liver abscess is primarily hypoechoic and has an irregular margin. Note the internal
hyperechoic strands. B. A pyogenic abscess of the left lobe of the liver,

test [14]. This test is positive in almost 100% of patients with
ultimately proven amebic infestation [9]. The test is considered
diagnostic of tissue invasion at a 1:512, suspicious at 1 : 128, and
negative at 1 : 32. Titers may reach levels in the thousands. The
initial determination may be less than l : 128 in the early phases
of development of an amebic hepatic abscess. If the diagnosis of
amebic abscess is strongly suspected and the initial test is
negative, the test should be repeated. The titer may then be
elevated. Amebae will be found in the stools in the minority of
cases with amebic liver abscess, including cases with diarrhea.
Diagnostic aspiration can be performed in amebic liver ab-
scess but with a strong presumptive diagnosis, diagnostic
aspiration is not necessary. Blood is drawn for the amebic
indirect hemagglutination test and treatment is initiated. A
prompt therapeutic response and the subsequent report of a
positive hemagglutination test confirm the diagnosis. There is
not a contraindication to aspiration. Indeed, the physician who
practices in a region where only a rare case of amebic liver
abscess is encountered may feel more comfortable if diagnostic
aspiration is performed. The character of the pus has already
been described--non-foul smelling, usually reddish-brown, and
without organisms on gram stain.
Identification of the bacteria involved in a pyogenic liver
abscess is an essential diagnostic step. The bacteria should be
166 World J. Surg. Vol. 15, No. 2, Mar./Apr. 1991

Table 2. Bacteriology of liver abscess, a

Abscess Blood
Organism 38 patients 20 patients
Aerobes 43 19
Escherichia coil 16 7
Klebsiella pneumoniae 6 4
Enterobacter species 3 1
Proteus speciesh 4 2
Other gram-negative rods " 3 2
Staphylococcus aureus 1 1
Streptococcia 6 1
Enterococci 4 1
Anaerobes 36 16
Microaerophilic streptococci 7 5
Peptococci 2 0
Peptostreptococci 5 2
Bacteroides fragilis 7 4
Bacteroides species" 8 3
Fusobacterium species 5 1
Other anaerobes./. 2 1
"This table summarizes bacteria isolated from a series of pyogenic
Fig. 4. An example of multiple liver abscesses--a Swiss cheese fiver. liver abscess at our medical center and is reproduced from a report by
This process may be seen in the immunocompromised patient. Barnes and associates [3].
bp. mirabilis (3), P. vulgaris (1).
"Pseudomonas aeruginosa (1), P. maltophilia (1), Citrobacter
fi'eundii (1), Eikenella species (1).
Ja and y streptococci (4),/3 streptococci not group A, B, or D (2).
eB. melaninogenicus (3), B. ruminicola (2), B. thetaiotaomicron
sought by cultures obtained from both blood and pus from the (2), B. species untyped (1).
abscess. Percutaneous aspiration is an important technique to tCIostridium per~'ingens (1), C. species (1), Veillonella species (1).
obtain pus from the abscess. Either ultrasound or CT scan
should guide percutaneous diagnostic aspiration. A " w i n d o w "
of accessibility is necessary. In contrast to amebic abscess, the
Treatment
pus will be foul-smelling if consequent to anaerobic infection,
and the gram stain should be positive for bacteria. A diagnostic The treatment of amebic liver abscess is by amebicides. The
aspiration is indicated in all cases with a presumptive diagnosis basic principles were established by Sir Leonard Rogers who
o f pyogenic liver abscess or with uncertainty as to the diagnosis served with the British Medical Service in India and who
o f pyogenic or amebic abscess, unless an exploratory celiotomy summarized these principles in the Lettsomian Lectures to the
will be performed for indications that will be discussed. Anti- Royal Medical Society in 1928 [15-17]. Rogers reported that, by
biotic selection will be based on sensitivity of all organisms avoiding open drainage with its attendant secondary bacterial
cultured from blood or pus. Identification of these organisms infection and with the use of amebicides, the mortality for
and of their antibiotic sensitivities will be critically important in treatment of amebic liver abscess had been reduced from 56.8%
the treatment of pyogenic liver abscess. to 14.49%. The original treatment described by Rogers con-
The organisms identified among 48 patients with pyogenic sisted of therapeutic as contrasted to diagnostic needle aspira-
liver abscess reported from our institution in 1987 are summa- tion in addition to the use of amebicidal agents. Needle aspira-
tion for purposes of treatment of amebic abscess is now very
rized in Table 2. Multiple organisms were identified in 25
rarely indicated. Exceptions may be either a large amebic
patients. In 7 of 17 cases where organisms were cultured from
abscess of the left lobe of the liver where there might be risk of
both the blood and the abscess, additional organisms were
pericardial rupture or the case that fails to respond to therapy
identified in the abscess that were not found on the blood
with amebicides [18].
culture [3]. An organism was not identified in either the blood or The amebicide of choice at the time of Rogers' report was
the abscess in only 3 of the 48 cases. The abscess content was emetine. This was replaced first by chloroquine and more
not cultured in 2 of 3 of the latter cases. Bacteria were identified recently by metronidazole. Metronidazole is currently the ame-
in the pus from 90% of abscesses that were cultured. bicide of choice. A randomized study from our institution by
Identification of bacteria in a liver abscess is considered Cohen and Reynolds in 1974 established that cbloroquine and
synonymous with the diagnosis of pyogenic liver abscess. metronidazole were equally efficacious--1 of 20 cases of failure
Spontaneous bacterial contamination of an amebic liver abscess with chloroquine and I of 16 with metronidazole [9]. Chloro-
is an event of such extraordinary rarity as to challenge its quine treatment was for 10 weeks; metronidazole was adminis-
existence [4, 10, 11]. Indeed, when bacterial growth occurs in a tered orally in a dose of 750 mg 3 times a day for 10 days.
liver abscess of a patient with an elevated amebic hemaggluti- Defervescence is usual by 72 hours and absence of fever and a
nation titer, we usually assume the abscess to be pyogenic and state of well-being is usually achieved within 7-10 days. Ease of
the elevated titer an indication of prior exposure to Entamoeba administration has dictated that metronidazole is the amebicide
histolytica. of choice.
A.J. Donovan et al.: Liver Abscess
When a presumptive diagnosis of amebic liver abscess is
made, we initiate therapy not only with metronidazole but also
with an aminoglycoside and penicillin. This will provide appro-
priate antibiotic coverage should the diagnosis be erroneous
and the lesion, indeed, be pyogenic. The results of the indirect
hemagglutination test should be available within 48 hours. If
that test is positive, the drugs other than metronidazole are
discontinued. The one exception to the above would be a young
male from an endemic area in whom the diagnosis of amebic
liver abscess is so certain that the aminoglycoside and penicillin
can be omitted and the therapeutic response to metronidazole
alone observed.
When a patient with a positive hemagglutination titer fails to
respond favorably after 7-10 days of therapy with either met-
ronidazole or chloroquine, as reflected by failure to defervesce
and lack of clinical improvement, crossover therapy is recom-
mended. Very rarely, the patient will not respond to either drug
in sequence and decompression of the amebic abscess is
advised [18]. This is best accomplished by a closed technique
with a percutaneously placed catheter under ultrasound or CT
guidance. The risk of secondary bacterial contamination is less
than with open surgical drainage. Therapeutic drainage was not
required in any of the last 96 cases of amebic liver abscess
reported from this institution [3].
Bacterial contamination is largely the result of open surgical
drainage. Open surgical drainage only serves to complicate the
course of the amebic liver abscess and is strongly contraindi-
cated. An intentional open drainage of an amebic liver abscess
has not been performed at our institution for the past 10 years.
During that period of time, we have treated approximately 500
patients with amebic liver abscess. For practical purposes,
unruptured amebic liver abscess is not a surgical disease.
Traditionally, it has been believed that surgical drainage was
required for rupture of amebic liver abscess. A study of 16 cases
of rupture of amebic liver abscess at this institution suggests
that amebicidal therapy alone may be effective in at least some
such cases [12]. Nonsurgical therapy of intraperitoneal rupture
was suggested in 1931 by Biggan and Ragab [19] who treated a
patient with emetine and aspiration. Among the 16 cases that
we reported, there were 3 in whom the diagnosis of intraab-
dominal rupture was made based on diagnostic needle aspira-
tion and in whom therapy with amebicidal agents was effective
without surgical drainage. We recommend this therapy under
such circumstances.
The diagnosis of intraperitoneal rupture of amebic liver
abscess is more often made at the time of abdominal exploration
with the mistaken preoperative diagnosis of acute appendicitis
or other acute abdominal illness. Leakage from the abscess
down the right lateral gutter leads to a right lower quadrant
syndrome suggestive of acute appendicitis. The principal ther-
apy for such a ruptured amebic liver abscess is metronidazole.
External drainage invites secondary bacterial contamination.
When a ruptured liver abscess is encountered at celiotomy,
therapy with metronidazole, an aminoglycoside, and penicillin
is initiated. This therapy is appropriate for either amebic or
pyogenic abscess. Closed suction drainage of the liver abscess
is established. The results of the gram stain, cultures, and
indirect amebic hemagglutination test will be available in less
than 48 hours. Therapy is then modified. Metronidazole is
continued if the abscess is amebic and drains are removed to
167
avoid the complications associated with secondary bacterial
contamination. Appropriate antibiotics are continued if the
abscess is pyogenic and the drains are left in place.
When rupture occurs into the pleural space, thoracentesis is
performed to confirm the diagnosis. Treatment is with metro-
nidazole. Should the abscess rupture into the lung or into the
gastrointestinal tract, appropriate antibiotics are administered
in addition to metronidazole because of the presumption of
bacterial contamination. With application of these policies, 1 of
16 cases of ruptured amebic liver abscess died [12]. The correct
diagnosis was delayed in this fatal case and amebicidal therapy
was not initiated in a timely fashion.
Pyogenic liver abscess has traditionally been considered to be
a disease demanding surgical intervention. The basic principles
of extraperitoneal drainage were established by Ochsner and
associates [20] in 1938. Avoidance of contamination of the
peritoneal cavity resulted in a sharp decline in mortality.
Ochsner and DeBakey [21], in their 1943 article on amebic liver
abscess, note a report of the treatment of a patient with
secondary infection of an amebic abscess with sulfonamides.
They refer to this "promising development," that might permit
"closed drainage even in the secondary infected cases." In
1953, McFadzean and associates [22] reported a series of 14
patients with pyogenic liver abscess treated in Hong Kong with
aspiration and antibiotics and without mortality. During the
past 10 years, we have had a major interest in nonsurgical
treatment of pyogenic liver abscess [3, 13].
Diagnostic aspiration of a pyogenic abscess has already been
discussed and its crucial role emphasized. The surgeon, inter-
nist, and interventional radiologist should all be involved in
therapeutic decisions. If the abscess is secondary to either
biliary tract disease or a focus of intraabdominal sepsis, a
celiotomy is usually indicated to correct that underlying pro-
cess. At that time, the pyogenic abscess is drained. Loculations
within the abscess should be opened. The above will represent
the minority of cases of pyogenic liver abscess. Our current
practice is to treat the remaining patients with a pyogenic liver
abscess with a 2 week course of intravenous antibiotics. Initial
therapy is with metronidazole, an aminoglycoside, and penicil-
lin. These drugs should provide coverage for any organism that
might be involved and the therapy is subsequently modified
based on the results of culture and antibiotic sensitivities. If the
patient's clinical progress is favorable, the course of intrave-
nous antibiotics is followed by 1 month of oral antibiotic
therapy. Antibiotic therapy is the only available therapy for
patients with multiple small abscesses.
Drainage is necessary in patients who have failed to respond
to a course of appropriate antibiotic therapy or for an extremely
large abscess that could possibly rupture--which is a very rare
occurrence. Clinical failure is usually manifest by lack of
defervescence, a sustained elevation of the white blood count,
and persistent malaise. Defervescence is somewhat slower in
pyogenic abscess than in amebic liver abscess and at least 2
weeks of therapy is indicated before considering that antibiotic
treatment is a failure, if there is progressive improvement
during this period. If not, earlier decompression is indicated. If
a "window" is available, percutaneous catheter drainage with
CT or ultrasound guidance is the preferred technique for
drainage; open surgical drainage is reserved for cases not
amenable to catheter drainage or that are failures of that
168
Table 3. Pyogenic liver abscess: Treatment of 48 casesY
Nonsurgical treatment: Sepsis controlled
Appropriate antibiotics only 18
Appropriate antibiotics and catheter 8 abdominale associ6e ou que le volume de l'abc~s soit extr~m-
drainage
Surgical drainage
Celiotomy for other pathology 4 comprend 2 semaines d'antibiotiques adapt6s par vole g6n6rale
Failure, appropriate antibiotics 2 suivies d'un mois d'antibiotiques par vole orale. La plupart des
Inappropriate antibiotics 2 abc6s g pyog6nes r6pondront bien ~t ce traitement. Si le drain-
Inadequate duration therapy, 11
appropriate antibiotics
Autopsy diagnosis
Total
aThis table summarizes the treatment of pyogenic liver abscess as
reported by Barnes and associates [3].
technique. Generally, the principle of extraperitoneal drainage
can be followed because there is no other intraabdominal
pathology to be treated surgically.
The results of treatment of 48 cases of pyogenic liver abscess
at our institution are summarized in Table 3. As will be noted,
only 2 of the cases who received adequate antibiotic therapy
required surgical drainage. Thirteen of 19 cases with surgical
drainage had either inadequate or inappropriate antibiotic treat-
ment.
Course
The mortality for amebic liver abscess that is diagnosed and
properly treated should be approximately zero [3, 9]. The
mortality for pyogenic liver abscess has remained as high as
40% [23]. This mortality is more related to the underlying
disease such as neglected intraabdominal sepsis or to the
immunocompromised host than it is to the hepatic abscess
itself. Seven of the 48 patients with pyogenic liver abscess
reported from our institution in 1987 died [3]. In 3, the diagnosis
was not suspected until autopsy and, in 2 of the remaining
cases, antibiotic coverage was inadequate. The mortality for
cases properly diagnosed, treated with appropriate antibiotics,
and selectively drained should not exceed 10%.
R6sum6
L'abc6s h6patique amibien ou ~t pyog6nes peut ~tre diagnos-
tiqu6 avec une grande precision soit par l'6chographie, soit par
la tomodensitom6trie. L'6chographie est la mEthode de choix et
d6tecte presque 100% des abc6s. On obtient la confirmation du
diagnostic d'abc~s amibien par le test d'h6magglutination indi-
recte qui est positive dans presque 100% des cas. On dolt faire
des cultures de pus provenanf de l'abc6s et des h6mocultures en
cas d'abc6s h pyog~nes du foie. Ces cultures ont 6t6 positives
dans 90% des cas. L'6chographie et la tomodensitom6trie
aident ~t guider le drainage de abc6s.
Dans le traitement de l'abc~s amibien du foie, le m6tronida-
zole est l'amibicide de choix. Le drainage h ciel ouvert est
contreindiqu6. Pour les cas qui ne r6pondent pas aux amibi-
cides, un drainage percutan6 guid6 par la tomodensitom6trie ou
l'6chographie est indiqu6. La surinfection d'un abc6s amibien
du foie est extr6mement rare.
L'identification et la d6termination de la sensibilit6 aux
World J. Surg. Vol. 15, No. 2, Mar./Apr. 1991
antibiotiques des organismes responsables de l'abc6s ~t pyo-
26 g6nes est une 6tape extr~mement importante. A moins qu'une
laparotomie soit n6cessaire pour traiter une infection intra-
ement important, le traitement initial d'un abe6s g pyog6nes
19
age d'un abc6s ~ pyog6nes s'av~re n6cessaire, la meilleure
3 technique est percutan6e avec un cath6ter ins6r6 sous contr61e
48 tomodensitom6trique ou 6chographique. On r6servera le drain-
age chirurgical ~ ciel ouvert aux cas o0 une laparotomie est
n6cessaire pour d'autres raisons et o2 le malade n'a pas
r6pondu ~ l'antibioth6rapie adapt6e et chez qui le drainage
percutan6 est impossible ~ faire.
La mortalitd de l'abc6s amibien trait6 devrait approcher 0%
et atteindre pour l'abc6s ~, pyog6nes moins de 10%.
Resumen
El absceso hepfitico--amibiano o piog6nico--puede ser diag-
nositicado con gran precisi6n mediante la ultrasonograffa (US)
o la tomografia computadorizada (TC). La ultrasonografia es la
modalidad de escogencia; detecta casi el 100% de los abscesos.
La confirmaci6n del diagn6stico de absceso amibiano del hf-
gado se hace por la prueba de hemaglutinacifn indirecta, la cual
debe resultar positiva en pr~cticamente el 100% de los casos.
Cultivos de[ pus y de la sangre deben ser realizados en los
pacientes con abscesos pi6genos. Se logran cultivos positivos
del pus del absceso en 90% de los casos. Se utiliza la guia
ultrasonogrfifica o de tomograf[a computadorizada para la as-
piraci6n del absces'o.
El metronidazol es el agente amebicida de preferencia en el
tratamiento del absceso amibiano del higado. El drenaje abierto
esfft contraindicado. En los casos en que falla la terapia con
amibicidos, se realiza el drenaje cerrado guiado por US o pot
TC. La infecci6n secundaria de un absceso amibiano del higado
es un fen6meno extraordinariamente raro.
La identificaci6n y determinaci6n de la sensibilidad antibi6t-
ica de los microorganismos responsables del absceso pi6geno
representa un paso crucial en su manejo. A menos que se haga
necesario realizar una laparotomia para la correci6n del algfin
proceso intraabdominal o porque el absceso es excesivamente
grande, el tratamiento inicial del absceso pi6geno es un ciclo de
antibi6ticos propiados de 2 semanas, seguidos de tratamiento
con antibi6ticos orales por un rues. La mayoria de los abscesos
pi6genos del h[gado responde a este tipo de tratamiento. Si se
requiere drenaje de un absceso pi6geno, la t6cnica de preferen-
cia es la punci6n percut~nea por medio de un cat6ter guiado pot
US o TC. E1 drenaje quirfirgico abierto debe reservarse para
aquellos casos en que la laparatomia es necesaria por razones
diferentes o en que hay falla en la respuesta a un ciclo de terapia
antibi6tica adecuada y el drenaje percutfineo no es factible.
La mortalidad en el manejo del absceso amibiano del hfgado
debe ser nula, y para el absceso pi6geno de menos de 10%.
References
1. Harch, J.M., Radin, R.D., Yellin, A.E., Donovan, A.J.:
Pylethrombosis: Serendipitous radiographic diagnosis. Arch. Surg.
122:1116, 1987
A.J. Donovan et al.: Liver Abscess
2. Yellin, A.E., Donovan, A.J.: Biliary lithiasis and helminthiasis.
Am. J. Surg. 142:128, 1981
3. Barnes, P.F., De Cock, K.M., Reynolds, T.N., Rails, P.W.: A
comparison of amebic and pyogenic abscess of the liver. Medicine
66:472, 1987
4. Berne, C.J.: Diagnosis and treatment of amebic liver abscess. Surg.
Gynecol. Obstet. 75:235, 1942
5. Rails, P.W., Colletti, P.M., Quinn, M.F., Halls, J.: Sonographic
findings in hepatic amebic abscess. Radiology 145:123, 1982
6. Ralls, P.W., Barnes, P.F., Radin, D.R., Colletti, P.M., Halls, J.:
Sonographic features of amebic and pyogenic liver abscesses: A
blinded comparison. Am. J. Radiol. 149:499, 1987
7. Radin, D.R., Ralts, P.W., Colletti, P.M., Halls, J.M.: CT of amebic
liver abscess. Am. J. Radiol. 150:1297, 1988
8. Ralls, P.W., Quinn, M.F., Boswell, W.D., Colletti, P.M., Radin,
D.R., Halls, J.M.: Patterns of resolution in successfully treated
hepatic amebic abscess: Sonographic evaluation. Radiology 149:
541, 1986
9. Cohen, H.G., Reynolds, T.B.: Comparison of metronidazole and
chloroquine for the treatment of amoebic liver abscess: A con-
trolled trial. Gastroenterology 69:35, 1975
10. Turrill, F.L., Burnham, J.R.: Hepatic amebiasis. Am. J. Surg.
111:424, 1966
11. Rosoff, L.: Amebic abscess of the liver. In Textbook of Surgery,
D.C. Sabiston, Jr., editor, Philadelphia, W.B. Saunders Co., 1981,
Ch. 35
12. Greaney, G.C., Reynolds, T.B., Donovan, A.J.: Ruptured amebic
liver abscess. Arch. Surg, 120:555, 1985
13. Maher, J.A., Reynolds, T.B., Yellin, A.E.: Successful medical
treatment of pyogenic liver abscess. Gastroenterology 77:618, 1979
14. Kessel, J.F., Lewis, W.P., Pasquel, C.M.: Indirect hemagglutina-
169
tion and complement fixation tests in amebiasis. Am. J. Trop. Med.
14:540, 1965
15. Rogers, L.: Lettsomian lectures on amoebic liver abscess: Its
pathology, prevention and cure: I. Aetiology and pathology of
amoebic liver abscess. Lancet 1:463, 1922
16. Rogers, L.: Lettsomian lectures on amoebic liver abscess: Its
pathology, prevention and cure: II. The varieties and treatment of
amoebic liver abscess. Lancet 1:569, 1922
17. Rogers, L.: Lettsomian lectures on amoebic liver abscess: Its
pathology, prevention and cure: III. The prevention of amoebic
liver abscess and the recent reduction in its prevalence and mor-
tality. Lancet 1:677, 1922
18. Singh, J.P., Kashyap, A.: A comprehensive evaluation of percuta-
neous catheter drainage for resistant amebic liver abscesses. Am. J.
Surg. 158:58, 1989
19. Biggam, A.G., Ragab, A.: Rupture of amoebic liver abscess into
the peritoneal cavity: Certain points concerning its diagnosis and
treatment. J. Trop. Med. Hygiene 34:285, 1931
20. Ochsner, A., DeBakey, M., Murray, S.: Pyogenic abscess of the
liver: II. An analysis of forty-seven cases with review of the
literature. Am. J. Surg. 40:292, 1938
21. Ochsner, A., DeBakey, M.: Amebic hepatitis and hepatic abscess.
Surgery 13:460, 1943
22. McFadzean, A.J.S., Chang, K.P.S., Wong, C.C.: Solitary pyo-
genic abscess of the liver treated by closed aspiration and antibi-
otics: A report of 14 consecutive cases of recovery. Br. J. Surg,
41:141, 1953
23. Conter, R.L., Pitt, H.A., Tompkins, R.K., Longmire, Jr., W.P.:
Differentiation of pyogenic from amebic hepatic abscesses. Surg.
Gynecol. Obstet. 162:114, 1986