669

STATE-OF-THE-ART CLINICAL ARTICLE

Spontaneous Bacterial Peritonitis

Jose Such and Bruce A. Runyon

From the Hospital General Universitario, Alicante, Spain; and

Transplantation Institute, Loma Linda University Medical Center,
Loma Linda, California, USA

Spontaneous bacterial peritonitis (SBP) is probably the bestcharacterized infectious complication that develops in patients
with cirrhosis and ascites [1, 2]. Since its first description in
1964, a large body of knowledge has accumulated regarding
the clinical presentation, diagnosis, pathogenesis, treatment,
and prevention of SBP, and the prognosis of patients who
develop this infection [1 3]. Although SBP has been described
as occurring in different clinical settings, such as nephrotic
syndrome or heart failure, most SBP episodes develop in patients with advanced cirrhosis as a manifestation of severe
derangement of hepatic function. Therefore, an episode of ascitic fluid (AF) infection has been proposed as an indication
for liver transplantation, in the absence of contraindications.
Variants of AF Infection
Several variants of AF infection have been described
(table 1).
SBP. SBP has been defined as an AF infection associated
with a positive bacterial culture and an AF polymorphonuclear
(PMN) cell count of 250/mm3, in the absence of a surgically
treatable intraabdominal source of infection. SBP was the first
AF infection described and is probably the most common variant. In a large series of AF infectious episodes, 67.8% met the
above criteria for SBP. Because this infection is almost always
monomicrobial, growth of more than one organism should raise
a suspicion of secondary peritonitis (see below).
Culture-negative neutrocytic ascites (CNNA). This variant
is diagnosed when cultures of AF are negative, a PMN cell
count is 250/mm3, and when there is no surgically treatable
intraabdominal source of infection [4]. Other possible causes

Publication of the State-of-the-Art Clinical Article has been made possible

by an educational grant from Roche Laboratories.
Received 8 April 1998; revised 5 June 1998.
J. S., a research scholar at Loma Linda University Medical Center, was
supported by grants from Conselleria de Educacion y Ciencia, Generalitat
Valenciana; Asociacion Espanola para el Estudio del HB gado (AEEH) and
Sociedad Espanola de PatologB a Digestiva (SEPD).
Reprints or correspondence: Dr. Bruce Runyon, Director of Hepatology and
Medical Director of Liver Transplantation, Loma Linda University Medical
Center, 11234 Anderson Street, Room 1405, P.O. Box 2000, Loma Linda,
California 92354.
Clinical Infectious Diseases 1998;27:66976
q 1998 by the Infectious Diseases Society of America. All rights reserved.
10584838/98/27040001$03.00

/ 9C57$$OC16

09-15-98 15:47:30

of neutrocytic ascites such as peritoneal carcinomatosis, pancreatitis, and tuberculous peritonitis must be ruled out. The
clinical, prognostic, and therapeutic characteristics of CNNA
are similar to that of SBP, and CNNA, therefore, is treated in
a similar fashion.
Monomicrobial nonneutrocytic bacterascites (MNB). This
variant is characterized by the isolation of bacteria in cultures
of AF and a PMN cell count of 250/mm3. The clinical course
of MNB is dependent on the presence or absence of associated
clinical symptoms. For patients who present with MNB and
with clinical signs or symptoms suggestive of infection, the
morbidity and mortality rates are similar to those for patients
with SBP or CNNA. In contrast, among patients with asymptomatic MNB the colonization is usually resolved without antibiotic therapy.
Secondary bacterial peritonitis. This entity is diagnosed in
cases for which AF cultures are positive (usually polymicrobial), PMN cell counts are 250/mm3, and for which there is a
surgically treatable intraabdominal source of infection. Clinical
signs and symptoms do not distinguish secondary from spontaneous peritonitis; however, the AF analysis is helpful in this
regard. The AF in secondary peritonitis usually meets at least
two of the following criteria: a total protein content of 1
g/dL, a glucose concentration of 50 mg/dL, and a lactate
dehydrogenase level of 225 U/mL (or higher than the upper
limit of normal for serum). The diagnosis of secondary peritonitis must be made early in the course of illness, since death is
the usual outcome in the absence of surgical correction.
Polymicrobial bacterascites. Polymicrobial bacterascites
is diagnosed when gram staining or cultures of AF demonstrate
multiple organisms and there is a PMN cell count of
250/mm3. This variant usually occurs as a result of inadvertent puncture of the intestines during attempted paracentesis.
Fortunately, this is a rare event, occurring in 1 of 1,000
paracenteses. Ileus, the presence of multiple surgical scars, and
inexperience of the operator are risk factors for this iatrogenic
variant of AF infection. If the AF protein concentration is 1
g/dL and the osponic activity of the fluid is adequate, this
colonization resolves spontaneously.

Flora
More than 60% of SBP episodes are caused by gram-negative enteric bacteria [5]. Escherichia coli and Klebsiella pneu-

cida

UC: CID

670

Such and Runyon

Table 1. Classification of ascitic fluid infections.

Type of infection

PMN cell
count
(/mm3)

Spontaneous bacterial
peritonitis
Culture-negative neutrocytic
ascites
Monomicrobial nonneutrocytic
bacterascites
Secondary bacterial peritonitis
Polymicrobial bacterascites

250

Bacterial culture result

250

Positive (usually 1
organism)
Negative

250

Positive (1 organism)

250
250

Positive (polymicrobial)
Positive (polymicrobial)

NOTE. PMN polymorphonuclear.

moniae are the organisms isolated most frequently. Gram-positive cocci account for 25% of episodes; streptococcal species
are isolated most frequently. Although the flora of the colon
is predominantly anaerobic, isolation of an anaerobic organism
as the cause of SBP is an infrequent event, probably because of
the high oxygen content of the intestinal wall and surrounding
tissues and because of the relative inability of anaerobes to
translocate across the intestinal mucosa (see below). This pattern of bacterial prevalence may differ for patients who are
receiving selective intestinal decontamination (SID), usually
with fluorinated quinolones, to suppress the gram-negative intestinal flora and reduce the incidence of SBP. SID reduces the
number of episodes caused by gram-negative bacteria, but can
increase the frequency of gram-positive SBP episodes.

CID 1998;27 (October)

of ascites appears to be an important risk factor for the development of bacterial translocation in cirrhotic rats. Bacterial translocation has also been observed in humans at the time of laparotomy. In healthy individuals, bacteria that colonize lymph
nodes are killed by local immune defenses. However, in the
setting of cirrhosis, several forms of immune deficiency (see
figure 1 and below) favor the spread of bacteria to the bloodstream.
Alterations in the systemic immune system. Bacteria that
enter the bloodstream of a healthy host are rapidly coated by
IgG and/or complement components and then engulfed and
killed by circulating neutrophils [2]. However, in the setting
of cirrhosis, several abnormalities have been described in the
humoral and cellular bactericidal systems including decreased
serum levels of complement factors, impaired chemotaxis, poor
function and phagocytic activity of neutrophils, and decreased
function of Fc-g-receptors in macrophages.
Reticuloendothelial system phagocytic activity. The stationary macrophages, such as the Kupffer cells of the liver,

Pathogenesis
Figure 1 schematizes our current knowledge concerning the
pathogenesis of SBP [3, 6].
Intestinal bacterial overgrowth (IBO). Among cirrhotic patients, 30% to 48% have colonization of the upper bowel with
colonic bacteria; patients with more advanced liver disease
have higher rates of colonization. Bacterial translocation is the
process by which intestinal bacteria exit the intestinal lumen,
cross the intestinal wall, and colonize intestinal and/or mesenteric lymph nodes. IBO has been shown to be a prerequisite
for the development of bacterial translocation in experimental
animals (figure 1). Possible explanations for IBO among patients with cirrhosis include an altered local IgA immune response and delayed intestinal transit.
Intestinal permeability. Intestinal structural abnormalities
characterized by vascular congestion and edema, as well as an
increased interepithelial cell space, are evident in patients with
cirrhosis. These abnormalities probably increase intestinal permeability and facilitate bacterial translocation.
Bacterial translocation. Once intestinal bacteria translocate across the mucosa and escape the intestines, they can then
spread to other tissues including the bloodstream. The presence

/ 9C57$$OC16

09-15-98 15:47:30

Figure 1. Mechanisms that may be involved in the pathogenesis

of spontaneous bacterial peritonitis. AF ascitic fluid; BA bactericidal activity; CNNA culture-negative neutrocytic ascites; RES
reticuloendothelial system; SBP spontaneous bacterial peritonitis.

cida

UC: CID

CID 1998;27 (October)

Spontaneous Bacterial Peritonitis

Table 2. Predisposing factors for the development of AF infection

in patients with cirrhosis and ascites.
1.
2.
3.
4.
5.
6.
7.

Severity of liver disease: Child-Pugh class C patients

AF total protein level 1 g/dL and/or AF C3 level 13 mg/dL
Gastrointestinal bleeding
Urinary tract infection
Intestinal bacterial overgrowth
Iatrogenic factors: urinary bladder and intravascular catheters
Previous spontaneous bacterial peritonitis episode(s)

NOTE. AF ascitic fluid; C3 third component complement.

assist the circulating neutrophils in the extraction and killing of

particulate matter (e.g., bacteria) from the systemic circulation.
These cells, and perhaps others, comprise the reticuloendothelial system. The function of this essential bactericidal system
may be severely impaired in the setting of cirrhosis. Patients
with the most severe dysfunction of this system have the highest risk of bacteremia and concomitant shortened survival, due
to sepsis. The presence of intrahepatic and extrahepatic portosystemic shunts as a consequence of portal hypertension, prevent circulating bacteria from encountering Kupffer cells. The
final consequence of these abnormalities is the prolongation of
bacteremia and eventual seeding of other sites, including AF.
AF defense mechanisms. The arrival of bacteria to the AF
does not guarantee that infection will develop. In fact, cirrhotic
AF is capable of humoral self-defense, mainly on the basis
of the effectiveness of the complement system. Patients with
adequate activity of this vital bactericidal system usually do
not develop AF bacterial infections [3]. However, it has been
demonstrated that among those with an AF third component
(C3) of complement level of 13 mg/dL and/or a protein level
of 1 g/dL, there is a predisposition to this infection [3, 7].
The complement levels may be deficient because of increased
consumption of these components or because of impaired synthesis. Most of the bacteria that colonize AF are intestinal
gram-negative bacteria. The presence of lipopolysaccharides
in their cell wall activates the alternative pathway of complement. If the complement levels are adequate to effectively kill
the bacteria, infection will not develop. However, if complement levels are consumed and depleted, killing may be ineffective. It is of interest to note that SID with norfloxacin reduces
the gram-negative intestinal flora and presumably reduces the
colonization of AF, and the agent has been shown to increase
the C3 complement levels in AF [7]. In summary, the frequent
colonization of AF by bacteria decreases its antimicrobial ability and can eventually lead to the development of infection
(figure 1).
Predisposing Factors
Factors that predispose to AF infection are outlined in table
2. The severity of the liver disease is probably the most important factor. Almost 70% of patients who develop SBP are

/ 9C57$$OC16

09-15-98 15:47:30

671

Child-Pugh class C, with the remainder being class B. A serum

total bilirubin level of 2.5 mg/dL is an independent predictive
factor of SBP [6]. Deficient AF bactericidal activity is the main
intraperitoneal predisposing factor for the development of AF
infection. A direct correlation between total protein level, complement components, and opsonic activity explains why an AF
total protein level of 1 g/dL is a risk factor for the development of AF infection [3].
Among cirrhotic patients with acute gastrointestinal bleeding, 20% have AF infection at the time of admission to the
hospital, and another 30% to 40% may develop bacterial infections (including extraperitoneal infection) during hospitalization. In an experimental model of cirrhosis, hemorrhagic shock
has been shown to increase bacterial translocation and intestinal
permeability, and to decrease the effectiveness of the reticuloendothelial system.
Bacteriuria is common among patients with cirrhosis, particularly among females, and it appears to be a predisposing factor
for the development of SBP. Screening for and treatment of
urinary tract infections, even in the absence of symptoms, may
be helpful in preventing AF infections. Urinary bladder catheters should be avoided if possible in patients with cirrhosis.
Intravascular catheters are commonly used in patients with
cirrhosis, especially in the intensive care unit (ICU). Among
these patients, between 4% and 20% of bacteremic episodes
may be caused by intravascular catheters, and their use should
be minimized in this population.
Infections of AF that occur after episodes of variceal bleeding appear to be related more to the bleeding itself, rather than
to therapeutic or diagnostic endoscopic procedures. Prophylactic antibiotics have been shown to be unnecessary for these
procedures, unless the patient has other risk factors for endocarditis.
Finally, those patients who survive an episode of SBP are
at high risk of recurrence: 43% at 6 months, 69% at 1 year,
and 74% at 2 years.
Prevalence and the Clinical Picture
In the past, SBP was considered an infrequent complication
of cirrhosis, probably because of the infrequency of paracentesis (because of unfounded fear of complications), and the low
diagnostic efficacy of bacterial cultures (resulting from the insensitivity of older methods). In reality, AF infection is the
most frequent infectious complication among patients with cirrhosis and those with ascites, comprising 31% of all bacterial
infections.
The symptoms observed most frequently are fever (69%)
and abdominal pain (59%). Contrary to popular belief, a rigid
abdomen does not occur in patients with infected ascites, even
if there is free perforation of the intestine into the fluid. The
presence of large-volume ascites prevents contact of the visceral and parietal peritoneal surfaces that is sufficient to elicit
the spinal reflex that causes rigidity. Other signs and symptoms

cida

UC: CID

672

Such and Runyon

Table 3. When to perform a paracentesis.

New-onset ascites
At the time of each admission to the hospital
When deterioration of the clinical status or laboratory test values is
evident
When the patient develops an associated complication, such as
hepatic encephalopathy or gastrointestinal bleeding

include hepatic encephalopathy (54%), abdominal tenderness

(49%), diarrhea (32%), ileus (30%), shock (21%), and hypothermia (17%). Among patients with SBP, 10% have no signs
or symptoms. Because of this lack of specificity and sensitivity
of clinical signs and symptoms, instances of unexplained deterioration in patients with cirrhosis should lead to a diagnostic
paracentesis. Prompt diagnosis and treatment maximize survival among patients with AF infections.

Diagnosis
Suspicion of infection is based on the clinical setting. However, the diagnosis of AF infection is based on AF analysis, and
to obtain fluid an abdominal paracentesis must be performed.
Paracentesis has been shown to be safe despite the predictable
coagulopathy in these patients; there is an 1% chance of
significant abdominal-wall hematoma, .01% chance of hemoperitoneum, and .01% chance of iatrogenic infection related to
paracentesis. Indications for paracentesis are outlined in table
3. Paracentesis should be avoided only in instances of clinically
evident fibrinolysis or disseminated intravascular coagulation.
Table 4 details some of the diagnostic tests that can be
ordered on AF. A cell count and differential should be ordered
for every specimen, even when a therapeutic paracentesis is
performed. The diagnosis of SBP is suspected when the AF
PMN cell count reaches 250/mm3. Patients with neutrocytic
ascites (i.e., PMN count, 250/mm3) should receive prompt
empiric antibiotic treatment (see below), without waiting for
the results of the AF culture, given that SBP and CNNA share
common clinical, prognostic, and therapeutic characteristics,
and a delay in antibiotic treatment may result in a significant
and potentially fatal deterioration in the clinical status of the
patient.
Serum and AF albumin levels should be obtained for calculation of the serum-ascites albumin gradient [8]. A serum-ascites
albumin gradient of 1.1 g/dL is nearly 100% accurate in
detecting the presence of portal hypertension. This test need
be performed only on the first specimen from a given patient.
AF should be inoculated into blood-culture bottles at the
bedside [5]. The volume of fluid used for cultures varies according to the manufacturers specifications; however, application of a 10-mL inoculum into each bottle has been shown to
optimize results in standard 100-mL bottles. Use of bloodculture bottles yields bacterial growth in 80% of episodes of

/ 9C57$$OC16

09-15-98 15:47:30

CID 1998;27 (October)

neutrocytic ascites, compared to 50% for the conventional

technique [5].
The optional tests (table 4) are ordered when there is
suspicion for something other than sterile, cirrhotic ascites,
and the unusual tests are ordered only when the pretest
probability of peritoneal tuberculosis or carcinomatosis, etc.,
is high enough to justify their use. The utility of other tests,
such as those for pH, lactate dehydrogenase, cholesterol, fibronectin, a-1 antitrypsin, glycosaminoglycans, etc., has not been
proven, and these tests are not recommended.
Treatment
Many years ago, the usual treatment for patients with SBP
was the combination of a b-lactam plus an aminoglycoside.
However, patients with SBP are very sensitive to the nephrotoxicity associated with use of aminoglycosides, and even if
toxic levels are avoided, fatal renal failure is common. In 1985,
results of a randomized trial demonstrated that cefotaxime, a
third-generation cephalosporin, achieved cures in SBP episodes
for 85% of patients, compared to 56% of patients who received
ampicillin plus tobramycin [9]. More importantly, neither renal
impairment nor side effects were reported in association with
cefotaxime. Since then, cefotaxime has become the empiric
antibiotic of choice for the treatment of SBP. More recent
studies have demonstrated that dosing regimens and duration
of treatment can be reduced with continued excellent results.
An intravenous dosage of 2 g of cefotaxime every 8 hours for
5 days is enough, even in patients with bacteremia. The dosage
need not be altered for cases of hepatic or renal failure. Despite
many years of use, 72% to 96% of AF SBP isolates remain
susceptible to cefotaxime.
Other antibiotics have been tested in the treatment of SBP,
such as ceftriaxone, aztreonam, cefonicid and, recently, amoxicillin/clavulanate. Although some of them, e.g., ceftriaxone
or amoxicillin/clavulanate, seem promising, more studies are
needed before their use can be recommended. Parenteral amoxicillin/clavulanate is not available in the United States.
A novel approach for the treatment of SBP has been the use
of oral ofloxacin. A randomized trial compared the efficacy of
oral ofloxacin with that of intravenous cefotaxime in a selected
group of cirrhotic patients with SBP in the absence of septic
shock, hepatic encephalopthy, azotemia, gastrointestinal bleed-

Table 4. Tests performed for analysis of ascitic fluid.

Routine tests
WBC count and
differential
Albumin levels
Cultures in blood-culture
bottles

cida

Optional tests

Unusual tests

Total protein levels

Glucose levels
Lactate dehydrogenase
levels
Gram staining
Amylase levels

Tuberculosis smear
and culture
Cytology
Triglyceride levels
Bilirubin levels

UC: CID

CID 1998;27 (October)

Spontaneous Bacterial Peritonitis

ing, or ileus. The infection resolution rate was 84% in the

ofloxacin group and 85% in the cefotaxime group, with a hospital survival rate of 81% in both groups. The results of this study
point to the possibility of outpatient treatment of uncomplicated
SBP. However, other issues, such as patient compliance and
duration of therapy must be closely evaluated before this option
can become routine.
Repeated paracenteses for follow-up of patients with SBP
are considered. However, this procedure does not appear to be
necessary if the clinical response to treatment is dramatic, the
setting is typical, and the infection is monomicrobial. Repeated
paracentesis is recommended if the first AF analysis has the
following characteristics: (1) polymicrobial infection and (2)
two or more of these factors total protein level, 1 g/dL;
glucose level, 50 mg/dL; or lactate dehydrogenase level
greater than the upper limit of normal for serum. Patients with
these characteristics frequently have secondary bacterial peritonitis; these patients should undergo an emergency radiological
evaluation to determine a surgical source for their peritonitis
followed by surgical intervention when appropriate.
Bacterascites represents the colonization of AF with bacteria
without a neutrocytic response. It has been shown that the
outcome of bacterascites is different in patients who are symptomatic vs. those without signs or symptoms of infection. Patients with signs or symptoms of infection are more prone to
progress to SBP and therefore should receive empiric antibiotic
treatment as detailed above for SBP. Therefore, treatment
should be initiated if there is new abdominal pain and/or temperature 1007F. In general, there is no growth demonstrated
in cultures for 12 48 hours. Therefore, at time of the paracentesis it is not known whether the cultures will yield bacteria and
treatment is initiated on the basis of clinical judgement. If
neither AF nor blood nor urine cultures yield bacteria by 48
hours, the antibiotic can be discontinued. In contrast to patients
with symptomatic bacterascites, only 15% of asymptomatic
patients with bacterascites progress to SBP. Therefore, it is
recommended that the paracentesis be repeated as soon as the
first culture yields bacteria. Antibiotics are initiated only if
signs or symptoms of infection develop or if the second paracentesis demonstrates neutrocytic ascites.
Prophylaxis
Several groups of patients with cirrhosis who are at high
risk for SBP have been identified: those with gastrointestinal
bleeding, those with AF protein levels of 1 g/dL, and those
who have survived a previous episode of SBP [10 12]. Prophylactic measures that are directed at decreasing the risk of
infection have been studied in these groups of patients (table 5).
General Measures

Long-term measures include abstinence from alcohol, improvement in nutrition and the general status of the patient,

/ 9C57$$OC16

09-15-98 15:47:30

673

Table 5. Prevention of ascitic fluid infection.

Selective intestinal decontamination should be initiated in the
following groups of cirrhotic patients with ascites:
Patients with gastrointestinal bleeding: norfloxacin, 400 mg
b.i.d. per os or via nasogastric tube for 7 days
Patients with ascitic fluid protein levels of 1 g/dL:
norfloxacin, 400 mg q.d. per os during hospitalization
Patients who have recovered from a previous episode of
spontaneous bacterial peritonitis: norfloxacin, 400 mg q.d. per
os indefinitely

and aggressive treatment and eradication of localized infections

before dissemination occurs. Measures directed at reducing the
risk of gastrointestinal bleeding or the development of ascites,
such as surgical portacaval shunts or trans-jugular intrahepatic
portasystemic stent-shunts, may help prevent SBP. Diuretic
therapy decreases the AF volume and has been shown to significantly increase the AF opsonic activity, theoretically helping to prevent the development of SBP.

SID

Low-protein AF. Because most SBP episodes are caused

by intestinal gram-negative flora, attempts have been made to
reduce this bacterial population. Norfloxacin is a poorly absorbed quinolone that selectively inhibits the gram-negative
flora without affecting the anaerobic population. The presence
of anaerobes is required to maintain the stability of the intestinal flora and to prevent overgrowth of problem organisms. SID
by use of norfloxacin, 400 mg q.d., during hospitalization has
proven useful in reducing the incidence of SBP as well as the
incidence of extraperitoneal infections in patients with lowprotein AF.
Gastrointestinal bleeding. Cirrhotic patients with gastrointestinal bleeding are at increased risk for developing SBP.
Among those patients admitted to the hospital with gastrointestinal bleeding, 22% are already infected and 30% to 40% may
develop infection during hospitalization. Norfloxacin has been
shown to be valuable in reducing bacterial translocation due to
gram-negative bacteria in cirrhotic rats exposed to hemorrhagic
shock. Results of a randomized trial demonstrate that norfloxacin, 400 mg orally b.i.d. for 7 days, significantly reduced the
incidence of infectious episodes caused by gram-negative bacteria in cirrhotic patients with gastrointestinal bleeding. Other
therapeutic alternatives have been considered, such as amoxicillin/clavulanic acid, trimethoprim-sulfamethoxazole, and parenteral antibiotics.
Survivors of a previous episode of SBP. Results of a randomized trial demonstrate that the administration of oral norfloxacin, 400 mg per day, to survivors of a previous episode
of SBP significantly reduced the rate of recurrence from 68%
to 20%.

cida

UC: CID

674

Such and Runyon

Adverse Effects of SID

The development of quinolone-resistant strains of gram-negative bacteria has been observed in patients receiving norfloxacin for primary or secondary prophylaxis; however, the clinical consequences remain to be determined. In a recent study,
9 of 10 patients who had undergone continuous SID for 43
weeks developed bacterial infections due to quinolone-resistant
bacteria. Shorter periods of SID, however, do not appear to be
associated with this problem. Therefore, long-term SID must
be used with caution. Fortunately, no cross-resistance has been
observed between cefotaxime and the quinolones. Therefore,
the clinical outcome of bacterial infections caused by quinolone-resistant bacteria that are treated with cefotaxime is similar
to that of infections that are caused by susceptible bacteria.
Another advantage of SID is that it has been reported to reduce
health care costs, in part, by preventing the need for hospitalizations for treatment of SBP.
Prognosis

2.
3.
4.
5.
6.

7.

8.

9.

10.

During the early 1970s, the mortality associated with hospitalization for SBP reached 80% to 90%. Since that time, the
widespread use of paracentesis; the higher index of suspicion
of infection; and the clarification of diagnostic criteria, together
with use of better and safer antibiotics, has significantly improved the short-term prognosis of these patients. Currently,
there are essentially no deaths as a result of this infection,
provided it is detected and treated before the development of
shock or renal failure. Unfortunately, the long-term prognosis
remains extremely poor among survivors of an episode of SBP,
a manifestation of severe impairment of liver function. Probabilities of survival of 1 and 2 years are in the range of 30%
and 20%, respectively. Therefore, liver transplantation should
be considered for patients who survive an episode of SBP.
References
1. Runyon BA. Ascites and spontaneous bacterial peritonitis. In: Feldman
M, Scharschmidt BF, Sleisenger MH, eds. Sleisenger and Fordtrans

/ 9C57$$OC16

09-15-98 15:47:30

11.

12.

CID 1998;27 (October)

gastrointestinal and liver disease. Pathophysiology/diagnosis/treatment.

6th ed. Philadelphia: WB Saunders, 1998; 1310 33.
Guarner C, Soriano G. Spontaneous bacterial peritonitis. Semin Liver Dis
1997; 17:203 17.
Runyon BA. Patients with deficient ascitic fluid opsonic activity are predisposed to spontaneous bacterial peritonitis. Hepatology 1988; 8:632 5.
Runyon BA, Hoefs JC. Culture-negative neutrocytic ascites: a variant of
spontaneous bacterial peritonitis. Hepatology 1984; 4:1209 11.
Runyon BA, Canawati HN, Akriviadis EA. Optimization of ascitic fluid
culture technique. Gastroenterology 1988; 95:1351 5.
Andreu M, Sola R, Sitges-Serra A, et al. Risk factors for spontaneous
bacterial peritonitis in cirrhotic patients with ascites. Gastroenterology
1993; 104:1133 8.
Such J, Guarner C, Soriano G, et al. Selective intestinal decontamination
increases serum and ascitic fluid C3 levels in cirrhosis. Hepatology
1990; 12:1175 8.
Runyon BA, Montano AA, Akriviadis EA, et al. The serum-ascites albumin gradient is superior to the exudate-transudate concept in the differential diagnosis of ascites. Ann Intern Med 1992; 117:215 20.
Felisart J, Rimola A, Arroyo V, et al. Randomized comparative study
of efficacy and nephrotoxicity of ampicillin plus tobramycin versus
cefotaxime in cirrhotics with severe infections. Hepatology 1985; 5:
457 62.
Gines P, Rimola A, Planas R, et al. Norfloxacin prevents spontaneous
bacterial peritonitis recurrence in cirrhosis: results of a double-blind,
placebo-controlled trial. Hepatology 1990; 12:716 24.
Soriano G, Guarner C, Teixido M, et al. Selective intestinal decontamination prevents spontaneous bacterial peritonitis. Gastroenterology 1991;
100:477 81.
Soriano G, Guarner C, Tomas A, et al. Norfloxacin prevents bacterial
infection in cirrhotics with gastrointestinal hemorrhage. Gastroenterology 1992; 103:1267 72.

The Conflict-of-Interest Policy of the Office of Continuing Medical Education, UCLA School of Medicine,
requires that faculty participating in a CME activity disclose to the audience any relationship with a pharmaceutical or equipment company which might pose a potential, apparent, or real conflict of interest with regard to
their contribution to the program. The author reports no
conflict of interest.

cida

UC: CID