Spontaneous bacterial peritonitis in adults: Management

INTRODUCTION — 

Spontaneous bacterial peritonitis (SBP) is an ascitic fluid infection that occurs in the absence of an obvious source of infection (eg, perforated viscus, urinary tract infection). SBP primarily occurs in patients with cirrhosis and ascites but can occur in any patient who has developed ascites. It is usually due to translocation of bacteria from the intestinal lumen. Patients may present with signs and symptoms such as abdominal pain, fever, or altered mental status. However, some patients with SBP are asymptomatic and are diagnosed when they undergo paracentesis for another reason (eg, acute illness requiring hospital admission, therapeutic paracentesis).

The diagnosis of SBP is established by an ascitic fluid absolute neutrophil count ≥250 cells/microL, regardless of whether the patient has symptoms. We begin antibiotic therapy for patients with an ascitic neutrophil count ≥250 cells/microL while awaiting ascitic fluid culture results. However, an ascitic fluid cell count ≥250 cells/microL is not specific for SBP and could be related to secondary peritonitis. The clinical features and diagnosis of SBP are discussed separately. (See "Spontaneous bacterial peritonitis in adults: Pathogenesis, clinical features, and diagnosis".)

This topic will review the management of SBP.

The evaluation of patients with ascites is discussed separately. (See "Evaluation of adults with ascites".)

Management of fungal peritonitis is discussed separately. (See "Candida infections of the abdomen and thorax".)

MICROBIOLOGY — 

Most cases of SBP are caused by enteric bacteria (mainly Escherichia coli, Klebsiella pneumonia, Enterococcus faecalis, and Enterococcus faecium) [1-3]. Studies suggest that E. coli is isolated in over one-half of patients with SBP and positive ascitic fluid culture, and K. pneumonia is the second most common pathogen [4-8].

Reports of infection with Gram-positive bacteria such as Enterococcus spp and Staphylococcus spp have been increasing [9,10]. Multidrug-resistant (MDR) gram-negative pathogens are increasingly prevalent causes of SBP, especially in regions with high rates of these pathogens [3]. Patients with cirrhosis often have risk factors for colonization with MDR bacteria due to frequent health care and antibiotic exposure [1]. Patients with a history of prior SBP who are on prophylactic antibiotics (eg, ciprofloxacin) have increased risk for fluoroquinolone-resistant infections. (See 'Prevention' below.)

PRETREATMENT EVALUATION — 

For patients with SBP, the pretreatment evaluation serves to assess the severity of illness and the risk of infection with a multidrug-resistant organism (MDRO).

Determine severity of illness — In addition to clinical assessment, we use the chronic liver failure-sequential organ failure assessment (CLIF-SOFA) score to determine the severity of illness. The CLIF-SOFA score is similar to the sequential (sepsis-related) organ failure assessment (SOFA) score, a predictive scoring system designed to assess severity of illness in patients with sepsis [11]. (See "Predictive scoring systems in the intensive care unit", section on 'Sequential (sepsis-related) Organ Failure Assessment (SOFA)'.)

More severe illness is a risk factor for infection with an MDRO. (See 'Assess risk for multidrug-resistant organism (MDRO)' below.)

The CLIF-SOFA severity score is based on the following measurements of organ function (calculator 1) [11]:

●Respiratory system – Ratio of arterial oxygen tension to fraction of inspired oxygen (PaO₂/FiO₂)

●Cardiovascular system – Amount of vasoactive medication necessary to prevent hypotension

●Liver – Bilirubin level

●Coagulation system – International normalized ratio (INR)

●Neurologic system – Grade of hepatic encephalopathy (table 1)

●Kidney – Serum creatinine or use of kidney replacement therapy

The CLIF-SOFA score includes subscores ranging from zero to four for each of these six components. Thus, the total score ranges from zero to 24, with higher scores indicating more severe organ impairment. A CLIF-SOFA score ≥7 suggests more severe illness [11].

Assess risk for multidrug-resistant organism (MDRO) — Patients with any of the following risk factors are at higher risk for infection due to an MDRO [1,3,9,12-15]:

●Known colonization or prior infection by an MDRO or a gram-negative bacillus resistant to any third-generation cephalosporin

●Onset of infection while hospitalized for another reason (ie, nosocomial SBP)

●Recent hospitalization or use of a beta-lactam antibiotic (ie, within the prior three months)

●Frequent contact with the health care system

●Exposure to regions with a high prevalence of multidrug-resistant gram-negative pathogens (eg, Russia, eastern Europe, India, southeast Asia) (see "Carbapenem-resistant E. coli, K. pneumoniae, and other Enterobacterales (CRE)", section on 'Overall prevalence of CRE' and "Extended-spectrum beta-lactamases")

Studies suggest that infection with MDROs is common in patients with cirrhosis [1,14]. In a global study including over 1300 patients with cirrhosis and infections, the prevalence of multidrug-resistant (MDR) bacteria was 34 percent [1]. Risk factors for MDR bacteria included use of antibiotics within the previous three months, prior health care exposure, acquiring infection in Asia (particularly India), and infection site (ie, urinary tract, lung). Compared with non-MDRO infection, MDRO infections were associated with higher rates of in-hospital mortality, septic shock, organ failure, and persistent infection despite empiric antimicrobial therapy. In another study including 455 European patients with cirrhosis and acute-on-chronic liver failure, the prevalence of MDR bacterial infections was 29 percent [14]. In a study from the United States, 25 percent of cases of SBP were caused by organisms with resistance to third-generation cephalosporins [16].

Obtain MRSA and VRE swabs — For patients at high risk for resistant organism carriage, such as those with recent antibiotic use, intensive care unit (ICU) admission, or transplantation waitlisting, we obtain a nares sample for methicillin-resistant Staphylococcus aureus (MRSA) and a perirectal sample for vancomycin-resistant enterococci (VRE). The results can inform the antibiotic regimen.  

Prevention and control of these health care-associated infections are discussed separately:

●(See "Methicillin-resistant Staphylococcus aureus (MRSA) in adults: Prevention and control".)

●(See "Methicillin-resistant Staphylococcus aureus (MRSA): Microbiology and laboratory detection".)

●(See "Vancomycin-resistant enterococci: Epidemiology, prevention, and control".)

Assess local resistance patterns — The pretreatment assessment includes reviewing the hospital’s local antibiotic resistance patterns (eg, the hospital antibiogram) to select the most appropriate antimicrobial therapy.

GENERAL MEASURES — 

General measures (in addition to antimicrobial therapy) for all patients with SBP include albumin infusion and assessment of other medications. (See 'Antimicrobial therapy' below.)

Albumin infusion — We administer 25 percent albumin solution, intravenously, within six hours of diagnosing SBP by ascitic fluid cell count (1.5 g/kg body weight; maximum dose: 100 g) and again on day 3 (1 g/kg body weight; maximum dose: 100 g) [17]. For patients with suspected or confirmed volume overload, we do not give albumin or we lower the dose (ie, 25 percent albumin solution, 0.5 mg/kg body weight). Patients with or at risk for volume overload include those with hypoxia, pulmonary edema, end-stage kidney disease on dialysis, or severe cardiac dysfunction.  

Acute kidney injury (AKI) develops in up to 40 percent of patients with SBP and is a major cause of mortality [17-19]. Data from clinical trials suggest that albumin infusion prevents kidney impairment. In a meta-analysis comparing albumin infusion with no albumin (three trials) or artificial colloid (one trial) in a total of 288 patients with SBP, administration of albumin resulted in lower risk of kidney impairment (8 versus 31 percent, odds ratio [OR] 0.21, 95% CI 0.11-0.42) and lower risk of mortality (16 versus 35 percent, OR 0.34, 95% CI 0.19-0.60) [20].

AKI is associated with activation of the renin-angiotensin system and a decrease in effective arterial volume. Thus, it has been hypothesized that in patients with liver disease, plasma volume expansion could restore effective arterial blood volume and minimize the adverse hemodynamic changes in SBP, thereby preserving kidney function [20].

The management of acute kidney injury and hepatorenal syndrome are discussed separately. (See "Hepatorenal syndrome: Treatment and prognosis".)

Adjusting medications — We hold nonselective beta blockers in patients with SBP who are hypotensive (systolic blood pressure <90 mmHg) or who have acute kidney injury. We restart the nonselective beta blocker in patients whose symptoms improve, kidney function normalizes, and antibiotic therapy is completed.

We do not routinely discontinue beta blockers in all patients because the benefit of such an approach has not been established [21,22].  

ANTIMICROBIAL THERAPY

Indications and timing — The indications for and timing of antimicrobial therapy depend on the level of suspicion for SBP and severity of illness (see "Spontaneous bacterial peritonitis in adults: Pathogenesis, clinical features, and diagnosis", section on 'Clinical features'):

●Symptomatic patients – For patients with symptoms of SBP (eg, abdominal pain or tenderness, fever, altered mental status), empiric antimicrobial therapy is indicated, and the timing depends on the severity of illness:

•Critically ill patients – For critically ill patients (eg, those with septic shock, those with hemodynamic instability), we initiate empiric therapy immediately if ascitic fluid, blood, and urine samples cannot be obtained promptly; we then obtain samples as soon as possible thereafter. In patients with cirrhosis and septic shock, delaying therapy has been associated with a higher risk of mortality [23,24].

•Patients without critical illness – For stable patients with suspected SBP, we start empiric antimicrobial therapy immediately after ascitic fluid, blood, and urine samples have been obtained for culture and ascitic fluid has been sent for neutrophil count and chemistries [23,25]. We do not wait for laboratory results to return before beginning antibiotic therapy.

If the ascitic fluid cell count does not suggest SBP (ie, neutrophil count <250 cells/microL), we discontinue empiric antibiotics for SBP and follow-up culture results. (See 'Duration of antimicrobial therapy' below.)

●Asymptomatic patients – For asymptomatic patients, we begin empiric antibiotics if the ascitic neutrophil count is ≥250 cells/microL. Determining the ascitic fluid’s neutrophil count typically requires no more than several hours from the time of the paracentesis.

Up to one-third of patients with SBP are asymptomatic, and their infection is detected when they undergo diagnostic paracentesis for other conditions (eg, hospitalization for acute gastrointestinal bleeding or acute kidney injury).

In some patients, SBP is detected at the bacterascites stage (ie, bacteria are present in the ascitic fluid, but the ascitic fluid neutrophil count is <250 cells/microL). For asymptomatic patients with bacterascites, we repeat the paracentesis after 48 hours, and we initiate therapy if the ascitic fluid neutrophil count has risen to ≥250 cells/microL.

Further details regarding the diagnosis of SBP are found separately. (See "Spontaneous bacterial peritonitis in adults: Pathogenesis, clinical features, and diagnosis", section on 'Diagnosis'.)

Initial empiric regimens — Initial antibiotic selection regimens should include coverage of the most common causes of SBP, mainly E. coli, K. pneumonia, and other gram-negative enteric pathogens (algorithm 1). Broader coverage for multidrug-resistant gram-negative and/or gram-positive pathogens is indicated for patients with critical illness or risk factors for MDROs. (See 'Pretreatment evaluation' above.)  

Once culture and susceptibility results are available, antibiotic regimens should be tailored. (See 'Subsequent management (tailoring therapy)' below.)

Patients with critical illness or risk factors for MDR gram-negative bacilli — Patients with critical illness or known risk factors for MDROs are at increased risk for adverse outcomes, including mortality, particularly if initial antibiotic coverage is inadequate [1,3,12,23,26-29] (see 'Pretreatment evaluation' above):  

●Risk factors for carbapenem-resistant gram-negative bacilli (eg, CRE) – For patients with a history of colonization or prior infection by a carbapenem-resistant pathogen, we suggest consultation with a specialist in infectious diseases. Novel antibiotics are available that have improved outcomes for treating these infections (algorithm 1). (See "Carbapenem-resistant E. coli, K. pneumoniae, and other Enterobacterales (CRE)", section on 'Antibiotic selection' and "Antimicrobial approach to intra-abdominal infections in adults".)

●Critical illness or risk factors for MDR gram-negative bacilli (other than CRE) – For these patients, we generally favor treatment with meropenem (1 g intravenously every eight hours) or imipenem-cilastatin (500 mg intravenously every six hours) (algorithm 1). (See 'Pretreatment evaluation' above.)

We add coverage for MRSA or VRE, if indicated. (See 'Indications for coverage of MRSA or VRE' below.)

If the ascitic fluid cell count does not suggest SBP (ie, neutrophil count <250 cells/microL), we discontinue empiric antibiotics for SBP and follow-up culture results. (See 'Duration of antimicrobial therapy' below.)

Carbapenems provide empiric coverage for gram-negative pathogens that produce extended-spectrum beta-lactamases (ESBLs) or AmpC beta-lactamases. ESBLs are the most common cause of multidrug-resistance in E. coli and most Klebsiella species, and carbapenems are the only beta-lactams with reliable activity against ESBLs [3,30]. (See "Extended-spectrum beta-lactamases".)

Some experts suggest empiric piperacillin-tazobactam for non-critically ill patients with risk factors for MDR gram-negative pathogens [25]. Carbapenems and piperacillin-tazobactam have not been directly compared in SBP studies. However, piperacillin-tazobactam and all cephalosporins have been shown to have worse outcomes than carbapenems in patients with various types of ESBL infections, as discussed separately. In patients with SBP, carbapenem use has been associated with lower risk of mortality compared with third-generation cephalosporins [12,13,31]. (See "Extended-spectrum beta-lactamases".)

We avoid ertapenem for treatment of SBP due to lack of clinical outcome data and limited data that suggest variable penetration of ertapenem into peritoneal fluid [32].  

Patients without critical illness or risk of MDROs — For patients at lower risk for MDROs (eg, those with community-acquired infection) who are not critically ill, we initiate a third-generation cephalosporin for empiric therapy [25,33-36]. We typically use ceftriaxone (2 g intravenously every 24 hours) or cefotaxime (2 g intravenously every eight hours) [35,37-39].

We add coverage for MRSA or VRE, if indicated (algorithm 1). (See 'Indications for coverage of MRSA or VRE' below.)

For patients who have a lower risk of MDRO but who cannot take a cephalosporin, an alternative is ciprofloxacin, although it does not penetrate into ascitic fluid to the same extent as cefotaxime [33,40]. We use ciprofloxacin 400 mg intravenously twice daily for patients with normal kidney function. We do not use fluoroquinolones for patients who had been receiving a fluoroquinolone for SBP prophylaxis because such patients are at risk for infection with fluoroquinolone-resistant bacteria.

If the ascitic fluid cell count does not suggest SBP (ie, neutrophil count <250 cells/microL), we discontinue empiric antibiotics for SBP and follow-up culture results. (See 'Duration of antimicrobial therapy' below.)

Indications for coverage of MRSA or VRE — Empiric MRSA and/or VRE coverage is not necessary for all patients with SBP.

We provide empiric coverage for patients at highest risk for these pathogens (algorithm 1):

●MRSA coverage – We add vancomycin (table 2) to cover MRSA for patients with any of the following risk factors:

•Critical illness

•Nosocomial SBP

•Known colonization or prior infection by MRSA (see 'Obtain MRSA and VRE swabs' above)

●VRE coverage – For patients with known colonization or prior infection by VRE, we add daptomycin (4 to 6 mg/kg intravenously every 24 hours). Daptomycin provides empiric coverage for MRSA as well as VRE. (See "Vancomycin-resistant enterococci: Epidemiology, prevention, and control".)

Subsequent management (tailoring therapy) — Subsequent therapy is informed by the patient’s clinical response to initial antimicrobial therapy, microbiologic results, and in some patients, analysis of a second sample of ascitic fluid obtained after 48 to 72 hours of antimicrobial therapy.

Patients with good clinical response should be expected to have improved symptoms within 48 hours of initiation of antibiotic therapy. For patients with good clinical response and positive ascitic fluid cultures, we tailor the antibiotic regimen and complete the course of therapy. (See 'Duration of antimicrobial therapy' below.)

Patients with symptomatic improvement

Pathogen identified — For patients with good clinical response to empiric antimicrobial therapy and positive ascitic fluid culture, we tailor therapy based on the specific pathogen and susceptibility test results. Narrowing antimicrobial coverage prevents the emergence of resistant pathogens caused by the extended use of broad-spectrum antibiotics [39].

Repeat paracentesis is not necessary for patients who improve symptomatically and have a pathogen isolated from the initial ascitic fluid culture.

●Antibiotic regimens for enteric gram-negative bacilli – These bacteria are becoming increasingly resistant to common antibiotics.

•Oral options – Transitioning patients to oral therapy is an option if the isolate is reported as susceptible and the patient has adequate gastrointestinal absorption.

Fluoroquinolones are favored for oral therapy based on clinical data [40,41]. For example, a trial including 123 patients with SBP found no difference in infection resolution with oral ofloxacin compared with intravenous cefotaxime (84 versus 85 percent, respectively) [41]. Another trial demonstrated comparable outcomes with a short course of intravenous ciprofloxacin (200 mg every 12 hours for two days) followed by oral ciprofloxacin therapy (500 mg every 12 hours for five days) compared with intravenous therapy alone [40]. However, such studies supporting oral fluoroquinolones were performed during an era of lower antimicrobial resistance.

•Intravenous options – Beta-lactam antibiotics are often favored for intravenous therapy based on clinical trials and high drug levels achieved in ascitic fluid. For patients intolerant of beta-lactams, intravenous fluoroquinolones are an option if the isolate is susceptible.

In most cases, beta-lactam selection should be based on susceptibility results. However, for certain gram-negative pathogens, susceptibility results for beta-lactam agents can be misleading, as described below:

-ESBL-producing isolates – Carbapenems and non-beta-lactam antibiotics are the best options for these organisms, if susceptible. ESBL-producing isolates are often identified by resistance to any third-generation cephalosporin or by a positive polymerase chain reaction (PCR) test; they may be reported as susceptible to numerous beta-lactams but clinical failure often occurs if beta-lactams other than carbapenems are chosen.

Details regarding antibiotic selection for ESBL infections are discussed separately. (See "Extended-spectrum beta-lactamases".)

-Enterobacter cloacae complex, Klebsiella aerogenes (formerly Enterobacter aerogenes), Citrobacter freundii, and other AmpC-producers – AmpC is a resistance mechanism that often cannot be detected on standard susceptibility testing or by PCR. The three listed bacteria are most affected, but other specific gram-negative bacilli are also known to be AmpC-producers, as discussed separately. (See "Extended-spectrum beta-lactamases".)

Cefepime, carbapenems, and non-beta-lactam antibiotics remain effective for these organisms, if susceptible. Beta-lactams other than cefepime and carbapenems are typically ineffective against these organisms even if they are reported as susceptible.  

-Carbapenem-resistant Enterobacterales (CRE) – These are among the most difficult-to-treat organisms, and consultation with an infectious diseases specialist is suggested. An isolate is identified as CRE if it is resistant to at least one carbapenem. (See "Carbapenem-resistant E. coli, K. pneumoniae, and other Enterobacterales (CRE)", section on 'Detection of CRE and carbapenemases'.)

These isolates are resistant to all beta-lactams, even if they are reported as susceptible. Non-beta-lactam agents are effective if the isolate is reported as susceptible, but many CRE isolates are resistant to these agents as well. (See "Carbapenem-resistant E. coli, K. pneumoniae, and other Enterobacterales (CRE)", section on 'Approach to treatment'.)

●Antibiotic regimens for gram-positive cocci

•S. aureus – Patients with cultures demonstrating methicillin-resistant S. aureus can complete treatment with vancomycin or daptomycin; patients with methicillin-sensitive S. aureus can be treated with a first-generation cephalosporin, such as cefazolin.

Patients with S. aureus bacteremia require longer durations of therapy, as described separately. (See "Clinical approach to Staphylococcus aureus bacteremia in adults".)

•Enterococcus species – Treatment of infections due to Enterococcus species is based on susceptibility results. Use of ampicillin or vancomycin is effective for susceptible isolates; for VRE, daptomycin is appropriate. (See "Treatment of enterococcal infections".)

No pathogen identified — For patients with symptomatic improvement but negative ascitic fluid cultures, we repeat the paracentesis (typically after 48 to 72 hours of initiating antimicrobial therapy) to further assess response. Patients with initial ascitic neutrophil count ≥250 cells/microL and negative ascitic fluid cultures have culture-negative neutrocytic ascites.

If repeat paracentesis reveals expected improvements (ie, a decrease in the ascitic neutrophil count by ≥25 percent), we complete the five-day course of therapy with the empiric antibiotic regimen. (See 'Duration of antimicrobial therapy' below.)

If repeat paracentesis shows minimal or no improvement (ascitic neutrophil count decreases by <25 percent), we obtain abdominal imaging (ie, contrast-enhanced CT scan of the abdomen and pelvis) to evaluate for secondary causes of peritonitis. (See "Spontaneous bacterial peritonitis in adults: Pathogenesis, clinical features, and diagnosis", section on 'Distinguishing SBP from secondary bacterial peritonitis'.)

Patients without symptomatic improvement — For patients who do not improve within 48 hours of initiating empiric antimicrobial therapy, we repeat the diagnostic paracentesis, assess the initial ascitic fluid culture results (if available) and broaden antimicrobial therapy to cover MDRO, if not done previously. (See 'Patients with critical illness or risk factors for MDR gram-negative bacilli' above.)

If the initial culture grows multiple organisms (especially if Enterococcus species or fungal organisms are present), secondary peritonitis related to an intra-abdominal process should be highly suspected. We also typically obtain a contrast-enhanced CT scan of the abdomen and pelvis to evaluate for secondary causes of peritonitis. (See "Spontaneous bacterial peritonitis in adults: Pathogenesis, clinical features, and diagnosis", section on 'Subsequent testing for selected patients'.)

Duration of antimicrobial therapy — For patients with SBP, the duration of antimicrobial therapy is informed by clinical response and analysis of ascitic fluid testing (see 'Subsequent management (tailoring therapy)' above):

●Patients with clinical response – We discontinue antibiotic therapy after five days for patients who have symptomatic improvement and expected findings on follow-up paracentesis (if performed). (See 'Patients with symptomatic improvement' above.)

Data from randomized clinical trials support this duration of treatment. In a trial comparing a five-day course of antimicrobial therapy with a 10-day course in 90 patients with SBP, there were no significant differences in rates of infection-related mortality, bacteriologic cure, or recurrent infection between groups [42].

●Patients without improvement in symptoms or ascitic fluid neutrophil count - If clinical response or follow-up ascitic fluid counts are suboptimal, further evaluation should be performed and antibiotics should be continued and possibly broadened. (See 'Patients without symptomatic improvement' above.)

For symptomatic patients who were started on empiric therapy prior to performing paracentesis, we discontinue antibiotics if the ascitic neutrophil count is <250 cells/microL, as discussed above. (See 'Indications and timing' above.)

PREVENTION

General preventive strategies — General measures to lower the risk of SBP include avoiding selected medications such as proton pump inhibitors when possible. (See "Proton pump inhibitors: Overview of use and adverse effects in the treatment of acid related disorders".)

Preventing SBP recurrence — We typically use long-term antibiotic prophylaxis in patients with cirrhosis who have recovered from an episode of SBP [43]. We tailor antimicrobial prophylaxis to the sensitivities of the specific organism that caused the prior episode(s) of SBP and consider local antibiotic resistance patterns when selecting the regimen. We consult with a specialist in infectious disease to inform decision-making, if necessary.

We generally choose from the following regimens in patients with normal kidney function:

●Ciprofloxacin (500 mg orally once daily)

●Trimethoprim-sulfamethoxazole (one double-strength tablet orally once daily).

Based on clinical experience, the benefits of secondary prophylaxis may be greater in patients whose underlying liver disease is expected to improve. Such patients include those receiving therapy for underlying liver disease (eg, antiviral therapy), those awaiting liver transplantation, or those undergoing transjugular intrahepatic portosystemic shunt (TIPS).

The use of antibiotic prophylaxis to lower the risk of recurrent SBP has been supported by society guidelines, although data are mixed [44-47]. In a trial including 80 patients with cirrhosis who had recovered from an episode of SBP, norfloxacin prophylaxis resulted in lower rates of SBP after 12 months (12 versus 35 percent) [44]. Of note, norfloxacin is not available in some countries, including the United States. In these areas, ciprofloxacin is a reasonable alternative. However, in a retrospective study including two cohorts of patients with cirrhosis who survived an episode of SBP, antibiotic prophylaxis was associated with a higher risk of developing SBP after two years compared with no prophylaxis (Veterans Administration Corporate Data Warehouse [n = 4673 patients]: 24 versus 14 percent, and TriNetX database [n = 6708 patients]: 23 versus 16 percent) [47].

Fluoroquinolones are generally safe and well tolerated, but rare, serious adverse events (eg, neuropathy, tendinopathy including tendon rupture) have been reported and are discussed separately. (See "Fluoroquinolones".)

Limited role for primary prophylaxis — We reserve primary prophylaxis for patients with cirrhosis and acute gastrointestinal bleeding because data from randomized clinical trials suggested a lower risk of mortality with antibiotic prophylaxis for such patients [48,49]. Antibiotic prophylaxis for patients with variceal bleeding is discussed separately. (See "Overview of the management of patients with variceal bleeding", section on 'Antibiotic prophylaxis'.)

The benefits of routinely using antibiotics for primary prophylaxis are uncertain because of mixed data, increasing rates of infection with MDROs, and increased risk of mortality related to the severity of underlying liver disease [50-54]. In a meta-analysis of 23 trials including 2587 patients with cirrhosis, antibiotic prophylaxis (various regimens) did not result in a survival benefit or lower risk of adverse events, although potentially important differences could not be excluded given the very low certainty of evidence [54]. In a trial comparing norfloxacin prophylaxis with placebo in 291 patients with decompensated cirrhosis, the risk of mortality after six months was not significantly different between groups (14.8 versus 19.7 percent, hazard ratio [HR] 0.69, 95% CI 0.38-1.23) [50]. Additional guidance on primary prophylaxis will be informed by an ongoing randomized trial of trimethoprim-sulfamethoxazole (cotrimoxazole) in patients with cirrhosis and ascites (ASEPTIC) [51].

PROGNOSIS — 

The infection-related mortality from SBP is low with appropriate treatment [18,24,55]. Several reports have found no infection-related deaths if therapy was started prior to developing septic shock or acute kidney injury [41,42]. In one systematic review, risk factors for in-hospital mortality included acute kidney injury and higher Model for End-stage Liver Disease (MELD) scores [18].

For patients who develop septic shock, mortality is high, but early initiation of appropriate antimicrobial therapy has been associated with improved outcomes. In a retrospective study of 126 patients with cirrhosis and SBP-associated septic shock, the overall in-hospital mortality was 82 percent [24]. Survivors received antimicrobial therapy earlier than those who died (median delay 1.8 versus 9.5 hours). The adjusted odds ratio for mortality was 1.9 for every hour delay in administering antimicrobial therapy (95% CI 1.1-3.1). (See 'Antimicrobial therapy' above.)

Patients with decompensated cirrhosis who develop SBP have a limited long-term prognosis [18,55-57]. In a large, nationwide database study of patients with cirrhosis, the three-year mortality rate for patients following hospitalization for SBP was 67 percent [55]. Similarly, in a cohort study including 178 patients with SBP, the mortality rate at 32 months was 76 percent [57]. Thus, liver transplantation should be considered for patients who recover from SBP and who are otherwise good transplantation candidates. (See "Liver transplantation in adults: Patient selection and pretransplantation evaluation".)

SOCIETY GUIDELINE LINKS — 

Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Portal hypertension and ascites".)

SUMMARY AND RECOMMENDATIONS

●Pretreatment evaluation – For patients with suspected spontaneous bacterial peritonitis (SBP), the pretreatment evaluation includes (see 'Pretreatment evaluation' above):

•Determine the severity of illness – We assess the severity of illness clinically and with the chronic liver failure-sequential organ failure assessment (CLIF-SOFA) severity score (calculator 1). Higher scores indicate more severe organ impairment. A CLIF-SOFA score ≥7 suggests more severe illness.

•Assess risk of infection with MDRO – Patients with recent hospitalization, frequent contact with the health care system, or recent antibiotic use (ie, within the previous three months) are at risk for multidrug-resistant organisms (MDROs).

•Obtain methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant enterococci (VRE) swabs because they inform the antibiotic regimen. Swabs may be most helpful for patients at high risk for MRSA and VRE colonization (eg, recent antibiotic exposure, ICU admission, on liver transplantation waiting list). (See "Methicillin-resistant Staphylococcus aureus (MRSA) in adults: Prevention and control" and "Vancomycin-resistant enterococci: Epidemiology, prevention, and control".)

•Assess local resistance patterns – We consult with the microbiology laboratory (eg, local antibiogram) to select the most appropriate antimicrobial therapy according to local resistance patterns.

●General measures – General measures (in addition to antimicrobial therapy) include (see 'General measures' above):

•Albumin infusion – We recommend administering albumin solution to most patients with SBP to reduce the risk of acute kidney injury (Grade 1B). We administer 25 percent albumin solution, intravenously, within six hours of diagnosis of SBP (1.5 g/kg body weight; maximum dose: 100 g) and again on day 3 (1 g/kg body weight; maximum dose: 100 g).

For patients with suspected or confirmed volume overload, we do not give albumin or we lower the dose (ie, 25 percent albumin solution, 0.5 mg/kg body weight).

•Adjusting medications – We typically hold nonselective beta blockers in selected patients with SBP (eg, those who are hypotensive [systolic blood pressure <90 mmHg], those with acute kidney injury).

●Empiric antimicrobial therapy – For patients with suspected SBP (eg, those with abdominal pain or tenderness, fever, altered mental status), empiric antimicrobial therapy is indicated, and the timing and selection of therapy depends on the severity of illness (algorithm 1). (See 'Antimicrobial therapy' above.)

•Timing of initiating therapy:

-For critically ill patients, we initiate empiric therapy immediately and obtain samples as soon as possible thereafter, if ascitic fluid, blood, and urine samples cannot be obtained promptly.

-For symptomatic patients without critical illness, we start antimicrobial therapy immediately after ascitic fluid, blood, and urine samples have been obtained.

-For asymptomatic patients, we begin empiric antibiotics if the ascitic fluid neutrophil count is ≥250 cells/microL.

•Selecting empiric therapy (algorithm 1):

-For patients with critical illness or at higher risk for MDROs, we suggest empiric therapy with a carbapenem (eg, meropenem 1g intravenously every eight hours) rather than other antibiotic regimens (Grade 2C). (See "Antimicrobial approach to intra-abdominal infections in adults".)

-For patients who are not critically ill and have a lower risk for MRDOs (eg, those with community-acquired infection), we suggest empiric therapy with a third-generation cephalosporin rather than other antibiotic regimens (Grade 2C). Options include ceftriaxone (2 g intravenously every 24 hours) or cefotaxime (2 g intravenously every eight hours).

For patients at higher risk for MRSA or VRE, we suggest empiric coverage with vancomycin (table 2) to cover MRSA or with daptomycin (4 to 6 mg/kg intravenously every 24 hours) to cover VRE (Grade 2C). (See 'Indications for coverage of MRSA or VRE' above.)

If the ascitic fluid cell count does not suggest SBP (ie, neutrophil count <250 cells/microL), we discontinue empiric antibiotics for SBP and follow-up culture results. (See 'Duration of antimicrobial therapy' above.)

●Tailoring therapy – Subsequent management is informed by the clinical course and the ascitic fluid culture results (see 'Subsequent management (tailoring therapy)' above):

•For patients with symptomatic improvement and a known pathogen, we narrow antimicrobial therapy when sensitivity data are available. We discontinue antimicrobial therapy after five days.

•For patients with symptomatic improvement and no known pathogen, we continue initial antimicrobial therapy and repeat the paracentesis. If the ascitic fluid neutrophil count does not decrease by at least 25 percent from the baseline neutrophil count, we evaluate for secondary causes of peritonitis with abdominal imaging. (See 'No pathogen identified' above.)

If the ascitic fluid neutrophil count decreases by ≥25 percent, we discontinue antimicrobial therapy after five days.  

•For patients without symptomatic improvement within 48 hours despite empiric antimicrobial therapy, we repeat the diagnostic paracentesis, compare the ascitic fluid neutrophil count to the pretreatment value, and assess the initial ascitic fluid culture results (if available). If initial coverage did not include MDRO, we broaden antibiotic coverage. (See 'Patients without symptomatic improvement' above and 'Initial empiric regimens' above.)  

●Prevention – For patients with cirrhosis who have recovered from an episode of SBP, we suggest long-term antibiotic prophylaxis rather than no prophylaxis (Grade 2C). We tailor the choice of antimicrobial prophylaxis to the sensitivities of the specific organism that caused the episode(s) of SBP. We consult with a specialist in infectious disease to inform decision-making by including local infection and antibiotic resistance patterns when selecting antimicrobial therapy. (See 'Prevention' above.)

We reserve primary prophylaxis for patients with cirrhosis and acute gastrointestinal bleeding because data from randomized clinical trials suggested a lower risk of mortality with antibiotic prophylaxis for such patients. (See "Overview of the management of patients with variceal bleeding", section on 'Antibiotic prophylaxis'.)

ACKNOWLEDGMENT — 

The UpToDate editorial staff acknowledges Bruce Runyon, MD, who contributed to earlier versions of this topic review.